Feeding Form,. Function, and Evolution in Tetrapod Vertebrates
This Page Intentionally Left Blank
Feeding Form, Function, and Evolution in Tetrapod Vertebrates Edited by
Kurt Schwenk Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut
:#) San Diego
San Francisco
ACADEMIC PRESS A Harcourt Science and Technology Company
New York
Boston
London
Sydney
Tokyo
Cover photographs: (Inset) Harry W. Greene, © 1999. (Background) Nirvana Filoramo, © 2000. This book is printed on acid-free paper.
^^
Copyright © 2000 by ACADEMIC PRESS All lights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Requests for permission to make copies of any part of the work should be mailed to: Permissions Department, Harcourt Inc., 6277 Sea Harbor Drive, Orlando, Florida 32887-6777
Academic Press A Harcourt Science and Technology Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http: / / www.academicpress.com
Academic Press Harcourt Place, 32 Jamestown Road, London NWl 7BY, UK http: / /www.academicpress.com Library of Congress Catalog Card Number: 99-63490 International Standard Book Number: 0-12-632590-1 PRINTED IN THE UNITED STATES OF AMERICA 00 01 02 03 04 05 EB 9 8 7 6 5
4
3
2
1
To my teachers: Warren F. Walker, Jr., James R. Stewart, and Marvalee H. Wake
This Page Intentionally Left Blank
Contents
Contributors Preface xiii
IV. Kinematics of Feeding: Feeding Stages V. Concluding Remarks 55 References 55
xi
49
S E C T I O N S E C T I O N
I
n
INTRODUCTION
AMPHIBIA
C H A P T E R C H A P T E R
Tetrapod Feeding in the Context of Vertebrate Morphology
Aquatic Feeding in Salamanders STEVEN M. DEBAN AND DAVID B. WAKE
KURT SCHWENK
I. II. III. IV. V.
I. Introduction 3 11. Approaches to the Study of Tetrapod Feeding 5 III. Concluding Comments 16 References 16
C H A P T E R
Introduction 65 Morphology 68 Function 82 Diversity and Evolution 88 Opportunities for Future Research References 92
92
C H A P T E R
An Introduction to Tetrapod Feeding
Terrestrial Feeding in Salamanders
KURT SCHWENK
DAVID B. WAKE AND STEPHEN M. DEBAN
I. Introduction 21 II. Morphology of the Feeding Apparatus III. Kinematics of Feeding:The Gape Cycle
I. Introduction 95 II. Morphology 97 III. Function 101
26 47
Vll
Contents
Vlll
IV. Diversity and Evolution 111 V. Opportunities for Further Research References 114
S E C T I O N
114
IV REPTILIA: LEPIDOSAURIA
C H A P T E R
C H A P T E R
Feeding in Frogs
8
KIISA C. NISHIKAWA
I. II. III. IV. V. VI. VII.
Introduction 117 Morphology of the Feeding Apparatus 119 Function of the Feeding Apparatus 124 Neural Control of Prey Capture 135 Evolution of the Feeding Apparatus 139 Conclusions 143 Current and Future Directions 144 References 144
C H A P T E R
Feeding in Lepidosaurs KURT SCHWENK
I. Introduction 175 II. Lepidosaurian Phylogeny and Classification 176 III. Natural History 178 IV. Morphology of the Feeding Apparatus 189 V. Feeding Function 220 VI. Specialized Feeding Systems 257 VII. Evolution of Feeding in Lepidosaurs 264 VIII. Future Directions 277 References 278
C H A P T E R
Feeding in Caecilians JAMES C. O'REILLY
I. 11. III. IV. V
Feeding in Snakes
Introduction 149 Morphology 150 Function 155 Evolution 161 The Future 163 References 164
DAVID CUNDALL AND HARRY W. GREENE
I. II. III. IV. V.
Introduction 293 Form and Function 301 Performance and Size 322 Evolution 322 Concluding Remarks 326 References 327
S E C T I O N
III REPTILIA: TESTUDINES
S E C T I O N
V REPTILIA: ARCHOSAURIA
C H A P T E R C H A P T E R
A Bibliography of Turtle Feeding KURT SCHWENK
I. Introduction II. Bibliography
169 169
10
Feeding in Crocodilians
JOHAN CLEUREN AND FRITS DE VREE
I. Introduction II. Morphology
337 340
Contents III. Function 347 IV. Evolution 354 References 357
IX
V. The Feeding Apparatus 421 VI. Feeding Function 439 VII. Control of Feeding Behaviors 444 References 444
C H A P T E R C H A P T E R
11 Feeding in Paleognathous Birds CAROLE A. BONGA TOMLINSON
14 The Ontogeny of Feeding in Mammals R. Z. GERMAN AND A. W. CROMPTON
I. 11. III. IV. V
Introduction 359 Materials and Methods 360 Morphology of the Hyolingual Apparatus 361 Function of the Hyolingual Apparatus 373 Evolution of the Feeding System 384 References 390
C H A P T E R
"12
I. II. III. IV. V.
Introduction 449 Morphology 449 Function and Mechanics of Suckling 450 Rhythmicity and Control of Suckling 453 Coordination of Swallowing and Respiration 455 VI. Transition from Suckling to Drinking at Weaning 455 VII. Evolutionary Considerations 456 References 456
Feeding in Birds: Approaches and Opportunities
C H A P T E R
MARGARET RUBEGA
I. Introduction 395 II. Patterns of Analysis III. Conclusion 406 References 406
396
S E C T I O N
VI MAMMALIA
15 Feeding in Myrmecophagous Mammals KAREN ZICHREISS
I. II. III. IV. V.
Introduction 459 Foraging Ecology 462 Morphology of the Feeding Apparatus 464 Functional Morphology 475 Evolution of Myrmecophagous Specializations 478 VI. Directions for Future Research 480 References 481
C H A P T E R
C H A P T E R
13
16
Feeding in Mammals
Feeding in Marine Mammals
KAREN M.HIIEMAE
ALEXANDER WERTH
I. Introduction 411 II. Mammalian Feeding System 414 III. The "Process Model" for Mammalian Feeding 416 IV. Mechanical Properties and Textures of Foods 419
I. Introduction 487 II. Feeding Strategies 492 III. Conclusions 521 References 521 Index
527
This Page Intentionally Left Blank
Contributors
Kiisa C. Nishikawa (117) Department of Biological Sciences, Northern Arizona University, Flagstaff, Arizona 86011. James C. O'Reilly (149) Program in Organismic and Evolutionary Biology, University of Massachusetts, Amherst, Massachusetts 01003. Karen Zich Reiss (459) Department of Biological Sciences, Humbolt State University, Areata, California 95521. Margaret Rubega (395) Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, Connecticut 06269. Kurt Schwenk (3, 21,169,175) Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, Connecticut 06269. Carole A. Bonga Tomlinson (359) Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138. David B. Wake (65,95) Museum of Vertebrate Zoology, University of California, Berkeley, California 94720. Alexander Werth (487) Department of Biology, Hampden-Sydney College, Farmville, Virginia 23901.
Numbers in parentheses indicate the pages on which the authors' contributions begin.
Johan Cleuren (337) Department of Biology, University of Antwerp, B-2610 Antwerp, Belgium. A. W. Crompton (449) Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138. David Cundall (293) Department of Biological Sciences, Lehigh University, Bethlehem, Pennsylvania 18015. Frits De Vree (337) Department of Biology, University of Antwerp, B-2610 Antwerp, Belgium. Stephen M. Deban (65, 95) Museum of Vertebrate Zoology, University of California, Berkeley, California 94720. Rebecca Z. German (449) Department of Biological Sciences, University of Cincinnati, Cincinnati, Ohio 45221. Harry W. Greene (293) Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, New York 14853. Karen M. Hiiemae (411) Department of Bioengineering and Neuroscience, Institute for Sensory Research, Syracuse University, Syracuse, New York 13244.
XI
This Page Intentionally Left Blank
Preface
This book addresses the first of these challenges. It examines in depth and breadth the myriad solutions to the essential problem of feeding in one clade of animals. It summarizes and synthesizes for the first time in 15 years our burgeoning knowledge of tetrapod feeding systems and how they have evolved. It explores the "variations on a theme" in this system and in so doing provides grist for evolutionary theorists. However, its proximate goals are more modest. The book is intended to instruct novice morphologists and others interested in animal form and function. It is both an introduction to the field and a presentation of the "state of the art." It is aimed at advanced undergraduate and gradutate students, as well as experts in the field who wish to delve outside their taxonomic bounds. It provides an accessible entree into an exploding literature and showcases our impressive knowledge—but it highlights also our great ignorance. As Rubega points out in Chapter 12, there are many dissertations on tetrapod feeding left to be written. The greatest possible outcome I can imagine for this book is that it will stimulate and provoke the next generation of morphologists to fill in the gaps and shoot down the dogma. In an effort to promote its utility to students, the book begins with two introductory chapters that establish the conceptual, historical, and factual contexts within which the empirical chapters can be interpreted. The empirical chapters provide a more-or-less phylogenetic survey of tetrapod vertebrate feeding systems. Although a phylogenetic approach is emphasized throughout, there are some cases in which I judged other criteria to be more useful in organizing current knowledge. Hence, some chapters are not limited to a monophyletic taxon, but are based on functional types (e.g., "marine mammals," Chapter 16), dietary types (e.g., "myrmecophagous mammals," Chapter 15), or the medium in which feeding occurs (e.g., "aquatic
Vertebrate morphology stands accused of failing to contribute meaningfully to the neo-Darwinian evolutionary synthesis. Although I strongly doubt this (see Chapter 1), I nonetheless take it as a challenge for the future. Our evolutionary theory is at a cusp—the power and efficacy of reductionism are undeniable, but equally so is its failure to deal effectively with intrinsic, organismal attributes. Despite leaps and bounds in our understanding of genetic- and population-level evolutionary phenomena, we remain almost embarrassingly ignorant about the fundaments of phenotypic evolution. Answers to the most basic questions are beyond our grasp: Why do some lineages evolve rapidly while others remain static? How do complex systems full of interacting characters evolve? And once evolved, how can they change? Our approaches to these questions are often simplistic and too facile. We point to "phylogenetic constraint," for example, as if it explains the failure of a lineage to evolve in some expected way when, in fact, it does little more than describe our ignorance. What is the mechanistic basis of phylogenetic constraint? Is it all just genetic background, or are there phenotypic processes that interfere with diversification and adaptive evolution or that facilitate it? Organisms are multihierarchical, complex systems, and as in other such systems, each level expresses emergent properties that are unpredictable, even unknowable, from the vantage point of other levels. If we want to understand the principles governing the evolution of phenotypes, it is logical, indeed, necessary to study the phenotype directly. Who better to do this than morphologists? The challenge to vertebrate morphology in the next century is therefore twofold: to develop the empirical database and conceptual tools needed to create a phenotypebased evolutionary theory, and to forge a new evolutionary synthesis by integrating this theory with the gene-based, neo-Darwinian paradigm.
Xlll
XIV
Preface
feeding in salamanders/' Chapter 3). Each chapter is authored by an expert or experts on the group, including both veteran and younger workers. I am very pleased to be able to include chapters on little-known groups, such as caecilian amphibians (Chapter 6), crocodilians (Chapter 10), paleognathous birds (Chapter 11), myrmecophagous mammals (Chapter 15), and marine mammals (Chapter 16). However, my goal of complete taxonomic coverage of all tetrapods was not quite achieved. Owing to many factors, a chapter on turtle feeding could not be completed. To mitigate this taxonomic breach, I have prepared a brief bibliography of turtle feeding to serve as an entree into the literature (Chapter 7). This book has a long and tortured history—even longer and more tortured than most edited works! It was inspired by a symposium on the ecology and evolution of feeding systems in lower vertebrates presented at the annual meeting of the American Society of Ichthyologists and Herpetologists in Austin, Texas, in 1993, to which I was a contributor. The symposium was organized by Drs. Peter Wainwright and Kiisa Nishikawa. Dr. Charles Crumly, a systematic herpetologist and editor at Academic Press, was in attendance. I had been toying with the idea of editing a book in the area of feeding, so when Chuck approached me with the idea I was thrilled to take it on. By the end of the meeting several authors were already lined up. That was the easy part! The project ebbed and flowed over the years as the author roster grew and shifted. Consequently, there is a large span of time over which chapters were completed and submitted. Although I have tried to update the literature where necessary, some chapters are inevitably not as current as others. Thus, authors who worked most dilligently to complete their manuscripts in time for early deadlines should not be held to blame for editorial shortcomings. There are many people to thank for their contributions, direct and indirect, to this project. I must begin by expressing my deep gratitude for the inspiration of my teachers, Warren R Walker, Jr., James R. Stewart, and Marvalee H. Wake, to whom this book is dedicated. Warren Walker first taught me vertebrate biology and comparative anatomy as a junior at Oberlin College and it was his deep knowledge and masterful teaching that led me to embark on a career in vertebrate morphology. I remain in awe of his knowledge of comparative anatomy; his course serves as the benchmark from which I measure my own feeble attempts. Warren had the poor taste to take a sabbatical leave my senior year at Oberlin, but this sad event (for me) had a positive side—James Stewart was hired to replace him that year. Jim came to Oberlin fresh out of Berkeley with a
new set of experiences and ideas. I watched firsthand as he put together his own terrific course on comparative anatomy and I was given the opportunity to assist teaching in the lab. Jim supervised my senior thesis research (on lizard feeding!) and became a friend as well as a mentor. His calm, philosophical, and scholarly approach to both life and science deeply impressed me and continues to inspire me now. At Oberlin, Jim regaled me with stories of Berkeley, the Museum of Vertebrate Zoology, and the "Herp Lab,'' so after a short hiatus as a zookeeper at the Bronx Zoo, I was thrilled to be accepted into Marvalee Wake's lab at Berkeley for graduate study. Marvalee, to me, is the quintessential vertebrate morphologist—painstaking, detailed, thorough, and a scholar of the highest order. She was also the perfect advisor. She knew unerringly when to leave me on my own and when to push me. She supported my work and my psyche. Most important, she set a high standard in the lab and maintained it by example. The depth and breadth of her work on caecilian amphibians are a model of achievement and a personal source of inspiration. I have depended on Marvalee's wisdom for the last 20 years and still turn to her when I am in need of counsel. I am profoundly grateful to each of these people who have contributed so critically to my professional, intellectual, and personal development—often in ways they cannot imagine. Whatever strengths my work has shown since are owed to their mentor ship. I thank my friend and editor at Academic Press, Chuck Crumly, for seeing this project through from the beginnning and for alternately holding my hand and kicking my butt, as required. Donna James and Joanna Dinsmore at AP provided much-needed help in the final stages of manuscript preparation, for which I am very grateful. Mary Jane Spring not only prepared some wonderful original artwork for my chapters, but also slaved over a hot scanner to produce many composite plates and other figures for reproduction. My father, George Schwenk, generously produced the penand-ink illustrations that introduce each section of the book. A number of people critically read chapters in whole or in part, offered comments, checked facts, and/or helped with bibliographic sources: William E. Bemis, A. W. Crompton, Nirvana I. Filoramo, Leo J. Fleishman, Harry W. Greene, Susan W. Herring, Dominique G. Homberger, Parish A. Jenkins, Jr., Kenneth V. Kardong, Nate Kley, John H. Larsen, Matthias Ott, Margaret Rubega, Carl D. Schlichting, Adam Summers, Carole Tomlinson, Giinter P. Wagner, Marvalee H. Wake, and Kentwood D. Wells. David Cundall, Harry Greene, Carl Schlichting, and Giinter Wagner supported this effort with their friend-
Preface ship, beer, and a high tolerance for whining. I thank my graduate students. Nirvana Filoramo and Charles Smith, for their forbearance in dealing with a busy and distracted advisor. My family—George Schwenk, Elizabeth Schwenk, Deborah Schwenk, John Schwenk, and Natalia Schwenk—have always been there for me and don't even seem to mind when I lapse into soliloquies about tongues and lizards. Finally, I thank my wife, Sandford von Eicken, and my son, Colton Schwenk, for their love and incredible patience.
XV
Work in my lab and preparation of the manuscript were made possible by grants from the University of Connecticut Research Foundation and the National Science Foundation (IBN-9601173) whose financial support is gratefully acknowledged. Kurt Schwenk Storrs, Connecticut April 2000
This Page Intentionally Left Blank
S E C T I O N
I
ff^^y^/^^y^,'
This Page Intentionally Left Blank
C H A P T E R
1 Tetrapod Feeding in the Context of Vertebrate Morphology KURT SCHWENK Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut 06269
I. INTRODUCTION A. Why Study Feeding? B. Delimiting the Topic II. APPROACHES TO THE STUDY OF TETRAPOD FEEDING A. Themes in Vertebrate Morphology B. Schools of Vertebrate Morphology C Techniques of Vertebrate Morphology III. CONCLUDING COMMENTS References
L INTRODUCTION A. Why Study Feeding? Despite huge strides in our understanding of gene structure and function, we remain largely ignorant about how phenotypes evolve. This apparent paradox arises from the fact that phenotypes embody emergent properties not directly codified in the genes. Fience, these properties are unknowable even with an exact and complete knowledge of the genetic system. It is possible to deduce them only from direct study of the phenotype itself. Knowledge of such properties is essential because it is the phenotype, after all, and not the genotype, that performs in an environmental context and it is this phenotype-environment interaction which determines lifetime reproductive success (fitness). A comprehensive view of organismal evolu-
FEEDING (K.SchwenKed.)
tion and diversity must, therefore, integrate genebased (bottom-up) and phenotype-based (top-down) approaches. Because morphology and function are the direct objects of selection, their study can contribute uniquely to the formulation of such a comprehensive view. Indeed, morphological studies have, historically, spawned many of the fundamental concepts of comparative biology: homology, analogy, adaptation, constraint, and Bauplan, for example. Furthermore, morphological principles might be critical to our understanding of how lineages navigate "phenotype space" through evolutionary time. Why do some systems remain static while others diversify? Do certain phenotypes confer intrinsically stable configurations that resist modification? How are segmental or modular body plans functionally integrated and how does the degree of their integration affect their ability to evolve? Such questions are best addressed through phenotypic analysis and it is in this context that a deep and detailed knowledge of form and function gains its greatest value. General principles of form-function evolution can be approached through a process of induction from specific systems. Tetrapod feeding is attractive in this regard because it offers several attributes that enhance its utility as a model system. First, tetrapod feeding systems are hugely variable, ranging from the edentulous jaws and extraordinarily protrusible tongues of ant- and termite-feeding mammals that feed frequently on many minute prey, to the syringe-like fangs and
Copyright © 2000 by Academic Press. All rights of reproductiori in any form reserved.
Kurt Schwenk venom glands of some caenophidian snakes that feed infrequently on few, very large prey. Within this array of systems, however, there has been the repeated acquisition of certain types. Lingual prehension of food, for example, is found in frogs, salamanders, turtles, lizards, birds, and mammals, and truly projectile tongues have evolved independently within salamanders, frogs, and lizards. Such phenotypic diversity allows us to address basic Darwinian questions regarding pattern and process of evolution and biodiversity; why are there so many types of feeding system? Conversely, the repeated acquisition of functionally analogous systems permits us to examine the evolutionary dynamic between extrinsic factors (e.g., environmental selection, adaptation) and intrinsic factors (e.g., developmental and functional constraint) in determining form. Despite extreme variation in form and function, tetrapod feeding systems are amenable to comparative analysis because they represent modifications of the same basic apparatus, comprising, for the most part, a set of unequivocally homologous parts. Many of the relevant skeletal components, for example, derive from the ancestral splanchnocranium (visceral skeleton), a defining set of vertebrate characters significant in the origin of the group and present in the tetrapod head under every variety of form and function (Hanken and Hall, 1993). Indeed, splanchnocranial elements of the feeding apparatus in one clade are modified to function as part of the auditory system in another, or the chemosensory system in a third. Such phenotypic recycling well illustrates the evolutionary truism that novelties are most often realized through the modification of preexisting forms. Finally, the relative functionality of the feeding system has, without a doubt, a large impact on individual survival and hence lifetime reproductive success. It is reasonable, therefore, to presume that feeding systems are under strong selection and that variations in feeding performance will have significant fitness consequences. Such a presumption allows us to apply optimality criteria and engineering principles in analyses of feeding system design and to identify trade-offs and constraints on design modification because very little about the feeding system is likely to be a result of random processes (e.g., fixation through drift). As such, the feeding system is well suited to analyses that address the relative contributions of adaptation and historical contingency to the phenotype. The importance of the feeding system to survival and fitness is underscored by its large impact on the tetrapod body plan, primarily through its influence on cranial form. The concentration of both feeding and
sensory components onto one portion of the body reflects a general, historical trend of increasing cephalization in vertebrate evolution. This, in turn, has set the stage for a complex integration and coevolution of feeding and sensory systems in a number of tetrapod clades through competition of these systems for limited cranial space and their shared use of certain anatomical parts. The characteristics of tetrapod feeding systems outlined earlier—ample variation, homology of parts, reasonable presumption of fitness consequences for variation, large impact on body form, and integration with other systems—promote the utility of feeding systems for comparative, evolutionary studies. Specific insights will emerge from detailed analyses of cladespecific patterns and cross-clade comparisons, as considered in individual chapters of this book. B. D e l i m i t i n g the Topic The decision to limit the coverage of this book to tetrapod vertebrates (as opposed to all vertebrates, including fish) was driven both pragmatically and conceptually. Pragmatically, issues of length and the tradeoff of depth versus breadth had to be considered. As editor, I felt strongly that tetrapod feeding systems were in need of a more in-depth treatment than they had been accorded heretofore. Although various aspects of feeding in fishes have been treated to several overviews (e.g., Lauder, 1982b, 1985; Bemis, 1986; Schaefer and Lauder, 1986; Motta, 1988; Westneat, 1990,1995a; Sanderson et al, 1991; Aerts and De Vree, 1993; Lauder and Shaffer, 1993; Wainwright and Lauder, 1992; Sanderson and Wassersug, 1993; Gerking, 1994; Frazzetta, 1994; Stouder et al, 1994; Vandewalle et al, 1994; Wu, 1994; Gosline, 1996; Drost et al, 1998), the rapidly growing literature in tetrapod feeding has barely been harnessed since the seminal papers of Bramble and Wake (1985) and Hiiemae and Crompton (1985). Only two major overviews of tetrapod feeding have appeared (Smith, 1993; Bels et al, 1994), but their coverage is too broad to achieve the depth for tetrapods aimed for here. One book provides in-depth coverage of "eating" in the context of human biology (Linden, 1998). Perhaps the most important reason to exclude fish from consideration, however, is because tetrapod feeding represents a real and significant departure from fish feeding—so much so that research in each area largely operates within a different paradigm. This departure arises from the simple fact that tetrapods, by definition, evolved to feed on land from ancestors that fed exclusively within water. The radically different
1. Tetrapod Feeding in Vertebrate Morphology physical properties of air versus water required an equally radical remodeling of the feeding system. The impact of the medium (air vs water) is so great that feeding system phenotype, overall, is widely viewed as "medium dependent" (Lauder, 1985; Bramble and Wake, 1985; Liem, 1990; Denny, 1990). Fish (presumably including those ancestral to tetrapods) rely almost universally on either suspension or suction mechanisms for feeding (e.g., Sanderson and Wassersug, 1993; Lauder and Shaffer, 1993), which exploit the high density of food particles in water, their buoyancy, and the frictional forces generated by the high viscosity of water. Faced with a low viscosity medium capable of producing only nominal frictional forces, a relatively much lower density of prey, and the need to lift and support the weight of any food item actually obtained, tetrapod ancestors could rely neither on suction nor on suspension feeding and thus required an entirely new approach to food acquisition and manipulation. They achieved this through invention of a true evolutionary novelty: a mobile, muscular tongue. A second area largely neglected in the present volume is the postcranial contribution to feeding function. Obviously, once captured, processed, and transported to the esophagus, food is further digested, assimilated, and reduced to excreta in the remainder of the gut, and these latter processes can have profound effects on body form. The reason for this neglect is largely historical; postcranial digestion and processing are mostly considered within the province of comparative physiology, whereas this book is rooted in the traditions of functional and evolutionary morphology (see later) (DuUemeijer, 1994). From this perspective, most feeding function occurs within the mouth and pharynx, hence the chapters herein tend to be limited to this region. Nonetheless, it is worth remembering that postcranial feeding adaptations can be significant and must be tightly integrated with cranial specializations for particular diets or feeding modes. An obvious example of such integration is mammalian herbivory, which evinces striking adaptations of the gut, including complex stomachs and enlarged caecae associated with the presence of a symbiotic microfauna capable of digesting cellulose (no vertebrate produces its own cellulase enzyme). A less well-known example is the relationship between feeding mode and digestive physiology in snakes. Most "advanced" (macrostomatan) snakes eat only rarely, but when they do they are likely to eat extremely large prey relative to their body size (Chapter 9). Due to a previous evolutionary commitment to an elongate, narrow-diameter body and small head [probably driven by a period of fossoriality and locomotory adaptation (e.g., Greene, 1997)], there was a
considerable mechanical challenge to be met before large prey could be engulfed. The cranial and mandibular apparatus of such snakes was radically modified for production of a wide gape, thus permitting passage of 'large prey through a small-diameter head into an extensible gut. The gut in such snakes is capable of very rapid and large-scale physiological upregulation when the occasional food item presents itself—for example, the intestine actually doubles in mass, mostly through mucosal growth (Secor and Diamond, 1998). Return to an atrophied, or quiescent, condition during long intervals between meals is energetically advantageous to these ectothermic animals. Thus in snakes, body form and the cranial apparatus are tightly integrated, and these aspects of the phenotype are correlated with less obvious, but equally dramatic adaptations of the intestine and its physiology. This example serves to remind us that feeding form and function are more than head deep, a fact the reader is urged to bear in mind.
II. APPROACHES TO THE STUDY OF TETRAPOD FEEDING As for any aspect of the phenotype, the feeding system can be studied in several complementary ways. Although the authors of the various chapters in this book were asked to cover certain general areas, differences in their contributions reflect not only peculiarities of the taxa they treat, but also differences in philosophy, technique, and approach. There is no agreed upon schema for these different approaches and most studies and investigators defy characterization by simple, typological labels. However, it is worth considering, in a broad sense, the different philosophical bases and "schools of thought" characterizing vertebrate morphology as an intellectual realm. It is this realm that has provided the context for this book and its individual contributions (see also Liem and Wake, 1985).
A. T h e m e s in Vertebrate Morphology Animal anatomy is certainly one of the oldest, if not the oldest, biological science, tracing its origins to preAristotelian times (Cole, 1944; Singer, 1957); however, the term "morphology" was only coined in 1800 (Nyhart, 1995). It was during the 19th century that morphology, as a discipline, achieved its heyday, becoming a dominant area of biological research intellectually, if not institutionally (Nyhart, 1995). Certainly, Darwin's
Kurt Schwenk (1859) formulation of the theory of natural selection grew out of an intellectual milieu in which the extensive and painstaking documentation of morphological variation within and among species figured prominently. It is therefore ironic that morphology is seen as having contributed little to the 20th century evolutionary synthesis, commonly referred to as neo-Darwinism (Ghiselin, 1980; Mayr, 1980; but see Waisbren, 1988). While this may be true, it is equally true that neoDarwinism has failed to incorporate many of the lessons of morphology and there is growing dissatisfaction among some morphologists with the ability of neo-Darwinian theory to deal adequately with the totality of phenotypic evolution (see later). In any case, since Darwin, few fields have embraced evolutionary theory as enthusiastically as morphology {contra Ghiselin, 1980), for it is in evolution that the morphological concepts of homology, analogy, Bauplan, and "unity of type," for example, become sensible and elevated in significance. Indeed, the entire field of systematics as currently practiced is based on the character concept, a pre-Darwinian morphological construct melded with neo-Darwinian evolutionary theory. One cannot overstate the centrality of morphology to the development of modern biological thought. What follows is a subjective and largely overlapping list of dichotomies that attempt to clarify and situate the discipline and subdisciplines of vertebrate morphology. The dynamic and protean nature of the field precludes even the pretense of total agreement in this formalization, but one hopes it might be of some heuristic value to new students of vertebrate form and function. 1. Morphology vs
Anatomy
Unlike botanists, who make a clear and formal distinction between plant morphology and plant anatomy (the former referring to whole-plant form and the organization of parts, the latter to the fine structure, or histology, of parts), zoologists are uncharacteristically vague in their usage of these terms. Morphology "deals with the form of living organisms, and with relationships between their structures" (from the Greek stem morpho), whereas anatomy is "the science of the structure of the bodies of humans, animals, and plants" (derived from the Greek stems ana- and -tomy, meaning "repeated cutting") (Oxford English Dictionary; Brown, 1993). Although these definitions would appear to be more-or-less synonymous, in current zoological usage they connote somewhat different things, the sense of which is hinted at in the etymology. Morphology is the study of "form," which can be generalized to all hierarchical levels, from organelle to whole
organism. It is also concerned with the relationships among structures, hence it includes emergent features of form such as relative size, allometry, and even function and physiology for some (see later). Thus, morphology is a more expansive term, subsuming anatomy (as toad is to frog, so anatomy is to morphology). Anatomy is largely limited to the hierarchical level of organs, or body parts, i.e., those elements of form revealed by dissection, and historically has been associated with human beings (Owen, 1866; Singer, 1957). For example, Gegenbauer (1878), epitomizing the late 19th century view, divided anatomy ("the doctrine of structure") into anthropotomy and zootomy, the dissection of humans and nonhuman species, respectively. According to Gegenbauer, anatomy is not a science because it is restricted to the empirical generation of descriptive data. However, these data achieve the status of science when they lead to synthesis and abstraction, only possible by comparing anatomical data among species, i.e., "comparative anatomy," or what Owen (1866) called "homological" and "zoological anatomy." Gegenbauer (1878) included anatomy within the larger field of morphology such that all of biology could be divided into physiology and morphology (see later) and morphology, in turn, into anatomy and embryology. Thus, while some treat the terms anatomy and morphology equivalently, most modern usage continues to reflect the 19th century view in which anatomy is restrictive, descriptive, based on dissection of body parts, and primarily (but not exclusively) anthropocentric, whereas morphology is holistic, multihierarchical, often synthetic and concerned not only with the structure of the parts, themselves, but relationships among the parts, and of parts to function. Although Owen (1866: vii) did not use the word morphology, his notion of "zoological anatomy" captures nicely the essence of modern animal morphology as "that which investigates the structure of an animal in its totality, with the view of learning how the form or state of one part or organ is necessitated by its functional connections with another, and how the co-ordination of organs is adapted to the habits and sphere of life of the species." The history of these ideas and disciplines is reviewed admirably by Russell (1916), Cole (1944), Singer (1957), and Nyhart (1995), among others. Perhaps the most important distinction between anatomy and morphology as formal disciplines is that anatomy is mostly unconcerned with the origin of structure, whereas the origin and generation of form are at the philosophical core of morphology (e.g., Gegenbauer, 1878; His, 1888; Russell, 1916; Thompson, 1942; Davis, 1960; Nyhart, 1995; Webster and Goodwin, 1996). Indeed, since the early 19th century mor-
1. Tetrapod Feeding in Vertebrate Morphology phologists have endeavored to discern "rules of form" that might underlie ontogenetic and phylogenetic morphological transformations. Interest in rules of transformation predates an evolutionary world view (e.g., Driesch, 1908; Nyhart, 1995). Elucidation of such rules was the goal of Geoffroy Saint-Hilaire (1818; in Russell, 1916) and the "transcendental" morphologists, and explored by Owen (e.g., 1848,1849) in the context of "archetypes." However, this fundamental interest in morphological rules of transformation is clearly evident in the modern structuralist movement (see Piaget, 1970; Rieppel, 1990), which, although it takes many forms, has at its core the notion that phenotypic hierarchies manifest emergent properties that certainly influence, if not dictate, directions of further phenotypic evolution (e.g., Russell, 1916; Whyte, 1965; DuUemeijer, 1974, 1980; Riedl, 1978; Lauder, 1982a; Ho and Saunders, 1979, 1984; Roth and Wake, 1985; Rieppel, 1986; Wagner, 1986; Wake and Larson, 1987; Wake and Roth, 1989; D. Wake, 1991; Smith, 1992; van der Weele, 1993; Schwenk, 1995, 2000; Amundson, 1996; Hall, 1996, 1998; Raff, 1996; Webster and Goodwin, 1996; Arthur, 1997; Wagner and Schwenk, 1999; Schwenk and Wagner, in preparation). Indeed, radical "process structuralists" (see Smith, 1992) go so far as to suggest that random variation, natural selection, and phylogenetic history (the tenets of Darwinism) are secondary players in the generation of hierarchically organized phenotypic diversity. They seek instead a "rational," predictive science of form based on rules of selforganization and organismal development (e.g.. Ho and Saunders, 1979,1984; Goodwin, 1989; Webster and Goodwin, 1996). Whether radical or moderate, inherent in the structuralist, morphological view is the sense that the atomistic, neo-Darwinian, gene-based paradigm of phenotypic evolution is incomplete and that the organism creates the context for its own further evolution, thus setting the stage for a top-down chain of evolutionary causality in phenotypic evolution (Whyte, 1965; Wagner and Schwenk, 1999). 2. Form vs Function The conceptual dichotomy of form and function is ancient, at least as old as Aristotle (Russell, 1916; Lauder, 1982a; Padian, 1995). It was identified by Russell (1916) as a major theme in the history of morphology, as evident in the title of his classic book {Form and Function), and it has continued to be a central theme in the study of phenotype throughout this century (e.g., Woodger, 1929; Bock and von Wahlert, 1965; Lauder, 1981,1982a; Gans, 1969,1988; Wake, 1992; Lauder et a/., 1995; Padian, 1995; Weibel, 2000). Historically, this dichotomy was discussed in terms of primacy (e.g., Rus-
sell, 1916; Woodger, 1929; Lauder, 1982a, 1995, 1996; Appel, 1987; Amundson, 1996): is it form that determines function or function that determines form? In the early 19th century, George Cuvier, for example, believed in the primacy of function, positing that form was imposed by the functional demands of the environment and that similar forms reflected similar functional "adaptation." He believed that form and function were so tightly bound in this causal sense that one could, through the "principle of correlation," predict the whole form of an animal from a single part. In contrast, his contemporary, Etienne Geoffroy St. Hilaire, sought a "pure morphology of organic forms" wherein function emerged as a mere by-product of structure. Indeed, Geoffroy eschewed any consideration of function in comparative anatomy and believed that changes in function followed secondarily upon primary changes in structure (Russell, 1916). Although the debate between Cuvier and Geoffroy epitomizes the contention in the field regarding form and function early in the 19th century, the dichotomy was formalized on a grander scale later in the century when the whole of biology was cleaved into physiology and morphology (e.g., Owen, 1866; Gegenbauer, 1878; see Russell, 1916; Nyhart, 1995). Physiology comprised all aspects of organismal function, whereas "the investigation of the material substratum of those functions, and accordingly of the phaenomena of form of the body and its parts, as well as the explanation of the phaenomena of form by reference to function, is the business of Morphology" (Gegenbauer, 1878:1). Physiology already had a long history separate from morphology and the two fields continued through the 19th century and into the 20th to develop in isolation (Russell, 1916; Woodger, 1929). They have remained distinct traditions until quite recently and even now a complete rapprochement remains elusive (e.g., Weibel et al, 1998; Weibel, 2000). Although there is obvious overlap in domain, comparative physiologists, for the most part, continue to study function of a different sort and in a different context than functional morphologists. Woodger (1929) regarded the splitting of form and function as an abstraction not manifest in nature. The "antithesis" results from employing "different modes of apprehension" and from artificially separating space and time (anatomy focusing on the former, physiology the latter) which are simultaneous attributes of living structures. Certainly, most modern evolutionary morphologists would consider the question of formfunction primacy moot, recognizing the chicken-andegg nature of the dichotomy in a historical context. Indeed, evolutionists do not acknowledge a distinction at all, capturing the seamlessness of form and function in
8
Kurt Schwenk
the inclusive term "phenotype." Nonetheless, vestiges of the dichotomy persist within morphology. This is most apparent in the question of how one studies function: must one study function directly or is it possible to infer function from structure? Certainly pure, descriptive, anatomical studies are not much interested in function and are therefore not relevant here, but there is, at the same time, a long tradition in vertebrate morphology of inferring animal function from static form. This is most apparent in the 20th century paleontological literature where functional speculation is commonplace (e.g., Rainger, 1989; Hopson and Radinsky, 1980; Thomason, 1995), but it is also rife in studies of living species. Indeed, the ability to predict function from form is at the conceptual core of "ecomorphology" (see later). The presumption that function can be inferred from form stems more or less directly from Cuvier's principle of correlation (see earlier discussion). It depends absolutely on the degree to which form and function are integrated so that each can stand as a proxy for the other. However, Lauder (1995, 1996) has shown that structure and function are not always tightly matched in a predictive sense, thus falsifying the assumption upon which the principle of correlation is based. He advocates a direct, experimental approach to function based on analysis of living specimens and emphasizes the need for quantification of functional data. Experimental, quantitative approaches to functional morphology are relatively recent. Przibram (1931:14), for example, noted, "Whereas in Physics, Chemistry or even Physiology nobody would nowadays try to state laws without experimental evidence, the general statements in animal Morphology have been mostly based on speculations. Only of late has the experimental method been gaining ground in Biology and research is being carried on in a large scale at institutions adapted to the p u r p o s e . . . Again I would refer to Physics and Chemistry as brilliant examples of what may be achieved in the way of unraveling the laws of nature by quantitative experiment and mathematical formulation based thereon." Thompson (1942:2) was, likewise, intent on a more quantitative approach to form: "But the zoologist or morphologist has been slow, where the physiologist has long been eager, to invoke the aid of the physical or mathematical sciences." Thus, while the inseparability of form and function may be intellectually acknowledged by most practicing morphologists, the ancient dichotomy remains manifest in investigators' approach to function: inferential and qualitative or experimental and quantitative. These different approaches are elaborated further. Although the dangers of inferring function from static form seem clear, seldom, if ever, is concern ex-
pressed for the converse: can the detailed study of function in the absence of equally detailed structural data be misinformed? Unquestionably (see also Cans, 1986). I am frequently impressed by the near total lack of morphology evident in some studies of putative "functional morphology" in which quantitative aspects of function are analyzed in a structural vacuum, almost wholly uninformed by knowledge of the relevant anatomy. At the very least, morphological data can, in such cases, eliminate from consideration alternative mechanistic hypotheses. Further, functional conclusions can be revealed as false, or even absurd, when held in the light of form. The dichotomy of form and function is an undeniably useful, heuristic tool. However, we must acknowledge that this dichotomous view is a philosophical construct, simplifying for us the complex notion oiphenotype in which form and function are interwoven dynamically and infrangibly (Woodger, 1929). As Ruffini (1925) noted, "form is the plastic image of function" (in DiDio, 1986:197). There can be no argument for primacy, only relationship. Therefore, all approaches to the study of phenotype are valid, if not complete, so long as conclusions follow from data and speculation is labeled as such. If only for pragmatic reasons, most individual studies will continue to emphasize either form or function, with syntheses relatively rare. Nonetheless, form and function, experiment and description, qualitative and quantitative data must be held as equally important, complementary, and ideally, "reciprocally illuminating" elements in the study of morphology [Lombard (1991) and Lauder (1995) offer related perspectives on these issues]. 3. Idiographic vs
Nomothetic
Idiographic studies are those that characterize a specific case without regard to its typicality or generality. Studies which treat the morphology or function of a single species are of this type. An organism, or part thereof, can be treated purely as mechanism and one can ask, simply, what does it look like? what are its component parts? and how does it function? without regard to whether the results are applicable to a larger group or peculiar to the case at hand. Such studies can have high intrinsic interest and are valuable in proportion to the quality of their data, but they do not inform us about the next system we might study, nor how one form evolves into another. Pattern, causality, and prediction derive, instead, from nomothetic studies, which strive to elucidate higher principles, rules, or, in the strict sense, scientific laws. Comparative morphology, like Owen's (1866) zoological anatomy (see earlier discussion), falls within
1. Tetrapod Feeding in Vertebrate Morphology this realm because it seeks generalities beyond the specific (Gegenbauer, 1878). However, nomothetic, comparative, vertebrate morphology does not require an evolutionary world view. As noted earlier, early 19th century morphologists, such as Cuvier and Geoffroy, were very much interested in general rules of form without reference to historical descent. Indeed, some modern structuralists (see earlier discussion) are also interested in an organism-based theory of form independent of evolution, arguing that * 'evolution provides only limited insight into the problem of form as regards both the causal explanation of form and the relations between forms . . . what is required is the development of a specific causal-explanatory theory of form, a theory of morphogenesis in the most comprehensive sense . . . such a theory will be as fundamental to biology, if not more so, at least as the theory of evolution" (Webster and Goodwin, 1996 :ix; see also His, 1888). Nonetheless, most modern, comparative studies in vertebrate morphology are rooted in the neo-Darwinian tradition, but framed in an explicit cladistic, phylogenetic context (e.g., Lauder, 1981, 1990; Greene, 1986a), whereas still others seek some middle ground (e.g.. Wake and Larson, 1987; Wagner and Schwenk, 1999). Whether organism centered or phylogenetic, such studies are part of a long, nomothetic tradition in vertebrate morphology that has contributed fundamentally to the principles of modern biology. 4. Laboratory vs Field Most vertebrate morphology must, of necessity, be undertaken within the laboratory. However, any consideration of function, whether it is inferential or experimental, is at risk if it fails to consider the actual behavior of organisms in the field (Greene, 1986b, 1994). For the inference of function from form, natural history data can circumscribe the universe of possibilities. For example, one need not waste time speculating about the role of certain skull attributes in processing flesh if it is known from observation and stomach content analyses that the organism in question is an insect specialist. Likewise, a treadmill study of locomotion in gibbons would be largely uninformative about the evolution of its form, whereas initial field observations of its natural locomotory mode (brachiation) would suggest a different experimental approach. These examples are obvious and might appear silly, but there is no doubt that functional and evolutionary morphology can be led astray in failing to consider the natural behavior of animals. A classic case is that of monitor lizard (Varanus) morphology and diet. A persistent theme in functional-morphological (and other) studies is that varanid lizards are adaptively special-
ized for carnivory and the ingestion of large prey (e.g., Rieppel, 1979; Smith, 1982). As such, experimental studies of feeding function have analyzed monitors during feeding on rodents (e.g.. Smith, 1982; Condon, 1987). However, a comparative study of diet in monitor lizards revealed that nearly all species are insectivorous, eating many small prey items rather than a few large ones (Losos and Greene, 1988). Varanus exanthematicus, the most commonly used species in liveanimal studies (e.g.. Smith, 1982), is one of the most extreme insect specialists within the family. Indeed, it is in many other respects unusual and derived— neither a typical lizard nor even a typical varanid. Thus, the experimental conditions of these studies were unnatural, or at least, atypical, and therefore the generality of their results is suspect. Certainly the interpretation of derived varanid cranial attributes as adaptations for eating large, vertebrate prey is without foundation. Rather, the typical monitor lizard cranial apparatus seems to represent a specialized pincer system for effectively nabbing elusive prey with the tips of the jaws (Frazzetta, 1983; see Chapter 8). B. Schools of Vertebrate Morphology The simple 19th century division of biology into physiology and morphology is no longer tenable. Experimentalism, technological advancement, new techniques, new theory, and increasing specialization have driven fragmentation of biology, generally, and vertebrate morphology, specifically. Because no investigator can master all possible approaches to morphology, individual studies tend to emphasize one approach, technique, or school of thought. The following list of approaches to vertebrate morphology is hardly exhaustive, but it represents a cross section of the field as currently practiced. Elements of these different approaches are evident in the chapters of this book. 1. Descriptive
Morphology
Description remains the foundation upon which morphological studies must ultimately rest. Good descriptive morphology is as rare as it is beautiful, virtually an art. Unfortunately, it may be a dying art because it is little valued within the context of modern science. This is tragically short-sighted because one thing is clear about high-quality, well-documented descriptive anatomy: it may represent the only truly hard, objective data in morphological research, as free of fashion and interpretation as possible. It is, therefore, timeless. This timelessness is critical because it means that descriptive data can be used in the service of other studies and new theories, including those not conceived at
10
Kurt Schwenk
the time the data were generated. As such, descriptive data are of value to future generations of biologists— it is common, for example, for 19th century and even older literature to be cited in modern morphological studies, often in contexts, such as cladistic, systematic analysis, never imagined by the original authors. For how many other fields, or even other approaches to morphology, can this be said? Of what value now, for example, are systematic analyses based on the once widespread and "cutting edge" method of phenetics? How many molecular studies more than 10 years old remain in currency? In contrast, a morphological treatise, such as D. Dwight Davis' (1964) classic study of the giant panda, is unceasingly relevant. Indeed, its value only increases with time as the availability of rare species declines. It is a telling fact that Davis' (1964) conceptual, evolutionary analysis of his descriptive work is dated and of dubious value in today's milieu of cladism and molecular genetics, but the 300 pages of painstaking, descriptive anatomy that precedes it remain—untainted, unfiltered, uninterpreted, and fully available for reanalysis in another context, now or in the future. Descriptive morphology can apply to any hierarchical level, but it is done most commonly at the anatomical level (see earlier discussion). Anatomies can be regional, such as Oelrich's (1954) Anatomy of the Head of Ctenosaura pectinata, or they can be systemic, as in Romer's (1956) Osteology of the Reptiles. Tissue-level morphology is usually referred to as histology, loosely limited to those aspects of morphology resolvable by light microscopy. Cellular and subcellular morphology studied by means of scanning and transmission electron microscopy is referred to as ultrastructure. In the parlance of vertebrate morphology, structure usually refers to organs and organ systems, although it can be in reference to any hierarchical level. However, in the biomedical community, "structural biology" is purely molecular. 2. Evolutionary
Morphology
Evolutionary morphology defies precise characterization because it is the most inclusive school of vertebrate morphology (D. Wake, 1982; M. Wake, 1991, 1992). It can be experimental, functional, or purely descriptive, but it has at its core an evolutionary intent. By definition, therefore, it is comparative. As such, it most clearly manifests the tradition of 19th century comparative anatomy in that it continues to seek higher level understanding of the generation and transformation of form (see earlier discussion; Davis, 1960). Evolutionary morphology is distinct from morpho-
logical systematics in that the latter uses morphology in the service of phylogeny, whereas the former uses phylogeny in the service of morphology. As such, the historical relationships of organisms are of secondary interest to the evolutionary morphologist whose main concern is the history of the characters themselves, i.e. character analysis in the purest sense. Whereas the systematist uses morphological characters as a matter of routine, in evolutionary morphology the character concept is, itself, a subject of study (Wagner, 2000). Furthermore, studies in evolutionary morphology may attempt not only to generate patterns of character evolution, but to address ultimate issues of causality, as such, the processes of phenotypic evolution (e.g., Galis, 1996; Wagner and Schwenk, 1999). 3. Functional
Morphology
The focus of functional morphology, transparently, is function. As discussed earlier, there is contention about whether studies that are limited to the inference of function from structure can be considered "functional." Nonetheless, purely inferential papers continue to appear with "functional morphology," or sometimes "functional anatomy" in their titles. Although the latter term may be an attempt at truth in advertising (i.e., it flags the paper as inferential), "functional anatomy" is to be discouraged because it is, in some sense, an oxymoron (see discussion earlier). Thus, in most recent and I would say preferred usage, functional morphology refers specifically to analyses based on the direct measurement of function in live, behaving animals. Functional morphology may be evolutionary; it is often purely mechanistic and idiographic (D. Wake, 1982; M. Wake, 1992). Like good descriptive morphology, good functional studies of single species can be of great intrinsic interest, and to the extent that data are objective, they are available to other investigators for meta-analyses. It is worth noting here that while most modern functional morphology is "experimental," there are two quite different usages of this word in the field. The first is appropriate and consistent with its use in traditional, reductionist scientific method in referring to hypothesis testing. Such a hypothetico-deductive approach to function usually requires manipulations of experimental subjects and careful controls, thus in functional morphology such studies often require invasive techniques and the use of complex technology. Nonetheless, it is important to note that hypothesis testing, and therefore experimentation sensu stricto, is not restricted to functional studies. It is possible to test certain types of hypotheses, even functional hypotheses, by reference to morphology alone. However, such descriptive
11
1. Tetrapod Feeding in Vertebrate Morphology analyses are never labeled "experimental." Thus, in current usage, the application of hypothetico-deductive analysis is insufficient to qualify a study as experimental. Nor would the use of control and manipulated subjects seem sufficient. Rather, the principal criterion appears to be the application of technology itself. This has led to a second, diluted, and much more common (if inappropriate) use of the term to mean any analysis employing technologically based methods. For example, many studies characterize muscle activity patterns during normal behavior using electromyography (EMG) and are labeled "experimental" even though they are purely descriptive—they show X muscle to be active during Y behavior. There need be no control, manipulation of the conditions, nor testing of functional hypotheses. However, not all use of technology seems to merit an experimental epithet. For example, use of binoculars to describe the same behavior, Y, in the field would not be considered experimental, although it may serve to test a functional hypothesis and scientifically it may be the more valuable of the two studies! Therefore, the currency of the technology employed, or perhaps its degree of invasiveness, seems to be a deciding factor. I see no obvious solution to these terminological paradoxes, but highlight them here as a reminder to students of form that scientific merit is a quality independent of technique or titular fashion. 4.
Biomechanics
Biomechanics might be regarded as a subdiscipline of functional morphology—certainly many functional morphological studies contain elements of biomechanical analysis (e.g., Gans, 1976; Rayner and Wootton, 1991). However, biomechanics is, in the strict sense, directly inspired by various fields of engineering and is far more mathematical and less anatomical than is typical of functional morphology (e.g., Fung, 1993). Indeed, biomechanics quickly blends into theory and modeling (see later). In addition, its more-or-less pure design approach to form (see Lauder, 1996) means that biomechanical studies generally eschew comparative or evolutionary issues. In fact, most of the biomechanics literature is biomedical and sports oriented. This can be contrasted with transformation morphology (see later) which, though often highly biomechanical, is concerned with understanding ontogenetic and phylogenetic transformations in form and function. Biomechanics in (nonbiomedical) morphology is especially concerned with the physics of biological materials and surrounding media, and their consequences for organismal form, function, and, occasionally, evolution (e.g., Wainwright et al, 1976; Alexander, 1985; Wainwright, 1988; Vogel, 1988, 1994; Vincent, 1990;
Denny, 1993). For example, it was noted at the outset of this chapter that the tetrapod feeding apparatus is, to a large extent, medium dependent in the sense that, depending on whether the organism lives in air or in water, the physical properties of the medium will impose a limited realm of phenotypic solutions to the problem of ingesting, processing, and swallowing food. Thus the convergence of feeding mechanisms in fish and secondarily aquatic tetrapods could be predicted from biomechanical first principles, i.e., the viscosity and fluid dynamics of water. Indeed, it is often found that certain aspects of the phenotype conform to the expectations of mechanical optimization in some parameter or other in a way that clarifies their functional and evolutionary significance. For example, although the tongue of squamate reptiles was well known to be a chemical sampling device, the functional significance of its forked form in snakes and some other species was not understood. An engineering approach in a comparative context, however, suggested that the fork provides a two-point sampling device for the detection of chemical gradients useful for following pheromonal trails (Schwenk, 1994). Fluiddynamic theory further suggested that the rapid oscillation of the tongue in the air characteristic of these species is an adaptation to enhance the molecular diffusion of environmental chemicals into the fluid on the tongue's surface, thereby amplifying the chemical signal carried into the mouth (Schwenk, 1996). Thus, while biomechanics can be pursued idiographically without reference to evolution, it can powerfully inform questions about the evolution of form and function (Lauder, 1991; M. Wake, 1992). 5. Developmental
Morphology
Like evolutionary morphology, developmental morphology is a broad field that grew out of the 19th century tradition of embryology (see earlier discussion). It implies no particular technique or approach, but is concerned with the ontogenetic transformation of form. Most often such studies are limited to the embryonic period, but they may treat any life stage so long as their concern is ontogenetic transformation. Because the focus of developmental morphology is on intrinsic attributes of individual organisms, it is the part of vertebrate morphology most clearly identified with biological structuralism (e.g., Wagner, 1988; Hall, 1994; D. Wake, 1991; Webster and Goodwin, 1996). Developmental studies have revealed principals of selforganization and pattern formation at the organismic level that are likely to be relevant to the directions and dynamics of phenotypic change in lineages through evolutionary time. Yet neo-Darwinian theory springs
12
Kurt Schwenk
from population-level phenomena, hence it has neglected the role of the organism in directing patterns of phenotypic change. To a large extent, the modern notion of "evolutionary constraint" has emerged from the field of developmental morphology (e.g., Maynard Smith et al, 1985; Wagner, 1986, 1988; D. Wake, 1991; Schwenk, 1995; Schwenk and Wagner, in preparation). 6. Ecological
Morphology
Ecological morphology (or ecomorphology) is a rather amorphous hybrid of several morphological schools (as suggested in the title of one review; Wainwright, 1991). Nominally it owes its origin to van der Klaauw (1948), but conceptually it is rooted in the ancient observation of the "fit" between organisms and their environments. As such, the prevailing definition of ecomorphology is "the study of the relationship between the morphology of the organism and its environment" (Wainwright and Reilly, 1994a: 3). Indeed, it is difficult to say how the basic premise of ecomorphology differs from Cuvier's "principle of correlation"—both assume that organismal form is tightly correlated to environmental conditions. Like Cuvier, ecomorphology aspires to predict form from ecology and ecology from form (e.g., Emerson, 1991; Motta and Kotrschal, 1992; Norton and Brainerd, 1993; Wainwright and Richard, 1995; Losos et al, 1998). However, unlike Cuvier, modern ecomorphologists invoke natural selection and adaptation to account for the formenvironment fit. Thus, they espouse the need for integrating historical approaches into ecomorphological studies (e.g., Motta and Kotrschal, 1992; Wainwright and Reilly, 1994a; Losos and Miles, 1994; Westneat, 1995b). Wainwright and Reilly (1994b) provide an entree into the ecomorphological literature. As they noted, ecomorphological studies usually are either mostly ecological or mostly morphological (Wainwright and Reilly, 1994a). However, at its best, ecomorphology attempts a true synthesis by addressing questions that are fundamentally ecological through the phenotypic analysis of organisms. The key point is that the study of phenotype is pursued at the population level. I think it is only here that ecomorphology clearly distinguishes itself. Otherwise it tends to suffer one of three fates: it is simply ecology with a few superficial measurements (e.g., limb length) thrown in; it is simply good functional or evolutionary morphology, which should, after all, incorporate ecological and natural history data into its analyses (see earlier discussion); or, most commonly and worst of all, it suffers from naive adaptationism {sensu Gould and Lewontin,
1979), which finds that organisms are, indeed, adapted to their environments, e.g., animals with big mouths can, in fact, eat big prey. Although such studies are not without merit, they evince little conceptual advancement over "natural theology" (e.g., Amundson, 1996)—things are as they must be (see also Liem, 1993). Rather than using the form-environment fit as a point of departure, ecomorphology can and should probe the limits of this assumption by incorporating notions of constraint and other potentially limiting factors into its paradigm (e.g., Barel et ah, 1989; Liem, 1993; Losos and Miles, 1994). Certainly the relationship between the organismal phenotype and the environment is far more complex than usually assumed (e.g., Simpson, 1953; Whyte, 1965; Greene, 1982; Wagner and Schwenk, 1999). 7. Transformation
Morphology
Transformation morphology is related to evolutionary and developmental morphology, but it is concerned specifically with the process of ontogenetic and phylogenetic phenotypic transformation (Barel, 1993; Galis, 1996). Its focus is less on the pattern of transformation (e.g., documentation of form-function complexes in comparative, or ontogenetic series) than on the rules of transformation underlying observed patterns. According to Galis (1996:128), these rules can be elucidated "by constructing biomechanically feasible transformation schemes, by studying key structural changes that break important constraints enabling a cascade of changes, and by studying the mechanisms that preserve the match between form and function during ontogenetic and evolutionary change." Transformation morphology has grown out of the Leiden school of morphology (e.g., van der Klaauw, 1945; DuUemeijer, 1959, 1974, 1980, 1989) and, as such, its emphasis is functional. Furthermore, as suggested by Galis' (1996) quote, modeling and engineering approaches to evolutionary transformation (e.g., Zweers, 1991; Zweers and Vanden Berge, 1997; van Leeuwen and Spoor, 1992; Galis, 1992,1993; Galis and Drucker, 1996; see later) are preferred to the comparative, phylogenetic methods typical of North American workers (e.g., Lauder, 1981,1990; Lauder et al, 1995; Larson and Losos, 1996). 8, Constructional
Morphology
Like transformation morphology, with which it overlaps, constructional morphology is largely a European tradition. It was formalized by Seilacher in the 1970s (e.g., 1970, 1973, 1979), but has earlier roots in Leiden (see references given earlier) (Reif et al., 1985).
1. Tetrapod Feeding in Vertebrate Morphology Constructional morphology deals explicitly with constraints on adaptive evolution arising from the physical properties of materials and the mechanisms of their deposition and growth, although it is occasionally more broadly construed as pertaining to any sort of structural constraint (e.g., Reif et ah, 1985; Barel et at, 1989; see also papers in Schmidt-Kittler and Vogel, 1991). Whereas ecomorphology takes as its starting point the adaptive nature of organismal form, constructional morphology focuses on nonadaptive aspects of form; it therejfore disputes the notion that all morphological variation reflects adaptive responses to different environments. Roth (1989), for example, showed that differences in dental form among nominal species of fossil elephant are attributable to "fabricational noise," reflecting individual differences in masticatory stresses during postnatal development of the teeth. As such, quite large differences in the form of the elephantid dental battery are a consequence of the uniquely retarded growth of elephant teeth as compared to other mammals. While the phenotypic plasticity that permits such individual responses to juvenile stresses might, itself, be an adaptation (e.g., Schlichting and Pigliucci, 1998), the particular morphcJiogical variants that arise are the result of nonadaptive processes stemming from mechanistic aspects of tooth development and growth. It is these processes, and their effects on the evolution of phenotype, that are the focus of constructional morphology. 9.
Morphometries
Morphometries is a catch-all term for a variety of methods that seek to capture and quantify complex shapes based on external measurements. It grew out of the emerging field of biometry in the late 19th century (Bookstein, 1994) and later, the pioneering work of D'Arcy Thompson (1917, revised and expanded in 1942) who sought a mathematization of zoology and morphology. Although best known for his application of Cartesian coordinates to the study of shape transformation, Thompson (1917) actually dealt more extensively with other quantitative aspects of form, especially patterns of growth. This work strongly influenced the subsequent, formal study of allometry, as evident in the dedication of Huxley's (1932) classic book to Thompson. Allometry, therefore, might be considered a part of the morphometric tradition (Huxley, 1932; Gould, 1966). In light of his stated goal, it is ironic that Thompson's (1942) use of deformed Cartesian coordinate planes to describe shape change was largely intuitive and not particularly quantitative. However, a num-
13
ber of methods have been described that accomplish Thompson's (1942) intent with greater rigor (e.g., Rohlf and Bookstein, 1990; Bookstein, 1991; Marcus et al, 1996; McLellan and Endler, 1998). However, these are limited in application to two-dimensional shapes, or two-dimensional projections of three-dimensional forms, which diminishes their utility to morphologists interested in complex anatomy (however, some threedimensional morphometric methods are being developed, e.g.. Roth, 1993). Perhaps of greater concern is the question of using coordinate-based, landmark data in evolutionary character analysis. As noted previously, the issue of atomizing organisms and delimiting characters is an area of active empirical and conceptual work (e.g., Wagner and Schwenk, 1999; Wagner, 2000; Schwenk, 2000). It is by no means clear that landmark data fulfill the criteria of homology necessary to establish them as "characters" in the sense of semiautonomous units of phenotypic evolution. Bookstein (1994), for example, has argued that biometrical shape characters are not commensurate with traditional, phenotypic characters as used in systematic analyses, whereas others suggest that, in the proper context, morphometric characters can be used this way (e.g., Zelditchefa/.,1992,1993). 10, Theory and Modeling Although most morphological theory is conceptual (e.g., DuUemeijer, 1974; Gould and Lewontin, 1979; Wake et al, 1983; Wagner and Schwenk, 1999), it can also be highly mathematical (e.g.. Burger, 1986; Van Leeuwen, 1991, 1997). As for other biological disciplines (e.g., ecology and evolution), theory in morphology is most effective in reciprocity with empiricism. The case of chameleon tongue projection offers an excellent example of this synergism. The mechanistic basis of the explosive, ballistic projection of the chameleon tongue defied understanding for centuries (see Chapter 8). However, an incisive, experimental study by Wainwright and Bennett (1992) implicated the use of hydrostatic elongation of the lingual accelerator muscle in this function. The Wainwright-Bennett model, based on an in vitro study, was beautifully supported by a theoretical model of the biomechanics of the accelerator muscle, which accurately predicted details of hyolingual form and function (Van Leeuwen, 1997). The combined results of these studies have, in turn, been corroborated and extended by an in vivo study using high-speed X-ray movies (cineradiography) of chameleon tongue projection (Schwenk et al, in preparation). As a result of this interplay among descriptive anatomy, functional inference, in vitro and
14
Kurt Schwenk
in vivo experimentation, and mathematical modeling, the mechanism of chameleon tongue projection is now reasonably well understood. Because many functional-morphological studies focus on musculoskeletal systems, it is no surprise that most of the inspiration for functional models comes fron\ mechanical engineering and materials science (e.g., Gans, 1974; Wainwright et ah, 1976; Alexander, 1983) (see also Section II, B, 4). Frazzetta's (1962) application of a "quadric-crank" model to the amphikinetic lizard skull is an early example of this approach, but mechanical modeling approaches are varied (see references given earlier; other examples include de Jongh et a/., 1989; Otten, 1989; Galis, 1992, 1993; Russell and Thomason, 1993; Weishampel, 1993, 1995; Greaves, 1995; Herrel et al, 1998). Models are especially powerful when they make predictions that can be tested empirically or with natural history data. For example, Galis' (1992) model of bite forces in a cichlid fish showed that ontogenetic changes in the pharyngeal jaws accounted for the inability of small fish to pierce the integument of certain prey types and the absence of such prey in the natural diets of the fish until they achieved a certain size (GaUs, 1993). The model accurately predicted the size at which fish started to include the harder prey type in their diets. Galis (1993) suggested that such patterns establish causality, as opposed to mere correlation, although this assertion is open to debate. C. Techniques of Vertebrate Morphology One of the great beauties of vertebrate morphology is that meaningful results can be obtained with minimal expense and laboratory sophistication. Many studies begin with dissection and, as argued earlier, highquality descriptive anatomy remains at the core of any morphological analysis. Beyond dissection and description, a variety of techniques are used to reveal increasingly reductionist hierarchical levels of anatomy and to record, measure, and quantify components of function. Many of these techniques are revealed in the individual chapters of this book and it is worth reviewing some of them here. I. Anatomical
Techniques
The art of dissection is little changed since the time of Aristotle, with the exception that tissues and whole organisms can now be fixed and preserved indefinitely for later examination. While vivisection is rarely necessary in morphological studies, occasionally fresh, unfixed tissue is required, as in some histochemical
procedures. Otherwise it is desirable to "stiffen" the tissue through fixation, most often in 10% formalin (formaldehyde solution), which not only allows ample time for one's study, but also facilitates the manipulation of the organs to be dissected, as in the separation of muscles along facial planes. The latter process can be made easier through differential staining of muscle and connective tissue with a topically applied stain (Bock and Shear, 1972). Very fine-scale dissection {microdissection) is possible with finely machined tools and a dissecting scope. Skeletal anatomy is often a central element in morphological studies. It is most often examined from dried skeletons, usually held in the collections of major research museums. Although they can be prepared manually, skeletons, especially those of small, delicate species, are best prepared with colonies of dermestid beetles, which conveniently consume any remaining, dried flesh. If the skeleton is retrieved from the beetles at the appropriate time, a clean, but articulated skeleton results. Study of skeletal anatomy can be enhanced through preparation of dried hone-ligament or honemuscle preparations (Hildebrand, 1968). The skeleton in situ can be studied indirectly with radiography (X-ray pictures). This is most useful for simple measurements such as long bone length and width. Stereo radiographs can be made to study more complex structures, such as the skull. Another very powerful technique for studying skeletons in situ is clearing and staining (Wassersug, 1976; Hanken and Wassersug, 1981). In this technique the surrounding flesh is enzymatically macerated and rendered transparent while the bone and cartilage of the skeleton are differentially stained red and blue, respectively. This method can be extended to include simultaneous staining and visualization of the nervous system (Filipski and Wilson, 1984,1985; Bloot et al., 1985) and the circulatory system (Russell et al., 1988). Minute elements of anatomy are revealed through light and electron microscopy. These techniques are used so widely that they fall into the realm of "standard technique" (e.g., Presnell and Schreibman, 1997), although to do them well usually requires skill learned through long experience. Most light microscope histology is carried out with tissues embedded in paraffin or paraffin-plastic polymers and sectioned on the order of 5 to 20 ^tm. Embedding in harder n\edia, such as epoxy resin, allows for thinner sections: 0.5 to 2 /xm for light microscopy and 0.01 to 0.05 yam for transmission electron microscopy (TEM). Scanning electron microscopy (SEM) reveals only surface features of cells and organs and therefore involves coating and examining whole, three-dimensional forms rather than sections.
1. Tetrapod Feeding in Vertebrate Morphology Both Ught microscopy and TEM require differential staining of cells and organelles in order to visualize the structure of interest. Both techniques dissolve, denature, or otherwise destroy some chemical constituents of the cells or tissues that one might like to reveal. Thus, instead of fixing the tissue chemically, it can be hardened and stabilized through freezing and sectioned on a microtome encased in a freezing compartment (a cryostat). In this way the chemical inclusions of the tissue are more or less preserved for visualization by treating frozen sections with stains that react with the desired compounds (histochemistry). It should be noted that all histological staining is histochemical in this sense, but usually histochemistry refers to frozen tissue techniques, which offer a greater variety and, more important, specificity of staining. 2. Functional
Techniques
Perhaps the most basic, and underrated, functional technique is observation. As His (1888) noted, "observation, though generally well marked in children, is more and more neglected, or even suppressed, by the usual school education." Careful observation of unrestricted animals performing natural behaviors is the first step of any functional analysis. Ideally, such observations are made in the field as well as in the laboratory, although this is remarkably rare in functional studies. Many functional techniques represent little more than enhancements of our powers of observation. First among these is standard photography, which captures and freezes for observation a rapid behavior, particularly if short shutter speeds and stroboscopic illumination are used; 35-mm and larger format film provides high-resolution images. Of course, photography has the disadvantage of being unable to record an entire behavioral sequence (the most rapid 35-mm motor drives are usually capable of no more than three to five frames per second). High-speed cinematography (cine film), in contrast, can capture a complete, rapid behavior (frame rates of 500 per second and higher are possible, although 50 to 250 fps is typical), but the rapid speed of the film through the camera and the small, 8- or 16-mm format reduces the resolution possible for individual frame analysis (few biologists have access to commercial 35-mm cine equipment). Furthermore, short shutter speeds and the chemistry of film emulsions require high light levels for proper exposure. Synchronized, stroboscopic illumination again can greatly enhance the resolution of individual cine frames and reduce the heat generated by photo floods. Cine is only infrequently used now that high-speed videography systems are available. Video has many ad-
15
vantages over film, including very high frame rates (up to 1000 fields per second), long recording times, high photosensitivity (only moderate light levels are necessary), and, in some cases, digital image capture and storage, but it still lacks the single-frame resolution of cine and so is not appropriate for some applications. It also has the virtue of being inexpensive to operate, although initial equipment cost is very high, whereas cine equipment is relatively less expensive, but has a high operational cost. Images generated by photography can be subjected to computer-based image analysis. Many commercial and publicly available (e.g., NIH Image) software packages permit numerous types of measurement (e.g., distance, area, image density) directly from imported images. Similarly, motion or kinematic analysis can be performed on sequential images obtained through cinematography or videography. Most often, the X-Y coordinates of relevant points on sequential images are obtained by digitizing them and these data are used for kinematic analysis. Many examples of kinematic plots are found in the chapters of this book. They show relative motion of desired points (e.g., the tips of the jaws) relative to time so that movements of various parts can be shown in synchrony and quantified throughout an entire behavioral sequence. Kinematic analyses can be coupled to simultaneous electromyography (Loeb and Cans, 1986; Cans, 1992) for information about muscle activity. Most often bipolar electrodes are used and these are inserted through a needle into the muscle of interest. The electrode measures the electrical potential across the dipole and when the muscle is active this voltage spikes. Simultaneous recording of activity in many muscles is possible; in some cases a reference muscle that exhibits rhythmic or otherwise predictable behavior is used to synchronize muscle activity patterns in separate groups of muscles measured in different experiments. Muscle activity patterns (motor patterns) are often illustrated in a summary bar diagram that shows relative onset and offset, and relative activity level in all muscles measured in conjunction with relevant kinematic plots. A variety of electronic transducers are now used to measure various functional parameters. Strain gages convert deformations along their axes into microvoltages, which can be resolved into compression and tension. Other transducers measure force, pressure, flow, etc. A unique, analogue method of goniometry for measuring minute flexion across limited-motion joints was employed by Condon (1987). In this technique a pointer and a protractor were glued to two bones
16
Kurt Schwenk
across a putative kinetic joint in a lizard skull. The technique was able to measure deflections as small as T, far better resolution than that provided by cineradiographic and strain gage techniques used in previous studies. This technique deserves to be exploited more widely. Lauder has introduced the use of Digital Particle Image Velocimetry in the analysis of fish locomotion (Lauder et al, 1996; Drucker and Lauder, 1997). This method is a great advancement over traditional flow visualization techniques, exploiting the use of laser light to illuminate a precise plane of particles in a flow tank, which are recorded on high-speed video. Software permits a three-dimensional reconstruction of fluid direction and velocity over time. This review has touched on only some of the anatomical and functional techniques employed by modern morphologists. There are many others, particularly if one includes specialized forms of light and electron microscopy and neurological techniques; however, this brief overview should provide an adequate introduction to the field and subsequent chapters of this book. III. C O N C L U D I N G C O M M E N T S If nothing else, the preceding overview should reveal that modern vertebrate morphology is a diverse and intellectually vibrant field. It is unique among the biological sciences in its combination of ancient knowledge with cutting-edge technological and conceptual advances. At its best it is the most integrative of sciences, moving fluidly among hierarchical levels and drawing insight from the interplay of disciplines as divergent as molecular genetics and community ecology. It is hard to imagine a more synthetic field, nor one more fundamentally relevant to the the cornerstones of comparative biology: systematics, evolutionary biology, and ecology. Nonetheless, as a discipline it has suffered its share of indignities, periods of professional and institutional stagnation, and even the sneering disregard of its neophytic, reductionist cousins. We have seen the early, impudent promises of molecular genetics to obviate morphology by "answering" the ultimate questions of phenotypic evolution fade as the emergent complexity of genetic, developmental, and functional systems becomes ever more apparent. I therefore conclude this chapter as I began it—by proclaiming that phenotypic approaches to comparative biology are not only deeply interesting, they are essential. If nothing else, the property of emergence requires that this is so. It is fruitless and petty to argue for primacy of one hierarchical level over another. Morphology shows us clearly that organisms are webs of interaction
and integration, not linear chains of cause and effect. Our task is to connect hierarchies, to hlur their boundaries, not to separate them. Morphologists are uniquely poised to contribute in this regard. Dramatic advances will emerge from explicit attempts to integrate topdown and bottom-up approaches in the study of form and function. Acknowledgments I am grateful to Nirvana Filoramo, Ken Kardong, Carl Schlichting, and Giinter Wagner for commenting on the manuscript. Willy Bemis pointed out some useful references. Preparation of the manuscript was supported by grants from the University of Connecticut Research Foundation and the National Science Foundation (NSF IBN-9601173) to the author.
References Aerts, P., and F. DeVree (1993) Feeding performance and muscular constraints in fish. J. Exp. Biol. 177:129-147. Alexander, R. M. (1983) Animal Mechanics, 2nd Ed. Blackwell, Oxford. Alexander, R. M. (1985) The ideal and the feasible: physical constraints on evolution. Biol. J. Linn. Soc. 26:345-358. Amundson, R. (1996) Historical development of the concept of adaptation. Pp. 11-53. In: Adaptation. M. R. Rose and G. V. Lauder (eds.). Academic Press, San Diego. Appel, T. A. (1987) The Cuvier-Geoffroy Debate: French Biology in the Decades Before Darwin. Oxford Univ. Press, Oxford. Arthur, W. (1997) The Origin of Animal Body Plans: A Study in Evolutionary Developmental Biology. Cambridge Univ. Press, Cambridge. Barel, C. D. N. (1993) Concepts of an architectonic approach to transformation morphology. Acta Biotheor. 41:345-381. Barel, C. D. N., G. C. Anker, F. Witte, R. J. C. Hoogerhoud, and T. Goldschmidt (1989) Constructional constraint and its ecomorphological implications. Acta Morphol. Neerl.-Scand. 27:83-109. Bels, V. L., M. Chardon, and P. Vandewalle (1994) Biomechanics of Feeding in Vertebrates. Adv. Comp. Environ. Physiol. 18. SpringerVerlag, Berlin. Bemis, W. E. (1986) Feeding systems of living Dipnoi: anatomy and function. J. Morph. (Suppl.) 1:249-275. Bloot, J., A. B. Boekelaar, G. J. Groen, and B. Baljet (1985) A mesoscopic technique for the study of the development of the peripheral nervous system in rat foetuses. Acta. Morphol. Neerl.-Scand. 23:35-46. Bock, W. J., and C. R. Shear (1972) A staining method for gross dissection of vertebrate muscles. Anat. Anz. 130:222-227. Bock, W. J., and G. von Wahlert (1965) Adaptation and the formfunction complex. Evolution 19:269-299. Bookstein, F. L. (1991) Morphometric Tools for Landmark Data. Cambridge Univ. Press, Cambridge. Bookstein, F. L. (1994) Can biometrical shape be a homologous character? Pp. 197-227. In: Homology: The Hierarchical Basis of Comparative Biology. B. Hall (ed.). Academic Press, San Diego. Bramble, D. M., and D. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Brown, L. (1993) The New Shorter Oxford English Dictionary, Vols. 1 and 2. Clarendon Press, Oxford.
1. T e t r a p o d F e e d i n g in Vertebrate M o r p h o l o g y Biirger, R. (1986) Constraints for the evolution of functionally coupled characters: a nonlinear analysis of a phenotypic model. Evolution 40:182-193. Cole, F. J. (1944) A History of Comparative Anatomy From Aristotle to the Eighteenth Century. MacMillan and Co., London. Condon, K. (1987) A kinematic analysis of mesokinesis in the Nile monitor {Varanus niloticus). Exp. Biol. 47:73-87. Darwin, C. (1859) On the Origin of Species by Means of Natural Selection. John Murray, London (1964 facsimile reprint of the first edition. Harvard Univ. Press, Cambridge, MA). Davis, D. D. (1960) The proper goal of comparative anatomy. Pp. 4 4 50. In: Proceedings of the Centenary and Bicentenary Congress of Biology, University of Malay, Singapore. R. D. Purchon (ed.). Univ. of Malay Press, Singapore. Davis, D. D. (1964) The giant panda: a morphological study of evolutionary mechanisms. Fieldiana: Zool. Mem. 3:1-339. de Jongh, H. J., R. Dantuma, and H. M. J. Sluijsmans (1989) The shape of the mandible in the domestic sheep: a biomechanical analysis using E.M.G. as an estimator of muscle force. Acta Morphol. Neerl.-Scand. 27:63-73. Denny, M. W. (1990) Terrestrial versus aquatic biology: the medium and its message. Am. Zool. 30:111-121. Denny, M. W. (1993) Air and Water. The Biology and Physics of Life's Media. Princeton Univ. Press, Princeton. DiDio, L. J. A. (1986) Anatomy as art and science. Anat. Anz. 161: 197-208. Driesch, H. (1908) The Science and Philosophy of the Organism. Black, London. Drost, M. R., M. Muller, and J. Osse (1998) A quantitative hydrodynamical model of suction feeding in larval fishes: the role of frictional forces. Proc. Roy Soc. Lond. B 234:263-281. Drucker, E., and G. V. Lauder (1997) Aquatic propulsion in fishes by vortex ring production. Am. Zool. 37:77A. Dullemeijer, P. (1959) A comparative functional-anatomical study of the heads of some Viperidae. Morph. Jb. 99: 881-985. Dullemeijer, P. (1974) Concepts and Approaches in Animal Morphology. Van Gorcum, Assen, The Netherlands. Dullemeijer, P. (1980) Functional morphology and evolutionary biology. Acta Biotheor. 29:151-250. Dullemeijer, P. (1989) On the concept of integration in animal morphology. Pp. 3-18. In: Trends in Vertebrate Morphology. H. Splechtna and H. Hilgers (eds.). Gustav Fischer, Stuttgart. Dullemeijer, P. (1994) Conclusion: a general theory for feeding mechanics? Pp. 347-358. In: Biomechanics of Feeding in Vertebrates. Adv. Comp. Environ. Physiol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Emerson, S. B. (1991) The ecomorphology of Bornean tree frogs (family Rhacophoridae). Zool. J. Linn. Soc. 101:337-357. Filipski, G. T., and M. V. H. Wilson (1984) Sudan black B as a nerve stain for whole cleared fishes. Copeia 1984:204-208. Filipski, G. T., and M. V. H. Wilson (1985) Staining nerves in whole cleared amphibians and reptiles using Sudan black B. Copeia 1985:500-502. Frazzetta, T. H. (1962) A functional consideration of cranial kinesis in lizards. J. Morph. Ill:287-319. Frazzetta, T. H. (1983) Adaptation and function of cranial kinesis in reptiles: a time-motion analysis of feeding in alligator lizards. Pp. 222-244. In: Advances in Herpetology, and Evolutionary Biology: Essays in Honor of Ernest E. Williams. Museum of Comparative Zoology, Cambridge, MA. Frazzetta, T. H. (1994) Feeding mechanisms in sharks and other elasmobranchs. Pp. 31-57. In: Biomechanics of Feeding in Vertebrates. Adv. Comp. Environ. Physiol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin.
17
Fung, Y. C. (1993) Biomechanics: Mechanical Properties of Living Tissues, 2nd Ed. Springer-Verlag, New York. Galis, F. (1992) A model for biting in the pharyngeal jaws of a cichlid fish: Haplochromis piceatus. J. Theor. Biol. 155:343-368. Galis, F. (1993) Interactions between the pharyngeal jaw apparatus, feeding behaviour, and ontogeny in the cichlid fish, Haplochromis piceatus: a study of morphological constraints in evolutionary ecology J. Exp. Zool. 267:137-154. Galis, F (1996) The application of functional morphology to evolutionary studies. Trends Ecol. Evol. 11:124-129. Galis, F, and E. G. Drucker (1996) Pharyngeal biting mechanics in centrarchid and cichlid fishes: insights into a key evolutionary irmovation. J. Evol. Biol. 9:641-670. Gans, C. (1969) Functional components versus mechanical units in descriptive morphology. J. Morph. 128:365-368. Gans, C. (1974) Biomechanics: An Approach to Vertebrate Biology. Lippincott, Philadelphia. Gans, C. (1986) Functional morphology of predator-prey relationships. Pp. 6-23. In: Predator-Prey Relationships: Perspectives and Approaches from the Study of Lower Vertebrates. M. E. Feder and G. V. Lauder (eds.). Univ. of Chicago Press, Chicago. Gans, C. (1988) Adaptation and the form-function relation. Am. Zool. 28:681-697. Gans, C. (1992) Electromyography. Pp. 175-204. In; Biomechanics— Structures and Systems: A Practical Approach. A. A. Biewener (ed.). IRL/Oxford Univ. Press, Oxford. Gegenbauer, C. (1878) Elements of Comparative Anatomy, 2nd Ed. MacMillan and Co., London. Gerking, S. D. (1994) Feeding Ecology of Fish. Academic Press, San Diego. Ghiselin, M. T. (1980) The failure of morphology to assimilate Darwinism. Pp. 180-193. In: The Evolutionary Synthesis: Perspectives on the Unification of Biology. E. Mayr and W. B. Provine (eds.). Harvard Univ. Press, Cambridge, MA. Goodwin, B. (1989) Evolution and the generative order. Pp. 89-100. In: Theoretical Biology: Epigenetic and Evolutionary Order from Complex Systems. Edinburgh Univ. Press, Edinburgh (1992 paperback edition by The Johns Hopkins Univ. Press, Baltimore). Gosline, W. A. (1996) Structures associated with feeding in three broad-mouthed, benthic fish groups. Environ. Biol. Fish. 47: 399-405. Gould, S. J. (1966) Allometry and size in ontogeny and phylogeny. Biol. Rev. 41:587-640. Gould, S. J., and R. C. Lewontin (1979) The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc. Roy. Soc. Lond. B 205:581-598. Greaves, W. S. (1995) Functional predictions from theoretical models of the skull and jaws in reptiles and mammals. Pp. 99-115. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Greene, H. W. (1982) Dietary and phenotypic diversity in lizards: why are some organisms specialized? Pp. 107-128. In: Environmental Adaptation and Evolution. D. Mossakowski and G. Roth (eds.). Gustav Fischer, Stuttgart. Greene, H. W. (1986a) Diet and arboreality in the emerald monitor, Varanus prasinus, with comments on the study of adaptation. Fieldiana Zool. New Ser. No. 31:1-12. Greene, H. W. (1986b) Natural history and evolutionary biology. Pp. 99-108. In: Predator-Prey Relationships: Perspectives and Approaches from the Study of Lower Vertebrates. M. E. Feder and G. V. Lauder (eds.). Univ. of Chicago Press, Chicago. Greene, H. W. (1994) Systematics and natural history, foundations for understanding and conserving biodiversity. Am. Zool. 34: 48-56.
18
Kurt Schwenk
Greene, H. W. (1997) Snakes: The Evolution of Mystery in Nature. Univ. of California Press, Berkeley, CA. Hall, B. K. (1994) Homology: The Hierarchical Basis of Comparative Biology. Academic Press, San Diego. Hall, B. K. (1996) Bauplane, phylotypic stages and constraint: why there are so few types of animals. Pp. 215-261. In: Evolutionary Biology. M. K. Hecht, R. J. Maclntyre, and M. T. Clegg (eds.). Plenum Press, New York. Hall, B. K. (1998) Evolutionary Developmental Biology, 2nd Ed. Chapman and Hall, London (distributed by Kluwer Academic). Hanken, J., and B. K. Hall (1993) The Skull, Vols. 1-3. Univ. of Chicago Press, Chicago. Hanken, J., and R. Wassersug (1981) The visible skeleton. Func. Photog. July/August: 22-26,44. Herrel, A., P. Aerts, and F. De Vree (1998) Ecomorphology of the lizard feeding apparatus: a modelling approach. Neth. J. Zool. 48:1-25. Hiiemae, K. M., and A. W. Crompton (1985) Mastication, food transport and swallowing. Pp. 262-290. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Hildebrand, M. (1968) Anatomical Preparations. Univ. of California Press, Berkeley, CA. His, W. (1888) On the principles of animal morphology. Proc. Roy. Soc. Edinburgh 15:287-298. Ho, M.-W., and P. T. Saunders (1979) Beyond neo-Darwinism: an epigenetic approach to evolution. J. Theor. Biol. 78:573-591. Ho, M.-W., and P. T. Saunders (1984) Beyond Neo-Darwinism: An Introduction to the New Evolutionary Paradigm. Academic Press, London. Hopson, J. A., and L. B. Radinsky (1980) Vertebrate paleontology: new approaches and new insights. Paleobiology 6:250-270. Huxley, J. S. (1932) Problems of Relative Growth. Methuen and Co., London (1972 reprint by Dover Publications, New York). Larson, A., and J. B. Losos (1996) Phylogenetic systematics of adaptation. Pp. 187-220. In: Adaptation. M. R. Rose and G. V. Lauder (eds.). Academic Press, San Diego. Lauder, G. V. (1981) Form and function: structural analysis in evolutionary morphology. Paleobiology 7:430-442. Lauder, G. V (1982a) Introduction. Pp. xi-xlv. In: Form and Function: A Contribution to the History of Animal Morphology, by E. S. Russell. Univ. of Chicago Press, Chicago. Lauder, G. V. (1982b) Patterns of evolution in the feeding mechanism of actinopterygian fishes. Am. Zool. 22:275-285. Lauder, G. V. (1985) Aquatic feeding in lower vertebrates. Pp. 210229. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Lauder, G. V. (1990) Functional morphology and systematics: studying functional patterns in an historical context. Annu. Rev. Ecol. Syst. 21:317-340. Lauder, G. V. (1991) Biomechanics and evolution: integrating physical and historical biology in the study of complex systems. Pp. 119. In: Biomechanics and Evolution. J. M. V. Rayner and R. J. Wootton (eds.). Cambridge Univ. Press, Cambridge. Lauder, G. V. (1995) On the inference of function from structure. Pp. 1-18. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Lauder, G. V. (1996) The argument from design. Pp. 55-91. In: Adaptation. M. R. Rose and G. V. Lauder (eds.). Academic Press, San Diego. Lauder, G. V, C. Connon, and D. Dunn-Rankin (1996) Visualization of flow behind the tail of swimming fish: new data using DPIV techniques. Am. Zool. 36:7A. Lauder, G. V, R. B. Huey, R. K. Monson, and R. J. Jensen (1995) Sys-
tematics and the study of organismal form and function. BioScience 45:696-704. Lauder, G. V, and H. B. Shaffer (1993) Design of feeding systems in aquatic vertebrates: major patterns and their evolutionary interpretations. Pp. 113-149. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Liem, K. F. (1990) Aquatic versus terrestrial feeding modes: possible impacts on the trophic ecology of vertebrates. Am. Zool. 30: 209-221. Liem, K. R (1993) Ecomorphology of the teleostean skull. Pp. 422452. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Liem, K. F., and D. B. Wake (1985) Morphology: current approaches and concepts. Pp. 366-377. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Linden, R. A. (1998) The Scientific Basis of Eating. Taste, Smell, Mastication, Salivation, and Swallowing and Their Dysfunctions. Frontiers of Oral Biology Series, Vol 9. Karger, Basel. Loeb, G. E., and C. Gans (1986) Electromyography for Experimentalists. Univ. of Chicago Press, Chicago. Lombard, R. E. (1991) Experiment and comprehending the evolution of function. Am. Zool. 31:743-756. Losos, J. B., and H. W. Greene (1988) Ecological and evolutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35: 379-407. Losos, J. B., and D. B. Miles (1994) Adaptation, constraint, and the comparative method: phylogenetic issues and methods. Pp. 6 0 98. In: Ecological Morphology: Integrative Organismal Biology. P. C. Wainwright and S. M. Reilly (eds.). Univ. of Chicago Press, Chicago. Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and L. Rodriquez-Schettino (1998) Contingency and determinism in replicated adaptive radiations of island lizards. Science 279:21152118. Marcus, L. R, M. Corti, A. Loy, G. J. P Naylor, and D. E. Slice (1996) Advances in Morphometries. Plenun\ Press, New York. Maynard Smith, J., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin, R. Lande, D. Raup, and L. Wolpert (1985) Developmental constraints and evolution. Quart. Rev. Biol. 60:265287. Mayr (1980) Morphology. P. 173. In: The Evolutionary Synthesis: Perspectives on the Unification of Biology. E. Mayr and W. B. Provine (eds.). Harvard Univ. Press, Cambridge, MA. McLellan, T., and J. A. Endler (1998) The relative success of some methods for measuring and describing the shape of complex objects. Syst. Biol. 47:264-281. Motta, P. J. (1988) Functional morphology of the feeding apparatus of ten species of Pacific butterflyfishes (Perciformes, Chaetodontidae): an ecomorphological approach. Env. Biol. Fish. 22: 39-67. Motta, P. J., and K. M. Kotrschal (1992) Correlative, experimental, and comparative evolutionary approaches in ecomorphology. Neth. J. Zool. 42:400-415. Norton, S. F., and E. L. Brainerd (1993) Convergence in the feeding mechanics of ecomorphologically similar species in the Centrarchidae and Cichlidae. J. Exp. Biol. 176:11-29. Nyhart, L. K. (1995) Biology Takes Form: Animal Morphology and the German Universities 1800-1900. Univ. of Chicago Press, Chicago. Oelrich, T. M. (1954) The anatomy of the head of Ctenosaura pectinata (Iguanidae). Misc. Publ. Mus. Zool. Univ. Michigan, No. 94. Often, E. (1989) Mathematical and philosophical reflections on motor control systems. Acta Morphol. Neerl.-Scand. 27:163-169. Owen, R. (1848) On the Archetype and Homologies of the Vertebrate Skeleton. John Van Voorst, London.
1. T e t r a p o d F e e d i n g in Vertebrate M o r p h o l o g y Owen, R. (1849) On the Nature of Limbs. John Van Voorst, London. Owen, R. (1866) On the Anatomy of Vertebrates, Vol. I. Fishes and Reptiles. Longmans, Green, and Co., London. Padian (1995) Form versus function: the evolution of a dialectic. Pp. 264-277. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Piaget, J. (1970) Structuralism. Basic Books, New York. Presnell, J. K., and M. P. Schreibman (1997) Humason's Animal Tissue Techniques, 5th Ed. The Johns Hopkins Univ. Press, Baltimore, MD. Przibram, H. (1931) Connecting Laws in Animal Morphology. Univ. of London Press, London. Rainger, R. (1989) What's the use: William King Gregory and the functional morphology of fossil vertebrates. J. Hist. Biol. 22: 103-139. Raff, R. A. (1996) The Shape of Life: Genes, Development, and the Evolution of Animal Form. Univ. of Chicago Press, Chicago. Rayner, J. M. V, and R. J. Wootton (1991) Biomechanics and Evolution. Cambridge Univ. Press, Cambridge. Reif, W.-E., R. D. K. Thomas, and M. S. Fischer (1985) Constructional morphology: the analysis of constraints in evolution. Acta Biotheor. 34:233-248. Riedl, R. (1978) Order in Living Organisms: A Systems Analysis of Evolution. Wiley, New York. Rieppel, O. (1979) A functional interpretation of the varanid dentition. Gegenbaurs Morphol. Jahr. 125:797-817. Rieppel, O. (1986) Atomism, epigenesis, preformation and pre-existence: a clarification of terms and consequences. Biol. J. Linn. Soc. 28:331-341. Rieppel, O. (1990) Structuralism, functionalism, and the four Aristotelian causes. J. Hist. Biol. 23:291-320. Rohlf, F. J., and F. L. Bookstein (1990) Proceedings of the Michigan Morphometries Workshop. Special Publication No. 2, Mus. of Zool., Univ. of Michigan. Romer, A. S. (1956) Osteology of the Reptiles. Univ. of Chicago Press, Chicago. Roth, B., and D. Wake (1985) Trends in the functional morphology and sensorimotor control of feeding behavior in salamanders: an example of the role of internal dynamics in evolution. Acta Biotheor. 34:175-192. Roth, V. L. (1989) Fabricational noise in elephant dentition. Paleobiology 15:165-179. Roth, V L. (1993) On three-dimensional morphometries, and on the identification of landmark points. Pp. 41-61. In: Contributions to Morphometries. L. F. Marcus, E. Bello, and A. Garcia-Valdecasas (eds.). Museo Nacional de Ciencias Naturales CSIC, Madrid. Russell, A. P., and J. J. Thomason (1993) Mechanical analysis of the mammalian head skeleton. Pp. 345-383. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Russell, A. P , R. L. Walker, and A. M. Bauer (1988) A technique for visualization of the circulatory system in small lizards. Copeia 1988:797-800. Russell, E. S. (1916) Form and Function: A Contribution to the History of Animal Morphology. John Murray, London (1917 American edition, E. P. Dutton and Co., New York). Sanderson, S. L., and R. Wassersug (1993) Convergent and alternative designs for vertebrate suspension feeding. Pp. 37-112. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Sanderson, S. L., J. J. Cech, and M. R. Patterson (1991) Fluid dynamics in suspension-feedingblackfish. Science 251:1346-1348. Schaefer, S. A., and G. V. Lauder (1986) Historical transformation of functional design: evolutionary morphology of feeding mechanisms in loricarioid catfishes. Syst. Zool. 35:489-508. Schlichting, C. D., and M. Pigliucci (1998) Phenotypic Evolution: A Reaction Norm Perspective. Sinauer Assoc, Sunderland, MA.
19
Schmidt-Kittler, N., and K. Vogel (1991) Constructional Morphology and Evolution. Springer-Verlag, Berlin. Schwenk, K. (1994) Why snakes have forked tongues. Science 263: 1573-1577. Schwenk, K. (1995) A utilitarian approach to evolutionary constraint. Zoology 98:251-262. Schwenk, K. (1996) Why snakes flick their tongues. Am. Zool. 36:84A. Schwenk, K. (2000) Functional units and their evolution. In: The Character Concept in Evolutionary Biology. G. P. Wagner (ed.). Academic Press, San Diego. Schwenk, K., and G. P. Wagner (in prep.) The reconciliation of selection and constraint. Schwenk, K., F. A. Jenkins, Jr., and J. Sheen (in prep.) The hydrostatic trigger hypothesis of chameleon tongue projection: an in vivo test. Secor, S. M., and J. Diamond (1998) A vertebrate model of extreme physiological regulation. Nature 395:659-662. Seilacher, A. (1970) Arbeitskonzept zur Konstructions-Morphologie. Lethaia 3:393-396. Seilacher, A. (1973) Fabricational noise in adaptive morphology. Syst. Zool. 22:451-465. Seilacher, A. (1979) Constructional morphology of sand dollars. Paleobiology 5:191-221. Simpson, G. G. (1953) The Major Features of Evolution. Columbia Univ. Press, New York. Singer (1957) A Short History of Anatomy from the Greeks to Harvey, 2nd Ed. Dover, New York. Smith, K. K. (1982) An electromyographic study of the function of the jaw adducting muscles in Varanus exanthematicus (Varanidae). J. Morph. 173:137-158. Smith, K. K. (1993) The form of the feeding apparatus in terrestrial vertebrates: studies of adaptation and constraint. Pp. 150-196. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Smith, K. C. (1992) Neo-rationaHsm versus neo-Darwinism: integrating development and evolution. Biol. Phil. 7:431-451. Stouder, D. J., K. L. Fresh, and R. J. Feller (1994) Theory and Application in Fish Feeding Ecology. Belle W. Baruch Institute for Marine Biology, Univ. of South Carolina Press, Columbia. Thomason, J. J. (1995) Functional Morphology in Vertebrate Paleontology. Cambridge Univ. Press, Cambridge. Thompson, D. W. (1942) On Growth and Form: A New Edition. Cambridge Univ. Press, Cambridge (1992 facsimile reprint. On Growth and Form: The Complete Revised Edition. Dover, New York). van der Klaauw, C. J. (1945) Cerebral skull and facial skull: a contribution to the knowledge of skull structure. Arch. Neerl. Zool. 7:16-37. van der Klaauw, C. J. (1948) Ecological studies and reviews. IV Ecological morphology. Biblio. Biotheor. 4:27-111. van der Weele, C. (1993) Metaphors and the privileging of causes: the place of environmental influences in explanations of development. Acta Biotheor. 41:315-321. Vandewalle, P., A. Huyssene, P. Aerts, and W. Verraes (1994) The pharyngeal apparatus in teleost feeding. Pp. 59-92. In: Biomechanics of Feeding in Vertebrates. Adv. Comp. Environ. Physiol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. van Leeuwen, J. L. (1991) Optimum power output and structural design of sarcomeres. J. Theor. Biol. 149:229-256. van Leeuwen, J. L., and C. W. Spoor (1992) Modelling mechanically stable muscle architectures. Phil. Trans. Roy. Soc. Lond. B 336: 275-292. van Leeuwen, J. L. (1997) Why the chameleon has spiral-shaped muscle fibres in its tongue. Phil. Trans. Roy. Soc. Lond. B 352: 573-589.
20
Kurt Schwenk
Vincent, J. (1990) Structural Biomaterials, 2nd Ed. Princeton Univ. Press, Princeton. Vogel, S. (1988) Life's Devices: The Physical World of Animals and Plants. Princeton Univ. Press, Princeton. Vogel, S. (1994) Life in Moving Fluids. The Physical Biology of Flow, 2nd Ed. Princeton Univ. Press, Princeton. Wagner, G. P. (1986) The systems approach: an interface between development and population genetic aspects of evolution. Pp. 149165. In: Patterns and Processes in the History of Life. D. M. Raup and D. Jablonski (eds.). Springer-Verlag, Berlin. Wagner, G. P. (1988) The influence of variation and of developmental constraints on the rate of multivariate phenotypic evolution. J. Evol. Biol. 1:45-66. Wagner, G. P. (2000) The Character Concept in Evolutionary Biology. Academic Press, San Diego. Wagner, G. P., and L. Altenberg (1996) Complex adaptations and the evolution of evolvability. Evolution 50:967-976. Wagner, G. P., and K. Schwenk (1999) Evolutionarily stable configurations: functional integration and the evolution of phenotypic stability. Pp. 155-217. In: Evolutionary Biology, Vol. 31. M. K. Hecht, R. J. Maclntyre, and M. T. Clegg (eds.). Kluwer Academic/ Plenum Press, New York. Wainwright, P. C. (1991) Ecomorphology: experimental functional anatomy for ecological problems. Am. Zool. 31:680-693. Wainwright, P. C., and A. F. Bennett (1992) The mechanism of tongue projection in chameleons. II. Role of shape change in a muscular hydrostat. J. Exp. Biol. 168:23-40. Wainwright, P C., and G. V. Lauder (1992) The evolution of feeding biology in sunfishes (Centrarchidae). Pp. 472-491. In: Systematics, Historical Ecology, and North American Freshwater Fishes. R. Mayden (ed.). Stanford Univ. Press, Stanford, CA. Wainwright, P C., and S. M. Reilly (eds.) (1994a) Introduction. Pp. 19. In: Ecological Morphology: Integrative Organismal Biology. Univ. of Chicago Press, Chicago. Wainwright, P. C , and S. M. Reilly (1994b) Ecological Morphology: Integrative Organismal Biology. Univ. of Chicago Press, Chicago. Wainwright, P. C , and B. A. Richard (1995) Predicting patterns of prey use from morphology of fishes. Environ. Biol. Fishes 44: 97-113. Wainwright, S. A. (1988) Axis and Circumference: The Cylindrical Shape of Plants and Animals. Harvard Univ. Press, Cambridge, MA. Wainwright, S. A., W. D. Biggs, J. D. Currey, and J. M. Gosline (1976) Mechanical Design in Organisms. Edward Arnold, London. Waisbren, S. J. (1988) The importance of morphology in the evolutionary synthesis as demonstrated by the contributions of the Oxford group: Goodrich, Huxley, and de Beer. J. Hist. Biol. 21:291-330. Wake, D. B. (1982) Functional and evolutionary morphology. Persp. Biol. Med. 25:603-620. Wake, D. B. (1991) Homoplasy: the result of natural selection, or evidence of design limitations? Am. Nat. 138:543-567. Wake, D. B., and A. Larson (1987) Multidimensional analysis of an evolving lineage. Science 238:42-48. Wake, D. B., and G. Roth (1989) Complex Organismal Functions: Integration and Evolution in Vertebrates. Wiley, Chichester. Wake, D. B., G. Roth, and M. Wake (1983) On the problem of stasis in organismal evolution. J. Theor. Biol. 101:211-224.
Wake, M. H. (1991) The impact of functional morphology and biomechanics on studies of evolutionary biology. Pp. 555-557. In: The Unity of Evolutionary Biology: Proceedings of the Fourth International Congress of Systematic and Evolutionary Biology. E. C. Dudley (ed.). Dioscorides Press, Portland. Wake, M. H. (1992) Morphology, the study of form and function, in modern evolutionary biology. Pp. 289-346. In: Oxford Surveys in Evolutionary Biology, Vol. 8. D. J. Futuyma and J. Antonovics (eds.). Oxford Univ. Press, Oxford. Wassersug, R. J. (1976) A procedure for differential staining of cartilage and bone in whole formalin-fixed vertebrates. Stain Tech. 51: 131-134. Webster, G., and B. Goodwin (1996) Form and Transformation. Generative and Relational Principles in Biology. Cambridge Univ. Press, Cambridge. Weibel, E. R. (2000) Symmorphosis: On Form and Function in Shaping Life. Harvard Univ. Press, Cambridge, MA. Weibel, E. R., C. R. Taylor, and L. Bolis (1998) Principles of Animal Design: The Optimization and Symmorphosis Debate. Cambridge Univ. Press, Cambridge. Weishampel, D. B. (1993) Beams and machines: modeling approaches to the analysis of skull form and function. Pp. 303-344. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Weishampel, D. B. (1995) Fossils, function and phylogeny. Pp. 3 4 54. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Westneat, M. W. (1990). Feeding mechanics of teleost fishes (Labridae: Perciformes): a test of four-bar linkage models. J. Morph. 205: 269-295. Westneat, M. W. (1995a) Feeding, fimction, and phylogeny: analysis of historical biomechanics in labrid fishes using comparative methods. Syst. Biol. 44:361-383. Westneat, M. W. (1995b) Phylogenetic systematics and biomechanics in ecomorphology. Environ. Biol. Fishes 44:263-283. Whyte, L. L. (1965) Internal Factors in Evolution. George Braziller, New York. Woodger, J. H. (1929) Biological Principles: A Critical Study. Kegan Paul, Trench, Trubner and Co., London. Wu, E. H. (1994) Kinematic analysis of jaw protrusion in orectolobiform sharks: a new mechanism for jaw protrusion in elasmobranchs. J. Morph. 222:175-190. Zelditch, M. L., R L. Bookstein, and B. L. Lundrigan (1992) Ontogeny of integrated skull growth in the cotton rat Sigmodon fulviventer. Evolution 46:1164-1180. Zelditch, M. L., R L. Bookstein, and B. L. Lundrigan (1993) The ontogenetic complexity of developmental constraints. J. Evol. Biol. 6:621-641. Zweers, G. (1991) Pathways and space for evolution of feeding mechanisms in birds. Pp. 530-547. In: The Unity of Evolutionary Biology: Proceedings of the Fourth International Congress of Systematic and Evolutionary Biology. E. C. Dudley (ed.). Dioscorides Press, Portland. Zweers, G., and J. C. Vanden Berge (1997) Evolutionary transitions in the trophic system of the wader-waterfowl complex. Neth. J. Zool. 47:255-287.
C H A P T E R
2 An Introduction to Tetrapod Feeding KURT SCHWENK
Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut 06269
11.
III. IV. V.
A phylogeny of the Tetrapoda and its relatives is given in Fig. 2.1. All taxon names used in this chapter are in reference to this phylogeny. More detailed phylogenies for particular clades are given in subsequent chapters.
INTRODUCTION A. Feeding Form and Function: Background B. Anatomical Terminology C. Phylogenetic Terminology MORPHOLOGY OF THE FEEDING APPARATUS A. Skull and Mandible B. Teeth C. Keratinous Structures D. Hyobranchial Apparatus E. Jaw Musculature F. Hyobranchial Musculature G. Tongue H. Pharynx I. Cheeks, Lips, and Probosces KINEMATICS OF FEEDING: THE GAPE CYCLE KINEMATICS OF FEEDING: FEEDING STAGES A. Overview B. Stages of the Feeding Cycle CONCLUDING REMARKS References
A. Feeding Form and Function: Background Feeding and excreting are common to all animal life. Feeding is required by virtue of the fact that animals, unlike plants (and other autotrophic organisms), cannot harness directly the radiant energy of the sun—the ultimate source of all energy on Earth. Excretion occurs because in every mouthful of material ingested there is some portion of it that an animal cannot assimilate. Solid matter is excreted via the gut, and following intracellular energy extraction, other molecular detritus is eliminated via the urinary system. Apart from the obvious waste implied by excretion, there is a more subtle inefficiency built into every digestive process—the conversion of energy from one form to another. Food that is processed and digested is ultimately reduced to its molecular components and these are circulated throughout the body via the blood-vascular system. Different food molecules have different fates and even the same molecule can be treated differently depending on the physiological state of the individual, but most food is eventually broken down into glucose molecules. These are transported across membranes into the cytoplasm of every living cell and are then systematically dismantled and oxidized in a process known as cellular respiration.
I. INTRODUCTION This chapter serves as a primer for the study of tetrapod feeding systems. It provides a brief overview of feeding form and function, introducing general concepts and a basic vocabulary. Its goal is to prepare the reader for the more detailed, taxon-based chapters that follow, and for the primary literature in functional and evolutionary morphology of tetrapod feeding.
FEEDING (K.SchwenKed.)
21
Copyright © 2000 by Academic Press. All rights of reproductioi\ in any form reserved.
22
Kurt Schwenk
^^- <^^^-
o^"-"^
GNATHOSTOMATA CRANIATA VERTEBRATA
CHORDATA
^
.x<>"
.^^
/^ .^
\
/
THERIA
MAMMALIA
AMNIOTA
TETRAPODA
FIGURE 2.1. (a) Generally accepted phylogenetic relationships among chordates, including vertebrates. The tetrapod sister taxon, Osteolepiformes, is an extinct group of sarcopterygian fish (e.g., Ahlberg et ah, 1996). Urochordata includes the tunicates, or sea squirts, and Cephalochordata contains the amphioxus, Branchiostoma. Myxini and Petromyzontia are the hagfishes and lampreys, respectively, representing the only extant jawless fishes. Chondrichthians are the cartilaginous fishes and actinopterygians are the bony, ray-finned fishes. Dipnoi are lungfish. (b) Relationships among tetrapod vertebrates. There is some contention about relationships in each of the major clades. Amphibia, Reptilia, and Mammalia (turtles are discussed in the text); however, the phylogeny shown represents a traditional, conservative view.
Cellular respiration releases the energy held in molecular bonds so that it can be used to create new molecules and to drive various chemical reactions constituting the work of the cell. However, not all energy released can be converted into a usable form. Some of it is relinquished as heat. For most ectothermic vertebrates (fish, amphibians, and nonavian reptiles) this small amount of heat is of little consequence—they must depend on extrinsic sources of warmth to elevate their body temperatures above ambient, if at all. Although they might regulate their body temperatures behaviorally during periods of activity, terrestrial ectotherms nevertheless tolerate significant diurnal and
seasonal fluxes. For endotherms (birds and mammals), in contrast, cellular work proceeds fast enough that "waste" heat can be exploited to warm the body core so that a nearly constant high body temperature is maintained. As a consequence, the cellular biochemistry of endotherms has been evolutionarily optimized for these conditions. Indeed, large fluxes in body temperature can be fatal. Primarily physiological, but also morphological and behavioral, mechanisms serve to maintain core temperature within very narrow limits, even when ambient temperature is much lower or higher. These processes are relevant in the context of tetrapod feeding systems in two ways: first, each time energy is converted—initially from solar energy to the chemical bonds within plant cells, then from plant tissue to animal tissue when herbivores feed on plants, and finally, from one type of animal flesh to another type when carnivores feed on other animals—significant amounts of energy are lost in the process. The inherent inefficiency of energy conversion has far-reaching ramifications for the organization of life on earth. At the most basic level it explains, for example, why herbivorous species are far more numerous than carnivorous species. As one ascends trophic (energy conversion) levels, less and less of the solar energy originally harnessed by plants is available to convert into individual animal bodies. In any given food web, therefore, there is far more biomass represented in herbivorous individuals than in carnivorous. This translates into the fact that there are more herbivorous than carnivorous individuals in a given ecosystem, and ultimately there are more herbivorous than carnivorous species of animal. Thus, in considering the functional morphology and evolution of feeding systems among tetrapods, we might expect a greater diversity of herbivorous forms than carnivorous. However, such patterns reflect ecological, not phylogenetic generalities, hence a particular clade may not conform to this expectation, particularly if it derived from a diet-specialized ancestor. Herbivory, for example, is quite unusual among lepidosaurian reptiles, possibly because they descended from an insectivorous ancestor (Chapter 8), whereas in turtles, herbivory is the rule. Certainly many mammalian orders are committed to one particular food type, but if one considers the mammalian or synapsid clade in toto, the predominance of herbivores becomes apparent (Chapter 13). Second, thermal physiology has had significant consequences for the evolution of tetrapod feeding systems. While endotherms might seem to benefit dramatically from a metabolic life in the fast lane, there is a cost—the constant need to fuel high rates of intracellular chemical work. After only hours without
2. An Introduction to Tetrapod Feeding food, the blood glucose levels of an endotherm are exhausted and it begins to metabolize fatty reserves— and eventually, its own flesh—to meet the unceasing demand of its cells for energy. The perpetual need to provision its metabolic machinery requires an endotherm to feed regularly and frequently. Except for a few species that hibernate or aestivate, few endotherms can tolerate lengthy fasting. This problem is exacerbated in some herbivorous species whose high-fiber, low-nutrition forage requires them to spend practically all waking hours procuring and processing food. In contrast, ectotherms need not squander calories solely to maintain a constant, high-body temperature. Thus, ectotherms decrease their daily caloric demand by permitting body temperature to fall during periods of inactivity. Many species can go for days, weeks, or even months without food, virtually shutting down metabolically and entering into a sort of suspended animation until conditions are once again favorable for activity and feeding. This allows them to survive and even to reproduce in places, and through periods of time, no bird or mammal could. Hence, ectothermy is a distinct and cost-effective evolutionary strategy, not merely an inefficient step on the way to endothermy (e.g., Pough, 1983). A consequence of these patterns is that an endotherm must eat far more than a like-sized ectotherm and it must digest its food as efficiently as possible in order to liberate its caloric content quickly and completely. The key to meeting this need is not an increase in digestive efficiency (i.e., the amount of energy liberated per unit mass of food), which is similar in ectotherms and endotherms, but an increase in the rate at which the food is passed through the gut and processed (Karasov and Diamond, 1985, 1988). A critical component of rapid digestion in endotherms is that they initiate significant mechanical and chemical reduction of the food before it enters the stomach. In mammals these adaptations are primarily associated with the teeth and jaws of the masticatory apparatus, and in birds the crop and gizzard serve analogous functions. Ectothermic tetrapods, in contrast, reduce their food comparatively little before it enters the stomach. Indeed, most amphibians and reptiles engulf and swallow food items whole and virtually all digestion occurs chemically within the gut. While some ectotherms (particularly some lizard species) do puncture, slice, or crush their food, they achieve only a paltry level of comminution as compared to mammals whose precise dental occlusion and masticatory orbit allows them to reduce food to a slurry of minute particles and salivary enzymes before it even reaches the esophagus! A profound evolutionary consequence of this fact is that mammalian and, to a lesser extent, avian feeding
23
systems show high dietary specificity in their phenotypes, whereas the feeding systems of amphibians and reptiles generally reflect not what food is eaten, but how it is eaten. In other words, in endotherms there is a relatively tight fit between phenotype and diet, but in ectotherms the fit is loose—a given feeding system accommodates a variety of food types. In sum, the biophysics of energy conversion at the molecular level leads to the basic ecological observation that herbivorous animals are more numerous than carnivorous. The independent evolution of endothermy in birds and mammals was associated in each case with sweeping morphological changes, including adaptations in the feeding apparatus that enhance rapid mechanical and chemical reduction of food before (and after) it enters the stomach. B. Anatomical Terminology Standard anatomical terminology is used freely throughout the book so it is advisable to review it very briefly here (Fig. 2.2). This information is available in any good comparative anatomy text or dissection guide (e.g.. Walker and Homberger, 1992; other citations later). Vertebrate animals are bilaterally symmetrical and most descriptors are defined with reference to the natural planes and axes of the body. Unfortunately (but not surprisingly), most of the earliest anatomical work was concerned with humans (Chapter 1) which are unique in their upright stance and bending of the cranial axis. Thus some human-based anatomical terminology is nonstandard in a comparative context. The standard nomenclature reviewed here is based on quadrupedal animals, and points of disagreement with human-based terms are noted where relevant. "Vertical" and "horizontal" are used with reference to the surface of the earth. Anterior and posterior refer to the head and tail ends of the body, respectively. A limb, for example, might be said to swing anteroposteriorly, meaning from front to back. Sometimes cranial and caudal are used instead. A structure or movement toward the head is craniad, toward the tail, caudad. If one's point of reference is within the head, then rostral (toward the nose) is substituted. A structure on the under or belly side is ventral and one on the upper side, or back, is dorsal. In humans, superior and inferior are used instead of cranial and caudal, and anterior (front) and posterior (back) are substituted for ventral and dorsal. Hence anterior and posterior have completely different meanings in human vs comparative anatomy (Fig. 2.2). Left and right are always with reference to the animal's left and right, not the viewer's (analogous to
24
Kurt Schwenk
Caudal
Anterior
(Posterior)
Transverse plane
Frontal plane
Inferior
F I G U R E 2.2. Standard anatomical terminology for quadrupedal and bipedal tetrapods. See text for discussion. From ''Functional Anatomy of the Vertebrates: An Evolutionary Perspective/' Second Edition, by W. Walker and K. Liem, © 1994 by Saunders College Publishing, reproduced by permission of the publisher.
"stage left" and "stage right"). Relative to the longitudinal axis, lateral means "to the side" and medial means "toward the axis," or midline. If soniething lies exactly in the midline it is median (e.g., median fins as opposed to paired, lateral fins). Distal and proximal are relative terms that mean farther from or nearer to (respectively) a given reference point (often, but not necessarily, the midline of the body). For example, the hand is distal to the upper arm, the upper arm is a proximal limb segment. The longitudinal axis of symmetry also defines several planes of section. A sagittal section is a vertical, longitudinal plane through the axis, or midline, of the body. As such, it divides the body into left and right halves. Although, by definition, a sagittal section is in the midline, it is usually specified as such by referring to it as midsagittal to distinguish it clearly from sections in the sagittal plane but away from the midline, or parasagittal sections. A vertical section that is perpendicular to the sagittal plane, running across the body from left to right, is transverse. A frontal section lies in a horizontal, longitudinal plane. In humans, a transverse section corresponds to the quadrupedal condition in cutting left to right across the axis, but because of the vertical posture it is now a horizontal plane (Fig. 2.2). Similarly, a frontal section is now vertical (if a human were to assume a position on "all fours" the planes would revert to their standard orientation). A section in the frontal plane in humans is usually referred to as a coronal section because it
aligns with the coronal (frontoparietal) suture of the skull roof. With reference to the head, this usage becomes particularly problematic. This is because the evolution of human bipedality was accompanied by a radical reorientation of the cranial axis such that the ancestral (quadrupedal), longitudinal axis of the skull was bent 90°. The facial skeleton retains, more or less, its original orientation, but the skull base is deformed (relative to quadrupeds) so that the spinal cord emerges from the occiput vertically (ventrally), rather than horizontally (posteriorly), to accommodate the upright spine. Thus, a coronal section in a human head would correspond to a transverse section in a quadruped, whereas a transverse section in a human head would correspond to a frontal section in a quadruped! The term "coronal" is only applicable to humans and other anthropoid primates. Its general use in comparative anatomy is strongly discouraged. Elevation refers to raising a structure relative to a reference point and depression to lowering it. Muscles that elevate are called elevators (or sometimes, levators) and muscles that depress are depressors (similarly for the following terms). Flexion refers to bending something or, specifically, to decreasing the angle between two skeletal elements, such as limb bones, whereas extension extends, or increases, the angle. Adduction is movement toward a point of reference (usually the ventral midline of the body) and abduction is movement away from the reference. Jaw closure, for example, is jaw adduction and the muscles that accomplish this are
2. An Introduction to Tetrapod Feeding usually referred to as jaw adductors (occasionally, jaw levators). Opening the lower jaw (mandible) is called jaw depression and the muscles that accomplish this in nonmammalian tetrapods are the mm. depressor mandibulae (the prefix m. is an abbreviation for the Latin term musculus and is used in formal descriptions of muscles; mm. is the plural form, i.e., the depressor mandibulae muscles). Protraction moves something forward, or out (when the tongue is protruded out of the mouth it is protracted), and retraction is the opposite. Rotation refers to movement around an axis, such as jaw rotation about the jaw joint. Twisting rotation around a central, longitudinal axis is axial rotation. C. Phylogenetic Terminology As for anatomy, a phylogenetic vocabulary is necessary to appreciate fully most current evolutionary morphology. In-depth treatments of these and related topics can be found in the following references, but a very brief review of terms is included here for convenience (see Hennig, 1966; Wiley, 1981; Forey et al, 1992; de Queiroz and Gauthier, 1994; Sanderson and Hufford, 1996). A taxon (pL, taxa) is a formally recognized group of organisms representing any hierarchical or categorical level in a classification. The lizard species Varanus niloticus is a taxon, as is the genus Varanus, the family Varanidae, and the suprafamilial group, Varanoidea. In cladistic usage, a taxon is expected to be natural, or monophyletic, but many traditionally named taxa are not (e.g., Invertebrata, or the traditional use of Reptilia, which excludes birds). A monophyletic group, or taxon, is one which includes an ancestor and all of its descendants (hence, the cladistic use of Reptilia includes birds because birds and "typical" reptiles share a common ancestor). A monophyletic group of organisms is also called a clade. A group that includes an ancestor and some, but not all, of its descendants is paraphyletic. The traditional notion of Reptilia again serves as an example: if we trace all living reptiles to their common ancestor, this same ancestor would also be shared by birds. Hence, to exclude birds from the taxon Reptilia would be to define a group that excludes some of the descendants of a common ancestor, i.e., a paraphyletic group. There is disagreement on how best to define a polyphyletic group. It is perhaps most easily thought of as a taxon in which the ancestor is not included (Hennig, 1966). This usually results when a taxon is defined not on the basis of its evolutionary descent, but on the basis of superficial similarity. For example, if the class Mammalia is defined as a taxon containing only those species with a dentary-squamosal jaw joint (as was traditionally done), then this taxon would be polyphyletic because such a jaw joint
25
evolved independently in several synapsid lineages. To make this group monophyletic we would have to trace these separate lineages back to their common ancestor and include it and all intermediate species in Mammalia, but not all mammals thus defined would have a dentary-squamosal jaw joint, the feature regarded as definitive of the taxon. In order to base modern classifications on monophyletic groups, it is preferable that taxa be defined according to their cladistic relationships (i.e., on the basis of a common ancestor and all of its descendants) rather than according to the presence of some "key" character (de Queiroz and Gauthier, 1994). A character is a homologous trait or attribute of a group of organisms (e.g., Wagner, 1996). It is a unit of phenotypic evolution that is developmentally individuated from other characters around it (Wagner, 1999). As such it is assumed that its evolution is largely, or entirely, uncorrelated with the evolution of other characters. This assumption can only be partially true because all characters in an organism are to some extent correlated with all others (e.g., Schwenk, 2000b); however, it is also true that some features of organisms evolve largely independent of change in other features. All phylogenetic analyses begin with a character analysis: identifying characters, documenting their various forms, or states in different taxa, and then determining which state of the character is plesiomorphic {primitive or representative of the ancestral condition) and which states are apomorphic {derived, or newly evolved). The character analysis is done with reference to a particular, monophyletic group of interest, the in-group. "Derived" and "primitive" are relative terms only applicable to character states with reference to a particular in-group. In other words, a character cannot in any absolute sense be considered "primitive"—a form, or state, of the character is only primitive in the context of a comparative analysis applied to a particular ingroup. The nearest clade to the in-group taxon is its sister group, i.e., the in-group and the sister group share a nearest common ancestor. More distantly related taxa are referred to as out-groups. Character states within the in-group are usually determined to be primitive or derived based on comparison to outgroup character states {out-group comparison). The character state inferred to exist at the common ancestor of the in-group and the first out-group defines the plesiomorphic state and other states of the character are, therefore, regarded as derived within the in-group. A derived character state shared by two or more taxa is said to be a synapomorphy. A shared primitive character state in two or more taxa is called a symplesiomorphy. Character similarity that is independently evolved, either convergently or in parallel, is known as homoplasy. Only synapomorphies are taken as evidence
26
Kurt Schwenk
of evolutionary relationship. Sharing a trait implies that the taxa in question share a common ancestor in which the trait evolved uniquely. In contrast, similar appearance stemming from characters that are symplesiomorphic simply means that the character failed to evolve from the ancestral condition. Such similarity gives no information about relationship. For example, the presence of hair is of little use in reconstructing the relationships of an in-group including only living mammals because this trait is symplesiomorphic for the group. However, if our in-group were expanded to included both reptiles and mammals, then hair would be a synapomorphy uniting the mammal species because it evolved once in their common ancestor. Hence, phylogenetic hypotheses, depicted as branching patterns representing putative genealogical relationships {cladograms), are constructed from patterns of synapomorphy only and monophyletic taxa are likewise diagnosed by lists of synapomorphies. Homoplasies potentially confound our ability to discern evolutionary relationships based on character analysis, but a key assumption in phylogenetic analysis is that the simplest solution is the correct one. This is known as the principle of parsimony. In other words, it is assumed that character evolution is a more-or-less direct process that does not usually reverse itself; nor is detailed similarity expected to evolve independently. Thus, a phylogenetic tree (cladogram) that minimizes the number of homoplasies necessary in character evolution is usually the goal of phylogeny reconstruction.
II. MORPHOLOGY OF THE FEEDING APPARATUS As noted in Chapter 1, the functional-morphological approach to vertebrate feeding is primarily concerned with the acquisition, processing, and swallowing of food (see later). Hence, our treatment of morphology is limited to the structures which participate in these functions, for the most part located in the head and neck. Much greater detail for many of these topics will be found in the following chapters; however, not all structures (and functions) reviewed here are relevant to every taxon or functional type. A. Skull and Mandible 1, Origin and
Development
Perhaps no single structure is more quintessentially vertebrate, nor more evocative of the whole, living organism, than the skull. It concentrates in one place most of the feeding apparatus, the braincase and the sensory
capsules, and it is this combination of attributes that makes the skull so revealing about an animal's habits. This is particularly true for mammal skulls whose jaw configurations and dental batteries are especially revelatory about diet and feeding function. The skull is a diagnostic vertebrate (or craniate) character. Indeed, the vertebrate head, or at least its anterior part, is widely regarded as an entirely novel structure built around the contributions of two, new, embryonic, ectodermal tissues, the neural crest and epidermal placodes (Northcutt and Cans, 1983; Gans, 1989,1993). Evolution of a head (a trend called "cephalization") and other novel aspects of vertebrate complexity seem to have been permitted by duplication of the ancestral Hox gene array, as inferred from the condition in Branchiostoma (the amphioxus), a member of the vertebrate sister group, Cephalochordata (e.g., Garcia-Fernandez and Holland, 1994; P. Holland et al, 1992; P. Holland and Garcia-Fernandez, 1996; L. Holland et al, 1996; L. Holland and Holland, 1998) (Fig. 2.1). Gene duplication is an important, if not the most important, mechanism for the introduction of novelty and complexity in evolving lineages (e.g., Arthur, 1997; Nowak et al, 1997), and the Hox (and related) genes are especially important in this regard because their expression controls patterning along the anteroposterior body axis, including the head (e.g., Hanken and Thorogood, 1993). The fact that the tetrapod skull is usually a unitary, structural entity belies its terrific complexity and composite nature (Fig. 2.3). It is, first of all, composed of many individual bones or cartilages that fuse or become sutured together, and these derive from different embryonic and evolutionary sources. Patterns of loss, fusion, and integration among elements vary considerably among lineages so that skull form is widely divergent among tetrapods. Skull is a nontechnical term that usually refers to the braincase and facial skeleton, including the upper jaw, but it must be kept in mind that the total head skeleton also includes the lower jaw and the hyobranchial apparatus (Fig. 2.4). The latter is touched upon here, but is treated more fully in a separate section later. Minimally, morphologists recognize three basic components of the vertebrate head skeleton (Fig. 2.3): the chondrocranium (or neurocranium), splanchnocranium (or viscerocranium), and dermatocranium (or dermal skull). Elements of each are present in the head skeletons of adults in all craniate (vertebrates excluding hagfish; e.g., Janvier, 1996) groups except the lampreys and Chondrichthyes, which, lacking bone in their skeletons, do not form dermatocranial elements. These separate sources of cranial elements are most clearly evident in early developmental stages. In later
2. A n I n t r o d u c t i o n to T e t r a p o d F e e d i n g
a. Chondrocranium
Otic capsule Opisthotic
Prootic
Vertebrae (
Basi-ExOrbitosphenoid Presphenoid Mesethmoid Supraoccipitals Basisphenoid Sphenethmoid (lower tetrapods)
Hyomandibula (
3d-5th arches
Hyoid arch
27
^ Meckel's cartilage
Palatoquadrate
Mandibular arch
Columella (stapes - ^ ^ of mammals) Quadrate (Incus Articular Epipterygoid of mammals) (malleus of mammals)
The chondrocranium forms first in cartilage. It provides a basin for the brain at the anterior end of the notochord and "capsules," which contain most of the special sensory structures (e.g., nose, eyes, ears). In cartilaginous fishes the basin is roofed with cartilage and the sensory capsules are fused to its sides to form a solid, enclosed structure. This chondrocranial structure remains as the adult skull in Chondrichthyes. In bony vertebrates, however, the cartilaginous model of the chondrocranium eventually ossifies (a process known as endochondral ossification). The brain basin is roofed by dermal bone (see later), which also forms most of the facial skeleton and palate. These dermal bones fuse insensibly with the endochondral bones of the chondrocranium to form the unitary skull. Sometimes even sutural junctions between ossifications centers are lost so that an apparently single bone in the adult skull, such as the mammalian temporal bone, is actually a composite structure. In adults of most species the chondrocranial contribution to the skull is most apparent at the occiput, surrounding the foramen magnum and forming the floor and walls of the braincase. h. Splanchnocranium
Vault series Orbital series Temporal series' Palatal series — Mandibular
FIGURE 2.3. A series of ontogenetic stages in the development of the tetrapod head skeleton showing contributions of the chondrocranium (dark shading), splanchnocranium (no shading), and dermatocranium (medium shading). Bones and cartilages derived from each ontogenetic source join and merge to form the adult skull. Reproduced from Kardong (1998), with permission of The McGrawHill Companies.
stages of growth, individual cranial elements become reduced, lost, ossified, fused, or invested by new bone so that the composite origin of the adult skull is obscured. The information reviewed briefly here is covered in greater depth in most comparative anatomy texts (e.g.. Walker and Liem, 1994; Hildebrand, 1995; Kardong, 1998) and in other, more specialized sources (e.g.. Flower, 1885; Bolk et al, 1936; de Beer, 1937; Romer, 1956; lordansky, 1973; Langston, 1973; Trueb, 1973; Moore, 1981; Hanken and Wake, 1982; Duellman and Trueb, 1986; Kuhn and Zeller, 1987; Hanken and Hall, 1993).
The splanchnocranium also forms initially as cartilage. In the early chordate ancestors of vertebrates the splanchnocranium supported the walls of the pharynx and is presumed to have served as a filter feeding basket, as in living ammocoetes (larval lampreys) and amphioxus. It consisted of a series of U-shaped arches (visceral arches) with openings in between through which water was expelled to create a one-way feeding current through the pharynx. It is believed that the visceral arches eventually developed joints and an intrinsic musculature capable of flexing them, putatively to create a more efficient, muscular pump for filter feeding. Gills also became associated with the arches so that water flow through the pharynx served both to transport food and to oxygenate the blood as it passed out the pharyngeal (now gill) slits. In one group of vertebrates, the gnathostomes ("jaw mouths"), the first arch became modified to open and close the mouth, thus effectively transforming the first arch into jaws (Forey and Janvier, 1993; Janvier, 1996). The jaws, or mandibular arch, became attached to the chondrocranium along its upper portion. Although the origin of jaws is traditionally attributed to feeding function and the evolution of a more active, predaceous lifestyle (e.g., Northcutt and Gans, 1983), the initial impetus may have been respiratory in that jaw closure could seal the pharynx to prevent reflux of water
28
Kurt Schwenk
pd
p'a
of
pi pa
frm
FIGURE 2.4. Skull of a gymnophione amphibian {Dermophis mexicanus) shown in dorsal (a), ventral (b), and lateral (c) views. The left mandibular ramus (one-half of the mandible) is shown in lateral (d) and medial (e) views. From Wake and Hanken (1982), Journal of Morphology, copyright © 1982, John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.
during its forceful expulsion posteriorly out the gill slits (Mallatt, 1996). In any case, jaws and, shortly thereafter, teeth became associated with the ability to grasp and consume larger, more active prey, and the ancestral mode of filter feeding was lost in gnathostomes (many living fish and some aquatic tetrapods have reevolved filter feeding systems secondarily). The second visceral arch was modified along with the first. Its upper elements connected the jaw joint to the otic (ear) region of the chondrocranium, bracing or suspending the posterior end of the first arch jaw. Its lower elements supported the floor of the mouth and, through rapid depression, could expand the pharynx to produce suction. More posterior arches supported the gills. The second visceral arch is referred to as the hyoid arch, and those posterior as branchial arches. In living fishes, particularly cartilaginous fishes, the splanchnocranium is little changed from the condition just described. However, in bony fishes many of the arch elements either ossify directly or become overlain by dermal bone (see later). The branchial arches are hidden beneath a large, bony operculum and other superficial, dermal bones of the head skeleton. In
tetrapods, some aquatic (mostly larval) forms retain gills and these continue to be supported by branchial arches. However, in terrestrial tetrapods the splanchnocranium is highly modified and often difficult to detect in adults. Some arch elements are entirely lost, as are portions of others; some ossify into separate bones, some fuse, some are covered over, and virtually all are transformed to serve various functions. The upper and lower jaws, for example, continue to form as cartilaginous, visceral arch elements in embryos, but these are soon ensheathed by dermal bones as development proceeds. The upper jaw cartilage, in particular, is covered over and fused to the ossifying chondrocranium and the dermal bones of the facial skeleton. However, various bits of the visceral arches ossify directly to create small bones, which contribute in various ways to the adult skull (Fig. 2.3). The posterior ends of upper and lower jaw cartilages, for example, ossify into the bones of the jaw joint in all nonmammalian tetrapods (a lower articular and an upper quadrate bone). The dorsal component of the hyoid arch (the hyomandibula) is robust in early tetrapods and putatively served to transmit low-frequency vibrations from the lower jaw
2. An Introduction to Tetrapod Feeding to the ear. During tetrapod evolution the hyomandibula is increasingly reduced in size, becoming more sensitive to airborne, high-frequency vibration, particularly in association with the evolution of a tympanum (ear drum), and is thus modified into a middle ear ossicle (now called the stapes). The ventral component of the hyoid arch fuses, or becomes articulated with, elements of the first and second branchial arches to form the hyobranchial apparatus (tongue and throat skeleton; see later). Other branchial arch elements contribute to the cartilages of the larynx and trachea. In mammals, a new jaw joint evolved between two dermal bones of the skull (the dentary of the lower jaw and the squamosal of the upper) and the original, endochondral jaw joint bones (articular and quadrate) were incorporated into the middle ear along with the stapes to form a chain of three middle ear ossicles (malleus, incus, stapes, respectively) (e.g., AUin, 1975, 1986). Other small ossifications of the first arch cartilages are incorporated into the mammalian skull as well. The evolution of the mammalian middle ear system from components of the ancestral jaws is one of the most dramatic and best documented gradual, morphological transformations in vertebrate history. c. Dermatocranium The dermatocranium is represented by a series of bones that ossify directly without a cartilage precursor. Most dermal ossification is relatively superficial, tending to form sheet-like ("intramembranous") bones and struts that either stand alone or invest chondro- or splanchnocranial cartilages (hence dermal bone is often referred to as "membrane" bone). Most of what gives shape to the adult, tetrapod skull is dermal bone. In mammals, the dermatocranium is a particularly dominant feature of the skull and most cranial adaptation is manifested here. In general the skull can be divided topologically into the braincase and the facial skeleton (van der Klaauw, 1945). This is particularly true for mammals. The facial skeleton comprises the snout and jaws, which are almost entirely dermal. The base of the braincase forms from the chondrocranium, but its sides and roof form from dermal bone. The dermatocranium, and skull generally, can be further subdivided according to various parameters. Some anatomical subdivisions of the skull are considered next. 2. Anatomical
Units
We can decompose the tetrapod skull into a number of anatomical units or, more accurately, topological regions because it is unlikely that these regions are individuated in developmental, functional, or evo-
29
lutionary senses (Schwenk, 2000b; see van der Klaauw, 1945). However, it is convenient to consider them separately because, to some extent, clade-specific adaptations and trends are circumscribed within each. a. Braincase The braincase surrounds and protects the brain. It is derived primarily from the chondrocranium. It is surprising that in many tetrapods, such as lizards, the braincase is not enclosed entirely in bone; much of its inner, anterior portion is membranous. However, in fossorial (burrowing) members of the clade, or in those with a fossorial ancestry, such as snakes, the braincase becomes completely osseous, as it is in most tetrapods. The posterior end of the braincase is perforated by a large foramen magnum, the opening for the spinal cord. One or two occipital condyles protrude lateral or ventral to the foramen magnum and these articulate with the first vertebra, the axis. The ability to rotate the head is a tetrapod novelty, as is a neck, both reflecting enhancement of head mobility characteristic of terrestrial vertebrates. In general, the largest portion of the jaw-closing musculature takes it origin on the side, and sometimes the top, of the braincase, passing ventrally to insert along the lower jaw (see later). In some taxa, this musculature and most of the braincase are obscured by lateral expansions of the dermal roofing bones so that a cavity is formed between the braincase and the dermal roof, which is filled in life by the jaw musculature (e.g., many turtles). The braincase incorporates the otic capsules of the chondrocranium into its base, and this dense, bony region contains the semicircular canals and other sensory structures of the inner ear. Air- or ground-borne vibrations are carried to the inner ear by one (most tetrapods) or three (mammals) tiny ear ossicles that bridge an air-filled cavity called the middle ear. In most tetrapods the ear ossicles are connected laterally to a tympanum (ear drum), but in some species a tympanum is absent and low-frequency vibrations are collected by the lower jaw or the skin at the side of the face (a tympanum evolved several times independently among tetrapods). The middle ear cavity forms from the embryonic first pharyngeal pouch (e.g., Kardong, 1998). Continuity of the middle ear cavity and the pharynx is maintained in most tetrapods (in some lizards light held up to one ear can be seen through the other!), but in mammals the middle ear is a separate cavity, which, in most species, is expanded into a bubble-like chamber called the bulla. The mammalian pharynx is highly modified relative to other tetrapods (see later) and its connection to the middle ear cavity is restricted to a narrow, soft tissue slit (the Eustachian tube).
30
Kurt Schwenk
b. Jaws While the upper jaw is, in most cases, fused to the facial skeleton by dermal bone, the lower jaw is articulated with it at the jaw joint and is referred to as the mandible (sometimes in the bird literature, upper and lower bills are both referred to as mandibles, but this is an idiosyncratic usage). Although the upper jaw is rarely as movable as the lower, in some species with kinetic skulls, independent movement of the upper jaw is possible, which may have significant functional consequences. In typhlopid snakes, for example, each maxilla of the upper jaw can pivot through a significant arc, sweeping ant and termite larvae into the mouth with large, anteriorly directed teeth (Parker and Grandison, 1977; Kley, 1998, unpublished data), and in snakes, generally, the upper jaws are independently movable within the highly kinetic skull (Chapter 9). Most tetrapod jaws contain teeth of various sorts (see later) that form a margin around the mouth. The premaxilla and maxillae of the upper jaws and dentaries of the mandible are the dentigerous (tooth-bearing) bones of the jaws. The jaw joint lies between two splanchnocranial bones, the (lower) articular and the (upper) quadrate, in all tetrapods except mammals in which it is between two dermal bones, the dentary and squamosal. As noted earlier, the articular and quadrate were incorporated into the middle ear of mammal ancestors. The jaw joint is highly variable in position, form, and function, but in most cases it acts primarily as a hinge, permitting vertical rotation of the mandible (depression and elevation). Species whose feeding mode requires limited rotational movement of this type have a joint morphology restricting lateral and anteroposterior deviation of the mandible so that it is virtually 'Tocked" into position (e.g., many carnivoran mammals). However, in herbivorous mammals, some turtles, and several other tetrapod taxa, considerable movement is possible at the joint so that the mandible is capable of more complex motion, including side-to-side rotation and anteroposterior sliding. In rodents, for instance, the mandible slides forward to engage upper and lower incisor teeth for gnawing and then slides back for molar occlusion during grinding. In general, jaw mechanics are far more complex than the simple scissors action implied by a hinge joint (Greaves, 1995). Three-dimensional considerations, such as the height of the jaw joint relative to the tooth row and the fact that chewing usually occurs on one side at a time, have important consequences for the form of the jaw, the nature of the mandibular symphysis, the placement of the teeth on the jaws, the optimal bite point, and the activity of the jaw muscles (e.g.. Bramble, 1978; see Greaves, 1995, for an overview).
Left and right halves of the mandible are joined anteriorly at a joint of varying firmness. In most species this joint is extremely tight and fibrous and called a symphysis. The mandibular symphysis is particularly well developed in mammals (e.g., Scapino, 1965,1981; Beecher, 1979) and well suited to resisting the tortional stresses associated with asymmetrical chewing (Hylander, 1985). In other groups a true symphysis does not form and some movement is possible. This is particularly true for most snakes and some salamanders in which there is a high degree of bilateral asymmetry in feeding movements (e.g., Bellairs, 1984; Cundall et ah, 1987; Elwood and Cundall, 1994). In advanced snakes, left and right halves of the mandible are virtually unattached, allowing them to expand the oral apparatus in order to engulf prey larger than the head (Gans, 1961; Bellairs, 1984; Young, 1998). c. Facial Skeleton The facial skeleton is a system of dermal bones that protrudes anteriorly off the braincase. It comprises the snout, upper jaws, and palate (the latter two are considered further in separate sections) and, in life, contains the eyes and nose. In mammals, the form of the facial skeleton is largely influenced by the latter two sensory systems: nocturnal species tend to have very large, prominent orbits, whereas those species with very sensitive olfactory systems tend to have long snouts. However, these are rough correlations and the shape of the facial skeleton is affected by many other factors. For example, ant- and termite-feeding mammals nearly always acquire an elongate, tapered snout that accommodates a long, slender prehensile tongue, together used to probe nests (Chapter 15). In birds, bill form is indicative of feeding mode (Chapter 12), and although the bill is a keratinous investment of the upper and lower jaws, the facial skeleton largely mirrors the form of the bill (Bock, 1966). Fossorial species of all kinds, such as gymnophione amphibians (Fig. 2.4) and amphisbaenian squamates, exhibit heavily ossified snouts with adjacent bones tightly sutured to withstand the stresses of burrowing (M. Wake, 1993). In contrast, the facial skeleton of most other amphibians and squamates is quite a delicate affair of struts and vacuities. Facial shortening is a prominent feature of anthropoid primate evolution and is noteworthy in a variety of other taxa, such as feliform (cat-like) carnivoran mammals and algae-scraping marine iguanas {Amblyrhynchus). Shortening of the facial skeleton is often related to feeding mode. As noted earlier, the facial skeleton is not a distinctly separate moiety in the adult skull (van der Klaauw, 1945). Posteriorly it is continuous with the dermal skull roof overlying the braincase, and laterally and ventrally, with the endochondral bones of the
2. An Introduction to Tetrapod Feeding braincase, itself, to create the unitary structure of the skull (Fig. 2.4). d. Palate The palate is represented by a series of dermal bones filling in the area between left and right upper jaws (Fig. 2.4b). Dermal bones of the palate join posteriorly with endochondral bones of the braincase. In most taxa the palate is relatively flat, or planar. In all tetrapods the palate is perforated by the internal nares {choanae), openings that allow the passage of air inspired through the nostrils into the nasopharynx (and from there, through the glottis and into the trachea and lungs). In squamate reptiles and many mammals, anterior palatal openings allow access to the vomeronasal organs from the oral cavity (called the incisive foramina in mammals). Many taxa are characterized by additional palatal vacuities; anuran amphibians, for example have a particularly reduced palate dominated by large vacuities ventral to the orbits. The eyeballs are actually retracted through these vacuities into the oral cavity during swallowing (see later)! Palatal form has been exploited as a taxonomic character in birds (e.g., Pycraft, 1900, in Goodrich, 1930; Witmer and Martin, 1987) and squamates (e.g., Lakjer, 1927; Bellairs and Boyd, 1950); however, Bellairs (1949) cautioned that similarity in the superficial (soft tissue) palate may belie significant differences in the hard (bony) palate. Teeth are found on the palates of some amphibians and reptiles. In some taxa a secondary palate forms ventral to the primary palate. As such, the choanae open into a separate chamber above the mouth that exits posteriorly into the mouth or pharynx. The secondary palate forms from medially projecting shelves of dermal jaw and palatal bones. It is best known in mammals where it is complete and robust. It evolved in synapsid ancestors of mammals and is presumed to be correlated with the evolution of endothermy, suckling, and the masticatory apparatus (e.g.. Walker and Liem, 1994; Hildebrand, 1995). It effectively isolates the respiratory passage from the oral cavity, allowing breathing to continue uninterrupted by bouts of feeding. However, initially the secondary palate was incomplete in the midline and is unlikely to have served this function. Probably the initial advantage to medial flanges on the palate was mechanical, serving to resist bending and torsional stresses during biting (Thomason and Russell, 1986; Russell and Thomason, 1993). Also, it should be noted that in mammals a significant part of the secondary palate posteriorly is composed of soft tissues and it is possible that a more extensive soft palate closed off the nasal passage in mammal ancestors with incomplete secondary palates. The presence of a complete, bony secondary palate
31
in crocodilians supports the notion of a respiratory function. Crocodilians forage with their mouths submerged and their nostrils elevated above the water. Because all nonmammalian tetrapods lack lips, cheeks, and other soft tissues that could seal the mouth cavity (see later), a separate respiratory passage allows crocodilians to breath with water-filled mouths—analogous to mammals that breath while suckling or masticating. Less well known is that some turtles (Gaffney, 1979) and scincid lizards (Goppert, 1903; Greer, 1989) have partial or complete bony, secondary palates. In turtles the secondary palate is best developed in aquatic taxa and is thus likely to serve a respiratory function, but in scincids its functional significance is unknown. 3. Temporal Fenestrae Fenestra is Latin for "window" (architects refer to the pattern of windows in a building as fenestration). Temporal fenestrae are windows, or openings, in the temporal region of the dermal roof of the skull. As such, the fenestrae reveal the underlying braincase. They have a characteristic pattern, position and relationship to particular dermal bones in different amniote taxa. Four basic patterns are recognized: anapsid, diapsid, synapsid, and euryapsid (Fig. 2.5). In anapsids (putatively the most primitive) there are no temporal fenestrae. The dermal skull roof forms a solid vault over the braincase with the jaw adductor musculature filling the cavity between the two. Many turtles exemplify this condition, but most living species are derived in having the dermal roof emarginated posteriorly or ventrally so that the braincase is partially, or wholly, exposed (Romer, 1956). Nonetheless, the earliest amniotes were clearly anapsid (e.g., Carroll, 1988). Diapsids are those amniotes with two temporal fenestrae, one above the other. The ventral border of each opening is called an "arch" (in Latin, apsis, hence the names for fenestration patterns actually refer to the number and position of the arches rather than the fenestrae themselves). The ventral border of the upper arch is formed by the squamosal and postorbital bones, and this arch is used as an anatomical landmark for establishing putative homologies in the fenestration pattern. The ventral border of the lower arch is formed by the quadrate and jugal bones. Although primitively both temporal arches are present in diapsids, some derived diapsids have lost the lower (squamates and euryapsids) or upper (birds) arch and their skulls are sometimes referred to as "modified diapsid." Although once considered a separate type, euryapsid skulls are also now believed to represent a modified diapsid condition with a single fenestra corresponding to the upper fenestra of the diapsid skull (i.e., its lower border is formed from the postorbital and squamosal bones).
32
Kurt Schwenk Sq P Po Emargination Modified Modified anapsid testudines diapsid snake
Modified diapsid lizard
Qj Modified Modified diapsid bird synapsid mammal
F I G U R E 2.5. Types of temporal fenestration in tetrapod skulls and a traditional view of skull evolution. Reproduced from Kardong (1998), with permission of The McGraw-Hill Companies. J, jugal; P, parietal; Po, postorbital; Qj, quadratojugal; Sq, squamosal.
Euryapsid taxa comprise a diverse assemblage of extinct, mostly marine reptiles (e.g., ichthyosaurs, plesiosaurs, and pliosaurs). Synapsids are those amniotes with a single temporal fenestra that corresponds to the lower opening of the diapsid condition. This condition is restricted to the clade comprising the mammals and all their extinct relatives forming the sister group to the Reptilia (Synapsida). In most living mammals the synapsid temporal fenestra is difficult to discern. This is because in the evolution of recent mammals two parallel trends have tended to obscure it: the enlargement of the fenestrae and the expansion of the braincase. As a result, the braincase has, in effect, swollen into the vacuity ancestrally present between it and the dermal skull roof while the dermal cranial vault itself has receded. The temporal fenestra can best be visualized in those taxa with well-developed temporal crests and ridges, which serve as attachment sites for the temporalis adductor muscle. The fenestra is evident as the space occupied by this muscle, bordered dorsally by the sagittal crest, posterodorsally by the lambdoidal, crest and ventrally by the zygomatic arch (cheekbone). The pattern of temporal fenestration in amniotes has long been held as indicative of phylogenetic relationship (reviewed by Goodrich, 1930). However, recent cladistic and molecular analyses have thrown into question this venerable contention. Although retained as a formal taxon, Diapsida is diagnosed as having an upper temporal fenestra (among other things), the presence of a lower fenestra being plesiomorphic for this group (e.g., Gauthier, 1984; Evans, 1988). As such, Diapsida includes Euryapsida within it. Monophyly
of the euryapsid skull condition itself is in question (Rieppel, 1994); however, some studies have allied ichthyosaurs with plesiosaurs and their relatives, thus retaining a monophyletic taxon Euryapsida within the Diapsida. Most surprising, perhaps, is the suggestion that turtles are also derived from within Diapsida (Rieppel, 1993, 1994; Rieppel and deBraga, 1996; deBraga and Rieppel, 1997; Platz and Conlon, 1997; Zardoya and Meyer, 1998; Hedges and Poling, 1999). If true, the traditional notion of a monophyletic Anapsida uniting turtles with fossil forms (e.g., Gaffney and Meylan, 1988) would be shattered. Furthermore, the anapsid condition of the turtle skull, long considered a retention of the primitive amniote condition, would represent a reversal from a diapsid ancestor, i.e., the loss of a fenestrated skull and reevolution of an uninterrupted dermal vault! However, at the time of this writing the character evidence supporting this contention is in dispute and molecular data are being reanalyzed (e.g., Wilkinson et a/., 1997; Lee, 1997; M. Lee, personal communication), hence the phylogenetic position of turtles and, therefore, evolutionary patterns of temporal fenestration remain open questions. The functional basis for the origin of temporal fenestration is unknown, but it is generally thought to be related to the distribution of stress on the dermal skull roof generated by the jaw adductor musculature (Olson, 1961; Frazzetta, 1968). According to Olson (1961 : 214), "the skull tended to become a network of lines of stress which were narrowly confined and precisely delimited." Attachment of muscle to bone is putatively more effective on edges and ridges than on smooth surfaces, particularly at sutures, hence there may have
2. An Introduction to Tetrapod Feeding
been initial selection for the formation of ridges on the underside of the skull roof. As adductor origin became increasingly restricted to marginal ridges, there would have been consequent selection to reduce the deposition of bone in intervening areas to conserve both weight and metabolic cost. Both these putative evolutionary trends are paralleled by plastic responses of bone in living individuals—bone tends to be deposited in areas of stress and resorbed in areas lacking it. Carroll (1988) further suggested that the typically rounded form of temporal fenestrae might serve to dissipate stress energy around the periphery of the cheek region to prevent the development of cracks. 4. Cranial Kinesis Cranial kinesis refers to any type intracranial mobility. As such, at least one component of the cranium is capable of moving relative to another, implying the presence of joints or flexure zones within the skull (Fig. 2.6). Cranial kinesis is widely distributed among vertebrates, but systems of kinesis are diverse and Parietal unit
Snout unit
FIGURE 2.6. Putative system of cranial kinesis in a lizard {Varanus) modeled as a system of joints and linkages. Reproduced from Kardong (1998), with permission of The McGraw-Hill Companies. Based on Frazzetta (1962).
33
should not be regarded as homologous. An intracranial joint was present in the sarcopterygian fishes ancestral to tetrapods (retained in the coelacanth) (reviewed by Jarvik, 1980), but was lost in the transition to tetrapods (Ahlberg et ah, 1996). Early tetrapods remained akinetic, a condition retained in living amphibians (with the possible exception of some caecilians; Wake and Hanken, 1982; J. O'Reilly, personal communication), but many amniote groups evolved some form of kinesis (see Goodrich, 1930). In most groups the inception of cranial kinesis was directly related to the evolution of temporal fenestration, which transformed the ancestral, massively bony, anapsid skull into a system of slender struts and vacuities. The significance of this transformation is most evident in diapsid reptiles, such as the living squamates, that have lost the lower temporal arch. In such modified diapsid skulls the quadrate is no longer fixed in place but is free to swing anteroposteriorly, a condition known as streptostyly (Versluys, 1912; Robinson, 1967). Streptostyly adds, in effect, an additional element to the lower jaw that allows anteroposterior translation of the mandible (Fig. 2.6). In advanced snakes the quadrate becomes very long and slender and can swing laterally, as well as anteroposteriorly, as part of the mechanism that permits an exceptionally wide gape for engulfing very large prey relative to body size (Chapter 9). In squamates, additional types of kinesis have evolved that allow the snout and facial skeleton to move relative to the braincase and the upper jaw to flex dorsoventrally (e.g., Frazzetta, 1962; lordansky, 1966; Arnold, 1998; see Chapter 8). In advanced snakes, kinesis can be extreme, including independent movement of left and right sides of the upper jaws and palate and flexure of the anterior snout (e.g., Albright and Nelson, 1959a,b; Kardong, 1977; Cundall and Shardo, 1995). Similarly, in birds, skull kinesis is variable and often extensive (e.g.. Bock, 1964; Biihler, 1981; Zusi, 1993). Much of the kinesis in birds relates to movement of the upper jaw, including independent movement of the tip. The skulls of mammals and their synapsid antecedents are singularly akinetic due to extensive fusion of the facial skeleton to the braincase and the evolution of a secondary palate. Nonetheless, it has been argued that some therapsid synapsids (ancestral to living mammals) had a connective tissue joint between the braincase and the facial skeleton that acted as a shockabsorbing device. Loss of this kinesis in living mammals is associated with expansion of the braincase (Crompton, 1955). In modern hares a zone of connective tissue divides the braincase into two, potentially kinetic moieties. Bramble (1989) suggested that this joint absorbs impact energy when the front legs hit the ground during saltatory locomotion. Inertial
34
Kurt Schwenk
movement of the large, extended pinnae (external ears) "resets" the mechanism during hind limb push-off. Caution must be exercised in the interpretation of cranial kinesis because its presence in the vast majority of cases is inferred only from anatomical evidence, such as the histology of putative joints or the manipulation of dead specimens. Functional demonstrations of kinetic movements in living animals are rare. Consequently, the functional significance of cranial kinesis in most groups remains unclear. However, with the exception of the hare example given earlier, nearly all scenarios identify kinetic mechanisms with feeding function. In birds, for example, kinesis increases gape for in-flight scooping of insect prey (see Zusi, 1993) and in other cases it may function as part of a mandibular lock important for shock absorption (e.g.. Bock, 1964; van den Heuvel, 1992). In lizards, simultaneous movement of upper and lower jaws may produce a pincer apparatus that enhances the efficacy of prey prehension with the tips of the jaws (Frazzetta, 1983; Condon, 1987; Schwenk, 1987) or it may aide flattening of the skull in crevice-dwelling forms (Arnold, 1998). B. Teeth Teeth, like the skull itself, are a vertebrate innovation associated with the presence of neural crest tissue. They are integumentary derivatives composed of several hard tissues (enamel, dentine, and cementum), but their origin is unclear. Traditionally they were thought to have arisen from hard tissues found in the dermal body armor of ancient, jawless fishes, but the presence of teeth in the even more ancient, soft-bodied conodonts, now believed to be basal vertebrates, throws this hypothesis into question. The discovery of dentine (a neural crest derivative) in conodont teeth supports the contention that conodonts were vertebrates and that their teeth are homologous to those of other groups (Sansom et al, 1994). Typically, vertebrate teeth have an exposed crown of very hard enamel with an inner core of somewhat softer dentine (Fig. 2.7a). The tooth is attached to a skeletal element by a layer of cementum. Tooth form, number, and position are highly variable among tetrapods. In most taxa teeth are restricted to the marginal elements of the jaws, primarily the premaxilla and maxillae of the upper jaw and the dentary of the lower jaw (in bony vertebrates). Many amphibians and lepidosaurian reptiles have palatal teeth as well. In gymnophione (caecilian) amphibians, for example, there are two upper rows of teeth, one marginal and one palatal (Fig. 2.4b). The teeth of the lower jaw fit in between those of the upper (Duellman and Trueb, 1986), probably enhancing the shearing function of a bite. The up-
per teeth of the lepidosaur, Sphenodon, are similarly arranged with two rows between which fit the lower teeth (Edmund, 1969). During chewing the lower teeth are slid longitudinally between the upper tooth rows for effective shearing (Gorniak et al, 1982). In advanced snakes the maxillary teeth are reduced and palatal teeth on the pterygoid bones are used unilaterally to pull the head of the snake over the prey item. In venomous viperids the maxilla is very short and contains a single, elongated fang (plus several replacements) that is erected when the maxilla is rotated during a strike. The fang is hollow and serves as a conduit to inject venom, forcefully expelled from the venom gland, into the prey (Chapter 9). Crocodilians and mammals lack palatal teeth, but many mammals have transverse, cornified ridges on the hard palate called palatal rugae that function analogously to hold food items in place during mastication and tongue movement. Turtles and birds lack teeth entirely and have replaced them with homy (keratinous) bills or beaks (e.g., Homberger, 1999; see later). Some frogs and many mammals also lack teeth. In all cases, the loss of dentition is a secondarily derived condition. In mammals, tooth reduction is typically associated with ant- and termitefeeding as part of a suite of adaptations enhancing lingual prehension of prey (Chapter 15). Filter-feeding baleen whales have also lost their teeth, as have some "toothed" whales (Chapter 16). In nearly all living amphibians the tooth crown is separated from the base, or pedicel, by a constricted zone of uncalcified dentine or connective tissue (Fig. 2.7g). Such teeth are called pedicellate (Parsons and Williams, 1962; Duellman and Trueb, 1986). Most amphibian teeth are conical and unicuspid, although bicuspid and even spatulate teeth occur. All living crocodilians are carnivorous and have conical teeth deeply rooted in the jaw bones {thecodont; Fig. 2.7d) (Edmund, 1969); however, a fossil species has been described with multicuspate, molariform teeth, suggesting a herbivorous diet (Wu et al, 1995). Lepidosaurian reptiles have a variety of tooth types. Most attach to the medial sides of the jaws {pleurodont; Figs. 2.7b and 2.7e), but in Sphenodon, some iguanians (Agamidae), and some amphisbaenians (Trogonophidae) the teeth are cemented to the apex of the jaws {acrodont; Figs. 2.7c and 2.7f) (Edmund, 1969; Cans, 1960). Cusp shape is highly variable. Carnivorous and insectivorous species tend to have conical, unicuspid crowns while in herbivorous species they are usually spatulate and multicuspate (Hotton, 1955; Edmund, 1969). Species whose diets include hard prey (durophagy) typically have teeth with broad, flat, or rounded crowns (e.g., Dalrymple, 1979; Rieppel and Labhardt, 1979). Tooth form and phylogeny are often closely correlated in squamates so not
2. An Introduction to Tetrapod Feeding
35
enamel dentine
bone pulp cavity
. .•;;
dental ligament cement a.
FIGURE 2.7. Tetrapod teeth, (a) Section through mandible and tooth of a mammal illustrating basic tooth anatomy; from Wake (1979), with permission; (b-d) types of tooth attachment in tetrapods: pleurodont, acrodont, and thecodont, respectively; from Wake (1979), with permission; (e) medial view of left mandibular ramus in a lizard (Morunasaurus) with pleurodont dentition; from Estes et ah (1988), with permission; (f) upper right tooth row of a lizard (Uromastix) with acrodont dentition; from Edmund (1969), with permission; (g) pedicellate teeth of a salamander (Amphiuma); from Duellman and Trueb (1986), with permission; and (h) lateral view of dentition in a female chimpanzee (Pan) in full occlusion; note that the teeth vary in form along the tooth row (heterodonty) and that uppers and lowers fit together very precisely (occlusion); from Owen (1868). Parts a - d copyright © 1979, University of Chicago Press. Reprinted from Hyman's Comparative Vertebrate Anatomy, 3rd ed.
all dental differences necessarily reflect adaptation to diet. In most amphibians and reptiles the teeth are polyphyodont, meaning that they are replaced throughout
life as they are worn or broken. Upper and lower teeth do not meet precisely, i.e., they do not occlude as they do in mammals, although in some amphisbaenians
36
Kurt Schwenk
upper and lower teeth do mesh to create a potential "cookie-cutter" shearing mechanism (Gans, 1960, 1974). Mammalian teeth are extremely variable and especially indicative of diet. This is because, unique among tetrapods, mammals masticate their food. Mastication requires a precise alignment of upper and lower tooth rows for the teeth to be mechanically effective in grinding, crushing, or slicing. Consequently their form is very sensitive to the mechanical attributes of the food processed (e.g., Hiiemae and Crompton, 1985; Strait, 1997). So that this alignment, or occlusion (Fig. 2.7h), is maintained, many (eutherians) or most (marsupials) teeth are never replaced. Some are replaced (the deciduous or milk dentition), but only once when suckling is completed and the oral apparatus nears adult size {diphyodonty). In most mammals, teeth have specialized functions along the tooth row, hence the teeth in a single animal vary significantly in form {heterodonty; Fig. 2.7h)—anterior incisors nip, crop, or gnaw, canines puncture and hold, premolars puncture and slice, and molars crush or grind (teeth with similar form or function in nonmammalian taxa are sometimes referred to as "incisiform," "caniniform," or "molariform"). In carnivorous, most insectivorous, and omnivorous mammals, molar tooth crowns are entirely covered with enamel, but in many herbivorous species, apical surfaces of the molars are dentine or dentine traversed by ridges of enamel. The softer dentine wears faster then the enamel so each tooth forms a corrugated grinding surface. Similarly, the gnawing incisors of rodents lack enamel on their posterior side so that they wear unevenly front to back, creating a chisel-like cusp. Mammal teeth are deeply rooted (thecodont). C. Keratinous Structures In the edentulous turtles and birds the jaws are covered by a keratinous or horny investment known as the rhamphotheca. In birds the rhamphotheca is called either the bill or the beak and in turtles it is called a beak. The beak is derived from the outer layer of dead, keratinized cells characteristic of all tetrapod integument (e.g., Homberger and Brush, 1986). In both taxa, beak form largely conforms to the shape of the bony jaws, but in some birds it extends significantly beyond the jaw tips. In others, keratinous ornaments occur seasonally during courtship and reproduction, but are later shed (Bock, 1966; Gaffney, 1979; Welty and Baptista, 1988). In both groups, the horny beak wraps around the jaw margins and extends onto the palate and floor of the mouth, especially anteriorly, to form a triturating surface for processing food (Fig. 2.13b). In turtles the beak is hard with a very sharp apical edge. When ele-
vated, the mandible fits snugly inside the upper jaw so that the lateral surface of the lower beak slides past the inner surface of the upper in a scissors-like shearing action. Some crushing of food may occur inside the mouth between the triturating surfaces. In birds, beak form is extremely variable among species and is closely correlated with diet (Fig. 2.8) (Bock, 1966; Ziswiler and Farner, 1972; Welty and Baptista, 1988; Witmer and Rose, 1991; Gill, 1995). There are beaks designed for probing, spearing, crushing, shearing, prying, and filtering. In spoonbills the cross-sectional shape of the broad, flat bill forms a hydrofoil that sheds a vortex as it sweeps through the water, disturbing the substrate and lifting buried prey items (Weihs and Katzir, 1994). In mallard ducks, flamingos, and some other aquatic birds the lateral margins of upper and lower bills form a series of lamellae that interact with each other and the tongue to form a filtering device (Jenkin, 1957; Zweers et al, 1977). In some shorebirds, the narrow bill is used to pluck minute prey out of the water like forceps and to move it along the beak to the mouth within a droplet of water using a unique mechanism of surface tension transport (Rubega and Obst, 1993). In many birds the beak is horny and hard, as in turtles, but in others, part or all of the beak is only lightly keratinized and remains relatively soft and pliant (Ziswiler and Farner, 1972; Gill, 1995). Furthermore, the beak and palate of many species are well supplied with proprioceptors (100-400 per mm^) and, in some cases, taste buds so that the beak is an extremely sensitive tactile and chemosensory organ used to locate, recognize, and manipulate food items (e.g., Berkhoudt, 1985a,b; Zweers and Berkhoudt, 1991; Dubbeldam, 1992). The so-called bills and beaks of monotreme mammals (platypus and echidnas) are only superficially bird-like. They are covered by naked skin that is not exceptionally keratinized (Griffiths, 1978). However, the tongues of all three living species of monotreme possess heavily keratinized pads and spines that abrade against similar horny structures on the palate (Owen, 1868; Doran and Baggett, 1972; Griffiths, 1978). These keratinous structures function as tooth analogues to process food. In the long-beaked echidna (Zaglossus), the anterior part of the tongue is also endowed with a series of barbed, keratinous spines that lie within a groove at rest (Griffiths, 1978). When the tongue is extended the spines are exposed and used to impale and apprehend slimy earthworm prey. The tongues of many birds and mammals are also covered by keratinous spines and projections, although none as well developed as in monotremes. In felid mammals, for example, heavily keratinized, spinous lingual papillae are used to rasp the flesh off bones, and in many nectar-feeding birds and mammals the tongue tip is elaborated into a
37
2. A n I n t r o d u c t i o n to T e t r a p o d F e e d i n g
MERGANSER Fish grasper
EAGLE
OYSTER
Moot fearer
CATCHER
Mollusc opener
WOODCOCK Earth probe
HUMMINGBIRD Flower probe
WOODPECKER Wood cutter
RAVEN Generolized
bill
CARDINAL
CROSSBILL
Seed cracker
Pine seed extroctor
FIGURE 2.8. Diversity of beak form in birds and its association with feeding types (see Chapter 12). Bony elements of the jaws largely, but not entirely, correspond in shape to their keratinous investment. From ' T h e Life of Birds,'' Second Edition, by J. C. Welty, © 1975 by Saunders College Publishing, reproduced by permission of the publisher.
filamentous brush (Figs. 2.13d and 2.13e) (e.g., Richardson et al, 1986; Homberger, 1988; Gill, 1995; Birt et ah, 1997). In many birds the underside of the tongue tip is heavily keratinized and stiff. This may serve to return the tongue tip to its original shape after deformation during feeding (Homberger and Brush, 1986; Homberger, 1988). Baleen whales derive their name from keratinous structures known as haleen, which they use to filter small prey from within or at the surface of the water (Chapter 16). The baleen grows as transverse plates suspended from the palate on each side of the midline (Pivorunas, 1979). The plates overlap and are frayed along their terminal margins to form a dense, comblike mat of fibers. Although they grow continuously, they are worn away during use. Baleen form corresponds to both phylogeny and feeding mode. Species with very long baleen plates (e.g., right whales) surface skim, whereas species with shorter baleen (e.g., blue whales) typically engulf large schools of prey beneath the surface (Pivorunas, 1979; Lambertsen, 1983). Water is expelled from the mouth, through the baleen, either by compression of an expanded pharynx and hollowed
tongue as in some species or by the piston-like action of a large, muscular tongue in others. In either case, minute prey are trapped in the frayed baleen, which acts as a sieve (Chapter 16). In some artiodactyl mammals (e.g., deer and antelope), there are no upper incisors. Instead, the underlip and anterior palate are covered by a thick, keratinous pad that the lower incisors contact to crop vegetation (e.g., Owen, 1868). Sirenians (manatees and dugongs) have thick keratinous pads at the ends of both upper and lower jaws, which, at least in some cases, are used to crop and process vegetation (e.g., Forsten and Youngman, 1982; see Chapter 16). Rhinoceroses lack both upper and lower incisors and some species have heavily keratinized lips used for cropping (Nowak, 1991). As noted previously, keratinous palatal rugae serve as anchor points for food during processing and tongue movement in most mammals. D . Hyobranchial Apparatus As noted previously, the hyobranchial apparatus (or hyobranchium) is the most obvious remnant of the
38
Kurt Schwenk
visceral skeleton (splanchnocranium) in tetrapods. It is derived from the ventral elements of the second (hyoid) visceral arch and as many as three branchial arches (visceral arches 3, 4, and 5). It lies within the tissues of the throat (pharynx) with its anterior part usually between the mandibular rami where it supports the tongue in the floor of the mouth (Fig. 2.9). The hyobranchial apparatus is often referred to, imprecisely, as the hyoid apparatus, or simply the hyoid. This usage is most appropriate for mammals, in which the branchial arch contribution is reduced. In fishes the
hyobranchial apparatus supports the gills and is an essential part of the suction-generating mechanism by means of pharyngeal expansion. Visceral arches also form an inner set of "pharyngeal jaws" in some fishes (notably cichlids) that bear teeth and process food (e.g., Liem, 1973; Liem and Sanderson, 1986), and in filterfeeding groups the branchial arches form part of the food-trapping system (Northcott and Beveridge, 1988; Sanderson and Wassersug, 1993). In tetrapods, only larval amphibians and some aquatic, paedomorphic adults retain branchial arch
ptat proc
B hyoid plate I—pmed proc
I—hyoid plate pmed proc plat proc
phyd bone
— den -Mc c (mand sym)
^ angsp '—hyog memb '—ant proc hyl Mc c
-angsp -antlai proc -hyale -hyog memb -den -mmc (mand sym)
vocal slit
FIGURE 2.9. Spatial relationships among the mandible, hyobranchial apparatus, and tongue in two species of frog, shown in anterodorsal view. The hyobranchium lies embedded in the throat musculature, its anterior portion between the mandibular rami. In most tetrapod taxa, as here, the hyobranchial skeleton supports and moves the tongue to which it is attached by muscle and connective tissue. However, the degree to which the tongue can move independent of the hyobranchium varies a great deal among tetrapod taxa. On the left is Rhinophrynus dorsalis (A and C); on the right is Bufo marinus (B and D). After Trueb and Gans (1983), with permission of Cambridge University Press.
39
2. An Introduction to Tetrapod Feeding gills. Larval salamanders, in particular, maintain an extensive set of arches. However, in adult amphibians and all other tetrapods the hyobranchial apparatus functions primarily to support and move the tongue and pharynx. In some secondarily aquatic tetrapods the hyobranchium has reverted to its role in pharyngeal expansion to generate suction. In these species the hyobranchial apparatus is typically large and robust (Bramble and Wake, 1985). A respiratory function for the hyobranchial apparatus is maintained in many tetrapods lacking gills, particularly in salamanders that use a "buccal p u m p " to ventilate the lungs (e.g., Brainerd and Monroy, 1998). Some lizards and other reptiles also use a buccal pump for lung ventilation (Owerkowicz ei a\., 1999; E. L. Brainerd, unpublished results) and to move air in and out of the nasal chamber for olfaction (Dial and Schwenk, 1996; Schwenk and Brainerd, in preparation). Gular or buccal movements and pharyngeal expansion are also associated in squamate reptiles with various displays and with evaporative cooling (e.g., Bels, 1992; Deban ei a\., 1994; Bels ei a\., 1995). Extreme examples of hyobranchialmediated display behavior occur in iguanian lizards, notably the erection of the dewlap in anoloid lizards (Font and Rome, 1990; Bels, 1990), the frill in the frilled lizard (Beddard, 1905), and the beard in the bearded dragon (Throckmorton ei a\., 1985). Hyobranchial morphology is diverse (Fiirbringer, 1922) (Fig. 2.10) and the names for individual elements vary among clades and workers. Indeed, homologies among elements in different taxa are not always certain and are sometimes controversial (e.g., Reilly and Lauder, 1988). However, there is always an anterior, median element derived from the hyoid arch, usually referred to as the body or basihyal The basihyal lies at the base of the tongue and is sometimes embedded within it. In some taxa, such as lepidosaurs and birds, the basihyal has an anterior extension that projects into the tongue. In lepidosaurs this lingual process is a slender rod, but in birds the tongue is more or less filled by an additional, apparently neomorphic cartilage called the paraglossal (ossified in some species, notably parrots; Homberger, 1986,1999). A first pair of hyoid arch elements curves laterally from the basihyal. These are usually called ceraiohyals. Subsequent (more posterior) arches are derived from ancestral branchial arches. They are given a variety of names, but their proximal elements are usually called hasibranchials and their apical elements, epibranchials. Arch elements can be fused, articulated with, or separate from the basihyal, depending on the arch and the taxon. Elements of the hyobranchial arches usually extend toward the larynx and the otic region of the skull base. Embryologically the association of the stapes to the second (hyoid) arch
is clearly evident. In many mammals hyobranchial elements form ligamentous or bony chains that attach the hyobranchium to the larynx and the styloid process of the skull (Hiiemae and Crompton, 1985; Homberger, 1994, 1999). A reduction in the branchial arch contribution to the hyobranchium in mammals is correlated with their greater contribution to the cartilages of the larynx. This increase in laryngeal complexity may be associated with more complex sound production, an important part of the sociobiology of mammals (Walker and Liem, 1994). With regard to mammals, it is typical that hyobranchial terminology is distinct and confusing relative to other tetrapods. In most mammalian species there are only two arches represented (Fig. 2.10i). The anterior arch corresponds to the ceratohyals, which are second visceral arch derivatives. These are referred to as the anierior cornua (horns). A second arch represents a basibranchial (third visceral arch) contribution called the posierior cornua (or thyrohyals). In most mammals the anterior cornua are prominent and comprise a chain of bones with a ligamentous connection to the styloid process of the skull base. The posterior cornua are smaller and associated with the larynx. However, in humans and other anthropoid primates it is the posterior cornua that are prominent while the anterior cornua are reduced to nubbins to which the stylohyoid ligament attaches (e.g., Shipman ei al, 1985), thus the anterior cornua are referred to as the "lesser horns" and the posterior cornua as the "greater horns" (Fig.2.10j). In some taxa virtually all tongue movement is a result of hyobranchial movement (e.g., most birds). However, in tetrapods with complexly muscled tongues (lepidosaurs, some turtles, parrots, and mammals), the tongue is also capable of complex, intrinsic movement independent of hyobranchial movement (see later). Many tetrapods capture prey by protruding or projecting the tongue out of the mouth. Tongue protrusion is usually a combination of hyobranchial protraction, intrinsic shape change of the tongue, and, sometimes, sliding of the tongue on the lingual process of the basihyal. During tongue protraction in salamanders, most or all of the hyobranchial apparatus is projected out of the mouth along with the tongue (Chapter 4), whereas in chameleons the lingual process of the basihyal is protruded and stabilized while the tongue is projected off of it (Chapter 8) (see later). E. Jaw Musculature As visceral arch derivatives, the jaws are moved by the muscles of the visceral arches, collectively known as the branchiomeric musculaiure. A deeply entrenched
40
Kurt Schwenk
FIGURE 2.10. Diversity in the form of the hyobranchial apparatus in amphibians (a-c), reptiles (d-h), and mammals (i-j). (a) Caecilian (Gymnophiona, Gymnopis); from Duellman and Trueb (1986), with permission; (b) frog (Anura, Leptodactylus); from Duellman and Trueb (1986), with permission; (c) salamander (Urodela, Amides); from Duellman and Trueb (1986), with permission; (d) hzard (Lepidosauria, Basiliscus); from Romer (1956), with permission; (e) snake (Lepidosauria, Vipera); from Romer (1956), with permission; (f) turtle (Testudines, Clemmys); from Romer (1956), with permission; (g) bird (probably Picaformes); from Proctor and Lynch (1993), with permission; (h) crocodile (Crocodylia, Crocodylus); from Romer (1956), with permission; (i) dog (Camivora, Canis); from Flower (1885); and (j) monkey (Primates, Papio); from Flower (1885). All hyobranchia are shown in ventral view with anterior toward the top, except for (i), which is an anterior view. Parts a—f, h copyright © 1956, University of Chicago Press. Reprinted from Osteology of the Reptiles.
notion in vertebrate morphology is that the branchiomeric musculature is "viscerar' in the sense that it is associated with the gut tube (pharynx) and therefore represents an anterior extension of unsegmented musculature investing the gut (the hypomere) (e.g., Goodrich, 1930; see Kardong, 1998; Walker and Liem, 1994). Recent developmental and neuroanatomical work, however, has cast doubt on this dogma and suggests the possibility that the branchiomeric muscula-
ture has a segmental origin from the paraxial mesoderm of the cranial somitomeres rather than deriving from the unsegmented hypomere (Northcuttt, 1990; D. Wake, 1993; see Jacobson, 1993, for a discussion of somitomeres). In the trunk the hypomere forms involuntary, smooth muscle of the gut and visceral organs, but in the visceral arches the branchiomeric musculature is striated and mostly voluntary. Furthermore, the branchiomeric musculature is innervated by the
41
2. A n I n t r o d u c t i o n to T e t r a p o d F e e d i n g
cranial nerves, which are serially homologous with the more posterior, segmented spinal nerves innervating the striated body musculature derived from somitic mesoderm. Thus, we can recognize serial, presumably segmental homologies in the branchiomeric muscles that move the visceral arches. The jaw musculature includes muscles that open and close the jaws and, in some cases, protract and retract them (Fig. 2.11). This apparent functional simplicity is belied by the remarkable variation and complexity of the jaw musculature among tetrapods. Ancestrally, and as retained in many fishes and amphibians, a large, relatively simple adductor mandibulae muscle elevates
the mandible to close the jaws (all muscles are bilaterally paired unless otherwise noted). It arises from the side of the braincase and upper jaw and inserts on the mandible. In gymnophione amphibians an additional large (and functionally more important), ventral interhyoideus muscle inserts on an exceptionally long retroarticular process behind the jaw joint, pulling it down to elevate the mandible (Bemis et al., 1983). In some plethodontid salamanders the mandible is connected by a stout ligament, over the skull, to the anterior cervical vertebrae. Because the ligament crosses the neck joint it is stretched when the head is sharply ventroflexed, pulling up on the mandible to effect a
iVI. temporalis
aponeurosis covering M temporalis
M. zygomaticomandibularis
M digastricus M masseter pars profunda /
3,^,10-be|,y
posterior belly
pars superficialis aponeurosis covering
M. zygomaticomandibularis
M temporalis M. pterygoideus extern us
d.
postenor belly M digastncus anterior belly
M. pterygoideus internus
FIGURE 2.11. Jaw musculature in a herbivorous mammal (Artiodactyla, Ovis). (a-c) Progressively deeper views in left lateral view; (d) a ventrolateral view with the left mandibular ramus removed except for its condylar process (shown in section next to the insertion of the m. pterygoideus externus). Note the pinnate form of the temporalis muscle and the highly complex and subdivided masseter complex (many workers would also include the m. zygomaticomandibularis as part of the masseter). Also note that the combination of jaw muscles supports the mandible in a "sling/' which can be moved in virtually any direction. The digastric muscle (a and b) inserts on the anteromedial surface of the mandibular ramus and serves to open the jaw. After Turnbull (1970), with permission.
42
Kurt Schwenk
powerful bite (Schwenk and Wake, 1993). In this case it is large, anterior trunk muscles ventroflexing the head that contribute to jaw closing. In arrmiotes, the adductor mandibulae is variously subdivided. In reptiles the adductor mass is internally partitioned into many fascicles or nominal muscles of varying degrees of separation and confluence, achieving its greatest complexity in squamates (e.g., Haas, 1973; Wineski and Gans, 1984). These muscles insert onto a vertical process on the mandible formed by the coronoid bone. In addition, a wholly separate muscle (m. pterygoideus) arises from the base of skull and forms a sling around the posterior end of the mandible, extending onto the retroarticular process. As one might imagine, the adductor musculature is particularly complex in snakes with highly kinetic skulls (e.g., Haas, 1973; Zaher, 1994). In mammals the principal adductor mass is relatively simple and is called the m. temporalis (Fig. 2.11). It lies in the position typically filled by the adductor mandibulae and inserts onto the coronoid process of the dentary bone (mandible). Mammals also have a well-developed pterygoideus muscle and a novel adductor (putatively differentiated from the temporal adductor mass) called the m. masseter, which takes its origin from the zygomatic arch. The masseter muscle is often apportioned into several, largely separate parts, particularly in rodents and other herbivorous mammals that rely on complex masticatory movements to grind plant food (e.g.. Fig. 2.11). Each of the three main adductor muscles in mammals has a different line of action, including anterior, posterior, lateral, and medial components, so that the lower jaw is held completely within a muscular sling capable of producing the finely controlled movements necessary for dental occlusion and mastication (TurnbuU, 1970; Crompton, 1989). The arrangement of the jaw adductors in mammals also helps to control reaction forces at the jaw joint during biting, a condition that may have permitted the reduction and eventual loss of the ancestral jaw joint bones to the middle ear (Crompton and Hylander, 1986; Crompton, 1989,1995). The apparent simplicity of mammalian adductors grossly is often in startling contrast to their high degree of internal, histological complexity. Most adductors are characterized by complex patterns of pinnation and some (particularly the masseter) are subdivided into many, independently controlled motor units, resulting in fine-scaled functional partitioning of the muscle (e.g.. Herring, 1980,1992). In all nonmammalian tetrapods the jaws are opened primarily by the m. depressor mandibulae, which arises from the back of the skull and neck and lifts the retroarticular process behind the jaw joint to depress the mandible. Although it acts on the jaw (first visceral
arch), the depressor mandibulae is derived from the ancestral hyoid (second) arch levator to which it corresponds topologically, if not functionally. In nonmammalian tetrapods, jaw opening is also probably aided by strap-like muscles in the ventral midline. These arise from the hyobranchial apparatus, which is itself connected by serially homologous muscles to the sternum. These muscles are hypohranchial, rather than branchiomeric, developing as anteroventral extensions of more posterior somites (striated, segmented body musculature). In mammals the depressor mandibulae is lost and the lower jaw is depressed primarily by the digastric muscle running ventrally from the back of the braincase to the mandible. Although the brief description just given summarizes significant complexity in the jaw musculature, it remains a gross simplification. For example, taxa with kinetic skulls have additional muscles that affect upper jaw movement independent of the braincase. As noted, in snakes that use unilateral movements of the upper jaw elements and which erect fangs, this musculature is intimidating indeed (e.g., Kochva, 1962; Haas, 1973; Kardong, 1973; Cundall, 1983,1987; Cundall and Rossman, 1993; Zaher, 1994). Cranial elevation at the atlanto-occipital joint contributes to mouth opening in many tetrapods, in addition to jaw depression, hence dorsal axial muscles inserting on the cranium may also function as "jaw muscles" (e.g., Bemis et ah, 1983; Larsen and Beneski, 1988). As noted earlier, head depression affected by ventral axial muscles in some salamanders causes jaw closing by means of a tendinous link between the mandible and the first vertebra (Schwenk and Wake, 1993). F. Hyobranchial Musculature The hyobranchial musculature is sufficiently complex and variable that relevant details are left for subsequent chapters. However, of note here is that, for the most part, the hyobranchial apparatus lacks direct connections to other skeletal components and is therefore free-floating within the musculature of the throat. In addition to its own intrinsic musculature that moves its various elements relative to one another, it has an extrinsic musculature that connects it to the mandible anteriorly, the skull dorsally, and the shoulder and sternum posteriorly (Fig. 2.12). Thus the hyobranchium is held within a three-dimensional array of muscular slings that permit movement in virtually any direction. The kinematic potential of this arrangement is particularly important regarding the hyobranchium's association with the tongue. In nearly all tetrapods the tongue is directly coupled to the basihyal with varying degrees of intimacy and most gross tongue movement
2. A n Introduction to Tetrapod Feeding
43
FIGURE 2.12. Muscles acting on the hyoid body of an opossum (Marsupialia, Didelphis) represented as vectors. The hyobranchium is thus suspended in the throat region by a series of paired, straplike muscles that can move it in any direction with precise control. Due to its connection to the tongue, hyobranchial movements usually translate into lingual movements. In addition, several extrinsic lingual muscles arise from the hyobranchium and insert within the tongue. From Crompton et al. (1977), with permission. A Dig, m. digastricus, anterior belly; GnGl, m. genioglossus; GnHy, m. geniohyoideus; OmHy, m. omohyoideus; P Dig, m. digastricus, posterior belly; StHy, m. sternohyoideus.
is a direct result of hyobranchial movement. Fine control of hyobranchial-mediated tongue position is particularly important during rhythmic feeding cycles in which hyolingual movements are used to transport and position food in the mouth and pharynx (see later). In addition to tongue movement, the hyobranchial musculature is largely responsible for pharyngeal compression associated with swallowing (see later). Movement of hyobranchial elements in display behavior usually involves the action of its intrinsic musculature, whereas gross hyolingual positioning, as during feeding, is the result of extrinsic muscle function. In species with highly protrusible or projectile tongues the hyobranchial musculature is often highly specialized, particularly in salamanders (Lombard and Wake, 1976, 1977; Chapters 3 and 4). G. Tongue Tongue form is enormously variable among tetrapods (Fig. 2.13) and its diversity is largely congruent with phylogeny rather than ecology (e.g., Schwenk, 1988; Homberger, 1989). Hence, superficial tongue morphology has been exploited in taxonomic and phylogenetic treatments of several groups (e.g., salamanders: Regal, 1966; Ozeti and Wake, 1969; Lombard and
Wake, 1977,1986; frogs: Regal and Cans, 1976; Horton, 1982; Grant et al, 1997; squamates: Camp, 1923; McDowell and Bogert, 1954; Schwenk, 1988; birds: Gardner, 1925; Bock, 1978; Homberger, 1980; mammals: Sonntag, 1925; Stadtmiiller, 1938b; Doran, 1975; see also specific chapters). The tongue is an essential part of the feeding apparatus of terrestrial tetrapods. It is used as a prehensile organ during the ingestion of food in most amphibians, most reptiles (including birds), and many mammals. It almost universally functions to manipulate food once it is in the mouth and to transport it posteriorly to the esophagus for passage into the gut. Hence the tongue is important in virtually every stage of feeding (see later). Within some clades, among-taxon differences in tongue morphology can largely be attributed to differences in the functional role played by the tongue during each feeding stage (Schwenk and Throckmorton, 1989). In aquatic tetrapods that have reverted to suction or filter feeding, the tongue is secondarily reduced and simplified, and in suction feeders the hyobranchial apparatus is large and robust due to its role in expanding the pharynx (Bramble and Wake, 1985). In many tetrapods the tongue is relatively simple. In some taxa it is little more than an epithelial elaboration
44
Kurt Schwenk
Tongue anchored to posterior end of stemtnn
FIGURE 2.13. Diversity of tongue form in tetrapods. (a) Salamander {Salamandra), showing mucous pad and pleated mucosa of the floor of the mouth indicative of the ability of the tongue to be protruded far beyond the mandible for lingual prey capture; from Stadtmiiller (1938), with permission; (b) tortoise {Testudo), showing papillose surface of the tongue; also shown is the triturating surface of the lower jaw covered in the hard keratin of the beak, which forms sharp cutting edges along its margins; from Stadtmuller (1938), with permission; (c) lizard (Cordylus), showing scale-like and plicate papillae, and notched tongue tip typical of squamates; from Schwenk (1995), with permission; (d) nectar-feeding bat (Choeronycteris), showing cylindrical form typical of highly extensible tongues, which employ a muscular-hydrostatic mechanism for elongation; also shown is a keratinous ''brush'' tip used to collect nectar; from Stadtmuller (1938), with permission; (e) nectar-feeding bird (Coereba), showing curled margins and bifurcate, brush tip; from Orinthology by Gill, © 1990, 1995 by W. H. Freeman and Company, used with permission; and (f) myrmecophagous mammal (Tamandua), showing an extremely long tongue originating on the sternum and hypertrophied salivary glands; from Feldhamer et al. (1999), with permission of The McGrawHill Companies.
overlying a median element of the hyobranchium. This is most evident in birds in which the tongue is essentially a superficial, largely keratinous adornment of
the paraglossal cartilage. In crocodilians the tongue is more substantial, but it is mostly fatty, lacks intrinsic musculature, and is attached to the floor of the mouth along its entire length (Schumacher, 1973; Ferguson, 1981), hence it provides little more than a pad used to seal the pharynx during submersion. Thus, both living archosaurian clades show lingual simplification, although this is undoubtedly independent and secondarily derived. Most frogs similarly lack intrinsic lingual musculature, although their tongues are highly protrusible (Chapter 5). In caecilian amphibians the tongue is an nonprotrusible muscular pad in the floor of the mouth that, at least in some species, is riddled with vascular sinuses (M. Wake and Schwenk, unpublished results). Salamanders, likewise, have tongues that are primarily mucous pads (Fig. 2.13a), but unlike caecilians, salamanders are characterized by extensive tongue protrusion and even ballistic tongue projection used for prey prehension (Chapter 4). Most of the morphological complexity of this system, however, is associated with the hyobranchial apparatus and its musculature extrinsic to the tongue. In contrast to the taxa just noted, terrestrial turtles, lepidosaurs, parrots, and mammals are characterized by highly complex, muscular tongues. Lepidosaurs and mammals, in particular, have large, intrinsically mobile tongues. In most lepidosaurs the tongue is supported by a median process of the basihyal {lingual, or entoglossal process), but in mammals there is no such support and the basihyal is restricted to the very base of the tongue. Muscles that contribute to the corpus of the tongue are typically divided into two types: extrinsic fibers originate from outside the tongue (usually from the mandible and elements of the hyobranchium) and insert within it and intrinsic fibers arise and insert entirely within the tongue. This simple dichotomy is belied by the complex organization of the fibers within the tongue where both types of fiber sometimes merge into a single, organized system so that identification of a particular lingual "muscle" as either "intrinsic" or "extrinsic" is problematic (Schwenk, 2000a). In any case, such muscular complexity reflects the unusual biomechanical properties of such tongues that function as muscular hydrostats (Kier and Smith, 1985; Smith and Kier, 1989). Tongues are similar in this sense to elephant trunks and squid tentacles, which are also complexly muscled organs lacking skeletal support. Such organs exploit the incompressibility of water (in this case intracellular fluid within muscle cells) in a constant volume structure to effect length and shape changes. Although the term "muscular hydrostat" was only coined recently, this biomechanical attribute of the tongue was recognized long ago by Owen (1868), if not earlier, who described
2. An Introduction to Tetrapod Feeding it in reference to the prehensile giraffe tongue. In all muscular hydrostats there are some muscle fibers oriented longitudinally, and others transversely, obliquely or vertically. The latter type serve to decrease the tongue's diameter which, due to the maintenance of a constant volume, translates into an increase in length. In addition, localized muscle contractions can cause complex shape changes in the tongue. Mammal tongues are especially protean in this regard. Tongue shape is maintained overall by an elastic, connective tissue tunic that may also help restore the tongue to its resting conformation through elastic recoil (Schwenk et al, manuscript). Although virtually all lepidosaur and mammal tongues are capable of some hydrostatic shape change, some groups, such as snakes and myrmecophagous mammals, show a high degree of lingual specialization for hydrostatic elongation (Fig. 2.13f). Parrots are unique among birds in their possession of a large, mobile tongue containing extensive intrinsic musculature (Homberger, 1986). They also contain cavernous vascular sinuses that may function hydraulically. These neomorphic modifications of the tongue are associated with parrots' unique feeding mechanism, specifically their ability to manipulate seeds for processing by the bill. Turtle tongues (Fig. 2.13b) are little studied, but terrestrial turtles are capable of significant tongue movement, including limited tongue protrusion for food prehension. Their tongues are supported by a robust basihyal with a small lingual process and include some intrinsic musculature (Schumacher, 1973). Although frogs generally lack intrinsic lingual musculature and therefore the ability to cause hydrostatic volume changes in the tongue, one exceptional species has been found. An African ranoid, Hemisus marmoratum, was shown to use an unusual form of slow, linear tongue extension during prey capture distinctly different from the typical anuran pattern of "flipping" the tongue over the mandibular symphysis (Ritter and Nishikawa, 1995). A morphological study found that Hemisus is unique among frogs in its possession of an intrinsic, vertical muscle fiber system, which is presumed to decrease the diameter of the tongue to effect hydrostatic elongation (Nishikawa et al, 1999). Hemisus is an ant and termite specialist. A similar system may operate in another, unrelated myrmecophagous frog, Rhinophrynus (Trueb and Gans, 1983), although this is unconfirmed. The association of hydrostatic elongation during lingual feeding in myrmecophagous tetrapods is striking: as noted earlier, ant- and termitefeeding mammals exhibit extreme modifications of the tongue in this regard (Fig. 13f), and even among lizards, which as a rule show little dietary specialization of the tongue (Schwenk, 1988), ant-eaters such as Phry-
45
nosoma and Moloch show adaptations for hydrostatic tongue protrusion (Schwenk and W. Sherbrooke, in preparation). In most species the dorsal surface of the tongue is elaborated into a series of papillary projections. These are highly variable in form and function (Fig. 2.13). In mammals, for example, they are usually quite keratinized (see earlier discussion), whereas in many squamates they are soft, filamentous, and covered with a simple mucous epithelium. In taxa that use the tongue for prehension of food, the tongue surface is usually very papillose and always well endowed with mucous or seromucous secretions. These fluids originate either intrinsically, as noted earlier, or from extrinsic salivary glands. The salivary glands of myrmecophagous species are usually hypertrophied (Fig. 2.13f), which may reflect their role in producing copious mucus for prey prehension, or may be related to the production of secretions that digest, detoxify, or immobilize noxious ant prey (which may bite, sting, and contain both formic acid and venom). In snakes, papillae are lost and the tongue surface is smooth, presumably due to exclusive use of the tongue in chemosensory tongue flicking with concomitant loss of tongue participation in feeding function. In addition to its role in the mechanics of feeding, the tongue also serves as a sensory organ. In particular, the tongue is the site of gustation, or taste, a chemical sense mediated by small organs in the lingual (and oral) epithelium known as "taste buds" (e.g., Gilbertson, 1998). In mammals, taste buds are restricted to specialized fungiform, circumvallate, and foliate papillae (e.g., Murray, 1973), but in other tetrapods they are distributed more generally (e.g., Schwenk, 1985). In frogs they take the form of large disc-like taste organs (Sato, 1976; Barlow, 1998). Lingual taste buds are a plesiomorphic feature of tetrapods, hence their absence in some species, such as snakes and monitor lizards (Schwenk, 1985; Young, 1997), is unusual. However, in these taxa loss of lingual taste buds probably follows secondarily on extreme modification of the tongue to subserve the vomeronasal system, a second chemical sense related to olfaction and mediated by paired vomeronasal organs in the snout above the palate (Halpern, 1992; Schwenk, 1994, 1995). In this case the tongue plays no direct sensory role, but functions to sample environmental chemicals and to delivery them to the vomeronasal organs. In most tetrapods the tongue tip is also probably utilized in tactile exploration of the environment (e.g., Kubota et al, 1962). Sensory use of the tongue is important during foraging in many taxa, especially squamate reptiles, and in the assessment of palatability of food once it is brought into the mouth (e.g.. Chapters 8 and 13).
46
Kurt Schwenk H. Pharynx
The pharynx is an ill-defined portion of the gut tube that lies between the mouth {buccal cavity) and the esophagus. Its anterior margin corresponds roughly to the posterior end of the tongue. In mammals, particularly, it is sometimes subdivided into an anterior oropharynx, a dorsal nasopharynx, and a laryngopharynx dorsal to the larynx. However, in most tetrapods these distinctions are unwarranted. This is because in nonmammalian species the pharynx is often a capacious cavity that blends indistinguishably with the buccal cavity anteriorly and the middle ear (tympanic) cavities laterally. In such species the pharynx is better described by the inclusive term oropharyngeal cavity (e.g.. Smith, 1992). In nonman\malian species the pharynx serves as a collection site for food that is transported there inertially or by the tongue and hyobranchial apparatus. Food does not remain long here before being moved further posterior into the esophagus for peristaltic transport to the stomach. Swallowing in nonmammalian tetrapods is properly called "pharyngeal emptying." In mammals, swallowing takes a special form and is referred to as "deglutition" (see later). Of particular relevance here is the condition of the pharynx in mammals (Smith, 1992). Unlike all other vertebrates, mammals possess a muscular pharynx that is associated with the evolution of facial musculature, a soft palate, modifications of the tongue, and several novel behaviors, notably suckling in infants and an explosive, muscular swallow that is integrated into the feeding cycle. Smith (1992) noted that the mammalian muscular pharynx seems to be entirely neomorphic, without homologues in related tetrapods. The muscular pharynx and its neuromuscular control are part of a sweeping reorganization of the feeding system associated with the evolution of the mammalian masticatory system. In effect, mammals have inserted a muscular sleeve into the ancestral pharynx that is continuous posteriorly with the esophagus and anteriorly with the soft palate and extrinsic muscles of the tongue. Thus in mammals the entrance to the pharynx (the pyriform recess) is marked by a sphincter-like orifice of muscles that run from the tongue base to the soft palate and from the soft palate into the pharyngeal musculature (Smith, 1992). Contraction of these muscles elevates the base of the tongue and creates a seal between the tongue and the soft palate. Creation of this seal may be critical to generating suction during suckling (Chapter 14). Masticated food collected in the pharynx is then swallowed during a reflexive contraction of the pharyngeal musculature coordinated with posterior tongue movement (e.g., Koltai and Gates, 1981; Miller, 1982; Hiiemae and Crompton, 1985; Crompton, 1989; Thexton and Crompton, 1998;
see later). The nature of deglutition in mammals, as well as its coordination with the rest of the feeding cycle, is unique to mammals (Smith, 1992). I. Cheeks, Lips, and Probosces In addition to the novel origin of pharyngeal musculature, mammals also uniquely possess facial muscles that endow them with facial mobility and their characteristic use of facial expression in communication (e.g., Darwin, 1896). Unlike the pharyngeal musculature, facial muscles have clear homologues in their expressionless kin, the reptiles (including birds), which have a set of thin, sheet-like, superficial muscles (mm. constrictor colli, intermandibularis posterior, and depressor mandibulae) innervated by the same cranial nerve (VII, the facial) (e.g.. Ruber, 1930; Smith, 1992). These muscles primarily elevate and constrict the pharynx in reptiles (helping to empty the pharynx, among other things), but in mammal ancestors they migrated onto the front and sides of the face where they function not only in facial expression, but in feeding as well. Of most importance in mammalian feeding are the cheeks and lips (e.g., Stadtmiiller, 1938a). The presence of these structures is a significant departure from other tetrapods because they provide walls and a flexible, anterior seal to the buccal cavity. This is in striking contrast to amphibians and reptiles whose mouths must gape when the jaws are open. The ability to seal the mouth permits therian mammals to generate intraoral suction, which is a fundamental part of the suckling mechanism used by infants (Chapter 14). Although monotremes lack lips and cheeks, Griffiths (1978) maintained that echidnas also suckle, and do not lap, their milk, and was convinced that they can generate intraoral suction for this purpose. It is unknown how platypus infants imbibe their milk, but some observations suggest suckling also. Owen (1868) noted that, as infants, monotremes have fleshy lips "transitorily manifested . . . at the suckling period." Some adult mammals, such as walruses and many cetaceans, also exploit suction during feeding (e.g., Gordon, 1984; Kastelein et al, 1994) (Chapter 16). In addition to helping seal the mouth during suction, the cheek (buccinator) muscles also function during mastication to maintain food between the upper and lower tooth rows for processing. Tongue and cheek movements coordinate to move food from one side to the other (side switching) during mastication (Hiiemae and Crompton, 1985). In some whales an expansive upper "lip" forms a lateral flap covering the baleen, overhanging the mandible when the mouth is closed (Owen, 1868). Finally, as noted previously, a mobile, prehensile upper lip is used by rhinoceroses
2. An Introduction to Tetrapod Feeding and other mammals for prehension of grasses and leaves. Some other structures to which the facial muscles contribute are significant in mammal feeding. Notable among these is the elephant trunk. Elephant trunks are muscular hydrostats that are stiffened and moved without the aid of internal skeletal support (Kier and Smith, 1985; Smith and Kier, 1989; see earlier discussion of the tongue). They are used as prehensile organs to grasp food items and to bring them into the mouth. The trunk terminates in one (Asian elephant) or two (African elephant) finger-like projections that can manipulate even very tiny objects (Shoshani, 1997). The subtlety of movement possible by the trunk is remarkable. Other mammals also use muscular hydrostatic, facial structures in feeding. Manatees use a large, mobile snout to evert stiff, perioral bristles that are used to sweep aquatic vegetation into the mouth (Marshall et ah, 1998a,b; Keep et al, 1998). The mobile snout of walruses may also function as a muscular hydrostat. The snout moves a dense mat of bristles, which are inferred to function, in part, during food identification, excavation, manipulation, and processing (e.g., Kastelein and Mosterd, 1989; Kastelein et al, 1991, 1994). Some mammals, including bovid antelope (Artiodactyla) and tapirs (Perissodactyla), have very mobile probosces that may function during feeding; in one species of dikdik, at least, the mobile snout is implicated in olfaction-mediated food selectivity, if not in feeding itself (e.g., Frey and Hofmann, 1996, 1997). Somewhat mobile anterior snouts {rhinaria) are characteristic of many mammals, particularly pigs and peccaries, which use them to root and to manipulate elements of the substrate during foraging (Herring, 1972). As noted previously, the lips are often used as prehensile organs to grasp plant matter during ingestion. Black rhinoceroses, for example, have a highly mobile, prehensile upper lip that is rather long and pointed, whereas white rhinoceroses have a square, nonprehensile lip (Kingdon, 1979; Nowak, 1991). This difference corresponds to the fact that the former species is a browser that grasps succulent vegetation with the lip, pulling it into the mouth, whereas the latter is a grazer that uses its stiff, keratinized upper lip as a stiff surface for cropping grasses. Extreme lip mobility is presumably effected by a muscular hydrostatic mechanism.
III. KINEMATICS OF FEEDING: THE GAPE CYCLE Having dealt briefly with the various parts potentially involved in tetrapod feeding, we can turn to their movements, or kinematics. Obviously, the ways in which tetrapods procure and process food are varied.
47
but virtually all feeding involves some degree of coordinated movement among jaws, tongue, and hyobranchium. As noted in Chapter 1, the form and function of these oropharyngeal structures are the focus of functional-morphological studies of vertebrate feeding. During feeding, movements of these morphological elements are repeated cyclically throughout food prehension, processing, transporting through the oral cavity, and swallowing. The pattern of movement is often highly regular, or stereotyped. The patterning and rhythmicity of such cyclical movements, which typically occur during feeding and locomotion, are thought to be initiated and controlled by central pattern generators (e.g., Dellow and Lund, 1971; Thexton, 1973; Cohen et al, 1988; Cohen, 1992). Central pattern generators are neural circuits that generate rhythmic patterns of neural output that, in turn, create rhythmic motor patterns, often without benefit of higher brain function. The motor patterns can be initiated and sustained by the central pattern generator without peripheral sensory feedback, but increasingly it is believed that in active, behaving animals there is extensive sensory modulation of cyclical feeding movements (e.g., Jtich et ah, 1985; Schwenk and Throckmorton, 1989; Cohen, 1992; Anderson and Nishikawa, 1993, 1997; Huang et al, 1993; Hiiemae, 1993; Hiiemae et al, 1995, 1996; Maglia and Pyles, 1995; Deban, 1997; Thexton and Hiiemae, 1997). In analyses of feeding the standard kinematic "unit" is the gape cycle (Bramble and Wake, 1985; Hiiemae and Crompton, 1985). The gape cycle represents simply the sequential pattern of velocity changes that occur during a single cycle of jaw opening and closing. This can be represented graphically as a gape profile or plot (Fig. 2.14). A complete feeding sequence usually involves many such cycles repeated one after the other (see later). It is often useful, therefore, to depict several gape cycles in a single plot to capture a larger portion of the feeding cycle. Furthermore, using the gape cycle as the point of reference, motions of the tongue, hyobranchium, head, and other relevant parts can be plotted simultaneously to create a schematic snapshot of a complex kinematic sequence involving many parts. Muscle activity patterns can be added to these for an even more complete picture (Fig. 2.15). Hence, implicit in the use of "gape cycle" is the occurrence of a suite of coordinated movements among oropharyngeal structures within the time frame delimited by a single open-close sequence of the jaws. Although gape profiles vary in their shape and complexity, they often exhibit a characteristic pattern of velocity changes that suggest discrete, intrinsic kinematic phases. The phases are delimited by inflection points in the profile {contra Smith, 1994) and when they occur each phase is given a standard name (Fig. 2.14): slow
48
Kurt Schwenk
HEAD & NECK
MANDIBLE
SKULL
Extent. Retract.
Down
Up
Down
Up
HYO-LING. COMPLEX Post.
Ant.
Down
Up
GAPE
Open
Close
F I G U R E 2.14. Kinematic plot of a single "moder' gape cycle as proposed by Bramble and Wake (1985), with permission. FC, fast close; FO, fast open; SC-PS, slow close-power stroke; SO, slow open.
open I (SO I), slow open II (SO ll)Jast open {¥0)Jast close (FC), and slow close-power stroke (SC-PS). The ubiquity of this pattern among tetrapods led to the suggestion that there is a generalized "model/' or ancestral, gape cycle (which exhibits each of these phases along with characteristic patterns of synchrony in movements of the hyolingual complex and head) (Bramble and Wake, 1985). Some tetrapod taxa may not exhibit all phases of the model gape cycle, hence the generality of this model has been questioned (e.g., Bels and Goosse, 1990; Reilly and Lauder, 1990; Smith, 1994). Certainly, the model pattern is phylogenetically widespread and occurs variably even in taxa, such as salamanders, reputed to lack it (Larsen et al, 1989; Miller and Larsen, 1990; contra Reilly and Lauder, 1990). In any case, the model gape cycle is based primarily on intraoral transport cycles (see later). Cycles from other feeding stages tend to deviate from the model cycle in characteristic ways. As noted, the Bramble-Wake model predicts coordinated movements of the jaws, hyolingual complex, head, and neck, and therefore, a particular pattern of activity in associated muscles (Fig. 2.14). In general, the model suggests that as the jaws initially part, the tongue and hyobranchium are protracted, moving forward under the food item. In those species with muscular tongues, the tongue surface is "fitted" to the food item during this stage. At the initiation of FO, the tongue and hyobranchium are rapidly retracted with
the food adherent. During SC-PS the hyolingual complex is once again protracted to its rest position. The rapid increase in gape during FO serves to release the food item from contact with the teeth and palate so that it can be transported posteriorly by the tongue toward the pharynx. Lingual papillae create a "frictional surface" (McDowell, 1972) that interlocks with the food item to move it back. During hyolingual protraction to "reset" the mechanism, the food item is held in place by palatal teeth or rugae while the tongue is slid beneath it. SC-PS applies most appropriately to mammals whose jaw muscles are active during this phase to crush food between the tooth rows. In lizards and other tetrapods, jaw muscle activity does not necessarily occur during the intervening period between cycles (e.g., Throckmorton, 1980; Smith, 1984), hence in most taxa a short power stroke may be followed by a relatively longer stationary phase. In any case, a short slow-close phase is often apparent as the teeth once again contact the food item if still present between the jaws, slowing the rate of jaw closure. As noted earlier, a typical feeding sequence requires multiple gape cycles for completion that span multiple functions from prey capture to swallowing (see next section). A weakness of the Bramble-Wake model is that a single, general cycle (based primarily on transport cycles) cannot capture the variation one might expect in the kinematic relationships necessary to perform these different functions. Furthermore, although
49
2. An Introduction to Tetrapod Feeding
of the food as it is processed. Thus the Bramble-Wake generalized tetrapod gape cycle is a heuristic model that serves as a hypothesis for testing rather than a fixed reality of feeding kinematics in tetrapods (see Smith, 1994). Variation and evolution of the gape cycle are discussed further in Chapter 8. LOWER JAW
IV. KINEMATICS OF FEEDING: FEEDING STAGES A. Overview
R. DEEP TEMPORALIS
L. DEEP TEMPORALIS
L. GENIOHYOID
R. STERNOHYOID
ANTERIOR TONGUE MARKER (cm)
HYLOID MARKER (cm)
CONDYLE POSITION (cm)
FIGURE 2.15. Two successive masticatory cycles in a mammal (Tenrec), showing the synchronous representation of kinematic variables and muscle activity (EMG); from Oron and Crompton (1985), Journal of Morphology, copyright © 1985, John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Abbreviations as in Fig. 2.14.
a series of feeding cycles might be entrained by a central pattern generator, patterning is modified by sensory feedback. Thus, modifications of this basic cycle are expected depending on where in the sequence of feeding the cycle occurs (i.e., which feeding stage; see next), the type of food, and the changing physical state
A complete feeding sequence involves from one to many gape cycles, which subserve several different behaviors. First, food must be captured or otherwise apprehended and brought to the mouth. Second, the food must be transported through the oral cavity toward the pharynx. During this traverse it may be processed in some way by the oral apparatus. Third, food collected in the pharynx must be emptied into the esophagus where peristalsis eventually takes over, moving the food into the stomach. These feeding stages are known, respectively, as ingestion, intraoral transport, and swallowing (Bramble and Wake, 1985; Hiiemae and Crompton, 1985). Sometimes processing is recognized as a separate stage. Although the nominal feeding stages are described as if discrete kinematic events, they often blend one into another so that transitional cycles are possible. Nonetheless, each stage has different functional outputs and makes different biomechanical demands on oropharyngeal structures, hence they are separable along these lines. As such, kinematic patterns of morphological elements vary more or less among stages. A standard terminology is suggested here and is maintained throughout the book (Table 2.1). Lack of consistency in the primary literature heretofore suggests that it would be worthwhile for workers to adopt such a standard nomenclature. Although the terms proposed here are as consistent as possible with mainstream usage in the primary literature, as a rule I have tended to split rather than to lump terms. This preserves specificity in meaning of certain terms that have tended to be conflated—a practice that greatly reduces their utility in scientific discourse. B. Stages of the Feeding Cycle 1. Ingestion Ingestion is the initial stage of feeding in which food is procured and brought into the mouth. For animals that feed on live prey, this stage is often equivalent to prey capture or prey prehension (in many mammals and a few other tetrapods, however, prey capture is a
50
Kurt Schwenk TABLE 2.1 Stage
Stages of Tetrapod Feeding
Predominant types
Comments
1. Ingestion
Jaw prehension Lingual prehension Suction Forelimb
Usually equivalent to prey capture in nonherbivorous species, but in some taxa a separate capture or subjugation phase precedes ingestion
2. Intraoral transport
Hyolingual Inertial Hydrodynamic
In mammals, stage I and stage II transport are recognized
Chewing (teeth) Puncture- crushing Mastication Nondental reduction (e.g., beaks and other keratinous structures) Shearing Crushing Rasping
Not all tetrapods process their food
4. Pharyngeal packing
Hyolingual
Probably restricted to lizards and turtles with muscular tongues, especially those with pronounced posterior lingual lobes
5. Pharyngeal emptying (swallowing)
Gravity Pharyngeal compression Cervical flexion Internal concertina Deglutition
Deglutition is restricted to mammals
3. Processing (food reduction)
distinct behavior that precedes ingestion). It remains true that ingestion is the least studied and least understood stage of feeding in most tetrapods (Hiiemae and Crompton, 1985; Schwenk and Throckmorton, 1989). Mechanisms for ingestion are variable. Most often it involves prehension of food with the jaws, usually their anterior portions. The actual points of contact, in most cases, are with the teeth or beak (in mammals the lips sometimes participate, as well). Jaw prehension is characteristic of caecilian amphibians, many turtles, many lizards, all snakes, crocodilians, many birds, and many mammals (Fig. 2.16a). Often there are morphological and behavioral specializations that enhance the ability to apprehend food with the jaws, particularly if the prey are fast-moving. Cranial kinesis in lizards may be one such adaptation (see Chapter 8). An elongate neck used for rapid strikes is characteristic of most snakes, some lizards, some turtles, and many birds. Many shorebirds have long, sharp bills used for spearing prey. Carnivoran mammals rip flesh from subdued prey with the teeth and some herbivorous species crop vegetation with the incisor teeth. Many aquatic tetrapods use the jaws for prehension underwater. Some
In mammals, deglutition cycles may be interlaced with transport cycles
The forelimbs participate in some taxa, alone, or with the jaws, to reduce food to smaller pieces before intraoral processing
All other forms of pharyngeal emptying characterize various nonmammalian taxa
crocodilians, for example, have narrow snouts and many sharp, monocuspid teeth that are putative adaptations for catching fish using rapid, sideways strikes of the head. Tongue, or lingual prehension, is the second most common form of ingestion (Fig. 2.16b). It is characteristic of terrestrial feeding in most frogs and salamanders.
FIGURE 2.16. (A) Jaw prehension (ingestion) in a lizard (Elgaria); note that the tongue is retracted and depressed; based on Frazzetta (1983); and (B) lingual prehension in a lizard (Phrynocephalus); based on Schwenk and Throckmorton (1989).
2. An Introduction to Tetrapod Feeding many turtles, many lizards, many birds, and many mammals. In some taxa, tongue prehension of prey is extremely rapid and the tongue is capable of lengthy excursions outside the mouth. This is most extremely developed in those species employing some form of ballistic tongue projection, notably some salamanders, frogs, and chameleons (see later). Many herbivorous mammals strip leaves off branches with the tongue or wrap the tongue around a bunch of grass to pull it out of the ground and draw it into the mouth. In all cases, lingual ingestion requires some form of lingual prehension, i.e., the tongue acts as a prehensile organ that grasps or adheres to a food item and draws it into the mouth. In most taxa there is direct adhesion between the food item and the lingual surface. Lingual adhesion probably results from a combination of frictional interlocking and wet adhesion (Bramble and Wake, 1985; Wagner and Schwenk, 2000; Schwenk, unpublished results), i.e., interdigitation of lingual rugosities with food surface features in combination with sticky mucus. In rare cases, suction may be generated by the tongue to enhance adhesion (Schwenk, 1983; Schwenk and Wake, 1988). There are a variety of ways in which the tongue can be protruded to effect food prehension. In most cases the tongue is anatomically coupled to the hyobranchium; however, the nature of this coupling is highly variable. In some species the coupling is tight so that little tongue protraction independent of the hyobranchium is possible (e.g., in most birds), but in other taxa there is only a loose connection. In most mammals, for example, the tongue is tightly adherent to the basihyal at its base, but it is nonetheless capable of extensive protrusion by means of hydrostatic deformation (see earlier discussion). In some ant- and termite-eating mammals the tongue musculature bypasses the hyobranchium altogether (Chapter 15) so that tongue and hyobranchial movements are virtually uncoupled. In most frogs the tongue has a muscular connection to the hyobranchial apparatus, but it is projected some distance from the mouth by inertial (rather than hydrostatic) elongation (Chapter 5). In salamanders the tongue varies in the extent of its attachment to the hyobranchium, but tongue protrusion is effected primarily by hyobranchial protraction rather than by protrusion of the tongue independent of the hyobranchium (Chapter 4). In extreme cases the entire length of the hyobranchial apparatus is projected from the mouth (Deban et al, 1997). In lingual feeding lizards, most tongue protrusion is also caused by hyobranchial protraction (Schwenk and Throckmorton, 1989), but some anterior tongue movement, both translational and hydrostatic, occurs as well (Chapter 8). In chameleons, the tongue is protruded on the hyobranchium as
51
in other lizards, but it is then projected entirely off the lingual process. The term tongue projection should be restricted to such cases in which there is a true, ballistic (i.e., free-floating) phase in the protrusion sequence. This will be true only for chameleons, some frogs, and some salamanders. In all other cases, extraoral tongue movement during ingestion is more accurately referred to as protrusion. In some aquatic tetrapods, ingestion is accomplished by means of suction generated by rapid expansion of the mouth and pharynx. In some cases the prey is actually drawn into the mouth, but in other cases the suction is "compensatory" in that it prevents the food item from being pushed away by the pressure wave generated by the head of the approaching predator (e.g.. Summers et al, 1998). In a few exceptional salamander species the tongue is used for prehension under water (Schwenk and Wake, 1988; J. Larsen, S. Deban, personal communication). Finally, the limbs occasionally participate in ingestion in some species. Although unusual among tetrapods as a whole, use of the forelimbs is the dominant mode of ingestion in primates in which the hands are typically the organs of food prehension. Food is grasped in the hand and brought to the mouth for processing (I have observed small primates taking bites from a mealworm held in the hand like a banana). Some other mammals commonly use the forelimbs for ingestion, including squirrels and other rodents, and raccoons (Carnivora). Mongooses (Carnivora) have been observed to pick up hard-shelled millipedes in the hand and dash them against a rock to split them open (Eisner and Davis, 1967; Eisner, 1968), however, the jaws are used for ingestion. Other carnivoran mammals, such as felids, use the forelimbs for apprehension and killing, but the jaws for ingestion. Some frogs use the forelimbs to hold or manipulate prey and occasionally to deliver it into the mouth. Occasional grasping and ingestion with the forelimbs are restricted to arboreal species, suggesting that climbing adaptations preadapted the forelimb for use in ingestion (Gray et al., 1997). Terrestrial turtles sometimes use the forelimbs to hold or manipulate food during ingestion, but they are never used to deliver food into the mouth (personal observation). Parrots often use a hind limb to grasp seeds and carry them to the mouth for husking by the beak and tongue (Homberger, 1986). Kinematics of the gape cycle during ingestion often depart somewhat from the Bramble-Wake model gape cycle. In some jaw-feeding lizards the gape profile is bell shaped, without discrete phases (e.g., Bels and Goosse, 1990), and in lingual-feeding species, slow open often occurs at wider gape angles relative to transport cycles, presumable to accommodate the
52
Kurt Schwenk
thickness of the tongue during its extraoral excursion (e.g., Schwenk and Throckmorton, 1989). 2. Intraoral
Transport
Intraoral transport is the movement of food through the mouth following ingestion. Sometimes food is transported directly to the pharynx for swallowing, but often it is first moved to a position between upper and lower tooth rows for processing. In mammals that masticate, two stages of intraoral transport are recognized (e.g., Hiiemae et al., 1978; Hiiemae and Crompton, 1985). In stage I transport, food is moved posteriorly from the prehension point to a position between the postcanine teeth for processing. This movement usually involves lateral manipulation and positioning of the food item with the tongue, as well as propalineal transport. In stage 11 transport, food particles are transported further posteriorly to the pharynx during and after chewing. They are collected in the spaces behind the tongue, in front of, and to the sides of the larynx, called the vallecular and pyriform recesses. When a large enough bolus forms there a swallow cycle is interjected amidst the masticatory (chewing) cycles (see later). In nonmammalian species, stage I and stage II transport are not clearly distinguished, although analogous stages may occur in some species. In general, it should be recognized that food must initially be transported from its point of arrival in the mouth following ingestion to a position appropriate for processing, if it occurs. While this is most often a posterior movement of the food, it need not be. Following lingual ingestion in some salamanders, for example, some prey items initially come to lie posteriorly on the back of the tongue. They are then transported by the tongue anteriorly to a position between the teeth for crushing (e.g., Schwenk and Wake, 1993). In any case, during or following processing, food particles are next transported posteriorly into the pharynx in preparation for swallowing. In species that do not process their food (e.g., many amphibians), all transport cycles serve to move ingested food directly to the pharynx. In frogs, the tongue is uniquely attached at the front and flipped over the mandible during ingestion so that the end of the tongue that is posterior while at rest in the mouth becomes anterior and the point of prey contact while out of the mouth. Consequently, when the adherent food item is returned into the mouth it comes to rest far posterior in the oral cavity, nearly within the pharynx (see Chapter 5). Hence, in some cases, intraoral transport might be bypassed altogether and swallowing initiated immediately after ingestion. Hyolingual transport is the most common form of in-
traoral transport in tetrapods. Food is moved on the surface of the tongue, which, in turn, is moved by the hyobranchial apparatus. Hyolingual transport cycles most often approximate the model gape cycle of Bramble and Wake (1985). Usually discrete SO, FO, FC, and SC-PS phases are evident. During SO the tongue is moved anteriorly under the food item. In species with muscular tongues, the tongue is formed to the food, "cupping" it for more effective transport (e.g.. Smith, 1984; Franks et al, 1984, 1985; Schwenk et al, unpublished). At the initiation of FO the tongue is retracted rapidly, carrying the food item with it. Hyolingual retraction continues through jaw closing and reverses at the onset of the next transport cycle. During protraction the tongue moves ventrally, disengaging from the food item while it is held in place by palatal teeth, palatal rugae, or some other obstruction so that the tongue can be slid beneath it. Thus, caudad progression of the food item during transport is ratchet-like. In some salamanders and possibly other amphibians, protraction is a lengthy process that occurs during prolonged recovery and preparatory phases between gape cycles (Reilly and Lauder, 1990). Hyolingual transport in amniotes involves cyclical movements of the tongue body and hyobranchium, which are often evident as oval or "figure 8" trajectories in lateral view (e.g., Crompton et al, 1977; Hiiemae et al, 1981,1995; Crompton, 1989; Hiiemae and Crompton, 1985). In some species, inertial, rather than hyolingual, transport is utilized. Most often, as in many birds and some lizards, inertial transport involves a posterodorsal jerk of the head accompanied by jaw opening so that the food item is disengaged from the jaws and carried back toward the pharynx by its inertia (Cans, 1969). Following release the head is usually moved anteriorly over the food item to increase the extent of posterior movement. It is noteworthy that inertial transport is most often characteristic of taxa with reduced tongues. Crocodilians and some mammals that bolt large pieces of food also employ inertial transport. A rather different form of inertial transport is used by macrostomatan snakes (Chapter 9). In this case, the inertia of a large prey item is used to hold it in place while the snake uses asymmetrical, kinetic movements of its jaws to pull itself over the stationary prey. This is not inertial transport as typically conceived, but rather represents an unusual form of jaw-mediated intraoral transport. Nonetheless, inertia plays a significant role in its execution. In some constricting snakes, the prey item is pinned or held by the coils while the head is literally pushed over it by the elevated cervical region (Cundall, 1995). The prey is thus transported far posteriorly into the pharynx. Thus in snakes, distinctions among feeding stages are particularly blurred.
2. An Introduction to Tetrapod Feeding Finally, some aquatic tetrapods have secondarily evolved hydrodynamic transport in which water movements are used to transport prey items following suction feeding. Because the food is buoyant in the water, it need not be supported by the tongue. Indeed, the tongue is apparently a hindrance in most aquatic feeders and is usually reduced or lost (Bramble and Wake, 1985; see earlier discussion). As for suction during ingestion, expansion of the pharynx is used to generate negative pressures within the buccal and pharyngeal cavities to draw food items posteriorly. Food position in the mouth is manipulated by modulating intraoral pressure through pharyngeal expansion and contraction so that the food item can be drawn backward and also forward for positioning during processing and then swallowing. 3. Processing Processing is the mechanical breakdown of food within the oral cavity before it is swallowed. As noted earlier, it can be considered a subset of intraoral transport cycles, but because it is kinematically distinct from transport in the strict sense, it is considered separately here. There are different sorts of food processing and it is worth establishing a standard nomenclature to maintain clarity. First, any kind of intraoral food reduction or breakdown can be referred to generically as processing. Most processing involves the use of teeth to puncture, crush, grind, or slice the food. Food reduction with the teeth is referred to as chewing. Chewing usually results in the breakdown of the food into smaller pieces, a process known as comminution; however, in some cases the food item is merely punctured and crushed, softening the food and introducing salivary enzymes without its separation into smaller bits. This type of chewing is common in many lizards, for example, and is called, aptly enough, puncture-crushing. Puncture-crushing is typical of chewing in nonmammalian tetrapods in which upper and lower teeth do not occlude. Most often it involves the use of the marginal teeth of the jaws, but it can employ palatal teeth as well. Chewing in nonmammalian taxa is not limited to puncture-crushing. It was noted earlier, for example, that tuatara (Sphenodon) use double upper tooth rows and propalineal movements of the lower tooth row between them to create a shearing action that effectively slices food. In most mammals the upper and lower teeth fit precisely together so that tightly apposed surfaces slide past one another, cusps are aligned with basins, etc., a condition known as occlusion (Fig. 2.7h). Often, occlusion is so precise that tooth wear is used literally to hone
53
the fit between upper and lower teeth. Occlusion enables mammalian teeth to perform a variety of biomechanical tasks very effectively so that extreme comminution is typical of mammalian chewing (Hiiemae and Crompton, 1985). Chewing in mammals with occlusion is called mastication. The efficacy of occluding teeth is enhanced, in most cases (especially herbivores), by lateral or anteroposterior movements of the lower jaw that create a masticatory orbit (Hiiemae, 1978). Occlusion and masticatory orbits uniquely characterize mammalian chewing (although some mammals have secondarily lost one or both). Mastication is sometimes used indiscriminately in reference to any kind of chewing, but this is inaccurate and undesirable. Mammal species that do not masticate their food (e.g., cetaceans and myrmecophages) typically lack dental occlusion and often exhibit simplified or reduced teeth, or no teeth at all. Finally, processing may involve structures other than the teeth. In many turtles and birds, for example, the tight fit of the lower jaw into the upper jaw creates a shearing action of the beak. Additional crushing can occur anteriorly and laterally in the beak between upper and lower triturating surfaces (see earlier discussion). Some monotreme mammals reduce food by rasping it with keratinous structures on the tongue and palate (see earlier discussion). 4.
Swallowing
Swallowing is the movement of food particles collected in the pharynx into the esophagus. Once in the esophagus, peristalsis transports the bolus to the stomach (once mechanically reduced or swallowed, the food item is referred to as a bolus). In lizards (and possibly other nonmammalian tetrapods), true swallowing is preceded by a separate stage called pharyngeal packing (Smith, 1984, 1986). This is treated later. In mammals, these actions are more or less combined into a single, reflexive contraction of the pharynx that empties it of the food that has collected during mastication and transport. Smith (1992) suggested that the general term "pharyngeal emptying" be applied to nonmammalian tetrapods and that "swallowing" and "deglutition" be restricted to mammals whose mechanism of pharyngeal emptying is highly derived. I use Smith's (1992) terms here with slightly modified meaning: swallowing is the generic, nontechnical term for pharyngeal emptying in all tetrapods; pharyngeal emptying is the generic, but technical term for swallowing in tetrapods; and deglutition is the specialized form of pharyngeal emptying unique to mammals. Although I am reluctant to deviate terminologically from Smith (1992), I believe that retention of "swallowing" as a
54
Kurt Schwenk
general tetrapod term is more consistent with common usage. Likewise, "pharyngeal emptying" describes a general tetrapod feeding stage of which "deglutition" is a specialized sort. Very little is known about swallowing in amphibians. In terrestrial salamanders and some frogs, intraoral transport is hyolingual, so presumably food is moved into the pharynx by means of the tongue, possibly followed by pharyngeal constriction in a manner similar to lizards (later). However, there are potentially at least two unique features of amphibian swallowing (Duellman and Trueb, 1986). First, the pharynx in many species is ciliated, suggesting that very small food particles might be moved into the esophagus by ciliary transport. Second, in most frogs and salamanders, food is pushed backward into the esophagus by retraction of the large eyeballs into the buccal cavity! This action is mediated by the retractor bulbi muscles, which pull the eyes down through large palatal vacuities, possibly forcing food posteriorly. In caecilians, the lack of tongue (and eye) mobility suggests a different mechanism for swallowing. Tongue inflation and elevation by means of engorgement with blood may help pack food into the pharynx or to empty it (some caecilian tongues contain large vascular sinuses; see earlier discussion). In reptiles, pharyngeal emptying is variable. In lizards and turtles with muscular tongues, food is initially pushed into the capacious pharynx by posterodorsal movements of the tongue and hyobranchial apparatus (Smith, 1984; Bramble and Wake, 1985) during a discrete kinematic stage known as pharyngeal packing. Pharyngeal packing is followed by pharyngeal emptying (swallowing in the strict sense), which is accomplished by contraction of the superficial, sling-like constrictor colli muscles that compress the pharynx to force food posteriorly into the esophagus. In some cases, ventroflexion of the head at the atlanto-occipital joint serves analogously to constrict the pharynx (personal observation). In varanid lizards (and possibly other species that lack posterior, muscular lobes of the hindtongue; see Fig. 2.13c) pharyngeal packing is accomplished by movements of the unusually complex hyobranchial apparatus (Smith, 1986). In snakes, the initial stages of swallowing are no more than a continuation of intraoral transport by means of skull kinesis. Large food items are simply forced into the esophagus by movements of the jaws or palate, or constricted prey are pinned by body coils and the head is forced over the immobilized prey item (see earlier discussion). However, in many, if not most, snakes, swallowing is facilitated in later stages by sinuous, concertina-like movements of the body axis within the body wall (i.e., the "concertina" flexion is internal), which
take over from the kinetic skull in pushing the bolus through the elongated esophagus (Cundall, 1995; Kley and Brainerd, 1996). In archosaurs (birds and crocodilians) the pharynx is emptied by means of gravity (e.g., Busbey, 1989; Zweers, 1982) and hyobranchial compression (Chapter 10). Long-necked birds frequently exhibit cervical flexure during swallowing as well. In nonmammalian tetrapods, swallowing is usually manifest as a discrete kinematic stage subsequent to intraoral transport. During pharyngeal packing in lizards, the posterodorsal movements of the hyolingual apparatus occur during SO, whereas in other feeding stages the tongue always moves anteriorly during SO (Smith, 1984). Thus, packing and swallowing cycles present distinct kinematic profiles that deviate significantly from the model cycle. In mammals, however, the kinematics of pharyngeal emptying are quite different. As noted, mammals employ a unique mechanism of swallowing known as deglutition. Smith (1992) enumerated several derived aspects of mammalian deglutition: (1) it is driven primarily by the action of internal pharyngeal muscles, which are lacking in all other tetrapods; (2) the swallow is a stereotyped reflex involving an explosive, but highly coordinated contraction of the palatal, lingual, pharyngeal, and hyobranchial musculature; (3) cranial nerve X (trigeminal) is primarily involved in deglutition through innervation of the pharyngeal muscles, whereas in other tetrapods, cranial nerve VII (facial) is primarily involved through its innervation of the constrictor colli muscles; and (4) deglutition cycles are integrated into intraoral transport cycles, which in turn are integrated into processing cycles. In reptiles, pharyngeal packing and emptying are discrete stages that follow upon a series of transport cycles. The essential characteristic of mammalian deglutition is the participation of the unique pharyngeal musculature (discussed earlier). The transformation of the mammalian pharynx into a muscular "tube" continuous with a soft, muscular palate, and the muscles of the tongue and hyobranchium, is part of a sweeping reorganization of the mammalian feeding apparatus relative to other tetrapods. A mammalian swallow represents an explosive, sphincter-like contraction of this musculature that restricts the lumen of the pharynx, forcing the food bolus into the esophagus. The reflex is apparently driven by sensory feedback from the pharynx that alerts the central nervous system to the presence of a sufficiently large bolus of processed food in the vallecular and pyriform recesses (Hiiemae and Crompton, 1985; Crompton, 1989; Thexton and Crompton, 1998). The extreme rapidity of this reflex allows it to be interjected into cyclical transport and chewing cycles, as opposed to the relatively slower and more involved pha-
2. An Introduction to Tetrapod Feeding
ryngeal packing and emptying cycles of lizards, which must follow a discrete stage of intraoral transport.
V. C O N C L U D I N G REMARKS If nothing else, this chapter serves to highlight the remarkable diversity of the tetrapod feeding apparatus that is all the more remarkable for its exploitation of a common set of parts (see Chapter 1). Perhaps the most important underlying theme to tetrapod feeding is the evolution of the tongue and its integration with the hyobranchial apparatus in feeding function. Modification and adaptation of the hyolingual apparatus are prominent features of tetrapod feeding evolution. As lengthy as it is, this introduction provides only a very superficial view of tetrapod feeding diversity. Few overarching patterns will be evident to the novice reader. This is partly an artifact of this cursory treatment, but it is also reflects the fundamental fact that many general attributes of the feeding system are clade specific. As such, although diverse overall, particular systems are typically characteristic of large, monophyletic species groups represented taxonomically at the family or suprafamilial level, within which they remain somewhat uniform. While this might also be an artifact resulting from the loss of intermediate forms through extinction, it more likely reflects real processes that limit the ease with which complex, functionally integrated systems can be evolutionarily reorganized (e.g.. Wake et al, 1983; Roth and Wake, 1989; Wagner and Schwenk, 2000; Schwenk, 2000b). Thus, we should expect that the kind of phenotypic remodeling necessary to create a novel feeding system to be a relatively rare evolutionary event associated phylogenetically with the origin of new clades and adaptive radiations. A prediction of this line of reasoning is that local adaptation to transient ecological conditions is not a general feature of tetrapod feeding evolution. Rather, feeding systems seem to reflect more fundamental aspects of phenotypic organization and adaptation. It is a challenge, therefore, to discern commonalties among the related, but divergent types of feeding systems among tetrapods. Individual, taxonspecific chapters address the issues of feeding system form, function, and evolution on a finer and more tractable scale. References Ahlberg, P. E., J. A. Clack, and E. Luksevics (1996) Rapid braincase evolution between Panderichthys and the earliest tetrapods. Nature 381:61-64. Albright, R. G., and E. M. Nelson (1959a) Cranial kinetics of the gen-
55
eralized colubrid snake Elaphe obsoleta quadrivittata. I. Descriptive morphology. J. Morph. 105:193-240. Albright, R. G., and E. M. Nelson (1959b) Cranial kinetics of the generaHzed colubrid snake Elaphe obsoleta quadrivittata. II. Functional morphology. ]. Morph. 105:241-292. AUin, E. F. (1975) Evolution of the mammalian ear. ]. Morph. 147: 403-438. Allin, E. F. (1986) The auditory apparatus of advanced mammal-like reptiles and early mammals. Pp. 283-294. In: The Ecology and Biology of Mammal-like Reptiles. N. Hotton, III, P. D. MacLean, ]. ]. Roth, and E. C. Roth (eds.). Smithsonian Institution Press, Washington, DC. Anderson, C. W., and K. C. Nishikawa (1993) A prey-type dependent hypoglossal feedback system in the frog Rana pipiens. Brain Behav.Evol. 42:189-196. Anderson, C. W., and K. C. Nishikawa (1997) The functional anatomy and evolution of hypoglossal afferents in the leopard frog, Rana pipiens. Brain Res. 771:285-291. Arnold, E. N. (1998) Cranial kinesis in lizards: variations, uses, and origins. Pp. 323-357. In: Evolutionary Biology, Vol. 30. M. K. Hecht, R. ]. Maclntyre, and M. T. Clegg (eds.). Plenum Press, New York. Arthur, W. (1997) The Origin of Animal Body Plans: A Study in Evolutionary Developmental Biology. Cambridge Univ. Press, Cambridge. Barlow, L. A. (1998) The biology of amphibian taste. Pp. 743-782. In: Amphibian Biology, Vol. 3. H. Heatwole and E. M. Dawley (eds.). Surrey Beatty and Sons, Chipping Norton, Australia. Beddard, F. E. (1905) A contribution to the anatomy of the frilled lizard (Chlamydosaurus kingi) and some other Agamidae. Proc. Zool.Soc.Lond. 1:9-22. Beecher, R. M. (1979) Functional significance of the mandibular symphysis. J. Morph. 159:117-130. Bellairs, A. d A . (1949) Observations on the snout of Varanus, and a comparison with that of other lizards and snakes. J. Anat. 83: 116-147. Bellairs, A. d A . (1984) Closing address: with comments on the organ of lacobson and the evolution of Squamata, and on the intermandibular connection in Squamata. Pp. 665-683. In: The Structure, Development and Evolution of Reptiles. M. W. J. Ferguson (ed.). Zoological Society of London and Academic Press, London (Symp. Zool. Soc. Lond. No. 52). Bellairs, A. d A . , and Boyd (1950) The lachrymal apparatus in lizards and snakes. II. The anterior part of the lachrymal duct and its relationship with the palate and with the nasal and vomeronasal organs. Proc. Zool. Soc. Lond. 120:269-310. Bels, V. L. (1990) The mechanism of dewlap extension in Anolis carolinensis (Reptilia: Iguanidae) with histological analysis of the hyoid apparatus. ]. Morph. 206:225-244. Bels, V. L. (1992) Functional analysis of the ritualized behavioural motor pattern in lizards: evolution of behaviour and the concept of ritualization. Zool. lb. Anat. 122:141-159. Bels, V. L., I.-P Case, V. Goosse, S. Renous, and R. Vernet (1995) Functional analysis of the throat display in the sand goanna Varanus griseus (Reptilia: Squamata: Varanidae). ]. Zool. Lond. 235: 95-116. Bels, V. L., and V. Goosse (1990) Comparative kinematic analysis of prey capture in Anolis carolinensis (Iguania) and Eacerta viridis (Scleroglossa). I. Exp. Zool. 255:120-124. Bemis, W. E., K. Schwenk, and M. H. Wake (1983) Morphology and function of the feeding apparatus in Dermophis mexicanus (Amphibia: Gymnophiona). Zool. I. Linn. Soc. 77:75-96. Berkhoudt, H. (1985a) Structure and function of avian taste receptors. Pp. 463-496. In: Form and Function in Birds, Vol. 3. A. S. King and I. McLelland (eds.). Academic Press, New York. Berkhoudt, H. (1985b) The role of exteroceptive sense organs in
56
Kurt Schwenk
avian feeding behavior. Pp. 269-272. In: Functional Morphology in Vertebrates. H.-R. Duncker and G. Fleischer (eds.). Gustav Fischer Verlag, Stuttgart (Fortschritte der Zoologie, Vol. 30). Birt, P., L. S. Hall, and G. C. Smith (1997) Ecomorphology of the tongues of Australian Megachiroptera (Chiroptera: Pteropodidae). Aust. J. Zool. 45:369-384. Bock, W. J. (1978) Tongue morphology and affinities of the Hawaiian honeycreeper M6/flm/:7rosops phaeosoma. Ibis 120:467-479. Bock, W. J. (1964) Kinetics of the avian skull. J. Morph. 114:1-41. Bock, W. J. (1966) An approach to the functional analysis of bill shape. Auk 83:10-51. Bolk, L., E. Goppert, E. Kallius, and W. Lubosch (1936) Handbuch der Vergleichenden Anatomie der Wirbeltiere, Vol. 4. Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Brainerd, E. L., and J. A. Monroy (1998) Mechanics of lung ventilation in a large aquatic salamander. Siren lacertina. J. Exp. Biol. 201: 673-682. Bramble, D. M. (1978) Origin of the mammalian feeding complex: models and mechanisms. Paleobiology 4:271-301. Bramble, D. M. (1989) Cranial specialization and locomotor habit in the lagomorpha. Am. Zool. 29:303-317. Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Biihler, P. (1981) Functional anatomy of the avian jaw apparatus. Pp. 439-468. In: Form and Function in Birds, Vol. 2. A. S. King and J. McLelland (eds.). Academic Press, London. Busbey, A. R. (1989) Form and function of the feeding apparatus in Alligator mississippiensis. J. Morph. 202:99-127. Camp, C. L. (1923) Classification of the lizards. Bull. Am. Mus. Nat. Hist. 48:289-481. Carroll, R. L. (1988) Vertebrate Paleontology and Evolution. Freeman, New York. Cohen, A. H. (1992) The role of heterarchical control in the evolution of central pattern generators. Brain Behav. Evol. 40:112-124. Cohen, A. H., S. Rossignol, and S. Grillner (1988) Neural Control of Rhythmic Movements in Vertebrates. Wiley, New York. Condon, K. (1987) A kinematic analysis of mesokinesis in the Nile monitor {Varanus niloticus). Exp. Biol. 47:73-87. Crompton, A. W. (1955) A possible explanation for the origin of the mammalian brain and skull. S. Afr. J. Sci. 52:130-133. Crompton, A. W. (1989) The evolution of mammalian mastication. Pp. 23-40. In: Complex Organismal Functions: Integration and Evolution in Vertbrates. D. B. Wake and G. Roth (eds.). Wiley, Chichester. Crompton, A. W. (1995) Masticatory function in nonmammalian cynodonts and early mammals. Pp. 55-75. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Crompton, A. W., and W. L. Hylander (1986) Changes in mandibular function following the acquistion of a dentary-squamasal jaw articulation. Pp. 263-282. In: The Ecology and Biology of Mammal-like Reptiles. N. Hotton, III, P D. MacLean, J. J. Roth, and E. C. Roth (eds.). Smithsonian Institution Press, Washington, DC. Crompton, A. W., A. J. Thexton, P Parker, and K. Hiiemae (1977) The activity of the jaw and hyoid musculature in the Virginian opossum, Didelphis virginiana. Pp. 287-305. In: The Biology of Marsupials. B. Stonehouse and D. Gilmore (eds.). Macmillan, London. Cundall, D. (1983) Activity of head muscles during feeding by snakes: a comparative study. Am. Zool. 23:383-396. Cundall, D. (1987) Functional morphology. Pp. 106-140. In: Snakes: Ecology and Evolutionary Biology. R. A. Seigel, J. T. Collins, and S. S. Novak. Macmillan, New York.
Cundall, D. (1995) Feeding behaviour in Cylindrophis and its bearing on the evolution of alethinophidian snakes. J. Zool. Lond. 237: 353-376. Cundall, D., and D. A. Rossman (1993) Cephalic anatomy of the rare Indonesian snake Anomochilus weberi. Zool. J. Linn. Soc. 109: 235-273. Cundall, D., and J. Shardo (1995) Rhinokinetic snout of thamnophiine snakes. J. Morph. 225:31-50. Cundall, D., J. Lorenz-Elwood, and J. D. Groves (1987) Asymmetric suction feeding in primitive salamanders. Experientia 43:12291231. Dalrymple, G. H. (1979) On the jaw mechanism of the snail-crushing lizards, Dracaena Daudin 1802 (Reptilia, Lacertilia, Teiidae). J. Herp. 13:303-311. Darwin, C. (1896) The Expression of the Emotions in Man and Animals. Appleton, New York. Deban, S. M. (1997) Modulation of prey-capture behavior in the plethodontid salamander Ensatina eschscholtzii. J. Exp. Biol. 200: 1951-1964. Deban, S. M., J. C. O'Reilly, and T. Themier (1994) Mechanism of defensive inflation in the chuckwalla, Sauromalus obesus. J. Exp. Zool. 270:451-459. Deban, S. M., D. B. Wake, and G. Roth (1997) Salamander with a ballistic tongue. Nature 389:27-28. de Beer, G. R. (1937) The Development of the Vertebrate Skull. Oxford Univ. Press, Oxford. deBraga, M., and O. Rieppel (1997) Reptile phylogeny and the interrelationsips of turtles. Zool. J. Linn. Soc. 120:281-354. Dellow, P. G., and J. P Lund (1971) Evidence for central timing of rhythmical mastication. J. Physiol. 215:1-13. de Queiroz, K., and J. Gauthier (1994) Toward a phylogenetic system of biological nomenclature. Trends Ecol. Evol. 9:27-31. Dial, B. E., and K. Schwenk (1996) Olfaction and predator detection in Coleonyx brevis (Squamata: Eublepharidae), with comments on the functional significance of buccal pulsing in geckos. J. Exp. Zool. 276:415-424. Doran, G. A. (1975) Review of the evolution and phylogeny of the mammalian tongue. Acta Anat. 91:118-129. Doran, G. A., and H. Baggett (1972) The specialized lingual papillae of Tachyglossus aculeatus. I. Gross and light microscopic features. Anat. Rec. 172:157-166. Dubbeldam, J. L. (1992) Cranial nerves and sensory centres: a matter of definition? Hypoglossal and other afferents of the avian sensory trigeminal system. Zool. Jb. Anat. 122:179-186. Duellman, W. E., and L. Trueb (1986) Biology of Amphibians. McGraw Hill, New York. Edmund, A. G. (1969) Dentition. Pp. 117-200. In: Biology of the Reptilia, Vol. 1. C. Gans, A. d'A. Bellairs, and T. S. Parsons (eds.). Academic Press, New York. Eisner, T. (1968) Mongoose and millipedes. Science 160:1367. Eisner, T., and J. A. Davis (1967) Mongoose throwing and smashing millipedes. Science 155:577-579. Elwood, J. R., and D. Cundall (1994) Morphology and behavior of the feeding apparatus in Cryptobranchus alleganiensis (Amphibia: Caudata). J. Morph. 220:47-70. Estes, R., K. de Queiroz, and J. Gauthier (1988) Phylogenetic relationships within Squamata. Pp. 119-281. In: Phylogenetic Relationships of the Eizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Evans, S. E. (1988) The early history and relationships of the Diapsida. Pp. 221-260. In: The Phylogeny and Classification of the Tetrapods, Vol. 1. M. J. Benton (ed.). Clarenden Press, Oxford. Feldhamer, G. A., L. C. Drickamer, S. H. Vessey, and J. F. Merritt (1999) Mammalogy. WCB/McGraw Hill, Boston.
2. A n I n t r o d u c t i o n to T e t r a p o d F e e d i n g Ferguson, M. W. J. (1981) The structure and development of the palate in Alligator mississippiensis. Arch. Oral Biol. 26:427-443. Flower, W. H. (1885) An Introduction to the Osteology of the Mammalia, 3rd Ed. Macmillan, London (1966 reprint, A. Asher and Co., Amsterdam). Font, E., and L. C. Rome (1990) Functional morphology of dewlap extension in the lizard Anolis equestris (Iguanidae). J. Morph. 206: 245-258. Forey, P., and P. Janvier (1993) Agnathans and the origin of jawed vertebrates. Nature 361:129-134. Forey, P L., C. J. Humphries, I. L. Kitching, R. W. Scotland, D. J. Siebert, and D. M. Williams (1992) Cladistics: A Practical Course in Systematics. Clarendon Press, Oxford. Forsten, A., and P. M. Youngman (1982) Hydrodamalis gigas. Am. Soc. Mammalogists, Mamm. Species 165:1-3. Franks, H. A., A. W. Crompton, and R. Z. German (1984) Mechanism of intraoral food transport macaques. Am. J. Phys. Anthro. 65: 275-282. Franks, H. A., R. Z. German, A. W. Crompton, and K. M. Hiiemae (1985) Mechanism of intraoral transport in a herbivore, the hyrax {Procavia syriacus). Arch. Oral Biol. 30:539-544. Frazzetta, T. H. (1962) A functional consideration of cranial kinesis in lizards. J. Morph. I l l :287-320. Frazzetta, T. H. (1968) Adaptive problems and possibilities in tbe temporal fenestration of tetrapod skulls. J. Morph. 125:145-158. Frazzetta, T. H. (1983) Adaptation and function of cranial kinesis in reptiles: a time-motion analysis of feeding in alligator lizards. Pp. 222-244. In: Advances in Herpetology and Evolutionary Biology: Essays in Honor of Ernest E. Williams. A. G. J. Rhodin and K. Miyata (eds.). Museum of Comparative Zoology, Cambridge, Mass. Frey, R., and R. R. Hofmann (1996) Evolutionary morphology of the proboscideal nose of Guenther's dikdik {Rhynchotragus guentheri Thomas, 1894) (Mammalia, Bovidae). Zool. Anz. 235:31-51. Frey, R., and R. R. Hofmann (1997) Skull, proboscis musculature and preorbital gland in the saiga antelope and Guenther's dikdik (Mammalia, Artiodactyla, Bovidae). Zool. Anz. 235:183-199. Fiirbringer, M. (1922) Das Zungenbein der Wirbeltiere insbesondere der Reptilien und Vogel. Abh. der Heidelberger Akad., math.naturw.Kl. 11:1-164. Gaffney, E. S. (1979) Comparative cranial morphology of recent and fossil turtles. Bull. Am. Mus. Nat. Hist. 164:65-376. Gaffney, E. S., and P. A. Meylan (1988) A phylogeny of turtles. Pp. 157-219. In: The Phylogeny and Classification of the Tetrapods, Vol. 1. M. J. Benton (ed.). Clarenden Press, Oxford. Gans, C. (1960) Studies on amphisbaenids (Amphisbaenia, Reptiliia). I. A taxonomic revision of the Trogonophinae, and a functional interpretation of the amphisbaenid adaptive pattern. Bull. Am. Mus. Nat. Hist. 119:133-204. Gans, C. (1961) The feeding mechanism of snakes and its possible evolution. Am. Zool. 1:217-227. Gans, C. (1969) Comments on inertial feeding. Copeia 1969:855-857. Gans, C. (1974) Biomechanics: An Approach to Vertebrate Biology. Lippincott, Philadelphia. Gans, C. (1989) Stages in the origin of vertebrates: analysis by means of scenarios. Biol. Rev. 64:221-268. Gans, C. (1993) Evolutionary origin of the vertebrate skull. Pp. 1-35. In: The Skull, Vol. 2. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Garcia-Fernandez, J., and P. W. H. Holland (1994) Archetypal organization of the amphioxus Hox cluster. Nature 370:563-566. Gardner, L. L. (1925) The adaptive modifications, and the taxonomic value of the tongue in birds. Proc. U. S. Nat. Mus. 67, No. 19:1-49. Gauthier, J. (1984) A Cladistic Analysis of the Higher Systematic Cate-
57
gories of the Diapsida. Ph.D. dissertation, Univ. of California, Berkeley. Gilbertson, T. A. (1998) Peripheral mechanisms of taste. Pp. 1-28. In: The Scientific Basis of Eating. R. A. Linden (ed.). Frontiers of Oral Biology Series, Vol 9, Karger, Basel. Gill, F B. (1995) Ornithology. 2nd Ed. Freeman, New York. Goodrich, E. S. (1930) Studies on the Structure and Development of Vertebrates. Macmillan, London. Goppert, E. (1903) Ueber die Bedeutung der Zunge fiir die Entstehung des sekundaren Gaumens. Verhandl. Anatom. Gessellsch. 17:75-83. Gordon, K. R. (1984) Models of tongue movement in the walrus (Odobenus rosmarus). J. Morph. 182:179-196. Gorniak, G. C , H. I. Rosenberg, and C. Gans (1982) Mastication in the tuatara, Sphenodon punctatus (Reptilia: Rhynchocephalia): structure and activity of the motor system. J. Morph. 171:321-353. Grant, T., E. C. Humphrey, and C. W. Myers (1997) The median lingual process of frogs: a bizarre character of Old World ranoids discovered in South American Dendrobatids. Am. Mus. Nov. No. 3212:1-40. Gray, L. A., J. C. O'Reilly, and K. C. Nichikawa (1997) Evolution of forelimb movement patterns for prey manipulation in anurans. J. Exp. Zool. 277:417-424. Greaves, W. S. (1995) Functional predictions from theoretical models of the skull and jaws in reptiles and mammals. Pp. 99-115. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Greer, A. E. (1989) The Biology and Evolution of Australian Lizards. Surrey Beatty and Sons, Chipping Norton, Australia. Griffiths, M. (1978) The Biology of the Monotremes. Academic Press, New York. Haas, G. (1973) Muscles of the jaws and associated structures in the Rhynchocephalia and Squamata. Pp. 285-490. In: Biology of the Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Halpern, M. (1992) Nasal chemical senses in reptiles: structure and function. Pp. 423-523. In: Biology of the Reptilia, Vol. 18. C. Gans and D. Crews (eds.). Univ. of Chicago Press, Chicago. Hanken, J., and B. K. Hall (1993) The Skull, Vols. 1-3. Univ. of Chicago Press, Chicago. Hanken, J., and P. Thorogood (1993) Evolution and development of the vertebrate skull: the role of pattern formation. Trends Ecol. Evol. 8:9-14. Hedges, S. B., and L. L. Poling (1999) A molecular phylogeny of reptiles. Science 283:998-1001. Hennig, W. (1966) Phylogenetic Systematics. Univ. of Illinois Press, Urbana, IL. Herring, S. W. (1972) The facial musculature of the Suoidea. J. Morph. 137:49-62. Herring, S. W. (1980) Functional design of cranial muscles: comparative and physiological studies in pigs. Am. Zool. 20:283-293. Herring, S. W. (1992) Muscles of mastication: architecture and functional organization. Pp. 541-548. In: The Biological Mechanisms of Tooth Movement and Craniofacial Adaptation. Z. Davidovitch (ed.). Ohio State Univ. College of Dentistry, Columbus, OH. Hiiemae, K. M. (1978) Mammalian mastication: a review of the activity of the jaw muscles and the movements they produce in chewing. Pp. 359-398. In: Develoment, Function and Evolution of Teeth. P. M. Butler and K. A. Joysey (eds.). Academic Press, New York. Hiiemae, K. M. (1993) Process and mechanism: mechanoreceptors in the mouth as the primary modulators of rhythmic behavior in feeding? Pp. 263-284. In: Sensory Research: Multimodal Perspectives. R. T. Verrillo (ed.). Lawrence Erlbaum, Hillsdale, NJ. Hiiemae, K. M., and A. W. Crompton (1985) Mastication, food
58
Kurt Schwenk
transport and swallowing. Pp. 262-290. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Hiiemae, K. M., S. M. Hayenga, and A. Reese (1995) Patterns of tongue and jaw movement in a cinefluorographic study of feeding in the macaque. Arch. Oral Biol. 40:229-246. Hiiemae, K. M., A. J. Thexton, and A. W. Crompton (1978) Intra-oral food transport: the fundamental mechanism of feeding. Pp. 181208. In: Muscle Adaptation in the Craniofacial Region. Monograph No. 8, Craniofacial Growth Series. D. Carlson and J. MacNamara (eds.). Univ. of Michigan, Ann Arbor, MI. Hiiemae, K. M., A. Thexton, J. McGarrick, and A. W. Crompton (1981) The movement of the cat hyoid during feeding. Arch. Oral Biol. 26:65-81. Hiiemae, K. M., M. R. Heath, G. Heath, E. Kazazoglu, J. Murray, D. Sapper, and K. Hamblett (1996) Natural bites, food consistency and feeding behaviour in man. Arch. Oral Biol. 41:175-189. Hildebrand, M. (1995) Analysis of Vertebrate Structure, 4th Ed. Wiley, New York. Holland, L. Z., and N. D. Holland (1998) Developmental gene expression in amphioxus: new insights into the evolutionary origin of vertebrate brain regions, neural crest, and rostrocaudal segmentation. Am. Zool. 38:647-658. Holland, L. Z., P W. H. Holland, and N. D. Holland (1996) Revealing homologies between body parts of distantly related animals by in situ hybridization to developmental genes: amphioxus versus vertebrates. Pp. 267-282, 473-483. In: Molecular Zoology: Advances, Strategies, and Protocols. J. D. Ferrraris and S. R. Palumbi (eds.). Wiley, New York. Holland, P. W. H., and J. Garcia-Femandez (1996) Hox genes and chordate evolution. Dev. Biol. 173:382-395. Holland, P W. H., L. Z. Holland, N. A. Williams, and N. D. Holland (1992) An amphioxus homeobox gene: sequence conservation, spatial expression during development and insights into vertebrate evolution. Development 116:653-661. Homberger, D. G. (1980) Funktionell-morphologische Untersuchungen zur Radiation der Emahrungs- und Trinkmethoden der Papageien (Psittaci). Bonn. Zool. Monogr. No. 13:1-192. Homberger, D. G. (1986) The lingual apparatus of the African grey parrot, Psittacus erithacus Linne (Aves: Psittacidae): description and theoretical mechanical analysis. Ornith. Monogr. (Am. Ornith. Union) No. 39:1-233. Homberger, D. G. (1988) Comparative morphology of the avian tongue. Pp. 2427-2435. In: Acta XIX Congressus Internationalis Ornithologici, Vol. 2. H. Ouellet (ed.). Univ. of Ottawa Press, Ottawa. Homberger, D. G. (1989) Correlations between morphology of the lingual apparatus and feeding mechanics in birds. Pp. 148-149. In: Trends in Vertebrate Morphology. H. Spechtna and H. Hilgers (eds.). Gustav Fischer, Stuttgart (Fortschritte der Zoologie, Vol. 35). Homberger, D. G. (1994) The hyoid suspension apparatus as a structural constraint of feeding mechanisms in birds and mammals. J. Morph. 220:355. Homberger, D. G. (1999) The avian linguo-buccal system: multiple functions in nutrition and vocalization. Pp. 94-113. In: Proc. 22nd Int. Ornithol. Congr. N. Adams and R. Slotow (eds.)., Univ. of Natal, Durban. Homberger, D. G., and A. H. Brush (1986) Functional-morphological and biochemical correlations of the keratinized structures in the African grey parrot, Psittacus erithacus (Aves). Zoomorphology 106:103-114. Horton, P. (1982) Diversity and systematic significance of anuran tongue musculature. Copeia 1982:595-602. Hotton, N. (1955) A survey of adaptive relationships of dentition to diet in the North American Iguanidae. Am. Midi. Nat. 53:88-114.
Huang, X., G. Zhang, and S. W. Herring (1993) Effects of oral sensory afferents on mastication in the miniature pig. J. Dent. Res. 72:980-986. Hylander, W. L. (1985) Mandibular function and biomechanical stress and scaling. Am. Zool. 25:315-330. lordansky, N. N. (1966) Cranial kinesis in lizards: contribution to the problem of the adaptive significance of skull kinesis. Zoologicheskii Zhurnal 45:1398-1410 [in Russian; 1968 English translation by the Smithsonian Herpetological Information Services]. lordansky, N. N. (1973) The skull of the Crocodilia. Pp. 201-262. In: Biology of the Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Jacobson, A. G. (1993) Somitomeres: mesodermal segments of the head and trunk. Pp. 42-76. In: The Skull, Vol. 1. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Janvier, P. (1996) Early Vertebrates. Clarendon Press, Oxford. Jarvik, E. (1980) Basic Structure and Evolution of Vertebrates, Vols. 1 and 2. Academic Press, New York. Jenkin, P. M. (1957) The filter-feeding and food of flamingoes (Phoenicopteri). Phil. Trans. Roy. Soc. Lond. Ser. B 240:401-493. Jiich, P. J. W., J. D. Van Willigen, M. L. Broekhuijsen, and C. M. Ballintijn (1985) Peripheral influences on the central pattern-rhythm generator for tongue movements in the rat. Arch. Oral Biol. 30: 415-421. Karasov, W. H., and J. M. Diamond (1985) Digestive adaptations for fueling the cost of endothermy. Science 228:202-204. Karasov, W. H., and J. M. Diamond (1988) Interplay between physiology and ecology in digestion. Bioscience 38:602-611. Kardong, K. V. (1973) Lateral jaw and throat musculature of the cottonmouth snake Agkistrodon piscivorus. Gegenbaurs Morph. Jahrb. 119:316-335. Kardong, K. V. (1977) Kinesis of the jaw apparatus during swallowing in the cottonmouth snake, Agkistrodon piscivorus. Copeia 1977:338-348. Kardong, K. V. (1998) Vertebrates: Comparative Anatomy, Function, Evolution, 2nd Ed. WCB/McGraw Hill, Boston. Kastelein, R. A., N. M. Gerrits, and J. L. Dubbeldam (1991) The antomy of the walrus head {Odobenus rosmarus). 2. Description of the muscles and of their role in feeding and haul-out behaviour. Aqua. Mamm. 17:156-180. Kastelein, R. A., and P. Mosterd (1989) The excavation technique for molluscs of Pacific walrusses {Odobenus rosmarus divergens) under controlled conditions. Aqua. Mamm. 15:3-5. Kastelein, R. A., M. Muller, and A. Terlouw (1994) Oral suction of a Pacific walrus {Odobenus rosmarus divergens) in air and under water. Zeit. Saugertier. 59:105-115. Kier, W. M., and K. K. Smith (1985) Tongues, tentacles and trunks: the biomechanics of movement in muscular-hydrostats. Zool. J. Linn. Soc. 83:307-324. Kingdon, J. (1979) East African Mammals: An Atlas of Evolution in Africa, Vol. IIIB. Univ. of Chicago Press, Chicago. Kley, N. J. (1998) Kinematics and mechanics of prey transport in scolecophidian snakes. Am. Zool. 38:202A. Kley, N. J., and E. L. Brainerd (1996). Internal concertina swallowing: a critical component of alethinophidian feeding systems. Am. Zool. 35:81 A. Kochva, E. (1962) On the lateral jaw musculature of the Solenoglypha with remarks on some other snakes. J. Morph. 110:227-284. Koltai, P. J., and G. A. Gates (1981) The swallow reflex: a neuromuscular cascade. Pp. 601-606. In: Proc. Xllth ORL World Congress, Budapest, Hungary. L. Surjan and G. Bodo (eds.). Kubota, K., J. Kubota, T. Nakamura, N. Fukuda, S. Asakura, S. Nakagawa and M. Masui (1962) Comparative anatomical and neurohistological observations on the tongue of the pangolin {Manis petadactyla, Linneus). Anat. Rec. 144:43-55.
2. An Introduction to Tetrapod Feeding Kuhn, H.-J., and U. Zeller (1987) Morphogenesis of the Mammalian Skull. Verlag Paul Parey, Hamburg. Lakjer, T. (1927) Studien iiber die Gaumenregion bei Sauriern im Vergleich mit Anamniern und primitiven Sauropsiden. Zool. Jb. Abt. Anat. 49:57-356. Lambertsen, R. H. (1983) Internal mechanism of rorqual feeding. J. Mamm. 64:76-88. Langston, W., Jr. (1973) The crocodilian skull in historical perspective. Pp. 263-284. In: Biology of the Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Larsen, J. H., Jr., and J. T. Beneski, Jr. (1988) Quantitative analysis of feeding kinematics in dusky salamanders (Desmognathus). Can. J. Zool. 66:1309-1317. Larsen, J. H., Jr., J. T. Beneski, Jr., and D. B. Wake (1989) Hyolingual feeding systems of the Plethodontidae: comparative kinematics of prey capture by salamanders with free and attached tongues. J. Exp. Zool. 252:25-33. Lee, M. S. Y. (1997) Reptile relationships turn turtle. Nature 389:245246. Liem, K. F. (1973) Evolutionary strategies and morphological innovations: cichlid pharyngeal jaws. Syst. Zool. 22:425-441. Liem, K. F, and S. L. Sanderson (1986) The pharyngeal jaw apparatus of labrid fishes: a functional morphological perspective. J. Morph. 187:143-158. Lombard, R. E., and D. B. Wake (1976) Tongue evolution in the lungless salamanders, family Plethodontidae. I. Introduction, theory and a general model of dynamics. J. Morph. 148:265-286. Lombard, R. E., and D. B. Wake (1977) Tongue evolution in the lungless salamanders, family Plethodontidae. II. Function and evolutionary diversity. J. Morph. 153:39-80. Lombard, R. E., and D. B. Wake (1986) Tongue evolution in the lungless salamanders, family Plethodontidae. IV. Phylogeny of plethodontid salamanders and the evolution of feeding dynamics. Syst. Zool. 35:532-551. Maglia, A. M., and R. A. Pyles (1995) Modulation of prey-capture behavior in Plethodon cinereus (Green) (Amphibia: Caudata). J. Exp. Zool. 272:167-183. Mallatt, J. (1996) Ventilation and the origin of jawed vertebrates: a new mouth. Zool. J. Linn. Soc. 117:329-404. Marshall, C. D., L. A. Clark, and R. L. Reep (1998a) The muscular hydrostat of the Florida manatee {Trichechus manatus latirostris): a functional morphological model of perioral bristle use. Mar. Mamm. Sci. 14:290-303. Marshall, C. D., G. D. Huth, V. Edmonds, D. L. Halin, and R. L. Reep (1998b) Prehensile use of perioral bristles during feeding and associated behaviors of the Florida manatee {Trichechus manatus latirostris). Mar. Mamm. Sci. 14:274-289. McDowell, S. B. (1972) The evolution of the tongue of snakes, and its bearing on snake origins. Pp. 191-273. In: Evolutionary Biology, Vol. 6. T. Dobzhansky, M. K. Hecht, and W. C. Steere (eds.). Appleton-Century-Crofts, New York. McDowell, S. B., Jr., and C. M. Bogert (1954) The systematic position of Lanthanotus and the affinities of the anguinomorphan lizards. Bull. Am. Mus. Nat. Hist. 105:1-142. Miller, A. J. (1982) Deglutition. Physiol. Rev. 62:129-184. Miller, B. T, and J. H. Larsen, Jr. (1990) Comparative kinematics of terrestrial prey capture in salamanders and newts (Amphibia: Urodela: Salamandridae). J. Exp. Zool. 256:135-153. Moore, W. J. (1981) The Mammalian Skull. Cambridge Univ. Press, Cambridge. Murray, R. G. (1973) The ultrastructure of taste buds. Pp. 1-81. In: The Ultrastructure of Sensory Organs. I. Friedmann (ed.). North Holland, Amsterdam. Nishikawa, K. C , W. M. Kier, and K. K. Smith (1999) Morphology and mechanics of tongue movement in the African pig-nosed
59
frog, Hemisus marmoratum: a muscular hydrostatic model. J. Exp. Biol. 202: 771-780. Northcott, M. E., and M. C. M. Beveridge (1988) The development and structure of pharyngeal apparatus associated with filter feeding in tilapias (Oreochromis niloticus). J. Zool. Lond. 215:133-149. Northcuttt, R. G. (1990) Ontogeny and phylogeny: a re-evaluation of conceptual relationships and some applications. Brain Behav. Evol. 36:116-140. Northcutt, R. G., and C. Gans (1983) The genesis of neural crest and epidermal placodes: a reinterpretation of vertebrate origins. Quart. Rev. Biol. 58:1-28. Nowak, R. M. (1991) Walker's Mammals of the World, 5th Ed., Vol. 2. Johns Hopkins Univ. Press, Baltimore, MD. Nowak, M. A., M. C. Boerlijst, J. Cooke, and J. Maynard Smith (1997) Evolution of genetic redundancy. Nature 388:167-171. Olson, E. C. (1961) Jaw mechanisms: rhipidistians, amphibians, reptiles. Am. Zool. 1:205-215. Oron, U., and A. W. Crompton (1985) A cineradiographic and electromyographic study of mastication in Tenrec ecaudatus. J. Morph. 185:155-182. Owen, R. (1868) On the Anatomy of Vertebrates, Vol. III. Longmans, Green and Co., London. Owerkowicz, T, C. G. Farmer, J. W. Hicks and E. L. Brainerd (1999) Contribution of gular pumping to lung ventilation in monitor lizards. Science 284:1661-1663. Ozeti, N., and D. B. Wake (1969) The morphology and evolution of the tongue and associated structures in salamanders and newts (family Salamandridae). Copeia 1969:91-123. Parker, H. W., and Grandison, A. G. C. (1977) Snakes: A Natural History. Cornell Univ. Press, Ithaca, NY. Parsons, T. S., and E. E. Williams (1962) The teeth of Amphibia and their relation to amphibian phylogeny. J. Morph. 110:375-389. Pivorunas, A. (1979) The feeding mechanisms of baleen whales. Am. Sci. 67:432-440. Platz, J. E., and J. M. Conlon (1997) Reptile relationships turn turtle . . . and turn back again. Nature 389:246. Rough, F H. (1983) Amphibians and reptiles as low-energy systems. Pp. 141-188. In: Behavioral Energetics. W. P. Aspey and S. I. Lustick (eds.). Ohio State Univ. Press, Columbus, OH. Proctor, N. S., and P. J. Lynch (1993) Manual of Ornithology. Yale Univ. Press, New Haven, CT. Reep, R. L., C. D. Marshall, M. L. Stoll, and D. M. Whitaker (1998) Distribution and innervation of facial bristles and hairs in the Florida manatee {Trichechus manatus latirostris). Mar. Mamm. Sci. 14:257-273. Regal, P. J. (1966) Feeding specializations and the classification of terrestrial salamanders. Evolution 20:392-407. Regal, P. J., and C. Gans (1976) Functional aspects of the evolution of frog tongues. Evolution 30:718-734. Reilly, S. M., and G. V Lauder (1988) Atavisms and the homology of hyobranchial elements in lower vertebrates. J. Morph. 195: 237-245. Reilly, S. M., and G. V. Lauder (1990) The evolution of tetrapod feeding behavior: kinematic homologies in prey transport. Evolution 44:1542-1557. Rieppel, O. (1993) Euryapsid relationships. N. Jb. Geol Palaont. Abh. 188:241-264. Rieppel, O. (1994) Osteology of Simosaurus gaillardoti and the relationships of stem-group Sauropterygia. Fieldiana, Geol. New Ser. No. 28:1-85. Rieppel, O., and M. deBraga (1997) Turtles as diapsid reptiles. Nature 384:453-455. Rieppel, O., and L. Labhardt (1979) Mandibular mechanics in Varanus niloticus (Reptilia: Lacertilia). Herpetologica 35:158-163. Ritter, D., and K. Nishikawa (1995) The kinematics and mechanism
60
Kurt Schwenk
of prey capture in the African pig-nosed frog (Hemisus marmoratum): description of a radicalUy divergent anuran tongue. J. Exp Biol. 198:2025-2040. Robinson, P. L. (1967) The evolution of the Lacertilia. Pp. 395-407. In: Problemes Actuels de Paleontologie (tvolution des Vertebres). Colloques Intemationaux du Centre National de la Recherche Scientifique. No. 163. Paris. Romer, A. S. (1956) Osteology of the Reptiles. Univ. of Chicago Press, Chicago. Roth, C , and D. B. Wake (1989) Conservatism and innovation in the evolution of feeding in vertebrates. Pp. 7-21. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, New York. Rubega, M. A., and B. S. Obst (1993) Surface-tension feeding in phalaropes: discovery of a novel feeding mechanism. The Auk 110: 169-178. Russell, A. P., and J. J. Thomason (1993) Mechanical analysis of the mammalian head skeleton. Pp. 345-383. In: The Skull, Vol. 3. J. Hanken, and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Sanderson, M. J., and L. Hufford (1996) Homoplasy: The Recurrence of Similarity in Evolution. Academic Press, San Diego. Sanderson, S. L., and R. Wassersug (1993) Convergent and alternative designs for vertebrate suspension feeding. Pp. 37-112. In: The Skull Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Sansom, I. J., M. P. Smith, and M. M. Smith (1994) Dentine in conodonts. Nature 368:591. Sato, M. (1976) Gustatory system. Pp. 576-587. In: Frog Neurobiology. R. Llinas and W. Precht (eds.). Springer-Verlag, Berlin. Scapino, R. P. (1965) The third joint of the canine jaw. J. Morph. 116: 23-50. Scapino, R. P. (1981) Morphological investigation into functions of the jaw symphisis in camivorans. J. Morph. 167:339-375. Schumacher, G.-H. (1973) The head muscles and hyolaryngeal skeleton of turtles and crocodilians. Pp. 101-199. In: Biology of the Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Schwenk, K. (1985) Occurrence, distribution and functional significance of taste buds in lizards. Copeia 1985:91-101. Schwenk, K. (1987) Evolutionary determinants of cranial form and function in lizards. Am. Zool. 27:105A. Schwenk, K. (1988) Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny. Pp. 569-598. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford University Press, Stanford, CA. Schwenk, K. (1994) Why snakes have forked tongues. Science 263: 1573-1577. Schwenk, K. (1995) Of tongues and noses: chemoreception in lizards and snakes. Trends Ecol. Evol. 10:7-12. Schwenk, K. (2000a) Extrinsic versus intrinsic lingual muscles: a false dichotomy? Bull. Mus. Comp. Zool. Schwenk, K. (2000b) Functional units and their evolution. In: The Character Concept in Evolutionary Biology. G. P. Wagner (ed.). Academic Press, San Diego. Schwenk, K., E. Brainerd, and A. W. Crompton (2000) Biomechanics of the mammal tongue: role of the lingual tunic. Schwenk, K., and G. S. Throckmorton (1989) Functional and evolutionary morphology of lingual feeding in squamate reptiles: phylogenetics and kinematics. J. Zool. Lond. 219:153-175. Schwenk, K., and D. B. Wake (1988) Medium-independent feeding in a plethodontid salamander: tongue projection and prey capture underwater. Am. Zool. 28:115A. Schwenk, K., and D. B. Wake (1993) Prey processing in Leurognathus marmoratus and the evolution of form and function in desmognathine salamanders (Plethodontidae). Biol. J. Linn. Soc. 49:141-162.
Shipman, R, A. Walker, and D. Bichell (1985) The Human Skeleton. Harvard Univ. Press, Cambridge, MA. Shoshani, J. (1997) It's a nose! It's a hand! It's an elephant's trunk! Nat. Hist. 106 (Nov.): 36-45. Smith, K. K. (1984) The use of the tongue and hyoid apparatus during feeding in lizards (Ctenosaura similis and Tupinambis nigropunctatus). J. Zool. Lond. 202:115-143. Smith, K. K. (1986) Morphology and function of the tongue and hyoid apparatus in Varanus (Varanidae, Lacertilia). J. Morph. 187: 261-287. Smith, K. K. (1992) The evolution of the mammalian pharynx. Zool. J. Linn. Soc. 104:313-349. Smith, K. K. (1994) Are neuromuscular systems conserved in evolution? Brain Behav. Evol. 43:293-305. Smith, K. K., and W. M. Kier (1989) Trunks, tongues, and tentacles: moving with skeletons of muscle. Am. Sci. 77:29-35. Sonntag, C. F. (1925) The comparative anatomy of the tongues of the Mammalia. XII. Summary, classification and phylogeny. Proc. Zool. Soc. Lond. 1925:701-762. Stadtmiiller, F. (1938a) Mundoffnung, Lippen, Wangen, Vestibulum oris. Pp. 895-954. In: Handbuch der Vergleichenden Anatomic der Wirbeltiere, Vol. 5. L. Bolk, E. Goppert, E. Kallius, and W. Lubosch (eds.). Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Stadtmiiller, F (1938b) Zunge, Mundhohlenboden. Pp. 955-1010. In: Handbuch der Vergleichenden Anatomic der Wirbeltiere, Vol. 5. L. Bolk, E. Goppert, E. Kallius, and W. Lubosch (eds.). Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Strait, S. G. (1997) Tooth use and the physical properties of food. Evol. Anthro. 5:199-211. Thexton, A. J. (1973) Some aspects of neurophysiology of dental interest. I. Theories of oral function. J. Dent. 2:49-54. Thexton, A. J., and A. W. Crompton (1998) The control of swallowing. Pp. 168-222. In: The Scientific Basis of Eating. R. A. Linden (ed.). Frontiers of Oral Biology Series, Vol 9. Karger, Basel. Thexton, A., and K. M. Hiiemae (1997) The effect of food consistency upon jaw movement in the macaque: a cineradiograhic study. J. Dent. Res. 76:552-560. Thomason, J. J., and A. P. Russell (1986) Mechanical factors in the evolution of the mammalian secondary palate, a theoretical analysis. J. Morph. 189:199-213. Throckmorton, G. S. (1980) The chewing cycle in the herbivorous lizard Uromastix aegyptius (Agamidae). Arch. Oral Biol. 25:225-233. Throckmorton, G. S., J. de Bavay, W. Chaffey, B. Merrotsy, S. Noske, and R. Noske (1985) The mechanism of frill erection in the bearded dragon Amphibolurus barbatus with comments on the jacky lizard A. muricatus (Agamidae). J. Morph. 183:285-292. Trueb, L. (1973) Bones, frogs and evolution. Pp. 65-132. In: Evolutionary Biology of the Anurans. J. L. Vial (ed.). Univ. of Missouri Press, Columbia. Trueb, L., and C. Gans (1983) Feeding specializations of the Mexican burrowing toad, Rhinophrynus dorsalis (Anura: Rhynophrynidae). J. Zool. Lond. 199:189-208. TumbuU, W. D. (1970) Mammalian masticatory apparatus. Fieldiana Geol. 18:149-356. van den Heuvel, W. F. (1992) Kinetics of the skull in the chicken (Callus gallus domesticus). Neth. J. Zool. 42:561-582. van der Klaauw, C. J. (1945) Cerebral skull and facial skull: a contribution to the knowledge of skull-structure. Arch. Neerl. Zool. 7:16-37. Versluys, J. (1912) Das Streptostylie-Problem und die Bewegungen im Schadel bei Sauropsiden. Zool. Jahrb. Suppl. 15:545-716. Wagner, G. P. (1996) Homologues, natural kinds and the evolution of modularity. Am. Zool. 36:36-43.
2. A n I n t r o d u c t i o n to T e t r a p o d F e e d i n g Wagner, G. R (1999) A research program for testii\g the biological homology concept. In: Homology. G. R. Bock and G. Cardew (eds.). Wiley, Chichester (Novartis Foundation Symposium No. 222). Wagner, G. P., and K. Schwenk (2000) Evolutionarily stable configurations: functional integration and the evolution of phenotypic stability. In: Evolutionary Biology, Vol. 31. M. K. Hecht, R. ]. MacIntyre, and M. T. Clegg (eds.). Kluwer Academic /Plenum Publishers, New York. Wake, D. B. (1993) Brainstem organization and branchiomeric nerves. Acta Anat. 148:124-131. Wake, D. B., G. Roth, and M. H. Wake (1983) On the problem of stasis in organismal evolution. I. Theor. Biol. 101:211-224. Wake, M. H. (1979) Hyman's Comparative Vertebrate Anatomy, 3rd Ed. Univ. of Chicago Press, Chicago. Wake, M. H. (1993) The skull as a locomotor organ. Pp. 197-240. In: The Skull, Vol. 3. ]. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Wake, M. H., and J. Hanken (1982) Development of the skull of Dermophis mexicanus (Amphibia: Gymnophiona), with comments on skull kinesis and amphibian relationships. ]. Morph. 173:203223. Walker, W. R, and D. G. Homberger (1992) Vertebrate Dissection, 8th Ed. Sauners/Harcourt Brace, Fort Worth, TX. Walker, W. R, Ir., and K. F Liem (1994) Functional Anatomy of the Vertebrates: An Evolutionary Perspective. Saunders/Harcourt Brace, Fort Worth, TX. Weihs, D., and G. Katzir (1994) Bill sweeping in the spoonbill, Platalea leucordia: evidence for a hydrodynamic function. Anim. Behav. 47:649-654. Welty, I. C. (1975) The Life of Birds, 2nd Ed. Saunders, New York. Welty, I. C , and L. Baptista (1988) The Life of Birds, 4th Ed. Saunders, New York. Wiley, E. O. (1981) Phylogenetics: The Theory and Practice of Phylogenetic Systematics. Wiley, New York. Wilkinson, M., ]. Thorley, and M. ]. Benton (1997) Uncertain turtle relationships. Nature 387:466. Wineski, L., and C. Gans (1984) Morphological basis of the feed-
61
ing mechanics in the shingle-back lizard Trachydosaurus rugosus (Scincidae, Reptilia). ]. Morph. 181:271-295. Witmer, L. M., and L. D. Martin (1987) The primitive features of the avian palate, with special reference to Mesozoic birds. Docum. Lab. Geol. Lyon, No. 99:21-40. Witmer, L. M., and K. D. Rose (1991) Biomechanics of the jaw apparatus of the gigantic Eocene bird Diatryma: implications for diet and mode of life. Paleobiology 17:95-120. Wu, X.-c, H.-D. Sues, and A. Sun (1995) A plant-eating crocodyliform reptile from the Cretaceous of China. Nature 376:678-680. Young, B. A. (1997) On the absence of taste buds in monitor lizards (Varanus) and snakes. ]. Herp. 31:130-137. Young, B. A. (1998) The comparative morphology of the intermandibular connective tissue in snakes (Reptilia: Squamata). Zool. Anz. 237:59-84. Zaher, H. (1994) Comments on the evolution of the jaw adductor musculature of snakes. Zool. ]. Linn. Soc. I l l : 339-384. Zardoya, R., and A. Meyer (1998) Complete mitochondrial genome suggests diapsid affinities of turtles. Proc. Natl. Acad. Sci. USA 95:14226-14231. Ziswiler, V, and D. S. Earner (1972) Digestion and the digestive system. Pp. 343-430. In: Avian Biology, Vol. 2. D. S. Earner and I. R. King (eds.). Academic Press, New York. Zusi, R. L. (1993) Patterns of diversity in the avian skull. Pp. 391-437. In: The Skull, Vol. 2. ]. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Zweers, G. A. (1982) The feeding system of the pigeon (Columba livia L.) Adv. Anat. Embryol. Cell Biol. 73:1-105. Zweers, G. A., and H. Berkhoudt (1991) Recognition of food in pecking, probing, and filter feeding birds. Pp. 897-901. In: Acta XX Congressus Internationalis Ornithologici. New Zealand Ornithological Congress Trust Board, Wellington. Zweers, G. A., A. F. C. Gerritsen, and P. ]. van Kranenburg-Voogd (1977) Mechanics of feeding of the mallard {Anas platyrhynchos L.; Aves, Anseriformes): the lingual apparatus and the suctionpressure pump mechanism of straining. Pp. 1-109. In: Contributions to Vertebrate Evolution, Vol. 3. M. K. Hecht and F. S. Szalay (eds.). Karger, Basel.
This Page Intentionally Left Blank
S E C T I O N
II
n^M/ /\//
\v/
This Page Intentionally Left Blank
C H A P T E R
3 Aquatic Feeding in Salamanders STEPHEN M. DEBAN AND DAVID B. WAKE Museum of Vertebrate Zoology and Department of Integrative Biology University of California Berkeley, California 94720
canum. Some undergo partial metamorphosis and possess both adult and larval features when reproductive, such as the fully aquatic Cryptobranchus. Others are primarily terrestrial and become secondarily aquatic as metamorphosed adults, notably during the breeding season. The family Salamandridae contains the most representatives of this type, known commonly as newts. Salamanders covered in this chapter may be terrestrial, semiaquatic, or fully aquatic. They may return to the water only periodically and may feed on both land and in water. Discussion here focuses on the aquatic feeding biology of these taxa. Terrestrial feeding of semiaquatic and terrestrial salamanders is discussed in the next chapter. Here we describe the various feeding behaviors: foraging, ingestion (prey capture), prey processing, intraoral prey transport, and swallowing. We review the relevant morphology and function of the sensory and motor systems and analyze the biomechanical function of the feeding apparatus. Finally, we consider the evolution of aquatic feeding systems within and among the major taxa of salamanders.
I. INTRODUCTION A. Systematics B. Natural History C. Feeding Modes and Terminology II. MORPHOLOGY A. Larval Morphology B. Adult Morphology C. Sensory and Motor Systems III. FUNCTION A. Ingestion Behavior and Kinematics B. Prey Processing C. Functional Morphology D. Biomechanics E. Metamorphosis F. Performance G. Variation IV. DIVERSITY AND EVOLUTION A. Features of Salamander Families B. Phylogenetic Patterns of Feeding Form and Function V. OPPORTUNITIES FOR FUTURE RESEARCH References
L INTRODUCTION
A. Systematics In this chapter and the next we use a recent phylogenetic hypothesis (Fig. 3.1) based on combined morphological and molecular data (Larson and Dimmick, 1993). The suborder Sirenoidea (Duellman and Trueb, 1986) is the basal clade in this phytogeny, containing only the Sirenidae, followed by a split between the Cryptobranchoidea (which contains the families Cryptobranchidae and Hynobiidae, both of which are external fertilizers) and the Salamandroidea (which has
Aquatic feeding is w^idespread among salamanders. All 10 families include members that are aquatic during part of their lives and all have members that are aquatic or semiaquatic as adults. Aquatic feeding behavior, morphology, and function in salamanders are accordingly diverse. Most aquatic adult salamanders are perennibranchiate or paedomorphic forms that forego metamorphosis and function as larvae w^hen sexually mature, such as the axolotl, Ambystoma mexi-
FEEDING (K.SchwenKed.)
65
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
66
S t e p h e n M. D e b a n a n d D a v i d B. W a k e
Sirenidae
_J
Sirenoidea
Cryptobranchidae Cryptobranchoidea Hynobiidae Amphiumidae Plethodontidae Rhyacotritonidae Salamandridae
Salamandroidea
Ambystomatidae Dicamptodontidae Proteidae
F I G U R E 3.1. Phylogenetic hypothesis of the relationships of the salamander families, from Larson and Dimmick (1993). This tree is based on a combined set of morphological and molecular data.
internal fertilization). The interrelationships of the three main clades within the Salamandroidea are uncertain. The first clade contains the Rhyacotritonidae, and the second the Plethodontidae and Amphiumidae as sister taxa. Within the third clade, the Proteidae is the outgroup to the Salamandridae, Dicamptodontidae, and Ambystomatidae, with the last two being sister taxa. B. Natural History Intimately tied to the form and function of the feeding systems of salamanders is their life history. The ancestral life history of salamanders is complex, like that of many frogs and caecilians, and includes an aquatic larval stage and a terrestrial or semiterrestrial adult stage, separated by a period of concentrated postembryonic development known as metamorphosis. Throughout this biphasic life history a salamander must capture and subdue living prey, first in water and then on land, and uses different means in these two environments. Aquatic salamanders typically ingest prey by rapidly expanding the mouth and throat, drawing prey in by suction, whereas terrestrial salamanders project a sticky tongue from the mouth to ensnare prey. These behaviors are performed rapidly, in a fraction of a second, ensuring that even highly evasive prey are captured. Six families of extant salamanders have members with the ancestral life history: Hynobiidae, Salamandridae, Rhyacotritonidae, Dicamptodontidae, Ambystomatidae, and Plethodontidae. The most frequent
departure from the ancestral life history is paedomorphosis, in which sexual maturity occurs while larval morphology is retained. This pattern is present in most families of salamanders in varying degrees of expression. It is most apparent in the Sirenidae and Proteidae, where it is termed perennibranchiation because of the retention of external gills and posterior gill bars, and less apparent in the Amphiumidae and Cryptobranchidae, whose members possess a mixture of larval and adult features. Both perennibranchiate forms and those with the ancestral life history are found in the Hynobiidae, Ambystomatidae, Dicamptodontidae, Plethodontidae, and Salamandridae. A second common departure from the ancestral life history is direct development, in which the terrestrial adult lays eggs on land, the larval stage remains encapsulated or is bypassed altogether, and a terrestrial juvenile emerges from the egg. Direct development is found only in Plethodontidae, but has evolved repeatedly in this family. This life history characterizes most plethodontids (i.e., all members of the tribes Plethodontini and Bolitoglossini) and thus over half of all salamanders. The third departure is viviparity, in which embryonic development occurs inside the oviducts of the mother, and is present only in the genera Salamandm and Mertensiella of the Salamandridae. In taxa with the ancestral life history, the aquatic larva ingests and manipulates prey using suction generated in the mouth, whereas the terrestrial adult uses the tongue. Perennibranchiate forms which retain the larval morphology also retain the larval suction feeding behavior. Direct developers either pass through a nonfeeding larval stage while encapsulated or bypass the larval stage altogether and emerge as tongueflipping terrestrial juveniles. Viviparous salamandrids give birth either to larvae, which suction feed, or to metamorphosed, terrestrial juveniles, which use tongue protrusion. During development in the oviducts of the mother, larvae and metamorphs feed on unfertilized eggs or smaller developing siblings (Alcobendas et al, 1996); the intraoviductal feeding behavior remains undescribed. Aquatic salamanders inhabit diverse freshwater environments, including lakes, ephemeral ponds, bogs, swamps, drainage ditches, springs, rivers, streams, and mountain brooks. Larvae can be divided into three types based on external morphology (Valentine and Dennis, 1964): pond type, stream type, and mountainbrook type. Pond-type larvae have large, bushy gills, deep tail fin that extends onto the body, peculiar paired organs called balancers protruding from the head during early larval development, and are found in standing or slowly flowing water where they float and swim
3. Aquatic Feeding in Salamanders in the water column. Stream-type larvae generally have small external gills and a shallow tail fin that does not extent onto the back, are found in quickly flowing water, and locomote by walking on the substrate and swim in short bursts. Mountain brook types have tiny gills and a shallow tail fin that does not extend onto the body and live in torrents and steep, cold mountain streams. Both stream and mountain brook types may be associated with a stream habitat, making use of different microhabitats. Most larval and perennibranchiate salamanders can be classified into one of these groups, although intermediates do exist. These three types of salamander larvae do not differ markedly in feeding biology. The literature on salamander diet is enormous and cannot be reviewed here in any detail, however, somie generalizations are presented. Salamanders are carnivorous in all stages of life, eating primarily live prey, including arthropods (mostly insects: Diptera, Ephemeroptera, and Trichoptera; as well as crustaceans: Isopoda, Ostracoda, Amphipoda, and Decapoda), mollusks, worms, fish, and amphibians (Martof and Scott, 1957; Peck, 1973; Joly, 1981; Petersen et ah, 1989; Tumlison et al., 1990). Some forms (e.g., cryptobranchids; Nickerson and Mays, 1973) may occasionally scavenge. Cannibalistic morphs oiAmbystoma eat primarily other salamanders, including members of their own species (e.g., Collins and Holomuzki, 1984). Terrestrial and aerial prey are occasionally eaten as they happen to fall or land in the water. The stomach contents of aquatic salamanders are generally representative of the prey available in the microhabitat, suggesting that little dietary specialization occurs (Duellman and Trueb, 1986). Avoidance of distasteful prey is more probable than specialization, although salamanders may select prey based on various parameters such as size, movement pattern, and nutrient value (Avery, 1968; see Roth, 1987, for a review). Larvae of a given species of salamander at different stages of growth may show differences in diet due to prey capture abilities (Bell, 1975; Brophy, 1980) or habitat partitioning (Leff and Bachmann, 1986), and diet may vary with seasonal changes in prey abundance (Burton, 1977). In general, the diet contains prey items of the appropriate size (White, 1977) in the frequency that they are likely to be encountered within the microhabitat. Depending on prey density and prey type, salamanders exhibit either active foraging in which they seek out and pursue prey or sit-and-wait foraging where they remain stationary and let prey come to them (Anthony et al, 1992; Jaeger and Barnard, 1981; Uiblein et ah, 1992); some salamanders have been shown to switch between foraging modes in the course of larval
67
development (Leff and Bachmann, 1986). Salamanders, as ectotherms, do not require much food to sustain themselves. A larval or adult salamander can be maintained in captivity on little food, and some remain healthy after months without eating. Simply obtaining enough food to stay alive is probably not a challenge for most salamanders; however, salamanders living in marginal conditions such as caves probably experience a scarcity of prey. The amount eaten influences growth rate and energy stores, and body size and the amount of stored fat are directly related to reproductive output. These relationships make competition for prey and, more generally, the trophic ecology of salamanders potentially important in understanding salamander feeding and may help explain the outstanding prey-capture abilities of some species. C. Feeding M o d e s and Terminology Aquatic salamanders capture or ingest prey by using the tongue, the jaws, or water flow (i.e., suction). The prey is then transported into the mouth and manipulated by water flow or movements of the tongue. Processing of prey is minimal and usually does not involve reduction or chewing, although this may occur to some degree in large taxa (e.g., amphiumids). Swallowing is the final stage of feeding in which the prey enters the esophagus. Salamanders capture prey by rapid movements of the body, jaws, and hyobranchial apparatus. A salamander can cover the initial distance between itself and the prey by (1) bringing the prey toward itself as in suction feeding, (2) extending an appendage to grasp the prey as in lingual prehension, or (3) moving its whole body toward the prey as in jaw prehension. A combination of these methods is often used to get the prey into the mouth. Suction feeding involves expanding the buccal cavity during mouth opening and drawing water and prey into the mouth while the gill slits are held closed. The buccal cavity is expanded by depression of the hyobranchial apparatus. Water is then expelled through the gill slits or mouth by elevation of the hyobranchial apparatus. Suction feeding is also called "gape-and-suck'' feeding. Lingual prehension involves grasping prey with the tongue and also relies on movements of both the jaws and hyobranchial apparatus. The hyobranchial apparatus, with the sticky tongue pad at its rostral tip, is protruded from the mouth toward the prey. The tongue pad adheres to the prey and the hyobranchial apparatus is then retracted, bringing prey either between the jaws or entirely into the buccal cavity. This behavior is also called tongue prehension, tongue flipping, tongue projection.
68
S t e p h e n M. D e b a n a n d D a v i d B. W a k e
or tongue protrusion and is accomplished by forward sHding and folding movements of hyobranchial apparatus. Jaw prehension is simply closing the mouth around the prey item so that it is trapped between the upper and the lower jaws. The salamander brings the jaws near the prey by lunging forward or sweeping the head laterally. The hyobranchial apparatus plays little role in jaw prehension. While jaw and lingual prehension are performed both on land and in water, suction feeding can only be performed aquatically because it relies upon the incompressibility of water to create enough flow to entrain the prey. The additional term ram feeding denotes ingestion behavior in which the salamander lunges forward and engulfs the prey without grasping it between the jaws. Suction feeding is the most conimon mode of ingestion among aquatic salamanders—it is used by all larvae and most species of aquatic adult salamanders. In addition, it is used by most bony fish (MuUer and Osse, 1984). The widespread use of suction feeding among aquatic vertebrates implies that it is an extremely effective way to capture prey in water (see also Chapter 16). Suction feeding is rapid; a large Cryptobranchus completes the behavior in less than one-tenth of a second. Suction feeding is also forceful and can draw even large, elongate, or highly elusive prey entirely into the mouth. Most secondarily aquatic adult salamanders are paedomorphic and perennibranchiate, and suction feed as their larvae do. Many aquatic salamanders, such as aquatic salamandrids and some hynobiids, undergo metamorphosis and lose larval features, but continue to suction feed with their adult morphology. Some salamandrids, for example Pachytriton, do so with great proficiency (Miller and Larsen, 1989) (Fig. 3.2). Others, such as cryptobranchids and amphiumids, appear to undergo partial metamorphosis, developing unusual adult structures while retaining some larval features such as posterior branchial cartilages, labial lobes, gill slits, and external gills; these groups suction feed. Some metamorphosed aquatic and semiaquatic salamanders, including all plethodontids and rhyacotritonids and some hynobiids, do not use suction to capture prey in water. Instead they use ancestral terrestrial ingestion behaviors, lingual and jaw prehension, to capture aquatic prey (Deban and Marks, 1992; Schwenk and Wake, 1993; Larsen, personal communication). Because lingual prehension and jaw prehension are performed more commonly by terrestrial forms and because they are performed in water in the same way they are on land, they will not be stressed here. Instead they will be discussed in the following chapter on terrestrial feeding in salamanders. Suction feeding behav-
F I G U R E 3.2. Suction-feeding behavior of an adult Pachytriton sp., a fully aquatic newt of the family Salamandridae. Note the rapid disappearance of the worm into the salamander's mouth, the extensive upper labial lobes occluding much of the lateral gape, and the pronounced buccal expansion and hyobranchial depression. The video frames are 8.3 msec apart, and the background is a 5-mm grid.
ior, morphology, and function in both larval and adult salamanders will be emphasized in this chapter. II. MORPHOLOGY Morphology is one of the most completely known aspects of the feeding biology of salamanders. Exten-
3. A q u a t i c F e e d i n g in S a l a m a n d e r s
sive anatomical data have been gathered over the last century and descriptions of head morphology exist for many salamander taxa (Parker, 1882; Driiner, 1901, 1904; Francis, 1934; Edgeworth, 1935; Piatt, 1938; Schumacher, 1958; Severtsov, 1964). Still, all morphological data relevant to feeding are not available for every group. Feeding behavior relies on morphological features of the skull, jaws, hyobranchial apparatus, tongue pad, gill slits, and labial lobes. Both larval and adult aquatic salamanders must feed in the same environment, and thus share many of these features. This section describes the general pattern of the feeding morphology of larval and adult salamanders. Most features of the larval feeding system presented here persist in the adults of perennibranchiate forms, i.e., perennibranchiate adults retain an essentially larval morphology. Metamorphic forms lose many larval morphological features, but may retain the larval feeding behavior while supplementing it with terrestrial feeding behavior. Metamorphic forms are discussed in a later section on adult morphology. A. Larval Morphology The head of larval salamanders is tapered anteriorly with an anterior mouth opening and posterior gill slits. Bushy external gills protrude posterolateral^ and have up to four main branches, or rami, each with numerous vascularized fimbriae. The eyes are lidless and do not protrude. Neuromasts of the lateral line system are abundant around the mouth and sides of the head.
69
and prominent labial lobes occlude the sides of the mouth. The teeth of larval salamanders are sharply pointed, typically nonpedicellate, and monocuspid and consist of a cylindrical base fused to a conical apex. They lack the suture plane of adult pedicellate teeth (Beneski and Larsen, 1989a,b). Dentition is generally homodont, and teeth are borne on the premaxilla, maxilla, vomer, and palatine or palatopterygoid of the upper jaw and the dentary and coronoid of the lower jaw (Fig. 3.3). Teeth grow in size and number during larval life. New tooth loci are added with the enlargement of jaw bones (Bonebrake and Brandon, 1971). Teeth are added posteriorly along the maxillae and dentaries and posterolaterally to the vomers (Parker and Dunn, 1964; Worthington and Wake, 1971). At metamorphosis, teeth of the coronoid and palatopterygoid usually disappear with the disintegration of these bones, and vomerine teeth extend onto the parasphenoid (Wilder, 1928). The tongue is usually absent or poorly developed in larvae; if present, it possesses few of the glands or papillae of the tongue pad of postmetamorphic individuals. Larvae often possess a fold in the floor of the mouth just rostral and ventral to the anterior elements of the hyobranchial apparatus, called a buccal fold or anterior fold. This region of overlap is just beneath the tongue and it accommodates ventral expansion of the buccal floor during suction feeding. A ventral gular slit is often present at the level of the external gills and accommodates extension of the skin of the throat during buccal expansion. Extensions of skin on the upper and lower jaws
premaxilla maxilla lacrimal pterygoid
1 cm
F I G U R E 3.3. Larval skull of the aquatic hynobiid Batrachuperus mustersi in dorsal, lateral, and ventral views. Note the presence of maxillae, lacrimals, septomaxillae, and prefrontals, as well as separate prootic, opisthotic, and exoccipital ossifications.
70
Stephen M. Deban and David B. Wake
called labial lobes block the sides of the mouth and direct the gape anteriorly. Upper and lower lobes are fused posteriorly and interlock or simply overlap anteriorly; in larvae the regions where upper and lower lobes meet are called labial folds. Labial lobes have long been recognized as important for effective suction feeding (Matthes, 1934). These interlocking lobes and their folds allow for expansion during mouth opening and, at the same time, prevent water from entering the mouth from the sides. During suction feeding, the labial lobes narrow the gape and direct it anteriorly, thus restricting, focusing, and accelerating water flow and increasing the likelihood that prey are captured. Two or three gill slits (also called gill apertures or gill clefts) are located on each side of the head below the external gill rami. The gill slits lie between the distal elements (epibranchials) of the hyobranchial apparatus and open into the mouth. Water that enters the mouth during suction feeding is expelled through the gill slits after the prey has been captured. Interlocking fingers of tissue or mineralized tooth-like elements called gill rakers are present on the pharyngeal surfaces of the gill bars of some taxa and help prevent prey from escaping as water is expelled. 1. Larval Skull, Jaws, and Hyobranchial
Apparatus
The larval skull consists of the posterior neurocranium and auditory capsules connected via parachordals and trabeculae to the anterior nasal capsules, invested by dermal bones dorsally and by palatal bones and attached ventrally to the suspensorium (Fig. 3.3). Posterior ossifications include the prootic, opisthotic, and exoccipital, which together comprise the occipitootic (Trueb, 1993). Dorsal bones include the paired frontals and parietals. Marginal tooth-bearing bones are the paired maxillae and the premaxillae, which are paired or fused medially. Bones of the palate include the paired vomers, the single parasphenoid, and the paired palatines or palatopterygolds. Palatines are present as separate elements only in the Sirenidae, positioned caudal to the vomers and bearing many teeth. Palatopterygoids are present caudal to the vomers in the Proteidae (where they articulate with the vomers and quadrates to form robust components of the upper jaw) and in larval plethodontids (where they are freefloating) and bear teeth on their rostrolateral edges. The orbitosphenoids form around the trabecular cartilages. The suspensorium, which connects the lower jaw to the skull, consists of the cartilaginous palatoquadrate and osseous quadrate invested by the bony pterygoid (which is absent in all plethodontids) and squamosal. The quadrate articulates with the articular of the lower jaw.
The skull bones of larval salamanders are loosely articulated and less robust than in the adult. Sutures between bones are lacking or incomplete, or gaps exist between bones. During growth, the bones increase in relative size and their degree of overlap and fusion. Larval skull morphology appears more conservative taxonomically than that of adults, and the phylogenetic diversity of larval skull morphology is not as well documented. The upper jaw consists of the paired or fused premaxilla, vomer, palatopterygoid, and, in some cases, the late-appearing paired maxilla, all of which bear teeth. The maxilla, when present, is loosely attached to the skull. The maxillae are absent in the larva of many taxa and appear during metamorphosis. In some perennibranchiate taxa the maxilla never forms (e.g., Necturus and Pseudobranchus) or appears late and remains a small, free-floating element (e.g.. Siren; Reilly and Altig, 1996). In these, the palatine or palatopterygoid forms a robust, dentate component of the upper jaw. In all taxa the medial parasphenoid forms the palate. The lower jaw consists of Meckel's cartilage invested by the tooth-bearing dentary, the prearticular (also called the gonial), usually the toothed coronoid (also called the splenial), and the angular (in cryptobranchoids). The articular cartilage of the lower jaw forms a joint with the quadrate of the skull, which is positioned between the orbit and the otic capsule. The jaw joint is far anterior in most larvae and perennibranchiate forms, reducing the gape. The hyobranchial apparatus of larval salamanders is composed of interconnected skeletal elements that lie in the floor of the mouth and throat and border the gill slits. An unpaired medial element, the basibranchial (sometimes called the copula), forms the main axis of the larval hyobranchial apparatus (Fig. 3.4) with which the remaining bilaterally paired elements articulate. The basibranchial articulates rostrally with the ceratohyals and caudally with the first and second ceratobranchials. These elements are sometimes called hypobranchials (Reilly and Lauder, 1988b), but likely represent the fusion of the ancestral hypobranchial and ceratobranchial, and here are called ceratobranchials. A third ceratobranchial is present in cryptobranchoids, but does not meet the basibranchial (see Batrachuperus, Fig. 3.4). The ceratobranchials are in turn attached to the more distal epibranchials (sometimes called ceratobranchials; Reilly and Lauder, 1988b), which number three or four on each side of the main axis and which curve dorsally to border the gill slits and support the external gills. An unpaired, medial element, the urohyal (also called the second basibranchial or os thyroideum), lies caudal to the basibranchial and is often continuous with it. Hypohyals lie between the ceratohyals and the
71
3. A q u a t i c F e e d i n g in S a l a m a n d e r s
Dicamptodon ensatus (L)
Pseudobranchus striatus (A)
Gyrinophilus porphyriticus
Proteus anguinus (A)
(L)
Batrachuperus mustersi (L)
Amphiuma means (L)
F I G U R E 3.4. Hyobranchia (in ventral view) of larval and perennibranchiate salamanders of six families. Cartilage is shown in white and ossification or mineralization is shaded. Scale bars equal 3 mm. See Table 3.1 for abbreviations. Proteus drawing is adapted from Marche and Durand (1983), with permission.
basibranchial in some taxa (e.g., Dicamptodon and Amphiuma, Fig. 3.4). The basibranchial and ceratobranchials are fused to form a branchial plate in some taxa (e.g., plethodontids, see Gyrinophilus, Fig. 3.4). The shapes of the hyobranchial elements show variation among taxa (Fig. 3.4), but with some underlying common themes. The basibranchial is usually elongate and circular or triangular in cross section. The ceratohyals are flat and broad, and the ceratobranchials are often cylindrical. The epibranchials are triangular in cross section and taper to points or terminate with knobs at the posterior tips, which are attached by muscles to the skull and dorsolateral neck region. The ceratohyals and first epibranchials are usually the largest elements of the larval hyobranchial apparatus; more posterior elements are progressively smaller. Articulations between elements are generally loose.
and elements are held together by shared membranes rather than by ligaments. The ceratohyals, however, are linked at their midpoints to the mandibles by the hyomandibular ligaments or at their caudal ends to the prearticulars or quadrates by hyoquadrate ligaments. A single branched ligament may be present in some taxa (Elwood and Cundall, 1994) or at some stages of development (Reilly and Altig, 1996). The elements of the hyobranchial apparatus are mostly cartilaginous, although ossification or mineralization occurs in some elements and is usually more extensive in older larvae. The basibranchial and ceratohyals are entirely cartilaginous, whereas the urohyal is usually the first element to ossify. The second and third ceratobranchials may be ossified or calcified in cryptobranchoids, and the epibranchials are ossified in some taxa (e.g., dicamptodontids and plethodontids). Adults
72
S t e p h e n M. D e b a n a n d D a v i d B. W a k e
of perennibranchiate species may show the metamorphic (i.e., adult) pattern of ossification, but a larval number and configuration of hyobranchial elements (e.g., Proteus and Fseudohranchus, Fig. 3.4). 2. Larval Head
Musculature
The muscles of the head can be divided into four broad functional groups: jaw muscles, hyobranchial muscles, skull muscles, and throat muscles. The general pattern of larval muscle morphology is presented in this section. Some taxa may not possess all of the muscles listed. Current names of muscles are used; some synonyms are listed in parentheses and are not mentioned thereafter (for more thorough lists of homo-
logues and synonyms, see Edgeworth, 1935; Francis, 1934). Only muscles likely to function in feeding are presented, and discussion of their functions is restricted to their expected actions when they contract. A later section on feeding function discusses in more detail the roles of various muscles in producing coordinated feeding movements. Mouth opening and closing are performed by antagonistic sets of muscles: the depressors, which open the mouth, and the levators, which close the mouth (Fig. 3.5A). The mouth-opening muscle is the depressor mandibulae (DM), which is sometimes divided into two parts: depressor mandibulae anterior (DMA) and posterior (DMP) (see Table 3.1 for abbreviations). DM muscles originate on the skull in the vicinity of the
F I G U R E 3.5. Larval (A) and adult (B) cranial muscles of the salamandrid Taricha granulosa in ventral and lateral views. In ventral views, deep muscles are depicted on the left side and more superficial muscles are shown on the right side. See Table 3.1 for abbreviations and the text for descriptions of muscles.
3. Aquatic Feeding in Salamanders TABLE 3.1 Abbreviations, Homologs and Synonyms of Skeletal Elements and Muscles Homologs and Synonyms" Skeletal elements A Angular BB Basibranchial BP Branchial plate BH C CB CH CO D EB
Basihyal Comua Ceratobranchial Ceratohyal Coronoid Dentary Epibranchial
HH L M OG PA R UH
Hypohyal Lingual Meckel's cartilage Otoglossal Prearticular Radial Urohyal
Muscles AA
Adductores arcuum
BH
Branchiohyoideus
CDSP
Cephalodorsosubpharyngeus
BM
Branchiomandibularis
DM DMA DMP DT G GG GGL GH GHL GHM HOYP IMA IMP IH IHP
Depressor mandibulae Depressor mandibulae anterior Depressor mandibulae posterior Dorsalis trunci Gularis Genioglossus Genioglossus lateralis Geniohyoideus Geniohyoideus lateralis Geniohyoideus medialis Hebosteoypsiloideus Intermandibularis anterior Intermandibularis posterior Interhyoideus Interhyoideus posterior
lOQ ITCS
Interossaquadrata Intertransversarius capitis superior Intervertebral epaxial Levatores arcuum branchiarum Levator mandibulae externus Levator mandibulae internus Levator mandibulae internus anterior Levator mandibulae internus posterior Lateral subvertebralis Medial subvertebralis Quadratopectoralis Rectus cervicis Rectus cervicis profundus Rectus cervicis superficialis Subarcualis rectus Subhyoideus Transversus ventralis
IVE LAB LME LMI LMIA LMIP LSV MSV QP RC RCP RCS SAR SH TV
^Synonyms are in parentheses.
(Copula) Splits into BB and CBs on metamorphosis (Hypobranchial) (Splenial) (Ceratobranchial); fused larval CB and EB in some taxa
(Gonial) R 1 formed from larval H H (Second Basibranchial, Os thyroideum) (Subarcualis obliquus, Transversus ventralis 1) (Ceratohyoideus, Ceratohyoideus externus) Formed from fusion of larval LAB and TV (Ceratomandibularis, Hyomandibularis)
Formed from larval IHP Formed from larval GG (Coracomandibularis) Formed from larval G H Formed from larval G H Formed from larval RC (Submentalis) (Sphincter colli, Interbranchialis) Formed from larval IH
Formed from larval IHP (Abdominohyoideus) (Sternohyoideus) Formed from larval IH
73
squamosal and quadrate (the DMP originates on the ceratohyal in Siren) and insert on the ventrocaudal surface of the articular or prearticular just beneath the jaw joint. The branchiomandibularis (BM, also called ceratomandibularis or hyomandibularis) originates on the distal tip of the first epibranchial and either merges with the DMP to insert on the prearticular or inserts on the hyomandibular ligament. In addition to depressing the lower jaw, the BM is in a position to aid in hyobranchial depression by drawing the epibranchial tips dorsorostrally. All exert dorsal and caudal components of force on the caudal side of the jaw joint. Contraction of these muscles swings the lower jaw downward by rotating it about its articulation with the quadrate. In addition to those jaw depressors which span the jaw joint, elongate ventral muscles may contribute to mouth opening by pulling the tip of the lower jaw posteroventrally. These include the geniohyoideus [GH, also called the coracomandibularis (Lauder and Shaffer, 1985)] and the rectus cervicis superficialis and profundus (RCS and RCP). The GH originates on the mandibular symphysis and inserts on the urohyal. The RCS (also called the sternohyoideus) originates from the sternum and the RCS originates on the rectus abdominus muscles, which in turn originate on the pelvis. The RCP (also called the abdominohyoideus) and RCS may be indistinct from one another at their rostral ends, where they insert on the basibranchial, first ceratobranchials, or the branchial plate. These muscles can affect mouth opening if the GH contracts simultaneously and the skull is prevented from flexing ventrally. The connection between mandibles and ceratohyals by the hyomandibular ligament may also affect mouth opening as the hyobranchial apparatus is depressed (Reilly and Lauder, 1990). Mouth-closing muscles include the levator (or adductor) mandibulae, which is divided into two main parts: the levator mandibulae externus (LME) and the levator mandibular internus (LMI). The internus is sometimes further divided into the internus anterior (LMIA) and internus posterior (LMIP). The LMIA and LME originate on the skull and the LMIP originates on the atlas. All insert primarily on the prearticular, but also on the dentary of the lower jaw. These muscles are positioned to produce primarily dorsal, but also medial components of force on the rostral side of the jaw joint, and swing the lower jaw upward or apply biting force. Suction feeding requires movements of the hyobranchial apparatus in addition to movements of the jaws. Muscles of the hyobranchial apparatus are divided here into (1) hyobranchial depressors and levators, which produce primarily dorsoventral and rostrocaudal movements, (2) branchial adductors and abductors, which produce lateral movements, and
74
Stephen M. Deban and David B. Wake
(3) hyobranchial stabilizers, which anchor or adjust the branchial tips. Hyobranchial depressors swing the hyobranchial apparatus ventrally and caudally and are responsible for enlarging the buccal cavity during suction feeding. The RCS and RCP exert a caudal force on the center of the hyobranchial apparatus. The branchiohyoideus (BH, also called ceratohyoideus or ceratohyoideus externus) originates on the ventral surface of the ceratohyal, inserts on the tip of the first epibranchial, and may share fibers with the adjacent BM (Fig. 3.5A). The BH pulls the epibranchial tip closer to the rostral end of the ceratohyal, buckling the hyobranchial apparatus ventrad and pulling the epibranchial tip ventrad. As mentioned previously, the BM may similarly aid in hyobranchial depression. Hyobranchial levators either cradle the hyobranchial apparatus or attach directly to it and function to compress the buccal cavity and return the hyobranchial apparatus to its resting position. The geniohyoideus (GH) originates on the urohyal, inserts at the mandibular symphysis, and pulls the center of the hyobranchial apparatus rostrad. As mentioned earlier, because of its attachment to the mandibular symphysis, this muscle may also act as a mouth opener if the hyobranchial apparatus is stabilized in its depressed position. The intermandibularis posterior (IMP) originates on the medial aponeurosis of the floor of the mouth and runs laterally to insert on the dentary and prearticular. A small intermandibularis anterior (IMA, also called submentalis) is usually present, spanning the mandibular symphysis. The interhyoideus (IH) originates with the IMP on the medial aponeurosis but runs caudally as well as laterally to insert on the caudal tip of the ceratohyal or on the hyoquadrate ligament. An interhyoideus posterior (IHP, also called the interbranchialis or sphincter colli) is often distinct from the IH, lies just caudal to the IH, and runs from the distal tip of the first epibranchial medially to insert on the medial aponeurosis of the throat. The IMP, IH, and IHP all raise the floor of the mouth, cradling and pushing the hyobranchial apparatus upward to its resting position. The IHP may also function in branchial adduction. Branchial abductors move the epibranchials apart, opening the gill slits. The BH is in a position to abduct the first epibranchial laterally in some species (e.g., Ambystoma tigriunum), whereas in others (e.g., plethodontids) it appears to move the epibranchial medially. The tranversus ventralis (TV) muscles, which vary in number among taxa, originate at the midline on the ventral surface of the pharynx and insert on or near the medial edge of the most medial epibranchial. The TV
muscles are in a position to move the last epibranchial medially. Branchial adductors move the epibranchial tips toward one another. The BH may serve this function in some taxa. The subarcularis rectus 1 (SAR 1), which originates on the ventral surface of the ceratohyal and inserts on the first epibranchial, can adduct the epibranchial medially. Subarcualis rectus muscles 2 through 4 span the epibranchials and can move them toward one another. The single or double adductores arcuum (AA, also called the subarcualis obliquus or transversus ventralis 1) runs from the ceratobranchial or branchial plate to the second and third epibranchials and can adduct the epibranchials medially. This may close or open the gill slits, depending on the relative position of the first epibranchial, or may swing the basibranchial relative to the epibranchial to return the hyobranchial apparatus to its resting position. Hyobranchial stabilizers include the levatores arcuum branchiarum (LAB), which originate broadly on the dorsalis trunci muscles (DT) of the neck or on the otic region of the skull and insert on the epibranchial tips (Fig. 3.5A). They are in a position to move the epibranchial tips dorsally or rostrally or to anchor them and prevent caudal excursion. The head is moved dorsally, ventrally, or laterally to direct the gape during suction feeding and to aid in prey manipulation and swallowing. Four sets of muscles can produce these movements. The dorsalis trunci (DT) originates on the vertebrae and inserts on the occipito-otic of the skull and can flex the head dorsolaterally. The intervertebral epaxial muscle (IVE), sometimes not distinct from the DT, can also raise the cranium. The intertransversarius capitis superior (ITCS) originates on the atlas, inserts on the occipitootic, and can deflect the cranium dorsolaterally. The levator scapulae and cucuUaris muscles originate on the pectoral girdle and insert on the skull and can flex the head laterally. The skull can be flexed ventrolaterally by the medial subvertebralis muscle (MSV), which originates on the ribs of the first few vertebrae and inserts on the parasphenoid. The lateral subvertebralis (LSV) originates from the lateral body wall musculature, inserts in the otic region, and can flex the skull laterally. The IMP, IH, and IHP can raise the floor of the mouth and constrict the pharynx and may function in prey manipulation or swallowing. In addition to branchial adduction, the TV can constrict the pharynx and aid in swallowing. The skull levators (DT, IVE, ITCS) and depressors (MSV, LSV), as well as the levator scapulae and cucullaris muscles, can aid in swallowing by moving the head.
75
3. Aquatic Feeding in Salamanders B. Adult Morphology Adult aquatic salamanders represent the full range of metamorphic patterns, from completely larviform, perennibranchiate species, such as the axolotl, to completely metamorphosed forms, such as most newts. We divide this spectrum of developmental patterns into three groups, based on the overall degree of metamorphosis of the feeding system. (1) "Perennibranchiates" possess all or most larval features, including gill slits, external gills, posterior branchial elements, and interlocking labial lobes. They possess few adult features, such as maxillae and a tongue pad. This group includes all of the Sirenidae and Proteidae and some members of other families. (2) "Partial metamorphs'' possess a mixture of larval and adult features. They may lose external gills and gill slits, but may retain labial lobes and posterior branchial elements. They possess maxillae and a rudimentary tongue pad. This group includes all members of the Cryptobranchidae and Amphiumidae, as well as ambystomatids formerly placed in the genus Rhyacosiredon. (3) "Metamorphs'' have lost external gills, gill slits, posterior branchial elements, interlocking labial lobes, and have fully formed maxillae and tongue pads with intrinsic tongue muscles. This group includes the newts and all terrestrial and semiaquatic salamanders. The adult morphology of metamorphs and partial metamorphs is the focus of this section. Perennibranchiate adults are considered here as well, but are emphasized in the previous section on larval morphology. Unlike the heads of larval and perennibranchiate salamanders, which have pointed or tapered rostra, the heads of metamorphosed and partially metamorphosed salamanders are more rounded and blunt. The external gills and tail fin are lost, the eyes protrude, and eyelids form. In addition to these superficial changes, which occur at metamorphosis, deeper alterations are made in most components of the feeding system, including the teeth, tongue, jaws, skull, hyobranchial apparatus, and many muscles of the head. The teeth of metamorphosed salamanders consist of a bony pedicel attached by fibrous connective tissue to an enamel-coated crown. Teeth of metamorphosed animals are either monocuspid or dicuspid, with various crown shapes, including clubs, discs, and spikes. Older perennibranchiate and metamorphosing animals possess biscupid, subpedicellate teeth. The dentition is generally monostichous and homodont (Beneski and Larsen, 1989a,b). Partially metamorphosing species may possess bicuspid teeth (Erdman and Cundall, 1984; Elwood and Cundall, 1994). Perennibranchiate forms generally lack a tongue
pad, whereas partially metamorphosing forms possess a raised glandular field in the floor of the mouth. Metamorphosing species have tongues of various sizes and degrees of complexity. Suction feeding forms tend to have smaller, simpler tongues, whereas those that use tongue prehension in water have larger, more complex tongues such as those of terrestrial species. Adult tongue morphologies of those species that use tongue prehension will be presented in the next chapter. The labial lobes of adult salamanders are of two types: those that are essentially retained larval features and those that form after metamorphosis. Labial lobes of the larval type are found in all larval and perennibranchiate salamanders and in many partially metamorphosing forms. Sirenids have by far the most extensive labial lobes of this type; the lateral gape is entirely occluded and the mouth forms a small, rectangular anterior opening. Cryptobranchus is unusual in that it retains larval labial lobes on the lower jaws, but not on the upper jaws. The fully aquatic, suction feeding ambystomatids formerly placed in the genus Rhyacosiredon (now placed in Ambystoma) retain larval labial lobes and otherwise appear to be partially metamorphosed (Reilly and Brandon, 1994). The labial lobes of aquatic adult salamandrids and hynobiids are of the second type and only form on the upper jaw. These adult labial lobes are simple flaps and are not homologous to the more complex, interlocking larval lobes. All perennibranchiate and many partially metamorphosing forms possess gill apertures. Cryptobranchus and Amphiuma possess one gill aperture, or spiracle on each side of the head. Andrias has no gill openings. Perennibranchiate forms may possess a gular slit, whereas some partially or completely metamorphosing forms possess a fold in this location. Cryptobranchoids and the fully aquatic salamandrid, Pachytriton, are unusual in having longitudinal folds in the buccal floor that accommodate ventral expansion. Cryptobranchids also retain the larval anterior fold beneath the hyoid and are capable of perhaps the greatest increases in buccal volume of any salamander. 1. Adult Skull, Jaws, and Hyobranchial
Apparatus
Metamorphosis from the larval to adult form involves changes in the skull, jaws, and hyobranchial apparatus. These changes are presented here in order to introduce the adult form of these skeletal systems. Elements that do not change substantially during metamorphosis are not mentioned in this section. During metamorphosis, elements of the skull are lost, appear, or are remodeled, and further ossification
76
Stephen M. Deban and David B. Wake
and expansion occur. In addition to the bones present in the larva, some of which enlarge and overlap, prefrontals, lacrimals, nasals, and septomaxillae may develop in the rostrum of the skull, paralleling the developmental appearance of rostral elements in the hyobranchial apparatus. Some taxa, particularly hynobiids, already possess many of these elements as larvae (Fig. 3.3). During metamorphosis, the jaws are strengthened. The maxilla appears or broadens its attachments to the skull. The vomer and premaxilla enlarge and reinforce their connections with the skull, and the orientation of vomerine tooth rows changes with the change in shape of the vomer. Dentate extensions of the vomer migrate posteriorly onto the parasphenoid, and the toothbearing, palatine portion of the palatopterygoid is lost. The dentaries become more heavily ossified and abut at the symphysis, and the coronoid is lost (except in Dicamptodon in which only the teeth are lost). Meckel's cartilage is reduced and the dentary and prearticular may fuse to one another. The jaw articulation is repositioned posteriorly so that it comes to lie behind the orbit as the squamosal and quadrate swing posteriorly, increasing the gape. The hyobranchial apparatus of perennibranchiate species is essentially larval in configuration. In some taxa, such as Proteus, extensive ossification occurs, but all larval elements are retained (Marche and Durand, 1983). In metamorphosing taxa, the transformation from the larval to the adult hyobranchial configuration involves loss, remodeling, and often ossification of elements. The posterior branchial arches are lost, typically leaving two pairs of ceratobranchials and one or two pairs of epibranchials in the adult. Some terrestrial members of the Salamandridae lack discrete first ceratobranchials as adults, and the epibranchials appear elongated. These adult elements are probably the product of fusion of the larval first ceratobranchials and first epibranchials and are called the epibranchials. The epibranchial and ceratobranchial persist as a separate element in all aquatic members of the Salamandridae. In the Plethodontidae the larval epibranchials disintegrate completely and a new adult epibranchial develops at metamorphosis from undifferentiated tissue (Alberch and Gale, 1986). In the Hynobiidae, two pairs of epibranchials persist, the first cartilaginous and the second bony. The Cryptobranchidae show variation in the degree of metamorphosis of the hyobranchial system. Cryptobranchus has a larval number of posterior elements, but with more ossification and mineralization than in the larva, whereas Andrias loses the more posterior epibranchials (Cox and Tanner, 1989). The ceratohyal becomes thinner and more blade-like anteriorly and narrowed and elongate posteriorly dur-
ing metamorphosis. Ligamentous and connective tissue attachments to surrounding skeletal elements are variable. A hyoquadrate ligament or hyomandibular ligament persists or develops in adults of some taxa, such as plethodontids. This ligament may be replaced in some of the Salamandridae (e.g., Taricha) by a hyosuspensory ligament between the upper quadrate and the posterior tip of the ceratohyal (Findeis and Bemis, 1990). The hypohyals, when present in the larva, may vanish entirely or give rise to adult elements, such as the radial cartilages found in the Ambystomatidae, Rhyacotritonidae, Dicamptodontidae, and Hynobiidae. The ceratohyal loses its connection to the basibranchial, except in hynobiids, ambystomatids, and dicamptodontids. In these taxa, the ceratohyals and basibranchial are connected by narrow radial cartilages (see Batrachuperus, Fig. 3.6). In some hynobiids, these cartilages form elongate loops, which cross in a figure-eight fashion over the midline and connect the two ceratohyals (Cox and Tanner, 1989; Larsen et al., 1996). The developmental appearance of new elements at the rostral end of the hyobranchial apparatus is common among salamanders, and these new elements may articulate with the basibranchial, for example, the second radii, otoglossal cartilage, or lingual plate of the Ambystomatidae and Dicamptodontidae (Krogh and Tanner, 1972) and Salamandridae (Ozeti and Wake, 1969), or may lie adjacent to it, like the basihyals and hypohyals of the Amphiumidae and Cryptobranchidae. In some taxa, anterior projections of the basibranchial, called cornua, form and support the adult tongue pad (Fig. 3.6). In some plethodontids a separate lingual cartilage forms at the tip of the basibranchial and functions in rotation of the tongue pad (Lombard and Wake, 1977). The articulations between the basibranchial and ceratobranchials are altered at metamorphosis. The branchial plate of the larvae of some taxa gives rise to separate basibranchial and ceratobranchials. The urohyal loses its attachment to the basibranchial and may disappear altogether, as in many salamandrids. Complete metamorphosis of the hyobranchial apparatus makes tongue protrusion possible. The ceratohyals are freed somewhat from the remaining elements, allowing folding and forward movements of the rest of the apparatus to which the tongue pad is attached. The ossification pattern of the hyobranchial apparatus varies taxonomically (Fig. 3.6). Only the urohyal, when present, is universally ossified. Aquatic plethodontids, which do not suction feed, such as Desmognathus marmoratus, possess slender cartilaginous hyobranchia; only portions of the basibranchial are ossified. In other taxa, the basibranchial, epibranchial tips.
77
3. A q u a t i c F e e d i n g in S a l a m a n d e r s
Batrachuperus mustersi
Rhyacosiredon rivularis
Cryptobranchus
alleganiensis
Pachytriton sp.
Leurognathus marmoratus
Pleurodeles waltl
FIGURE 3.6. Hyobranchia (in ventral view) of adults of metamorphosing and partially metamorphosing salamanders from five families. Bone or mineralization is shaded and cartilage is unshaded. Some elements are cut away to reveal underlying elements or the outlines of underlying elements are shown as dashed lines. Scale bars equal 1 cm. See Table 3.1 for abbreviations. Cryptobranchus drawing is adapted from Elwood and Cundall (1994). /. Morph. 220, 47-70. Copyright © 1994. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.
and posterior portion of the ceratohyal are frequently ossified. Almost the entire apparatus may ossify in suction-feeding members of the Salamandridae, such as Pachytriton, with only the rostral portion of the ceratohyal remaining cartilaginous (Ozeti and Wake, 1969). Rhyacosiredon has a more robust hyobranchial apparatus than other ambystomatids and shows extensive ossification of the epibranchials and ceratohyal, as well as ossification of the first ceratobranchial. Cryptobranchoids show an unusual, but consistent pattern of ossification of the second ceratobranchial and epibranchial, which may be indicative of an unusual method of force transmission during feeding (see Section III,D).
In partially metamorphosing taxa, such as cryptobranchids and amphiumids, the hyobranchial apparatus of the adult retains some larval features. Hypohyals, as well as basihyals, are present in the adult. Articulations between hyobranchial elements are strengthened during development. The anterior hyobranchial elements of cryptobranchids remain cartilaginous, even in large animals, and are extremely broad, filling the space between the mandibles (Cox and Tanner, 1989; Elwood and Cundall, 1994). The hypohyals of Rhyacosiredon have a morphology intermediate between the larval configuration and that of the adults of fully metamorphosing ambystomatids (Reilly and Brandon, 1994).
78
Stephen M. Deban and David B. Wake
Posterior epibranchials persist in amphiumids, as well as in Cryptobranchus in which the third epibranchial calcifies. The first epibranchial and ceratobranchial fuse to one another to form a single element, the epibranchial, in amphiumids, cryptobranchids, and some hynobiids (Fig. 3.6). In amphiumids, the second ceratobranchial is absent, even in larvae. In sirenids and proteids, the anterior ceratohyal, the basibranchial, and ceratobranchials ossify. In sirenids the basibranchial is particularly unusual: the anterior tip is bulbous where it meets the ceratohyals (Cope, 1889). In proteids the epibranchials also ossify (Fig. 3.4). Some taxa show divergent skull and jaw morphologies. Sirenids, amphiumids, and Proteus have elongate, tapered heads and jaws, whereas cryptobranchids have broad, rounded heads and jaws. Amphiumids have enlarged jaws, teeth, and jaw musculature and are capable of a powerful bite; they can bite violently when handled. Members of the Salamandridae have particularly robust skulls; many possess a bony arch joining the frontal and squamosal and some have the maxilla and pterygoid abutting. Cryptobranchoids possess an ossified angular element in the lower jaw (Noble, 1931), and the mandibular symphysis forms a joint consisting of cartilaginous pads and collagen ligaments, allowing the mandibles to flex considerably ventrally and dorsally with respect to one another (Cundall et al, 1987; Elwood and Cundall, 1994). This unilateral jaw depression is possible only because the mandibles are unusually strongly curved. Their curvature allows them to swing independently of one another, producing an anterolateral gape when one side is depressed relative to the other. Desmognathine plethodontids have specializations associated with a strong bite, including stalked occipital condyles, modified anterior vertebrae, and robust skulls (see Schwenk and Wake, 1993, and references therein). 2. Adult Head
Musculature
Muscles change in size, orientation, and function during metamorphosis and may be lost, form, or give rise to new muscles. In general, more muscles are lost than appear, mostly those associated with gill and epibranchial movements, including the SAR muscles (except SAR 1), AA, and LAB (Fig. 3.5B). Those feeding muscles that typically change during metamorphosis are emphasized here. The BM is lost or shifts its posterior attachment from the epibranchial tip to merge with the DMP and fan out posteriorly onto the fascia cephalodorsalis, thereby losing its role in hyobranchial depression. The new combined muscle is called the DMR In some taxa the DMA and DMP become indistinguishable at meta-
morphosis and are known as the DM thereafter. In desmognathine plethodontids, the tendon associated with the LMIP is enlarged at metamorphosis, forming the distinctive atlanto-mandibular ligament (Schwenk and Wake, 1993). The BH is lost at metamorphosis in metamorphosing and partially metamorphosing taxa. The SAR 1 enlarges, often forming a muscular bulb around the first epibranchial (or ceratobranchial in those species that lack a separate epibranchial), and shifts its function to tongue protrusion. In partially metamorphosing forms (i.e., cryptobranchids and amphiumids) and in many newts, the SAR 1 replaces the BH in position and function (Fig. 3.5B). The remaining SAR muscles are lost. In plethodontids, a geniohyoideus lateralis (GHL) arises from the GH, which then becomes known as the geniohyoideus medialis (GHM). The GHL originates on the mandible lateral to the symphysis and inserts on the lateral edge of the ceratohyal. The GHM retains its larval origin and insertion, but the urohyal separates from the basibranchial before metamorphosis is complete. The posterior attachment of the GH shifts to the RCS in those taxa lacking a urohyal. The GHM loses its function in drawing the hyobranchial apparatus anteriorly and becomes a buccal stabilizer in plethodontids. In other taxa, the GH similarly loses its attachment to the basibranchial and acts as a sling that can stabilize or raise the hyobranchial apparatus. The rectus cervicis (RC) muscles become more distinct as profundus (RCP) and superficialis (RCS). A superficialis lateralis (RCSL) slip may form in plethodontids. The entire group of RC muscles functions in tongue retraction or hyobranchial depression. A hebosteoypsiloideus (HOYP) muscle, the function of which is unknown, becomes distinct from the RC group at metamorphosis in plethodontids, originating on the urohyal (or basibranchial) and inserting on the RCP near the sternum (Lombard and Wake, 1977). Among the dorsal muscles of the head and neck, only the IVE changes significantly at metamorphosis, becoming distinct from the DT. This new muscle functions in cranial elevation. The metamorphic fate of the superficial throat muscles varies among taxa and has been the source of much nomenclatural confusion. In plethodontids, the IHP of the larva gives rise to both the quadratopectoralis (QP) and the gularis (G) of the adult. The G is the only remnant of the IHP to persist in most adult plethodontids. The desmognathines and Aneides possess both; the QP becomes a strong head depressor, shifting its dorsal attachment from the epibranchial to the quadrate, and shifting its caudal attachment to the pectoral fascia. The G in these taxa is a small strap muscle that lies against the QP. In adults of all other
3. Aquatic Feeding in Salamanders salamander taxa (i.e., nonplethodontids), the IHP retains its general larval position or shifts its origin to the fascia cephalodorsalis, but is still called the IHP (Piatt, 1940). In the Salamandridae and Hynobiidae, the IH of the larva splits to give rise to the anterior subhyoideus (SH) and the posterior interossaquadrata (lOQ). The SH originates from the posterior end of the ceratohyal and inserts anteriorly on the dorsal fascia of the IMP or on the mandible near the symphysis, and upon contraction draws the ceratohyal rostrally (Fig. 3.5B). The lOQ may shift its insertion from the ceratohyal to the quadrate. In adults of all other families, the Iti retains its larval position (Krogh and Tanner, 1972) and name. The cephalodorsosubpharyngeus (CDSP) forms from the fusion of the posterior larval TV and LAB muscles and is in a position to constrict the pharynx and aid in swallowing. A genioglossus (GG) forms or is elaborated during metamorphosis, originating on the mandibular symphysis and fanning out into the tongue pad. The GG draws the tongue pad toward the mandibular symphysis. In Cryptobranchus, this muscle inserts on the hyopohyals and acts as a hyobranchial levator (Elwood and Cundall, 1994). In ambystomatids, the GG gives rise to the genioglossus lateralis (GGL), which inserts on or in the vicinity of the lateral edge of the ceratohyal (such as the GHL in plethodontids) (Larsen and Guthrie, 1975). Rhyacotritonids may possess both SH and GGL (Krogh and Tanner, 1972). These muscles and the intrinsic tongue muscles, which also form at metamorphosis, function primarily in tongue prehension and will therefore be discussed in the next chapter. Feeding muscles are divided into two series, branchiomeric and hypobranchial, based on developmental origin and innervation. The branchiomeric (branchial) series includes the majority of the feeding muscles, is innervated by visceral motor components of the nervous system, and forms from the embryonic myomeres. Branchiomeric feeding muscles include IM, IMP, IH, SH, IHP, LM, DM, BH, LAB, and SAR. The hypobranchial series forms from rostral growths of myotomes behind the branchial area, is innervated by somatic motor components of the nervous system (i.e., the hypoglossal nerve), and includes the GH, GHL, GG, GGL, RCP, and RCS and the intrinsic muscles of the tongue pad. C. Sensory and Motor Systems Larval and adult salamanders foraging in water are presented with the same array of sensory stimuli and generally use the same set of sensory modalities to detect and localize prey, but differ in the degree to which
79
they use a particular modality. Larvae and perennibranchiate forms typically rely on mechanical, electrical, and olfactory senses, but can use vision as well (Luthardt-Laimer, 1983; Besharse and Brandon, 1974). Adults of metamorphosing forms rely to a lesser degree on olfactory and lateral line stimuli; they use primarily vision and have well-developed eyes. The eyes of salamanders have all the components of typical vertebrate eyes, including a cornea, lens, iris and pupil, and multilayered retina of cells and fibers. The retina includes photoreceptors (rods and cones), horizontal cells, bipolar cells, amacrine cells, and ganglion cells. The cornea is flat and the lens is spherical, suited for an aquatic environment in which only the lens and not the cornea is responsible for focusing the image (Roth, 1987). Salamanders are capable of color vision, and larvae and aquatic adults are more sensitive to longer wavelengths of light (which predominate in water) than terrestrial adults (Himstedt, 1973a,b). Focusing on near objects such as prey is accomplished by contraction of the protractor lentis muscle, which moves the lens away from the retina. The eyes are reduced and covered with skin in many cave-dwelling salamanders, many of which are perennibranchiate. These species show a decrease in overall eye size and a degeneration of the retina and optical structures. Eyelids are lacking in larval and perennibranchiate salamanders, cryptobranchids and amphiumids; they are reduced in some fully aquatic salamandrids and hynobiids, and Rhyacosiredon, and are present in fully metamorphosed salamanders. The optic nerve, which consists of the axons of the retinal ganglion cells, crosses the optic nerve of the contralateral eye at the optic chiasm and enters the diencephalon from below. Fibers terminate in, or project to, various parts of the diencephalon (e.g., thalamus and pretectum), the optic tectum in the dorsal portion of the mesencephalon, and the tegmentum in the ventral portion. Most fibers of the optic nerve travel to the contralateral side of the thalamus, pretectum, and tectum; however, some ipsilateral projections exist (Fig. 3.7). The proportion of ipsilateral projections varies among taxa and developmental stages. Proportionally more ipsilateral projections are present after metamorphosis in those taxa with larvae. Ipsilateral projections are most abundant in completely terrestrial forms, especially those with greatly overlapping visual fields and binocular vision, such as bolitoglossine plethodontids (Roth, 1987). Information from the optic tectum is relayed through three descending tecto-bulbar tracts to the brain stem where motor nuclei controlling feeding muscles reside (Dicke and Roth, 1994). Salamanders possess two olfactory systems: the
80
Stephen M. Deban and David B. Wake
Olfactory Bulb Vomeronasal
I I Muscles I I Skeletal elements Lateral line system Olfactory system Vomeronasal system Visual system Motor nerves Brain tracts XI Motor nuclei
Lateral Line System
F I G U R E 3.7. Sensory-motor schematic of the feeding system of a generalized salamander, showing both larval and adult features. Sensory information entering the eyes, nostrils, and lateral line organs is integrated in the telencephalon, diencephalon, and mesencephalon. These centers activate motor nuclei (numbered ovals) in the brain stem, which control the contraction of muscles (striped gray) through cranial nerves (black lines) and consequently produce feeding movements of the jaws and hyobranchial apparatus (solid light gray). Nerves and brain tracts are solid black lines or gray channels and muscle innervations are shown as black dots, and sense organs and parts of the brain are shown in white or dark gray. Motor nuclei are numbered according to the cranial nerves they supply See Table 3.1 for muscle abbreviations.
3. Aquatic Feeding in Salamanders primary olfactory system and the vomeronasal organ or accessory olfactory system. The main olfactory system includes the nasal epithelium with chemosensory olfactory cells, each of which has many filaments exposed to the nasal chamber which are capable of responding to molecules in the environment. Axons of olfactory cells form the first cranial nerve, which projects to the olfactory bulbs of the forebrain (Fig. 3.7). The vomeronasal system opens into the nasal chamber and has receptors in a separate sac, lateral to those of the main olfactory system (Dawley, 1988), and has a separate accessory bulb in the brain (Eisthen ei a\., 1994). Plethodontids possess the additional feature of nasolabial grooves, which form at metamorphosis and which conduct substrate-borne molecules to the vomeronasal sac. This vomeronasal system is important in courtship behavior and may be used in prey tracking. Taste buds on the tongue and pharynx are used to reject distastful or noxious prey. The glossopharyngeal nerve (cranial nerve IX) conducts impulses from the taste buds to the brain. Additional taste buds are present on the skin and may be used to locate prey. The lateral line system is present in all aquatic salamanders and is used to detect water currents and electrical charges. The lateral line system is closely associated with the auditory system and the two are collectively called the acousticolateralis system. Lateral line detectors form series of small depressions or raised "stitches" on the head and along the body, containing ampuUary organs, neuromasts, or pit organs. Ampullary organs are present on the head and function as electroreceptors, which can detect the tiny electrical fields generated by the muscle contractions of prey up to a few centimeters away (Bartels ei al., 1990). Neuromasts and pit organs extend down the head, body, and tail and are mechanoreceptive, containing sensory hair cells that are responsive to water movements and pressure changes. Larval salamanders possess all three types of receptors, as do some aquatic adult salamanders (Fritzsch and Wahnschaffe, 1983). Lateral line organs become covered with epidermis in salamanders that spend time out of water, such as Notophthaltnus and Siren, but become exposed to the skin surface and become functional again when the salamander returns to water (Dawson, 1936; Reno and Middleton, 1973). Axons of sensory cells of the lateral line system enter the brain through five separate nerves in the vicinity of the auditory nerve (VIII) and project to the acousticolateral area of the somatic sensory column of the medulla (Northcutt and Brandle, 1995). Lateral line information is relayed to the tectum of the midbrain, where it is processed and where the lateral line sensors
81
are represented as a topographical map (Bartels et al., 1990). Salamanders use lateral line information to detect predators as well as to localize and orient to moving prey and to direct the predatory strike (Himstedt et al, 1982). Salamanders are capable of detecting prey using visual, olfactory, tactile, and lateral line receptors and are most responsive to live, active prey. Aquatic salamanders use olfactory and visual cues for initial orientation to prey and to direct the prey-capture strike (Martin et ah, 1974), but also use mechanoreception and electroreception (Joly, 1981; Himstedt,1967; Griffiths, 1993). Cave salamanders may have slender, elongate legs, which are used during foraging to lift the body above the substrate and enhance mechanoreception by exposing more of the lateral line system to water movements (Peck, 1973). Blind urodeles, such as Proteus, are sensitive to olfactory stimuli and will respond readily to immobile, dead prey, whereas salamanders that use vision are less responsive to dead prey and olfactory cues and rely on prey movement (Durand et al., 1982; Uiblein et al., 1992). In light, vision dominates prey localization and inhibits olfactory detection; darkness (or blindness) appears to release this inhibition (Roth, 1987). Feeding movements are controlled by clusters of motor neurons or motor nuclei of the brain stem (Roth et al., 1990). Motor nuclei of the feeding muscles are arranged in two columns along each side of the brain stem from the medulla oblongata to the second spinal nerve (Fig. 3.7). The motor nuclei supply axons to the cranial and first and second spinal nerves, which innervate the feeding muscles (Piatt, 1938). The trigeminal (cranial nerve V) innervates the IMP and LM. The facial (VII) innervates the IH, SH, BH, and DM. The glossopharyngeal (IX) innervates the SAR 1. The vagus (X) innervates the AA, LAB, SAR 2-SAR 4, TV, CDSP, and the SAR 1 of adults of some taxa (e.g., plethodontids). The accessory (XI) serves the cucuUaris. The hypoglossus (XII, formed from first and second spinal nerves) innervates the GG, GGL, GHM, GHL, RCP, RCS, and muscles of the tongue pad. Salamanders must gather sensory information about their prey and produce the approprialte behavior to make a successful capture. Information Is perceived by the sense organs and is processed in the brain, which then activates motor nuclei in the brain stem, contracting muscles in the proper sequence to generate coordinated movements of the body, head, jaws, and throat. Most if not all of the sensory and motor components of the feeding system are known. Flowever, the specific nuclei and connections of the brain that are involved in integrating the various sensory inputs (olfaction.
82
Stephen M. Deban and David B. Wake
lateral line, vision) and in producing the appropriate activation of motor nuclei are as yet undetermined. Sensory-motor integration in amphibians in general remains a rich area of research. III. FUNCTION A. Ingestion Behavior and Kinematics The feeding behavior of salamanders has been divided into four stages: orientation, approach, fixation, and snapping (Roth, 1987). Orientation involves lifting the head from the substrate and directing it toward the prey item. Approach involves walking or swimming toward the prey item until it is within reach. Fixation is a period during which the salamander assesses the prey, usually while the prey is stationary. Snapping, or striking, is the ingestion behavior: suction feeding or tongue or jaw prehension, sometimes combined with a lunge. Snapping is typically triggered, or "released," by prey movement. Snapping usually occurs so rapidly that it is effectively invisible to the eye and has been difficult to describe or analyze until recent technological advances in high-speed cinematography and videography. The other stages occur at a speed that is observable to the unaided eye. The four stages are often discrete, but some may be absent or blend together. For example, if the prey is already within reach, the approach may be omitted, or if the prey is moving continuously the fixation period may be skipped. Salamanders may also perform an olfactory test, in which the snout is pressed downward onto or near the prey, after the fixation stage but before striking at stationary prey. 1. Suction Feeding Snapping in most aquatic salamanders involves generating intraoral suction to capture prey. During suction feeding, the mouth is opened while the buccal cavity is expanded. This expansion creates a drop in buccal pressure relative to ambient pressure, causing water and prey to flow into the mouth. The gill slits are closed and labial lobes restrict the mouth opening to the front. The head is raised during mouth opening. This allows for ventral excursion of the lower jaw and also serves to aim the gape at the prey. The buccal cavity continues expanding as the mouth is closed and the head is lowered, resulting in continuous water flow throughout the gape cycle. The gill slits open before mouth closure and before maximum buccal expansion. After the mouth closes, water is expelled slowly through the open gill slits by compression of the buccal
cavity. Salamanders that lack or have a reduced number of gill slits expel water slowly out of the mouth, which is held ajar. Gill rakers or teeth prevent the prey from escaping during water expulsion. Suction feeding requires special morphology to be successful, including a robust hyobranchial apparatus and associated musculature to generate rapid and forceful buccal expansion and labial lobes or other means of restricting water flow to the front of the mouth. The coordination of buccal expansion and mouth opening is also important for successful suction feeding, although some variation in these two behaviors exists. Variation in suction feeding behavior across taxa or developmental stages involves changes in the extent, timing, and duration of jaw and hyobranchial movements during the prey-capture strike. Comparative studies of feeding behavior make use of kinematic analysis of the movements of the jaws and throat during prey capture. These movements are confined primarily to the sagittal plane, so are best visualized in lateral view, and are typically rapid and must be captured by high-speed videography or cine. Gape cycle durations for some representative taxa are as follows: 17-30 msec in larval Desmognathus quadramaculatus (Deban, personal observation), 35-47 msec in larval Salamandra salamandra (Reilly, 1995), 40 to over 80 msec in adult Cryptobranchus (Elwood and Cundall, 1994) and Amphiuma tridactylum (Erdmann and Cundall, 1984), and 80-110 msec in larval Ambystoma mexicanum (Lauder and Shaffer, 1985). Quantification of the strike by kinematic analysis involves recording the positions of various anatomical points throughout the course of the behavior (i.e., on each frame of a video sequence). Displacements, distances, angles, and velocities can be calculated from these position data, as can the timing and duration of events. Quantification of behavior in this way facilitates comparisons across taxa or developmental stages. Such data can then be visualized by kinematic profiles, which depict distances or the positions of points through time (Fig. 3.8). Two measures typically made on suction feeding salamanders are gape distance, or the distance between the tips of the jaws, and hyobranchial depression distance, or the distance between the top of the neck and the ventralmost point on the throat. Gape distance and hyobranchial depression distance both increase and then decrease during the suction feeding behavior. The kinematic profile of gape distance (the gape profile) typically appears as a bellshaped curve (Fig. 3.8). The hyobranchial profile appears as a skewed curve, showing a rapid increase and a slower decrease. By plotting both kinematic profiles
83
3. A q u a t i c F e e d i n g in S a l a m a n d e r s
2.0
1
1
1
1
1
Hyobranchial Depression E 1.5 E CD O
c
r
1.0
1
/
CO
-t—'
Q 0.5
r-"
h-D-^
t^^.^ ^
J
r
4
h
L
Gape /Distance
0.0 0
4
8
12
16
20
24
28
32
36
40
44
0
4
8
12
16
20
24
28
32
36
40
44
Time (ms) F I G U R E 3.8. Kinematic profiles of jaw, hyobranchial, and head movements through the course of a single suction-feeding event in a larval Desmognathus quadramaculatus (snout-vent length is 30 mm). Hyobranchial depression that continues into mouth closing is typical of suction feeding. The head angle increases as the head is raised during mouth opening, directing the gape at the prey, and later decreases as the head is tucked during and after mouth closing.
against the same time axis, the temporal relationship of the two movements can be seen. Figure 3.8 shows the typical pattern for suction feeding: mouth opening and closing occur much more rapidly than hyobranchial depression and elevation, and maximum gape is reached before maximum hyobranchial depression. Kinematic profiles of this type are used to examine how the relationship between the two movements might vary among taxa or with different prey types. In addition to gape and hyobranchial distance, other kinematic variables differ among suction feeding salamanders, including the angle between head and body (Fig. 3.8) and the distance of forward lunging. For example, Cryptobranchus shows especially large buccal expansion for its size, and both Cryptobranchus and Ampliiuma display pronounced head dipping on mouth closing (Reilly and Lauder, 1992). Cryptobranchids are unique in that they often display asymmetrical kinematics during capture and intraoral transport in which one side of the lower jaw or hyobranchial apparatus is abducted more rapidly or farther than the other side (Elwood and Cundall, 1994).
2. Tongue Prehension Tongue prehension is used to capture prey by all terrestrial salamanders and by several aquatic salamanders (e.g., some plethodontids and rhyacotritonids) (Schwenk and Wake, 1988; personal observation). The behavior involves forward movement of the hyobranchial apparatus which propels the tongue out of the mouth. The SAR and SH muscles protract the tongue, and the RCS and RCP muscles retract it. Several muscles of the tongue tip rotate and shape the tongue pad. The gape profile is usually not bell shaped as in suction feeding, but often has a plateau region of constant gape corresponding with tongue protrusion. Gape increases to its maximum not as the tongue is protruded, but rather as the tongue returns with the prey to the mouth, producing a three- or four-part gape profile. 3. Jaw Prehension Prey capture by jaw prehension is used in both aquatic and terrestrial feeding and involves movements of
84
Stephen M. Deban and David B. Wake
the jaws and Httle movement of the hyobranchial apparatus. A forward lunge of the body or lateral turn of the head is used to bring the jaws to the prey. The gape cycle produces a bell-shaped profile similar to suction feeding, but the hyobranchial apparatus shows extremely small excursions relative to those seen during suction feeding. B. Prey Processing After prey is captured, it must be prepared for swallowing. Salamanders usually swallow their prey whole, without chewing or reducing it, although there are some exceptions, such as cryptobranchids and amphiumids. Processing prey usually involves only positioning it in the oral cavity and lubricating it with mucus. Occasionally, prey that protrudes from the mouth after capture is pushed against the substrate to orient it or to move it into the mouth. The head is sometimes thrashed laterally to orient the prey, to reduce it, or to break it free from the substrate. Prey is sometimes moved into and out of the mouth across the teeth repeatedly by water flow to center it; this may have the effect of macerating the prey. The forelimbs are not used in prey manipulation. Like prey capture, intraoral transport of prey involves coordinated movements of the jaws and hyobranchial apparatus and can be accomplished by suction or tongue movements. Suction-based transport of prey within the mouth to the esophagus is achieved by movements similar to suction-feeding movements (Gillis and Lauder, 1994). Prey is moved toward the esophagus by currents of inflowing water and held with the teeth or pressed to the roof of the mouth until the next cycle of intraoral transport. This process is repeated until the prey is in a position to be swallowed. Lingual-based prey transport (hyolingual transport) is accomplished by rapid cyclical movements of the hyobranchial apparatus and tongue. The prey adheres to the tongue pad and is moved caudally and ventrally in the mouth. As the tongue is moved forward for the next cycle, the teeth and jaws prevent the prey from being pushed forward. This ratchet-like process slowly moves prey toward the esophagus. Movements of the cranium, for example, head tucking in desmognathine plethodontids (Schwenk and Wake, 1993), may assist in forcing prey toward the rear of the throat. The tongue is first moved forward in the mouth, pressed against the prey, and then retracted rapidly, producing a drop in the floor of the buccal cavity similar to that seen in suction feeding. This process is repeated until the prey is swallowed. Swallowing involves forcing prey down the esophagus by contraction of the pharyngeal musculature, re-
traction of the eyeballs, and movements of the head. The tongue may be used as well in salamanders that possess it. Swallowing is similar in both terrestrial and aquatic salamanders, involving contraction of the transverse throat musculature (IM, IH, G, QP, CDSP, TV) and retractor bulbi, which squeeze prey down the esophagus. The dilator and constrictor laryngis muscles open and close the larynx and glottis except in plethodontids, which lack these muscles. Peristaltic contractions of the smooth muscles of the esophagus then move prey to the stomach. C. Functional Morphology Prey capture, intraoral transport, and processing are accomplished by movements of the jaws and hyobranchial apparatus. The type of hyobranchial movements and relative timing of jaw and hyobranchial movements differ among the various feeding behaviors (suction feeding, tongue prehension, and jaw prehension). The hyobranchial apparatus is thrust forward (tongue prehension), swung backward (suction feeding), or held relatively stationary (jaw prehension) by different patterns of muscle activity. The muscles involved in each of these behaviors, as well as their hypothesized roles in producing feeding movements, are discussed in this section. Feeding movements are discussed as though they are symmetrical, although some taxa, such as cryptobranchids, are known to move asymmetrically (e.g., Cundall et ah, 1987). Again, suction feeding is emphasized over tongue and jaw prehension; the latter behaviors are covered in the next chapter. The buccal expansion of suction feeding is produced by the hyobranchial apparatus and its associated musculature (Fig. 3.5). The hyobranchial apparatus at rest is folded and lies in the floor of the mouth and the throat and rather close to the palate. During buccal expansion, it is swung ventrally and caudally about its attachments to the lower jaw, skull, and neck. The attachments of the ceratohyal to the jaw joint and of the epibranchials to the neck and skull form the dorsal points of rotation for this hyobranchial depression. The basibranchial retains its horizontal orientation, but is translated downward and backward. In addition to rotation around joints, elasticity of the cartilaginous elements has been proposed to play a role in some movements of the hyobranchial apparatus, such as opening and closing of the gill slits (Severtsov, 1966). The entire apparatus may move slightly dorsoventrally as well because the points of rotation at the jaw joint and neck are somewhat mobile. Also, the distal tips of the ceratohyals and epibranchials may move laterally during buccal expansion, broadening the pharynx. Both the
85
3. Aquatic Feeding in Salamanders dorsoventral shifting of the hyobranchial apparatus and the lateral expansion of the pharynx contribute only slightly to buccal expansion, and will not be discussed further; ventrocaudal rotation of the hyobranchial apparatus is considered the primary movement producing buccal expansion. As the hyobranchial apparatus rotates and drops, it applies a downward force to the buccal lining and the skin of the throat, which unfold and expand ventrally to produce the volume change that draws water, and usually prey, into the mouth. When the buccal cavity has expanded fully, it begins to compress by elevation of the hyobranchial apparatus, which is accomplished like hyobranchial depression, only in reverse and much more slowly. Water is expelled through the gill slits or mouth and prey is retained by the gill rakers or teeth. Mouth opening and closing must be coordinated with hyobranchial depression and elevation to ensure capture and prevent subsequent escape of the prey. The mouth is opened and closed simply by rotation of the mandible about its articulation with the skull. The mouth is opened and begins to close during buccal expansion and is kept closed or ajar during buccal compression. If a prey item is not brought entirely into the mouth during the initial strike, it is gripped by the opposing teeth of the dentary on the lower jaw and the premaxilla and maxilla (when present) on the upper jaw. Some large salamanders, such as amphiumids and cryptobranchids, have robust jaws and teeth that can be used for crushing or tearing prey. Labial lobes are critical for effective suction feeding (Matthes, 1934; Miller and Larsen, 1989). They effectively prevent water from entering the sides of the mouth and restrict water flow to the front. By reducing the area of the aperture through which a given volume of water must flow in a certain period of time, labial lobes increase the velocity of the flow and thus increase the probability that elusive prey will be entrained. Labial lobes are so frequently present in suction feeders that the presence of these structures in the absence of behavioral data is evidence that the salamander suction feeds. Salamanders that lack labial lobes may, however, use suction when feeding in water, for example, breeding Ambystoma (Miller and Larsen, 1986). The importance of labial lobes in suction feeding is reflected in the poor prey-capture success of these salamanders that lack them (Reilly and Lauder, 1988a). Other morphological and functional features may also enhance suction-feeding ability. In amphiumids, the maxillae are attached loosely to the rest of the skull and undergo medial displacement during suction feeding, consequently reducing frontal gape and potentially improving capture success (Erdmann and Cundall, 1984). Cryptobranchids are unique in that they
can depress the mandibles asymmetrically, opening only one-half of the mouth at a time, a behavior observed most often during prey manipulation. This behavior may serve the same function as labial lobes in other taxa, restricting the gape and accelerating water flow. Cryptobranchids can also depress the hyobranchial apparatus asymmetrically during prey capture and manipulation (Cundall et al, 1987). The ability of cryptobranchids to move the jaws and hyobranchial apparatus unilaterally indicates that muscles may not be functioning symmetrically during feeding, as many functional studies tend to assume. The gill slits of many aquatic salamanders have also been proposed to contribute to suction-feeding performance (Lauder and Shaffer, 1986; Reilly and Lauder, 1988a). Gill slits have been shown to open during suction feeding before maximum buccal expansion. This timing of gill slit opening would allow the momentum of the primary flow to carry water posteriorly out the gill slits, preventing turbulence that could disrupt flow into the mouth (MuUer and Osse, 1984). Alternatively, the momentum of the primary water flow can be absorbed by a relatively enormous buccal expansion, in which water is stored for later expulsion during buccal compression. This latter mechanism likely operates in Cryptobranchids, which have no gill slits or only a tiny spiracle, and in metamorphosed newts, which lack gill slits. The exact biomechanical role of gill slits in suction feeding remains to be determined and likely differs in taxa with different morphologies. Gill slits are clearly not always necessary for successful suction feeding in salamanders, given the suction-feeding prowess of some postmetamorphic salamandrids that lack gill slits, many of which suction feed as effectively as larvae (Miller and Larsen, 1989). D . Biomechanics Analyses of feeding kinematics and anatomy of the feeding system can be combined to produce a model of the biomechanical events of suction feeding. This section presents such a model for salamanders, which is intended to be used as a heuristic device to identify developmental and evolutionary transitions in hyobranchial function. It is to be used as a foundation on which to build hypotheses of functional transformation, rather than as a precise representation of all of the events occurring during suction feeding. The skeletal elements of the skull and hyobranchial apparatus of larval salamanders form a system of rigid elements on which the muscles act to produce movments. Suction feeding involves primarily rotational movements of the jaws and hyobranchial apparatus (Fig. 3.9A) and can be modeled as a system of levers.
86
Stephen M. Deban and David B. Wake
IHP RCP F I G U R E 3.9. Mechanics of hyobranchial depression during suction feeding, showing hyobranchial apparatus and jaws in the elevated position on the left and the depressed position on the right. Larval skull, jaws, and hyobranchial elements, shown in lateral view (A), are modeled as systems of levers (B). Muscles (black lines) contract to produce movements of the levers. Muscles shown on figures on the left produce movements that place the elements in the positions shown on the right, and vice versa. Oblique views (C) show displacements of the hyobranchial apparatus in three dimensions, including twisting of vertical elements about their long axes.
The skull and jaw^s form a single lever system. The hyobranchial apparatus, together with the skull and neck, can be represented in simplified form as a fourbar mechanism (Fig. 3.9B) in w^hich rotation of any one element causes equal rotation of the parallel element. The skull and neck, taken together, can be considered a single component, which forms the dorsal element of the mechanism. The basibranchial forms the ventral element. These two horizontal elements are linked by the paired ceratohyals and a pair of epibranchials (either the first or the second pair), which form the rostral and caudal elements, respectively. Each of the four elements of the mechanism articulates with neighboring elements and rotates with respect to them. The buccal cavity is expanded by parallel rotation of the rostral
and caudal elements, anchored by the dorsal element and linked by the ventral element, which undergoes ventrocaudal translation without rotation. The rostral and caudal elements are bilaterally paired, but articulate with the single, medial ventral and dorsal elements. The entire system thus resembles two four-bar linkages functioning in parallel, sharing some components (Fig. 3.9C). The jaw and hyobranchial lever systems act in concert to produce the movements of suction feeding. Movement of the skeletal levers is accomplished by contraction of the associated musculature. In the case of the jaws, the DM and LM act as antagonists to open and close the mouth, respectively. The GH is in a position to open the mouth as well, via its attachment to
87
3. Aquatic Feeding in Salamanders the mandible (Fig. 3.9B). The mechanical advantage of the GH in mouth opening is increased when the hyobranchial apparatus is depressed. The muscles of the neck direct the gape dorsally, ventrally, or laterally, and some (the DT and ITCS) are in a position to provide resistance to skull depression when the GFi is used in mouth opening. The hyobranchial apparatus is swung ventrocaudally by the RCP, RCS, and BH and is swung dorsorostrally by the GH, IMP, IFF, and IFFP. The RCP is attached to the basibranchial and pulls that ventral element caudally, forcing it to pivot about its attachments to the rostral and caudal elements, which, in turn, pivot about their attachments to the skull and neck. The BH is attached to the tip of the first epibranchial and to the rostral tip of the ceratohyal and contracts to pull the diagonal corners of the mechanism toward one another, forcing the basibranchial ventrocaudally. The BH and RCP both contribute to the same resultant force vector and produce the same ventrocaudal movement of the basibranchial. These two muscles are likely to be coactive during feeding, together producing a powerful expansion of the buccal cavity. In addition to forcing the hyobranchial apparatus ventrocaudally, the BH has two further roles in buccal expansion. It is in a position to pull the tips of the epibranchials ventrally and to rotate the flat ceratohyals about their long axes so that their medial edges are directed ventrally when the hyobranchial apparatus is fully depressed. This latter action brings the broad surfaces of the ceratohyals into contact with the buccal walls, supporting the buccal lining and aiding in expansion. During the first part of buccal expansion the gill slits are held closed by the SAR muscles, which adduct the epibranchials. Acting in direct opposition to the RCP is the GH, which draws the basibranchial rostrally via its attachment to the urohyal. In adults in which the urohyal loses its attachment to the basibranchial, the GH retains its larval role but functions by cradling the hyobranchial apparatus rather than pulling directly on it. The IMP, IH, and IHP also do not attach directly to the linkage, but all cradle the ventral and caudal elements and exert dorsal components of force, moving the basibranchial rostrodorsally and reducing buccal volume. The LAB muscles are in a position to pull the epibranchial tips dorsally, further compressing the buccal cavity. The lever system of the jaws is common to all salamanders, whereas the four-bar linkage system of the hyobranchial apparatus is present in its purest form in larval salamanders. The first epibranchial is usually the most robust element of the larval hyobranchial apparatus and forms the caudal component of the linkage system. The ceratobranchials may contribute either to
the ventral element, as in the branchial plate of plethodontids, or to the caudal element as in amphiumids, in which they are fused with the epibranchials. Some larval salamanders possess additional skeletal elements, which lie between components of the linkage, but these act as part of one of the elements of the system and do not alter its mechanical function as proposed here. Adult salamanders may lack one or more connections in the linkage, making the system a series of levers. Some aquatic salamandrids, for example, lack a tight connection between the ceratohyals (rostral element) and the basibranchial (ventral element). Adults also lack the BH, which pulls diagonal corners toward one another; however, the SAR enlarges at metamorphosis to replace the BH in both position and function. All adults retain the primary muscle of the system, the RCP, which alone can swing the entire linkage ventrocaudally. In many larvae both the first epibranchial and the ceratohyal are equally robust and probably bear comparable loads during buccal expansion. However, the relative robustness of these elements is variable among taxa and across developmental stages, reflected in the extent of ossification of the elements or the crosssectional shape and area. In some groups, the ceratohyals are more robust than the epibranchials, whereas in others the reverse is true. During development and evolution, the primary axis of the linkage system may shift from one set of skeletal elements to another or from the rostral to the caudal components. Cryptobranchoids, for example, are unusual in having the caudal component of the linkage be the second ceratobranchial and epibranchial, rather than the first as in other taxa. This caudal component is ossified and forms the main load-bearing component; the rostral component, the ceratohyal, is flattened and cartilaginous. In larval salamandrids, the ceratohyal is more robust than the epibranchials and is thus likely to bear more load, whereas in adults that feed aquatically, the epibranchials form the primary load-bearing axis of the system and are more fully ossified than the ceratohyal. In taxa that also feed on land using tongue protraction, the rostral portions of the hyobranchial apparatus are generally more flexible and less capable of bearing compressive loads, whereas the caudal components are more robust, bearing loads during both suction feeding and tongue protraction. E. Metamorphosis Following metamorphosis, many salamanders leave the water to forage on land. Suction feeding, the primary mode of aquatic salamanders, relies on the incompressibility of water to move prey into the mouth
88
Stephen M. Deban and David B. Wake
and therefore must be abandoned and replaced by a mode that can operate in air. Terrestrial salamanders must capture prey by either tongue or jaw prehension, behaviors that make use of many of the same morphological components as suction feeding. The jaws are strengthened at metamorphosis and assume more importance in prey capture. The hyobranchial apparatus is remodeled extensively for propelling the tongue from the mouth, but in many cases can continue to function in buccal expansion. In general, the skull and jaws begin a gradual metamorphosis before the hyobranchial apparatus, which undergoes a more abrupt transformation (Rose, 1996). Suction-feeding performance can thus be maintained at a high level late into larval life in those taxa that make a developmental transition to tongue or jaw prehension. In some taxa, such as plethodontids, suction-feeding function is lost completely as the hyobranchial apparatus transforms and tongue function is emphasized. Even completely aquatic adult plethodontids lack the ability to suction feed and use tongue or jaw prehension to capture prey in water (Deban and Marks, 1992). F. Performance Many of the changes that occur during metamorphosis can potentially affect suction-feeding performance, including the loss of labial lobes, transformation of the hyobranchial apparatus, development of the tongue, reduction in buccal volume, and closure of the gill slits. Larvae and perennibranchiate salamanders, which undergo few of these changes, are the most adept at suction feeding. Larval Notophthalmus viridescens capture worms in up to 99% of their attempts (Reilly and Lauder, 1988a), larval Salamandra salamandra up to 91% of attempts (Reilly, 1995), and larval Amybstoma tigrinum capture worms in 93% of attempts and can catch even highly elusive prey such as fish 33% of the time (Lauder and Shaffer, 1986). Partially metamorphosing taxa such as amphiumids and cryptobranchids perform as well as larvae, even though they may lack some larval features. The suction-feeding performance of aquatic adult salamandrids is variable but comparable to that of larvae and is correlated with the degree of head tapering and hyobranchial ossification, and the size of the labial lobes (Miller and Larsen, 1989). Adult semiaquatic newts with rounded heads and modest labial lobes {Taricha, Cynops, Notophthalmus, Paramesotriton, and Pleurodeles) capture earthworm pieces 90-97% of attempts, fairy shrimp 3 8 58% of attempts, and fish 13-23% of attempts, whereas the fully aquatic Pachytriton, with a strongly tapered snout, fully ossified hyobranchial apparatus, and large labial lobes, captures these same prey 100,90, and 65%
of attempts, respectively. The absence of gill slits in the adult newts does not affect their suction-feeding performance (Miller and Larsen, 1988; Reilly and Lauder, 1988a). G. Variation Larger suction-feeding salamanders show considerable variation in the relative timing of events during prey capture. This variation across prey type and among individuals can be due to the salamander actively modulating its behavior or selecting from a repertoire of stereotyped behaviors based on an assessment of conditions before initiating a predatory strike. The high speed of suction-feeding movements is likely to preclude any adjustments via sensory feedback during the strike. Cryptobranchus and Amphiuma show variation in the timing of movements of the hyobranchial apparatus, jaws, and body. When capturing slowly moving or stationary prey, they show little forward movement and mouth opening occurs before buccal expansion begins. When feeding on elusive prey, these salamanders exhibit forward lunging and synchronous mouth opening and buccal expansion. Mouth opening and buccal expansion are greater when feeding on elusive prey compared to sluggish prey (Erdman and Cundall, 1984; Elwood and Cundall, 1994). Slight individual variation in muscle activity patterns has been recorded in larval Ambystoma feeding on the substrate, but variation due to prey type was not detected (Reilly and Lauder, 1989). Ambystoma larvae have been observed to forage in the water column and strike at prey from a distance (Hoff et ah, 1985), suggesting that additional variation in prey-capture behavior remains to be investigated in some taxa. Explicit studies of variation are few (Erdmann and Cundall, 1984; Shaffer and Lauder, 1985a,b; Elwood and Cundall, 1994), however, leading to the impression that salamander prey-capture is an invariant, stereotyped behavior. This notion is gradually being dispelled as more incidental kinematic evidence of variation is gathered. Nonetheless, studies directly examining the taxonomic and developmental diversity of both preycapture performance and variation are needed.
IV. DIVERSITY A N D EVOLUTION The preceding sections have presented the general pattern of morphology, function, and behavior of the feeding systems of aquatic salamanders. Characters have been emphasized over taxa. The following sections are designed to place these characters in the context of taxa that possess them. Unique features of each
3. Aquatic Feeding in Salamanders family, common patterns of evolution (convergence and homoplasy) within the Caudata, as well as synapomorphies of larger clades (above the family level) are discussed. Features common to most families have already been covered. A. Features of Salamander Families Each of the 10 families of salamanders possesses unusual or unique features of feeding biology. In the following accounts the developmental patterns (i.e., perennibranchiate, paedomorphic, metamorphic) of each family is discussed as well as its bearing on feeding form and function. 1. Sirenidae Sirenids are dwarf {Pseudohranchus, two species) or giant {Siren, two species) paedomorphic, perennibranchiate, fully aquatic salamanders that are elongate and lack hindlimbs entirely. They possess a fundamentally larval skull and hyobranchial configuration, which are unusual in many respects (Reilly and Altig, 1996). They possess premaxillae and robust tooth-bearing palatines in the upper jaw. The frontal processes of the premaxillae border the nasals laterally rather than medially as in most other salamanders. The skull is strongly tapered rostrally, and the lower jaw is nearly triangular and rather small, bearing teeth only on the coronoid (in Siren) and with the unusual ball-andsocket jaw joint far anterior. Maxillae are lacking in Pseudohranchus but are present in Siren as tiny, freefloating elements (Parker, 1882; Larsen, 1963; Duellman and Trueb, 1986). The hyobranchial apparatus is extensively ossified and the anterior tip of the basibranchial is bulbous where it articulates with the ceratohyals. Three epibranchials are present in Pseudohranchus and four in Siren. The labial lobes are the largest of any salamander, larval or adult, with upper lobes overlapping the lower jaw to such a degree that they almost meet ventrally. Sirenids are powerful suction feeders, due to the combination of the robust hyobranchial apparatus and the entirely frontal gape afforded by the immense labial lobes. The Sirenidae is the sister taxon to all other salamanders (Fig. 3.1), but its morphology is unsual and not representative of the ancestral condition. 2.
Hynobiidae
The sister group of the Cryptobranchidae, the Hynobiidae, is a large family with about 39 species in seven genera. Some perennibranchiate populations of Hynohius exist. Most species are metamorphic, how-
89
ever, and many are aquatic or semiaquatic. Little is known of their feeding behavior. The semiaquatic Salamandrella keyserlingii and Batrachuperus persicus use jaw prehension in water (Deban, personal observation), whereas the fully aquatic Batrachuperus and Pachyhynohius possess pronounced labial lobes and a pleated buccal lining, morphology that is consistent with suction feeding. Fiynobiids possess complex skulls and hyobranchia, with more elements than the other salamanders (Fig. 3.3), a condition that has led, in part, to the conclusion that they are basal within the Caudata (with the cryptobranchids) after the Sirenidae (Fig. 3.1). Two epibranchials are present in adults, compared to one in all other metamorphosing taxa. The second ceratobranchials and epibranchials are ossified whereas the first are cartilaginous, a feature shared with the Cryptobranchidae. Free radii are completely lacking, but unusual, elongate, radial cartilages connect the ceratohyals to one another or to the basibranchial. 3.
Cryptobranchidae
The Cryptobranchidae is a small family with two genera and three species, all aquatic. Cryptobranchids possess unusual behavioral and morphological features related to aquatic feeding. Not only do they reach enormous size as adults (up to 75 cm in Cryptohranchus and 150 cm in Andrias), but they also have the largest buccal expansion relative to body size of any salamander, due in part to the broad, flat head and pleated buccal lining. As a consequence they are supreme suctionfeeders and are unique in that they can direct their strike to the side using unilateral jaw and hyobranchial depression (Elwood and Cundall, 1994). This behavior is facilitated by the loose mandibular symphysis, strongly curved mandibles, and broad hyobranchial elements. The pattern of hyobranchial ossification is unusual and shared with the Hynobiidae: the first ceratobranchial and epibranchial (which are fused to one another) are cartilaginous and the second are bony. This second set acts as the primary lever in hyobranchial depression, as compared to the first set in most other salamanders. Cryptobranchids are essentially metamorphic in their skull morphology, but retain other larval features, such as lidless eyes, a poorly developed tongue pad, lower labial lobes, posterior branchial elements, and spiracles; the last two features are present in Cryptohranchus and absent in Andrias. Cryptobranchids will scavenge, but can also capture elusive prey such as fish and can handle aggressive prey such as crayfish. They are unusual among suction-feeding salamanders in that they can modulate their feeding behavior (the timing and extent of jaw
90
Stephen M. Deban and David B. Wake
and hyobranchial movements), depending on the elusiveness of the prey. 4. Proteidae The Proteidae includes the genera Necturus (five species) and Proteus (one species), completely aquatic paedomorphic perennibranchiates with an essentially larval morphology including labial lobes and two gill slits. The hyobranchial apparatus is heavily ossified and retains the larval configuration throughout life. Proteids have only three epibranchials, a feature shared only with some plethodontid larvae and Pseudobranchus (Fig. 3.4). The upper jaw lacks maxillae but contains massive, dentate palatopterygoids that replace them in function. Proteids are strong suction feeders. Proteus is elongate, blind, and cavernicolous and is adept at locating prey in complete darkness by olfactory and lateral line senses (Durand et ah, 1982; Uiblein and Parzefall, 1993). 5.
Amphiumidae
The Amphiumidae is a small family with three species, all aquatic, elongate, and with four tiny limbs and a reduced number of toes. Amphiumids resemble cryptobranchids in behavior and in their degree and pattern of paedmorphosis, despite marked differences in morphology and ecology. They retain some larval features (lidless eyes, small tongue pad, interlocking labial lobes, spiracles, and posterior branchial elements) but have solid and robust skulls with some metamorphic features. Amphiumids are strong suction feeders and, like cryptobranchids, can modulate their behavior depending on prey type (Erdmann and Cundall, 1984). They have powerful jaws and teeth and can capture and crush formidable prey, and can bite viciously (Conant and Collins, 1991). The skull is elongate and tapered anteriorly, quite unlike the rounded, flattened skull of cryptobranchids. 6.
Salamandridae
The Salamandridae is a large and diverse family with 14 genera and about 45 species, most of which are aquatic or semiaquatic. The majority of species are fully metamorphic and possess the adult configuration of hyobranchial elements, although some populations are perennibranchiate. Salamandrids possess the full range of hyobranchial morphologies, from almost entirely cartilaginous, slender apparati that are used in long-distance tongue protraction (e.g., Salamandrina and Chioglossa) to the almost fully ossified, robust apparati used for powerful suction feeding (e.g., Pachytriton and Cynops). Salamandrids of the latter type
are perhaps the most proficient suction feeders of all metamorphic salamanders. Taxa that use both suction feeding and tongue protraction (e.g., Taricha and Notophthalmus) possess an intermediate hyobranchial morphology. This morphology must play two disparate feeding roles and represents a functional compromise (Findeis and Bemis, 1990). Salamandrids that feed aquatically after metamorphosis develop labial lobes on the upper jaws that are functionally similar to larval lobes (Ozeti and Wake, 1969) and increase suctionfeeding performance. Those that return to the water to breed develop these lobes temporarily but resorb them rapidly (within 48 hr) on resuming a terrestrial existence (Matthes, 1934). A terrestrial eft stage is present after metamorphosis in many taxa that are aquatic as adults (e.g., Notophthalmus), during which time prey is captured with the tongue. A particularly unsual situation occurs in Pachytriton, in which adults have a tiny tongue pad and completely lack tongue protraction ability as do their larvae. However, a terrestrial eft stage occurs after metamorphosis in which tongue prehension presumably is used (Thiesmeier and Romberg 1997), suggesting that a secondary loss of tongue structures accompanies the return to a permanently aquatic lifestyle. The only viviparous salamanders are in this family. The fetuses of these species feed on unfertilized ova or smaller embryos in utero and are born as larvae or metamorphs; the terrestrial juveniles and adults do not feed in water. Chioglossa is unusual in that it possesses a slender hyobranchial apparatus specialized for tongue protraction, suggesting that it does not suction feed. It is known, however, to spend time in water (Arntzen, 1981) and may capture aquatic prey using tongue or jaw prehension. 7.
Rhyacotritonidae
The Rhyacotritonidae contains four species of one genus, Rhyacotriton, all semiaquatic and metamorphic with reduced lungs. Adults possess a generalized hyobranchial morphology similar to the Ambystomatidae and modest tongue protraction, which they use on land and in water (Larsen, personal communication). They suction feed only as larvae. 8,
Dicamptodontidae
The Dicamptodontidae contains four species of one genus, Dicamptodon, three metamorphic and one perennibranchiate (D. copei). The larvae and perennibranchiate adults suction feed in a typical manner. No information is available on aquatic feeding in adults, which are essentially terrestrial but which return to the water to breed and guard their eggs (Nussbaum, 1969).
3. Aquatic Feeding in Salamanders 9.
Ambystomatidae
The Ambystomatidae, the sister taxon of the Dicamptodontidae, is a large family with one genus, Ambystoma, and 31 species. Most are metamorphic and terrestrial, but some species (e.g., the axolotl A. mexicanum) and populations of other species (e.g., A. tigrinum) are perennibranchiate. Larvae and perennibranchiate adults suction feed. Adults may use suction (although poorly) to capture prey when returning to the water to breed. Species formerly placed in the genus Rhyacosiredon are partially metamorphic; the adults inhabit mountain streams and retain larval features such as labial lobes, lidless eyes, and a small tongue pad. They possess a midmetamorphic hyobranchial configuration, with broader, more heavily ossified elements than other adult ambystomatids and less elaboration of anterior lingual elements (Fig. 3.6) (Reilly and Brandon, 1994). These features strongly suggest they are suction feeders, although behavioral information is lacking. 10,
Plethodontidae
The Plethodontidae is the largest family of salamanders, with over 275 species in 28 genera. Most of these species are members of the tribes Plethodontini and Bolitoglossini of the subfamily Plethodontinae and are completely terrestrial and direct developing. Two basal clades within the family, the subfamily Desmognathinae and the tribe Flemidactyliini (within the subfamily Plethodontinae), have members that retain the ancestral life history of metamorphosis from an aquatic suction-feeding larva. Direct development has also evolved two to three times within the Desmognathinae. Morphologically, plethodontid larvae are typical in most respects, but possess dentate palatopterygoids and lack maxillae and ossified pterygoids. Most hemidactylines possess only three epibranchials as larvae and many are perennibranchiate and suction feed as adults with their larval morphology. Two of these taxa, Haideotriton and Typhlomolge, are blind cave salamanders with one and two species, respectively. Metamorphic plethodontids that feed in the water use tongue or jaw prehension, even the fully aquatic desmognathine Desmognathus marmoratus (Schwenk and Wake, 1988; Deban and Marks, 1992; note that previous studies refer to D. marmoratus as Leurognathus marmoratus; however, we follow the recommendation of Titus and Larson, 1996, based on their recent molecular study of desmognathine relationships). Suction feeding has not been documented in metamorphosing adult plethodontids. They have probably lost the ability to suction feed as a result of the elaboration
91
of the tongue protraction system that followed on the loss of robust hyobranchial elements that can function as a mechanism for expanding the buccal cavity. Instead they use terrestrial feeding behaviors in water. The completely terrestrial bolitoglossines have a unique pattern of one embryonic epibranchial. All plethodontids are lungless and possess nasolabial grooves, a feature unique to the family. B. Phylogenetic Patterns of Feeding Form and Function Many features of the feeding morphology and behavior of salamanders have evolved convergently. Paedomorphosis, or the retardation of development relative to the ancestral condition, is a common pattern among salamanders and, when taken to the extreme, results in perennibranchiation. Perennibranchiation appears in 7 of the 10 salamander families and has evolved more than once in some of these (notably the Ambystomatidae and Plethodontidae). Extensive ossification of the hyobranchial apparatus has evolved independently in the Sirenidae, Proteidae, Salamandridae, Amphiumidae, and Cryptobranchidae and contributes to the strong suction-feeding abilities of these taxa. Free hypohyals in larvae and first radii of adults (which we consider homologues) have been lost independently in the Sirenidae, Proteidae, and Plethodontidae, but are present at some stage of development in the other families. Four larval epibranchials appear to be the ancestral condition for salamanders, and the presence of only three has evolved at least three times: in Pseudohranchus, the Proteidae, and in the Hemidactyliini (and in the embryos of the Plethodontini). A composite first epibranchial representing the fusion of the first ceratobranchial to the first epibranchial has evolved convergently in the Amphiumidae and some members of the Cryptobranchoidea and Salamandridae. The presence of a third larval ceratobranchial is unique to the Cryptobranchoidea, as is the functioning of the second set of branchial elements as the primary lever in hyobranchial depression. Free-living larvae, and thus suction feeding, have been lost numerous times, at least twice with the evolution of viviparity among salamandrids that give birth to metamorphosed young and at least three times with the evolution of direct development in plethodontids. Suction feeding has been lost at least three times in adult salanianders that feed in water: among semiaquatic and aquatic rhyacotritonids, plethodontids, and hynobiids, which have exapted the terrestrial ingestion behaviors of tongue or jaw prehension to capture aquatic prey.
92
S t e p h e n M. D e b a n a n d D a v i d B. W a k e
V. OPPORTUNITIES FOR FUTURE RESEARCH The fundamental structure, function, and natural history of feeding in salamanders are well known. Studies of foraging behavior have been conducted on a number of species, and diet is known for many more taxa. Cranial anatomy has been described for many species from most of the major groups, although more detailed morphological studies are needed for a number of taxa, including the Sirenidae (especially Pseudobranchus), Dicamptodontidae, and Rhyacotritonidae. The Rhyacotritonidae, which possesses a generalized and perhaps ancestral morphology, is particularly interesting because of its phylogenetic position. There are some groups for which scant morphological information is available, such as the Hynobiidae. The basic function of the suction-feeding system common to most aquatic salamanders is fairly well understood, but the functional significance, if any, of the great morphological diversity of salamanders remains largely uninvestigated. Feeding function of only a handful of the approximately 400 species of salamanders has been studied in any detail, and even fewer species have been examined at more than one level of biological organization. Studies of prey-capture behavior and kinematics among the aquatic taxa have focused on the families Ambystomatidae and Salamandridae, and little information is available for the other families (Terrestrial taxa are more fully studied; see Chapter 4). Physiological studies of in vivo muscle function (i.e., electromyography) have been undertaken in only seven species of aquatic salamanders representing half of the families. Clearly more diversity remains to be explored, including those taxa that show pronounced modulation or asymmetry of movement, such as Cryptobranchus and Amphiuma. The development of feeding behavior and function has been studied in only two species of salamanders, and early development of larval feeding remains completely unstudied. Little information is available on the feeding behavior and function in larvae of the unusual Cryptobranchoidea, particularly the Hynobiidae. The diversity of life histories among aquatically feeding salamanders has not yet been exploited. Comparisons between 1-year and multiyear larvae, and among larvae, paedomorphic adults, and secondarily aquatic metamorphic forms, may reveal relationships among structure, function, and life history. Our understanding of salamanders as a group is very limited in areas of neural control mechanisms, learning abilities, developmental plasticity in behavior and morphology, and variation and modulation. Most major salamander taxa remain to be examined in these
areas. Taxonomic and developmental diversity among salamanders remain to be exploited by researchers of biomechanics, physiology, development, and neuroethology. As long as representatives of most of the major groups of salamanders can be obtained, technological advances in high-speed imaging, electromyography, muscle physiology, and neurophysiology will continue to provide opportunities to explore the form and function of the feeding systems of salamanders at multiple levels. Acknowledgments We thank members of the Wakelunch group and Marvalee Wake for providing constructive comments on the manuscript. Karen Klitz rendered Figs. 3.4 and 3.6.
References Alberch, P., and E. A. Gale (1986) Pathways of cytodifferentiation during the metamorphosis of the epibranchial cartilage in the salamander Eurycea bislineata. Dev. Biol. 117:233-244. Alcobendas, M., H. Dopazo, and P. Alberch (1996) Geographic variation in allozymes of populations of Salamandra salamandra (Amphibia: Urodela) exhibiting distinct reproductive modes. J. Evol. Biol. 9:83-102. Anthony, C. D., D. R. Formanowicz, Jr., and E. D. Brodie, Jr. (1992) The effect of prey availability on the search behavior of two species of Chinese salamanders. Herpetologica 48:287-292. Amtzen, J. W. (1981) Ecological observations on Chioglossa lusitanica (Caudata, Salamandridae). Amph.-Rept. 1:187-203. Avery, R. A. (1968) Food and feeding relation of three species of Triturus (Amphibia: Urodela) during the aquatic phases. Oikos 19: 408-412. Bartels, M., H. Miinz, and B. Claas (1990) Representation of lateral line and electrosensory systems in the midbrain of the axolotl, Ambystoma mexicanum. J. Comp. Physiol. A. 167:347-356. Bell, G. (1975) The diet and dentition of smooth newt larvae {Triturus vulgaris). J. Zool. Lond. 176:411-424. Beneski, J. T., Jr., and J. H. Larsen, Jr. (1989a) Interspecific, ontogenetic, and life history variation in the tooth morphology of mole salamanders (Amphibia, Urodela, and Ambystomatidae). J. Morph. 199:53-70. Beneski, J. T, Jr., and J. H. Larsen, Jr. (1989b) Ontogenetic alterations in the gross tooth morphology of Dicamptodon and Rhyacotriton (Amphibia, Urodela, and Dicamptodontidae). J. Morph. 199:165-174. Besharse, J. C., and R. A. Brandon (1974) Postembryonic eye degeneration in the troglobitic salamander Typhlotriton spelaeus. J. Morph. 144:381-406. Bonebrake, J. E., and R. A. Brandon (1971) Ontogeny of cranial ossification in the small-mouthed salamander, Ambystoma texanum (Matthes). J. Morph. 133:189-204. Brophy, T. E. (1980) Food habits of sympatric larval Ambystoma tigrinum and Notophthalmus viridescens. J. Herp. 14:1-6. Burton, T. M. (1977) Population estimates, feeding habits and nutrient and energy relationships of Notophthalmus v. viridescens, in Mirror Lake, New Hampshire. Copeia 1977:139-143. Collins, J. P., and J. R. Holomuzki (1984) Intraspecific variation in diet within and between trophic morphs in larval tiger salamanders {Ambystoma tigrinum nebulosum). Can. J. Zool. 62:168-174.
3. A q u a t i c F e e d i n g in S a l a m a n d e r s Conant, R., and J. T. Collins (1991) A Field Guide to Reptiles and Amphibians: Eastern and Central North America, 3rd Ed. Houghton Mifflin, Boston. Cope, E. D. (1889) The Batrachia of North America. Bull. U. S. Nati. Mus. 34:1-525. Cox, D. C , and W. W. Tanner, W. W. (1989) Hyobranchial apparatus of the Cryptobranchoidea (Amphibia). Great Basin Nat. 49: 482-490. Cundall, D., J. Lorenz-Elwood, and J. D. Groves (1987) Asymmetric suction feeding in primitive salamanders. Experientia 43:12291231. Dawley, E. M., and A. H. Bass (1988) Organization of the vomeronasal organ in a plethodontid salamander. J. Morph. 198:243-255. Dawson, A. B. (1936) Changes in the lateral line organs in the life of the newt, Triturus viridescens: a consideration of the endocrine factors involved in the maintenance of differentiation. J. Exp. Zool. 74:221-237. Deban, S. M., and S. B. Marks (1992) Aquatic prey capture in plethodontid salamanders. Am. Zool. 32:140A. Dicke, U., and G. Roth (1994) Tectal activation of premotor and motor networks during feeding in salamanders. Eur. J. Morph. 32: 106-116. Duellman, W. E., and L. Trueb (1986) Biology of Amphibians. McGrawHill, New York. Durand, J. R, J. Parzefall, and B. Richard (1982) fitude comparee de la detection chimique des proies par Proteus anguinus, cavernicole, et son parent de surface Necturus maculosus (Froteidae, Urodela). Behav. Frocess. 7:123-134. Driiner, L. (1901) Studien zur Anatomie des Zungebein-, Kiemenbogen- und Kehlkopfmuskeln der Urodelen. I. Theil. Zool. Jahrb. Abteil.Anat. 15:435-622. Driiner, L. (1904) Studien zur Anatomie des Zungebein-, Kiemenbogen- und Kehlkopfmuskulatur der Urodelen. II. Theil. Zool. Jahrb. Abteil. Anat. 19:361-690. Edgeworth, F. H. (1935) The Cranial Muscles of Vertebrates. Cambridge Univ. Fress, Cambridge. Eisthen, H. L., D. R. Sengelaub, D. M. Schroeder, and J. R. Alberts (1994) Anatomy and forebrain projections of the olfactory and vomeronasal organs in axolotls {Ambystoma mexicanum). Brain Behav. Evol. 44:108-124. Elwood, J. R. L., and D. Cimdall (1994) Morphology and behavior of the feeding apparatus in Cryptobranchus alleganiensis (Amphibia: Caudata). J. Morph. 220:47-70. Erdman, S., and D. Cimdall (1984) The feeding apparatus of the salamander Amphiuma tridactylum: morphology and behavior. J. Morph. 181:175-204. Findeis, E. K., and W. E. Bemis (1990) Functional morphology of tongue projection in Taricha torosa (Urodela: Salamandridae). Zool. J. Linn. Soc. 99:129-157. Francis, E. B. T. (1934) The Anatomy of the Salamander. Clarendon Press, Oxford. Fritzsch, B., and U. Wahnschaffe (1983) The electroreceptive ampullary organs of urodeles. Cell Tissue Res. 229:483-503. Griffiths, R. A. (1993) The effect of p H on feeding behavior in newt larvae (Triturus: Amphibia). J. Zool., Lond. 231:285-290. Himstedt, W. (1967) Experimentelle Analyse der optischen Sinnesleistungen im Beutefangverhalten der einheimischen Urodelen. Zool. Jb. (allg. Zool.) 73:281-320. Himstedt, W. (1973a) Die spektrale Empfindlichkeit von Urodelen in Abhangigkeit von Metamorphose, Jahreszeit und Lebensraum. Zool. Jahrb. Fhysiol. 77:246-274. Himstedt, W. (1973b) Die spektrale Empfindlichkeit von Triturus alpestris (Amphibia: Urodela) w a h r e n d d e s Wasser- und Landlebens. Ffliigers Arch. 341:7-14. Himstedt, W., J. Kopp, and W. Schmidt (1982) Electroreception
93
guides feeding behavior in amphibians. Naturwissenschaften 69: 552-553. Hoff, K. S., M. J. Larmoo, and R. J. Wassersug (1985) Kinematics of midwater prey capture by Ambystoma (Caudata: Ambystomatidae) larvae. Copeia 247-251. Jaeger, R. G., and D. E. Barnard (1981) Foraging tactics of a terrestrial salamander: choice of diet in structurally simple environments. Am. Nat. 117:639-664. Joly, F. (1981) Le comportement predateur du Triton alpestre (Triturus alpestris). I. fitude descriptive. Biol. Behav. 6:339-355. Krogh, J. E., and W. W. Tanner (1972) The hyobranchium and throat myology of the adult Ambystomatidae of the United States and Northern Mexico. Brigham Young Univ. Sci. Bull. 16:1-69. Larsen, J. H., Jr. (1963) The Cranial Osteology of Neotenic and Transformed Salamanders and Its Bearing on Interfamilial Relationships. Fh.D. Dissertation, Univ. of Washington, Seattle. Larson, A., and W. W. Dimmick (1993) Fhylogenetic relationships of the salamander families: an analysis of congruence among morphological and molecular characters. Herp. Monogr. 7:77-93. Larsen, J. H., Jr., and D. J. Guthrie (1975) The feeding system of terrestrial tiger salamanders (Ambystoma tigrinum melanostictum Baird). J. Morph. 147:137-154. Larsen, J. H., Jr., J. T. Beneski, Jr., and B. T. Miller (1996) Structure and function of the hyolingual system in Hynobius and its bearing on the evolution of prey capture in terrestrial salamanders. J. Morph. 227:235-248. Lauder, G. V., and H. B. Shaffer (1985) Functional morphology of the feeding mechanism in aquatic ambystomatid salamanders. J. Morph. 185:297-326. Lauder, G. V., and H. B. Shaffer (1986) Functional design of the feeding mechanism in lower vertebrates: unidirectional and bidirectional flow systems in the tiger salamander. Zool. J. Linn. Soc. 88:277-290. Leff, L. G., and M. D. Bachmann (1986) Ontogenetic changes in predatory behavior of larval tiger salamanders (Ambystoma tigrinum). Can. J. Zool. 64:1337-1344. Lombard, R. E., and D. B. Wake (1977) Tongue evolution in the lungless salamanders, family Flethodontidae. II. Function and evolutionary diversity. J. Morph. 153:39-80. Luthardt-Laimer, G., and G. Roth (1983) Reduction of visual inhibition to stationary prey by early experience in Salamandra salamandra (L.). Z. Tierpsychol. 63:294-302. Marche, C , and J. F. Durand (1983) Recherches comparatives sur Fontogenese et revolution de Fappareil hyobrachial de Proteus anguinus L., proteidae aveugle des eaux souterraines. Amph.Rept. 4:1-16. Martin, J. B., N. B. Witherspoon, and M. H. Keenleyside (1974) Analysis of feeding behavior in the newt Notophthalmus viridescens. Can. J. Zool. 52:277-281. Martof, B. S., and D. C. Scott (1957) The food of the salamander Leurognathus. Ecology 38:494 -501. Matthes, E. (1934) Bau und Funktion der Lippensaume wasserlebender Urodelen. Zeit. Morph. Okol. Tiere 28:155-169. Miller, B. T., and J. H. Larsen, Jr. (1986) Feeding habits of metamorphosed Ambystoma tigrinum melanostictum in ponds of high p H (<9). Great Basin Nat. 46:299-301. Miller, B. T., and J. H. Larsen, Jr. (1989) Feeding performance in aquatic postmetamorphic newts (Urodela: Salamandridae): are bidirectional flow systems necessarily inefficient? Can. J. Zool. 67:2414-2421. MuUer, M., and J. W. M. Osse (1984) Hydrodynamics of suction feeding in fish. Trans. Zool. Soc. Lond. 37:51-135. Nickerson, M. A., and C. E. Mays (1973) The hellbenders: North American ''giant salamanders." Milwaukee Public Mus. Pub. Biol. Geol. 1:1-106.
94
S t e p h e n M. D e b a n a n d D a v i d B. W a k e
Noble, G. K. (1931) The Biology of the Amphibia. McGraw-Hill, New York. Northcutt, R. G., and K. Brandle (1995) Development of branchiomeric and lateral line nerves in the axolotl. J. Comp. Neurol. 355: 427-454. Nussbaum, R. A. (1969) Nest and eggs of the Pacific giant salamander, Dicamptodon ensatus (Eschscholtz). Herpetologica 25: 257-261. Ozeti, N., and D. B. Wake (1969) The morphology and evolution of the tongue and associated structures in salamanders and newts (family Salamandridae). Copeia 91-123. Parker, H. W., and E. R. Dunn (1964) Dentitional metamorphosis in the Amphibia. Copeia 1964:75-86. Parker, W. K. (1882) On the structure and development of the skull of urodeles. Trans. Zool. Soc. Lond. 11:171-214. Peck, S. B. (1973) Feeding efficiency in the cave salamander Haideotriton wallacei. Int. J. Speleol. 5:15-19. Peterson, C. L., J. W. Reed, and R. F. Wilkinson (1989) Seasonal food habits of Cryptobranchus alleganiensis (Caudata: Cryptobranchidae). Southwestern Nat. 34:438-441. Piatt, J. (1938) Morphogenesis of the cranial muscles of Amblysioma punctatum. J. Morph. 63:531-587. Piatt, J. (1940) Correct terminology in salamander myology. II. Transverse throat musculature. Copeia 1940:9-14. Reilly, S. M. (1986) Ontogeny of cranial ossification in the eastern newt, Notophthalmus viridescens (Caudata: Salamandridae), and its relationship to metamorphosis and neoteny. J. Morph. 188: 315-326. Reilly, S. M. (1995) The ontogeny of aquatic feeding behavior in Salamandra salamandra: stereotypy and isometry in feeding kinematics. J. Exp. Biol. 198:701-708. Reilly, S. M., and R. Altig (1996) Cranial ontogeny in Siren intermedia (Caudata: Sirenidae): paedomorphic, metamorphic, and novel patterns of heterochrony. Copeia 1996:29-41. Reilly, S. M., and S. M. Brandon (1994) Partial paedomorphosis in the Mexican stream Ambystomatids and the taxonomic status of the genus Rhyacosiredon Dunn. Copeia 1994:656-662. Reilly, S. M., and G. V. Lauder (1988a) Ontogeny of aquatic feeding performance in the eastern newt, Notophthalmus viridescens (Salamandridae). Copeia 87-91. Reilly, S. M., and G. V. Lauder (1988b) Atavisms and the homology of hyobranchial elements in lower vertebrates. J. Morph. 195: 237-245. Reilly, S. M., and G. V. Lauder (1989) Physiological bases of feeding behavior in salamanders: do motor patterns vary with prey type? J. Exp. Biol. 141:343-358. Reilly, S. M., and G. V. Lauder (1990) Metamorphosis of cranial design in tiger salamanders (Ambystoma tigrinum): a morphometric analysis of ontogenetic change. J. Morph. 204:121-137. Reilly, S. M., and G. V. Lauder (1992) Morphology, behavior, and evolution: comparative kinematics of aquatic feeding in salamanders. Brain Behav Evol. 40:182-196. Reno, H. W., and H. H. Middleton (1973) Lateral line system of Siren intermedia Le Conte (Amphibia: Sirenidae), during aquatic activity and aestivation. Acta Zool. 54:21-29. Rose, C.S. (1996) An endocrine-based model for developmental and morphogenetic diversification in metamorphic and paedomorphic urodeles. J. Zool. Lond. 239:253-284.
Roth, G. (1976) Experimental analysis of the prey catching behavior of Hydromantes italicus Dunn (Amphibia, Plethodontidae). J. Comp. Physiol. 109:47-58. Roth, G. (1987) Visual Behavior in Salamanders. Springer-Verlag, Heidelberg. Roth, G., K. C. Nishikawa, D. B. Wake, U. Dicke, and T. Matsushima (1990) Mechanics and neuromorphology of feeding in amphibians. Netherlands J. Zool. 40:115-135. Schumacher, G. H. (1958) Zur Morphologie des Mundboden der urodelen: untersuchungen an Cryptobranchus japonicus. Gegenbaur Morph. Jahrbuch 99:344-371. Schwenk, K., and D. B. Wake (1988) Medium-independent feeding in a plethodontid salamander: tongue projection and prey capture underwater. Am. Zool. 28:115A. Schwenk, K., and D. B. Wake (1993) Prey processing in Leurognathus marmoratus and the evolution of form and function in desmognathine salamanders (Plethodontidae). Biol. J. Linn. Soc. 49:141162. Severtsov, A. S. (1964) Formation of the tongue in the Hynobiidae. DokL Biol. Sci. 154:34-37. Severtsov, A. S. (1966) Food-seizing mechanism in urodele larvae. Dokl. Biol. Sci. 168:230-233. Shaffer, H. B., and G. V. Lauder (1985a) Aquatic prey capture in ambystomatid salamanders: patterns of variation in muscle activity. J. Morph. 183:273-284. Shaffer, H. B., and G. V. Lauder (1985b) Patterns of variation in aquatic ambystomatid salamanders: kinematics of the feeding mechanism. Evolution 39:83-92. Thiesmeier, B., and C. Hornberg (1997) Paarung, Fortpflanzung und Larvalentwicklung von Pachytriton sp. {Pachytriton A) nebst Bemerkungen zur Taxonomie der Gattung. Salamandra 33: 97-110. Titus, T. A., and A. Larson (1996) Molecular phylogenetics of desmognathine salamanders (Caudata: Plethodontidae): a reevaluation of evolution in ecology, life history, and morphology. Syst. BioL 45:451-472. Trueb, L. (1993) Patterns of cranial diversity among the Lissamphibia. Pp. 255-343. In: The Skull Vol. 2. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. TumHson, R., G. R. Cline, and P Zwank (1990) Prey selection in the Oklahoma salamander {Eurycea tynerensis). J. Herp. 24:222-225. Uiblein, F., J. P Durand, C. Juberthie, and J. Parzefall (1992) Predation in caves: the effects of prey immobility and darkness on the foraging behaviour of two salamanders, Euproctus asper and Proteus anguinus. Behav. Process. 28:33-40. Uiblein, F., and J. Parzefall (1993) Does the cave salamander Proteus anguinus detect mobile prey by mechanic cues? Mem. Biospeol. 20:261-264. Valentine, B. D., and D. M. Dennis (1964) A comparison of the gillarch system and fins of three genera of larval salamanders, Rhyacotriton, Gyrinophilus, and Amby stoma. Copeia 1964:196-201. White, R. L., II (1977) Prey selection by the rough skinned newt {Taricha granulosa) in two pond types. Northwest Sci. 51:114-118. Wilder, I. W. (1925) The Morphology of Amphibian Metamorphosis. Smith College Fiftieth Anniversary Publications. Smith College, Northampton, MA. Worthington, R. D., and D. B. Wake (1971) Larval morphology and ontogeny of the ambystomatid salamander, Rhyacotriton olympicus. Am. Midi. Nat. 85:349-364.
C H A P T E R
4 Terrestrial Feeding in Salamanders DAVID B. WAKE AND STEPHEN M. DEB AN Museum of Vertebrate Zoology and Department of Integrative Biology University of California Berkeley, California 94720
A. Systematics
I. INTRODUCTION A. Systematics B. Natural History C. Feeding Modes and Terminology 11. MORPHOLOGY A. Skeleton B. Musculature C. Sensory and Motor Systems III. FUNCTION A. Foraging, Prey Detection, and Localization B. Prey Capture and Ingestion C. Biomechanics and Functional Morphology D. Prey Processing E. Modulation of Feeding Behavior IV. DIVERSITY AND EVOLUTION A. Origins and Outgroups B. Phylogenetic Diversity C. Feeding Biology and Evolution V. OPPORTUNITIES FOR FUTURE RESEARCH References
There are roughly 450 species of salamanders currently placed in 10 families (Fig. 4.1). Evidence exists for the monophyly of the Caudata and for the sistergroup relationship of Caudata to Anura (Clouthier and Trueb, 1992; Larson and Wilson, 1991; Larson, 1991). How^ever, relationships within the Caudata are unstable, and different results are obtained from DNA sequence analysis and from morphological characters (Larson, 1991). We use one of tw^o similar phylogenetic hypotheses resulting from a combined data analysis (Larson and Dimmick, 1993). In this tree (Fig. 4.1), Sirenidae (aquatic, gilled, elongate, lacking hind limbs, and presumed to have external fertilization) is basal, followed by a branch leading to the sister-taxa Hynobiidae and Cryptobranchidae (which have external fertilization). The more derived salamanders form a monophyletic group (all have internal fertilization), but the base is unstable. Whereas earlier workers placed plethodontids as one of the most derived taxa, Larson's (1991) DNA data place it as one of the most basal lineages, and only the combined data pull it up in the tree to the base of the derived taxa. The diverse and much studied ambystomatids and salamandrids are deeply nested and envisioned as highly derived. Most terrestrial salamanders belong to the Plethodontidae, which includes about two-thirds of all species of salamanders. About three-quarters of the plethodontid species have no aquatic larval stage and develop directly from eggs laid on land (Wake and Flanken, 1996). Relationships within the Plethodontidae are discussed in Lombard and Wake (1986). Flynobiidae and Salamandridae are both relatively speciose and display
L INTRODUCTION Chapter 3 presented a general introduction to the feeding biology of salamanders, w^ith a focus on feeding in aquatic habitats. This chapter deals with feeding by metamorphosed individuals, or the free-living stage of direct-developing species. Much basic information concerning morphology and diversity applies to both aquatic and terrestrial salamanders, and we have repeated information in this chapter only v\^here necessary. The previous chapter provides essential background information for the present one.
FEEDING (K. Schwenk, ed.)
95
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
96
David B. Wake and Stephen M. Deban Sirenidae
_| Sirenoidea
Cryptobranchidae
Cryptobranchoidea
Hynobiidae Amphiumidae Plethodontidae Rhyacotritonidae Salamandridae
| Salamandroidea
Ambystomatidae Dicamptodontidae Proteidae F I G U R E 4.1. Hypothesis of phylogenetic relationships of the salamander families, adapted from one of two most-parsimonious trees from Larson and Dimmick (1993) based on an analysis of combined molecular and morphological data. See text for discussion.
diverse morphology related to feeding. Relationships of hynobiid genera are discussed in Zhao and Hu (1988) and those of salamandrids by Titus and Larson (1995). Only one genus of Ambystomatidae is recognized, but it is relatively speciose and the species display morphological variation with respect to feeding. Species relationships within Ambystoma were studied by Shaffer et ah (1991). Only a single genus is recognized for the Dicamptodontidae and the Rhyacotritonidae, and each contains four species. Relationships of the species of Dicamptodontidae are discussed by Good (1988) and of the Rhyacotritonidae by Good and Wake (1992). B. Natural History The vast majority of salamanders either have a terrestrial phase during which they feed on terrestrial prey or are terrestrial throughout life. The Cryptobranchidae, Sirenidae, Proteidae, and Amphiumidae, which are strictly to essentially aquatic, are emphasized in Chapter 3 and will not be discussed here. Together these four families include only about 16 species. Most members of the families Ambystomatidae, Dicamptodontidae, Hynobiidae, and Salamandridae metamorphose from aquatic larvae to terrestrial adults, but some remain aquatic as adults, either in a perennibranchiate (e.g., Ambystoma mexicanum, Dicamptodon copei) or in a fully metamorphosed state. The four species of the Rhyacotritonidae metamorphose, but they remain closely associated with water throughout life. The situ-
ation is complicated in the Plethodontidae, which is the largest family. The Plethodontidae includes the most terrestrial of salamanders, but adults of some taxa are fully aquatic, either perennibranchiate (e.g., Haideotriton, Typhlomolge, and some species of Eurycea and Gyrinophilus) or metamorphosed (e.g., Desmognathus marmoratus). All members of the plethodontid tribes Plethodontini and Bolitoglossini (together accounting for well over 50% of the living species of salamanders) undergo direct development (in encapsulated eggs) and are strictly terrestrial throughout life. A few salamandrids have a stage with larval-like morphology that develops within the oviduct and these, too, are fully terrestrial. Terrestrial salamanders are strict carnivores, and the main prey are arthropods, especially insects (Lynch 1985; Keen 1979; Sites 1978; Kuzmin, 1991). Annelids are eaten by hynobiids (Kuzmin, 1991), ambystomatids (Bishop, 1941), and salamandrids (Roth, 1987). Some large salamanders eat other salamanders (Bishop, 1941; Hairston, 1987) and even such vertebrates as frogs, shrews, mice, lizards, and snakes (Dicamptodon, Bury, 1972; Gyrinophilus, Bishop, 1941). Snails and isopods are widely eaten. C. Feeding M o d e s and Terminology This chapter deals not only with all terrestrial feeding, but also salamanders that use essentially terrestrial mechanisms when feeding under water, such as the desmognathine plethodontids. Prey capture typically is by the tongue, which is extended from the mouth and returns to the buccal cavity with the prey attached, a behavior known as tongue protraction or tongue protrusion. We reserve the term tongue projection for prey-capture mechanisms in which the tongue is fired from the mouth ballistically and reaches the prey under its own momentum, a mechanism known only in bolitoglossine plethodontids. Many terrestrial salamanders have weak jaws and teeth, particularly those that rely primarily on the tongue to bring prey into the buccal cavity. There are, however, species that have robust jaws: Dicamptodon, Ambystoma tigrinum, and some species of Batrachuperus have very large heads and strong jaws as adults, the desmognathine plethodontids have well-ossified skulls, powerful jaws, and unique jaw mechanics (Schwenk and Wake, 1993), and the plethodontid genus Aneides has strong jaws and very large, saber-like teeth (Wake, 1963; Staub, 1993). These species and others that typically use tongue protraction to capture prey will resort to capturing prey with the jaws, in a behavior known as jaw prehension, when the tongue fails to return prey to the mouth.
4. Terrestrial Feeding in Salamanders II. MORPHOLOGY Salamanders have a generalized morphology with a relatively narrow head, well-developed eyes, and a short and broad neck. The head is flattened and is broader than high. The external nares are well separated at the front of a short face. The eyes are at the dorsolateral margin of the face, and head muscles lie behind them. There are varying degrees of eye frontality, but visual fields of the two eyes typically overlap to some degree (Roth, 1987). The mouth opening is large, characteristically extended to a point ventral to the eye. At the margins of the mouth, generally small, bicuspid, pedicellate teeth are found in opposing semicircular rows, one on the premaxillary and maxillary bones and the other on the dentaries. There are also palatal teeth on the vomer in various arrangments and highly varied in number among species; these teeth typically do not parallel the marginal teeth. The tongue is typically large and fleshy, varying among species in the degree to which the margins are attached to surrounding tissues (from completely attached to entirely free). The surface of the tongue is covered with sense organs and the openings of numerous mucous glands, and it is soft and spongy in texture. Internally the tongue has an extensive musculature, including both extrinsic muscles that originate outside the tongue pad and intrinsic muscles that are associated with tongue retraction as well as molding of the tongue pad. General surface morphology of the tongue in relation to the mouth and teeth is illustrated for a variety of species by Bishop (1941) and Lombard and Wake (1977).
97
tals and parietals. Prootic and opisthotic bones form the braincase posterolaterally and enclose the inner ear, and the opisthotics are fused with the exoccipitals and bear the occipital condyles. In most families and species there is a single ossification, termed occipitootic. The jaw suspensorium (largely cartilaginous) is attached to the ventrolateral sides of the occipito-otic and is covered dorsally by the squamosal. The distal part of the suspensorium is the ossified quadrate. The lower jaw includes two or three bones: dentary, prearticular (to which jaw muscles attach), and angular (Hynobiidae and Cryptobranchidae) or coronoid (which does not bear teeth in the only metamorphosed salamanders to have this bone, the Dicamptodontidae). Meckel's cartilage is persistent to some degree in all species, and the proximal portion is expanded to form the articular, which articulates with the quadrate. Teeth are generally small, pedicellate, and bicuspid, with a small medial and a larger labial cusp. The marginal teeth are borne on the premaxillary and maxillary bones of the skull and the dentary of the lower jaw. Vomerine teeth are always present, and in salamandrids and plethodontids they extend posteriorly under the parasphenoid bone. In most plethodontids the posterior vomerine teeth are disconnected from the anterior ones and form a pair of large patches of small teeth, often exceeding 100 in number. Some members of the plethodontid genus Aneides have large, unicuspid teeth that apparently are used more in aggressive interactions than for feeding (Staub, 1993). A few small tropical species of the bolitoglossine plethodontids lack maxillary dentition (Wake 1966). 2. HyoUngual Skeleton
A. Skeleton 1, Skull and Jaws Skulls of salamanders are weakly ossified and contain relatively few, thin bones. Premaxillary bones (single or paired) have rising, posteriorly oriented frontal processes that articulate with surrounding bones and provide support (Carroll and Holmes, 1980; see also Chapter 3, Fig. 3.3). Laterally the bones articulate with the much larger maxillary bones. The maxillaries have a frontal process that rises to articulate with surrounding bones (nasal, lacrimal, prefrontal, and frontal, depending on species). The premaxillaries and maxillaries have small, medially projecting, shelf-like processes that articulate with the expanded vomers to form the roof of the buccal cavity. The eyes lie lateral to the brain case, which is formed by a ventral parasphenoid, lateral orbitosphenoids, and dorsal, paired fron-
A hyolingual apparatus (also referred to as a hyobranchial apparatus, particularly in aquatic species) is associated with the tongue in all terrestrial salamanders and it varies substantially among taxa. The elements of the apparatus are cartilaginous to some degree, and in most taxa there is little or no bone, but the apparatus may ossify extensively in others. The general morphology of the hyolingual apparatus is discussed in detail in Chapter 3 (see Figs. 3.4 and 3.6, as well as Figs. 4.2 through 4.4). In metamorphosed salamanders, an elongate basibranchial and a urohyal lie along the ventral midline. The urohyal is continuous with the basibranchial in embryos but the connection is lost before metamorphosis. Unlike the basibranchial, which is usually cartilaginous, the urohyal ossifies early. Paired arches extend posterolaterally and somewhat dorsally from the basibranchial: a hyoid and two branchial arches. The homologies of the arch
98
David B. Wake and Stephen M. Deban
Onychodactylus japonicus
Hynobius leechii
Rhyacotriton olympicus
Dicamptodon ensatus
Amhystoma maculatum
Ambystoma texanum
Taricha granulosa
Salamandra salamandra
Chioglossa lusitanica
F I G U R E 4.2. Hyobranchial apparatus of terrestrial salamanders. Ventral views, showing relation of apparatus to lower jaw. Note the radial loops and two pairs of epibranchials in the two hynobiids shown, Onychodactylus japonicus and Hynobius leechii, the otoglossal cartilage in Rhyacotriton olympicus and Ambystoma maculatum, and the lingual plate in Dicamptodon ensatus and forked second radii in Ambystoma texanum. Compare the ossification of Taricha granulosa and the elongate basibranchial, radii, and epibranchials of Chioglossa lusitanica to the more generalized condition in Salamandra salamandra. Bone is stippled and cartilage is outlined. Scale is 5 mm. Abbreviations are explained in Chapter 3, Table 3.1. In some instances, an element has been interrupted or cut on one side to reveal the element dorsal to it. In other instances, the hidden element is indicated by a dotted line.
99
4. Terrestrial Feeding in Salamanders
elements are debated and terminology varies among authors. All authors use the same terms for the first arch elements: a small hypohyal (proximal; absent in most taxa, possibly fused to the remaining element) and a much larger ceratohyal (distal). The proximal elements of the next two arches are called hypobranchials by some authors and ceratobranchials by others. A serial homologue of the hypohyal would be expected to be short, but this branchial element is characteristically long; the element may well represent both hypobranchial and ceratobranchial or the hypobranchial may be absent (Fig. 4.2). The more distal elements are termed either ceratobranchials or epibranchials. We use the term ceratobranchial for the more proximal of the paired branchial elements and epibranchial for the more distal element. There are only two branchial arches in terrestrial salamanders. Although the two proximal elements are universally present, there may be one (most taxa) or two (hynobiids) epibranchials (Fig. 4.2). When one epibranchial is present, the two ceratobranchials on each side connect to the single epibranchial, either directly or indirectly (in some taxa the second ceratobranchial connects at its distal end to the first ceratobranchial instead of the epibranchial). The articulated hyolingual skeleton of terrestrial salamanders includes the basibranchial, various skeletal pieces connected near its anterior tip (radii, otoglossal, lingual cartilage), and elements of the first two branchial arches. Ceratohyals lack articulations, other than with the hypohyal (when present); they are attached proximally to the suspensorium by the hyoquadrate ligament, and there are quasi-ligamentous connections of the flattened, distal blade to the lower jaw. When present the urohyal is an ossified triangular or crescentic element well separated posteriorly from the other elements and embedded in musculature. Hynobiids have two epibranchials, but all other terrestrial salamanders have one (compare Onychodactylus and Hynobius in Fig. 4.2 to the others). The ceratohyals in hynobiids are continuous with one another across the midline by a slender, looped cartilage (radial cartilage; radial loop of Larsen et ah, 1996) of uncertain homology (but sometimes termed hypohyal, and considered by other authors to represent the first radials) arranged in the form of a flat spring with the loop extending posteriorly and crossing at the midline. At its most posterior limit it may be attached to the basibranchial. The basibranchial in hynobiids is of moderate length in some species but very short in others, whereas it is relatively long in the other taxa. There are two pair of radii in most salamandrids and in the ambystomatids, dicamptodontids, and rhyacotritonids. The last three families also have an unpaired, medial otoglossal associated with the tongue pad. There
Desmognathus monticola
Ensatina eschscholtzii
Gyrinophilus porphyriticus
Plethodon neomexicanus
FIGURE 4.3. Hyobranchial apparatus of semiaquatic and fully terrestrial plethodontid salamanders. Ventral views, showing relation of apparatus to lower jaw. Note the heavy jaw in Desmognathus monticola and the extended jaw and large urohyal in Gyrinophilus porphyriticus. The epibranchials are lengthened in Ensatina eschscholtzii compared to the more generalized Plethodon neomexicanus. Bone is stippled and cartilage is outlined. Scale is 5 mm. Abbreviations are explained in Chapter 3, Table 3.1. In some instances, an element has been interrupted or cut on one side to reveal the element dorsal to it. In other instances, the hidden element is indicated by a dotted line.
is a small interradial cartilage in most salamandrids that may be a homologue of the otoglossal. The first pair of radii has no muscles attached to it; we consider it to be the ontogenetic homologue of the hypohyals, as have other authors (Driiner, 1901). When only one pair
100
David B. Wake and Stephen M. Deban
Batrachosevs attenuatus
Thorius narisovalis
F I G U R E 4.4. Hyobranchial apparatus of bolitoglossine plethodontid salamanders. Ventral views, showing relation of apparatus to lower jaw. Note that apparati are entirely cartilaginous. The epibranchials of Hydromantes platycephalus are the longest of any salamander. The small, tapered first ceratobranchial in Thorius narisovalis disarticulates from the basibranchial during tongue protraction. Bone is stippled and cartilage is outlined. Scale is 5 mm. Abbreviations are explained in Chapter 3, Table 3.1. In some instances, an element has been interrupted or cut on one side to reveal the element dorsal to it. In other instances, the hidden element is indicated by a dotted line.
of radii is present, it is the second. Some plethodontids lack radii and in others they are elongate. The most elongate radii occur in the salamandrid genera Salamandrina and Chioglossa (Fig. 4.2). These genera, as well as Salamandra and Mertensiella, have an epibranchial
that results from the fusion of the epibranchial and first ceratobranchial of other genera. The two epibranchials of hynobiids are attached to the ends of the first and second ceratobranchials. In general, differences among taxa with respect to the hyolingual apparatus are
101
4. Terrestrial Feeding in Salamanders mainly due to differences in proportions, which in turn reflect functional diversity, for example, differences in the length of the epibranchials among plethodontids (Figs. 4.3 and 4.4). B. Musculature The musculature associated with feeding is derived from the so-called branchiomeric series (Wake, 1993) and the somitic hypobranchial series. Detailed descriptions of muscles common to all salamanders are presented in Chapter 3 (see Chapter 3, Fig. 3.5). Among the branchiomeric derivatives are the jaw levators and depressors and the tongue protractors; the tongue retractors are somitic in origin. The neck is poorly differentiated, but epaxial somitic muscles are involved in raising the skull, and the levator scapulae (most families) and/or cucuUaris (plethodontids) is associated with postural movements of the head. Details of tongue musculature relevant to feeding movements are discussed later in the context of tongue function (Section III,C,l,b). C. Sensory and Motor Systems 1. Vision, Olfaction, and the Brain Terrestrial salamanders locate prey by olfaction and vision, but vision predominates. The olfactory system is well developed but generalized. There is typically a rather elongate snout that houses the paired and cartilaginous olfactory capsules. Within each capsule is a large medial olfactory sac, lined with olfactory epithelium, and a small ventrolateral sac, lined with a specialized sensory epithelium (Dawley and Bass, 1989). The smaller sac is the vomeronasal organ. Plethodontids have a unique organ, the nasolabial groove, associated with each external nostril. The groove lies in the skin of the snout and extends to the margin of the jaw, where typically it opens broadly. In males, tissues associated with the groove swell and elongate during the courtship season (Eurycea) or throughout life (tropical plethodontids). The groove draws surface water into the nasal/vomeronasal sacs and is thought to be important in courtship behavior (Dawley and Bass, 1989), but it may also play a role in feeding. The lateral eyes are large relative to larvae and to permanently aquatic species, and vision is good. The eyes are laterally oriented in species with more generalized terrestrial feeding, but in those species with highly protrusible tongues the eyes become frontally oriented and the animals are able to determine distance of the prey from the head with great accuracy. Visual behavior has been studied extensively, with spe-
cial emphasis on its role in feeding (reviewed in Roth, 1987). The optic nerve projects to the well-developed midbrain, which has an enlarged optic tectum, the main center for the coordination of feeding in salamanders. Projections are both contralateral and ipsilateral, and ipsilateral projections are especially well developed in bolitoglossine plethodontids. The existence of this dual projection system contributes to visual acuity and the ability to accurately estimate distance. 2. Tectobulbar Pathways, Brain Stem, and Innervation Patterns Descending pathways from the optic tectum and tegmentum to the bulbar portion of the brain stem are generalized in salamanders and less clearly differentiated from one another spatially than those of frogs. In sharp contrast with the situation in frogs, there is involvement by more cranial nerves in feeding, and motor nuclei are less condensed and overlap more extensively. As yet there has been little demonstration of interneurons in the reticular formation of the brain stem, and there is some evidence of direct innervation of motor neurons. There are few neurons in the motor nuclei, and their cell bodies are characteristically large (Roth, 1987). The 4th, 5th, 7th, 9th, 10th, and 11th cranial nerves are all involved in feeding, as are the 1st and 2nd spinal nerves. The 4th nerve (abduscens) has well-differentiated lateral and medial subnuclei. The lateral serves extrinsic eye muscles, whereas the medial serves the retractor bulbi, muscles that function in part in aiding in swallowing. The 5th nerve (trigeminal) innervates the jaw levators and anterior constrictors of the buccal cavity, and the 7th nerve (facial) innervates the jaw depressors as well as throat constrictors. The 9th (glossopharyngeal) and 10th (vagal) nerves serve the tongue protractor muscles. The eleventh (accessory) nerve innervates musculature associated with movements of the head on the neck. The 1st and 2nd spinal nerves (which have trunks that unite to form the ramus hypoglossus) innervate the retractor muscles of the tongue, the muscles of the tongue pad, and the strap-like muscles paralleling the body axis that contribute to throat movements and throat stability.
IIL FUNCTION A. Foraging, Prey Detection, and Localization Foraging has been studied in plethodontids by Jaeger and associates (Jaeger and Barnard, 1981; Jaeger
102
David B. Wake and Stephen M. Deban
and Gergits, 1979; Jaeger and Rubin, 1982; Jaeger et al, 1982), but only incidentally for other groups (e.g.. Flower, 1927). Terrestrial salamanders forage actively at night under appropriate conditions of temperature and humidity, under suboptimal conditions the animals adopt a sit and wait strategy, remaining in burrows in the soil or other retreats with only the head exposed. Capture of prey by salamanders involves a sequence of activities (Roth, 1987). Orientation is elicited by visual or mechanical stimuli and involves a turning movement of the head, which permits binocular fixation of the prey. The salamander next approaches the prey by walking to a point within reach of the tongue, which varies substantially in length among taxa. If the prey are not evasive, an olfactory test then may take place in which the salamander places its snout directly on the prey. Once within attack distance, binocular fixation of the prey may occur again. Finally, snapping occurs, and in this phase the mouth opens and the tongue seizes the prey, which is drawn into the mouth. Typically, prey capture is by lingual prehension, although forward lunging of the whole body may accompany tongue protraction. Visual aspects of feeding have been studied in detail by Roth and associates (reviewed by Roth, 1987). Species differ greatly in the size of stimuli they prefer, which is correlated with natural food preferences. Species such as the plethodontid Hydromantes, which feed on insects, prefer smaller stimuli than Salamandra, which feeds mainly on worms. Both maximum and minimum size of response stimuli is much smaller in Hydromantes than in Salamandra. Salamandra feeds on prey moving at velocities from 0.5 to about 2 cm/sec, but bolitoglossine plethodontids have been recorded to feed on prey moving from 6 to 10 cm/sec. Motionless prey normally are ignored by terrestrial salamanders. Salamanders typically forage at night and are able to use visual cues at extraordinarily low levels of light (Roth, 1976, Himstedt, 1982). Laboratory experiments demonstrate that they are able to feed visually at illumination levels equivalent to open areas on clouded or rainy nights. Salamanders also feed from olfactory cues in the absence of visual stimuli, as in complete darkness. When light is present, odoriferous dead prey will be ignored for a long time, but if vision is ineffective the food is eaten quickly. Salamanders that live so deep in caves (e.g., Typhlotriton) that there is no light probably use a combination of smell and touch to obtain food. Salamanders have both binocular and monocular depth perception. They do not use lens accommodation, but rather rely on disparities between direct ipsilateral and contralateral retinotectal projections. They
apparently are able to make immediate calculations of depth and feed with great accuracy (Wiggers and Roth, 1991). B. Prey Capture and Ingestion Feeding involves mouth opening/closing, and tongue protraction/retraction. When prey are small relative to the salamander, jaws are not used in feeding, and marginal dentition does not contact prey. However, larger species feeding on larger prey (e.g., members of Dicamptodon and Ambystoma) do use their jaws. The role of vision in feeding is uncertain once the prey is localized. In general, the feeding act is so rapid that once feeding starts relatively little modulation of the strike itself is possible because of delays in response to sensory information (Thexton et al., 1977). Some species close their eyes during the strike, whereas others do not, and this variation occurs even among confamilial species (Roth, 1976; Larsen and Beneski, 1988; Larsen et al, 1989; Miller and Larsen, 1990). In general, those species with the fastest tongues and the greatest degree of protraction are the most likely to keep the eyes open during the feeding cycle. Correspondingly, those that lunge forward strongly during feeding often close their eyes, presumably to protect them. Because prey are out of sight during tongue protraction and lunging, these movements are thought to be preplanned and executed without modification once begun. C. Biomechanics and Functional Morpliology 1. Lingual Feeding a. Kinematics Lingual feeding has been observed in all families of terrestrial salamanders. The published literature is dominated by information concerning ambystomatids, hynobiids, plethodontids and salamandrids, and little information is available for the small families Rhyacotritonidae and Dicamptodontidae. The prey object is first contacted by a protracted tongue pad that is delivered rapidly through a gaping mouth into the buccal cavity. The mode of protraction has been characterized as a lift and thrust mechanism (Beneski et ah, 1995). Both a tongue cycle and a necessarily associated gape or mouth opening and closing cycle can be recognized. These combine to form the basis of a model of a generalized vertebrate feeding cycle, based on the premise that lingual prehension was ancestral in tetrapod feeding (Bramble and Wake, 1985). The four stages, first formally recognized in mammalian feeding (Hiiemae,
4. Terrestrial Feeding in Salamanders 1978), include slow open (SO), fast open (FO), fast close (FC), and slow close-power stroke (SC-PS). Subsequent research on salamanders has recognized patterns generally similar to these four stages. Subtle distinctions have been drawn, however, with some authors continuing to use these designations, some choosing strictly temporal terminology (e.g.. Phases I IV; Larsen and Beneski, 1988), and others choosing new descriptors (e.g., preparatory, fast open, close, recovery; Reilly and Lauder, 1990a). In part, the distinctions are based on the fact that different taxa were studied by different workers. Gape cycles for ambystomatid salamanders using lingual prehension consist of only three parts (lacking a second phase of mouth opening), whereas those of hynobiids, plethodontids and salamandrids generally have the standard fourpart cycle. Plethodontids may show variation in feeding kinematics which complicates the division of the gape cycle into distinct stages (e.g., see Fig. 4.5). The three-part cycle of ambystomatids involves rapid mouth opening, a period of stable positioning of the head during tongue protraction and retraction, and rapid mouth closing (Reilly and Lauder, 1989). The four-part cycle of other taxa includes an initial brief gape phase that opens the mouth only slightly, a second phase during which the mouth is held open just wide enough for tongue protraction, a third phase when gape is increased rapidly during tongue retraction, followed by mouth closure (Larsen and Beneski, 1988). Ambystomatids have modest abilities for tongue protraction, and the gape cycle has been simplified as a result. Beneski et ah (1995) and Larsen et al (1996) characterize the ambystomatid feeding pattern as "lift and roll" in reference to the modest protraction of the tongue in the family and the extensive involvement of tongue pad ventral rolling over the rostral end of the protracted basibranchial and the extensive movement and "fitting" of the tongue pad onto the prey. Members of the other families have varying degrees of tongue protraction, but prey captured by tongue prehension often do not contact the teeth of the jaws. Some species, including even such highly proficient tongue-protracting forms as the plethodontid Ensatina, may use teeth and jaws with relatively large prey. The power stroke of the fourth phase of the generalized vertebrate feeding cycle is modified or absent in salamanders. The extreme degree of reliance on tongue prehension in terrestrial salamanders may account for the departures noted from the general model. Some studies have been conducted on salamandrids feeding without tongue prehension, using only jaws, and these differ from the generalized model in having a two-phase gape cycle: rapid mouth opening followed immediately by rapid mouth closing (Miller and Larsen, 1990). However,
103
these exceptional cases are typically aquatic species feeding on land. The typical strike in salamanders is accompanied by a forward lunge of the entire body toward the prey. This is the ancestral mode and it is widely retained. Lunging effectively increases the strike distance and may also increase the force of the strike. This may be of significance in forcing the tongue pad to conform to the prey surface, thus increasing the area of contact (Larsen et al, 1996). Furthermore, the lunge may continue as tongue retraction occurs so that teeth are quickly brought into contact with the prey item (Larsen et ah, 1996). The lunge is absent in certain taxa that display high-speed tongue protraction, including bolitoglossine plethodontids (Larsen et al., 1989; Roth and Wake, 1985b) and the salamandrid Salamandrina (Miller et al, 1990), and is rarely present in some hemidactyliine plethodontids with long tongues (e.g., Eurycea). b. Protraction Lombard and Wake (1976) studied biomechanics of tongue protraction in plethodontids and presented a theoretical model of folding of the hyobranchial apparatus. The model was based on assumptions dealing with the nature of materials (that the central main elements, the basibranchial and the ceratobranchials, do not bend during protraction or retraction), the stability of joints (there is no disarticulation during protraction and retraction), and other considerations (e.g., the system folds dor sally rather than ventrally). The model was used to formulate hypotheses concerning biomechanical events and was generated with specific reference to plethodontids. Members of this family are sufficiently different from other salamanders in hyobranchial structure that it is unclear if the biomechanical model has generality, but folding occurs to some degree in other taxa as well [e.g., the salamandrid Taricha; Findeis and Bemis (1990) and the hynobiid, Hynobius; Larsen et al (1996)]. Hypotheses generated from the model were tested in various ways. The fundamental part of the model is that the skeletal parts of the tongue are folded into a kind of bundle, formed together with protractor and retractor muscles, other connective tissue, and nerves. This compact unit is protracted from the mouth, to some degree in a ballistic manner, carrying the attached tongue pad at its tip. The cylindrical epithelial sheath surrounding the bundle is heavily pleated when at rest, and as the bundle is protracted it unpleats. The folding of the apparatus is initiated with the contraction of the protractor muscles, the subarcualis rectus I, which originate on the ventral side of the anterior blade-like portion of the ceratohyals and extend to
3.0 2.5 J——Tongue Reach £
2.0
O
1.5
c
(0
a
1.0 Gape3 Distance 0.5 -0.<
0.0
28
56
83
111
139
i
V,
^OODO^iamojffiODOt HDOOCxj} 1 ^ 167 194 222 250 278 306
0.4
0.3
0
E o c
• •-.^ ^%4f^%^^Gape Distance
• 0.2
i
(0
0.1 ,-•' 0
^•1
1
N1
•
)•
28
•a
56
83
111
139
167
194
222
250
278
306
278
306
1.1
£ c o o
1.0
0.8
CL
0.7
r
0.6
>
0.5
0
^ ^ ] — Upper Jaw Tip
0.9
Lower Jaw Tip —M^^ yj^-'-^^^O
"OooCr^
J-
0.4 0
28
56
83
111
139
167
194
222
250
30 20 ^ ^
10 0
. . ^
.vv^f
lead An gle
L^^^ yW^
1
^^^
-10
'^'^^^
-20 28
56
83
111
139
167
194
222
250
278
306
Time (ms) F I G U R E 4.5. Kinematic profiles for the bolitoglossine plethodontid salamander Hydromantes platycephalus. The panels are, from top to bottom, tongue reach and gape distance, gape distance, jaw deflections, and head angle, all presented on the same time axis. Gape distance is presented in both the first and the second panels, on different ordinates, to illustrate the vast differences in excursion between the tongue and jaws. This preycapture event was performed with small prey and lacks a distinct second phase of mouth opening, which often occurs during tongue retraction.
4. Terrestrial Feeding in Salamanders form a complex, pimiate muscle wrapped around the tapered epibranchial. The protractor muscles are differentiated in the plethodontids into an anterior, parallel-fibered portion and a distinct, posterior, bulb-like muscle, which is greatly enlarged in relation to that found in members of other families. The hyolingual apparatus contains two triangular units, formed by the medial basibranchial and the ceratobranchials on either side, so in order to form a compact bundle without bending, it folds in three dimensions as it is protracted, moving along a morphological track having the geometrical form of a tractrix. This form has favorable attributes. It appears to act as an accelerator of movement and contributes to the rapidity of protraction. It also acts as a brake for the returning apparatus. The three-dimensional expansion of a tractrix is known geometrically as a bugle body, and it exists, in a limited degree, in the floor of the mouth. The sides of the bugle body are formed by the medial margins of the ceratohyals. The bottom is formed by intermandibularis muscles and the top by a strap-like, unpaired suprapeduncularis muscle (the last muscle unique to the Plethodontidae). The structural element thus formed, called the cylinder by Lombard and Wake (1977), controls the direction of tongue protraction. In species that are most proficient in tongue protraction, members of the plethodontid tribe Bolitoglossini, the cylinder is well formed and incorporates a number of muscles that serve different functions in other taxa. These muscles include the anterior fibers of the subarcualis rectus I, the geniohyoideus medialis, anterior fibers of the rectus cervicis superficialis, and the hebosteoypsiloideus. According to the biomechanical model of Lombard and Wake (1976, 1977) movement of the cylinder from side to side within the mouth is possible. The cylinder appears to rest in the floor of the mouth attached posterolateral^ to the mandible by a slender and poorly defined mandibulohyoid ligament and anteriorly by the geniohyoidius lateralis. Thus there appears to be a kind of "firing platform," and a relatively large contraction of the left geniohyoideus lateralis is hypothesized to direct the tongue toward the right. The biomechanical model also hypothesizes that a mechanical linkage between parts of the hyolingual apparatus accomplishes at least a partial rotation of the tongue pad during protraction. The very action of folding is biomechanically coupled to rotate the tongue pad around the tip of the basibranchial so that the fleshy, mucous-covered pad contacts the prey. A ligament-like bundle of connective tissue extends from the anterolateral part of the first ceratobranchial into the substance of the tongue pad. This fiber bundle extends anteriorly from each side, coalescing at the ventral midline of the anterior tip of the basibranchial and then
105
variously attaching to a flexible tip of the basibranchial, to a detached anterior part of the basibranchial known as the lingual cartilage, or extending dorsally around the tip of the basibranchial and then fanning out posteriorly into the substance of the tongue pad. When the skeleton folds as it is protracted, the fibers become taut and the tongue pad is pulled forward to rotate around the tip of the basibranchial in such a fashion that the sticky dorsal surface is presented to the prey. The cylinder of plethodontids is lined along its inner surface by serous glands, which apparently lubricate the bundle within it during protraction. Protraction can result in the bundle being rapidly propelled forward and can be so great as to result in the skeletal elements, totally evacuating the cylinder as well as the bulb formed by the subarcualis rectus muscles. Momentum of the projectile carries the epibranchial cartilages fully out of the mouth in species of the genus Hydromantes and other bolitoglossine species we have observed. In these taxa the tongue is fired ballistically from the mouth as a projectile (Fig. 4.6; Deban et ah, 1997). Auxiliary protraction mechanisms function in other, nonplethodontid taxa. The subhyoideus connects the posterolateral parts of the ceratohyals to the mandible (see Fig. 3.5B). When these paired muscles contract, the ceratohyals are moved forward as a first stage in tongue protraction. The subarcualis rectus are presumably firing at the same time so the two-stage protraction involves (1) the anterior movement of the ceratohyals carrying with them the entire tongue apparatus, and (2) the independent protraction of the articulated hyobranchial apparatus and attached tongue pad relative to the first segment (Findeis and Bemis, 1990; Miller and Larsen, 1990). This has been termed the "mobile ceratohyal system" by Findeis and Bemis (1990), who contrast it with the other main evolutionary trend involving a "stable ceratohyal system" in plethodontids (which we believe is the ancestral condition). Another auxiliary system found in salamandrids (Chioglossa and Salamandrina) involves the rotation of elongated radii (Fig. 4.2) in an arc around the tip of the strengthened, mineralized basibranchial, which is "T" shaped in cross section (Ozeti and Wake, 1969). This action carries the tongue pad forward, effectively flipping the free posterior flap of the pad well out of the mouth. c. Prehension and the Tongue Pad The tongue pad varies considerably in shape among terrestrial salamanders. In general, it is attached firmly to somewhat loosely at the front, has varying degrees of freedom along the sides, and has the greatest
106
D a v i d B. W a k e a n d S t e p h e n M. D e b a n
F I G U R E 4.6. Photograph of Hydromantes genei capturing a housefly, showing projection of the hyoHngual apparatus. The projectile includes the retractor muscle and the bundle of cartilages wrapped in a cylindrical membrane of connective tissue and epithelium. The tongue pad has enveloped the fly, but is not extended so maximal reach for the tongue would be considerably longer than is apparent in this photograph. The epibranchial has left the borders of the mouth. The bends near the distal tip of the projectile indicate the positions of the end of the basibranchial and the epibranchial-ceratobranchial joints.
degree of freedom posteriorly. The posterior part of the tongue may be extended into two limbs that are loose and form relatively long strands during full protraction. This is especially true in some salamandrids and some plethodontids. In two plethodontid clades the tongue has lost its anterior attachment (the genioglossus muscles are absent) and the tongue has a mushroom-like shape. In these species the pad typically becomes relatively small and round; these groups show the greatest distance of tongue protraction. Shaping of the tongue pad may be facilitated by the contraction of muscles inside the tongue pad that arise from the tip of the basibranchial and fan out into the posterior part of the tongue (hyoglossus), extend to the tips of the radii (basiradialis), extend between the tips of the radii (interradialis), or extend from the tips of the radii into the tongue pad (radioglossus). For the plethodontid salamanders, Lombard and Wake (1977) proposed three functional classes of tongue pad muscles: rotators (genioglossus, circumglossus, basiradialis, and intraglossus), molders (interradialis and hyoglossus), and restorers (rectus cervicis profundus). Families differ substantially with respect to relative numbers, sizes, and proportions of skeletal elements and muscles pres-
ent, and the tongue pad varies significantly among taxa; however, this diversity has not been the subject of detailed comparative analysis. The surface of the tongue is rotated by a combination of skeletal and muscular movements so that the glandular dorsal surface covered with relatively sticky mucous contacts the prey. The tongue pad shape is transformed and expanded during protraction so that it is relatively large and expansive when it contacts the prey. In those taxa having an otoglossal cartilage, movement of this element during protraction is postulated to carry the glandular field of the tongue pad dorsoanteriorly so that it contacts the prey (Larsen et ah, 1996). In ambystomatids, the tongue pad is rotated forward as protraction occurs via the connection between the first radii and the hypohyals. The mechanism of tongue protraction in plethodontids is somewhat simplified in relation to the other group that has been studied in detail, the ambystomatids (Reilly and Lauder, 1989, 1990b). In the ambystomatids the ventral throat constrictors (intermandibularis and interhyoideus) and two longitudinal muscles (geniohyoideus and genioglossus) are active during protraction. The constrictors and the geniohyoids are
4. Terrestrial Feeding in Salamanders implicated in provision of a lift vector to the hyolingual apparatus. The subarcualis rectus I provides a separate forward and upward vector, and when well protracted the geniohyoids and genioglossus provide a third forward and ventral vector to the protraction. The combined effect of these three vectors is a resultant horizontal vector that advances the entire hyolingual system, and flipping of the tongue pad is attributed in part to the action of the genioglossus. d. Retraction The articulated hyolingual skeleton has one retractor muscle attached to it at the posterior border of the anterior end of the first ceratobrancial. This is a muscle with several names, but is usually termed the sternohyoideus or the rectus cervicis lateralis (in some taxa the muscle is not well differentiated from the rectus cervicis superficialis). It arises from the sternum and is an anterior continuation of the rectus abdominis series. However, this muscle appears to be a secondary retractor, and the main retractor is a muscle that extends forward under the ventral surface of the hyolingual skeleton and passes through the triangular gap between the first and the second ceratobranchials and the basibranchial. From this position the paired muscles extend forward and then bend abruptly dorsally to enter the tongue pad, where they typically attach to a mass of connective tissue long known by the German term Sehnenplatte ("tendon plate"). This muscle is also known by various names, including the abdominohyoideus and the rectus cervicis profundus. This muscle arises from the lateral margin of the sternum, from a tendinous inscription lateral to the sternum that separates it from the rectus abdominis profundus, or as a direct and continuous anterior extension of the last named muscle, which in extreme cases (Plethodontidae) represents an undivided muscle that arises from the posterior border of the ischium and proceeds uninterrupted into the tongue pad. Retraction of the tongue appears to be exclusively the result of the action of the two different segments of the rectus cervicis series. In the most proficient tongue protractors, the bolitoglossine plethodontids, the rectus cervicis lateralis muscle is absent in two of the three major lineages, and retraction is exclusively by the rectus cervicis profundus. The omohyoideus is also missing in these taxa, and the hebosteoypsiloideus, which is part of the general retractor system in most taxa, is incorporated into a muscular cylinder through which the tongue is folded as it is protracted. e. Speed and Distance of Lingual Feeding The effectiveness of lingual feeding has not been studied in most taxa. As a generalization, species dem-
107
onstrating true tongue projectility are the fastest, the most accurate, and have the greatest range. The fastest tongue recorded to date is that of Bolitoglossa occidentalis, which takes less than 10 msec from the start of electrical activity in the subarcualis rectus I muscles to the time the tongue strikes the target. Maximum tongue extension (from tip of snout) has been measured in a number of species. In six species of Ambystoma the largest average distance was 2.4 ± 0.3 mm {A. californiense) and the smallest was 0.3 ± 0.4 mm {A. niabeei){Beneski et al., 1995). Hynobiids can extend their tongues only a few millimeters beyond the tip of the snout, but as much as 7% of snout-vent length (Larsen et al, 1989, 1996). However, Hynobius kimurae was found to protract its tongue 4 - 6 mm beyond the symphisis (J. Larsen, personal communication), and in our laboratory a large (6 to 7 cm snoutvent length) Salamandrella keyserlingii has protracted its tongue 6.6 mm, thus some hynobiids are apparently far more proficient than those reported to date. Salamandrids are apparently the most variable with respect to maximal tongue extension. Pachytriton brevipes, an aquatic species that lacks a defined tongue pad (Ozeti and Wake, 1969), retracts rather than protracts the hyolingual apparatus during terrestrial feeding (Miller and Larsen, 1990), thus in essence performing an aquatic, suction-feeding behavior on land (see Chapter 3 for details of this behavior, including Fig. 3.2). In most species of salamandrids, tongue protraction is relatively short, from 1.1 to 2.7 mm (average extension beyond snout), the latter distance in Tylototriton verrucosus (Miller and Larsen, 1990). Salamandrina terdigitata has a complex tongue protraction involving an initial protraction of the hyolingual apparatus from the mouth and then flipping of the pad (Ozeti and Wake, 1969; Miller and Larsen, 1990). The average maximal tongue extension recorded by Miller and Larsen (1990) for this species is 7.4 mm, or 20% of snout-vent length (4.7 mm is reported in the paper, but a printer's error reversed the digits; J. Larsen, personal communication). Chioglossa lusitanica probably has the longest tongue extension of the salamandrids, based on its morphology (Ozeti and Wake, 1969), but it remains largely unstudied. Maximum tongue extension is variable in plethodontids; distances are about 7% of snoutvent length in species with attached tongues {Desmognathus quadramaculatus and Plethodon glutinosus), about 15% in a species with an attached protrusible tongue {Ensatina eschscholtzii), and 30-44% in Bolitoglossa occidentalis (Thexton et ah, 1977; Larsen et ah, 1989). This species is capable of firing the tongue effectively for at least 17 mm from the head, or as much as 44% of the snout-vent length of the salamander (Thexton et ah, 1977). Hydromantes italicus, a larger species, is reported
108
David B. Wake and Stephen M. Deban
to be capable of capturing prey 40 mm from the head, and the intermediate-sized Bolitoglossa subpalmata can project its tongue 30 mm (Roth, 1987). Adult Hydromantes supramontis can project the tongue accurately to a distance of about 80% of snout-vent length (over 60 mm from the head)(Fig. 4.6; Deban et al, 1997). The prey-capture success rate has been measured in only a few species, but is higher in the proficient tongue protractors than in more generalized species. Salamandra salamandra was successful in only 39% of attempts (Luthardt-Laimer, 1983) compared with above 50% in the generalized plethodontid genera Plethodon, Eurycea, and Batrachoseps. The very long tongued bolitoglossine genera Bolitoglossa and Hydromantes "rarely miss," although no exact figures are available (Roth, 1987). These last two genera may engage in "stalking" in the laboratory to move slowly into range (Roth, 1987; personal observation). The total length of the kinematic cycle is variable among the different taxa studied. The mean gape cycle (in species with tongue protrusion) ranges from 92.8 to 115.7 msec in hynobiids (Larsen et al, 1996), 78 to 214 msec in ambystomatids (Beneski et al, 1995), 100 to 238 in salamandrids (Miller and Larsen, 1990), and 87 to 110 in plethodontids (Larsen et al, 1989; Beneski and Larsen, 1988). Gape cycle time is apparently rather similar across a wide array of morphologies. Speed of the tongue strike (tongue protraction, equivalent to phase II of Beneski and Larsen, 1988) varies greatly among taxa. The fastest tongues are found in Bolitoglossa, which has a mean duration of phase II of 5.0-7.7 msec, with Ensatina at 11.6 msec being only a little slower (although we have measured one individual Ensatina at ca. 7 msec from the time the tongue left the mouth until it touched the prey). Pseudotriton ruber, a hemidactyliine with a free tongue, has a protraction time of as little as 11 msec (Deban, 1997). The plethodontids with attached tongues are slower still: Plethodon glutinosus takes 19.3 msec and Desmognathus quadramaculatus takes 37.3 msec to protract the tongue (Larsen et al, 1989). Roughly comparable times are reported for other taxa (different papers use slightly different methods of reporting). Ambystomatids range from 16 msec for Ambystoma mabeei to 87 msec for Ambystoma cingulatum (Beneski et al, 1995). Maximum tongue protraction in Ambystoma tigrinum is reported by Reilly and Lauder (1989) to take 45 msec, and by Dockx and De Vree (1986) to take 39.6 ± 1 1 . 0 msec. Hynobiids are reported to range from 25.1 msec for Hynobius kimurae to 36.0 msec for H. nebulosus by Larsen et al (1996) [27.4 msec and 36.0 msec for the same species by Larsen et al (1989)]. Among salamandrids, tongue protraction times are reported by Miller and
Larsen (1990) to range from 22.0 msec in Salamandra salamandra to 111.9 msec in Paramesotriton hongkongensis. Dockx and De Vree (1986) report 84 ± 21.7 msec for tongue protraction in Salamandra salamandra, and Findeis and Bemis (1990) report a range of durations from 80 to 140 msec for Taricha torosa. f. Physiology Electromyographic (EMG) investigations of tongue movement (Thexton et al, 1977; Reilly and Lauder, 1990b) in two very different taxa (a bolitoglossine plethodontid with a highly protractible, free tongue and an ambystomatid with a weakly protrusible, attached tongue) show that the protractor (subarcularis rectus I) and the retractor (rectus cervicis) have essentially synchronous onset and similar motor patterns. Because electrical signals are delivered to the protractors and retractors simultaneously, the activity of the system is thought to be controlled by the peripheral organization of the system. The length-tension properties of the protractors favor immediate force generation, whereas those of the much longer (and in terrestrial plethodontids, somewhat lax and even looped) retractors have delayed biomechanical activity. Differences in contractile properties of the two muscles, such as time to peak tension, remain a possibility that is yet unstudied. The similarity in EMG patterns of the two species studied may suggest that motor patterns are phylogenetically conserved. While the electromyographic patterns of larval and metamorphosed ambystomatids suction feeding in water are remarkably similar, they are radically different from those of adults feeding on land (Shaffer and Lauder, 1988; Lauder and Shaffer, 1988). In general, there is little variability detected in experiments, and the motor patterns appear to be highly stereotyped. These facts suggest that changes in the peripheral arrangement of the muscles, skeletal elements, and other connective tissues are critical in determining biomechanical function during feeding. Most of the physiological work has involved ambystomatids (Reilly and Lauder, 1990b). The epaxial muscles connecting the neck to the head and the mandibular depressors are active prior to mouth opening (which involves raising the head slightly to greatly in all species studied) and they remain active until mouth closing. There is a second peak of activity in the epaxial muscles during tongue retraction. The jaw adductors are also active prior to mouth opening, but they have a large burst of activity again as the jaws close. The muscles responsible for tongue protraction (subarcualis rectus) and pad flipping (genioglossus) have peaks of activity during the tongue protraction phase, but unexpectedly the subarcualis rectus muscles have
109
4. Terrestrial Feeding in Salamanders a second peak during the retraction phase. Buccal muscles in the floor of the mouth are active throughout the cycle, but tend to peak early and then taper off. The main retractor muscle (rectus cervicis profundus) peaks at the time retraction begins and activity continues until after mouth closing. We take the coactivity of antagonistic muscles as a sign of fine control over tongue and jaw movements and the prolonged activity of the buccal floor muscles as an indication that they are performing a stabilizing function. In Bolitoglossa strain gauges were used to measure force of tongue impact with the prey (Thexton et ah, 1977). At relatively short distances, force varies considerably, probably indicating differences in motivation and concomitant muscle activity level. However, at distances exceeding 10 mm from the snout (animals have a maximal head-body size of about 44 mm), force fell off dramatically as a function of distance from the snout, finally failing to register at the greatest distance (19 mm). This supports the idea that there is a ballistic phase (at the end of muscular protraction) to long-distance tongue projection in the species most proficient in tongue protraction. 2. Jaw Feeding Use of the jaws to capture prey is unusual in terrestrial salamanders, but it does occur, at least in members of the families Ambystomatidae and Salamandridae (Larsen and Guthrie, 1975; Miller and Larsen, 1990). We have also observed it infrequently in various terrestrial species of the Plethodontidae in circumstances when the tongue fails to apprehend the prey. In addition, we have observed the semiaquatic plethodontid Desmognathus quadramaculatus and the hynobiids Salamandrella keyserlingi and Batrachuperus persicus using jaw prehension in water and the fully aquatic plethodontid Desmognathus marmoratus using either tongue or jaw prehension in water (see also Schwenk and Wake, 1993). Kinematics of the gape profile of salamanders using jaw prehension differs from either the generalized four-part or the specialized three-part pattern of other taxa and consists of a bell-shaped, two-part gape profile. The action is relatively rapid (on the order of 60 msec) and resembles that of fully metamorphosed ambystomatids that feed underwater without using tongue protraction (Reilly and Lauder, 1989). While the standard pattern of feeding in salamanders involves participation of the tongue in apprehending the prey, species such as Pachytriton hrevipes, which effectively lacks a tongue pad (Ozeti and Wake, 1969), and other mainly aquatic salamanders feeding on land (Miller and Larsen, 1989) can use jaw prehension, al-
though not very effectively. Jaws are used in aggressive encounters in salamanders (Staub, 1993), and it is evident that biting without tongue protraction is possible. Nonetheless, we expect that salamanders feeding in terrestrial situations typically will display a pattern of feeding that involves tongue prehension (Bramble and Wake, 1985). D . Prey Processing Prey immobilization is significant when the prey are too large to be fully engulfed at the time of capture. Large-bodied salamanders are capable of eating long and slippery (e.g., earthworms in ambystomatids) and even very large prey (e.g., mice in dicamptodontids) relative to body size. In these species, strength of the jaw-closing muscles is important, as well as size and strength of the marginal tooth-bearing bones. The mouth is closed at the end of the strike and if the prey protrudes from the mouth there may be a delay of a few milliseconds to several seconds before processing and/or swallowing proceeds. This process has been most thoroughly studied in ambystomatids and is described later. Dicamptodontids have heads that are the largest of terrestrial salamanders both absolutely and relatively, and they seize (it is not recorded if this is by lingual prehension or jaw prehension) prey such as mice and hold them in the buccal cavity until the prey suffocates before further processing proceeds. Two groups of plethodontids have special prey-processing features. Aneides includes species (e.g., A. lugubris) that have greatly enlarged adductor muscles, jaws and marginal teeth (Wake, 1963). These may well be associated mainly with aggressive behavior (e.g., Staub, 1993), but they also enable these species to eat larger food than co-occurring species of similar size (Lynch, 1985). Desmognathine plethodontids engage in a unique behavior, cranial ventroflexion or head tucking (Dalrymple et al, 1985; Larsen and Beneski, 1988; Schwenk and Wake, 1993), that is enabled by the presence of a number of morphological modifications of the head and neck regions, including a ligamentized tendon extending over either side of the skull from a specialized ridge on the atlas to the lower jaw. This behavior is characteristically performed after lingual capture of the prey and involves a sharp ventroflexion of the head relative to the neck with the prey caught in the jaws. Head tucking may occur as a final component of mouth closing following capture, or without mouth opening, and it represents an extreme form of a static pressure feeding system (Olson, 1961) in salamanders. The result is the penetration of the prey by the teeth and the effective immobilization of the prey. Both Aneides and the
110
David B. Wake and Stephen M. Deban
desmognathines have enlarged quadrapectoralis and an additional muscle (often termed gularis) that contribute to their ability to deliver a strong bite. Prey transport has been studied most extensively in Ambystoma tigrinum (Reilly and Lauder, 1990a, 1991b; Gillis and Lauder, 1994), and only fragmentary information is available for other taxa (Dockx and De Vree, 1986; Thexton et al, 1977). Four phases of intraoral prey transport are recognized in metamorphosed ambystomatids feeding on worms on land (as well as in the water as larvae): preparatory, fast operung, closing, and recovery. Transport per se is accomplished during fast opening and closing. Electromyographic recordings reveal that the preparatory phase involves electrical activity in muscles of the buccal floor (both the longitudinal genioglossus and geniohyoideus, and the transverse intermandibularis and interhyoideus) and then there is a short period of electrical silence before a fast opening of the mouth that is accompanied by simultaneous activity of all muscles studied, even antagonists such as the mandibular depressors and adductors (Reilly and Lauder, 1991b). Reilly and Lauder (1990b) divided the preparatory phase into two parts. In the first part the prey item is pressed against the roof of the mouth by the elevated hyoid apparatus. Gape increases slowly but is never great during phase one. During the second phase, which is shorter in duration, the gape is held constant. Then the fast opening phase occurs, followed quickly by a closing phase, and in each cycle of opening and closing between 4 and 8 mm of prey is transported. Intraoral transport of the prey occurs at the beginning of mouth opening, as the rectus cervicis contract, thus retracting the hyobranchial apparatus and moving the tongue pad (to which the prey is sticking) posteriorly into the pharynx. The muscle that protracts the tongue during prey capture, the subarcualis rectus I, is also active during this time, and Reilly and Lauder (1991a) postulate that it may act antagonistically to the retractors so as to stabilize the interaction of the articulated hyobranchial apparatus and the ceratohyal and to enable the system to function as a whole in aiding the rapid posterior movement of the tongue. There is no inertial component to prey transport. The mouth closes as the prey moves backward. During recovery, hyobranchial protraction carries the tongue forward under the prey, for now the prey is held by the combination of palatal and marginal teeth. The intraoral prey transport system in Ambystoma is very similar during aquatic and terrestrial feeding. However, electromyographic patterns differ between prey capture and transport (reviewed by Lauder and Gillis, 1997). Variation in feeding kinematics is event specific rather than reflecting the environment in
which it occurs (Gillis and Lauder, 1994). Reilly and Lauder (1991) hypothesize that prey transport in terrestrial feeding retains an ancient (extending to fishes) motor pattern associated with suction feeding by larvae. The facts that prey transport behaviors are similar in fishes and salamanders and that they are faster and involve less excursion than prey capture have been used to hypothesize that the prey transport system of terrestrial salamanders may have been directly inherited from the aquatic transport behavior based on suction in aquatic ancestors (Gillis and Lauder, 1994; Lauder and Gillis, 1997). Many salamanders, especially plethodontids but also members of other families, eat small prey that are ingested fully on capture. In these species the marginal dentition does not contact the prey, and transport within the buccal cavity differs from the pattern seen in ambystomatids. Frequently the mouth is not opened again, but evidence shows that the tongue is repositioned and then retracted further, moving the prey into the pharynx. Thexton et al. (1977) recorded two to three bursts of activity in the rectus cervicis muscles following prey capture in the diminutive plethodontid Bolitoglossa occidentalis. This implies that the tongue is being protracted and retracted repeatedly with the mouth closed, and that swallowing follows. E. Modulation of Feeding Behavior Little attention has been given to modulation of feeding behavior under different conditions. Substantial modulation is possible under different circumstances, especially with respect to differences in prey (Deban, 1997; unpublished observations). The species that have been most intensively studied to date are Salamandra salamandra (earlier work reviewed by Roth, 1987; see also Reilly, 1995) and Ensatina eschscholtzii (Deban, 1997); other species studied less intensively include various salamandrids (e.g.. Miller and Larsen, 1990) and plethodontids {Hydromantes italicus, Roth, 1987; Plethodon cinereus, Maglia and Pyles, 1995). The greatest modulation occurs in newts that feed both on land (using tongue protraction) and in the water (using suction feeding) (Miller and Larsen, 1990). Bolitoglossa occidentalis showed little evidence of modulation in early studies (Thexton et al, 1977; Larsen et al, 1989), but our observations show a high degree of modulation of the timing and extent of tongue movements in several species of bolitoglossines. Among ambystomatids, some species (e.g., Ambystoma tigrinum) are apparently highly stereotyped in tongue protraction, whereas others {Ambystoma macrodactylum) show evidence of sonie modulation (Larsen and Guthrie, 1975).
4. Terrestrial Feeding in Salamanders When tested with two distinctly different kinds of prey (waxworms and termites), Ensatina eschscholtzii modulated both the timing and the magnitude of tongue and jaw movements with respect to different prey. When feeding on waxworms, the larger prey, feeding took less time and tongue and jaw movements attained a higher velocity than when feeding on termites (Deban, 1997). In Plethodon cinereus, maximal tongue extension was as great as 17% of snout-vent length when feeding on adult Drosophila (mean 10.4%), but as little as 1% (mean 4.5%) when feeding on larval Drosophila (Maglia and Pyles, 1995). In Ensatina, distance of protraction correlated with distance of the prey from the head, not prey type. In both species, prey capture is completed more quickly on large than on small prey.
IV. DIVERSITY A N D EVOLUTION A. Origins and Outgroups Extant out-group taxa for the order Caudata include the orders Anura and Gymnophiona, both of which differ profoundly with respect to feeding mechanisms from all salamanders. There has been a general consensus that the three orders of living amphibians (Lissamphibia) were derived from a temnospondyl labyrinthodont stock; one widely accepted hypothesis is that lissamphibians might be a sister taxon of the temnospondyl group Dissorophoidea (Bolt, 1977). Another alternative is that lissamphibians are derived from a lepospondylous ancestral stock, perhaps somewhere in the microsaur radiation (Laurin and Reisz, 1996). A further alternative is that lissamphibians do not form a monophyletic group with respect to fossils (Carroll and Holmes 1980). Regardless of phylogenetic considerations, living salamanders have a hyobranchial apparatus that is more generalized in morphology and more similar to known fossils (whether temnospondyls or lepospondyls) than either frogs or caecilians, and we (in accord with Lauder and Reilly, 1994) consider salamanders to be an appropriate model for the first terrestrial feeding system. B. Phylogenetic Diversity Six families of salamanders have species that metamorphose and feed on land: Hynobiidae, Rhyacotritonidae, Dicamptodontidae, Ambystomatidae, Salamandridae, and Plethodontidae. The last three families have been studied in greatest detail with respect to feeding mechanisms and the last two display great diversity.
1.
111
Hynobiidae
There is a general consensus that hynobiids are the most basal of the terrestrial taxa of the Caudata (Larson and Dimmick, 1993), and certain features of their feeding mechanism retain apparent ancestral states. The most evident of these is the retention of two branchial arches, with two epibranchials. All other terrestrial salamanders have a single epibranchial, supported by two ceratobranchials. There are, however, indications that hynobiids do not retain the ancestral structure, because they have a unique feature, the radial loop. The ceratohyals are connected to the basibranchial by means of elongated, attenuate hypohyal derivatives arranged in the form of a flat spring (Fig. 4.2). This arrangement is associated with a modest degree of tongue protraction, which occurs with great speed (Larsen et al, 1989,1996; unpublished data). The tongue pad itself has been described as "sac-like" (Larsen et al., 1996) and lacks differentiated musculature, and the basibranchial is extended forward in the radial loop, pulling it along as the hyobranchium is protracted. The extent of protraction is apparently limited by the structural connection of the the hyoid arch to the basibranchial, a connection that exists to a more limited extent in other terrestrial salamanders that do not protract their tongues so far. Tongue protraction is accompanied by a strong forward lunge of the entire body of the salamander, so the effective strike distance is relatively great. Apart from the radial loop, which represents the first radii of other families, the structural arrangement of the tongue in hynobiids is generalized. The second basibranchials are ossified, whereas the first remain cartilaginous, suggesting that the second basibranchials are the main force-transmitting elements from the protractile musculature to the tongue pad. There is a well-developed subhyoideus muscle in hynobiids, and the subarcualis rectus I muscle is wrapped around both first and second epibranchials. These two muscles, working together, serve to protract the hyobranchial apparatus, but functional morphology of the musculature has not been well studied (Severtsov, 1971). A pair of "cornua" are found at the anterior end of the basibranchial in many hynobiids (Cox and Tanner, 1989); these are paired anteriolaterally directed projections continuous with the anterior end of the basibranchial, and they are drawn out into long processes in Onychodactylus (Fig. 4.2). Cornua resemble the radii of plethodontids, but phylogenetic analysis suggests that the resemblance is homoplastic. The hynobiid condition has been proposed to be the most basal salamander feeding mechanism, based on the combination of its morphology and the fact that it
112
David B. Wake and Stephen M. Deban
displays a generalized four-part feeding cycle (Larsen et al, 1996; Beneski et al, 1995). However, the specialized nature of the radial loop precludes the possibility that it retains the ancestral morphology or function entirely. 2.
Rhyacotritonidae
The Rhyacotritonidae has only recently been recognized as a distinct taxon that is relatively basal phylogenetically within the Caudata. Rhyacotritonids have a generalized hyobranchial morphology (Fig. 4.2). The first radii are present but not elongate as in the hynobiids. There is an otoglossal cartilage, an unpaired, dorsal, medial element that either joins the tips of the two second radii or lies freely between them. The articulated elements of the hyolingual skeleton remain cartilaginous, and there is only a single pair of very short epibranchials. The musculature is also generalized. The family resembles the Hynobiidae and the Salamandridae in having a subhyoideus muscle, which may function in a two-phase tongue protraction system. 3.
Dicamptodontidae
The four species comprising this family are large, robust animals, three of which have a terrestrial stage. The tongue is unique and largely unstudied. It combines elements of the tongues of hynobiids, rhyacotritonids, and ambystomatids. The first radii are drawn into an attenuated loop that recalls the situation in hynobiids but is less extreme. The second radii are joined by a thin cartilaginous plate (Fig. 4.2) that appears to be the homologue of the otoglossal cartilage of rhyacotritonids and ambystomatids and supports the relatively enormous tongue pad that is used to capture prey of a wide range of sizes. There is no subhyoideus muscle. The single pair of epibranchials are short and ossified near their proximal ends. 4.
Ambystomatidae
This is a large family of North American salamanders that has been relatively well studied. Both otoglossal cartilages and two pairs of radii are present. However, in contrast to the situation in hynobiids, rhyacotritontids, and dicamptodontids, the first radii are neither attenuate nor drawn into a radial loop that is continuous with the hyoid arch, but their distal tips are loosely connected to the free anterior ends of the ceratohyals. The second radii may be forked and elongate in the subgenus Linguaelapsis (Fig. 4.2). Ambystomatids lack a subhyoideus muscle. They have a muscle termed the genioglossus lateralis that lies in the same general area as the subhyoideus but that has an entirely different origin and innervation
(Piatt, 1940). This muscle may be used to move the ceratohyals medially, thus orienting the tongue. 5.
Salantandridae
Basal salamandrids have two pairs of radii. The first radii are typically rather short and tapered, with free distal ends. The first radii are apparently lacking in Salamandrina and Chioglossa, in which the second radii have moved to the distal tip of the basibranchial and rotate around it as the tongue pad is flipped. The radii are moderately long and robust in Salamandrina, but very long and attenuate in Chioglossa (Fig. 4.2). In both genera the basibranchial is stout and is the only part of the articulated hyobranchial apparatus that is mineralized. In Salamandrina it is T shaped in cross section. In other salamandrid genera the second radii are located behind the first radii, and an interradial cartilage (possibly a homologue of the otoglossal of other taxa) extends between them. Tongue pad musculature is relatively complicated in the Salamandridae, and several muscles occur that are not found in other families (e.g., radioglossus) whereas others are larger than in other families (e.g., basiradialis). The most complicated tongue pad is found in Salamandrina, in which tongue pad flipping is thought to be accomplished by rapid rotation of the radii by the basiradialis muscles (Ozeti and Wake, 1969). Salamandrid genera were divided into two functional groups based on tongue structure and use of the tongue in feeding by Ozeti and Wake (1969). Most genera have a "water tongue" and they have been known as "Wassermolchen" in the German literature (Wolterstorff and Herre, 1935). In these genera the hyolingual apparatus is used for suction feeding in the water and for tongue prehension on land. Characteristically the skeleton of the hyolingual apparatus is relatively heavy and well ossified. At least one genus, Pachytriton, is permanently aquatic, has no tongue prehension ability, and has a very reduced tongue pad. Paradoxically, Pachytriton has a terrestrial eft stage (common in other newts) in which tongue prehension likely occurs (Thiesmeier and Hornberg, 1997), prior to the return as an adult to a permanently aquatic existence in which tongue prehension is impossible. The most terrestrial salamandrids, Salamandra, Mertensiella, Chioglossa, and Salamandrina, have "land tongues." They either do not enter water or feed in water using terrestrial behaviors and they have specialized tongues used for prehension of prey. Typically the hyolingual apparatus is either entirely cartilaginous or only the basibranchial is ossified. A significant part of tongue protraction in Chioglossa and Salamandrina is accomplished by flipping and extension of the large tongue pad (Miller and Larsen, 1990).
4. Terrestrial Feeding in Salamanders Salamandrids have a subhyoideus muscle (see Fig. 3.5 in Chapter 3) that is used to protract the ceratohyals and the tongue in general (Findeis and Bemis, 1990). Some salamandrids have some lateral muscle fibers arising near the genioglossus (e.g., Ozeti and Wake, 1969), and these may be homologues either of the genioglossus lateralis of ambystomatids or possibly the geniohyoideus lateralis of plethodontids. 6.
Plethodontidae
Otoglossals are absent in the Plethodontidae, and some taxa have a medial unpaired "lingual cartilage'' (Fig. 4.4). Rose (1996) has shown that the lingual cartilage may have some connection with the hypohyals of experimental animals treated with thyroxin. Wake (1966) argued that the ,cartilage is derived from the anterior extension of the basibranchial that lies in front of the attachment of the radii. Out-group taxa that have an anterior extension also have either a pair of first radii attached to it or bear a pair of cornua. It is possible that the anterior extension represents the fusion of the lingual cartilage to the basibranchial; perhaps the lingual cartilage should be considered a homologue of a long missing basihyal. The first radii of out-group taxa that have two pairs of radii lack muscular attachments, and because the single radii of plethodontids have muscular attachments they are best considered to be homologues of the second radii of other taxa. An alternative interpretation is that the radii of plethodontids may be homologues of the cornua of hynobiids, because in many plethodontids the radii are homocontinuous with the cartilage of the basibranchial, although in others they are articulated with the basibranchial, as are the second radii in members of other families. If, as argued earlier, the first radii represent hypohyals of out-group taxa, the lingual cartilage represents the basihyal, the distal (with respect to distance from the midline) element found in rhyacotritonids represents the epihyal, and portions of the middle ear complex are derived from the arch as well, all components of the entire hyoid arch are present in the Caudata, but not in any single taxon. There is no subhyoideus muscle in plethodontids. A well-developed geniohyoideus lateralis may play an important role in controlling lateral to medial movements of the ceratohyal and thus direct the firing of the tongue, but this has not been demonstrated behaviorally (Lombard and Wake, 1977). Feeding mechanisms in plethodontids have been studied extensively, and summary discussions of feeding in a phylogenetic context accompanied by evolutionary scenarios are presented by Roth and Wake (1985b), Lombard and Wake (1986), and Wake and Larson (1987). Phylogenetic analyses (Jackman et ah, 1997)
113
support the hypothesis (Wake, 1966; Lombard and Wake, 1986) that the freely projectile tongues of the supergenera Hydrotnantes and Bolitoglossa are derived independently. Both are derived independently of the even more phylogenetically remote members of the tribe Hemidactyliini, which also have evolved freely projectile tongues. Thus, within the Plethodontidae there are three lineages that have evolved highly specialized tongue projection mechanisms. The hemidactyliine tongue is folded according to one of two hypothetical possibilities (Lombard and Wake, 1976, 1977), called option 1. This folding pattern involves holding the first ceratobranchial and epibranchial coplanar during folding. It results in a relatively bulky projectile that is postulated to have more limited projectability than is involved in option 2 (Wake, 1982). This option has been hypothesized as a necessary consequence of the retention of an aquatic larval stage and associated larval suction feeding in hemidactyliines (Wake, 1982). In contrast, bolitoglossines, which all have direct terrestrial development and no aquatic larval stage, all use option 2 folding, in which the second ceratobranchial and the epibranchial are held coplanar. This option results in a slenderer projectile that is less limited in length than in hemidactyliines. There are two modifications of the option, that of Hydrotnantes, which has apparently optimized for distance by having evolved extremely long epibranchials, and that of the supergenus Bolitoglossa, which also has elongated epibranchials but substantially shorter than in Hydromantes, and is apparently optimized for speed (Larsen et ah, 1989). C. Feeding Biology and Evolution 1. Ecology and Selective Regime The evolution of terrestrial feeding in salamanders has proceeded in a great diversity of habitats and microhabitats, and in the absence of information about close sister taxa of Caudata it is impossible to reconstruct the habitat occupied by the first terrestrial feeders. However, it seems likely that early salamanders had a biphasic life cycle, involving feeding in water and on land, and that courtship and mating probably occurred in the water. Accordingly, feeding most likely entailed suction feeding both as larvae and as adults, and tongue protraction, apprehension, and capture on land. The most specialized terrestrial feeding mechanisms tend to be associated with species that do not spend much time in the water, or that are entirely terrestrial. The exception to this generalization is the family Hynobiidae, in which species with aquatic larvae and semiaquatic habits as adults have a hyolingual system that is fast and biomechanically specialized (e.g..
114
D a v i d B. W a k e a n d S t e p h e n M. D e b a n
Larsen et al, 1996). In the hemidactyhine plethodontids and in salamandrids such as Chioglossa and Salamandrina, tongues are also highly specialized for speed and distance of protraction, and these forms all have aquatic larvae. However, all are lungless or have reduced lungs, and this appears to be a necessary, but not sufficient, precondition for the high specialization of the tongue (Roth and Wake, 1985b). Rhyacotritonids, whose tongue and feeding are relatively unstudied but which have a generalized morphology, also have reduced lungs. The greatest specialization in tongue speed and protrusion is foimd in the bolitoglossine plethodontids, all of which lack an aquatic larval stage. These species have diverged far beyond the ancestral home of not only salamanders as a group but of plethodontids, and occur in a great diversity of terrestrial habitats and microhabitats, including waterless caves, cavities in trees, epiphytes, scrublands, and other settings in which foraging is restricted by environmental considerations. Whatever the environment in which tongue specialization evolved, it is a highly effective feeding strategy. 2.
Homoplasy
Homoplasy is common in the evolution of terrestrial feeding mechanisms of urodeles and has been discussed extensively (e.g.. Wake, 1966, 1982; Lombard and Wake, 1986). What has attracted the greatest attention is tongue protrusion. While some degree of tongue protrusion occurs in all salamanders that metamorphose and have a terrestrial feeding stage, longdistance tongue protrusion has evolved independently along very different biomechanical pathways within at least three families (Hynobiidae, Salamandridae, and Plethodontidae). In hynobiids, tongue protrusion is associated with modest increases in length of the two pairs of epibranchials and with the flat spring arrangement of the hyoid loop, which is attached to the articulated lingual skeleton in some way. However, the basibranchial remains short. Paired, relatively elongate cornua appear in Onychodactylus (Fig. 4.2), and while these may represent retained ancestral elements homologous to the second radii of other taxa, phylogenetic analysis suggests homoplasy. Tongue protrusion is never as great as in the other two families, but the tongue is fast and maneuverable. Larsen et al (1996) argue that the radial loops and the attachment of the ceratohyals to the suspensorium constitute a functional constraint that mechanically limits the extent of tongue protrusion. Wntiether there has been homoplasy within the Hynobiidae awaits a modem phylogenetic analysis of the family and further morphological studies.
V. OPPORTUNITIES FOR FUTURE RESEARCH Research effort has been unevenly distributed with respect to salamander clades. Biomechanical models of tongue protraction have been produced and tested for plethodontids and ambystomatids, but are largely lacking for other taxa. Detailed morphological studies date to the early part of this century and do not include several families, including Dicamptodontidae, Hynobiidae, and Rhyacotritonidae. However, comparative anatomical studies of some of the larger families, including Ambystomatidae, Plethodontidae, and Salamandridae, are relatively complete. What is needed is a broad and integrated comparative anatomical analysis of the musculoskeletal system of all the families, with special attention given to the establishment of homologies. Quantitative studies of kinematics are limited to only a few species and in general have been conducted in artificially controlled conditions and with limited, sometimes unnatural, prey. The modulation of behavior has been investigated in only a handful of species, even though most species feed on a diversity of prey and under a variety of conditions. Diet and foraging are in need of further study in most taxa. Two families, the Dicamptodontidae and the Rhyacotritonidae, have been especially understudied with respect to kinematics. References Beneski, J. T., Jr., and J. H. Larsen, Jr. (1989) Interspecific, ontogenetic, and life history variation in the tooth morphology of mole salamanders (Amphibia, Urodela, and Ambystomatidae). J. Morph. 199(1): 53-70. Beneski, J. T., Jr., J. H. Larsen, Jr. and B. T. Miller (1989) Ontogenetic alterations in the gross tooth morphology of Dicamptodon and Rhyacotriton (Amphibia, Urodela, and Dicamptodontidae). J. Morph. 199(2): 165-174. Beneski, J. T, Jr., and J. H. Larsen, Jr. (1995) Variation in the feeding kinematics of mole salamanders (Ambystomatidae: Ambystoma). Can. J. Zool. 73:353-366. Bishop, S. C. (1941) The salamanders of New York. New York State Mus. Bull. No. 324:1-365. Bolt, J.R. (1977) Dissorophoid relationships and ontogeny, and the origin of the Lissamphibia. J. Paleo. 51:235-239. Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms in lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. E Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Bury, R. B. (1972) Small mammals and other prey in the diet of the Pacific Giant Salamander (Dicamptodon ensatus). Am. Midi. Nat. 87:524-525. Carroll, R. L. and R. Holmes (1980) The skull and jaw musculature as guides to the ancestry of salamanders. Zool. J. Linn. Soc. 68:1-40. Dalrymple, G. H., J. E. Juterbock, and A. L. La Valley (1985) Function
4. Terrestrial F e e d i n g in S a l a m a n d e r s of the atlanto-mandibular ligaments of desmognathine salamanders. Copeia 1985:254-257. Dawley, E. M., and A. H. Bass (1989) Chemical access to the vomeronasal organs of a plethodontid salamander. J. Morph. 200: 163-174. Deban, S. M. (1997) Modulation of prey-capture behavior in the plethodontid salamander Ensatina eschscholtzii. J. Exp. Biol. 200: 1951-1964. Deban, S. M., D. B. Wake, and G. Roth 1997. Salamander with a ballistic tongue. Nature 389:27-28. Dockx, P., and F. De Vree (1986) Prey capture and intra-oral food transport in terrestrial salamanders. Stud. Herpetol. 521-524. Driiner, L. (1901) Studien zur Anatomie des Zungebein-, Kiemenbogen- und Kehlkopfmuskeln der Urodelen. I. Theil. Zool. Jahrb. Abteil. Anat. 15:435-622. Findeis, E. K., and W. E. Bemis (1990) Functional morphology of tongue projection in Taricha torosa (Urodela: Salamandridae). Zool. J. Linn. Soc. 99:129-157. Flower, S. S. (1927) Loss of memory accompanying metamorphosis in amphibians. Proc. Zool. Soc. Lond. 1:155-156. Francis, E. B. T. (1934) The Anatomy of the Salamander. Clarendon Press, Oxford. Good, D. A. (1988) Hybridization and cryptic species in Dicamptodon (Caudata: Dicamptodontidae). Evolution 43:728-744. Good, D. A., and D. B. Wake (1992) Geographic variation and speciation in the torrent salamanders of the genus Rhyacotriton (Caudata, Rhyacotritonidae). Univ. Calif. Publ. Zool. 126:1-91. Hairston, N. G., Sr. (1987) Community Ecology and Salamander Guilds. Cambridge Univ. Press, Cambridge. Hiiemae, K. M. (1978) Mammalian mastication: a review of the activity of the jaw muscles and the movements they produce in chewing. Pp. 359-398. In: Develoment, Function and Evolution of Teeth. P. M. Butler and K. A. Joysey (eds.). Academic Press, New York. Himstedt, W. (1982) Prey selection in salamanders. Pp. 47-66. In: Analysis of Visual Behavior. D. Ingle, M. A. Goodale, and R. J. W. Mansfield (eds.). MIT Press, Cambridge, MA. Jaeger, R. G., and D. E. Barnard (1981) Foraging tactics of a terrestrial salamander: choice of diet in structurally simple environments. Am. Nat. 117:639-664. Jaeger, R. G., D. E. Barnard, and R. G. Joseph (1982) Foraging tactics of a terrestrial salamander: assessing prey density. Am. Nat. 119: 885-890. Jaeger, R. G., and A. M. Rubin (1982) Foraging tactics of a terrestrial salamander: judging prey profitability. J. Anim. Ecol. 51:167-176. Keen, W. H. (1979) Feeding and activity patterns in the salamander Desmognathus ochrophaeus (Amphibia, Urodela, Plethodontidae). J. Herp. 13:461-467. Kuzmin, S. L. (1991) Feeding of the salamander Ranodon sibiricus. Alytes 9:135-143. Larsen, J. H., Jr., and J. T. Beneski, Jr. (1988) Quantitative analysis of feeding kinematics in dusky salamanders (Desmognathus). Can. J. Zool. 66:1309-1317. Larsen, J. H., Jr., J. T. Beneski, Jr., and B. T Miller (1996) Structure and function of the hyolingual system in Hynobius and its bearing on the evolution of prey capture in terrestrial salamanders. J. Morph. 227:235-248. Larsen, J. H., Jr., J. T Beneski, Jr., and D. B. Wake (1989) Hyolingual feeding systems of the plethodontidae: comparative kinematics of prey capture by salamanders with free and attached tongues. J. Exp. Zool. 252:25-33. Larsen, J. H., Jr., and D. J. Guthrie (1975) The feeding system of terrestrial tiger salamanders (Amhystoma tigrinum melanostictum Baird). J. Morph. 147:137-154.
115
Larson, A. (1991) A molecular perspective on the evolutionary relationships of the salamander families. Evol. Biol. 25:211-277. Larson, A., and A. C. Wilson (1989) Patterns of ribosomal RNA evolution in salamanders. Mol. Biol. Evol. 6:131-154. Larson, A., and W. W. Dimmick (1993) Phylogenetic relationships of the salamander families: an analysis of congruence among morphological and molecular characters. Herp. Monogr. 7:77-93. Lauder, G. V., and S. M. Reilly (1994) Amphibian feeding behavior: comparative biomechanics and evolution. Adv. Comp. Environ. Physiol. 18:163-195. Lauder, G. V., and H. B. Shaffer (1985) Functional morphology of the feeding mechanism in aquatic ambystomatid salamanders. J. Morph. 185:297-326. Lauder, G. V., and H. B. Shaffer (1988) Ontogeny of functional design in tiger salamanders {Amhystoma tigrinum): are motor patterns conserved during major morphological transformations? J. Morph. 197:249-268. Laurin, M., and R. R. Reisz (1997) A new perspective on tetrapod phylogeny. Pp. 9-59. In: Amniote Origins: Completing the Transition to Land. S. S. Sumida and K. L. M. Martin (eds.). Academic Press, San Diego. Lombard, R. E., and D. B. Wake (1976) Tongue evolution in the lungless salamanders, family Plethodontidae. I. Introduction, theory and a general model of dynamics. J. Morph. 148:265-286. Lombard, R. E., and D. B. Wake (1977) Tongue evolution in the lungless salamanders, family Plethodontidae. II. Function and evolutionary diversity. J. Morph. 153:39-80. Lombard, R. E., and D. B. Wake (1986) Tongue evolution in the lungless salamanders, family Plethodontidae. IV. Phylogeny of plethodontid salamanders and the evolution of feeding dynamics. Syst.Zool.35(4):532-551. Luthardt-Laimer, G. (1983) Distance estimation in binocular and monocular salamanders. Zeit. Tierpsychol. 63:233-240. Luthardt-Laimer, G., and G. Roth (1983) Reduction of visual ir\hibition to stationary prey by early experience in Salamandra salamandra (L.). Z. Tierpsychol. 63:294-302. Lynch, J. F. (1985) The feeding ecology of Aneides flavipunctatus and sympatric plethodontid salamanders in northwestern California. J. Herp. 19:328-352. Maglia, A. M., and R. A. Pyles (1995) Modulation of prey-capture behavior in Plethodon cinereus (Green) (Amphibia: Caudata). J. Exp. Zool. 272:167-183. Maiorana, V. (1978) Behavior of an unobservable species: diet selection by salamander. Copeia 1978(4): 664-672. Miller, B. T, and J. H. Larsen, Jr. (1990) Comparative kinematics of terrestrial prey capture in salamanders and newts (Amphibia: Urodela: Salamandridae). J. Exp. Zool. 256:135-153. Olson, E. C. (1961) Jaw mechanisms: rhipidistians, amphibians, reptiles. Am. Zool. 1:205-215. Ozeti, N., and D. B. Wake (1969) The morphology and evolution of the tongue and associated structures in salamanders and newts (family Salamandridae). Copeia 1969:91-123. Reilly, S. M. (1995) The ontogeny of aquatic feeding behavior in Salamandra salamandra: stereotypy and isometry in feeding kinematics. J. Exp. Biol. 198:701-708. Reilly, S. M., and G. V. Lauder (1988) Atavisms and the homology of hyobranchial elements in lower vertebrates. J. Morph. 195:237245. Reilly, S. M., and G. V. Lauder (1989) Physiological bases of feeding behavior in salamanders: do motor patterns vary with prey type? J. Exp. Biol. 141:343-358. Reilly, S. M., and G. V. Lauder (1990a) The evolution of tetrapod feeding behavior: kinematic homologies in prey transport. Evolution 44:1542-1557.
116
D a v i d B. W a k e a n d S t e p h e n M. D e b a n
Reilly, S. M., and G. V. Lauder (1990b) The strike of the tiger salamander: quantitative electromyography and muscle function during prey capture. J. Comp. Physiol. A 167:827-839. Reilly, S. M., and G. V. Lauder (1991a) Experimental morphology of the feeding mechanism in salamanders. J. Morph. 210:33-44. Reilly, S. M., and G. V. Lauder (1991b) Prey transport in the tiger salamander: quantitative electromyography and muscle function in tetrapods. J. Exp. Zool. 260:1-17. Rose, C. S. (1996) An endocrine-based model for developmental and morphogenetic diversification in metamorphic and paedomorphic urodeles. J. Zool. Lond. 239:253-284. Roth, G. (1976) Experimental analysis of the prey catching behavior of Hydromantes italicus Dunn (Amphibia, Plethodontidae). J. Comp. Physiol. 109:47-58. Roth, G. (1987) Visual Behavior in Salamanders. Springer-Verlag, Heidelberg. Roth, G., K. Nishikawa, U. Dicke, and D. B. Wake (1988) Topography and cytoarchitecture of the motor nuclei in the brainstem of salamanders. J. Comp. Neurol. 278:181-194. Roth, G., K. C. Nishikawa, D. B. Wake, U. Dicke, and T. Matsushima (1990) Mechanics and neuromorphology of feeding in amphibians. Neth. J. Zool. 40:115-135. Roth, G., and D. B. Wake (1985a) The structure of the brainstem and cervical spinal cord in relation to feeding of lungless salamanders, family Plethodontidae. J. Comp. Neurol. 241:99-110. Roth, G., and D. B. Wake (1985b) Trends in the functional morphology and sensorimotor control of feeding behavior in salamanders: an example of the role of internal dynamics in evolution. Acta Biotheor. 34:175-192. Roth, G., and D. B. Wake (1989) Conservatism and innovation in the evolution of feeding in vertebrates. Pp. 7-21. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, Chichester. Schwenk, K., and D. B. Wake (1993) Prey processing in Leurognathus marmoratus and the evolution of form and function in desmognathine salamanders (Plethodontidae). Biol. J. Linn. Soc. 49:141162. Severtsov, A. S. (1971) The mechanism of food capture in tailed amphibians. Doklady Akademii Nauk SSSR 197:728-731. Sites, J. W. (1978) The foraging strategy of the dusky salamander, Desmognathus fuscus (Amphibia, Urodela, Plethodontidae): an empirical approach to predatiion theory. J. Herp. 12:373-383. Shaffer, H. B. and G. V. Lauder (1988) The ontogeny of functional design: metamorphosis of feeding behaviour in the tiger salamander (Amhystoma tigrinum). J. Zool. London 216:437-454. Shaffer, H. B., J. M. Clark, and R Kraus (1991) When molecules and morphology clash: a phylogenetic analysis of the North American
ambystomatid (Caudata: Ambystomatidae) salamanders. Syst. Zool. 40:284-303. Thexton, A. J., D. B. Wake, and M. H. Wake (1977) Tongue function in the salamander Bolitoglossa occidentalis. Arch. Oral Biol. 22: 361-366. Thiesmeier, B., and C. Romberg (1997) Paarung, Fortpflanzung und Larvalentwicklung von Pachytriton sp. {Pachytriton A) nebst Bemerkungen zur Taxonomie der Gattung. Salamandra 33:97-110. Titus, T. A., and A. Larson (1995) A molecular phylogenetic perspective on the evolutionary radiation of the salamander family Salamandridae. Syst. Biol. 44:125-151. Trueb, L., and R. Cloutier (1991) A phylogenetic investigation of the inter- and intrarelationships of the Lissamphibia (Amphibia: Temnospondyli). Pp. 223-313. In: Origins of the Higher Groups of Tetrapods: Controversy and Consensus. H.-P. Schultze and L. Trueb (eds.). Comstock Publ. Assoc, Ithaca. Wake, D. B. (1963) Comparative osteology of the plethodontid salamander genus Aneides. J. Morph. 113:77-118. Wake, D. B. (1966) Comparative osteology and evolution of the lungless salamanders, family Plethodontidae. Mem. Southern Calif. Acad. Sci. 4:1-111. Wake, D. B. (1982) Functional and developmental constraints and opportunities in the evolution of feeding systems in urodeles. Pp. 51-66. In: Environmental Adaptation and Evolution. D. Mossakowski and G. Roth (eds.). Gustav Fischer, Stuttgart. Wake, D. B. (1991) Homoplasy: the result of natural selection, or evidence of design limitations? Am. Nat. 138(3): 543-567. Wake, D.B. (1993) Brainstem organization and branchiomeric nerves. Acta Anat. 148:124-131. Wake, D. B., and A. Larson (1987) Multidimensional analysis of an evolving lineage. Science 238:42-48. Wake, D. B., and J. Hanken (1996) Direct development in the lungless salamanders: what are the consequences for developmental biology, evolution and phylogenesis? Int. J. Dev. Biol. 40:859-869. Wake, D. B., K. C. Nishikawa, U. Dicke, and G. Roth (1988) Organization of the motor nuclei in the cervical spinal cord of salamanders. J. Comp. Neurol. 278:195-208. Wiggers, W., and G. Roth (1991) Anatomy, neurophysiology and functional aspects of the nucleus isthmi in salamanders of the family Plethodontidae. J. Comp. Physiol. A 169:165-176. Wolterstorff, W., and W. Herre (1935) Die Gattungen der Wassermolche der Familie Salamandridae. Arch. Naturgesch. N.F. 4: 217-229. Zhao, E., and Q. Hu (1988) Studies on Chinese tailed amphibians. English translation. Pp. 1-44. In: Studies on Chinese Salamanders. E. Shao, Q. Hu, Y. Jiang, and Y. Yang (eds.). Society for the Study of Amphibians and Reptiles, Miami, OH.
C H A P T E R
5 Feeding in Frogs KIISA C NISHIKAWA Department of Biological Sciences Northern Arizona University Flagstaff, Arizona 86011
animals. One aspect of this goal is to compare the function of the feeding apparatus across the major clades of tetrapod vertebrates. For each clade, the function of the feeding apparatus can be addressed by asking the following questions: (1) Which muscles are involved in producing feeding movements and v^hat is the specific contribution of each? (2) What are the spatial and temporal patterns of muscle activation and how do they relate to movement? and (3) What are the neural mechanisms that are responsible for producing observed patterns of muscle activity? Since the early 1990s, my research has focused on understanding the evolutionary relationships among morphology, biomechanics, and neural control of movement using prey capture of anurans as a model system. This chapter provides a summary of this work and a historical perspective on hypotheses concerning the mechanisms of anuran prey capture. The chapter begins by briefly reviewing the phylogeny and natural history of anurans (Section I). Subsequent sections describe the morphology (Section II) and function (Section III) of the anuran feeding apparatus, the neural control of prey capture movements (Section IV), and the evolution of mechanisms of tongue protraction and neural control of prey capture (Section V). The chapter ends with a summary of conclusions and a description of current and future directions. It is hoped that by describing the pitfalls that have been encountered in attempting to understand the mechanisms of prey capture in frogs, this chapter will help future workers avoid similar problems in the future. It is also hoped that our attempts to understand the biomechanics and neural control of prey capture in frogs will stimulate functional morphologists to undertake similar studies
I. INTRODUCTION A. Phylogeny of Anurans B. Natural History of Anurans II MORPHOLOGY OF THE FEEDING APPARATUS A. Craniun\ and Jaw Muscles B. Mandible and Buccal Floor Muscles C. Hyoid and Associated Muscles D.Tongue and Associated Muscles III. FUNCTION OF THE FEEDING APPARATUS A. Methods for Studying the Function of the Feeding Apparatus B. Hypotheses for the Mechanism of Tongue Protraction C. Functional Diversification among Anuran Taxa IV. NEURAL CONTROL OF PREY CAPTURE A. Visual Analysis of Prey Features B. Role of Tongue Afferents C. Interactions between Tongue Afferents and Visual Input V. EVOLUTION OF THE FEEDING APPARATUS A. Evolutionary Transitions in Mechanisms of Tongue Protraction B. Morphological Correlates of Tongue Protraction Mechanisms C. Evolution of Tongue Afferents D. Evolutionary Transitions in Mechanisms of Neural Control VI. CONCLUSIONS VII. CURRENT AND FUTURE DIRECTIONS References
I. INTRODUCTION A major goal of functional morphology is to understand the evolution of functional differences among
FEEDING (K.SchwenKed.)
117
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
118
Kiisa C. N i s h i k a w a
in other groups of animals so that the generality of our results can be tested in other groups.
Ascaphidae Leiopelmatidae Bombinatoridae
A. Phylogeny of Anurans On superficial examination, frogs appear to be a morphologically homogeneous group. They are easily distinguished from other amphibians by their welldeveloped hind legs and the absence of a tail in adults. Frogs are, however, quite diverse morphologically, particularly in terms of their feeding apparatus, which has played an important role in anuran classification schemes for more than a century (Giinther, 1859; Griffiths, 1963). Recent workers recognize 21-27 families of frogs (Duellman and Trueb, 1986; Ford and Cannatella, 1993). A number of phylogenetic hypotheses have been published for the anuran families, based on morphological (e.g.. Ford and Cannatella, 1993) or genetic characters (e.g., Hillis et al, 1993; Hay et al, 1995). In these studies, the relationships of some anuran families are agreed upon generally, but relationships of the majority of families remain to be resolved definitively (Hillis efflZ., 1993). This chapter uses a recent phylogenetic hypothesis of anuran familial relationships (Fig. 5.1) that was developed using morphological characters by Ford and Cannatella (1993). In this hypothesis, the "archaeobatrachian" families (Ascaphidae, Leiopelmatidae, Bombinatoridae, and Discoglossidae) represent the most basal lineages of frogs. These families share many ancestral morphological characteristics, but are not closely related to each other and do not constitute a monophyletic group (Ford and Cannatella, 1993). On the basis of genetic characters, however. Hay et ah (1995) concluded that the Archaeobatrachia may, in fact, be monophyletic. In Ford and Cannatella's (1993) phylogeny, the Mesobatrachia (Fig. 5.1) is a monophyletic group that includes two lineages: the pelobatoids (families Pelobatidae, Pelodytidae, and Megophryidae) and pipoids (families Pipidae and Rhinophrynidae). The remaining families are placed in the clade Neobatrachia. Within the Neobatrachia, the families Hylidae, Centrolenidae, and Pseudidae appear to form a clade, as do the ranoids (families Ranidae, Arthroleptidae, Hyperoliidae, Rhacophoridae, Dendrobatidae, Hemisotidae, and Microhylidae), although the relationships among ranoid families are unresolved. The placement of the dendrobatids remains controversial and some authors believe that they are more closely related to bufonids than to ranoids (Hass, 1995; Hay et ah, 1995; Ruvinsky and Maxson, 1996). The relationships among the other neobatrachian families are also unresolved (Fig. 5.1). Most of the currently recognized anuran families
Pelobatidae Pelodytidae Megophryidae Rhinophrynidae Pipidae
_
Limnodynastinae Myobatrachlnae Sooglossidae Heleophrynidae "Leptodactylidae" Brachycephalidae Bufonidae Rhinodermatidae Centrolenidae Hylidae Pseudidae "Ranidae" Arthroleptidae Hyperoliidae Rhacophoridae Dendrobatidae Hemisotidae Microhylidae
F I G U R E 5.1. A phylogeny of the living anurans modified from Ford and Cannatella (1993). The archaeobatrachians are a grade group that share many ancestral characteristics but are not closely related to each other. The families Leptodactylidae and Ranidae are indicated in quotes because there is no evidence for their monophyly.
appear to be monophyletic, with the possible exception of Myobatrachidae, Leptodactylidae, and Ranidae (Ford and Cannatella, 1993). The family Myobatrachidae may not be monophyletic because its subfamily Myobatrachlnae shares several characteristics with the family Sooglossidae that are absent in the subfamily Limnodynastinae. For the families Ranidae and Leptodactylidae, monophyly has been rejected, but the relationships of the subfamilies to other taxa remain to be resolved (Ford and Cannatella, 1993). At present, the unresolved relationships among neobatrachian frogs are a serious impediment to understanding anuran evolution. B. Natural History of Anurans Nearly all frogs are carnivorous after metamorphosis, and invertebrates are the most common prey, although there are reports of frogs that eat fruits (Silva et ah, 1989) and leaves (Das and Coe, 1994; Das, 1995, 1996). Most anurans are also feeding generalists, eating any prey that will fit into their mouths in proportions that match the relative abundance of prey in
119
5. Feeding in Frogs nature (Toft, 1981). Dietary generalists tend to be widemouthed, sit-and-wait predators. A few species of anurans are known to specialize on a narrower range of prey types (Toft, 1981). Some forest floor species, including the genus Bufo (Smith and Bragg, 1949; Inger and Marx, 1961; Berry and Bullock, 1962; Clarke, 1974; Zug and Zug, 1979; Toft, 1981), and many burrowing anurans, such as Hemisus (Emerson, 1976a), specialize on small prey such as ants, coUembolans, termites, or mites (Simon and Toft, 1991; Toft, 1995; Caldwell, 1996), consuming them in greater proportions than their relative abundance in nature. Small prey specialists tend to lack teeth, have relatively narrow mouths, and search actively for prey (Toft, 1981). They also tend to have higher aerobic capacities and lower anaerobic capacities than dietary generalists (Taigen and Pough, 1983). Several anurans are known to consume rather peculiar types of prey, such as snails (Drewes and Roth, 1981), crabs (Premo and Atmowidjojo, 1987), fish (Knoepffler, 1976), or even other frogs (Emerson, 1985). However, these species also take a wide range of other prey types and so are not strict dietary specialists. There have been relatively few attempts to correlate the morphology of the feeding apparatus with natural diets in anurans. Emerson (1985) and colleagues (Emerson and Voris, 1992, Emerson and Bramble, 1993; Emerson et ah, 1994) have studied the relationships among skull morphology, feeding behavior, and diet among diverse species. She found that frogs that eat relatively large prey have relatively longer jaws and wider skulls than those that eat small prey. In addition, the feeding cycle is longer and more asymmetrical (i.e., mouth closing takes longer than mouth opening) in species that eat large prey than in species that eat small prey. Gray (1997) compared the diets of sympatric hylid frogs with short vs long tongues. She found that there was overlap in the types of prey taken by the two species, but long-tongued species consumed a greater proportion of rapidly moving prey than shorttongued species. The ecological significance of differences among frogs in the morphology of the feeding apparatus is a topic that deserves further study.
II. MORPHOLOGY OF THE FEEDING APPARATUS Feeding mechanisms of amphibians (frogs, salamanders, and caecilians) are extremely diverse. The tongues of caecilians are small and cannot be protruded from the mouth so jaw prehension is used to capture prey and some species possess a highly specialized jawclosing mechanism (Bemis et ah, 1983; Nussbaum,
1983). In contrast to caecilians, terrestrial salamanders and frogs use lingual protraction to catch prey. They depend heavily on lingual adhesion for prey capture. In salamanders, the articulated hyobranchial skeleton forms an internal support for the tongue, which leaves the mouth with the tongue during protraction (Lombard and Wake, 1976; Thexton et ah, 1977). Salamanders vary widely in their tongue morphology, especially with regard to projection distance, muscle attachment, and hyobranchial folding (Ozeti and Wake, 1969; Regal, 1966; Wake, 1982; Lombard and Wake, 1977, 1987). In contrast to salamanders, the anuran tongue consists only of muscles and connective tissue, with no intrinsic cartilaginous or bony skeleton. The hyobranchial apparatus is an unarticulated plate formed from the ontogenetic fusion of the cartilaginous larval elements. It cannot be folded and does not leave the mouth with the tongue during feeding (Gans, 1961, 1962; Gans and Gorniak, 1982a,b; Roth et ah, 1990). Frogs also vary in their tongue morphology, especially with regard to the anatomy of the protractor muscle (Regal and Gans, 1976; Horton, 1982) and projection distance (Deban and Nishikawa, 1992; Nishikawa et ah, 1992). Variation in the morphology of the feeding apparatus has been studied extensively in frogs (MagimelPelonnier, 1924; Regal and Gans, 1976; Emerson, 1976b, 1985; Horton, 1982; Cannatella, 1985). Until recently, there have been few attempts to map morphological characteristics of the feeding apparatus onto a cladogram of anuran taxa, particularly in a functional context. In general, although the size and shape of elements of the feeding apparatus may differ dramatically among species, there is a great deal of overall similarity in the bones and muscles of the feeding apparatus among frogs (Figs. 5.2A-5.2C). The functional significance of anatomical variation remains unknown in most cases. The feeding apparatus of frogs consists of the cranium, mandibles, hyoid, and tongue, each of which is associated with a series of muscles. The morphology of each of these elements is described next, along with brief descriptions of hypothesized functions of the musculoskeletal elements and a brief review of diversity among taxa, where known. A more detailed examination of the function of the feeding apparatus is given in Section III. A. Cranium and Jaw Muscles The morphology of the anuran cranium was studied extensively by Trueb (1973), and a fairly detailed summary can be found in Duellman and Trueb (1986). Cranial morphology is highly variable among species.
120
Kiisa C. N i s h i k a w a Hyla cinerea
Hemisus marmoratum
Bufo marinus M. submentalis
M. intermandibularis
geniohyoideus
M. intermandibularis
M. geniohyoideus
hyoid M. interhyoideus
M. intermandibularis
l\^. submentalis M. geniohyoideus
^ interhyoideus
M. intertiyoideus
M. omohyoideus
M. hyoglossus M. stemohyoideus
M. genioglossus
M. genioglossus
M. geniohyoideus medialis M. geniohyoideus lateralis
M. hyoglossus
M. hyoglossus M. hyoglossus
M. omohyoideusM. stemohyoideus
M. stemohyoideus
tongue tip " I ^ M. hyoglossus \ lingual sinus
,„„„«oo mucosa I
M. hyoglossus "^^X^iJ"^^
^- hyoglossus
M. genioglossus
u inte,h„«iHai,c / // M. intertiyoideus / (/ . ^ ._, M. submentalis ^ M. geniohyoideus
mentomeckelian \io\)Q%
hyoid plate
F I G U R E 5.2. Musculature of the feeding apparatus in Hyla cinerea (A), Bufo marinus (B), and Hemisus marmoratum (C). Superficial muscles of the feeding apparatus include M. intermandibularis, M. interhyoideus, and M. submentalis. Deep muscles include the extrinsic tongue muscles M. hyoglossus and M. genioglossus, and the hyoid protractor M. geniohyoideus and retractor M. stemohyoideus. A sagittal section of the tongue of Hemisus (D) shows the dorsoventral and longitudinal compartments of the M. genioglossus.
which exhibit a continuum from reduced ossification (e.g., Ascaphus) to hyperossification (e.g., pipid frogs). Reduction or loss of bones has occurred in numerous taxa, and a few taxa possess neomorphic cranial elements (Trueb, 1973). Bones of the neurocranium include the sphenethmoid and the paired prootics and exoccipitals. The dermal roofing bones usually include only the nasals and the frontoparietals. The frontoparietals and prootics provide attachment sites for the M. levator mandibulae anterior longus and M. levator mandibulae posterior longus (= M. temporalis). The bones of the palate include the paired vomers, which are often absent, and the parasphenoid, which is always present and covers the neurocranium ventrally. The prevomers, palatines, and pterygoids are absent in many taxa. Although quite variable in shape, the squamosal is always present. It acts as an important attachment site for the jaw levators. Maxillary arch bones always include the premaxilla and maxilla, and sometimes the quadratojugal. The premaxilla is highly
variable and is important in determining the shape of the snout (Trueb, 1973). The teeth of frogs are reduced compared to most other vertebrates (Duellman and Trueb, 1986). Maxillary and premaxillary teeth are usually, but not always, present. Vomerine teeth are usually present if the vomers are present. No other bones of the upper jaw bear teeth in frogs. Small prey specialists, such as toads (family Bufonidae), lack teeth entirely (Duellman and Trueb, 1986). The jaw muscles consist of one main depressor for opening the mouth and a complex of six levators (= adductors) for closing it. The M. depressor mandibulae originates from the dorsal fascia and/or otic capsule and inserts on the proximal tip of the mandible (Emerson, 1977), close to the jaw joint and therefore in a position of relatively low mechanical advantage. The muscles of the M. levator mandibulae complex extend from the otic capsules and squamosal to the mandible as follows: (1) the M. levator mandibulae anterior longus (= M. pterygoideus) originates on the frontoparietal
121
5. Feeding in Frogs and prootic and inserts via a tendon on the medial side of the angulosplenial; (2) the massive M. levator mandibulae posterior longus (= M. temporalis) originates from the median raphe on the skull roof, the lateral surface of the frontoparietal, and the dorsal surface of the prootic and it inserts via a tendon on the medial side of the angulosplenial; (3) the M. 1. m. posterior lateralis originates on the ventral arm of the squamosal and inserts on Meckel's cartilage and the lateral surface of the angulosplenial; (4) the M. 1. m. posterior articularis originates on the quadrate and inserts on the mandible; (5) the M. 1. m. externus originates on the zygomatic process of the squamosal and inserts laterally on the mandible; and (6) the M. 1. m. posterior subexternus also originates on the zygomatic process of the squamosal and inserts on the posterior end of the mandible (Duellman and Trueb, 1986). All of the Mm. levator mandibulae insert farther from the jaw joint, in a position of relatively greater mechanical advantage, than the M. depressor mandibulae. The Mm. levator
Hyla cinerea
mandibulae are innervated by the trigeminal nerve, whereas the M. depressor mandibulae is innervated by the facial nerve (Figs. 5.3D-5.3F). Although there is much variation in the morphology of the skull and jaws of frogs (Trueb, 1973) and considerable variation among species in the jaw musculature (Starrett, 1968), the functional significance of interspecific differences remains largely unknown (Duellman and Trueb, 1986). James Birch (personal communication) is using morphometric methods to analyze the functional consequences of changes in anuran skull shape during development as well as differences among species. B. Mandible and Buccal Floor Muscles The mandibles of frogs generally consist of three paired bony elements associated with Meckel's cartilage: the angulosplenial, dentary, and mentomeckelian bones (which are absent in pipoids and a few other
Bufo marinus
Hemisus marmoratum
F I G U R E 5.3. Camera lucida drawings of the peripheral nerves of adult Hyla cinerea (A,D), Bufo marinus (B,E), and Hemisus marmoratum (C,F) stained with Sudan black B. (Top row) The glossopharyngeal nerve is shown on the left and the hypoglossal nerve is shown on the right. (Bottom row) The trigeminal nerve is shown on the left and the facial nerve is shown on the right. The stippled area indicates the tongue pad. The glossopharyngeal nerve provides only sensory innervation of the tongue. The hypoglossal nerve innervates the Mm. genioglossus basalis and medialis (if present), the M. hyoglossus, and the M.m. geniohyoideus, sternohyoideus, and omohyoideus. The M. genioglossus basalis is innervated by the most proximal branches of the hypoglossal nerve at the base of the tongue, whereas the M. genioglossus medialis is innervated by the more distal branches. The trigeminal nerve crosses over the mandible and innervates the M. intermandibularis and M. submentalis. At the base of the mandible, the facial nerve innervates the M. interhyoideus.
122
Kiisa C. Nishikawa
species). The presence of a movable joint between the mentomeckelian and dentary is an unusual feature of the mandibles of most anurans (Regal and Gans, 1976; Nishikawa and Roth, 1991). In most species, depression of the mandibular tips results not only from downward movement of the mandibles relative to the cranium (mandibular depression), but also from downward movement of the mentomeckelian bones relative to the rest of the mandible (mandibular bending). Most frogs lack teeth on the mandible. Only one species (Amphignathodon guentheri) is known to possess mandibular teeth (Duellman and Trueb, 1986), although some species (e.g., Ceratophrys) possess tooth-like processes on the dentary. A series of three transversely oriented muscles form the floor of the buccal cavity: the M. submentalis, M. intermandibularis, and M. interhyoideus (Figs. 5.2A5.2C). The M. submentalis connects the anterior ends of the mandibles. During feeding, it bends the mandibles downward by depressing their tips. During breathing, the M. submentalis closes the nares by lifting the mentomeckelian bones upward. This upward movement deforms the alary cartilages, which closes the nares (Gans and Pyles, 1983; Nishikawa and Gans, 1996). The M. intermandibularis extends between the posterior ends of the mandibles (Figs. 5.2A-5.2C). In some species, the M. intermandibularis has supplementary elements of unknown function (Tyler, 1974; Emerson, 1976b). The M. interhyoideus lies posterior to the M. intermandibularis and supports the vocal sacs (Figs. 5.2A-5.2C). The Mm. submentalis and intermandibularis are innervated by the trigeminal nerve, whereas the M. interhyoideus is innervated by the facial nerve (Figs. 5.3D-5.3F). Contraction of the M. intermandibularis and M. interhyoideus raises the buccal floor (Gans and Gorniak, 1982b). C. H y o i d and Associated Muscles The hyobranchial apparatus of adult anurans is a broad plate that forms from the fusion of larval branchial elements (De Jongh, 1968). The slender hyalia extend from the anterior end of the plate and usually attach to the prootic bones, although the hyalia are disjunct in most mesobatrachians and are missing altogether in pelodytids (Cannatella, 1985). Alary processes and posterolateral processes are usually present on the lateral margins of the hyoid plate, and posteromedial processes flank the larynx. In most anurans, only the posteromedial processes are ossified. A pair of muscles, the M. geniohyoideus and M. sternohyoideus, serve to protract and retract the hyoid, respectively (Emerson, 1977; Gans and Gorniak, 1982b). The M. geniohyoideus originates on the pos-
terolateral processes of the hyoid and inserts near the mandibular symphysis (Figs. 5.2A-5.2C). In most species, it consists of separate medial and lateral compartments. The M. sternohyoideus originates on the sternum and inserts on the posterolateral edge of the hyoid plate (Figs. 5.2D-5.2F). The Mm. petrohyoidei anterior et posteriores and the M. omohyoideus elevate and depress the hyoid, respectively (de Jongh and Gans, 1969; Emerson, 1977). The glossopharyngeal nerve innervates the M. petrohyoideus anterior and the vagus nerve innervates the Mm. petrohyoidei posteriores, whereas the hypoglossal nerve innervates the Mm. omohyoideus, sternohyoideus and geniohyoideus (Figs. 5.3A-5.3C; Gaupp, 1896). A great deal of variation in hyoid morphology is found among anurans (Duellman and Trueb, 1986). Aquatic suction feeders (e.g., Hymenochirus, family Pipidae) possess large, articulated, ossified hyoids, whereas terrestrial lingual feeders possess small, fused, cartilaginous hyoids. A similar pattern is found among aquatic and terrestrial turtles, in which suction feeders possess large, ossified, articulated hyoids whereas terrestrial feeders possess small, cartilaginous ones (Bramble and Wake, 1985). An ossified and articulated hyoid appears to be associated with the ability to generate large forces, such as those necessary for moving large volumes of water during aquatic suction feeding. The role of the hyoid in lingual feeding varies widely among tetrapods. In salamanders, it forms a skeletal support for the tongue, and both tongue and hyoid are projected from the mouth as a unit (Lombard and Wake, 1977). In chameleons, the hyoid is a tapered rod around which an accelerator muscle contracts to project the tongue from the mouth (Gans, 1967; Wainwright et ah, 1991). In frogs, the hyoid forms a base on which the tongue rests. The hyoid does not leave the mouth with the tongue. The only connection between hyoid and tongue is the M. hyoglossus, which originates on the posterolateral process of the hyoid and inserts broadly in the tongue. Movements of the hyoid plate appear to play an important role in buccal expansion and contraction during breathing and calling in anurans (de Jongh and Gans, 1969; Martin and Gans, 1972; Emerson, 1977). The role of the hyoid in anuran feeding is less clear (see Section III,B). Cineradiographic recordings of hyoid movement during feeding in Bufo marinus show that the hyoid is stabilized in a retracted position during the initial phase of tongue protraction and that it moves anteriorly during tongue protraction (Emerson, 1977). Based on these observations, Emerson (1977) hypothesized that the hyoid acts as a stable platform for the tongue during the initial stages of protraction, that it stores potential energy during intermediate stages.
5. Feeding in Frogs and that the stored energy is imparted to the tongue during the final stages of protraction. However, this mechanism appears to be unlikely (see Section III,B). In contrast to the neobatrachians that have been studied, the hyoid appears to play a more important role in feeding in mesobatrachians. Based on anatomical observations and muscle stimulation experiments, Trueb and Gans (1983) suggested that the hyoid plays an important role in tongue protraction during feeding in the termite-eating frog, Rhinophrynus dorsalis. This hypothesis remains to be tested experimentally in feeding animals. In the spadefoot toad {Spea multiplicata), another mesobatrachian, tongue movements were impaired after bilateral denervation of the M. geniohyoideus, suggesting that hyoid protraction is necessary for normal tongue protraction in this species (O'Reilly and Nishikawa, 1995). The disjunct hyoid of mesobatrachians may be responsible for the greater role of hyoid protraction during feeding in this group. In most frogs, the cornua of the hyoid are fused to the prootic bones, which may limit forward excursion of the hyoid (Cannatella, 1985). In mesobatrachians, the cornua are continuous until metamorphosis, at which time a gap develops in the cornua, which frees the hyoid plate from its attachment to the skull (Ridewood, 1897), perhaps allowing the hyoid to move farther anteriorly during feeding than it can in other frogs. D . Tongue and Associated Muscles There are several problems that terrestrial frogs must overcome to capture prey successfully. These include contacting the prey with the tongue, ingesting prey, transporting it through the oral cavity, and finally swallowing it. In terrestrial anurans, the sticky tongue plays an important role in ingesting, transporting, and swallowing prey. There are no known instances of inertial feeding or transport among anurans, probably because they consume mostly small prey. Frogs do not masticate their food, and there is relatively little manipulation of the food by the tongue once it is in the oral cavity. Little is known about differences among anuran species in modes of oral transport and swallowing. All anurans possess a relatively simple tongue that consists only of two pairs of extrinsic muscles: the M. genioglossus and M. hyoglossus (Gaupp, 1901; Regal and Gans, 1976). In contrast to most other terrestrial vertebrates, intrinsic muscles are absent in most species. The M. genioglossus originates near the mandibular symphysis and inserts posteriorly into the tongue pad (Figs. 5.2A-5.2C). In many anurans, the M. genioglossus is subdivided by connective tissue
123
into a number of different compartments, which vary widely among species (Horton, 1982). The M. hyoglossus originates on the posteromedial process of the hyoid and inserts along the lateral margin of the tongue pad, often interdigitating with the fibers of the M. genioglossus (Figs. 5.2A-5.2C). The Mm. genioglossus and hyoglossus are innervated by the hypoglossal nerve (Figs. 5.3A-5.3C). In most frogs, the M. genioglossus is used to place the tongue on the prey (discussed in detail in Section III) and prey are returned to the mouth by the M. hyoglossus. In Bombina, Bufo, Phrynomerus, and Hemisus, denervation experiments demonstrate that the M. hyoglossus plays an important role in prey capture, oral transport, and swallowing (Ritter and Nishikawa, 1995; Tso et ah, 1995). Small prey are returned to the mouth and delivered to the esophagus by the tongue in a single movement, sometimes with the help of the forelimbs, whereas large prey are nearly always transported with the help of the forelimbs, presumably because lingual transport would be ineffective by itself (Gray et ah, 1997). Because most anurans retract the eyes into the orbit during swallowing, it has been suggested that the M. retractor bulbi also plays a role in anuran swallowing (Regal and Gans, 1976). However, this hypothesis has yet to be tested experimentally (Duellman and Trueb, 1986). There have been three broadly comparative morphological studies of the tongue musculature of anurans. Magimel-Pelonnier (1924) studied 45 species. Regal and Gans (1976) studied 12 species, and Horton (1982) studied 63 species, predominantly from Australia. In all, 61 genera are represented. From these studies, three different patterns of tongue morphology have been described. The first pattern consists of a round tongue that is broadly attached to the floor of the mouth, both anteriorly and posteriorly, so that there is no free flap posteriorly. In these tongues, the fibers of the M. hyoglossus radiate from the hyoid into the tongue pad and the ventralmost fascicles insert near the base of the tongue (Horton, 1982). The fibers of the M. genioglossus arise near the niandibular symphysis and radiate into the tongue, where they interdigitate with those of the M. hyoglossus (Regal and Gans, 1976; Horton, 1982). This morphology is found in all archaeobatrachians (i.e., Ascaphus, Leiopelma, Bombina, Alytes, and Discoglossus), some mesobatrachians {Pelobates, Pelodytes, and Rhinophrynus), and some neobatrachians, including Telmatobius, Litoria, Hyla, Gastrotheca, Rheobatrachus, and several genera of limnodynastines (Magimel-Pelonnier, 1924; Regal and Gans, 1976; Horton, 1982; Trueb and Gans, 1983). Based on their origins and insertions. Regal and
124
Kiisa C. Nishikawa
Gans (1976) hypothesized that, upon contraction, the M. genioglossus pulls the tongue toward the symphysis in these species, whereas the M. hyoglossus pulls the tongue toward the esophagus. These hypotheses have been confirmed by denervation of the M. genioglossus in Bombina (Nishikawa et ah, 1992), Discoglossus (Nishikawa and Roth, 1991), and Hyla (Deban and Nishikawa, 1992) and by denervation of the M. hyoglossus in Bombina (Tso et ah, 1995). The second pattern is found only among aquatic pipid frogs (e.g., Xenopus, Pipa, and Hymenochirus), which possess large articulated, ossified hyoids (Cannatella, 1985) and are secondarily tongueless (Cannatella and Trueb, 1988) although they possess vestiges of tongue musculature (Horton, 1982). These species use suction or ram feeding to ingest and transport prey (Sokol, 1969; Avila and Frye, 1977; O'Reilly et al, 1999). A third pattern consists of a muscular tongue that is attached anteriorly to the buccal floor, with a free posterior flap that varies in length among species (Regal and Gans, 1976) and varies to some extent among fixed specimens within a species (Magimel-Pelonnier, 1924). This type of tongue is found in some pelobatoids {Scaphiopus, Spea, and Megophrys) and some neobatrachians, such as Hyla and Limnodynastes (Magimel-Pelonnier, 1924; Regal and Gans, 1976; Horton, 1982). In these tongues, the fibers of the M. hyoglossus recurve to insert in the more distal parts of the tongue, rather than its anterior base. In contrast to the M. hyoglossus, which is relatively homogeneous among species, the arrangement of the M. genioglossus varies widely (Magimel-Pelonnier, 1924; Regal and Gans, 1976; Horton, 1982). All species retain the interdigitating element that is found in species with round, broadly attached tongues, but several additional elements also may be present (Regal and Gans, 1976; Horton, 1982). These elements fall into two groups: those with fibers that run parallel to the long axis of the tongue [i.e., the ventral, dorsomedial and superficial elements of Horton (1982)] and those with transverse fibers [i.e., the genioglossus basalis of Gaupp (1901)]. At present, there is no information concerning the functional significance of this variation in the arrangement of the M. genioglossus (but see Section III,C). As in species with round, broadly attached tongues, retraction of the tongue is accomplished by contraction of the M. hyoglossus (Gans and Gomiak, 1982a,b; Ritter and Nishikawa, 1995; Tso et al, 1995). On the basis of their anatomy, the tongues of species with free posterior flaps are presumed to rotate over the mandibular symphysis so that the dorsal surface of the tongue at rest becomes the ventral surface of the fully protracted tongue, as has been shown to occur in Rana (Gans, 1961, 1962) and Bufo (Gans and Gorniak,
1982a,b). It has been hypothesized that these "flipping tongues" have evolved convergently several times among frogs (Regal and Gans, 1976), and this idea is supported by phylogenetic analysis (Section V,A). Despite the variation in tongue morphology, most frogs (with the exception of pipids) share several important features of the tongue: (1) it is attached anteriorly near the mandibular symphysis; (2) most of the fibers in both the protractor and the retractor muscles are oriented nearly parallel to the long axis of the tongue so that their shortening would pull the tongue pad either toward the symphysis or toward the esophagus (Horton, 1982); (3) those fibers that are transverse (i.e., genioglossus basalis) are relatively short and are associated with large amounts of connective tissue (Horton, 1982); (4) the resting length of the tongue approximates the length of the mandibles; and (5) the mass of the tongue is approximately 0.5-1.0% of body mass, which is about as large in relative terms as the human heart.
III. FUNCTION OF THE FEEDING APPARATUS This section first gives an overview of methods that can be used to study the function of any morphological system, in this case the feeding apparatus. It next discusses previously published hypotheses for the mechanism of tongue protraction in anurans, including an analysis of which methods were informative, which were uninformative, and which were positively misleading. The section ends with a discussion of functional diversification of the mechanism of tongue protraction among anuran species. A. Methods for Studying the Function of the Feeding Apparatus Before undertaking a comparative and functional analysis of feeding, it is important to ask what types of experiments and observations can best be used to understand the function of the feeding apparatus. Obviously, the necessary data include a description of anatomy as well as a description of movement patterns, which usually involves some type of kinematic analysis. These often are supplemented with information about muscular and neural activity. Traditional methods include recordings of electrical activity from relevant muscles and nerves as well as electrical stimulation experiments. Kinematic studies are of three types: (1) description of films or still photos; (2) analysis of kinematic profiles in which values of kinematic variables, such as gape angle, are plotted over time (Fig. 5.4); and (3) trajectory
90 80
Hyla cinerea
Hemisus marmoratum
Bufo marinus
701
c
< Q. CO
o
60 50 40 30 20 10 0 -10 5.0
4.0 3.0 2.0 1.0 0.0 0> Q. CO
o
-1.0 -2.0 -3.0 200
M^
1901
c
<
180 170 160
O
-o
150 140 40 30 20 10 0 -10 -20 j -30 -40 -50 2.01
\J
1.5
c
.o
1.0 0.5
0.01 -0.5 -1.0 5.0 4.0 3.0
c a>
2.0 1.0
0.0 I -1.0 1.0 0.0 i
-1.0 -2.0
-3.01 -4.0 f — -200 -100
0
100 200
300
-200 -100
0
100 200
300
-200 -100
0
100 200 300
Time (ms) F I G U R E 5.4. Kinematic profiles for Hyla cinerea, Bufo marinus, and Hemisus marmoratum showing gape angle (°), gape velocity (°/msec), mandible angle (°), head rotation (°), tongue length (cm), lunge length (cm), and jaw position (cm) as a function of time. Mandible angle is the angle subtended by the jaw joint and the tip of the mandibles, with the midpoint of the mandible at the vertex. It is approximately 180° at rest and bends downward to approximately 140° under the contraction of the M. submentalis. Head rotation is the angle formed between a line connecting the jaw joint and upper jaw tips and the horizon line and measures the change in head position relative to the horizon. In this sequence, H cinerea uses jaw prehension to capture the prey, and the head is rotated downward to an angle of 45° below the horizon. Lunge length is the horizontal position of the upper jaw tip in each field. Jaw position shows the vertical position of the upper and lower jaw tips in each field.
126
Kiisa C. N i s h i k a w a
analysis, in which the positions of elements of the feeding apparatus are plotted in space relative to each other as well as relative to an external reference (Figs. 5.557). Trajectory analysis has the advantage that the displacements, velocities, and accelerations of elements relative to each other can be calculated over time and, if the masses of the elements are known, then the relative accelerations can be used to estimate the forces produced at each joint by muscular contraction. Traditionally, kinematic profiles have been used in studies of functional morphology (see, e.g.. Bramble and Wake, 1985; Reilly and Lauder, 1989), whereas trajectory analysis has been more commonly used in studies of motor control (e.g.. Bout and Ziegler, 1994). Whereas the anatomy of the feeding apparatus has been studied in 61 genera of frogs (Magimel-Pelonnier, 1924; Regal and Gans, 1976; Horton, 1982), kinematic studies have been performed for only a handful of genera. Among anurans, still photos are available for Discoglossus pictus, Pelobates fuscus and Rana temporaria (Vences, 1988). Descriptions of films are available for Ascaphus truei (Larsen and Guthrie, 1975), Bombina orientalis (Regal and Gans, 1976), Xenopus laevis (Avila and Frye, 1977), Hymenochirus boettgeri (Sokol,
F I G U R E 5.5. Movement trajectories of Hyla cinerea, a frog that uses mechanical pulling to protract the tongue. Anterior is toward the right. Trajectories A - C begin at the same point, but are offset for clarity. Arrows show the direction of movement in each segment of the trajectories, and open circles indicate the position of each element in each video field. In this sequence, the frog uses tongue prehension to capture the prey so there is little craniovertebral flexion. Trajectories of (A) the upper jaw tip, (B) lower jaw tip, and (C) tongue tip relative to an external reference point. Relative to an external reference point, the upper jaw tip moves forward during the lunge and backward during body recovery and the lower jaw tip moves forward during the lunge, then downward as the mouth opens and upward and back as the mouth closes during body recovery. (D) Relative to the upper jaw tip, the lower jaw tip moves downward and backward during mouth opening, then upward and forward during mouth closing. (E) Relative to the lower jaw tip, the tongue tip moves upward then forward relative to the lower jaw tip, then down and back. The tongue shortens during protraction and never protrudes more than a few millimeters beyond the tips of the mandibles.
F I G U R E 5.6. Movement trajectories of Bufo marinus, a frog that uses inertial elongation to protract the tongue. Anterior is toward the right. Trajectories A - C begin at the same point, but are offset for clarity. Arrows indicate the direction of movement in each segment of the trajectories, and open circles indicate the position of each element in each video field. Trajectories of (A) the upper jaw tip, (B) the mandibular tip, and (C) the lingual tip relative to an external reference. The lingual tip follows a nearly straight trajectory from mouth to prey during protraction as it elongates by up to 180% of its resting length. Relative to an external reference point, the upper jaw tip moves downward and forward during the lunge, then upward and back during body recovery. (D) Relative to the upper jaw tip, the movement of the lower jaw tip moves is mostly limited to the vertical plane. (E) The trajectory of the lingual tip relative to the mandibular tip is first upward and forward as the tongue shortens, but then changes abruptly to downward and forward as the tongue elongates and the lingual tip passes beyond the mandibles.
1969), B. marinus (Gans and Gorniak, 1982a,b), and Rana catesbeiana (Gans, 1961, 1962). Kinematic profiles have been published for a handful of genera, including Bufo (Lauder and Reilly, 1994), Rana, Bufo, Kaloula, Pyxicephalus, and Dyscophus (Emerson, 1985), Ascaphus truei (Nishikawa and Cannatella, 1991), Discoglossus pictus (Nishikawa and Roth, 1991), Hyla cinerea (Deban and Nishikawa, 1992), phyllomedusine hylids (Gray and Nishikawa, 1995), S. multiplicata (O'Reilly and Nishikawa, 1995), Rana pipiens (Anderson, 1993), and Hemisus marmoratum (Ritter and Nishikawa, 1995). The marine toad B. marinus is the only amphibian for which a trajectory analysis of feeding movements has been published (Nishikawa and Gans, 1996). Kinematic profiles (Fig. 5.4) are useful for analyzing the timing of movements of different elements of the feeding apparatus relative to each other. However, as a method for quantifying movement, kinematic profiles are limited by the fact that displacements are plotted in time rather than space. Important information about
127
5. Feeding in Frogs
5 mm
F I G U R E 5.7. Movement trajectories of Hemisus marmomtum, a frog that uses hydrostatic elongation to protract its tongue. Anterior is toward the right. Trajectories A - C begin at the same point, but are offset for clarity. Arrows indicate the direction of movement in each segment of the trajectories, and open circles indicate the position of each element in each video field. Trajectories of (A) the upper jaw tip, (B) the mandibular tip, and (C) the lingual tip relative to an external reference. There is relatively little head movement during feeding, although the head moves slightly upward and forward, then down and back during feeding. (D) The lower jaw tip is both depressed and retracted during mouth opening, and retraction of the lower jaw counteracts the upward rotation of the tongue tip relative to the mandibles so that the initial trajectory of the tongue (C) is straight out of the mouth. (E) Unlike inertial elongators, the tongue does not follow a straight line from mouth to prey during protraction.
head, jaw, and tongue movements and their coordination can often be obtained by performing a trajectory analysis. For example, from kinematic profiles, we can observe that the skull typically begins to rotate upward as the mouth opens during feeding in Hyla and Bufo (Fig. 5.4). By plotting the trajectory of the mandibular tip relative to the tip of the upper jaw, we also see that mouth opening involves nearly pure depression of the mandibles in Bufo (Fig. 5.6), whereas the mandibles are both depressed and retracted to some extent in Hyla (Fig. 55) and Hemisus (Fig. 5.7). From the trajectory analysis, we can see that in Bufo the upward rotation of the head during mouth opening balances the retraction of the mandibles, which is an inevitable consequence of downward rotation of the mandibles, whereas the retraction of the mandibles is not entirely balanced by upward head rotation in Hyla and Hemisus. As a second example, we can examine tongue movements during prey capture. Kinematic profiles (Fig. 5.4) show that anuran species protract their tongues to different degrees and at different rates of speed. In addition, trajectory analyses show that the tongue tip of Bufo (Fig. 5.6) follows a straight line from mouth to
prey, whereas the tongues of Hyla (Fig. 5.5) and Hemisus (Fig. 5.7) do not. This straight trajectory can be seen to emerge from the precise coordination of head, jaw, and tongue movements (Fig. 5.6). Electromyographic (EMG) recordings coupled with kinematic analyses may help generate hypotheses about the roles that individual muscles play during feeding (Gans and Gorniak, 1982a,b). For anurans, EMG studies of feeding behavior have been performed only in the toads B. marinus (Gans and Gorniak, 1982a,b) and B. japonicus (Matsushima et al., 1985) and the frog R. pipiens (Anderson and Nishikawa, 1993). Gans and Gorniak (1982a,b) obtained records of EMG activity from several muscles in freely behaving toads feeding on natural prey. Matsushima et al. (1985) stimulated the optic tectum with implanted electrodes to elicit tongue flipping in freely behaving animals. In addition to EMG recordings, muscle stimulation has often been used to study the function of muscles believed to be involved in feeding behavior. It is often possible to perform muscle stimulation on uncooperative species that refuse to eat in captivity (Bemis et ah, 1983; Trueb and Gans, 1983). Among anurans, muscle stimulation experiments have been performed only in B. marinus (Emerson, 1977; Gans and Gorniak, 1982a,b) and R. dorsalis (Trueb and Gans, 1983). In addition to these traditional techniques, a variety of other techniques may provide additional information that may be used to understand the roles of particular elements of the feeding apparatus. One method is to denervate individual muscles by surgically transecting the nerves that innervate them in order to study how movements change after a muscle or set of muscles has been inactivated. This technique can be used to demonstrate that a particular muscle is either necessary or sufficient for performing a given movement (Nishikawa and Roth, 1991). However, because most peripheral nerves contain both sensory and motor fibers, it is important to consider the possibility that some observed effects of nerve transection may be due to deafferentation rather than to muscle denervation (Nishikawa and Gans, 1992). B. Hypotheses for the Mechanism of Tongue Protraction Historically, numerous mechanisms have been proposed to explain tongue flipping in anurans (for a review, see Gans and Gorniak, 1982b). Two relatively recent hypotheses include the hyoid model of Emerson (1977) and the ballista model of Gans and Gorniak (1982a,b). Both of these studies used a variety of techniques to study feeding in the marine toad, B. marinus. Both studies proposed relatively complex mechanisms
128
Kiisa C. Nishikawa
for tongue protraction and, although based on accurate and repeatable observations, both have been shown to have problems with some aspects of data interpretation (Nishikawa and Gans, 1992, 1996). It is useful to describe these hypotheses and the evidence on which they were based in order to determine which techniques were helpful, which were relatively uninformative, and which were misleading in terms of understanding the mechanism of tongue protraction in anurans. Emerson (1977) used dissection, cineradiography, and electrical stimulation to develop a biomechanical model of tongue protrusion in toads. Cineradiographic observations revealed that the hyoid is held in the retracted position during mouth opening, moves downward and forward during tongue protraction, and upward and backward during retraction. When stimulated electrically, the tongue protractor muscle M. genioglossus "positioned the tongue in the preflip condition but did not cause the tongue to leave the mouth." Stimulation of the M. depressor mandibulae did not cause the jaws to open, although stimulation of the Mm. geniohyoideus and sternohyoideus resulted in mouth opening. From these and other observations, Emerson (1977) hypothesized that the hyoid plays a static role during mouth opening and a dynamic role during protraction and retraction of the tongue. Specifically, she proposed that the hyoid retractor M. sternohyoideus stabilizes the hyoid in a retracted position in the floor of the mouth and stores potential energy during the early stages of mouth opening, which she presumed to be caused by contraction of the hyoid muscles Mm. geniohyoideus and sternohyoideus. As the mouth opens more fully under the contraction of the M. depressor mandibulae and the M. sternohyoideus stops contracting, potential energy is released as the hyoid moves anteriorly, and this released energy is imparted to the tongue as kinetic energy. Posterior movement of the hyoid, perhaps due to elastic recoil, would initiate retraction of the tongue, which would then be fully retracted by contraction of the M. hyoglossus. When the hyoid is fully retracted, it is once again held in place by the action of the M. sternohyoideus. Emerson's (1977) model was refuted by Gans and Gorniak (1982a,b) who showed that (1) the pattern of EMG activity in the muscles of the hyoid and tongue is inconsistent with the model; (2) the tongue is stiff rather than flaccid when protracted; and (3) when the hyoid is wired to the sternum, and thus can impart no kinetic energy to the tongue, toads were observed to feed normally. As an alternative to Emerson's hyoid model, Gans and Gorniak (1982a,b) offered their ballista model of
tongue protraction in anurans, which is now widely accepted. This model was derived initially from kinematic analyses and muscle stimulation experiments in R. catesheiana (Gans, 1952,1961; Severtzov, 1961). Later, additional species were studied anatomically (Regal and Gans, 1976) and the model was eventually tested using dissection, electromyography, and muscle stimulation experiments in the toad B. marinus (Gans and Gorniak, 1982a,b). In the ballista model, the M. genioglossus medialis forms a stiffened rod and the M. genioglossus basalis forms a wedge at the anterior end of the rod near the mandibular symphysis. The main force-generating element is the M. submentalis, which depresses the mandibular symphysis, lifts the middle of the lingual rod, and rotates it about the symphysis (Gans and Gorniak, 1982a,b). The observations on which the model was based include: (1) electromyographic data showing that the M. submentalis and Mm. genioglossus basalis and medialis are active during tongue protraction and (2) the observation that stimulation of the M. genioglossus basalis resulted in rotation of the base of the tongue toward the mandibular symphysis, whereas stimulation of the M. genioglossus medialis resulted in stiffening of the tongue with no anterior movement. Results of muscle denervation experiments and detailed kinematic analyses have shown that the ballista hypothesis for tongue protraction also requires substantial revision (Nishikawa and Gans, 1996). Specifically, denervation of the M. submentalis, which is the major force-producing element in the ballista hypothesis, had no effect on tongue protraction in B. marinus, whereas denervation of the Mm. genioglossus basalis and medialis reduced tongue protraction significantly. From these experiments, Nishikawa and Gans (1996) concluded that the M. genioglossus medialis not only stiffens the tongue during protraction, but also is the main force-generating element that pulls the tongue forward out of the mouth. We can now examine why the models of Emerson (1977) and Gans and Gorniak (1982a,b), although based on accurate and repeatable observations, failed to identify the muscles that are responsible for tongue protraction in toads. First, neither study described the kinematics of tongue protraction in toads with either kinematic profiles or trajectory analyses, so both studies overlooked the fact that the tongue normally elongates by approximately 180% of its resting length during protraction (Nishikawa and Gans, 1996). Second, EMG studies were relatively uninformative because numerous muscles are activated simultaneously during feeding (Gans and Gorniak, 1982b), which makes it difficult to understand how each contributes to feeding movements. Furthermore, electromyographic data
129
5. Feeding in Frogs were positively misleading because some muscles that are active during tongue protraction (i.e., the M. submentalis) appear to have little or no effect on tongue movements as demonstrated by surgical denervation experiments (Nishikawa and Gans, 1996). Finally, the results of muscle stimulation experiments were also positively misleading. At least three repeatable results were misinterpreted by Emerson (1977), Gans and Gorniak (1982a,b), or both. First, Emerson (1977) reported accurately that stimulation of the M. depressor mandibulae is insufficient to open the mouth of a spinal pithed toad. The explanation for this result, however, is not that the force produced by the muscle is insufficient to open the mouth, but rather that, even in spinal pithed toads, tonic contractions of the Mm. levator mandibulae resist mouth opening (Nishikawa and Gans, 1992). During normal feeding, this tonic activity is inhibited during mouth opening so that the M. depressor mandibulae does not have to overcome the additional force produced by tonic activity. When the M. levator mandibulae is denervated, stimulation of the M. depressor mandibulae results in rapid opening of the mouth (Nishikawa and Gans, 1992). Second, Emerson (1977) accurately reported that stimulation of the M. geniohyoideus causes the mouth to open. Once again, however, the explanation is rather complex. In this case, the hypoglossal nerve runs through the body of the M. geniohyoideus and it contains sensory fibers from the tongue that, when stimulated, inhibit tonic contractions of the M. levator mandibulae (Nishikawa and Gans, 1992). When the M. geniohyoideus is stimulated, the hypoglossal nerve is also stimulated, and tonic contractions of the M. levator mandibulae are inhibited. In this case, stimulation of the M. geniohyoideus causes the mouth to open, but it takes less force to open the mouth because the tonic contractions of the M. levator mandibulae are inhibited by electrical stimulation of the hypoglossal nerve that runs through the belly of the M. geniohyoideus. When the hypoglossal nerve is transected before it enters the M. geniohyoideus, stimulation of the M. geniohyoideus no longer produces mouth opening because the tonic contractions of the M. levator mandibulae are no longer inhibited (Nishikawa and Gans, 1992). Both of these examples illustrate the fact that muscle stimulation experiments can be positively misleading because the neural pathways, both sensory and motor, may remain intact even after spinal pithing and affect the force necessary to achieve a given movement. In muscle stimulation experiments, tonic muscle activity may be present that is not present during normal behavior in the intact animal. Furthermore, electrical stimuli may activate not only muscle fibers but also
nearby nerve fibers. In turn, stimulation of sensory nerve fibers may activate central neural circuits that modulate tonic activity of other muscles, thereby changing the amount of force that is necessary to achieve a given movement. For these reasons, the results of muscle stimulation experiments should always be interpreted with caution. Finally, both Emerson (1977) and Gans and Gorniak (1982b) noted that stimulation of the Mm. genioglossus in a spinal pithed toad does not result in protraction of the tongue. Based on this and other observations, Emerson (1977) hypothesized that the hyoid played an important role in tongue protraction, whereas Gans and Gorniak (1982a,b) hypothesized that the M. submentalis was the major force-generating element for tongue protraction. In contrast, denervation experiments suggest that the Mm. genioglossus are the major forcegenerating elements for tongue protraction during feeding (Nishikawa and Gans, 1996). The resolution of these conflicting observations seems to be that the dynamics of tongue protraction, which in toads involves large, simultaneous displacements of the jaws and tongue in as little as 35 msec (Nishikawa and Gans, 1996), cannot be duplicated under the static conditions in which muscle stimulation experiments necessarily are conducted. In my opinion, our current understanding of the feeding mechanisms of anurans has come mostly from detailed anatomical descriptions, kinematic analyses of movement, including both kinematic profiles and trajectory analyses, and experimental denervation of feeding muscles. In the case of anuran feeding behavior, both electromyographic studies and muscle stimulation experiments have been mostly uninformative and sometimes even misleading. An understanding of the biomechanics of complex anatomical systems can only be achieved by using as many techniques as possible to develop and test hypotheses about function. C. Functional Diversification among Anuran Taxa Over the past several years, my students and I, along with several collaborators, have used many different techniques to study the function of the anuran feeding apparatus. Detailed studies have been published for several anuran species, including D. pictus (Nishikawa and Roth, 1991); S. multiplicata (O'Reilly and Nishikawa, 1995); Hyla cinerea (Deban and Nishikawa, 1992); Pachymedusa dacnicolor (Gray and Nishikawa, 1995); B. marinus (Nishikawa and Gans, 1996); Dendrohates, Phyllobates, and Epipedobates (Wiltenmuth and Nishikawa, 1994); R. pipiens (Anderson, 1993); H. marmoratum (Ritter and Nishikawa, 1995); and
130
Kiisa C. N i s h i k a w a
Phrynomerus bifasciatus (Jaeger and Nishikawa, 1993; Meyers et al, 1996). From these comparative studies, we have identified three different mechanisms that anurans use to protract their tongues during feeding in terrestrial environments. These mechanisms are mechanical pulling, inertial elongation, and hydrostatic elongation, and each is described in detail in the following sections. All of the nearly 150 species that have been examined so far fall into one of these categories, although additional mechanisms, such as hyoid pushing in mesobatrachians (see Section II,C; O'Reilly and Nishikawa, 1995), may also be present. 1, Mechanical
Pulling
The first mechanism of tongue protraction is mechanical pulling, here exemplified by H. cinerea (Fig. 5.8). It is found in all archaeobatrachian lineages, including Ascaphus (Ascaphidae) (Nishikawa and Cannatella, 1991), Leiopeltna (Leiopelmatidae), Bombina (Bombinatoridae), and Discoglossus (Discoglossidae) (Nishikawa and Roth, 1991), as well as in some mesobatrachians (e.g., Pelobates and Spea) and some neobatrachians (e.g., Hyla). Anatomically, the tongues of mechanical pullers tend to be round in shape, broadly attached to the floor of the mouth, and there is much interdigitation of the M. hyoglossus and M. genioglossus medialis throughout the tongue pad (MagimelPelonnier, 1924; Regal and Gans, 1976; Horton, 1982). The ventralmost fibers of the M. hyoglossus insert near the anterior tip of the tongue (Horton, 1982). In mechanical pullers, the function of the tongue protractor muscle M. genioglossus is most like that of typical vertebrate skeletal muscle. Both the M. genioglossus and the tongue as a whole shorten during protraction as the fibers of the M. genioglossus contract and pull the tongue tip toward the mandibular symphysis (Fig. 5.8). This mechanism is essentially similar to that proposed by Regal and Gans (1976). In mechanical pullers, the resting length of the tongue is approximately equal to the length of the mandibles (Figs. 5.3A-5.3C). The tongue shortens as the M.
F I G U R E 5.8. Selected frames from a normal prey-capture sequence for Hyla cinerea, showing mechanical pulling. The number in the right-hand corner of each picture indicates the time (msec) from the onset of mouth opening. The squares in the background are 1 X 1 cm. The short tongue can be seen at full protraction ait = 17 msec. In this sequence, the frog uses jaw prehension to capture the waxworm. The craniovertebral joint is flexed downward between t = 42 and t = 92 msec, and at f = 92 msec, the jaws close on the prey.
131
5. Feeding in Frogs TABLE 5.1 Characteristics of the Different Mechanisms of Tongue Protraction Characteristics
Hydrostatic elongation
Inertial elongation
Mechanical pulling
Example
Hyla cinerea
Bufo marinus
Hemisus marmoratum
Tongue movement
Tongue shortens
Tongue elongates 180%
Tongue elongates 200%
Velocity (cm/sec)
79
286
24
Acceleration (m/sec^)
66
310
3.5
Tongue/jaw synchrony
No
Yes
No
Aiming
Distance
None
Distance, azimuth, elevation
Accuracy
95%
33%
>99%
On-line correction
Yes
No
Yes
Feedforward control
Yes
Yes
Yes
Feedback control
Yes
No
Yes
Hypoglossal afferents
No
Yes
No
genioglossus contracts, pulling the tongue pad upward and forward toward the symphysis (Figs. 5.4, 5.5E, and 5.8). Thus, in mechanical pullers, the tongue is shorter at full protraction than it is at rest in the floor of the mouth. In H. cinerea, for example, the length of the tongue at maximum protraction is approximately 60% of its resting length (Fig. 5.4). Movements of the head, jaws, and tongue are small and rather asynchronous (Fig. 5.11) in mechanical pullers. Because the tongue shortens during protraction, the frogs must lunge forward with their entire bodies in order to place the tongue on the prey (Figs. 5.4,5.5A, and 5.8). When capturing relatively large prey, the frogs also exhibit head flexion in order to bring the tongue down onto the prey, and the prey are apprehended with the jaws (Figs. 5.4 and 5.8). For small prey, tongue prehension is used for prey capture and there is less flexion of the head (Fig. 5.5A). During feeding, the lower jaw tip is depressed and retracted slightly relative to the upper jaw tip (Fig. 5.5D). The short tongue can hardly be moved independently of the lower jaw (Fig. 5.5D) and the maximum reach of the tongue beyond the jaws typically is not more than a few millimeters (Fig. 5.4; Deban and Nishikawa, 1992). Relative to the lower jaw tip, to which the tongue is attached, the tongue reaches moderate velocities (79 cm/sec) and accelerations (66 m/sec^) during protraction (Table 5.1). Tongue movements can be corrected in progress within a single gape cycle and there is little, if any, ability to aim the tongue relative to the head (Deban and Nishikawa, 1992). In the two species of mechanical pullers that have been studied (Gray, 1997), prey capture success ranges from 95% in H. cinerea to 68% in H. arenicolor when feeding on relatively small, slow-moving prey such as waxworms.
2. Inertial Elongation In contrast to mechanical pullers whose tongues shorten during protraction, the tongues of many frogs elongate during protraction (Figs. 5.9 and 5.10). Given the fact that muscles can only contract to do work, and given that all frogs exhibit rather similar morphologies of the extrinsic tongue muscles in which most fibers run parallel to the long axis of the tongue, it is rather surprising that the tongues of these frogs can elongate beyond resting length during protraction. What mechanisms are responsible for tongue elongation and which muscles are responsible? Many anurans use inertia to elongate the tongue during protraction. Inertial elongation is found among several anuran lineages, including some leptodactyline leptodactylids {Physalaemus and Pleurodema), some eleutherodactyline leptodactylids {Eleutherodactylus), all bufonids, all phyllomedusine hylids, and all ranoids except hemisotids and microhylids (Nishikawa et ah, 1992). This chapter uses the marine toad, B. marinus, as an example of a typical inertial elongator (Fig. 5.9). The tongues of inertial elongators are similar morphologically to those of mechanical pullers, except that the fibers of the protractor and retractor muscles are relatively longer, which produces a posterior flap that is free from the floor of the mouth (Magimel-Pelonnier, 1924; Regal and Gans, 1976; Horton, 1982). In addition, there is less interdigitation between the M. hyoglossus and the M. genioglossus in the tongue pad, and the ventralmost fibers of the M. hyoglossus recurve to insert in the posterior, rather than the anterior, part of the tongue (Horton, 1982). The relative mass of the tongue is typically smaller in inertial elongators (0.5%) than in mechanical pullers (1.0%).
132
Kiisa C. N i s h i k a w a
The initial stages of tongue protraction are similar in mechanical pullers and inertial elongators. Inertially elongated tongues shorten at first as the M. genioglossus medialis contracts and accelerates the tongue pad upward and forward. In inertial elongators, however, the tongue elongates by as much as 180% of its resting length under its own inertia after the initial shortening phase (Nishikawa and Gans, 1996). Inertial elongation appears to be the fastest and least accurate mechanism of tongue protraction in anurans. In toads (B. marinus), the tongue tip reaches velocities of up to 270 cm/sec and accelerations of 310 m/sec^ (more than 30 times gravity) relative to the tips of the mandibles during protraction (Nishikawa and Gans, 1996). These velocities and accelerations are several times greater than those of mechanical pullers (Table 5.1). Toads are also much less accurate (30%) at capturing prey than mechanical pullers (Table 5.1; Gray, 1997). Because tongue protraction is rapid and ballistic, tongue movements cannot be corrected within the gape cycle (Nishikawa and Gans, 1996). Like mechanical pullers, toads possess little, if any, ability to aim the tongue relative to the head. The complexity of tongue and jaw movements during feeding in inertial elongators is demonstrated by the trajectories of the upper jaw tip, mandibular tip, and lingual tip (Fig. 5.6), which show how the individual movements of these elements contribute to tongue protraction (Nishikawa and Gans, 1996). A remarkable feature that emerges from this analysis is that the lingual tip follows an almost straight line from mouth to prey, despite rotation of the tongue over the mandibular symphysis during protraction, from top to bottom as well as rear to front, and despite substantial changes in the shape of the tongue during protraction (Fig. 5.9). The straight-line trajectory of the lingual tip relative to an external reference is the sum of the relative trajectories of all the elements to which the tongue is attached: (1) the trajectory of the upper jaw tip relative to an external reference, which is first downward and forward, then upward and back (Fig. 5.6A); (2) the trajectory of the mandibular tip relative to that of the upper jaw tip, which is mostly restricted to the vertical plane
F I G U R E 5.9. Selected frames from a normal prey-capture sequence for Bufo marinus, showing inertial elongation. The number in the right-hand corner of each picture indicates the time (msec) from the onset of mouth opening. The squares in the background are 1 X 1 cm. The tongue shortens at first (f = 8 msec) but then elongates by up to 180% of its resting length under its own momentum (f = 2 5 42 msec).
5. F e e d i n g in Frogs
133
(Fig. 5.6D); and (3) the trajectory of the lingual tip relative to the mandibular tip, which first rises as the tongue shortens and moves forward, but then changes direction abruptly as the lingual tip passes beyond the oral cavity and moves downward and forward as the tongue elongates (Fig. 5.6E). Trajectory analysis demonstrates that precise coordination among head, jaw, and tongue movements is responsible for the straight trajectory of the tongue tip from mouth to prey. First, the downward and forward and then the upward and back movement of the head counteracts the tendency for mouth opening to produce retraction of the lower jaw, which would occur without compensatory head movements because the jaw tip must transcribe a circular path about the jaw joint. Furthermore, a large upward acceleration of the lingual tip occurs simultaneously with a large downward acceleration of the mandibular tips (Fig. 5.11). The acceleration of the tip of the mandibles relative to that of the upper jaw is oriented downward, whereas the acceleration of the lingual tip relative to that of the mandibles is oriented upward and forward. These large, synchronous, partially opposed jaw and tongue movements add together precisely to generate the straight lingual trajectory (Nishikawa and Gans, 1996). Nishikawa and Gans (1996) used a simple matrix model to demonstrate that the linear trajectory of the tongue depends on the simultaneous accelerations of tongue and jaws. In this model, the relative timing of displacements of the jaws and tongue was varied systematically and the change in tongue trajectory was measured. This analysis showed that changes in the relative timing of the displacements had only a small effect on the horizontal position of the tongue tip. However, the vertical position of the lingual tip strongly depends on the simultaneous displacement of the mandibular and lingual tips. If the vertical displacements of the lingual tip are added to the vertical displacements of the upper jaw and mandibular tips 8 msec (i.e., one video field) earlier than they actually occur, then the trajectory of the lingual tip would intersect that of the upper jaw (i.e., the tongue would contact the palate). If the lingual displacements occur 8 msec too late, then the lingual tip would drop with the mandible before accelerating upward, which would
FIGURE 5.10. Selected frames from a normal prey-capture sequence for Hemisus marmoratum, showing hydrostatic elongation. The number in the right-hand corner of each picture indicates the time (msec) from the onset of mouth opening. The squares in the background are 1 X 1 cm. The tongue elongates slowly during protraction {t = 25-125 msec).
134
Kiisa C. N i s h i k a w a
Hemisus
>
-0.75 -200
-100
Time (ms) F I G U R E 5.11. Horizontal (above) and vertical (below) displacements between successive frames for Hyla, Bufo, and Hemisus (upper jaw tip, closed circles; mandibular tip, open circles; lingual tip, open squares). For Hyla, horizontal displacements of the upper jaw tips are greater than those of the mandibular and lingual tips, whereas for Bufo and Hemisus, horizontal displacements of the lingual tip are greater than those of the upper jaw and mandibular tips and are positive during protraction and negative during retraction. In Bufo, the downward displacement of the mandible during mouth opening occurs simultaneously with a large upward displacement of the lingual tip, whereas movements of the jaws and tongue are asynchronous in Hyla and Hemisus.
produce a nonlinear trajectory that deviates from the observed trajectory by more than 1 cm (approximately 20% of tongue length). Thus, the synchrony of the large displacements of the tongue and mandibles is necessary for successful prey capture in inertial elongators, but not in mechanical pullers (Table 5.1). 3. Hydrostatic
Elongation
The third mechanism of tongue protraction, hydrostatic elongation, is possessed only by members of the families Hemisotidae and Microhylidae. This chapter uses Hemisus marmoratum as a representative hydrostatic elongator (Fig. 5.10). Hemisus may elongate its tongue up to 200% of resting length (Ritter and Nishikawa, 1995). However, it protracts its tongue too slowly to use an inertial mechanism. The tongue tip reaches maximum velocities of 15 cm/sec (nearly 20 times slower than inertial elongators) and accelerations of 0.5 m/sec^ (more than 600 times slower than inertial elongators. Table 5.1). Instead of inertia, a hydrostatic mechanism is used to protract the tongue (Nishikawa et al, 1995). In hydrostatic elongators, the tongue protractor muscle (M. genioglossus) consists of two compartments, one in which the muscle fibers are oriented par-
allel to the long axis of the tongue as in other frogs and one in which the fibers are oriented vertically (Nishikawa et al, 1995; Fig. 5.2D). During protraction, the volume and the width of the tongue remain constant. When the vertical fibers contract, they decrease the thickness of the tongue and, because the volume of the tongue is constant, this change in shape is translated directly into tongue elongation (Nishikawa et ah, 1995). A decrease in tongue thickness of 50% (from 2 to 1 mm) is translated into a 100% increase in tongue length (from 5 to 10 mm). Ritter and Nishikawa (1995) favored a hydraulic mechanism of tongue elongation in Hemisus because they overlooked the presence of the dorsoventral compartment of the M. genioglossus in H. marmoratum. The presence is apparent only in sections, and not in gross dissection. In contrast to inertial elongators, which must protract their tongues rapidly in order to achieve elongation, hydrostatic elongators may protract their tongues either slowly or rapidly. In terms of prey capture, H. marmoratum is the slowest and most accurate of the anuran species that have been studied to date (Ritter and Nishikawa, 1995). Other hydrostatic elongators, such as Dyscophus insularis, protract their tongues more rapidly than Hemisus. Hydrostatic elongators exhibit high accuracy of prey capture (99%; Gray, 1997) and on-line
135
5. Feeding in Frogs correction of tongue movements within a gape cycle (Table 5.1). In contrast to both mechanical pullers and inertial elongators, there is very little head movement during feeding in hydrostatic elongators, although the head moves slightly upward and forward and then down and back during feeding (Fig. 5.7A). The lower jaw tip is both depressed and retracted during mouth opening (Fig. 5.7D), and retraction of the lower jaw counteracts the upward rotation of the tongue tip relative to the mandibles so that the initial trajectory of the tongue is straight out of the mouth, rather than upward and forward as in inertial elongators (Fig. 5.7C). In contrast to inertial elongators, the tongue does not follow a straight line from mouth to prey during protraction (Fig. 5.7E). As in mechanical pullers, movements of the head, jaws, and tongue are asynchronous (Fig. 5.11). All anurans studied to date can aim their heads relative to their bodies and do so readily during feeding. In addition, some mechanical pullers can modulate tongue protraction distance in response to prey distance (Deban and Nishikawa, 1992; O'Reilly and Nishikawa, 1995). All hydrostatic elongators can modulate tongue protraction distance in response to prey distance, and in addition are the only anurans that can aim their tongues relative to their heads in dimensions other than distance (Ritter and Nishikawa, 1995; Jaeger and Nishikawa, 1993; Meyers et al, 1996). Phrynomerus bifasciatus (family Microhylidae) aims its tongue relative to the head in distance and azimuth (Jaeger and Nishikawa, 1993; Meyers et al, 1996), whereas H. marmoratum aims its tongue relative to its head in distance, azimuth, and elevation (Ritter and Nishikawa, 1995). Insight into the mechanism of tongue aiming comes from denervation experiments in Hemisus (Ritter and Nishikawa, 1995) and Phrynomerus (Meyers et ah, 1996). In both species, when unilateral denervation of the M. genioglossus is performed, the tongue bends toward the inactivated side. In Phrynomerus, the tongue deviates by up to 90° from the target, whereas in Hemisus it deviates by more than 180°. These experiments are consistent with the proposed hydrostatic mechanism of tongue protraction, in which the tongue bends toward the inactivated side to equalize tensile stresses in the tongue. They also suggest that microhylids and hemisotids may regulate the azimuth of the tongue by differential recruitment of the right and left sides of the M. genioglossus (Ritter and Nishikawa, 1995). In contrast, the amplitude of tongue movement is reduced after unilateral denervation of the M. genioglossus in B. marinus, but the direction of tongue protraction is unaffected (personal observation). Finally, H. marmoratum is the only frog that is known to possess a truly prehensile tongue. In other frogs, the
tongue sticks to prey by wet adhesion. In Hemisus, the tongue not only sticks to prey but actually grasps it (Ritter and Nishikawa, 1995). If a termite is held with forceps, the tongue of H. marmoratum can generate enough tensile force to tear it in half. Muscle denervation experiments demonstrate that activation of the M. hyoglossus is necessary for this prehensile function. The M. hyoglossus sends a fascicle into each lobe of the bilobed tongue and, when inactivated, prehension is eliminated, although the frogs can still capture prey using lingual adhesion (Ritter and Nishikawa, 1995; Tso et al, 1995). These experiments also demonstrate that the M. hyoglossus plays an important role in swallowing. Intact Hemisus always ingests, transports, and swallows termites in a single movement, whereas the tongue is often protracted with a previously captured termite still adhering to it after M. hyoglossus denervation.
IV. NEURAL CONTROL OF PREY CAPTURE Until the 1970s, when in vitro intracellular recording techniques became feasible in the intact vertebrate central nervous system, patterns of motor output were widely believed to result from a chain of reflex-like interactions between sensory and motor neurons (Delcomyn, 1980). Since then, however, in vitro and deafferentation studies in a large number of animals, ranging from leeches to primates, have shown that intrinsic patterns of motor output are produced in the absence of sensory input (Delcomyn, 1980; Grillner, 1985). The neurons responsible for producing this output have been termed central pattern generators (CPGs). These CPGs are implicated in the production of numerous behavior patterns, including breathing, swimming, walking, and feeding. In frogs, a variety of evidence points to the medial reticular formation as a possible site of the CPG for prey capture (Matsushima et al, 1989; Weerasuriya, 1989). Unlike a spinal cord in a dish, however, all behaving animals must produce varied patterns of motor output that are exquisitely appropriate to the animal's changing conditions. Adaptive behavior involves the production of a motor response that is appropriate in the context of incoming sensory stimuli. In order to understand the neural basis of adaptive behavior, we need to know how central pattern generators and motor neurons interact with sensory receptors to produce motor output that tracks changes in an animal's external environment as well as its internal state. Current theories suggest that sensory input acts directly on CPGs to change the frequency, amplitude, and phase of
136
Kiisa C. Nishikawa
motor patterns (Rossignol et al, 1988). In theory, such changes result from the effects of neurotransmitters and neuromodulators on the membrane properties of CPG neurons (Harris-Warrick, 1988). Empirically, however, the mechanisms by which sensory input influences CPGs are not yet well understood, even in relatively well-studied systems such as lamprey swimming (Grillner et al, 1988). Comparative studies of the neural control of movement should be conducted in the context of the adaptive modification of CPG activity by incoming sensory input. Within this conceptual framework, we can investigate which sensory modalities are involved in modulating prey capture movements, how sensory information influences the pattern of motor activity produced by central pattern generators, and how sensory modalities interact to modulate movement. We can then proceed to ask whether species differ in these aspects of motor control. There are several problems that frogs must overcome to capture prey successfully. These include detecting the prey, locating it in space, and analyzing its relative size, shape, and speed of movement. Once prey are detected and located, a frog must respond to it before it escapes, place its tongue accurately on the prey, apprehend it, and bring it back to the mouth. Apprehension requires planning and execution of precisely coordinated movements of the head, jaws and tongue (Nishikawa and Gans, 1996). Some of these problems will be discussed later. Other problems, such as locating prey in space (Ingle, 1983), are beyond the scope of this chapter. In terms of neural control, prey capture in frogs is a goal-oriented movement that is similar to reaching in humans (Gottlieb et al, 1989; Flanders et al, 1992) or pecking in pigeons (Bermejo and Ziegler, 1989). Performance of these goal-oriented tasks requires sensory information about the target and sensory information about the animal's internal state, both of which are used to modulate the output of CPGs. Relevant information about the target is often obtained visually and includes target position in three dimensions (distance, azimuth, and elevation), size, shape, and velocity. Relevant information about the animal's internal state is obtained through a variety of proprioceptive sense organs distributed throughout the body and includes the length and mass of musculoskeletal elements, the force-velocity relationships and mechanical advantage of the muscles, the position of musculoskeletal elements before and during movement, and the action of forces such as gravity and inertia. Prey recognition has been studied extensively in amphibians (Ewert, 1987; Roth, 1987), and a review of these studies is beyond the scope of this chapter. To date, motor control of prey capture has been studied in
detail in only two anuran species, R. pipiens (family Ranidae) and B. marinus (family Bufonidae). Methods that have been used to study neural control of prey capture include behavioral studies, deafferentation experiments, electrophysiological recording, and neuroanatomical tracing. These studies have focused on the modulatory effects of visual analysis of prey features (Anderson, 1993; Valdez and Nishikawa, 1997), of proprioceptive tongue afferents (Nishikawa and Gans, 1992; Nishikawa et al, 1992; Anderson and Nishikawa, 1993, 1997), and of the interaction between vision and proprioception in controlling feeding movements (Anderson and Nishikawa, 1993,1996). The modulatory effects of visual input on prey capture movements, the modulatory effects of tongue afferents on prey capture movements, and the interaction between vision and proprioception in controlling feeding movements are described in detail next. A. Visual Analysis of Prey Features Some anurans will snap in response to tactile stimulation (Comer and Grobstein, 1981). Some species also can locate prey on the basis of olfactory cues alone (Dole et al, 1981; Shinn and Dole, 1978). Even auditory stimuli can be used to locate prey (Martof, 1962; Jaeger, 1976). However, vision appears to be the dominant sensory modality that most frogs use to detect prey. When vision is intact, frogs seldom lunge or snap at stationary prey (Lettvin et al, 1959; Kaess and Kaess, 1960; Ewert, 1985; Satou and Shiraishi, 1991). There have been numerous studies of the visual cues used in prey recognition by toads (reviewed in Ewert, 1987). These studies measured the rates of orientation and snapping elicited by artificial, two-dimensional, prey-like stimuli that differed in shape, size, and speed of movement. These studies showed that toads (B. hufo) exhibit higher orienting rates for rectangular objects which move in a direction that is parallel to their long axis (i.e., worm orientation) than for similar objects that move perpendicular to their long axis (i.e., antiworm orientation). Not only the tendency to respond, but also patterns of movement change in response to prey characteristics (Anderson, 1993; Anderson and Nishikawa, 1996; Valdez and Nishikawa, 1997). For example, R. pipiens (an inertial elongator, family Ranidae) exhibits different behavior patterns to capture small vs large prey (Anderson, 1993). Small prey are captured with the tongue and are transported to the esophagus without contacting the jaws (here termed "tongue prehension"), whereas large prey are first contacted with the tongue but the head rotates downward, the prey are captured in the closing jaws, and are transported into the oral cavity with the forelimbs (here termed "jaw
137
5. F e e d i n g in Frogs
prehension"). During tongue prehension, the head remains more nearly level with respect to the horizon, the lunge distance is shorter, the tongue is protracted to a greater distance, and the mouth remains open for a longer time than during jaw prehension (Anderson, 1993). Because they depend heavily on lingual adhesion to capture prey, it is not surprising that frogs with generalized diets use different strategies to capture prey of different sizes. Tongue prehension is more effective for capturing small prey than jaw prehension because the prey is transported to the esophagus in a single movement, which offers less chance for escape. However, tongue prehension is effective only if the mass of the prey is less than the adhesivity between tongue and prey. Thus, jaw prehension is more effective for capturing large prey (Anderson and Nishikawa, 1996; Valdez and Nishikawa, 1997). It has been shown that the decision to use jaw vs tongue prehension is made on the basis of a visual analysis of prey size (Anderson and Nishikawa, 1996). When offered pieces of earthworm ranging in size from 1.5 to 4.5 cm, adult frogs {R. pipiens) always use tongue prehension to capture 1.5-cm prey and jaw prehension to capture 2.0-cm and larger prey (Figs. 5.13A and 5.13B). Not surprisingly, the distinction between "small" vs "large" prey is relative to the size of the frog. Larger frogs switch behavior patterns at larger prey sizes than smaller frogs. Another study compared prey capture movements across five different types of live prey (earthworms, waxworms, newborn mice, crickets, and termites) in the Australian frog, Cyclorana novaehollandiae (a mechanical puller, family Hylidae) (Valdez and Nishikawa, 1997). This study showed that these frogs modulate their feeding movements in response to features of prey in addition to size, especially shape and speed of movement. It also showed that the ability to use a visual analysis of prey characteristics to modulate feeding movements is widespread among frogs. Distantly related frogs with different tongue morphologies (i.e., R. pipiens, family Ranidae, and C. novaehollandiae, family Hylidae) use tongue prehension to capture small prey and jaw prehension to capture large prey. However, some frogs appear to have lost the ability to switch between tongue prehension and jaw prehension. For example, B. marinus uses tongue prehension to capture large prey as well as small prey, whereas Leptopelis uses jaw prehension to capture both small and large prey. B. Role of Tongue A£f erents In both frogs (R. pipiens) and toads (B. marinus), mechanosensory afferents of the tongue, innervated by
the hypoglossal nerve, serve a variety of functions in feedforward (i.e., open loop, planned in advance) control of jaw and tongue movements during prey capture. One function is to modulate the phase of activity in the mouth opening and closing muscles (Nishikawa and Gans, 1992; Anderson and Nishikawa, 1993). In intact frogs and toads, the mouth opening muscles are active approximately 90 msec before the mouth closing muscles. After bilateral transection of the hypoglossal nerves, the mouth remains closed when Rana or Bufo attempt to feed because the M. levator mandibulae and the M. depressor mandibulae are activated simultaneously (Figs. 5.12E-5.12H and 5.13C). Thus, sensory input from the tongue coordinates jaw muscle activity by sending an afferent signal to the brain that delays activity of the jaw levators. This signal is produced
After
Before
200 ms FIGURE 5.12. The function of hypoglossal afferents in the marine toad, Bufo marinus. (Left) Normal feeding in an intact toad before deafferentation, illustrating inertial elongation. (A) The toad orients toward the prey, (B) the mouth opens and the tongue is protracted, (C) the tongue is retracted, and (D) the mouth closes. (Right) After deafferentation, (E) toads orient normally, but (F-H) the mouth fails to open during the feeding attempt. Electromyographic traces on the left show activity in the jaw muscles depressor mandibulae (DM) and levator mandibulae (LM) in intact toads before deafferentation. DM reaches its peak activity approximately 90 msec before LM. Right traces show that DM and LM reach their peak activity simultaneously after deafferentation. These results demonstrate that hypoglossal afferents modulate the phase of activity in the jaw muscles during feeding.
138
Kiisa C. N i s h i k a w a
Before
After
would be absent, or would have a different function, in mechanical pullers and hydrostatic elongators because these frogs lack synchronous acceleration of tongue and jaws, as well as a straight trajectory of the tongue tip from mouth to prey (Table 5.1). Studies on the evolution of hypoglossal afferents are described in more detail later (Section V,C). C. Interactions b e t w e e n Tongue Afferents and Visual Input
FIGURE 5.13. The function of hypoglossal afferents in the leopard frog, Rana pipiens. Top portion: (A) intact frogs use tongue prehension to capture small prey, (B) intact frogs use jaw prehension to capture large prey, (C) when deafferented frogs attempt to feed on small prey, the mouth fails to open; but (D) when deafferented frogs attempt to feed on large prey, the mouth opens normally. Bottom portion: The four panels on the left show intact frogs feeding on earthworm pieces that vary in size from 1.5 to 4.5 cm (size of prey is indicated by numbers in upper right corner of each picture). The frogs always exhibit tongue prehension for 1.5-cm prey and jaw prehension for 2.0-cm and larger prey. The five panels on the right show deafferented frogs feeding on earthworm pieces. After deafferentation, the mouth always fails to open when the frogs feed on 1.5-cm prey and the mouth always opens when the frogs feed on 2.5-cm and larger prey. However, when the frogs feed on 2.0-cm prey, they alternate randomly between opening and not opening the mouth. This result demonstrates that hypoglossal afferents in the tongue influence motor program choice.
before the onset of mouth opening and may result from stimulation of tongue mechanoreceptors by retraction of the hyoid during the preparatory stage of feeding. In intact frogs and toads, we hypothesize that this afferent signal coordinates the simultaneous acceleration of the tongue and mandibles, which ensures that the lingual trajectory will fall on a straight path from mouth to prey (Nishikawa and Gans, 1996). If this hypothesis is correct, then we would expect that these afferents
It is somewhat surprising that afferent input from the tongue interacts with visual input in controlling prey capture movements. In R. pipiens, the modulatory effect of tongue afferents depends on attributes of the visual stimulus that is presented to elicit feeding (Anderson and Nishikawa, 1993). When presented with small prey, deafferented frogs attempt to use tongue prehension to capture the prey and the mouth remains closed, as noted earlier. However, when the same deafferented frogs are presented with large prey, they use jaw prehension to capture the prey and their mouths open normally (Fig. 5.13; Anderson and Nishikawa, 1993). These results demonstrate that the modulatory effect of hypoglossal afferents on feeding movements is itself modulated by visual input, and they suggest that visual input has a gating effect on the hypoglossal afferents. Hypoglossal afferents also interact with visual input during motor program choice in R. pipiens (Anderson and Nishikawa, 1996). As mentioned previously, intact adult frogs use tongue prehension to capture 1.5-cm pieces of earthworm, but switch to jaw prehension for 2.0-cm and larger prey. Based on the results of the hypoglossal deafferentation experiments described earlier, we would expect that the mouth would never open for 1.5-cm prey and always open for 2.0-cm and larger prey. When the tongue afferents are inactivated, the mouth never opens for 1.5-cm prey and always opens for 2.5-cm and larger prey, as expected. However, the frogs alternate randomly between tongue prehension and jaw prehension when 2.0-cm prey are offered (Fig. 5.13). Thus, the ability to choose between motor programs for tongue prehension and jaw prehension is impaired by hypoglossal transection. In R. pipiens, hypoglossal afferents not only subserve typical motor control functions, such as modulating motor output, but also have become involved in behavioral decision making. Neural network modeling is being used in conjunction with neuroanatomical tracing studies to understand the premotor circuits that underlie interactions between vision and proprioception in the control of feeding movements (Corbacho et al, 1996).
5. Feeding in Frogs V. EVOLUTION OF THE FEEDING APPARATUS A major goal of comparative and functional morphology is to understand the evolution of complex functional systems. This goal can be achieved using cladistic analysis, in which a phylogeny is obtained for the group, data are collected on characters of interest, and the distribution of these characters is mapped onto the phylogeny. A cladistic analysis permits identification of the ancestral condition for the character(s) of interest, which represents the starting point for evolutionary diversification. This type of analysis also locates transitions in character state on the phylogeny that indicate when and how the condition has changed. Studies in my laboratory have used a cladistic approach to study the evolution of mechanisms of tongue protraction and neural control of feeding behavior in anurans, and the results of these studies are described next. A. Evolutionary Transitions in Mechanisms of Tongue Protraction A phylogenetic analysis (Fig. 5.14) of feeding behavior in 148 species of frogs representing 15 of approximately 330 described genera (23%) and 20 of 27 families (74%) was conducted by mapping characters derived from high-speed video motion analysis (Nishikawa ei al., 1992) onto the most recent hypothesis of frog phylogeny (Ford and Cannatella, 1993). The results of this analysis show that all archaeobatrachian frogs that have been studied {Ascaphus, Ascaphidae; Leiopelma, Leiopelmatidae; Bomhina, Bombinatoridae; and Discoglossus, Discoglossidae) possess mechanical pulling tongues that shorten during protraction (Fig. 5.14). Some members of the clades Mesobatrachia (e.g., Pelobates) and Neobatrachia (e.g., hyline hylids) also possess short tongues (Fig. 5.14). The most parsimonious interpretation of the observed character distribution across taxa is that anurans ancestrally possess tongues that shorten during protraction and therefore can be protracted only a few millimeters beyond the tips of the mandibles. Other ancestral characteristics include (1) downward bending of the lower jaw during mouth opening and (2) use of the whole body as a projectile, which brings the short tongue into contact with the prey (Nishikawa et al, 1992). Tongues that elongate during protraction have evolved many times independently from mechanical pulling tongues that shorten during protraction (Fig. 5.14). It is difficult to estimate the exact number of independent evolutionary events due to the lack of resolution of family relationships (Cannatella et al, 1992).
139
However, at least seven lineages possess elongating tongues (Fig. 5.14). Among mesobatrachians, the most parsimonious hypothesis is that mechanical pulling is ancestral and inertial elongation has evolved once in the common ancestor of Megophrys montana and Leptobrachium. Megophrys is the only known genus in which the constituent species differ in tongue length, with M. aceras using mechanical pulling and M. montana using inertial elongation to protract the tongue (Fig. 5.14). Among neobatrachians, mechanical pulling also appears to be the ancestral method of tongue protraction based on parsimony analysis. The most parsimonious hypothesis is that mechanical pulling is ancestral for the group and that elongating tongues evolved independently in as many as six different lineages: (1) Ceratophrys and Chacophrys, (2) Physalaemus and Fleurodema, (3) Eleutherodactylus and Syrrhophus, (4) the phyllomedusine hylids, (5) the bufonids, and (6) the ranoids (Nishikawa et al, 1992). The most variable families are the Leptodactylidae (which is probably not monophyletic; Ford and Cannatella, 1993) and Hylidae, both of which possess species with both long and short tongues. Most frogs with elongating tongues share many derived behavior patterns, particularly a reduction in lunge length (Nishikawa et al, 1992). The phyllomedusine hylids are exceptional in retaining a long lunge despite possessing a long tongue (Gray and Nishikawa, 1995). However, there are many differences among long-tongued lineages as well. Long tongues appear to have evolved to enhance crypsis in Bufo (Gray, 1997), for catching rapidly moving prey in the phyllomedusines (Gray and Nishikawa, 1995), for capturing large prey in Megophrys (Emerson, 1985; Gans et al, 1991), and for capturing prey accurately in H. marmoratum (Ritter and Nishikawa, 1995). Among species with elongating tongues, all use inertial elongation to protract the tongue, except members of the families Microhylidae and Hemisotidae, all species of which use hydrostatic elongation. Given the phylogeny of Ford and Cannatella (1993), it appears that mechanical pulling is the ancestral tongue protraction mechanism, that inertial elongation evolved up to seven times independently from mechanical pulling, and that hydrostatic elongation evolved once or maybe twice from inertial elongation within the ranoid clade (Fig. 5.14). This hypothesis is somewhat counterintuitive, however, because hydrostatic elongation is the slowest mechanism of tongue protraction whereas inertial elongation is the fastest (Table 5.1). However, these tongue types appear to represent a morphocline in terms of the amount of connective tissue in the tongue that would restrict elongation (see next section; Webster, 1996).
140
Kiisa C. N i s h i k a w a Ascaphus Leiopelma Bombina Discoglossus Scaphiopus Spea Pelobates Pelodytidae Megophrys aceras Megophrys montana Leptobrachium Rhinophrynidae Pipidae Limnodynastes Myobatrachinae Sooglossidae Heleophrynidae Ceratophrys Chacophiys Lepidobatrachus Leptodactylus Pleurodema Physalaemus Telmatobius Caudiverbera Eleutherodactylus Syrrhophus ] Brachycephalidae Bufonidae Rhinodermatidae Centrolenidae Pseudidae Hylinae
Protraction Mechanism • • •
mechanical pulling unknown none inertial elongation hydrostatic elongation I '
Hemiphractinae Pelodryadinae Phyllomedusinae Ranidae Arthroleptidae Hyperoliidae I Rhacophoridae Dendrobatidae Hemisotidae Microhylidae
F I G U R E 5.14. Cladogram illustrating the evolution of tongue protraction mechanisms among frogs. For mechanisms of tongue protraction, mechanical pulling is the ancestral condition. Inertial and hydrostatic elongation are derived conditions that evolved several times independently among frogs.
B. Morphological Correlates of Tongue Protraction Mechanisms It is interesting to ask what morphological changes are responsible for the profound functional changes in
the biomechanics of the feeding apparatus that have occurred during anuran evolution (Table 5.1). Comparative studies have demonstrated considerable diversity in the feeding behavior of frogs (compare Figs. 5.8, 5.9 and 5.10; Table 5.1). Anuran species vary
141
5. Feeding in Frogs considerably in the length of the fully protracted tongue, which ranges from less than 10% of jaw length in semiaquatic species such as Lepidobatrachus llanensis (unpublished observation) to nearly 300% of jaw length in H. marmoratum (Ritter and Nishikawa, 1995). Even among frogs with relatively long tongues, species vary in the speed of tongue protraction from 24 cm/sec in the slowest hydrostatic elongators to 270 cm/sec in the fastest inertial elongators (Table 5.1). In anurans, these functional changes appear to have involved relatively small quantitative and qualitative changes in the anatomy and physiology of the muscles of the feeding apparatus. The transition from mechanical pulling to inertial elongation appears to have involved (1) a decrease in relative tongue mass; (2) an increase in the relative length of muscle fibers in both the M. genioglossus and M. hyoglossus; (3) a change in insertion of the M. hyoglossus from the anterior to the posterior part of the tongue pad (Horton, 1982); (4) a decrease in the amount of connective tissue (as measured subjectively from scanning electron micrography of muscles digested with NaOH), which appears to restrict tongue elongation in mechanical pullers (Webster, 1996); and (5) an increase in contraction velocity compared to mechanical pullers (Peters and Nishikawa, 1999). The transition from inertial to hydrostatic elongation involved only (1) the addition of a new dorsoventral compartment in the M. genioglossus, which elongates the tongue when it contracts; and (2) a further decrease in the amount of connective tissue in the tongue. Peters and Nishikawa (1999) completed a study of the contractile properties of the tongue protractor and retractor muscles of mechanical pullers, inertial elongators, and hydrostatic elongators. The studies show that although there are some differences (e.g., in contraction and half relaxation time of the muscles), the contractile properties of the muscles alone cannot account for differences in function (i.e., differences in length or velocity of tongue protraction). A major difference among frog tongues is in the amount of connective tissue they contain, as well as the orientation of collagen fibers within the tongue (Webster, 1996). Mechanical pullers have the most connective tissue, inertial elongators are intermediate, and hydrostatic elongators have the least. Furthermore, the orientation of collagen fibers also differs among taxa. Mechanical pullers have the largest proportion of fibers with low orientation angles (<55°) relative to the long axis of the tongue, inertial elongators are intermediate, and hydrostatic elongators have the most fibers with high orientation angles (>55°). In mechanical pullers, these parallel connective tissue fibers would
actively resist elongation of the tongue. In H. marmoratum, most of the collagen fibers are oriented nearly perpendicular to the long axis of the tongue (modal angle = 80°), where they do not resist tongue elongation but instead resist increases in tongue diameter (Nishikawa et fl/., 1999). The differences in tongue morphology among mechanical pullers, inertial elongators, and hydrostatic elongators are so small and subtle that it has not proved possible to predict differences in function on the basis of anatomical differences. Ironically, the functions of most of the larger and less subtle anatomical differences that have been described among anuran species remain obscure. Based on the preliminary data described earlier, it seems quite likely that relatively small quantitative differences in the amount and orientation of connective tissue in the tongue, via their effects on passive properties, may have a greater effect on tongue function than anatomical and physiological differences in tongue muscles among anurans. C. Evolution of Tongue Afferents We next conducted a cladistic analysis of the evolution of hypoglossal afferents among anurans (Nishikawa et ah, 1993). For the cladistic analysis, 57 species of anurans representing 39 genera and 15 families were filmed while feeding before and after bilateral transection of the hypoglossal nerves (Fig. 5.15). Results of this analysis show that hypoglossal afferents that modulate the phase of activity in the jaw muscles are absent in all mechanical pullers and all hydrostatic elongators (Fig. 5.15). Some but not all inertial elongators possess hypoglossal afferents that modulate the phase of activity in the jaw muscles, and it appears that these afferents have evolved convergently at least four times independently: (1) in the leptodactyline leptodactylids Pleurodema and Physalaemus; (2) in the phyllomedusine hylids; (3) in the bufonids; and (4) in ranoids except for hemisotids and microhylids, in which hypoglossal afferents appear to have been lost (Fig. 5.15). Although the function of these afferents is very similar across species at the behavioral level, both cladistic analyses and comparative neuroanatomical studies show that the circuits are convergent rather than homologous in the different anuran lineages (Nishikawa et al, 1993; Anderson and Nishikawa, 1997). Frogs ancestrally lack afferents in the hypoglossal nerve, as do most vertebrates (Ariens-Kappers et ah, 1936). In toads of the family Bufonidae, sensory fibers from the glossopharyngeal nerve have invaded the tongue via the hypoglossal nerve and these fibers ascend to higher brain centers in the tractus solitarius (Nishikawa et al., 1993). In frogs of the family Ranidae, large myelinated
142
Kiisa C. N i s h i k a w a Ascaphus Leiopelma Bombina Discoglossus Scaphiopus Spea Pelobates Pelodytidae Megophrys aceras Megophrys montana Leptobrachium Rhinophrynidae Pipidae Limnodynastes ] Myobatrachinae ] Sooglossidae ] Heleophrynidae Ceratophrys ] Chacophrys ] Lepidobatrachus Leptodactylus ] Pleurodema ] Physalaemus ] Telmatobius ] Caudiverbera ] Eleutherodactylus ] Syrrhophus ] Brachycephalidae ] Bufonidae ] Rhinodermatidae ] Centrolenidae ] Pseudidae
Hypoglossal Afferents
• • •
absent J present ] unknown
Hylinae Hemiphractinae Pelodryadinae Phyllomedusinae Ranidae Arthroleptidae Hyperoliidae Rhacophoridae Dendrobatidae Hemisotidae Microhylidae
F I G U R E 5.15. Cladogram illustrating the evolution of hypoglossal afferents among frogs. Hypoglossal afferents are ancestrally absent among frogs. These afferents have evolved several times independently, but only in frogs that use inertial elongation to protract the tongue. These afferents appear to have been lost in hydrostatic elongators (families Hemisotidae and Microhylidae).
afferents of the most anterior cervical spinal nerve have invaded the tongue via the hypoglossal nerve (Anderson and Nishikawa, 1997). These fibers ascend and descend in the dorsomedial funiculus and project to the granular layer of the cerebellum and the medial reticu-
lar formation. The source of hypoglossal afferents remains unknown in leptodactylids and hylids. During invasions of new territory in Rana and Bufo, sensory fibers have changed their peripheral pathways as well as their central connections, although the location
143
5. Feeding in Frogs of their cell bodies and the basic class of cutaneous mechanoreceptors that they innervate appear to have been conserved. Current studies are exploring how these convergent neural circuits differ physiologically. D . Evolutionary Transitions in Mechanisms of Neural Control An important question that remains largely unanswered concerning the function and evolution of tetrapod feeding systems is whether evolution of the morphology and mechanics of the feeding apparatus affect mechanisms of neural control. A related question simply asks whether there is variation in the neural control of the feeding apparatus among tetrapods and, if so, what is the nature and significance of the variation. Because anuran species exhibit variation in the morphology and mechanics of the feeding apparatus, they offer a unique opportunity to investigate these questions. Motor control differs in several ways among frog species that use different mechanisms to protract their tongues. The first difference is that inertial elongators use only feedforward, open loop control to coordinate jaw and tongue movements. In inertial elongators, there is no opportunity for on-line, feedback correction after the tongue is launched because tongue protraction is ballistic (Nishikawa and Cans, 1996). In contrast, mechanical pullers and hydrostatic elongators can rely on both feedforward and feedback control of tongue movements because there is no inertial stage of tongue elongation (Table 5.1). A second difference in motor control is that, in inertial elongators, accurate placement of the tongue on the prey requires precise coordination of the extremely rapid, simultaneous movements of the tongue and jaws. Precise coordination is not necessary in mechanical pullers because the movement of the short tongue pad relative to the lower jaw is restricted to a few millimeters so that the tongue pad will always end up in nearly the same location as the tips of the mandibles. Precise coordination is unnecessary in hydrostatic elongators because the tongue is moved slowly and can be moved independently in three dimensions relative to the head. In inertial elongators, tongue afferents that are innervated by the hypoglossal nerve have evolved convergently in at least four independent lineages for the precise coordination of tongue and jaw movements. Small changes in the central and peripheral connections of cranial (Bufonidae) or spinal (Ranidae) mechanosensory afferents have led to the emergence of novel functions in coordinating feeding behavior in inertial elongators, including modulating the phase of activity in jaw muscles and influencing motor program choice during feeding.
Finally, the three-dimensional aiming ability of hydrostatic elongators implies several changes in neural control that have yet to be investigated. For both threedimensional aiming and high accuracy, we would expect that motor units should be smaller and more numerous in hydrostatic elongators than in other species. VL C O N C L U S I O N S In summary, the feeding behavior of anurans has proved to be an interesting model system for understanding the process of functional diversification during evolution. Comparative studies have demonstrated that frogs exhibit at least three different mechanisms for protracting their tongues. These are mechanical pulling, inertial elongation, and hydrostatic elongation. These mechanisms differ in the extent of tongue movement relative to the head, in the velocity, acceleration, and trajectories of tongue movements, in aiming ability, and in the accuracy of prey capture (Table 5.1). Morphologically, the feeding apparatus of frogs is rather homogeneous. All species possess similar sets of muscles and bones, although there is variation in the presence of compartments within muscles, for example, the dorsoventral compartment of the M. genioglossus in hydrostatic elongators, which has a major effect on tongue function. There are also differences among species in the contractile properties of tongue muscles as well as in the connective tissue that is responsible for transmitting the forces produced by the tongue muscles during feeding. In general, numerous small qualitative and quantitative morphological differences among species appear to be responsible for the rather large differences in the biomechanics of tongue protraction across anuran species. Species that differ in the biomechanics of tongue protraction also differ in mechanisms of neural control. Species differ in the relative importance of feedforward vs feedback control, in the requirement for precise coordination of multijoint movements, in the afferents that are used to coordinate these movements, and finally in aiming ability and accuracy of prey capture. There has been considerable rewiring of cranial or cervical spinal afferents in inertial elongators to provide a mechanism for precise coordination of tongue and jaw movements. In summary, these studies show that small changes in the anatomy of the feeding apparatus may lead to large changes in biomechanics and that small changes in neuroanatomy may lead to large changes in sensorimotor coordination. Potential precursors for novel sensory pathways appear to be prevalent and may change readily in response to natural selection. In anurans, high levels of neural and behavioral evolution
144
Kiisa C. N i s h i k a w a
are found, even among closely related species and in parts of the brain usually thought to be evolutionarily conservative. A major implication of these studies on the evolution of frog tongues is that, for any given species, the neural networks that subserve sensorimotor coordination are fine-tuned to its particular morphology and environment (Nishikawa, 1997). VII. CURRENT A N D FUTURE DIRECTIONS Numerous questions remain to be studied concerning the biomechanics and neural control of prey capture in anurans. In terms of biomechanics, a few differences in morphology and biomechanics of the feeding apparatus have been quantified in just a few species, and for many anatomical differences, their functional significance remains to be studied. Current research in my laboratory is focused on using biomechanical modeling to investigate the diversification of function in frog tongues (Nishikawa et al, 1997). With Eric Mallett and Gary Yamaguchi at Arizona State University, we have developed a planar, forward dynamic, multijoint, rigid body model of the anuran tongue to study the dynamics of tongue protraction and retraction during prey capture. The model contains four degrees of freedom, including the lower jaw (fixed length) and three segments of variable length in the tongue. In the model, the M. genioglossus and M. hyoglossus are used to actuate the tongue, while impulsive joint torques are applied at the jaw joint to open or close the mouth. Joint torques are obtained from a muscle model that incorporates nonlinear springs and dampers to simulate the forcelength-velocity characteristics of the tongue muscles. The model takes anatomical data on the size and shape of the feeding apparatus, as well as contractile properties of the tongue muscles, as input. Simulations involve varying the pattern of muscle activation to find the optimal pattern for each mechanism of tongue protraction. This model is being used to explore the following questions: (1) how much complexity must be incorporated into the model in order for it to predict movement trajectories accurately? (i.e., is a four degree of freedom model sufficient, or are more segments necessary? is the mentomeckelian joint needed? is a movable hyoid needed?) (2) how will differences in size and shape affect movement kinematics? (3) do changes in musculoskeletal design represent suites that together may enhance a particular aspect of kinematics? (4) can the same pattern of muscle activation be used for all species? and (5) how much can morphology change before a different muscle activation strategy becomes desirable?
In terms of neural control, differences in the neuroanatomy of proprioceptive pathways have been described in only two anuran species, and the neurophysiology of convergent afferent pathways remains to be explored. The anatomy and physiology of hypoglossal mechanosensory receptors also require further study. Finally, hypoglossal afferents are only one of many proprioceptive pathways that coordinate feeding movements, and the modulatory effects of additional afferent systems on feeding behavior, such as the lingual withdrawal reflex (Matsushima et ah, 1986, 1987,1988), remain to be described. Our studies of the neural control of prey capture in anurans have barely scratched the surface of understanding the evolutionary relationships among morphology, biomechanics, and neural control of movement. It is hoped that our attempts to understand the evolution of prey capture in frogs will stimulate functional morphologists to undertake similar studies in other groups of animals. Acknowledgments I thank the many undergraduate and graduate students from my laboratory who have contributed to this work and who made the process of doing it both fun and exciting. This research was supported by the National Science Foundation (IBN-9507497) and the National Instihites of Health (S06-GM08215). I thank James Birch, Lucile Gray, James O'Reilly, Philip Service, Steve Wainwright, Kentwood Wells, and Erika Wiltenmuth for constructive comments on the manuscript. Drawings were made by Tad Theimer (Fig. 5.2) and Virginia Coryell (Fig. 5.3), and Robyn O'Reilly prepared the figures on the computer.
References Anderson, C. W. (1993) Modulation of feeding behavior in response to prey type in the frog Rana pipiens. J. Exp. Biol. 179:1-11. Anderson, C. W., and K. C. Nishikawa (1993) A prey-type dependent hypoglossal feedback system in the frog, Rana pipiens. Brain Behav.Evol. 42:189-196. Anderson, C. W., and K. C. Nishikawa (1996) The roles of visual and proprioceptive information during motor program choice in frogs. J. Comp. Physiol. A 179:753-762. Anderson, C. W., and K. C. Nishikawa (1997) The functional anatomy and evolution of hypoglossal afferents in the leopard frog, Rana pipiens. Brian Res. 771:285-291. Ariens-Kappers, C. U , G. C. Huber, and E. C. Crosby (1936) The Comparative Anatomy of the Nervous System of Vertebrates, including Man. MacMillan, New York. Avila, V. L., and P. G. Frye (1977) Feeding behavior in the African clawed frog {Xenopus laevis Daudin). Herpetologica 33:152-161. Bemis, W. E., K. Schwenk, and M. H. Wake (1983) Morphology and function of the feeding apparatus in Dermophis mexicanus (Amphibia: Gymnophiona). Zool. J. Linn. Soc. 77:75-96. Bermejo, R., and H. P. Ziegler (1989) Prehension in the pigeon. II. Kinematic analysis. Exp. Brain Res. 75:577-585. Berry, P., and J. Bullock (1962) The food of the common Malayan toad, Bufo melanostictus. Copeia 1962:736-741.
5. F e e d i n g in Frogs Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. R Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Bout, R., and H. P Ziegler (1994) Jaw muscle (EMG) activity and amplitude scaling of jaw movements during eating in pigeon (Columba livia). J. Comp. Physiol. A 174:433-442. Caldwell, J. P. (1996) The evolution of myrmecophagy and its correlates in poison frogs (family Dendrobatidae). J. Zool. London 240: 75-101. Cannatella, D. C. (1985) A phytogeny of primitive frogs (archaeobatrachians). University Microfilms, Ann Arbor, MBA86-08380. Cannatella, D. C , and L. Trueb (1988) Evolution of pipoid frogs: intergeneric relationships of the aquatic frog family Pipidae (Anura). Zool. J. Linn. Soc.94:1-38. Cannatella, D. C , K. C. Nishikawa, and J. C. O'Reilly (1992) Patterns of convergence in tongue projection in frogs. Am. Zool. 32:134A. Clarke, R. D. (1974) Food habits of toads, genus Bufo (Amphibia: Bufonidae). Am. Midi. Natur. 91:140-147. Comer C , and P. Grobstein (1981) Factually elicited prey acquisition behavior in the frog, Rana pipiens, and a comparison with visually elicited behavior. J. Comp. Physiol. 142:141-150. Corbacho, R J., K. C. Nishikawa, J.-S. Liaw, and M. A. Arbib (1996) An expectation-based model of adaptable and flexible prey-catching in anurans. Soc. Neurosci. Abstr. 22:164. Das, I. (1995) Comparative morphology of the gastrointestinal tract in relation to diet in frogs from a locality in South India. Amphibia-Reptilia 16:289-293. Das, I. (1996) Folivory and seasonal changes in diet in Rana hexadactyla (Anura: Ranidae). J. Zool. London 238:785-794. Das, I., and M. Coe (1994) Dental morphology and diet in anuran amphibians from South India. J. Zool. London 233:417-427. da Silva, H. R., C. D. Biottopereira, and U. Caramaschi (1989) Frugivory and seed dispersal by Hyla truncata, a neotropical frog. Copeia 1989:781-783. Deban, S. M., and K. C. Nishikawa (1992) The kinematics of prey capture and the mechanism of tongue protraction in the green tree frog, Hyla cinerea. J. Exp. Biol. 170:235-256. De Jongh, H. J. (1968) Functional morphology of the jaw apparatus of larval and metamorphosing Rana temporaria L. Neth. J. Zool. 18:1-103. De Jongh, H. J., and C. Cans (1969) On the mechanism of respiration in the bullfrog, Rana catesbeiana: a reassessment. J. Morphol. 127: 259-290. Delcomyn, F. (1980) Neural basis of rhythmic behavior in animals. Science 210:492-498. Dole, J. W., B. B. Rose, and K. H. Tachiki (1981) Western toads {Bufo boreas) learn odor of prey insects. Herpetologica 37:63-68. Drewes, R., and B. Roth (1981) Snail eating frogs from the Ethiopian highlands: a new anuran speciaUzation. Zool. J. Linn. Soc. 73: 267-287. Duellman, W. E., and L. Trueb (1986) Biology of Amphibians. McGraw Hill, New York. Emerson, S. B. (1976a) Burrowing in frogs. J. Morphol. 149:437-458. Emerson, S. B. (1976b) A preliminary report on the superficial throat musculature of the Microhylidae and its possible role in tongue action. Copeia 1976:546-551. Emerson, S. B. (1977) Movement of the hyoid in frogs during feeding. Am. J. Anat. 149:115-120. Emerson, S. B. (1985) Skull shape in frogs: correlations with diet. Herpetologica 41:177-188. Emerson, S. B., and D. M. Bramble (1993) Scaling, allometry, and skull design. Pp. 384-421. In: The Skull Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Emerson, S. B., and H. Voris (1992) Competing explanatins for sexual dimorphism in a voiceless Bornean frog. Funct. Ecol. 6:654-660.
145
Emerson, S. B., H. W. Greene, and E. L. Charnov (1994) AUometric aspects of predator-prey interactions. Pp. 123-139. In: Ecological Morphology. P. C. Wainwright and S. M. Reilly (eds.), Univ. of Chicago Press, Chicago. Ewert, J.-P. (1985) Concepts in vertebrate neuroethology. Anim. Behav.33:l-29. Ewert, J.-P. (1987) Neuroethology of releasing mechanisms: prey catching in toads. Behav. Brain Sci. 10:337-405. Flanders, M., S. I. H. Tillery, and J. F Soechting (1992) Early stages in a sensorimotor transformation. Behav. Brain Sci. 15:309-362. Ford, L., and D. C. Cannatella (1993) The major clades of frogs. Herpetologica 7:94-117. Gans, C. (1952) The tongue protrusion mechanism in Rana catesbeiana. Am.Zool.2:524A. Gans, C. (1961) The bullfrog and its prey. Nat. Hist. 70:26-37. Gans, C. (1962) The tongue protrusion mechanism in Rana catesbeiana. Am. Zool. 2:524. [abstract] Gans, C. (1967) The chameleon. Nat. Hist. 76:52-59. Gans, C , and G. C. Gorniak (1982a) How does the toad flip its tongue? Test of two hypotheses. Science 216:1335-1337. Gans, C , and G. C. Gorniak (1982b) Functional morphology of lingual protrusion in marine toads {Bufo marinus). Am. J. Anat. 163: 195-222. Gans, C , and R. Pyles (1983) Narial closure in toads: which muscles? Respiration Physiol. 53:215-223. Gans, C , K. C. Nishikawa, and D. C. Cannatella (1991) The frogMegophrys montana: specialist in large prey. Am. Zool. 31:52A. Gaupp, E. (1896) A. Ecker's und R. Wiedersheim's Anatomie des Frosches. Friedrich Vieweg und Sohn, Braunschweg. Gaupp, E. (1901) Uber den Muskelmechanismus bei den Bewegungen der Froschzunge. Anat. Anz. 19:385-396. GottHeb, G. L., D. M. Corcos, and G. C. Agarwal (1989) Strategies for the control of voluntary movements with one mechanical degree of freedom. Brain Behav. Sci. 12:189-210. Gray, L. A. (1997) Tongue Morphology, Feeding Behavior and Feeding Ecology in Anurans. Ph. D. Dissertation, Northern Arizona University. Gray, L. A., and K. C. Nishikawa (1995) Feeding kinematics of phyllomedusine tree frogs. J. Exp. Biol. 198:457-463. Gray, L. A., J. C. O'Reilly, and K. C. Nishikawa (1997) Evolution of forelimb movement patterns for prey manipulation in anurans. J. Exp. Zool. 277:417-424. Griffiths, I. (1963) The phytogeny of the Salientia. Biol. Rev. 38:241292. Grillner, S. (1985) Neurobiological bases of rhythmic motor acts in vertebrates. Science 228:143-149. Grillner, S., J. T. Buchanan, P Wallen, and L. Brodin (1988) Neural control of locomotion in lower vertebrates: from behavior to ionic mechanisms. Pp. 1-40. In: Neural Control of Rhythmic Movements in Vertebrates. A. H. Cohen, S. Rossignol, and S. Grillner (eds.). Wiley, New York. Giinther, A. C. L. G. (1859) Catalogue of the Batrachia Salientia in the Collection of the British Museum. British Museum, London. Harris-Warrick, R. (1988) Chemical modulation of central pattern generators. Pp. 285-332. In: Neural control of rhythmic movements in vertebrates. A. H. Cohen, S. Rossignol. and S. Grillner (eds.), Wiley, New York. Hass, A. (1995) Cranial features of dendrobatid larvae (Amphibia: Anura: Dendrobatidae). J. Morphol. 224:222-241. Hay, J. M., I. Ruvinsky, S. B. Hedges, and L. R. Maxson (1995) Phylogenetic relationships of amphibian families inferred from DNA sequences of mitochondrial 12S and 16S ribosomal RNA genes. Mol. Biol. Evol. 12:928-937. Hillis, D. M., L. K. Ammerman, M. T. Dixon, and R. O. de Sa (1993) Ribosomal DNA and the phytogeny of frogs. Herpetologica 7: 118-131.
146
Kiisa C. N i s h i k a w a
Horton, P. (1982) Diversity and systematic significance of anuran tongue musculature. Copeia 1982:595-602. Inger, R., and H. Marx (1961) The food of amphibians. Expl. du Pare National de I'Upemba, Fasc. 64:1-86. Ingle, D. J. (1983) Brain mechanisms of visual localization by frogs and toads. Pp. 177-226. In: Advances in Vertebrate Neuroethology. J.-P. Ewert, R. R. Capranica, and D. J. Ingle (eds.). Plenum Press, New York. Jaeger, R. G. (1976) Possible prey-call window in anuran auditory perception. Copeia 1976:833-834. Jaeger, J. C , and K. C. Nishikawa (1993) The feeding behavior and tongue aiming ability of the African snake-necked frog, Phrynomerus bifasciatus. J. AZ-NV Acad. Sci. 28:59. Kaess W., and F. Kaess (1960) Perception of apparent motion in the common toad. Science 132:953. Knoepffler, L.-P (1976) Food habits of Aubria subsigillata in Gaboon (Anura: Ranidae). Zool. Afr. 11:369-371. Larsen, J. H., Jr., and D. J. Guthrie (1975) The feeding mechanism of terrestrial tiger salamanders {Ambystoma tigrinum melanostictum Baird). J. Morph. 147:137-154. Lauder, G. V., and S. M. Reilly (1994) Amphibian feeding behavior: Comparative biomechanics and evolution. Pp. 163-195. In: Advances in Comparative and Environmental Physiology, Vol. 18, Springer Verlag, Berlin. Lettvin, J. Y., H. R. Maturana, W. S. McCulloch, and W. H. Pitts (1959) What the frog's eye tells the frog's brain. Proc. Inst. Radio Engrs. 47:1940-1951. Lombard, R. E., and D. B. Wake (1976) Tongue evolution in lungless salamanders, family Plethdontidae. I. Introduction, theory, and a general model of dynamics. J. Morph. 148:265-286. Lombard, R. E., and D. B. Wake (1977) Tongue evolution in the lungless salamanders, family Plethodontidae. II. Function and evolutionary diversity. J. Morph. 153:39-80. Lombard, R. E., and D. B. Wake (1987) Tongue evolution in the lungless salamanders, family Plethodontidae. IV. Phylogeny of plethodontid salamanders and the evolution of feeding dynamics. Syst. Zool. 35:532-551. Magimel-Pelonnier, O. (1924) La langue des Amphibiens. These, Fac. Sci. Paris. A. Saugnac et E. Provillard, Bordeaux. Martin, W., and C. Gans (1972) Muscular control of the vocal tract during release signaling in the toad, Bufo valliceps. J. Morphol. 137:1-28. Martof, B. S. (1962) Some observations on the feeding of Fowler's toad. Copeia 1962:439. Matsushima, T., M. Satou, and K. Ueda (1985) An electromyographic analysis of electrically-evoked prey-catching behavior by means of stimuli applied to the optic tectum in the Japanese toad. Neurosci. Res. 3:154-161. Matsushima, T., M. Satou, and K. Ueda (1986) Glossopharyngeal and tectal influences on tongue-muscle motoneurons in the Japanese toad. Brain Res. 365:198-203. Matsushima, T., M. Satou, and K. Ueda (1987) Direct contact between glossopharyngeal afferent terminals and hypoglossal motoneurons revealed by double labeling with cobaltic lysine and horseradish peroxidase in the Japanese toad. Neurosci. Lett. 80:241245. Matsushima, T., M. Satou, and K. Ueda (1988) Neuronal pathways for the lingual reflex in the Japanese toad. Comp. Physiol. A 164: 173-193. Matsushima, T., M. Satou, and K. Ueda (1989) Medullary reticular neurons in the Japanese toad: morphologies and excitatory inputs from the optic tectum. J. Comp. Physiol. A. 166:7-22. Meyers, J. J., K. C. Nishikawa, and J. C. O'Reilly (1996) Tongue aiming in the microhylid frog Phrynomerus bifasciatus. Am. Zool.36:81 A. Nishikawa, K. C. (1997) Emergence of novel functions during brain evolution. Bioscience 47:341-354.
Nishikawa, K. C , and D. C. Carmatella (1991) Kinematics of prey capture in the tailed frog, Ascaphus truei. Zool. J. Linn. Soc. 103: 289-307. Nishikawa, K. C , and C. Gans (1992) The role of hypoglossal sensory feedback during feeding in the marine toad, Bufo marinus. J. Exp. Zool. 264:245-252. Nishikawa, K. C , and C. Gans (1996) Mechanisms of prey capture and narial closure in the marine toad, Bufo marinus. ]. Exp. Biol. 199:2511-2529. Nishikawa, K. C , and G. Roth (1991) The mechanism of tongue protraction during prey capture in the frog Discoglossus pictus. J. Exp. BioL 159:217-234. Nishikawa, K. C , C. Anderson, S. Deban, and J. O'Reilly (1992) The evolution of neural circuits controlling feeding behavior in frogs. Brain Behav. Evol. 40:125-140. Nishikawa, K. C , J. C. O'Reilly, B. W. P Sasongko, and C. W. Anderson (1993) Convergent evolution of hypoglossal afferents that influence jaw muscle activity in frogs. Soc. Neurosci. Abstr. 19:161. Nishikawa, K. C , W. M. Kier and K. K. Smith (1999) Morphology and mechanics of tongue movement in the African pig-nosed frog (Hemisus marmoratum): A muscular hydrostatic model. J. Exp. BioL 202:771-780. Nishikawa, K. C , E. S. Mallett, and G. T. Yamaguchi (1997) A biomechanical model for the simulation of prey capture in toads. Soc. Neurosci. Abstr. 23:2135. Nussbaum, R. A. (1983) The evolution of a unique dual jaw closing mechanism in caecilians (Amphibia: Gymnophiona) and its bearing on caecilian ancestry. J. Zool. (London) 199:545-554. O'Reilly, J. C , K. C. Nishikawa, and S. M. Deban (2000) Derived life history characteristics constrain the evolution of aquatic feeding behavior in amphibians. Zoology: Analysis of Complex Systems: In press. O'Reilly, S. R., and K. C. Nishikawa (1995) Mechanism of tongue protraction during prey capture in the spadefoot toad Spea multiplicata (Anura: Pelobatidae). J. Exp. Zool. 273:282-296. Ozeti, N., and D. B. Wake (1969) The morphology and evolution of the tongue and associated structures in salamanders and newts (family Salamandridae).Copeia 1969:91-123. Peters, S. E., and K. C. Nishikawa (1999) Comparison of isometric contractile properties of the tongue muscles in threee species of frogs, Litoria caerulea, Dyscophus guinetti and Bufo marinus. J. Morphol. 242:107-124. Premo, D. B., and A. H. Atmowidjojo (1987) Dietary patterns of the "crab-eating frog," Rana cancrivora, in West Java. Herpetologica 43:1-6. Regal, P. J. (1966) Feeding specializations and the classification of terrestrial salamanders. Evolution 20:392-407. Regal, P. J., and C. Gans (1976) Functional aspects of the evolution of frog tongues. Evolution 30:718-734. Reilly, S. M., and G. V Lauder (1989) Kinetics of tongue projection in Ambystoma tigrinum: quantitative kinematics, muscle function, and evolutionary hypotheses. J. Morphol. 199:223-243. Ridewood, W. G. (1897) On the structure and development of the hyobranchial skeleton and larynx in Xenopus and Pipa; with remarks on the affinities of the Aglossa. J. Linn. Soc, Zool. 26: 53-128. Ritter, D. A., and K. C. Nishikawa (1995) The kinematics and mechanism of prey capture in the African pig-nosed frog {Hemisus marmoratum): the description of a radically divergent anuran tongue. J. Exp. Biol. 198:2025-2040. Rossignol, S., J. P. Lund, and T. Drew (1988) The role of sensory inputs in regulating patterns of rhythmical movements in higher vertebrates: a comparison between locomotion, respiration and mastication. Pp. 201-284. In: Neural Control of Rhythmic Movements in Vertebrates. A. H. Cohen, S. Rossignol, and S. Grillner (eds.), Wiley, New York.
5. F e e d i n g in Frogs Roth, G. (1987) Visual Behavior in Salamanders. Springer Verlag, Berlin. Roth, G., K. C. Nishikawa, D. B. Wake, U. Dicke, and T. Matsushima (1990) Mechanics and neuromorphology of feeding in amphibians. Neth. J. Zool. 40:115-135. Ruvinsky, I., and L. R. Maxson (1996) Phylogenetic relationships among bufonid frogs (Anura: Neobatrachia) inferred from mitochondrial DNA sequences. Mol. Phylogen. Evol. 5:533-547. Satou, M., and A. Shiraishi (1991) Local motion processing in the optic tectum of the Japanese toad, Bufo japonicus. J. Comp. Physiol. A 169:569-589. Severtzov, A. S. (1961) On the mechanism of tongue protrusion in anuran amphibians. Doklad. Akad. Nauk SSSR 140:256-259 (translation). Shinn, E. A., and J. W. Dole (1978) Evidence for a role for olfactory cues in the feeding response of leopard frogs, Rana pipiens. Herpetologica 34:167-172. Simon, M. P., and C. A. Toft (1991) Diet specialization in small vertebrates: mite-eating frogs. Oikos 61:263-278. Smith, C., and A. Bragg (1949) Observations on the ecology and natural history of Anura. VII. Food and feeding habits of the common species of toads in Oklahoma. Ecology 30:333-349. Sokol, O. M. (1969) Feeding in the pipid frog Hymenochirus boettgeri (Tornier). Herpetologica 25:9-24. Starrett, P. H. (1968) The Phylogenetic Significance of the Jaw Musculature in Anuran Amphibians. Ph. D. Dissertation, University of Michigan. Taigen, T. L., and F. H. Rough (1983) Prey preference, foraging behavior, and metabolic characteristics of frogs. Am. Nat. 122: 509-520. Thexton, A. J., D. B. Wake, and M. H. Wake (1977) Tongue function in the salamander Bolitoglossa occidentalis. Arch. Oral Biol. 22:361366. Toft, C. A. (1981) Feeding ecology of Panamanian litter anurans: patterns in diet and foraging mode. J. Herpetol. 15:139-144. Toft, C. A. (1995) Evolution of diet specialization in poison-dart frogs (Dendrobatidae). Herpetologica 51:202-216.
147
Trueb, L. (1973) Bones, frogs, and evolution. Pp 65-132. In: Evolutionary Biology of the Anurans. J. Vial (ed.). Univ. of Missouri Press, Columbia. Trueb, L., and C. Gans (1983) Feeding specializations of the Mexican burrowing toad, Rhinophrynus dorsalis (Anura: Rhinophrynidae). J. Zool. London 199:189-208. Tso, T. A., J. C. O'Reilly, and K. C. Nishikawa (1995) Conservation of function of the m. hyoglossus during feeding in frogs. Am. Zool. 35:123A. Tyler, M. J. (1974) Superficial mandibular musculature and vocal sac structure of the Mexican burrowing toad, Rhinophrynus dorsalis. Herpetologica 30:313-316. Valdez, C. M., and K. C. Nishikawa (1997) Sensory modulation and motor program choice during feeding in the Australian frog, Cyclorana novaehollandiae. J. Comp. Physiol. A 180:187-202. Vences, M. (1988) Zum Beutefang Verhalten der europaischen Amphibien. Herpetofauna 10:6-10. Wainwright, P C , D. M. Kraklau, and A. F Bennett (1991) Kinematics of tongue projection in Chamaeleo oustaleti. J. Exp. Biol. 159: 109-133. Wake, D. B. (1982) Functional and developmental constraints and opportunities in the evolution of feeding systems in urodeles. Pp. 51-66. In: Environmental Adaptation and Evolution. D. Mossakowski and G. Roth (eds.) Gustav Fischer, Stuttgart. Webster, S. (1996) The Morphology of the Connective Tissues of Extensible Tongues. Unpublished Honors Thesis, College of Cardiff, University of Wales. Weerasuriya, A. (1989) In search of the pattern generator for snapping in toads. Pp. 589-614. In: Visuomotor Coordination, Amphibians, Comparisons, Models and Robots. J.-P. Ewert and M. A. Arbib (eds.). Plenum, New York. Wiltenmuth, E. B., and K. C. Nishikawa (1994) Scaling of feeding kinematics across four species of dendrobatid frogs. Am. Zool. 34:57A. Zug, G., and P. Zug (1979) The marine toad, Bufo marinus: a natural history resume of native populations. Smithsonian Contr. Zool. 284-1-58.
This Page Intentionally Left Blank
C H A P T E R
6 Feeding in Caecilians JAMES C. O'REILLY Organismic and Evolutionary Biology University of Massachusetts Amherst Amherst, Massachusetts 01003
I. INTRODUCTION 11. MORPHOLOGY A. Sensory Organs B. Musculoskeletal Anatomy III. FUNCTION A. Finding Prey B. Capturing Prey (Ingestion) C. Intraoral Transport IV. EVOLUTION V. THE FUTURE References
son, 1989). In Central and South America, they are known from southern Mexico to northernmost Argentina. In Asia, they range from India and Sri Lanka to Indochina. In southeast Asia they range from the southernmost Philippines and southern China south through the Greater Sunda Islands to Wallace's line. No caecilians are known from the rain forests of New Guinea or Australia. In Africa, they are abundant throughout the humid coastal regions of the west and from several disjunct areas in the east, including parts of Tanzania, Ethiopia, Kenya, Malawi, and Rwanda (Nussbaum and Hinkel, 1994). Only a single specimen is known from all of central Africa (Nussbaum and Pfrender, 1998) and no caecilians have been reported in Madagascar. Six species, which appear to be a monophyletic group, are found in the Seychelles Archipelago (Nussbaum, 1985a; Nussbaum and Ducey, 1988), islands approximately halfway between India and Madagascar. Most caecilians are fossorial but several species are semiaquatic, and at least four South American species are entirely aquatic (Moodie, 1978; Nussbaum and Wilkinson, 1989; Wilkinson, 1989; Wilkinson and Nussbaum, 1997). Fossorial species primarily dwell in highly organic, friable surface layers of the soil where they maintain tunnel systems (Tanner, 1971; Himstedt, 1991,1996; Ducey ei al, 1993; Wake, 1993b). These surface-dwelling burrowers display a high degree of mechanical independence of the vertebral column and skin (Gaymer, 1971; Summers and O'Reilly, 1997) and can generate extraordinary burrowing forces by using their entire bodies as single chambered hydrostatic skeletons (O'Reilly ei al., 1997). However, observations
L INTRODUCTION Caecilians are legless, burrowing amphibians that superficially resemble earthworms and are found only in the tropics. Currently, there are about 150 species and six families recognized (Nussbaum and Wilkinson, 1989). There is a growing consensus concerning the phylogenetic relationships among the basal caecilian families (Hedges et ah, 1993; Wilkinson and Nussbaum, 1996; Fig. 6.1); however, over two-thirds of the species and over three-quarters of the genera are currently placed in the paraphyletic Caeciliidae. There are currently no widely accepted phylogenetic hypotheses that include more than a small fraction of the genera within Caeciliidae, thus relationships among the vast majority of caecilians remain uncertain (Nussbaum and Wilkinson, 1989; Hedges et al, 1993; Wake, 1993a; Wilkinson and Nussbaum, 1996; Wilkinson, 1997). Caecilians have been found throughout most of the humid tropics (Taylor, 1968; Nussbaum and Wilkin-
FEEDING (K. Schwenk, ed.)
149
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
150
James C. O'Reilly Rhinatrematidae Ichthyophiidae Uraeotyphlidae Scolecomorphidae "Caeciliidae" Typhlonectidae
FIGURE 6.1. Current phylogenetic hypothesis of caecilian relationships modified from Hedges et al. (1993) and Wilkinson and Nussbaum (1996). Caeciliidae is almost certainly paraphyletic with respect to Typhlonectidae.
of captive specimens of some extremely elongate species that lack independence between the vertebral column and skin (e.g., Boulengerula) indicate that they seem to modify preexisting spaces rather than excavating tunnel systems of their own design. As far as is known, all of the members of the basal families (Rhinatrematidae, Ichthyophiidae and Uraeotyphlidae) have a life history that includes a free-living, aquatic larval stage and a terrestrial adult stage (Sarasin and Sarasin, 1887-1890; Parker, 1956; Taylor, 1968; Wake, 1977a; Wilkinson, 1992b). At least three genera of Caeciliidae {Grandisonia, Praslinia, and Sylvacaecilia) have species with a free-living larval stage (Parker, 1958; Largen et al, 1972; Nussbaum, 1992). The larval period of caecilians can last several months and is followed by a gradual metamorphosis to the adult stage (Breckenridge et al, 1987; Himstedt, 1991,1996). Members of the rest of Caeciliidae, Scolecomorphidae, and Typhlonectidae, for which there are data, have either direct development or are viviparous (Parker 1956; Wake, 1977a,b, 1992; Wilkinson and Nussbaum, 1998). After their yolk has been depleted, fetuses of viviparous species are matrotrophic, feeding on secretions of the epithelial cells of the oviduct and, at least inadvertently, the epithelial cells themselves (Parker, 1956; Welsch et al, 1977; Wake, 1977a,b). Unfertilized eggs, expelled from the ovaries well after pregnancy has begun, may also be consumed in some species. Gestation is known to last from 6 months in Typhlonectes (Exbrayat and Delsol, 1985) to as long as a year in Dermophis (Wake, 1980a). At least one viviparous species, Geotrypetes seraphini, guards its young for some time after birth (Sanderson, 1937). Little is known about the diet of caecilians relative to other tetrapods. As far as is known, all caecilians are carnivores. Wake (1980a) found that earthworms dominate the diet of one population of Dertnophis from Guatemala, and Moll and Smith (1967) reported a
specimen of Dertnophis having eaten a lizard. Barbour and Loveridge (1928) and Hebrard et al (1992) concluded that termites are the predominant prey of two different east African species of Boulengerula. Moodie (1978) and Wake (1978) found coleopteran larvae, shrimp, and small arthropods in the digestive tracts of Typhlonectes. Loveridge (1936) and Nussbaum and Pfrender (1998) reported termites, earthworms, and insects in the stomachs of Schistometopum. Remaining data on the caecilian diet are based on captive animals. In captivity, larvae of Epicrionops will eat only small aquatic arthropods when originally captured, but can be trained to eat earthworms (O'Reilly, 1995, 1996). Specimens of Siphonops and Dertnophis will readily eat newborn rodents. Several species that have been kept in captivity (e.g., species of Caecilia) have refused to eat anything but earthworms (O'Reilly, 1996). Boulengerula also appears to have a specialized diet as those kept in captivity will only eat termites and small crickets while ignoring earthworms (O'Reilly, 1996). In contrast, adults of some species (e.g., Hypogeophis rostratus) are apparently opportunists that will eat a wide variety of prey, including salmon eggs and pieces of veal, items that they presumably would never come across in nature (Tanner, 1971). The current dearth of information on caecilian phylogeny and feeding behavior limits the following discussion primarily to functional interpretations of sensory and musculoskeletal anatomy. It is my hope that this will be a starting point from which more complete studies of feeding behavior can be undertaken. I begin by summarizing the relatively abundant literature on caecilian cranial anatomy and reviewing the few functional studies on caecilian feeding. I will then propose some general hypotheses about the forces driving the evolution of caecilian head anatomy and feeding behavior. Finally, I will discuss areas that would be productive for future investigation. 11. MORPHOLOGY This section describes what is known about the sensory and musculoskeletal morphology of caecilians. In the section on sensory organs, the descriptions are organized by the functional system. In the musculoskeletal anatomy section, the anatomy of larvae, fetuses, and adults will be discussed in turn. A. Sensory Organs 1. Eye With the exception of some scolecomorphids that have protrusible eyes, caecilian eyes are small and buried either under skin or dermal bones (Wake, 1985;
151
6. Feeding in Caecilians Himstedt, 1995; O'Reilly et ah, 1996). All caecilians that have been examined in detail possess a retina that is linked to the brain with a complete, although sometimes extremely attenuate, optic nerve (Wake, 1985). Among those species for which data are published, the lens is variably developed with the exception of Boulengerula, which is lensless, and the extrinsic eye muscles have been coopted for other functions, reduced in size, or no longer develop (Wake, 1985). The eyes of caecilians are also reduced in terms of neuroanatomical features. The optic nerve of Ichthyophis has far fewer fibers than those of lungfish, frogs, and salamanders (Roth et ah, 1993). In Ichthyophis (both larvae and adults) and Typhlonectes, the visual fields of the optic tectum are virtually absent, but bilateral thalamic projections to the medial optic tracts are well developed (Clairambault et ah, 1980; Fritzsch et ah, 1985; Himstedt and Manteuffel, 1985). 2. Olfactory
System
In contrast to the visual system, the olfactory system of caecilians is well developed relative to frogs and salamanders. Like all anamniote sarcopterygians that possess lungs, caecilians use buccal pumping to breath (Carrier and Wake, 1995). In this method of breathing, air is sucked into the buccal cavity via the nares (and nasal epithelia) before it is pumped into the lungs by muscles of the hyobranchial apparatus. When not filling the lungs, less dramatic movements of the floor of the buccal cavity cause air to be flushed over the nasal epithelia. The main nasal cavity is lined with olfactory epithelium at its anterior end and respiratory epithelium at the caudal end (Schmidt and Wake, 1990). The olfactory epithelium lines the majority of the nasal cavity in most species, but is reduced in the typhlonectids in which the majority of the cavity is lined with respiratory epithelium. Still, both morphological and behavioral evidence indicate that typhlonectids use buccal pumping in order to smell in water (Wilkinson and Nussbaum, 1997). 3. Tentacle Organ Caecilians posses a unique sensory organ called the tentacle that is derived from the tear duct and other parts of the eye (Wiedershiem, 1879; Engelhardt, 1924; Billo and Wake, 1987) and is directly connected to the vomeronasal organ (Badenhorst, 1978; Fiimstedt and Simon, 1985). The tentacle is seen in all metamorphosed caecilians but is not present in larvae. Among rhinatrematids, the tentacle lies immediately in front of the eye and is not protractile. In all other caecilians, the tentacle is in a more anterior position and the degree to which it can be protracted varies. In some species (e.g., Ichthyophis kohtaoensis) the tentacle can be pro-
truded a considerable distance (several millimeters) out of the head. Other species (e.g., Dermophis mexicanus) are capable of only very limited protrusion and members of the genus Typhlonectes do not protrude the tentacle at all (Wilkinson and Nussbaum, 1997). Fox (1985) found evidence of tactile sensory cells on the surface of the tentacle of Ichthyophis but no evidence of chemosensory cells. The paired vomeronasal organs of typhlonectids are proportionally larger and have more extensive projections to the bulbus olfactorius accessorius than those of other caecilians that have been examined (Schmidt and Wake, 1990). 4. Lateral Line Caecilians can possess two types of lateral line organs: neuromasts (mechanoreceptor organs) and ampuUary organs (electroreceptors). Larvae of five genera {Epicrionops, Ichthyophis, Uraeotyphlus, Grandisonia, and Sylvacaecilia) have been described as having a lateral line system (Parker, 1958; Taylor, 1970; Largen et al, 1972; Hetherington and Wake, 1979; Fritzsch et al, 1985; Wahnschaffe et al, 1985; Fritzsch and Wake, 1986; Wake, 1987; Wilkinson, 1992a,b), but only that oi Ichthyophis has been examined in detail. Larvae of Ichthyophis possess a lateral line system with both neuromasts and ampuUary organs. The neuromasts are found both on the head and along the trunk, whereas ampuUary organs are restricted to the head (Fritzsch et al, 1985; Wahnschaffe et al, 1985). Larvae of Epicrionops possess two rows of neuromasts on the trunk, whereas other larvae possess only one row that is homologous to the dorsal row of Epicrionops (Wilkinson, 1992a). As in other anamniotes, the lateral line afferents of Ichthyophis project to the dorsal nucleus of the rhombencephalic alar plate (Fritzsch et al, 1985). Lateral line organs have not been reported in any fetal caecilians, but are presumably present at least late in development in fetal typhlonectids. With the exception of the semiaquatic Hypogeophis rostratus and the semiaquatic and aquatic typhlonectids, which have only ampuUary organs, lateral line organs are absent in adult caecilians (Fritzsch and Wake, 1986). 5. Auditory
System
An excellent comparative study and literature review of the anatomy of the caecilian auditory apparatus and its inervation has been provided by Fritzsch and Wake (1988). The inner ear of amphibians is generally composed of six regions: three semicircular canals (horizontal, anterior, and posterior), the utriculus, the sacculus, and the lagena. Each of these regions has distinct sensory epithelia innervated by branches of the eighth cranial nerve. These epithelia primarily function as vestibular sensors, but both the sacculus and
152
James C. O'Reilly
the lagena also play a role in the perception of low frequency sounds (Lewis et al, 1985). All of these sensory epithelia are found in all caecilians with the exception of the lagena, which has not been observed in Boulengerula, Oscaecilia, or typhlonectids (Fritzsch and Wake, 1988). In addition to the general inervation of the epithelia of the inner ear chambers, three specialized auditory patches of epithelial cells, termed papillae, are found in the utriculus and sacculus. The basilar papilla is found in Latimeria, most amniotes, frogs, and some salamanders. This papilla is present only in basal caecilians (rhinatrematids, ichthyophiids, and uraeoty phlids) where it located in a side pocket of the sacculus (Fritzsch and Wake, 1988). This region of epithelial cells is sensitive to high-frequency sounds in frogs (Lewis and Lombard, 1988) and presumably serves the same function in those caecilians that possess it. The papilla neglecta is a synapomorphy of living gnathostomes, but has been lost in both salamanders and frogs. It resides in the utriculus just earlier the open ing to the sacculus and is sensitive to low-frequency sounds (Lewis et al., 1985). The amphibian papilla, a synapomorphy of living amphibians, is found in the sacculus just later the utriculus and is sensitive to lowfrequency sounds in anurans (Lewis et al, 1985). Based on ontogenetic data from Dermophis, Fritzsch and Wake (1988) argue that the amphibian papilla is derived directly from part of the papilla neglecta. Despite the relatively reduced size of the papillae
spiracle
lateral line
eye
nares
of the caecilian ear relative to other amphibians, the eighth cranial nerve shows no appreciable degeneration. The auditory projections of Ichthyophis are very similar to those seen in salamanders and have been interpreted as primitive in organization by Fritzsch (1988) because they resemble those of Latimeria and Dipnoi more than those of amniotes and frogs. B. Musculoskeletal Anatomy 1. Larvae Among basal caecilians, young first feed as freeliving aquatic larvae. The only available detailed descriptions of larval cranial anatomy are based on Ichthyophis. It is difficult to assign the descriptions presently available to any particular species of Ichthyophis and it is likely that at least some of the series described include mixtures of more than one species. Fortunately, other than the proportions of the skull and the number of gill slits, there is little variation between the specimens described. Larval caecilians have well-developed labial lobes, which give the head a squared-off appearance relative to adults (O'Reilly, 1995; Fig. 6.2), but the skull is much more fusiform in shape. The skull of larval Ichthyophis has large temporal fenestrae and the large jaw levator muscles originate along the sagital midline before traveling through the fenestra to the lower jaw (Edgeworth, 1935; Visser, 1963). The lower jaw is well ossi-
annulus annular groove
eye nares tentacle
FIGURE 6.2. Illustration of the external cranial morphology of larval and adult Ichthyophis banannicus. Note the large labial folds and resulting blunt head shape of the larvae. Upon metamorphosis, the skull becomes ossified more heavily and the skin becomes connected more tightly to the dermal bones, resulting in the more fusiform head shape of the adult. Illustration by Loree Harvey.
6. Feeding in Caecilians fled, with two rows of recurved teeth and a prominent retroarticular process that curves dorsally posterior to the jaw articulation, and a well-developed interhyoideus posterior (Edgeworth, 1965; Visser, 1963). The upper jaws also have two rows of teeth, including a premaxillary-maxillary row and a vomerine-palatine row. In larval caecilians the maxillary and palatine are separate bones, but these are always fused into a maxillopalatine with a much extended row of maxillary teeth in adults (Reiss, 1996). The quadrate of larvae is relatively independent, becoming more firmly attached to the squamosal and maxillopalatine during metamorphosis (Visser, 1963). The hyobranchial apparatus and associated musculature is well developed in larval Ichthyophis and consists of several pairs of independent elements. A ventromedial basibranchial articulates with a pair of ceratohyals and four pairs of ceratobranchials, with the fourth ceratobranchial articulating with the third (Edgeworth, 1935). In Ichthyophis, the hyobranchial skeleton is cartilaginous, but in larval Epicrionops, it is well ossified (Wake, 1989). 2. Fetuses Many species of caecilians have a fetal developmental stage that is associated with a suite of derived cranial characteristics. In the earliest developmental stages of Dermophis pictured by Wake and Hanken (1982), the hyobranchial apparatus has the same design as that of adults (cartilaginous with left and right elements fused). During the early stages of this fetal period, the head is conspiciously larger and more advanced in development than the rest of the body. In contrast to free-living larvae, which have teeth like those of adult caecilians, fetuses possess up to 12 rows of spatulate teeth that are sometimes combed at the tips (Parker, 1956; Parker and Dunn, 1964; Wake, 1976). The teeth extend well outside the oral cavity and the bases of the teeth are sometimes fused into plates (Parker, 1956). The mouth is almost terminal (Wake and Hanken, 1982; Wilkinson, 1991; Wake, 1993c), with the teeth, quadrate, and jaws ossifying well ahead of the rest of the skull (Wake and Hanken, 1982). 3. Adults In all adult caecilians, the hyobranchial apparatus is cartilaginous and all of the paired elements that are independent in larvae fuse whereas most medial elements degenerate. The tongue is large but is tightly connected to the mandible along its margins and is not protrusible (Bemis et al, 1983). The teeth are recurved.
153
pedicellate, and can be either monocuspid or bicuspid (Lawson, 1965a; Wake and Wurst, 1979). Most caecilian teeth are sharply pointed, but at least some typhlonectids possess more blade-like tooth tips (Wake and Wurst, 1979; Wilkinson, 1991). Most caecilians possess a unique diminutive muscle, the levator quadrati, which originates on the os basal and inserts on the quadrate (Edgeworth, 1925, 1935; Wilkinson and Nussbaum, 1997). Caecilians are also unique in that the interhyoideus muscle inserts on the retroarticular process of the lower jaw rather than the tips of the ceratohyals (Fig 6.3) and contributes to jaw closing (Nussbaum, 1977,1983; Bemis et a/., 1983). The mandibles and associated musculature display significant variation among species. Descriptions of the five morphological groups that follow are summaries of data from Nussbaum (1983) and Wilkinson and Nussbaum (1997) unless otherwise noted. a. Group I Members of Rhinatrematidae possess skulls with large temporal fenestrae as adults. The lower jaw has a well-developed, mainly horizontal retroarticular process. The adductor mandibulae complex is larger than in other caecilians, extending through the temporal fenestra to the midline of the cranium. The interhyoideus posterior is smaller than that of other caecilians, in terms of cross section, but the fibers are of similar length to those of other groups. The depressor mandibulae has a broad origin with fibers ranging in orientation from horizontal to vertical with all fibers converging on a narrow insertion on the tip of the retroarticular process. h. Group II Members of Ichthyophiidae and Uraeotyphlidae have skulls that either have no temporal fenestra or a small fissure between the squamosal and the parietal. The retroarticular process is strongly recurved just posterior to the jaw articulation. The adductor complex of all adult members of this group is confined below the squamosal and fibers do not extend onto the surface of the skull. As a result, the fibers of the adductor complex are relatively much shorter than those seen in group I. The interhyoideus posterior of this group has fibers of similar length as those seen in group I, but the muscle extends much farther up the side of the body. This greatly increases the physiological cross section of the interhyoideus posterior relative to group I while maintaining a fusiform body cross section. In contrast to group I, all of the fibers of the depressor mandibulae have assumed a horizontal orientation.
154
James C. O'Reilly I me
FIGURE 6.3. (A) Dorsolateral view of the cranial anatomy of an adult Dermophis mexicanus. Note the large size of the interhyoideus and its insertion on the retroarticular process of the mandible, rl, rectus lateralis; seo, obliquus externus superficialis; lab, levator arcus branchiales; dm, depressor mandibulae; ih, interhyoideus; dt, dorsalis trunci; p, parietal; f, frontal; npm, nasopremaxilla; s, squamosal; mp, maxillopalatine; pa, pseudoangular; pd, pseudodentary. (B) Ventrolateral view of Dermophis with superficial muscles removed. Note the diminutive size of the levators mandibulae. Ic, longus capitis; imp, levator posterior; Ime, levator mandibulae externus; Ima, levator mandibulae anteior; im, intermandibularis; gh, geniohyoideus; gg, genioglossus; re, rectus cervicus; pa, pseudoangular; pd, pseudodentary. From Bemis et ah (1983).
c. Group III According to Nussbaum (1985b), members of Scolecomorphidae can have solidly roofed skulls {Crotaphatrema) or skulls with large temporal fenestrae (Scolecomorphus). Regardless of the presence or absence of the fenestrae, the adductor mandibulae complex is diminutive and restricted to the adductor chamber. Like in group II, the retroarticular process of the mandible is strongly recurved. However, unlike the re-
troarticular process of group II, the mandible curves anterior to the jaw articulation (Nussbaum, 1985b). Thus, the mandibular fossa opens caudally in scolecomorphids in contrast to most other caecilians. Like group II, the interhyoideus posterior is massive and many of the fibers extend horizontally along the first several segments of the trunk. In contrast to all other groups, some fibers of the interhyoideus posterior insert on the lower jaw anterior to the jaw joint. The de-
155
6. Feeding in Caecilians pressor mandibulae is diminutive, oriented obliquely, and inserts on the retroarticular process.
we can start to understand how caecilians go about localizing and subduing potential food items.
d. Group IV With the exception of the typhlonectid Atretochoana, all of the members of Caeciliidae and Typhlonectidae can generally be grouped together because they share moderate curvature of the retroarticular process, diminutive adductors, and a well-developed interhyoideus posterior (Fig. 6.3). The skulls of this group range from solidly roofed (e.g., Caecilia and Dermophis) to having large temporal fenestrae {Geotrypetes and typhlonectids) (Taylor, 1969). The adductor mandibulae complex is restricted to the adductor chamber even in those species with large temporal fenestrae. The interhyoideus posterior varies in size, reaching its most extreme form in Microcaecilia and Praslinia where it extends many segments down the body and is pinnate in form. The depressor mandibulae is obliquely oriented and inserts well posterior to the jaw joint. e. Group V The final morphological group contains a single species, Atretochoana eiselti. Although long recognized as distinct species (Taylor, 1968), the morphology oi Atretochoana has only recently been described in detail (Nussbaum and Wilkinson, 1995; Wilkinson and Nussbaum, 1997). This relatively large (only two known specimens are over 70 cm in length and robust in proportions) caecilian is unique among Gymnophiona in that it completely lacks lungs and functional choanae. The jaw apparatus oi Atretochoana has several unique characteristics. The most obvious is the extreme dorsoventral compression relative to other caecilians. The jaw joint is placed well behind the rest of the skull, due to extreme elongation of quadrate. The retroarticular process is very small and recurved sharply directly before the jaw articulation (as in type III). The temporal fenestra is proportionally much larger than in any other known caecilian. The jaw levators are proportionally larger than those seen in type II, III, and IV skulls and are oriented horizontally between the skull and lower jaws, rather than dorsoventrally as in other caecilians. A detailed description of the anatomy of this unique species with comparisons to other caecilians can be found in Wilkinson and Nussbaum (1997). III. F U N C T I O N With the exception of two experimental behavioral studies and several reports that include observations on feeding, we have little direct evidence of how caecilians find and capture their prey. However, by combining these studies with available morphological data.
A. Finding Prey In both larvae and adults, the sense of sight probably plays little or no role in prey identification or localization. There are currently no data to suggest that the eyes of any caecilian are capable of forming images or play any role in localizing prey. Negative phototaxis has been reported in Hypogeophis (Taylor, 1968) and Ichthyophis (Himstedt and Manteuffel, 1985) and has been observed in many other species that have been kept in captivity. However, in larval and adult Ichthyophis, the number of optic nerve fibers and the afferent projections of the optic nerve suggest that the retina functions only as a light receptor (Himstedt and Manteuffel, 1985; Fritzsch et ah, 1985; Himstedt, 1995). The use of the eye as a photoreceptor is probably widespread among caecilians (Wake, 1985) except in those species where the eyes are completely covered by bone and may no longer serve any sensory function. As with the eye, the morphology of the ear of caecilians suggests that it plays little or no role in prey detection. The lack of a tympanum, the reduction of the auditory epithelia of the inner ear, and the poorly developed auditory centers of the alar plate of caecilians indicate that the sense of hearing is not well developed (Fritzsch and Wake, 1988). The basal papilla, which is sensitive to high-frequency sounds, is not present in most caecilians, whereas those epithelia thought to be involved in the detection of low-frequency sounds (sacculus, lagena, papilla neglecta, and the amphibian papilla) are largely retained. Available data (from Ichthyophis, Geotrypetes, and Dermophis) on the sensitivity of caecilian ears to airborne sounds of different frequencies indicate that they are not very effective in air but might function well in water or mud (Wever, 1975; Wever and Gans, 1976). A functional role of lowfrequency sound detection for prey localization has yet to be demonstrated. Both lateral line organs and olfaction apparently play an important role in prey detection in aquatic species. Results of behavioral experiments carried out on larvae of Ichthyophis indicate that the ampuUary organs of the head are fully functional electroreceptors that are used to locate prey (Himstedt and Fritzsch, 1990). The functional role of the neuromasts of the lateral line have not been tested, but they may sense prey movement, detect the approach of predators, or act as proprioreceptors during locomotion. Most adult caecilians primarily feed on land and lack ampullary organs. The presence of ampullary organs in the heads of aquatic foraging typhlonectids and Hypogeophis (Fritzsch and
156
James C. O'Reilly
Wake, 1986) implies that they may be used to locate aquatic prey in these species (Wilkinson, 1992a). The behavior of typhlonectids indicates that olfaction probably also plays an important role in prey detection (Wilkinson and Nussbaum, 1997) The single behavioral study of how terrestrial caecilians find prey was performed by Himstedt and Simon (1995) on Ichthyophis. They found that Ichthyophis can usually locate pieces of bovine heart at a distance of 12 cm in less than 2 min (significantly faster than the salamander Triturus alpestris). When the nares were blocked, this species was not only incapable of locating prey, but apparently could not identify prey items when they are in direct contact with the snout. These results suggest that olfaction via the nares is the principle sensory modality used by terrestrial caecilians to find food. The tentacle organ appears to play little or no role in the detection of airborne substances given off by prey, but is more likely important for following trails of molecules left in tunnels by passing prey items (Himstedt and Simon, 1995). The lack of chemosensory cells on the tentacle of Ichthyophis (Fox, 1985) suggests that it must aid in chemoreception by gathering molecules from the substrate and delivering them to the chemosensory cells of the vomeronasal organ (Badenhorst, 1978; Himstedt and Simon, 1995). B. Capturing Prey (Ingestion) After locating a prey item it must be captured before it can be eaten. In water, larval and adult salamanders generally use suction feeding to apprehend prey items (Reilly and Lauder, 1992; O'Reilly, 1995). Suction feeding is relatively rare among frogs and caecilians (O'Reilly, 1995). The only larval caecilian for which kinematic data are available (an unidentified species of Epicrionops from Ecuador) suction feeds (Fig. 6.4, left). However, anatomical evidence suggests that suction feeding is the general method of prey capture used by larval caecilians (O'Reilly, 1995). The jaw movements are highly stereotyped in larval Epicrionops (Fig. 6.5A) and display a pattern similar to the jaw movements of suction-feeding salamanders (Reilly and Lauder, 1992). In many viviparous caecilians, the young feed for extended periods in the oviduct before being born (Parker, 1956; Wake, 1977b). The subject of what fetal caecilians feed on and the mechanism by which they feed has been the subject of some inference, but direct observations of feeding have yet to be reported. The extended gestation period and the fact that fetal caecilians increase many times in body mass after they have resorbed their yolk indicate that they are gaining considerable nutrition from the mother during development (Parker, 1956). The gut contents of fetal caecili-
ans suggest that they are feeding orally (histotrophy), rather than absorbing nutrients through their skin or gills (Parker, 1956; Wake, 1977b). Fetuses appear to be eating "uterine milk" from secretory cell beds in the uterine epithelium and some shed epithelial cells (Parker, 1956; Wake, 1977b) and may also eat unfertilized eggs, which are produced by the ovaries of at least some species well into fetal development. Parker and Dunn (1964) concluded that the dentition of fetal caecilians was not functional, but effective arguments to the contrary have been made by Wake (1977b) and Wilkinson (1991). The fact that the epithelial lining of the oviduct is no longer present in the immediate vicinity of advanced fetuses but is intact elsewhere (Wake, 1977b) suggests that fetal caecilians are wearing down the epithilium while feeding and lend support to the conclusion that fetal teeth are an aptation for oviductal feeding. In contrast to salamanders and frogs, which generally use their tongues to capture prey on land (Lauder and Reilly, 1993; Nishikawa et al, 1992), all caecilians observed thus far {Boulengerula, Caecilia, Dermophis, Epicrionops, Grandisonia, Hypogeophis, Ichthyophis, Schistometopum, Scolecomorphus, and Siphonops) use jaw prehension for terrestrial prey capture (Figs. 6.4, right, and 6.6, left). Caecilians generally contact prey before a gape cycle is initiated (although excited individuals have been observed to wave their heads back and forth with the mouth open after smelling food), implying that tactile information is usually needed to trigger a feeding sequence. Data from Ichthyophis with blocked nares suggest that olfactory cues are also needed for prey capture to be initiated (Himstedt and Simon, 1995). Relative to frogs and salamanders, the feeding movements of adult caecilians are extremely slow (Table 6.1) and appear strongly modified by sensory information gathered midstrike (O'Reilly. 1995). Once prey is contacted, the jaws are slowly opened while precisely tracing the outline of the prey item. The process of opening the jaws and placing them on the prey can take well over a second in Ichthyophis (Fig. 6.5B). Jaw closing is often initiated and then interrupted repeatedly in apparent attempts to gain better jaw placement (Fig. 6.5). However, at the slightest sign of attempted escape, the jaws are closed relatively rapidly. Jaw movements are much more rapid when Ichthyophis is presented with a cricket rather than an earthworm, even if the cricket is pinned to the substrate with forceps so that it cannot escape. These observations suggest that the gape cycle is modified both by a priori olfactory information about prey type and by tactile information received during the strike itself. Aquatic feeding in adult caecilians is qualitatively very similar to terrestrial feeding with some minor
6. F e e d i n g in Caecilians
F I G U R E 6.4. (Left) High-speed cinematography of an aquatic prey-capture sequence in larval Epicrionops eating a piece of an earthworm. Note the movement of the prey toward the feeding individual, the occluded lateral margin of the gape, and hyobranchial expansion after mouth closing. (Right) Selected video frames of terrestrial prey capture in an adult Ichthyophis kohtaoensis. Note the forward lunge, lack of tongue protraction, and lack of hyobranchial depression.
157
158
James C. O'Reilly B.
\'' '. /
/
v^
\
0.0 0.01 0.02 0.03 0.04 0.05 0.06
60-^
c
CD 0> CL CQ
I
40 20 /
/
/
/\ A
\-,
ii
//
0.5
1.0
1.5
2.0
Time (sec)
0.0
0.2
0.4
0.6
0.8
1.0
Time (sec)
F I G U R E 6.5. Selected sample gape cycles from (A) aquatic larval Epicrionops {n = 5), (B) terrestrial adult Ichthyophis {n = 3), (C) terrestrial adult Hypogeophis {n = 3), and (D) aquatic adult Typhlonectes {n = 3). Note the extreme length of some adult gape cycles and how jaw closing can be interupted and the gape cycle restarted multiple times before the mouth is finally closed on the prey item.
differences. The two aquatic foraging caecilians for which data are available {Typhlonectes and Hypogeophis) use jaw prehension to capture aquatic prey and cannot suction feed (O'Reilly, 1995). However, Typhlonectes does depress the hyobranchium in conjunction with mouth closing (Fig. 6.6, Right), which probably keeps the lunging caecilian from pushing prey items away ("compensatory suction" of Van Damme and Aerts 1997). Compared to terrestrial caecilians, Typhlonectes has relatively fast jaw movements for its body size (O'Reilly, 1995). Individuals of Hypogeophis increase the speed of their jaw movements when fed the same prey item in water than on land. However, this facultative increase in speed does not result in movements as quick as those seen in Typhlonectes, suggesting that the timing of feeding behavior has diverged significantly from that of its terrestrial ancestors. The degree to which the retroarticular process is re-
curved and whether it curves antiororly or posteriorly to the jaw articulation is likely to have significant functional consequences (Fig. 6.7). When the retroarticular process is straight, the interhyoideus will be most effective at large gape angles (Fig. 6.7A). As the retroarticular process becomes more recurved, bite force performance at low gape angles will improve, but at the expense of performance at high gape angles (Fig. 6.7B). The decrease in performance with increasing gape is more pronounced if the jaw curves upward anterior to its articulation with quadrate (Fig. 6.7C). Available data on maximum gape angles support this model, as Ichthyophis, which has a strongly recurved retroarticular process, has only been observed to open the jaws during feeding to about 50°, whereas Hypogeophis and Typhlonectes, with less recurved retroarticular processes, open their mouths to over 70° (O'Reilly, 1995).
159
6. Feeding in Caecilians TABLE 6.1 Species Gymnophiona Epicrionops sp.—larvae^ Hypogeophis rostratus' Ichthyophis kohtaoensis^ Typhlonectes nutans ^ Anura Ascaphus truei" Bombina orientalist Bufomarinus^ Discoglossus pictus^ Hyla cinerea" Hymenochirus curtipes^ Rana pipiens (waxworm)^ R. pipiens (earthworm)" Caudata Ambystoma dumerilii^ A. mexicanum^ A. mexicanum^ A. ordinarium^ A. tigrinum—larvae^ A. tigrinum^ A. tigrinum^ Amphiuma means ^ Bolitoglossa occidentalism' B. mexicana^ Cryptobranchus allegheniensis ^ Cynops pyrrhogaster^ Desmognathus fuscus ^' D. marmoratus^ D. quadramaculatus^ D. quadramaculatus' Dicamptodon tenebrosus—larvae^ Ensatina eschscholtzii^ Hynobius nebulosus^ H. kimurae^ Necturus maculosus ^ Pachytriton brevipes^ P. brevipes^ Plethodon glutinosus^ Pleurodeles waltl^ P. waltl' Paramesotriton hongkongensis" Salamandra salamandra" Salamandrina terdigitata^ Siren intermedia ^ Taricha torosa'' Tylototriton verrucosus''
Duration of the Gape Cycle in Amphibians^ Source
Range (msec)
Mean (± SE)
42-58 400-1733 367-4734 300-1033
51.8 ± 1.8 1028 ± 73 1819 ± 189 572 ± 41
O'Reilly O'Reilly O'Reilly O'Reilly
80-280 83-158
-140 126.5 ± 7.6 143 ± 22 134 152 ± 8.1 62.7 ± 4.2 100.8 ± 6.4 163.2 ± 9.8
Nishikawa and Cannatella (1991) O'Reilly (1995) Cans and Gorniak (1982) Nishikawa and Roth (1991) Deban and Nishikawa (1992) O'Reilly (1995) Anderson (1993) Anderson (1993)
88 ± 3.5 59 ± 1.8 69.7 ± 1.2 73 ± 2.6 83 75 107 72 ± 1.9 99.6 ± 1 . 4 108 ± 2 53.3 ± 1.1 162 ± 16.3 92.6 ± 19.6 128 136.1 ± 6.1 109.9 ± 1.1 54.4 ± 2.5 87.4 ± 4.9 115.7 ± 7.2 94.4 ± 5.7 51 ± 1 79 ± 10.9 44.4 ± 2.4 96.3 ± 6.3 75.9 ± 3.2 220 ± 21.2 211 ± 30.3 100 ± 7.1 240 ± 23.1 61 ± 2.8 246 160 ± 10.7
Shaffer and Lauder (1985) Reilly and Lauder (1992) Shaffer and Lauder (1985) Shaffer and Lauder (1985) Shaffer and Lauder (1988) Shaffer and Lauder (1988) Shaffer and Lauder (1988) Reilly and Lauder (1992) Larsen et al (1989) Larsen et al (1989) Reilly and Lauder (1992) Miller and Larsen (1990) Larsen and Beneski (1988) Schwenk and Wake (1993) Larsen and Beneski (1988) Larsen et al (1989) Reilly and Lauder (1992) Larsen et al (1989) Larsen et al (1989) Larsen et al (1989) Reilly and Lauder (1992) Miller and Larsen (1990) O'Reilly (1995) Larsen et al (1989) O'Reilly (1995) Miller and Larsen (1990) Miller and Larsen (1990) Miller and Larsen (1990) Miller and Larsen (1990) Reilly and Lauder (1992) Findeis and Bemis (1990) Miller and Larsen (1990)
92-325 42-108
84-288
50-112 25-58 50-125 144-312 64-468 60-136 90-326
72-248
(1995) (1995) (1995) (1995)
^Note how long the gape cycles of adult caecilians are relative to frogs and salamanders. ^Aquatic feeding. '^Terrestrial feeding.
C. Intraoral Transport After prey is seized, caecilians use different methods of prey transport, depending on the medium in which they are feeding (air vs water) and the size and type
of prey being eaten. Larvae of Epicrionops have only been observed using hydraulic transport (O'Reilly, 1995), despite the presence of a well-developed tongue (Wake, 1989). Metamorphosed caecilians employ two or three methods of transport, depending on the
160
James C. O'Reilly
F I G U R E 6.6. (Left) Selected video frames of terrestrial prey capture in an adult Hypogeophis rostratus. Note the forward lunge, lack of tongue protraction, and lack of hyobranchial depression. (Right) Selected video frames of aquatic prey capture in an adult Typhlonedes nutans. Note the forward lunge, lack of tongue protraction, and substantial hyobranchial depression during mouth closing.
161
6. F e e d i n g in Caecilians
AAAAAA
^
AAAAAA
AAAAAA
45°
^
90°
E o CQ
40
^
20 0,
Gape Angle (")
20
Gape Angle (°)
40
(
Gape Angle {°)
F I G U R E 6.7. Model of the relationship of mechanical advantage and gape angle in the three major arrangements of the retroarticular process and interhyoideus found in living caecilians. Dashed lines with arrows represent force applied by the interhyoideus. Arrangement A has its best mechanical advantage at high gape angles. Arrangements B and C reach peak performance at low gape angles. Model assumes identical force input and retroarticular process length across all systems.
situation. On land, caecilians use a combination of inertial transport (releasing prey and lunging over it, using the mass of the prey as a counter weight) and lingual transport on larger prey items, while using lingual transport alone on smaller items (Bemis et ah, 1983; O'Reilly, 1995). In water, adults add hydraulic transport sequences to the mix (O'Reilly, 1995). During transport, caecilians often use longitudinal spinning to help subdue prey (Tanner, 1971; Bemis et al, 1983), the function of which is still not well understood. Larger arthropods (such as large crickets) are often rubbed against the substrate to disable or remove their limbs.
IV. EVOLUTION Living caecilians represent an ancient lineage and began to diverge from one another over 100 million
years ago (Hedges et al, 1993). The only fossil material that has been attributed to Gymnophiona that includes cranial material is Eocaecilia micropodia from the early Jurassic. This species is thought to be the closest known outgroup to living caecilians Jenkins and Walsh, 1993). Unlike living caecilians, Eocaecilia had limbs and limb girdles. Its skull was solidly roofed, lending some support to arguments that the ancestor of living caecilians lacked fenestrae in the skull (Carroll and Currie, 1975; Jenkins and Walsh, 1993). However, read Wake and Hanken (1982) and Nussbaum (1983) for counterarguments and reviews of the available evidence. The issue of whether the solidly roofed skull has secondarily evolved in living caecilians as an adaptation for burrowing or is homologous to the condition seen in the first tetrapods will only be settled by a vast improvement in the available fossil record of the ancestors of living amphibians. For the purposes of this
162
James C. O'Reilly
chapter, I will largely follow the scenario of Nussbaum (1977,1983) while acknowledging that the fenestrated condition of adult rhinatrematids may have evolved within that lineage from a solid skulled ancestor through paedomorphosis. The common ancestor of living caecilians was most likely oviparous, with the female guarding her clutch of eggs until they hatched into free-living, aquatic larvae. These larvae localized prey by the combined use of olfaction and electroreception, after which they used suction feeding for prey capture. The skull had large temporal fenestrae and was strongly kinetic, especially in the case of movement of the quadrate relative to the braincase (streptostyly). The hyobranchial apparatus may have been cartilaginous or ossified, but almost certainly possessed many independent elements that articulated such that posterior-directed force applied to the basibranchial series was translated into ventrally directed force expanding the buccopharynx. The flow of water that carried prey into the mouth was focused by well-developed labial lobes. The teeth where pedicellate and best developed in the front of the mouth, where two rows where present on both the upper and the lower jaws. On the upper jaw, the outer (premaxillary-maxillary) series of teeth did not extend as far caudally as the inner (vomer-palatine) series. In contrast, on the lower jaw, the outer row of teeth was more extensive than the inner row. After capture, hydraulic transport was used to transport prey to the pharynx. After spending many months as a larva, this common ancestor metamorphosed into a terrestrial, semifossorial adult. During metamorphosis, the skull became more solidly constructed and less kinetic due to the fusion of some dermal elements that were independent in larvae. However, the skull still had large temporal fenestrae, components of the levator mandibulae complex originating on the midline of the skull and passing through a large gap between the squamosal and the parietal before inserting on the lower jaw. Although the quadrate was now connected more firmly to an expanded squamosal, it could still move mediolaterally relative to the neurocranium during prey transport. The lateral line system, not being useful on land, was lost at metamorphosis. The lacrimal (tear) ducts were used to transport chemicals directly to the vomeronasal organ from the eye, possibly allowing the animals to smell even when the nostrils were pressed into the soil during burrowing. This adult used jaw prehension to capture prey, with the levator mandibulae complex being aided during mouth closing by an unusually oriented interhyoideus muscle inserting on the ventral edge of the retroaricular process of the lower jaw. The dentition consisted of two rows of numerous
teeth in both the upper and the lower jaws, with all four rows extending caudally to the corner of the gape. The mouth opening was terminal, looking like that of an adult salamander. The first evolutionary changes from the just-described ancestral pattern concerned the terrestrial adult stage, with the characteristics of the larval stage remaining relatively constant. As the adult stage became more specialized for burrowing, some signficant changes took place in head anatomy. The mouth opening became sub terminal, with the lower jaw being recessed into the upper jaw. The tear duct chemosensory system became more elaborate, with extrinsic eye muscles and other parts of the eye being incorporated into the now protusible tentacle apparatus with an independent aperture located well in front of the orbit. Important transitions during the evolution of the caecilian feeding apparatus include changes in the arrangement of the jaw levators and the degree to which the adductor chamber is roofed by dermal bones. The ancestor of living caecilians apparently possessed a skull that was broadly open between the squamosal and the temporal. Prominent jaw levators originated on the top of the skull and traveled down through these fenestrae and inserted on the lower jaw. This primitive condition is still seen in larval caecilians and in adult rhinatrematids. In a common ancestor of all caecilians other than rhinatrematids, the squamosal became much larger, completely covering the temporal fenestra. Concurrently, the jaw levators were reduced in size, being restricted to the now din\inutive adductor chamber under the squamosal. Temporal fenestrae have subsequently reappeared convergently in three groups {Geotrypetes, Scolecotnorphus, and typhlonectids), but in all of these cases the jaw levators have remained restricted to the adductor chamber. Transitions in the arrangement of the retroarticular process and interhyoideus posterior were also important. A common ancestor of caecilians had a small retroarticular process and an interhyoideus muscle, which originated midventrally and inserted on the ceratohyals. The original shift of insertion to the retroarticular process and subsequent enlargement of the process and muscle probably originally occurred as an adaptation for breathing. As the ancestors of caecilians became more elongate in body form, the ratio of the volume of the buccal chamber to that of the lungs became progressively smaller. Moving the insertion of the interhyoideus to the retroarticular process would increase the stroke volume of the buccal pump and at least partially compensate for these changes (Carrier and Wake, 1995). This condition is closely approximated by the condition seen in living rhinatrematids.
163
6. Feeding in Caecilians After the initial shift in the insertion of the interhyoideus, ancestral caecilians made major rearrangements in the jaw apparatus as they became progressively more fossorial. First, there was a concurrent increase in the size of both the interhyoideus and the retroarticular process. At least three times (in ichthyophiids and uraeotyphlids, scolecomorphids, and Atretochoana), the relatively straight retroarticular process has been recurved radically upward. In ichthyophiids and uraeotyphlids, the process is curved posterior to the jaw articulation, whereas in scolecomorphids and Atretochoana the curve occurs before the jaw articulation, such that the articular surface lies in the vertical rather than the horizontal plane. The primitive arrangement of the jaw muscles and skull presents at least two problems for active burrowers that use their heads as the only means of penetration. First, large levators necessarily increase the diameter of the head. Because elongate limbless burrowers use compression to create tunnels, the cost of constructing a given length of tunnel is exponentially related to tunnel diameter (Gans, 1974) and they are most likely under strong selection to minimize body diameter. Because the interhyoideus is tucked behind the skull, relying on it as the primary jaw-closing muscle permitted a vast increase in maximum jawclosing forces with no increase in head diameter. Muscle is also more subject to injury when exposed to crushing forces than is bone. Moving the primary jawclosing muscles behind the skull and roofing the levators with bone would allow more violent burrowing movements than were possible in ancestral caecilians. While the unique arrangement of caecilian jaw-closing muscles is a spectacular adaptation for head-first burrowing, it may have also led to a severe constraint on caecilian cranial evolution. The interhyoideus, ancestrally functioning only as a breathing muscle, has at least one new function (jaw closing) and its position suggests that it plays a prominent role in head movements during burrowing as well (Lawson, 1965b). Condensing all of these functions to a single muscle potentially conserves an enormous amount of muscle volume that would otherwise increase body diameter. However, as single components take on multiple functions, their evolution tends to become severely constrained (Lauder and Liem, 1989). In most living caecilians the interhyodius is the dominant muscle in breathing, jaw closing, and side-to-side head movements during burrowing and, despite its great age, we see remarkably little variation in its arrangement. Interestingly, the only great experiment in caecilian head design is seen in Atretochoana, an animal that does not breathe and is apparently not an accomplished bur-
rower (Nussbaum and Wilkinson, 1995; Wilkinson and Nussbaum, 1997). V. THE FUTURE Caecilians are the most poorly known major group of tetrapods. Regardless of why this is the case, it makes them a gold mine of opportunity for any researcher willing to make the extra effort to acquire and work with these animals. The study of caecilian feeding is no exception. Despite a relatively good understanding of variation in caecilian cranial anatomy, there is a dearth of experimental studies on every aspect of caecilian feeding behavior. The neurophysiology and neuroethology of caecilian prey capture remain to be described. There are only two experimental studies on the sensory systems of caecilians (Himstedt and Fritzsch, 1990; Himstedt and Simon, 1995). Thus we still know very little about how any caecilian perceives its surroundings and finds prey, let alone how prey detection abilities vary among different species. The relative importance of different types of stimuli in different environments and the cellular physiology of various receptors are currently unknown. The motor control of feeding movements and how the neural connections of the interhyoideus have changed during its integration into the jaw closing system also await study. There is currently a single published paper (Bemis et ah, 1983) describing prey capture and transport in any detail and this treats only a single species {Dermophis). There are no detailed descriptions of the feeding behavior of larval or fetal caecilians. There are no quantitative studies on the functional significance of the differences among the five major arrangements of the adult caecilian feeding apparatus, including variation in feeding kinematics, bite force generation, or prey transport behavior. The variation in the contractile characteristics of the different jaw adductors and how this is related to the various biomechanical arrangements seen in living caecilians is also unknown. Although cranial kinesis is widespread among caecilians (Nussbaum, 1977; Wilkinson and Nussbaum, 1997), we know little about its potential functional significance. Ultimately, it would be ideal if the just-mentioned studies were placed in ecological, ontogenetic, and evolutionary contexts. Understanding how caecilians interact with their actual environment would be the ultimate test of the relevance of laboratory studies. Understanding how function varies during ontogeny will provide a complete picture of the demands placed on
164
James C. O'Reilly
the feeding system as well as the opportunities available through heterochrony during evolution. Finally, placing future studies in an explicit phylogenetic context will allow the synthesis of functional, ecological, and developmental data into a robust portrait of caecilian evolution. Acknowledgments First, I must extend my deep appreciation to Kurt Schwenk for his encouragement and patience during the writing of this chapter. Adam Summers and Nate Kley provided many helpful comments on earlier drafts of this manuscript. Ron Nussbaum and Mark Wilkinson introduced me to caecilian biology and have continued to provide mentoring and insight in all things caecilian for the last 12 years. Ron Nussbaum also supplied many of the animals on which the observations herein are based. Edmund Brodie, Jr., Jonathan Campbell, Louis Porras (Zooherp, Inc., Sandy, Utah), Rob Maclnnes (Glades Herp. Inc., Ft. Myers, Florida), and Ed Budziak aided in the acquisition of specimens. Carl Cans provided logistical and financial support during my initial forays into the topic of caecilian prey capture. Steve Deban and Dale Ritter helped in the videotaping of feeding sequences. Kiisa Nishikawa provided extensive advice and support throughout all aspects of this project. This work was supported by NSF Grants IBN-8909937 and IBN-9211310 to Kiisa Nishikawa, a grant-in-aid-of-research from Sigma Xi, and a Darwin Postdoctoral Fellowship from the Organismic and Evolutionary Biology Program at the University of Massachusetts Amherst.
References Anderson, C. W. (1993) The modulation of feeding behavior in response to prey type in the frog Rana pipiens. J. Exp. Biol. 179:1-12. Badenhorst, A. (1978) The development and the phylogeny of the organ of Jacobson and the tentacular apparatus of Ichthyophis glutinosus (Linne). Annale Universiteit van Stellenbosch, Serie A2 (Soologie) 1:1-26. Barbour, T., and A. Loveridge (1928) A comparative study of the herpetological faunae of the Uluguru and Usambara Mountains, Tanganyika Territory, with descriptions of new species. Mem. Museum Comp. Zool. 50:87-265. Bemis, W. E., K. Schwenk, and M. H. Wake (1983) Morphology and function of the feeding apparatus in Dermophis mexicanus (Amphibia: Gymnophiona). Zool. J. Linn. Soc. 77:75-96. Billo, R., and M. H. Wake (1987) Tentacle development in Dermophis mexicanus (Amphibia: Gymnophiona), with an hypothesis of tentacle origin. J. Morph. 192:101-111. Breckenridge, W. R., S. Nathanael, and L. Pereira (1987) Some aspects of the biology and development of Ichthyophis glutinosus (Amphibia: Gymnophiona). J. Zool. Lond. 211:437-449. Carrier, D. R., and M. H. Wake (1995) Mechanism of lung ventillation in the caecilian Dermophis mexicanus. J. Morph. 226:289-295. Carroll, R. L., and P. J. Currie (1975) Microsaurs as possible apodan ancestors. Zool. J. Linnean Soc. 57:229-247. Clairambault, R, M.-J. Cordier-Picouet, and C. Pairault (1980) Premieres donnees sur les projections visuelles d'un Amphibien Apode {Typhlonectes compressicauda), C. R. Acad. Sci. Ser. D. 291: 283-286. Deban, S. M., and K. C. Nishikawa (1992) The kinematics of prey capture and the mechanism of tongue protraction in the green tree frog, Hyla cinerea. J. Exp. Biol. 170:235-256.
Ducey, P. K., D. R. Formanowicz, L. Boyet, J. Mailloux, and R. A. Nussbaum (1993) Experimental examination of burrowing behavior in caecilians (Amphibia, Gymnophiona): effects of soil compaction on burrowing ability of four species. Herpetologica 49:450-457. Edgeworth, F. H. (1925) On the autosylism of Dipnoi and Amphibia. J. Anat. 59:225-264. Edgeworth, F. H. (1935) The Cranial muscles of Vertebrates. Cambridge University Press, Cambridge. Engelhardt, F. (1924) Tentakelapparat und auge von Ichthyophis. Jenaische Zeitschrift fuer Naturwissenschaft 60:241-304. Exbrayat, J.-M., and M. Delsol (1985) Reproduction and growth of Typhlonectes compressicaudus: a viviparous gymnophione. Copeia 1985:950-955. Findeis, E. K., and W. E. Bemis (1990) Functional morphology of tongue projection in Taricha torosa (Urodela: Salamandridae). Zool. J. Linn. Soc. 99:129-157. Fox, H. (1985) The tentacles of Ichthyophis (Amphibia: Caecilia) with special reference to the skin. J. Zool., Lond. (A) 205:223-234. Fritzsch, B. (1988) Phylogenetic and ontogenetic origin of the dorsolateral auditory nucleus of anurans. Pp. 561-585. In: The evolution of the amphibian auditory system. B. Fritzsch, M. Ryan, W. Wilczynski, W. Hetherington, and W. Walkowiak (eds.). Wiley, New York. Fritzsch, B., W. Himstedt, and M.-D. Crapon de Caprona (1985) Visual projections in larval Ichthyophis kohtaoensis (Amphibia: Gymnophiona). Dev. Brain. Res. 23:201-210. Fritzsch, B., U. Wahnschaffe, M.-D. Crapon de Caprona, and W. Himstedt (1985) Anatomical evidence for electroreception in larval Ichthyophis kohtaoensis. Naturwiss 72:102-104. Fritzsch, B., and M. H. Wake (1986) The distribution of ampuUary organs in Gymnophiona. J. Herp. 20:90-93. Fritzsch, B., and M. H. Wake (1988) The inner ear of gymnophione amphibians and its nerve supply: a comparative study of regressive events in a complex sensory system (Amphibia: Gymnophiona). Zoomorph. 108:201-217. Gans, C. (1974) Biomechanics: An Approach to Vertebrate Biology. The University of Michigan Press, Ann Arbor, MI. Gans, C , and G. C. Gorniak (1982) Functional morphology of lingual protrusion in marine toads {Bufo marinus). Am. J. Anat. 163: 195-222. Gaymer, R. (1971) New method of locomotion in limbless terrestrial vertebrates. Nature 234:150-152. Hebrard, J. J., G. M. O. Maloiy, and D. M. I. Alliangana (1992) Notes on the habitat and diet of Afrocaecilia taitana (Amphibia, Gymnophiona). J. Herp. 26:513-515. Hedges, S. B., R. A. Nussbaum, and L. Maxson (1993) Caecilian phylogeny and biogeography inferred from mitochondrial DNA sequences of the 12S and 16S rRNA genes (Amphibia: Gymnophiona). Herp. Monogr. 7:64-76. Hetherington, T. E., and M. H. Wake (1979) The lateral line system in larval Ichthyophis (Amphibia: Gymnophiona). Zoomorph. 93: 209-225. Himstedt, W. (1991) Zur biologie und larvenentwicklung der thailandischen blindwiihle Ichthyophis kohtaoensis (Amphibia: Gymnophiona: Ichthyophiidae). Veroff. Naturhist. Mus. Schleusingen 6:16-24. Himstedt, W. (1995) Structure and function of the eyes of the caecilian Ichthyophis kohtaoensis (Amphibia: Gymnophiona). Zool. (Jena) 99:81-94. Himstedt, W. (1996) Die Blindwiihlen. Die Neue Brehm-Bucherei, Vol. 360. Westarp/Spektrum, Madgeburg/Heidelberg. Himstedt, W., and B. Fritzsch (1990) Behavioural evidence for electroreception in larvae of the caecilian Ichthyophis kohtaoensis (Amphibia, Gymnophiona). Zoologische Jahrbiicher-Abteilung fiir Allgemeine Zoologie und Physiologie der Tiere 94:486-492.
6. F e e d i n g in Caecilians Himstedt, W., and G. Manteuffel (1985) Retinal projections in the caecilian Ichthyophis kohtaoensis (Amphibia: Gymnophiona). Cell Tissue Res. 239:689-692. Himstedt, W., and D. Simon (1995) Sensory basis of foraging behaviour in caecilians (Amphibia: Gymnophiona). Herp. J. 5:266-270. Jenkins, F. A., and D. M. Walsh (1993) An early Jurassic caecilian with limbs. Nature 365:246-250. Largen, M. J., P. A. Morris, and D. W. Yalden (1972) Observations on the caecilian Geotrypetes grandisonae Taylor (Amphibia: Gymnophiona) from Ethiopia. Monit. Zool. Ital. 4:185-205. Larsen, J. H., and J. T. Beneski (1988) Quantitative analysis of feeding kinematics in dusky salamanders (Desmognathus). Can. J. Zool. 66:1309-1317. Larsen, J. H., J. T. Beneski, and D. B. Wake (1989) Hyolingual feeding systems of the Plethodontidae: comparative kinematics of prey capture by salamanders with free and attached tongues. J. Exp. Zool. 252:25-33. Lauder, G. V., and K. F. Liem (1989) The role of historical factors in the evolution of complex organismal functions. Pp. 63-78. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, New York. Lauder, G. V., and S. M. Reilly (1993) Amphibian feeding behavior: comparative biomechanics and evolution. Pp. 163-195. In: Biomechanics of Feeding in Vertebrates. V. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Lawson, R. (1965a) The teeth of Hypogeophis rostratus (Amphibia, Apoda) and tooth structure in the Amphibia. Proc. Zool. Soc. London 145:321-326. Lawson, R. (1965b) The anatomy of Hypogeophis rostratus. II. The musculature. Proc. Univ. Newcastle Tyne Phil. Soc. 1:52-63. Lewis, E. R., E. L. Leverenz, and W. Bialek (1985). The Vertebrate Inner Ear. CRC Press, Boca Raton, FL. Lewis, E. R., and R. E. Lombard (1988) The amphibian inner ear. Pp. 93-123. In: The Evolution of the Amphibian Auditory System. B. Fritzsch, M. Ryan, W. Wilczynski, T. H. Hetherington, and W. Walkowiak (eds.). Wiley, New York. Loveridge, A. (1936) Scientific results of an expedition to rain forest regions in eastern Africa. VII. Amphibians. Bull. Mus. Comp. Zool. 79:369-430. Miller, B. T. and J. H. Larsen (1990) Comparative kinematics of terrestrial prey capture in salamanders and newts (Amphibia: Urodela: Salamandridae). J. Exp. Zool. 256:135-153. Moll, E. O., and H. M. Smith (1967) Lizards in the diet of an American caecilian. Nat. Hist. Miscell. Chicago Acad. Sci. 187:1-2. Moodie, G. E. E. (1978) Observations on the life history of the caecilian Typhlonectes compressicaudus (Dumeril and Bibron) in the Amazon Basin. Can. J. Zool. 56:1005-1008. Nishikawa, K. C , and D. C. Cannatella (1991) Kinematics of prey capture in the tailed frog, Ascaphus truei (Anura: Ascaphidae). Zool. J. Linn. Soc. 103:289-307. Nishikawa, K. C , and G. Roth (1991) The mechanism of tongue protraction during prey capture in the frog, Discoglossus pictus. J. Exp. Biol. 159:217-234. Nishikawa, K. C , C. W. Anderson, S. M. Deban, and J. C. O'Reilly (1992) The evolution of the neural circuits controlling feeding behavior in frogs. Brain Behav. Evol. 40:125-140. Nussbaum, R. A. (1977) Rhinatrematidae: a new family of caecilians (Amphibia: Gymnophiona). Occasional Papers of the Museum of Zoology, University of Michigan 682:1-30. Nussbaum, R. A. (1983) The evolution of a unique dual jaw-closing mechanism in caecilians (Amphibia: Gymnophiona) and its bearing on caecilian ancestry. J. Zool. 199:545-554. Nussbaum, R. A. (1985a) Amphibian Fauna of the Seychelles Archipelago. National Geographic Research Reports 18:53-62. Nussbaum, R. A. (1985b) Systematics of the caecilians (Amphibia:
165
Gymnophiona) of the family Scolecomorphidae. Occasional Papers of the Museum of Zoology, University of Michigan 713:1-49. Nussbaum, R. A. (1992) Caecilians. Pp. 52-59 In: Reptiles and Amphibians. H. G. Cogger and R. G. Zweifel (eds.). Smithmark, New York. Nussbaum, R. A., and P. K. Ducey (1988) Cytological evidence for the monophyly of the caecilians (Amphibia: Gymnophiona) of the Seychelles Archipelago. Herpetologica 44:290-296. Nussbaum, R. A., and H. Hinkel (1994) Revision of east African caecilians of the genera Afrocaecilia Taylor and Boulengerula Tornier (Amphibia: Gymnophiona: Caeciliidae). Copeia 1994:750-760. Nussbaum R. A., and M. E. Pfrender (1998) Revision of the African caecilian genus Schistometopum Parker (Amphibia: Gymnophiona: Caeciliidae). Miscellaneous Publications, Museum of Zoology, University of Michigan 187:1-32. Nussbaum, R. A., and M. Wilkinson (1989) On the classification and phytogeny of caecilians (Amphibia: Gymnophiona), a critical review. Herp. Monog. 3:1-42. Nussbaum, R. A., and M. Wilkinson (1995) A new genus of lungless tetrapod: a radically divergent caecilian (Amphibia: Gymnophiona). Proc. Royal Soc. Lond. Ser. B 261:331-335. O'Reilly, J. C. (1995) Comparative Studies on the Evolution of Prey Capture Behavior in Amphibians. M. S. Thesis. Northern Arizona University. O'Reilly, J. C. (1996) Keeping caecilians in captivity. Adv. Herpetocult. 1:39-45. O'Reilly, J. C , R. A. Nussbaum, and D. Boone (1996) A vertebrate with protrusible eyes. Nature 382:33. O'Reilly, J. C , D. A. Ritter, and D. R. Carrier (1997) Hydrostatic locomotion in a limbless tetrapod. Nature 386:269-272. Parker, H. W. (1956) Viviparous caecilians and amphibian phylogeny. Nature 178:250-252. Parker, H. W. (1958) Caecilians of the Seychelles Islands with description of a new subspecies. Copeia 1958:71-76. Parker, H. W., and E. R. Dunn (1964) Dentitional metamorphosis in the Amphibia. Copeia 1964:75-86. Reilly, S. M., and G. V. Lauder (1992) Morphology, behavior and evolution: comparative kinematics of aquatic feeding salamanders. Brain Behav. Evol. 40:182-196. Reiss, J. (1996) Palatal metamorphosis in basal caecilians (Amphibia: Gymnophiona) as evidence for lissamphibian monophyly. J. Herp. 30:27-39. Roth, G., K. C. Nishikawa, C. Naujoks-Manteuffel, A. Schmidt, and D. B. Wake (1993) Paedomorphosis and simplification in the nervous system of salamanders. Brain Behav. Evol. 42:137-70. Sanderson, I. (1937) Animal Treasure. Viking Press, New York. Sarasin, P., and F. Sarasin (1887-1890) Ergebnisse naturwissenschaftlichen Forschungen auf Ceylon in den Jahren 1884-1886. Zur Entwicklunggeschichte und Anatomie der Ceylonesischen Blindwuhle. Ichthyophis glutinosus. C. W. Kreidel, Weisbaden. Schmidt, A., and M. H. Wake (1990) Olfactory and vomeronasal systems of caecilians (Amphibia-Gymnophiona). J. Morph. 205:255268. Schwenk, K., and D. B. Wake (1993) Prey processing in Leurognathus marmoratus and the evolution of form and function in desmognathine salamanders (Plethodontidae). Biol. J. Linn. Soc. 49:141162. Shaffer, H. B., and G. V. Lauder (1985) Patterns of variation in aquatic ambystomatid salamanders: kinematics of the feeding mechanism. Evolution 39:83-92. Shaffer, H. B., and G. V. Lauder (1988) The ontogeny of functional design: metamorphosis of feeding behavior in the tiger salamander {Ambystoma tigrinum). J. Zool., Lond. 216:437-454. Summers, A. P., and J. C. O'Reilly (1997) A comparative study of the locomotor modes in two species of caecilian, Dermophis mexicanus
166
James C. O'Reilly
and Typhlonectes nutans (Amphibia: Gymnophiona). Zool. J. Linn. Soc. 121:65-76. Tanner K. (1971) Notizen zur pflege und zum verhalten einiger blindwiihlen (Amphibia: Gymnophiona). Salamandra 7 : 9 1 100. Taylor, E. H. (1968) The Caecilians of the World: A Taxonomic Review. Univ. of Kansas Press, Lawrence, KS. Taylor, E. H. (1969) Skulls of the Gymnophiona and their significance in the taxonomy group. Univ. Kansas Sci. Bull. 48:585-689. Taylor, E. H. (1970) The lateral-line system in the caecilian family Ichthyophiidae (Amphibia: Gymnophiona). Univ. Kansas Sci. Bull. 48:861-868. Van Damme, J., and P. Aerts (1997) Kinematics and functional morphology of aquatic feeding in Australian snake-necked turtles (Pleurodira: Chelodina). J. Morph. 233:113-125. Visser, M. H. C. (1963) The cranial morphology of Ichthyophis glutinosus (Linne) and Ichthyophis monochrous (Bleeker). Annale Universiteit van Stellenbosch, Ser. A 38:67-102. Wahnschaffe, U., B. Fritzsch, and W. Himstedt (1985) The fine structure of the lateral-line organs of larval Ichthyophis (Amphibia: Gymnophiona). J. Morph. 186:369-377. Wake, M. H. (1976) The development and replacement of teeth in viviparous caecilians. J. Morph. 148:33-63. Wake, M. H. (1977a) The reproductive biology of caecilians: an evolutionary perspective. Pp. 73-101. In: Reproductive Biology of Amphibians. E. H. Taylor and S. I. Guttman (eds.). Plenum, New York. Wake, M. H. (1977b) Fetal maintenance and its evolutionary significance in the amphibia: Gymnophiona. J. Herp. 11:379-386. Wake, M. H. (1978) Comments on the ontogeny of Typhlonectes obesus, particularly its dentition and feeding. Papeis Avulsos de Zoologia32:l-13. Wake, M. H. (1980a) Reproduction, growth and population structure of the Central American caecilian Dermophis mexicanus. Herpetologica 36:244-256. Wake, M. H. (1980b) Fetal tooth development and adult replacement in Dermophis mexicanus (Amphibia: Gymnophiona): fields versus clones. J. Morph. 166:203-216. Wake, M. H. (1985) The comparative morphology and evolution of the eyes of caecilians (Amphibia, Gymnophiona). Zoomorph. 105:277-295. Wake, M. H. (1987) A new genus of African caecilian (Amphibia: Gymnophiona). J. Herp. 21:6-15. Wake, M. H. (1989) Metamorphosis of the hyobranchial apparatus in Epicrionops (Amphibia, Gymnophiona, Rhinatrematidae): replacement of bone by cartilage. Ann. Sci. Nat. Zool. Biol. Anim. 10:171-182. Wake, M. H. (1992) Reproduction in caecilians. Pp. 112-120 In: Re-
productive Biology of South American Vertebrates. W. C. Hamlett (ed.). Springer-Verlag, New York. Wake, M. H. (1993a) Non-traditional characters in the assessment of caecilian phylogenetic relationships. Herp. Monog. 7:42-55. Wake, M. H. (1993b) The skull as a locomotor organ. Pp. 197-240 In: The Skull Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Wake, M. H. (1993c) Evolution of oviductal gestation in amphibians. J. Exp. Zool. 266:394-413. Wake, M. H., and J. Hanken (1982) The development of the skull of Dermophis mexicanus (Amphibia: Gymnophiona), with comments on skull kinesis and amphibian relationships. J. Morph. 173:203223. Wake, M. H., and G. Z. Wurst (1979) Tooth crown morphology in caecilians (Amphibia: Gymnophiona). J. Morph. 159:331-340. Welsch, U., M. Miiller, and C. Schubert (1977) Elektronenmikroskopische und histochemische beobachtungen zur fortpflanzungs vivaprer gymnophionen (Chthonerpeton indistinctum). Zoologische Jahrbuch von Anatomie 97:532-549. Wever, E. G. (1975) The caecilian ear. J. Exp. Zool. 191:63-72. Wever, E. G., and C. Gans (1976) The caecilian ear: further observations. Proc. Natl. Acad. Sci. USA 73:3744-3746. Wiedersheim, R. (1879) Die Anatomie der Gymnophionen. Verlag Gustav Fischer, Jena. Wilkinson, M. (1989) On the status of Nectocaecilia fasciata Taylor, with a discussion of the phylogeny of the Typhlonectidae (Amphibia: Gymnophiona). Herpetologica 45:23-36. Wilkinson, M. (1991) Adult tooth crown morphology in the Typhlonectidae (Amphibia: Gymnophiona): a reinterpretation of variation and its significance. Zeitschrift fiir zoologische Systematische un,d Evolutions-forschung 29:304-311. Wilkinson, M. (1992a) The phylogenetic position of the Rhinatrematidae (Amphibia: Gymnophiona): evidence from the larval lateral line system. Amphibia-Reptilia 13:74-79. Wilkinson, M. (1992b) On the life history of the caecilian genus Uraeotyphlus (Amphibia: Gyrrmophiona). Herp. J. 2:121-124. Wilkinson, M. (1997) Characters, congruence and quality: a study of neuroanatomical and traditional data in caecilian phylogeny. Biol. Rev. 72:423-470. Wilkinson, M., and R. A. Nussbaum (1996) On the phylogenetic position of the Uraeotyphlidae (Amphibia: Gymnophiona). Copeia 1996:550-562. Wilkinson, M., and R. A. Nussbaum (1997) Comparative morphology and evolution of the lungless caecilian Atretochoana eiselti (Taylor) (Amphibia: Gymnophiona: Typhlonectidae). Biol. J. Linn. Soc. 62:39-109. Wilkinson, M., and R. A. Nussbaum (1998) Caecilian viviparity and amniote origins. J. Nat. Hist. 32:1403-1409.
S E C T I O N
III
^--\.-M^^^:
This Page Intentionally Left Blank
C H A P T E R
7 A Bibliography of Turtle Feeding KURT SCHWENK Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut 06269
I. INTRODUCTION 11. BIBLIOGRAPHY
lonian Bauplan suggests some kind of uniformity. Our comparative ignorance of feeding function in turtles hampers our ability to discern general patterns in the history of tetrapod feeding. The fact that they are an ancient group that might retain ancestral amniote features in some aspects of their feeding system makes their further study all the more desirable.
I. INTRODUCTION The bibliography that follows is hardly exhaustive, but it does serve as an introduction to the literature on turtle feeding. General references on natural history and diet are not included, but a number of important papers on digestive physiology are. Also omitted are references on the systematics of turtles. The emphasis here is on the morphology of the feeding apparatus, feeding function, and behavioral observations of feeding in turtles. Historically, the functional morphology of feeding in turtles has been sadly neglected, but several recent contributions suggest that this is changing. In any case, turtle feeding is badly in need of investigation. One of the attractive features of turtles, as a group, is their diversity. Among other things, they span the range from fully terrestrial to fully aquatic, with everything in between. Thus, turtles, like salamanders, offer the opportunity to examine phenotypic changes in the feeding system associated with a change in the feeding medium (see Bramble and Wake, 1985; Lauder, 1985). Terrestrial turtles share with mammals and lepidosaurs a mobile, muscular tongue, but the intrinsic anatomy, biomechanics, and function of the turtle tongue are virtually unstudied. In general, there is a great deal of morphological diversity among turtle feeding systems that is largely unappreciated, perhaps because the che-
FEEDING (K. Schwenk, ed.)
II. BIBLIOGRAPHY Beisser, C J., J. Weisgram, and H. Splechtna (1995) Dorsal lingual epithelium of Platemys pallidipectoris (Pleurodira, Chelidae). J. Morph. 226:267-276. Beisser, C. J., J. Weisgram, H. Hilgers, and H. Splechtna (1998) Fine structure of the dorsal lingual epithelium of Trachemys scripta elegans (Chelonia: Emydidae). Anat. Rec. 250:127-135. Belkin, D. A., and C. Gans (1968) An unusual chelonian feeding niche. Ecology 49:768-769. Dels, V. L., and S. Renous (1991) Kinematics of feeding in two marine turtles {Chelonia mydas and Dermochelys coriacea). Pp. 73-78. In: Proceedings of the 6th Ordinary General Meeting of the Societas Europaea Herpetologica. Z. Korsos and I. Kiss (eds.). Hungarian Natural History Museum, Budapest. Bels, V. L., J. Davenport, and V. Delheusy (1997) Kinematic analysis of the feeding behavior in the box turtle Terrapene Carolina (L.), (Reptilia: Emydidae). J. Exp. Zool. 277:198-212. Bjorndal, K. A. (1980) Nutrition and grazing behavior of the green turtle Chelonia mydas. Mar. Biol. 56:147-154. Bjorndal, K. A. (1985) Nutritional ecology of sea turtles. Copeia 1985: 736-751. Bjorndal, K. A. (1986) Effect of solitary vs group feeding on intake in Pseudemys nelsoni. Copeia 1986:234-235. Bjorndal, K. A. (1987) Digestive efficiency in a temperate herbivorous reptile, Gopherus polyphemus. Copeia 1987:714-720. Bjorndal, K. A. (1989) Flexibility of digestive responses in two gen-
169
Copyright © 2000 by Academic Press. All rights of reproductior\ ir\ ar\y form reserved.
170
Kurt Schwenk
eralist herbivores, the tortoises Geochelone carbonaria and Geochelone denticulata. Oecologia 78:317-321. Bjomdal, K. A. (1990) Digestive processing in a herbivorous freshwater turtle: consequences of small-intestine fermentation. Physiol. Zool. 63:1232-1247. Bjorndal, K. A. (1990) Digestibility of the sponge Chondrilla nucula in the green turtle, Chelonia mydas. Bull. Mar. Sci. 47:567-570. Bjorndal, K. A. (1991) Digestive fermentation in green turtles, Chelonia mydas, feeding on algae. Bull. Mar. Sci. 48:166-171. Bjorndal, K. A. (1992) Body size and digestive efficiency in a herbivorous freshwater turtle: advantages of small bite size. Physiol. Zool. 65:1028-1039. Bjorndal, K. A. (1993) Digestive efficiencies in herbivorous and omnivorous freshwater turtles on plant diets: do herbivores have a nutritional advantage. Physiol. Zool. 66:384-395. Bjomdal, K. A. (1997) Fermentation in reptiles and amphibians. Pp. 199-230. In: Gastrointestinal Microbiology, Vol. 1. R. I. Mackie and B. A. White (eds.). Chapman and Hall, New York. Bjorndal, K. A. (1997) Foraging ecology and nutrition of sea turtles. Pp. 199-321. In: The Biology of Sea Turtles. P L. Lutz and J. A. Musick (eds.). CRC Press, Boca Raton, LA. Bramble, D. M. (1971) Functional Morphology, Evolution, and Paleoecology of Gopher Tortoises. Unpublished doctoral dissertation, Univ. of California, Berkeley. Bramble, D. M. (1973) Media dependent feeding in turtles. Am. Zool. 13:1342. [abstract] Bramble, D. M. (1974) Occurrence and significance of the Os transiliens in gopher turtles. Copeia 1974:102-109. Bramble, D. M. (1978) Functional analysis of underwater feeding in the snapping turtle. Am. Zool. 18:623. [abstract] Bramble, D. M. (1980) Feeding in tortoises and mammals: why so similar? Am. Zool. 20:931. [abstract] Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Carmignani, M. P. A., and G. Zaccone (1975) Histochemical distribution of acid mucopolysaccarides in the tongue of reptiles. I. Chelonia {Pseudemys scripta Clark). Ann. Histochim. 20:77-88. Dalrymple, G. H. (1977) Intraspecific variation in the cranial feeding mechanism of turtles of the genus Trionyx (Reptilia, Testudines, Trionychidae). J. Herp. 11:255-285. Dalrymple, G. H. (1979) Packaging problems of head retraction in trionychid turtles. Copeia 1979:655-660. Davenport, J., M. Spikes, S. M. Thornton, and B. O'Kelly (1992) Crabeating in the diamondback terrapin Malaclemys terrapin: dealing with dangerous prey. J. Mar. Biol. Assoc. U. K. 72:835-848. Davenport, J., T. M. Wong, and J. East (1992) Feeding and digestion in the omnivorous estuarine turtle Batagur baska (Gray). Herp. J. 2:133-139. Drummond, H., and E. R. Gordon (1979) Luring in the neonate alligator snapping turtle {Macroclemys temminckii): description and experimental analysis. Z. Tierpsychol. 50:136-152. Edgeworth, F. H. (1935) The Cranial Muscles of Vertebrates. Cambridge Univ. Press, Cambridge. Fenchel, T. M., C. P McRoy, J. C. Ogden, P Parker, and W. E. Rainey (1979) Symbiotic cellulose degradation in green turtles, Chelonia mydas L. Appl. Environ. Microbiol. 37:348-350. Ferdinand, L. Prinz von Bayern (1884) Anatomic der Zunge. Fine vergleichend-anatomische Studie. Literarisch-Artistische Anstalt (Theodor Riedel), Munich. Fuchs, H. (1907) IJber das Hyobranchialskelett von Fmys lutaria und seine Entwicklung. Anat. Anz. 31:33-39. Fiirbringer, M. (1922) Das Zungenbein der Wirbeltiere insbesondere der Reptilien und Vogel. Abh. der Heidelberger Akad., math.naturw.Kl. 11:1-164.
Gaffney, E. S. (1972) An illustrated glossary of turtle skull nomenclature. Am. Mus. Novit. No. 2486:1-33. Gaffney, E. S. (1979) Comparative cranial morphology of Recent and fossil turtles. Bull. Am. Mus. Nat. Hist. 164(2): 67-376. George, J. C , and R. V. Shah (1954) The myology of the head and neck of the common Indian pond turtle, Lissemys punctata granosa Schoepff. J. Anim. Morphol. Physiol. 1:1-12. George, J. C , and R. V. Shah (1955) The myology of the head and neck of the Indian tortoise, Testudo elegans. J. Anim. Morphol. Physiol. 2:1-13. Gnanamuthu, C. P. (1937) Comparative study of the hyoid and tongue of some typical genera of reptiles. Proc. Zool. Soc. Lond. Ser.B 1937:1-63 Graper, L. (1932) Die das Zungenbein und die Zunge bewegenden Muskeln der Schildkroten I. Jena. Z. Naturwiss. 66:169-198. Graper, L. (1932) Die das Zungenbein und die Zunge bewegenden Muskeln der Schildkroten II. Jena. Z. Naturwiss. 66:274-280. Hailey, A. (1997) Digestive efficiency and gut morphology of omnivorous and herbivorous African tortoises. Can. J. Zool. 75:787794. Iwasaki, S.-I. (1992) Fine structure of the dorsal epithelium of the tongue of the freshwater turtle, Geoclemys reevesii (Chelonia, Emydinae). J. Morph. 211:125-135. Iwasaki, S.-I., T. Asami, Y. Asami, and K. Kobayashi (1992) Fine structure of the dorsal epithelium of the tongue of the of the Japanese terrapin, Clemmys japonica (Cheloia [sic], Emydinae). Arch. Histol.Cytol. 55:295-305. Iwasaki, S.-I., T. Asami, and C. Wanichanon (1996) Ultrastructural study of the dorsal lingual epithelium of the soft-shelled turtle, Trionyx cartilagineus (Cheloia [sic], Trionychidae). Anat. Rec. 246: 305-316. Iwasaki, S.-I., T. Asami, and C. Wanichanon (1996) Fine structure of the dorsal lingual epithelium of the juvenile hawksbill turtle, Eretmochelys imbricata bissa. Anat. Rec. 244:437-443. Iwasaki, S.-I., C. Wanichanon, and T. Asami (1996) Histological and ultrastructural study of the lingual epithelium of the juvenile Pacific ridley turtle, Lepidochelys olivacea (Chelonia, Cheloniidae). Ann. Anat. 178:243-250. Iwasaki, S.-I., C. Wanichanon, and T. Asami (1996) Ultrastructural study of the dorsal lingual epithelium of the Asian snail-eating turtle, Malayemys subtrijuga. Ann. Anat. 178:145-152. Kochva, E. (1978) Oral glands of Reptilia. Pp. 43-161. In: Biology of the Reptilia, Vol. 8. C. Gans and K. A. Cans (eds.). Academic Press, New York. Korte, G. E. (1980) Ultrastructure of the tastebuds of the red-eared turtle, Chrysemys scripta elegans. J. Morph. 163:231-252. Lakjer, T. (1926) Studien Uber die Trigeminus-versorgte Kaumuskulatur der Sauropsiden. C. A. Rietzel, Copenhagen. Lauder, G. V. (1985) Aquatic feeding in lower vertebrates. Pp. 210229. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Lauder, G. V., and T. Prendergast (1992) Kinematics of aquatic prey capture in the snapping turtle Chelydra serpentina. J. Exp. Biol. 164:55-78. Lee, M. S. Y. (1997) The evolution of beaks in reptiles: a proposed evolutionary constraint. Evol. Theor. 11:249-254. Legler, J. M. (1962) The Os transiliens in two species of tortoises, genus Gopher us. Herpetologica 18:68-69. Legler, J. M. (1976) Feeding habits of some Australian short-necked tortoises. Victorian Nat. 93:40-43. Legler, J. M. (1978) Observations on behavior and ecology in an Australian turtle, Chelodina expansa (Testudines: Chelidae). Can. J. Zool. 56:2449-2453. Legler, J. M. (1989) Diet and head size in Australian chelid turtles, genus Emydura. Ann. Soc. Roy. Zool. Belgique 119, Suppl. 1:1-10.
7. A Bibliography of Turtle F e e d i n g Legler, J. M. (1993) Family Chelidae. Pp. 142-152. In: Fauna of Australia, Vol. 2A. C. J. Glasby G. J. B. Ross, and P. L. Beesley (eds.). Australian Government Publ. Service, Canberra. Legler, J. M. (1993) Morphology and physiology of the Chelonia. Pp. 108-119. In: Fauna of Australia, Vol. 2A. C. J. Glasby G. J. B. Ross, and P. L. Beesley (eds.). Australian Government Publ. Service, Canberra. Lemell, P., and J. Weisgram (1997) Feeding patterns of Pelusios castaneus (Chelonia: Pleurodira). Neth. J. Zool. 47:429-441. Lubosch, W. (1933) Untersuchungen iiber die Visceral muskulatur der Sauropsiden. (Der Untersuchungen iiber die Kaumuskulatur der Wirbeltiere 3. Teil.). Gegenbaurs Morph. Jb. 72:584-666. Lubosch, W. (1938) Muskeln des Kopfes: Viscerale Muskulatur. Pp. 1011-1106. In: Handbuch der Vergleichenden Anatomie der Wirbeltiere, Vol. 5. L. Bolk, E. Goppert, E. Kallius, and W. Lubosch (eds.). Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Marlow, R. W., and K. Tollestrup (1982) Mining and exploitation of natural mineral deposits by the desert tortoise, Gopherus agassizii. Anim. Behav. 30:475-478. Meyer, V., and L. Prutkin (1974) An ultrastructural study of the oral mucous membrane of the turtle, Pseudemys scripta elegans. Acta Anat. 89:89-99. Meylan, A. (1988) Spongivory in hawksbill turtles: a diet of glass. Science 239:393-395 Nalavalde, M. N., and A. T. Varute (1976) Histochemical studies on the mucins of the vertebrate tongue. VIII. Histochemical analysis of mucosubstances in the tongue of the turtle. Folia Histochem. Cytochem. 14:123-134. Owen, R. (1866) On the Anatomy of Vertebrates, Vol. 1. Longmans, Green, and Co., London. Parsons, T. S. (1968) Variation in the choanal structure of Recent turtles. Can. J. Zool. 46:1235-1263. Pevzner, R. A., and N. A. Tikhonova (1979) Fine structure of the taste buds of the Reptilia. I. Chelonia. Tsitologiya 21:132-138. Poglayen-Neuwall, I. (1953) Untersuchungen der Kiefermuskulatur und deren Innervation bei Schildkroten. Acta Zool. 34:241-291. Poglayen-Neuwall, I. (1953/54) Die Besonderheiten der Kiefermuskulatur von Dermochelys coriacea. Anat. Anz. 100:22-32. Pritchard, P. C. H. (1971) The leatherback or leathery turtle, Dermochelys coriacea. lUCN Monograph 1:1-39. Ray, C. E. (1959) A sesamoid bone in the jaw musculature of Gopherus polyphemus (Reptilia: Testudinidae). Anat. Anz. 107:85-91. Rhodin, A. G. J., F Medem, and R. A. Mittermeier (1981) The occurrence of neustophagia among podocnemine turtles. Br. J. Herp. 6:175-176. Romer, A. S. (1956) Osteology of the Reptiles. Univ. of Chicago Press, Chicago. Ruckes, H. (1937) The lateral arcades of certain emydids and testudinids. Herpetologica 1:97-103. Schumacher, G.-H. (1953/1954) Beitrage zur Kiefermuskulatur der Schildkroten. I. Mitteilung. Wiss. Z. Univ. Greifswald Math. Nat. 3:457-518. Schumacher, G.-H. (1954/1955) Beitrage zur Kiefermuskulatur der Schildkroten. II. Mitteilung. Wiss. Z. Univ. Greifswald Math. Nat. 4:501-518. Schumacher, G.-H. (1954/1955) Beitrage zur Kiefermuskulatur der Schildkroten. III. Mitteilung. Wiss. Z. Univ. Greifswald Math. Nat. 4:559-601. Schumacher, G.-H. (1956) Morphologische Studie zum Gleitmechanismus des. M. adductor mandibularis externus bei Schildkroten. Anat. Anz. 103:1-12. Schumacher, G.-H. (1956) Uber die Fascien des Kopfes der nebst einigen Bemerkungen zu der Arbeit von Tage Lakjer 1926. Zool. Anz. 156:35-54.
171
Schumacher, G.-H. (1973) Die Kopf- und Halsregion der Lederschildkrote. Anatomische Untersuchungen im Vergleich zu anderen rezenten Schildkroten. Abhandl. Akad. Wissensch. DDR, No. 2. Akademie-Verlag, Berlin. Schumacher, G.-H. (1973) The head muscles and hyolaryngeal skeleton of turtles and crocodilians. Pp. 101-199. In: Biology of the Reptilia, Vol. 4. C. Gans and T S. Parsons (eds.). Academic Press, New York. Secor, S. M., and J. Diamond (1999) Maintenance of digestive performance in the turtles Chelydra serpentina, Sternotherus odoratus, and Trachemys scripta. Copeia 1999:75-84. Sewertzoff, S. A. (1929) Zur Entwicklungsgeschichte der Zunge bei den Reptilien. Acta Zool. 10:231-341. Shah, R. V. (1963) The neck musculature of a cryptodire (Deirochelys) and a pleurodire (Ghelodina) compared. Bull. Mus. Comp. Zool. 129:343-368. Siebenrock, F (1898) Uber den Bau und die Entwicklung des Zungenbein-apparates der Schildkroten. Ann. Naturhist. Hofmus. Wien. 13:424-437. Smith, D. T. J. (1989) The Cranial Morphology of Fossil and Living Sea Turtles (Cheloniidae, Dermochelyidae and Desmatochelyidae). Unpublished doctoral dissertation, Kingston Polytechnic University, United Kingdom. Sondhi, K. C. (1958) The hyoid and associated structures in some Indian reptiles. Ann. Zool. Acad. Zool. (Lond.) 2:155-239. Spindel, E. L., J. L. Dobie, and D. F Buxton (1987) Functional mechanisms and histologic composition of the lingual appendage in the alligator snapping turtle, Macroclemys temmincki (Troost) (Testudines: Chelydridae). J. Morph. 194:287-301. Summers, A. P., K. F. Darouian, A. M. Richmond, and E. L. Brainerd (1998) Kinematics of aquatic and terrestrial prey capture in Terrapene Carolina, with implications for the evolution of feeding in cryptodire turtles. J. Exp. Zool. 281:280-287. Thompson, J. S. (1932) The anatomy of the tortoise. Sci. Proc. Roy. Soc. Dublin 20:359-461. Uchida, T (1989) Ultrastructural and histochemical studies on the taste buds in some reptiles. Arch Histol. Jap. 43:459-478. Van Damme, J., and P. Aerts (1997) Kinematics and functional morphology of aquatic feeding in Australian snake-necked turtles (Pleurodira; Chelodina). J. Morph. 233:113-125. Versluys, J. (1936) Kranium und Visceralskelett der Sauropsiden. 1. Reptilien. Pp. 699-808. In: Handbuch der Vergleichenden Anatomie der Wirbeltiere. Vol. 4. L. Bolk, E. Goppert, E. Kallius, and W. Lubosch (eds.). Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Vogt, R. C , D. M. Sever, and G. Moreira (1998) Esophageal papillae in pelomedusid turtles. J. Herp. 32:279-282. Weisgram, J. (1985) Feeding mechanics of Claudius angustatus Cope 1865. Pp. 257-260. In: Functional Morphology in Vertebrates (Fortschr. Zool. Vol. 30). H.-R. Duncker and G. Fleischer (eds.). Gustav Fischer, Stuttgart. Weisgram, J., and H. Splechtna (1990) Intervertebral movability in the neck of two turtle species {Testudo hermanni hermanni, Pelomedusa subrufa). Zool. Jb. Anat. 120:425-431. Weisgram, J., and H. Splechtna (1992) Cervical movement during feeding in Chelodina novaeguinaeae (Chelonia, Pleurodira). Zool. Jb. Anat. 122:331-337. Weisgram, J., H. Ditrich, and H. Splechtna (1989) Comparative functional anatomical study of the oral cavity in two turtle species. Plzen. Lek. Sborn., Suppl. 59:117-122. Winokur, R. M. (1988) The buccopharyngeal mucosa of the turtles (Testudines). J. Morph. 196:33-52. Wocheslander, R., H. Hilgers, and J. Weisgram (1999) Feeding mechanism of Testudo hermanni boettgeri (Chelonia, Cryptodira). Neth. J. Zool. 49:1-13.
This Page Intentionally Left Blank
S E C T I O N
IV
(p/sBo^e SCUWEMK t ^ ^
This Page Intentionally Left Blank
C H A P T E R
8 Feeding in Lepidosaurs KURT SCHWENK
Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut 06269
I. INTRODUCTION
I. INTRODUCTION 11. LEPIDOSAURIANPHYLOGENY AND CLASSIFICATION III. NATURAL HISTORY A. Diet B. Foraging Ecology C. Sensory Basis of Food Location and Identification IV. MORPHOLOGY OF THE FEEDING APPARATUS A. Skull and Mandible B. Dentition C. Hyobranchial Apparatus D. Jaw Musculature E. Tongue V. FEEDING FUNCTION A. Overview of Feeding B. Feeding Stages C. Feeding in Sphenodon D. Feeding in Iguania E. Feeding in Scleroglossa F. Biomechanics of Lingual Prey Capture G. Function of Cranial Kinesis VI. SPECIALIZED FEEDING SYSTEMS A. Chameleons B. Amphisbaenians C. Komodo Monitor D. Snakes VII. THE EVOLUTION OF FEEDING IN LEPIDOSAURS A. Evolution of Ingestion Mode B. Post-Ingestion Feeding Stages C. Evolution of the Gape Cycle D. Tongue Evolution E. Dietary Specialization F. Feeding Systems, Functional Units, and Evolutionary Constraint VIII. FUTURE DIRECTIONS References
FEEDING (K. Schwenk, ed.)
This chapter considers the structure, function, and evolution of the feeding system in nonophidian lepidosaurs—tuatara, lizards, and amphisbaenians. The latter two groups comprise, along with snakes, the squamate reptiles (Squamata). Although snakes are cladistically nested within squamates, their feeding systems have diverged sufficiently from other taxa to merit separate treatment (Chapter 9). They are, however, considered in this chapter in a general sense, as in the discussion of evolutionary patterns within Lepidosauria. Lepidosaurs offer a number of attributes that make them attractive subjects for study in the context of tetrapod feeding mechanisms. First, they are phylogenetically well positioned to be informative about evolutionary trends and patterns in the tetrapod clade. Perhaps more to the point is that many of them apparently retain a relatively primitive, or at least generalized, phenotype as compared to other living amniotes and so provide better structural analogues for reconstructing the ancestral-feeding mode. For example, most lepidosaurs retain an unspecialized, welldeveloped hyobranchial apparatus and a mobile, complexly muscled tongue. The latter trait they share with mammals and turtles and so we can infer its presence in the common anmiote ancestor. The muscular tongue is of considerable intrinsic interest. Squamate tongues haveprovided part of the empirical basis for the development of the muscular hydrostatic model of movement (e.g., Kier and Smith, 1985; Smith and Kier, 1989) and, due to their relatively more predictable
175
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
176
Kurt Schwenk
kinematics relative to mammal tongues, they are receiving increasing attention from modelers (e.g., Chiel et ah, 1992; van Leeuwen, 1997; van Leeuwen and de Groot, in preparation). Second, the feeding apparatus, particularly the tongue, is highly variable among lepidosaurs and so provides the grist for basic evolutionary studies, including phylogenetic analyses (e.g., Schwenk, 1988, and references therein), as well as studies of evolutionary pattern and process (e.g., Robinson, 1967; Schwenk, 1993b, 1995a,b; Bels et al, 1994; Kardong et al, 1997; Wagner and Schwenk, 2000). The lepidosaurian-feeding system is of particular interest in the latter case because there has been a fundamental shift in feeding mode within the group from tongue-based ingestion to jaw-based ingestion, corresponding cladistically to the origin of the Scleroglossa, one of two basal squamate clades (see later). The acquisition of jaw prehension, therefore, provides a model system for studying transitions in complex, functionally integrated systems (Schwenk, 1993b, 1995a, 2000b; Wagner and Schwenk, 2000). A related point is that the squamate tongue subserves two fundamentally different functions: feeding and vomeronasal chemoreception. A biomechanical consideration of tongue design suggests that a tongue phenotypically optimized for feeding is a poor chemoreceptor, whereas a tongue optimized for chemoreception is poorly designed for feeding function (Schwenk, 1993b, 1995a; Wagner and Schwenk, 1999). Hence, there is an evolutionary tension between these two functions that plays out in the evolution of tongue form. Clade-specific solutions to the "dilemma" posed by this evolutionary trade-off provide insight into underlying processes of phenotypic evolution (Schwenk, 1995a, 2000b, in preparation; Wagner and Schwenk, 2000).
IL LEPIDOSAURIAN PHYLOGENY A N D CLASSIFICATION Lepidosauria is a diverse clade of reptiles comprising approximately 7150 species of tuatara, lizards, snakes, and amphisbaenians (Pough et al, 1998). It is the sister group of Archosauria, which includes crocodilians, birds, and various extinct diapsid reptiles, such as the dinosaurs (Gauthier et al, 1988), or of turtles (Testudines) plus archosaurs (e.g.. Hedges and Poling, 1999; Kumazawa and Nishida, 1999) (Fig. 8.1). Lepidosauria is further divided into the Rhynchocephalia and Squamata, the former containing two species of tuatara, genus Sphenodon (Daugherty et al, 1990), and the latter, all remaining lepidosaurian species. Squamates are, themselves, divided into two basal clades.
the Iguania and Scleroglossa, and these, in turn, are subdivided into several suprafamilial groups (Fig. 8.1). A molecular study of nuclear and mitochondrial gene sequences suggested that Sphenodon is more closely related to archosaurs (+ turtles) than to squamates, thus splitting the Lepidosauria as presently construed (Hedges and Poling, 1999). Such a phylogenetic hypothesis is extremely unlikely in the face of morphological data. It would deny the 35 morphological synapomorphies identified by Gauthier et al (1988) uniting Sphenodon and Squamata relative to all other amniotes. Furthermore, one would find it very difficult to identify morphological synapomorphies uniting tuatara (+ fossil sphenodontids) with a clade including crocodilians, birds, and turtles. Morphology overwhelmingly supports a monophyletic Lepidosauria. Lepidosauria is an ancient group with fossil lizards known from the Upper Permian (approximately 250 mybp) and sphenodontids from the Triassic (Estes, 1983; Carroll, 1988b). Many Late Cretaceous (approximately 7b mybp) fossil species are assignable to modern families and some Late Jurassic (135+ mybp) taxa are recognizable as varanoids related to living monitor lizards and snakes (Estes, 1983). In traditional classifications, Lepidosauria is accorded the rank of subclass with Rhynchocephalia and Squamata as orders within it (e.g., Romer, 1956). Lizards (Lacertilia or Sauria), snakes (Serpentes or Ophidia), and amphisbaenians (Amphisbaenia) are given equal categorical ranking as suborders within the order Squamata, despite the fact that relationships among these groups remain poorly understood. Amphisbaenians were historically regarded as a family of "lizards" and called Amphisbaenidae in suit (e.g.. Camp, 1923), but later work suggested that these unusual, fossorial squamates were quite distinct from "typical" lizards and deserving of subordinal status in equality with lizards and snakes (e.g.. Cans, 1978; Crook and Parsons, 1980; Bellairs and Cans, 1982). Needless to say, such traditional classifications were crafted by workers in the context of "evolutionary taxonomy" rather than phylogenetic (cladistic) systematics, thus they were not overly concerned that the classification mirror phylogenetic relationships among the groups. Rather, recognition of Serpentes, Lacertilia, and Amphisbaenia as separate but equal ranks within Squamata calls attention to the relative morphological distinctness of each taxon. Indeed, this distinctness has been the basis for much of the ambiguity regarding the phylogenetic position of snakes and amphisbaenians relative to lizards. It has sometimes been suggested that each of these taxa is completely outside the others, i.e., monophyletic (e.g., Hoffstetter, 1968; Rieppel, 1978c, 1983; see Rieppel, 1988, for a review), but most
m IGUANIAI :::::!$:
% 1
a> C0
o a>
I
"D
P
®
5>
<
O
^?;: SCI^ R O 3LOSSA i'^^$^$^^^^
mmmmmjig^i^^iigiilggggii^^; X3
a>
0)
O
a>
5
CO
a 5 o
.t-
CD CO
to
^
CO
•o
a> E
r o(0
_j
T3
(0
CO
X
o c
;}5
•D
o
O
CO
"5 c
<
CO
o c X
:::$:::$;:;WS. CD CO
%
To
g
a> 5CD
I
en
O
®
cCO
x: CO -J
CD
s.
'c 2CO
CD
2
<^ >
LEPIDOSAURIA
e IGUANIA iiii
m scLEROGLossA w^^^^m
0) CO •o CD CO
•D
'c
CO 3
o>
0)
E CO O)
<
c o CD
CD
(D CO
CO
"O
CO JZ
CD
E
O
a >
2
^
o O
n
iS
d> (B
CD CO
TD (D CO
5 O) c
<
3
CO CO O
c CD X
I
E
CD
X
CD CO
2 I (0 -J
CO CO >
Q.
Thecoglossa
Anguimorpha Diploglossa
SQUAMATA I FIGURE 8.1. (A) Phylogeny of Lepidosauria based on Estes et ah (1988). Amphisbaenia and Dibamidae are not included because their positions are uncertain; however, most evidence suggests placement of Amphisbaenia within Scleroglossa. Estes et al. (1988) declined to place snakes (Serpentes) in the phylogeny, but most data support varanoid affinities. (B) Squamate relationships based on Lee (1998). Taxa that differ in their placement relative to (A) are shown in bold type.
178
Kurt Schwenk
studies suggest that snakes and amphisbaenians are nested within the "lizard" clade (e.g.. Camp, 1923; Estes et al, 1988; Schwenk, 1988; Caldwell and Lee, 1997; Lee, 1998; Caldwell, 1999), thus rendering "Lacertilia," as traditionally conceived, a paraphyletic taxon. "Lizard," however, remains useful as an informal term for nonsnake, nonamphisbaenian squamates. Current consensus accepts a monophyletic Squamata comprising two basal clades, Iguania and Scleroglossa (Estes et al., 1988), with snakes and probably amphisbaenians as members of Scleroglossa (Fig. 8.1). Unfortunately, molecular analyses of higher-level squamate relationships that might help resolve these issues are thus far unconvincing, having been poorly conceived and methodologically flawed [see Macey and Verma's (1997) reanalysis and discussion of Forstner et al. (1995)]. Despite general acceptance of Iguania and Scleroglossa as monophyletic taxa, details of relationships within each major clade remain contentious. Iguania, for example, is traditionally held to comprise three families, Iguanidae, Agamidae, and Chamaeleonidae, with the latter two sister taxa (Camp, 1923; Estes et al, 1988). However, morphological evidence for monophyly of Iguanidae was never compelling (Etheridge and de Queiroz, 1988; Estes et al, 1988; Schwenk, 1988; Frost and Etheridge, 1989) and the paraphyly of Agamidae relative to Chamaeleonidae has been suspected (Estes et al, 1988; Frost and Etheridge, 1989); hence, on the basis of a cladistic analysis. Frost and Etheridge (1989) dispensed with the traditional three-family taxonomy and erected nine separate iguanian families based on those taxa for which monophyly was strongly supported (Etheridge and de Queiroz, 1988; Frost and Etheridge, 1989; reviewed by Schwenk, 1994d). This resulted in a radical taxonomy that, unfortunately, has been widely adopted in the herpetological literature (e.g., Conant and Collins, 1998; Pough et al, 1998). Nonetheless, molecular evidence has provided strong support for iguanid monophyly and limited support for agamid monophyly, potentially vindicating the traditional taxonomy. I therefore choose to retain the threefamily system here, as I have elsewhere [see Schwenk (1994d) for additional arguments in favor of this system]. Nevertheless, it must be borne in mind that taxonomies are fluid and subject to change in accordance with the shifting weight of character evidence. The case of Scleroglossa is equally contentious, if not more so. The placement of several groups, notably Amphisbaenia, Serpentes, Dibamidae, and Xantusiidae, is especially problematic. Phylogenetic uncertainties are compounded by the fact that these taxa are limbless and/or fossorial and therefore prone to morphological convergence (Estes et al, 1988; Lee, 1998; Lee and Caldwell, 1998). Estes et al (1988) felt that
character evidence was too ambiguous to situate some taxa, so they designated Serpentes, Amphisbaenia, and Dibamidae as Scleroglossa incertae sedis, i.e., of uncertain position within the group. There is a growing consensus, however, that snakes are anguimorphans,. probably derived from within the varanoids and possibly the sister group of the living monitor lizards, Varanidae and Lanthanotidae (e.g., McDowell and Bogert, 1954; Schwenk, 1988; Caldwell and Lee, 1997; Lee, 1997, 1998). However, Caldwell (1999) suggested that snakes are the sister taxon of all other scleroglossans (although in some of his analyses, snakes are associated with various groups within Scleroglossa, including varanoids). Placement of other taxa remains even less certain [see Lee (1998) for one recent treatment]. Given these ambiguities, I illustrate two possible phylogenies for Squamata based on Estes et al (1988) and Lee (1998) that differ in their placement of some scleroglossan taxa (Figs. 8.1A and 8.1B, respectively). Throughout this chapter I use the former phylogeny, but the differences have little effect on interpreting major patterns of feeding evolution in Lepidosauria. Finally, it is worth noting that in recent, phylogenetic classifications of lepidosaurs, traditional Linnean ranks above the family level are rarely used (e.g., Pough et al, 1998). This reflects a current trend toward adopting the principles of "phylogenetic taxonomy" in classification, a system that assigns names according to shared ancestry, which does not, therefore, impose categorical "ranks." Such ranks are found to be unnecessary in a phylogenetic system (see de Queiroz and Gauthier, 1992,1994; de Queiroz, 1996,1997).
III. NATURAL HISTORY A. Diet Greene (1982) noted that lizard diets had not been reviewed to that time. Remarkably, this remains true today. Greene's brief treatment remains one of the best, single sources of lizard dietary information. Greer (1989) and Pianka (1986) summarized diets for the diverse Australian lizard fauna and desert lizard communities, respectively. However, to gain a complete picture of lepidosaurian diets, one must scour the primary, natural history literature through which relevant data are peppered. A coniplete review is far beyond the scope of this chapter; however, a short consideration of diet is required to place the analysis of feeding form and function in context. It can be said, first, that lepidosaurian diets are diverse—so diverse that the range of food types eaten is nearly as broad as the dietary breadth of tetrapods as a
8. Feeding in Lepidosaurs whole. Some individual species have very broad diets and eat a wide range of food (euryphagy), whereas others eat only one or a few things (stenophagy). One striking pattern of lepidosaurian feeding evolution is that dietary specialization is often not accompanied by phenotypic specialization in the trophic (feeding) apparatus (Greene, 1982). Phenotypic specialization for stenophagy is infrequent among lepidosaurs, in contrast to mammals and birds (see later and Chapters 12 and 13). For a given specialized diet, such as ants, some taxa are highly derived and phenotypically specialized [e.g., Phrynosoma (Iguanidae) and Moloch (Agamidae)], whereas others remain generalized in form [e.g., Liolaemus monticola (Iguanidae)]. The reasons why some organisms become trophically specialized while others do not remains a fundamental question in evolutionary morphology (e.g., Greene, 1982; Lauder, 1983a,b). The question of dietary specialization is pursued further in Section VII. Iguania and Scleroglossa differ little in the breadth of diets they encompass, with one exception—no iguanian regularly eats prey items that are very large relative to its body size, whereas within Scleroglossa, several taxa are specialized in this way. The evolution of macrostomatan snakes, in particular, is characterized by adaptations that enhance their ability to subdue and ingest prey, which, in some cases, are larger than the snake itself (Greene, 1997; Chapter 9). Komodo dragons (Varanus komodoensis) kill large mammals and reduce them to sizable chunks, which are bolted with minimal processing (Auffenberg, 1981; see Section VI, C). A few other lepidosaurs regularly consume vertebrates (e.g., Sphenodon punctatus, Walls, 1981; Gambelia wislizenii, Tanner and Krogh, 1974; ToUestrup, 1979; Parker and Pianka, 1976; some Chamaeleo, Loveridge, 1953; Bourgat, 1970; Burrage, 1973; Eremiascincusjam.es and Losos, 1991; Heloderma, Bogert and Martin del Campo, 1956; Pregill et al, 1986; Lialis, Patchell and Shine, 1986; most Varanus, Shine, 1986; Losos and Greene, 1988; James et al, 1992; see also Greene, 1982), but for most of these species vertebrate prey are rare or occasional dietary items and few evince morphological specialization of the feeding apparatus for carnivory (the pygopodid, Lialis, specializes on skinks and may be the most trophically specialized). Otherwise, the vast majority of lepidosaurs feed frequently on relatively small, invertebrate prey. This is true even for some of the largest lepidosaurs, the varanid lizards (Losos and Greene, 1988), and it may be the single generalization that can be made about lepidosaurian diets. In a highly influential paper, Pough (1973) suggested a relationship between lizard size and diet such that small species (< 100 g) are nearly all carnivorous/insectivorous and large species (> 300 g)
179
are nearly all herbivorous, a pattern he attributed to the greater relative mass-specific energy requirements of smaller lizards. Pough suggested that the capture of small, active prey items is energetically untenable for large lizards and found that those groups that do maintain carnivory at a large body size have physiological, morphological, or behavioral adaptations which apparently circumvent this putative energetic constraint (see also Wilson and Lee, 1974). Pough suggested that juveniles of large species should show a dietary shift from carnivory/insectivory to herbivory as they achieve larger body sizes and found some support in the literature for this contention. Pough's (1973) conclusions can no longer be considered valid, however, except in broadest outline. Several factors mitigate against accepting the generality of his principal findings: (1) The study was done before phylogenetic awareness had infiltrated comparative ecological studies, hence it failed to take into account the effects of phylogeny on the observed correlation; in fact, many large, herbivorous species are closely related, hence represent single origins of herbivory and large body size, and the phylogenetic retention of this correlation within a clade, rather than multiple, independently evolved correlations. Thus, the apparent evolutionary "connection" between large size and herbivory was inflated falsely. (2) Many exceptions to the body size-diet correlation have since been noted. In particular, many very small species have been found to be purely, or mostly, herbivorous (e.g., Greene, 1982; Iverson, 1982; Schoener et al, 1982; Jaksic and Schwenk, 1983; Whitaker, 1968, 1987; Rocha, 1989, 1998; Schall and Ressel, 1991; Troyer, 1991; PerezMellado and Corti, 1993,1997; Vitt and Morato de Carvalho, 1995; Vitt and de la Torre, 1996; Vitt et al, 1997a). (3) Some putative ontogenetic dietary shifts from insectivory to herbivory have not been supported. For example, Swanson's (1950) report of diet switching in Iguana iguana, cited by Pough (1973), was speculative rather than empirical, and subsequent studies have found that even juvenile iguanas are fully herbivorous (Troyer, 1984). (4) Pough (1973) believed that lizard herbivores mostly lacked phenotypic specializations that would make them more efficient at digesting plant matter, but subsequent work has documented several adaptations for herbivory, notably hindgut fermentation systems for cellulose digestion (e.g., Iverson, 1982; McBee and McBee, 1982; Troyer, 1991; Bjorndal, 1997). In conclusion, although many large lizards are herbivorous, many are not. Most small lizards are insectivorous, but many are herbivorous. Certainly the number of small species for which plant matter constitutes a substantial portion of the diet is legion and grossly underestimated in the literature historically.
180
Kurt Schwenk
It has been generally assumed that the ancestral lepidosaurian diet was insectivory (e.g., Carroll, 1977, 1988a); however, this conclusion was based on the presumption of a correlation between tooth form and diet (e.g., Hotton, 1955; Montanucci, 1968), body size and diet (see earlier discussion), and fossil evidence showing that the earliest lepidosaurs were small with generalized teeth. However, the correlation between diet and tooth form is now known to be tenuous at best (see Section IV, B) and, as noted, the correlation between diet and body size is unsupported, hence one cannot assume that small, fossil lepidosaurs with generalized dentitions were insectivorous (Greene, 1982). Nonetheless, it remains true that by sheer weight of species number, the "typical" lepidosaur is a small, generalized insectivore that feeds opportunistically. Only comparative, phylogenetic analyses of diet in specific lepidosaurian clades will provide reasonable inferences about ancestral diets. Such analyses have yet to be done and their absence represents a large gap in our knowledge of feeding evolution in lepidosaurs. 1. Sphenodon Diet in tuatara has been studied by Walls (1981). Sphenodon is almost entirely insectivorous or carnivorous, but fecal pellet analysis showed more than an incidental quantity of plant matter (seeds) in a small percentage (approximately 10%) of individuals (Walls, 1981; I. C. Southey, in Whitaker, 1987). One fossil species shows convincing morphological evidence for herbivory, but virtually all other known fossil sphenodontids are presumed to have been insectivorous or carnivorous (Throckmorton et al, 1981). Tuatara forage at night and feed opportunistically on whatever they encounter and can capture, with certain noxious and fast-moving prey absent from their diet. Unlike most lizards, however, a powerful biting apparatus and specialized dentition (e.g., Robinson, 1967, 1976; Rosenberg et ah, 1982; see later) permit them to process relatively large prey items, including whole petrel (sea bird) eggs and chicks. Large chicks are killed by biting and are reduced to manageable pieces by the teeth. Although eggs and chicks are seasonally very important components of the diet, large beetles and other invertebrates are consistently the most important prey types. Tuatara showed some preference for forest, as opposed to open area, prey species. 2.
Amphisbaenians
Natural diets of amphisbaenians are poorly known. Most data are available for the genera Amphisbaena (Riley et al, 1986; ColH and Zamboni, 1999; White et al, 1992; Cabrera and Merlini, 1990; Cruz Neto and Abe,
1993; Cusumano and Powell, 1991) and Blanus (Lopez et ah, 1991; Gil et ah, 1993), with some information on Cercolophia (Cruz Neto and Abe, 1993), Monopeltis and Dalophia (Broadley et al, 1976). An association between ants and amphisbaenians has long been assumed but rarely proven. However, such an ecological relationship has been established for Amphisbaena alba in Trinidad, with some evidence suggesting that the pattern holds for mainland South America as well (Riley et al, 1986). These workers found that A. alba follows the pheromone-marked foraging trails of leaf-cutter ants {Atta) across the surface to the nest where the amphisbaenian occupies the refuse chamber. Within this chamber it feeds primarily on arthropods, especially beetle larvae, without attracting the attention of resident soldier ants. Surprisingly, relatively few ants were found in the diet and many of these were individuals of a raiding species. The most important dietary items were beetle larvae and other insects that are, themselves, inhabitants of ant nests. Colli and Zamboni (1999)'s study of Brazilian A. alba did not determine whether the amphisbaenian occupied ant nests, but it did reveal a large number of ants in stomachs of preserved specimens. Termites were the most numerous item, however, but unidentified insect larvae were the most important items by volume. This study, therefore, agrees with Riley et al. (1986) in identifying insect larvae as the most important food type, but differs in finding large numbers of termites and ants, suggesting that the Brazilian population exhibits more opportunistic foraging. Other Amphisbaena species were also found to consume large numbers of termites (Cabrera and Merlini, 1990; White et al, 1992; Cruz Neto and Abe, 1993). Cruz Neto and Abe (1993) collected many of their specimens next to fallen trees where termite nests were abundant. These studies also found that ants and insect larvae were important food items. Termites were also the most frequent prey type for the African genera Monopeltis and Dalophia, although both taxa also consumed significant numbers of ants and beetle larvae as well (Broadley et al, 1976). Blanus cinereus was found to consume a variety of soil invertebrates that tend to be most dense and diverse under rocks (Lopez et al, 1991; Gil et al, 1993). However, large insect larvae (particularly dipterans) are preferentially selected, presumably because of their higher caloric payload. Ants are also frequently taken, but not in proportion to their environmental abundance. Riley et al (1986) reviewed the meager amphisbaenian dietary literature at that time and concluded that most amphisbaenians feed opportunistically on fossorial arthropods encountered within their own fixed tunnel systems, although some species, including
8. Feeding in Lepidosaurs A. alba, may also occupy ant nests. Dietary data published since seems to confirm this view, although it increasingly appears that termites and, to a lesser extent, ants are important prey types. The predominance of termites in amphisbaenian diets may reflect dietary specialization or it may result from chance discovery of clumped prey items in shared habitat (Cabrera and Merlini, 1990). The latter interpretation is supported by Colli and Zamboni's (1999) finding that a few individuals ate very large numbers of termites, but most individuals had consumed few or none. White et at. (1992) noted a single unidentified vertebrate in one specimen of A. gonavensis, and Colli and Zamboni (1999) found evidence for only three vertebrate prey items (a small lizard limb, a small anomalepid snake, and some hair) in 116 specimens of A, alba. These findings do not support the hypothesis that the imposing biting apparatus of amphisbaenians is an adaptation for eating large prey (Cans, 1966,1969a, 1974; see Section VI). 3. Other Squamates Although snake diets are diverse, there are no herbivorous species (Pough, 1983; Greene, 1997), which is also true for all other limbless tetrapods (Cans, 1975). Pough (1983) speculated that this pattern might obtain from a constraint in gut size relative to body mass in an elongate body. It should be kept in mind, as well, that snakes evolved from insectivorous/carnivorous lizards and went on to evolve specialized, jaw-based prehension mechanisms adapted for capturing animal prey. Thus, snakes were committed early in their ancestry to a specialized, highly integrated feeding apparatus that might have limited their phenotypic (and therefore, dietary) options—the loss of a hyolingual transport mechanism (see later), alone, may prevent the effective transport and processing of most plant foods. In any case, under the rubric of "animal prey," snake diets comprise an astonishing variety of food types, including foamy nests of frog eggs, worms, slugs, ant eggs and larvae, bird eggs, fish, frogs, lizards, other snakes, antelope, and even the occasional human (see Greene, 1997; Chapter 9). Although "advanced" (macrostomatan) snakes are clearly specialized for the consumption of extremely large prey (in some cases >100% body mass), basal scolecophidian snakes eat many tiny prey (ant eggs and larvae, and termites), thus there are few generalizations one can make about snake diet. Nevertheless, diet and feeding habits appear to be central to understanding the origin of snakes and patterns of diversification within the group (Greene, 1983, 1997; Savitzky, 1983; see Chapter 9).
181
The diversity of lizard diets has already been treated in the introduction to this section. Little need be added other than to reiterate a few important points. Although the "typical" lizard, in the sense of species number, is a small, opportunistic insectivore, numerous exceptions exist. Some taxa are dietary specialists, including herbivorous (plant-eating), myrmecophagous (ant- and termite-eating), carnivorous (vertebrateeating), and oophagous (egg-eating) types. Within each of these dietary categories further specialization is often represented. For example, some herbivores eat primarily fruit (frugivory; e.g., Varanus olivaceous, Auffenburg, 1988), others leaves (folivory; e.g., I. iguana, Troyer, 1984; Rand et al, 1990), nectar (nectivory; e.g., some Hoplodactylus, Whitaker, 1987; Podarcis lilfordi, Perez-Mellado and Casas, 1997), or seeds [e.g., Angolosaurus skoogii and Meroles (Aporosaura) anchietae, Greene, 1982], and one species eats marine algae (Amblyrhynchus cristatus. Carpenter 1966). Importantly, however, diets often vary temporally, ontogenetically, or individually (e.g., Burrage, 1973; Jackson and Telford, 1975; Capel-Williams and Pratten, 1978; Vitt et al, 1981; Bauer et al, 1989; Dessem, 1985; Pianka, 1986; Shine, 1986; James, 1991; Perry and Brandeis, 1992; PerezMellado and Corti, 1993; Brown and Perez-Mellado, 1994; Rocha, 1996, 1998; Perez-Mellado and Casas, 1997; Whiting and Greeff, 1997; Znari and Nagy, 1997; Duffield and Bull, 1998). Thus, stenophagy may be seasonal or opportunistic. In some cases, of course, a narrow diet is maintained throughout life. It is therefore important to recognize that diet analyses based on samples limited geographically or temporally do not necessarily reveal the actual (complete) diet of a population or species. Undiscovered temporal shifts in diet driven by seasonality of food resources may explain why some putative dietary specialists fail to become phenotypically specialized. In other words, some species regarded as specialists based on inadequate dietary sampling might actually manifest serial, or opportunistic, stenophagy so that they remain functional generalists. Finally, there is the possibility that in some herbivorous species, stenophagy may be entrained in an individual, i.e., it is a phenotypically plastic trait. Although I know of no equivalent example in lizards, it has been shown that, as a species, green sea turtles (the only herbivorous marine turtle) are capable of eating both marine algae and sea grasses, but individuals eat only one or the other, even when both are available (Bjorndal, 1980, 1985). According to Bjorndal this pattern probably reflects individual specialization of the hindgut fermentation (cellulose-digesting) system for one or the other food type. Once established, a particular hindgut microbial fauna may be suitable for digesting only one type of food. Alternatively, variation in
182
Kurt Schwenk
food preferences among individuals within a species could be genetically determined, as has been shown for some garter snakes (e.g., Arnold, 1981). Although as a general rule dietary types characterize larger taxonomic groups, closely related species occasionally differ radically in diet. For example, virtually all monitor lizards (Varanus) feed exclusively on other animals (Losos and Greene, 1988), but, as noted earlier, V. olivaceous is one of the most highly frugivorous lizards in the world (Auffenberg, 1988)! Nearly all lizards of the genus Cnemidophorus (Teiidae) are insectivorous and large species are sometimes carnivorous, but a few species are primarily herbivorous (e.g., C. aruhensis, Schall and Ressel, 1991; C. lemniscatus, Vitt et ah, 1997a). Likewise, the vast majority of tropidurine iguanids, genus Liolaemus, are insectivores, but a few species (e.g., L. magellanicus, Jacksic and Schwenk, 1983; L. lutzae, Rocha, 1989; L. fitzingeri; Troyer, 1991) are mostly, or entirely, herbivorous and others are known to include fruits in their diets (e.g., L. pictus; Willson et al, 1996). Often the dietary shift in close relatives is not attended by apparent phenotypic specializations, but such differences can be cryptic. For example, although there are no obvious, diet-related differences in the cranial feeding apparatus of the Liolaemus species, Troyer (1991) showed that L. fitzingeri has an expanded large intestine as compared to its insectivorous relatives, suggesting the presence of a hindgut fermentation system for cellulose digestion (Troyer, 1991; Bjomdal, 1997).
B. Foraging Ecology The literature on foraging ecology of lepidosaurs is an area of ambiguity and contention that I am loathe to enter, but its obvious relevance to feeding biology requires at least a cursory treatment. As for other tetrapods, lepidosaurs are generally regarded as either "ambush" ("sit-and-wait") or "active" ("wide") foragers (e.g., Pianka, 1966, 1986; Schoener, 1971; Regal, 1978, 1983; Huey and Pianka, 1981). In their extreme expressions these foraging modes are qualitatively distinct: ambush foragers remain more or less motionless in position and wait for active prey to come within reach of a strike, whereas active foragers move more or less widely through their environments in search of prey that may be stationary, clumped, or cryptic. Thus diets tend to differ between ambush and active foragers living in the same habitat (e.g., Huey and Pianka, 1981). In general, active foragers have higher field metabolic rates and higher rates of energy expenditure than ambush foragers, but their food (energy) intake rates are significantly greater (Huey and Pianka, 1981;
Anderson and Karasov, 1981; Nagy et al, 1984; Waldschmidt et al, 1987). However, active foragers suffer from a higher risk of predation and lower relative clutch volumes (e.g., Huey and Pianka, 1981). Active foragers tend to rely more heavily on vomeronasal chemoreception than ambush foragers and are putatively capable of distinguishing prey from nonprey using chemosensory cues (Cooper, 1994, 1995a). They are also more proficient at following pheromonal trails than ambushers and consequently are often equipped with forked tongues (Schwenk, 1994e; see later). This dichotomy in foraging modes, as for most biological dichotomies, is oversimplified. The reality is far more complex and the biological correlates presented do not universally apply (see Perry, 1999). Many workers have suggested that ambush and active foraging modes represent extremes of a continuum, or have introduced additional qualitatively distinct or intermediate foraging modes (e.g.. Regal, 1978,1983; Magnusson et al, 1985; Pietruszka, 1986; Waldschmidt et al, 1987; Perry et al, 1990; Cooper, 1994, 1995a; Werner et al, 1997; Perry, 1999). There is little agreement on how to define terms precisely and what the relevant criteria for such definitions are. This disagreement has led to a lack of uniformity in the attribution of foraging modes among studies and among taxa. The latter case may be especially problematic; a study of a closely related species might score one as an ambush forager on a relative, dichotomous scale, but in an absolute (quantitative) sense, the same species would be considered an active forager in the context of a broader phylogenetic sample [see discussion of Perry's (1999) study later]. Further, some studies have identified significant among-individual variation, as well as ontogenetic, diurnal, and seasonal variation in foraging mode within a species (e.g., Taylor, 1986; Pietruszka, 1986; Perry et al, 1990; Duffield and Bull, 1998). Some taxa simply defy categorization. For example, nearly all iguanid lizards studied conform to our expectations for ambush foragers, but most horned lizards (Phrynosoma) do not (e.g., Huey and Pianka, 1981). This seems to be related to dietary specialization on ants, which are a clumped and unpredictable prey of the sort usually preyed on by active foragers. Thus horned lizards move widely through their environments seeking locations with active ants. Once found, however, they station themselves in place, behaving like typical ambushers. Nonetheless, they restrict themselves to short feeding bouts so that nests remain active (high predation rates cause ant nests to become inactive for days) and to avoid incurring the wrath of soldier ants (Whitford and Bryant, 1979; Rissing, 1981; Shaffer and Whitford, 1981), hence they soon move on again in search of other locations, behaving once again
8. Feeding in Lepidosaurs like an active forager. This pattern may also apply to Phrynosoma's agamid counterpart, Moloch (Pianka and Pianka, 1970). Likewise, herbivorous iguanids and agamids appear to be qualitatively distinct from their typical ambush foraging relatives, for which Cooper (1994) erected a third, diet-based foraging mode. Scincids, as a rule, do not conform to expectation (Regal, 1983), although they are embedded in a typically wide-foraging, clade. Regal (1978, 1983) suggested an intermediate category of "cruising'' to indicate the relatively slow-moving, stop-and-go nature of skink foraging. In any case, scincids are so incredibly diverse that they are likely to vary significantly in foraging mode (e.g., Greer, 1989). Finally, I note that most snakes do not conform to simple categorization. Many caenophidian snakes, such as vipers, would appear to be archetypal ambush foragers because they sit nearly motionless beside prey trails for hours, days, or even weeks until a prey animal comes within striking distance (e.g., Greene, 1997). Nonetheless, such a snake initially moves extensively through its habitat, seeking chemical-laden prey trails and appropriate ambush sites. Similarly, scolecophidian snakes move widely through their environments until they locate ant trails, which they follow to the nest wherein they feed on eggs and larvae (e.g., Gehlbach et al, 1971; Webb and Shine, 1992). For these reasons (among others), snakes exhibit extreme chemosensory prowess, deeply forked tongues, and unsurpassed trailing abilities (Halpern, 1992; Schwenk, 1994e), traits typically associated with wide foragers (see earlier discussion). These examples hint at the complexity of foraging mode among lepidosaurs and explain why many taxa exhibit characteristics of both typical ambush and active foragers. Finally, a study using a broad taxonomic base and comparative (phylogenetically controlled) methods rejected the notion of bimodality in foraging mode among lizards and supported a foraging continuum (Perry, 1999). However, Perry's (1999) data reveal that the continuum of foraging modes is not uniformly distributed among clades—iguanians (only iguanids were included) and gekkotans lie overwhelmingly toward the "ambush" end of the foraging spectrum whereas the remaining taxa (Autarchoglossa) span the entire range. The continuum emerges from the summation of these two patterns. As such, there are no iguanians or gekkotans at the high (active) end of the foraging spectrum and there are relatively few autarchoglossans at the low (ambush) end. Thus I would interpret Perry's data more generously to indicate a continuum of foraging modes that are not randomly distributed among clades, with ambush strategies probably primitive for squamates. Autarchoglossans apparently have evolved features that permit, but do not require, active foraging modes.
183
In any case, there remain exceptions to all of these generalizations, as noted. Tuatara are characterized as "passive" (ambush) foragers (Walls, 1981); however. Walls implied that individuals move around seeking appropriate prey. For example, because eggs and hatchlings of nesting shorebirds form an important part of the diet (see earlier discussion), tuatara must search actively through the birds' underground nest burrows. Among lizards, most iguanians and gekkotans are ambush foragers (Cooper, 1995b; Perry, 1999), as noted, but there are exceptions. The case of herbivores and the myrmecophagous genus Phrynosotna were discussed earlier. In addition, Pietruszka (1986) found that the carnivorous iguanid Gambelia exhibited rates of foraging movement equivalent to active foraging Cnemidophorus (Teiidae) and Perry (1999) identified several other actively foraging iguanid species ("phrynosomatids" or sceloporines). Although most autarchoglossan species are active foragers, there are numerous exceptions, particularly among cordylids (Cooper, 1994) and lacertids (Huey and Pianka, 1981; Nagy et al, 1984; Perry et al, 1990). Scincids are certainly unusual active foragers, if they are active foragers at all. The unusual case of snakes was also discussed earlier; other anguimorphans, such as Komodo dragons (Varanus komodoensis), often exhibit a similar pattern of searching widely for ambush sites (e.g., Auffenberg, 1981). Despite these exceptions and cautionary statements, some general remarks can be proffered. There is a strong phylogenetic component to foraging mode so that large taxa (e.g., families) tend to be relatively uniform in this regard (e.g., Huey and Pianka, 1981; Cooper, 1994, 1995a; Perry, 1999). Tuatara, and iguanian and gekkotan squamates, are primarily ambush foragers, but within this mode a range of variation exists so that some species approach an active foraging style. It is noteworthy that three of the exceptional iguanian taxa cited earlier are all dietary specialists: herbivores, anteating Phrynosotna, and lizard-eating Gambelia. Chemosensory prey discrimination also putatively evolves in conjunction with foraging mode (Cooper, 1994,1995a). These observations suggest, unsurprisingly, that diet and foraging mode are coevolved, integrated components of a lepidosaur's phenotype, although foraging mode is usually presented as a determinant of diet (e.g., Huey and Pianka, 1981). Autarchoglossans are primarily active foragers, but again, many exceptions are known. What remains most unclear is whether a given scleroglossan species characterized as an ambush forager would be comparable, both qualitatively and quantitatively, to an iguanian ambush forager. Perry's (1999) study is the first to approach this issue globally.
184
Kurt Schwenk C. Sensory Basis of Food Location and Identification
Lepidosaurs have at their disposal a full arsenal of sensory systems to employ in the detection, identification, and capture of food. These include, in most cases, well-developed systems of vision, audition, touch, and chemoreception. Some snakes have specialized systems for thermo (infrared) reception as well. Among these sensory modes, vision and chemoreception are denionstrably the most important for feeding and foraging in most lepidosaurs. There is a long-standing dogma in the herpetological literature suggesting that "ascalabotans" [Camp's (1923) paraphyletic group comprising Iguania and Gekkota] are primarily visual and that autarchoglossans are primarily chemosensory Schwenk (1993a,b, 1994a) reviewed this issue in the context of squamate phylogeny and the merits of cladistic vs paraphyletic classifications and showed that the dichotomy is unsupported, even if updated to contrast the strictly monophyletic taxa Iguania and Scleroglossa. Like the dichotomy in foraging mode, however, this simple view does have merit in broad outline. Certainly it is heuristically useful as a starting point from which one might elaborate the exceptions. 1. Vision Lepidosaurian eye structure and function have been comprehensively reviewed by Walls (1942), Underwood (1970), and Peterson (1992). Reperant et ah (1992) treated the central projections of the visual system. Lepidosaurs, with the exception of most snakes, are unique in possessing a retina with a single layer of photoreceptor cells. All other tetrapods have what is known as a "duplex" retina with two layers of sensory cells. The functional significance of this difference is unknown. In any case, most lizards have an otherwise typical tetrapod eye with the exception that the retina apparently comprises only cones and no true rods. However, the notion of "rods" and "cones" is based on mammalian models and it is not clearly applicable to most reptiles. Some nocturnal lepidosaurs have rodlike receptor cells that have been interpreted alternatively as true rods or as modified cones. In particular, some geckos and snakes have very rod-like receptors (Underwood, 1970; Peterson, 1990), but the homology of these photoreceptors to other tetrapod rods is unlikely and has not been critically evaluated. In general, a high density of cones is taken to indicate high visual acuity (sharpness of vision) and the presence of color vision (Walls, 1942). Although it is generally assumed that diurnal lizards possess color vision, this is not certain. The neurophysiology of color vision is complex and poorly un-
derstood (Goldsmith, 1990; Fleishman et a/., 1998). It depends minimally on the varying spectral sensitivities of different pigments within classes of photoreceptor cells. Pigmented oil droplets within the cones may contribute to color vision by acting as filters (Peterson, 1992). Consistent with this observation is that the cones of nocturnal lizards (presumably lacking the need for color vision) usually lack oil droplets or the droplets are unpigmented (Underwood, 1970). Limited experimental evidence suggests that oil droplets do contribute to color vision in an agamid lizard (Peterson, 1992). In any case, chromatic information might also be encoded in the pattern of depolarization and hyperpolarization of the receptors or their connecting interneurons (e.g., Solessio and Engbretson, 1993). Only recently have the techniques become available to measure directly the absorption spectra of individual photoreceptors (e.g., Fleishman et al., 1993; Solessio and Engbretson, 1993; EUingson et al, 1995). These studies show that at least some diurnal lizards possess the anatomical basis for wavelength discrimination, although direct proof that they perceive colors is still lacking. Behavioral evidence does indicate that most diurnal lizards employ color vision, but most of this evidence is anecdotal or without convincing controls. Importantly, Fleishman et al. (1998) pointed out that colors vary not only in frequency, but in brightness (i.e., perceived intensity) contrast with the background and such differences confound the assumption of color discrimination in most experimental designs. Nonetheless, the evidence is suggestive. Numerous studies demonstrate, for example, that conspecific color cues are used by lizards in social contexts to assess sex, age, and breeding status (reviewed by Cooper and Greenberg, 1992; see also Watkins, 1997). However, the use of color cues in feeding is less well supported. Several studies show preferences of some lizards for foods of certain colors, especially yellow, orange, and red, which they differentiate from other colors (e.g., Benes, 1969; Rand et al, 1975; McGovern et al, 1984; Cans et al, 1985; personal observation). Two iguanian taxa, Anolis (Iguanidae) and Chamaeleonidae, show extreme adaptations for vision and visually directed predation that are worth exploring as exemplars. Anoles are extraordinary in their possession of a second fovea in the retina (Underwood, 1970; Fleishman, 1992). They are the only vertebrates other than some birds to have dual fovea (Underwood, 1970). The fovea is a specialized zone of the retina containing unusually high photoreceptor (usually cone) density (e.g., Makaretz and Levine, 1980; Fite and Lister, 1981) and other attributes that enhance visual acuity When an object of interest is spotted, the visual axis is usually
8. Feeding in Lepidosaurs adjusted so that it falls on the fovea. The typical, central fovea of anoles provides a zone of acuity for wide-field, monocular vision, whereas the second (novel) temporal fovea comes into play when the eyes fixate binocularly on a prey item directly in front of the lizard (Fite and Lister, 1981). Use of the temporal fovea in anoles is very similar to that in raptors (Sillman, 1973; Fite and Lister, 1981). Other nonavian tetrapods with welldeveloped binocular vision, including some squamates (e.g., the Asian vine snake, Ahaetulla) and humans, have a single fovea in the temporal position (Walls, 1942). The peripheral retina of anoles is extremely sensitive to certain kinds of movement so that even minute prey movements elicit monocular fixation with the fovea, potentially followed by binocular fixation and a predation attempt (Fleishman, 1986, 1992). A prey item that lies within a threshold distance will be approached and attacked, and this distance is modulated depending on prey size and abundance (Shafir and Roughgarden, 1998). The visibility of a prey item to an anole (and presumable other lepidosaurs) depends largely on the extent to which its movement differs from random background motion (Fleishman, 1986, 1988,1992; Persons et al, 1999). Anoles have also been shown to have visual sensitivity to ultraviolet light, which is used by some species in the detection of conspecific dewlap displays (Fleishman et ah, 1993). It is unknown whether ultraviolet cues play a role in predatory behavior, however. Chamaeleonids also exhibit unusual eyes and specialized, visually mediated predatory behavior. Upper and lower eyelids are fused and the eyeball is positioned extraorbitally so that the eyes are "turreted,'' each capable of independent movement and 360° of rotation. Although chameleons are certainly extreme in this regard, it is seldom appreciated that nearly all lizards, as well as other nonmammalian tetrapods, are capable of independent eye movement (Walls, 1942, 1961; Kirmse, 1988). Furthermore, some lizard species, particularly anoline iguanids and some agamines, have partial turreting of the eyes (e.g., Abel, 1952; personal observation). In Chamaeleolis, for example, an anoline form convergent with chameleons in many respects, upper and lower eyelids are fused and turreting is pronounced so that independent eye movement is clearly evident (personal observation). Chameleons scan their environments using saccadic (rapid, stop-go) movements of each eye independently. Only one eye at a time is in focus, however, alternating at approximately 1-sec intervals (Ott et al, 1998). Thus, a wide area of the environment can be scanned monocularly for prey, but apparently only one eye's image is processed at a time. When a prey item is spotted, the
185
head is turned and the eyes rotate forward to fixate it binocularly (Ott et al., 1998). Prey movement is then tracked with head, rather than eye movement (Flanders, 1985). In preparation for prey capture, the distance to the prey item is visually determined. Chameleons depend entirely on accommodation (focusing) cues to do this and thus differ from most mammals, and probably other tetrapods, which principally employ binocularity (stereopsis) for depth perception (Harkness, 1977; CoUett and Harkness, 1982; Ott and Schaeffel, 1995). The speed of accommodation is among the fastest of any vertebrate measured (Ott and Schaeffel, 1995; Ott et al, 1998). This nearly instantaneous and highly accurate assessment of prey distance is used to modulate precisely tongue projection distance for lingual prey capture (Harkness, 1977; Ott et ah, 1998; see Section VI, A). Another unique feature of the chameleon eye is that the lens is negatively powered; in conjunction with corneal refraction, the visual system is telephoto, i.e., it magnifies the image on the retina (Ott and Schaeffel, 1995; Land, 1995). Land (1995) suggested that the peculiar lens arrangement creates monocular disparity so that objects at different distances in the visual field move across the retina at different speeds as the eye is rotated. This might help chameleons to distinguish prey at varying depths within the complexly threedimensional, arboreal environment they inhabit. It is noteworthy that both anoles and chameleons are arboreal species with extremely reduced vomeronasal chemosensory function (Haas, 1937, 1947; Armstrong et al, 1953; Gabe and Saint Girons, 1976). Although arboreality has often been associated with visual predation and reduced chemoreception, other arboreal species, such as green iguanas and most geckos, are, in fact, highly chemosensory, thus there is no necessary, causal relationship between the two traits (see Schwenk, 1993a). Nonetheless, some shared component of anoline and chamaeleonid environments, or similarities in predatory behavior, might have driven the acquisition of analogous adaptations. Alternatively, ancestral reduction of chemoreception, for whatever reason, might strongly predispose any lepidosaurian lineage toward compensatory enhancement of its visual system, no matter what the environment. 2.
Chemoreception
Most squamates rely heavily on chemosensory, as well as visual, cues during foraging and feeding. There are at least three separate chemosensory systems available to most species: gustation, nasal olfaction, and the vomeronasal system (for reviews see Burghardt, 1970,1980; Halpern, 1992; Schwenk, 1995b; Font, 1996).
186
Kurt Schwenk
A possible fourth, trigeminal chemosensory system remains mostly unexplored in lepidosaurs (Tucker, 1971). Gustation is mediated by taste buds, which, in reptiles, are restricted to the oral cavity and pharynx, and are innervated by cranial nerve IX, the glossopharyngeal (e.g., Willard, 1915). Tuatara are well supplied with lingual taste buds and are unique among lepidosaurs in having specialized "gustatory papillae" (Schwenk, 1986). Although snakes lack lingual taste buds, many species have them within the oral mucosa, particularly along the tongue sheath (Schwenk, 1985; Young et ah, manuscript in preparation). Taste buds reach extremely high densities on the foretongues of some lizards, particularly iguanians (Schwenk, 1985; Herrel et ah, 1998c). They are stimulated by chemically diverse substances introduced to the surface of the tongue or into the mouth (see Schwenk, 1985), but the significance of taste in the biology of lepidosaurs remains almost wholly unexplored (Burghardt, 1970; Schwenk, 1985; see later). Olfaction is mediated by the nasal epithelium covering the nasal conchae and other surfaces within the nasal cavities. Volatilized chemicals inspired through the external nares stimulate the olfactory sensory cells, which project directly to the olfactory bulb of the brain. Air moves through the nasal cavities and exits posteroventrally through the palate via the internal nares (choanae) and into the glottis to the lungs. Inspiration is usually driven by typical rib-based ventilation of the lungs, but in some lizards a buccal or gular pumping mechanism drives airflow through the nasal cavities with or without lung ventilation (e.g., Schwenk, 1993a, personal observation; Dial and Schwenk, 1996; Owerkowicz et al, 1999). Hence, some lizards may employ an olfactory behavior analogous to mammalian "sniffing" (Dial and Schwenk, 1996). The vomeronasal system is served by paired vomeronasal (or Jacobson's) organs (VNO) that lie above the anterior palate, encapsulated by bone and cartilage within the nasal cavities. Each VNO has a sensory epithelium, the neurons of which project to the accessory olfactory bulb (at the base of the olfactory bulb in squamates). In Sphenodon, the VNO are tubular chambers open to the nasal cavities, but not directly to the mouth (Hoppe, 1934; Parsons, 1970). VNO stimulation is presumably by means of nasal inhalation, but virtually nothing is known of vomeronasal function in tuatara. In squamates, the VNO are sequestered from the nasal cavity and open directly to the mouth via two tiny openings in the palate, the vomeronasal fenestrae (Parsons, 1970). They are stimulated by chemicals brought into the oral cavity by the tongue and then drawn through the vomeronasal fenestrae (Halpern, 1992). Environmental chemicals are gathered by the tongue
during tongue flicking, a behavior characteristic of virtually all squamates. Tongue flicks vary from simple downward protrusion toward the substrate to rapid, multiple oscillations of the tongue in the air (e.g., Gove, 1970; Bels et al, 1994). In general it is presumed that VNO function is distinguished from nasal olfaction by its sensitivity to relatively nonvolatile, heavy molecular weight molecules (e.g.. Tucker, 1971); however, aerial tongue flicks can also saniple volatiles for VNO scrutiny (e.g., Halpern et al., 1997), thus confounding the biological roles of the two senses (Schwenk, 1995b). Similarly, the presence of taste buds on the tongue tip in most squamates makes it possible that tongue flicks directed to the substrate might be gustatory rather than serving to collect nonvolatile chemicals for the VNO, particularly in iguanians, which possess high taste bud densities on the tongue tip (Schwenk, 1985; Herrel et al, 1998c). However, strong circumstantial evidence supports the supposition that all tongue flicks, including substrate touches, mediate vomeronasal stimulation rather than gustation (Schwenk, 1993a; Dial and Schwenk, 1996). Gustation is most likely to be relevant for the assessment of palatability after a food item is held within the mouth (Schwenk, 1986,1995b). The manner in which the chemical senses are functionally and behaviorally interrelated remains poorly understood. Schwenk (1995b) proposed a heuristic model of chemically mediated predation in squamates consistent with current data: (1) Nasal olfaction initially senses a volatile food odor, which triggers tongueflicking behavior (Cowles and Phelan, 1958; Halpern et al, 1997). (2) Tongue flicking stimulates the vomeronasal system, which provides additional information about the odor or samples a different volatile component of the odor. (3) Exploratory and trailing behavior is initiated to localize the odor source using both aerial and substrate-directed tongue flicks; species with forked tongues can employ tropotactic mechanisms, but species with unforked tongues must rely on klinotactic mechanisms (see Schwenk, 1994e). (4) Tongue touches to a prey trail or the food item itself sample nonvolatile chemical components of the odor source for additional information. (5) Once the source is located, the chemical information gleaned so far may be sufficient to trigger ingestion, but often only in conjunction with appropriate visual cues (see later). (6) Following ingestion of the food item, taste buds provide gustatory assessment of palatability. Distasteful food is rejected at this point and a combination of visual and chemosensory cues associated with the prey item mediate learned avoidance in future encounters. This scenario clearly does not apply universally to all squamates, nor to any individual squamate during every feeding event. Some taxa have reduced or enhanced particular chemosensory modes, or empha-
8. Feeding in Lepidosaurs size nonchemosensory cues. For example, geckos and some other groups may substitute nasal olfaction for vomeronasal function at any point in this sequence (Schwenk, 1993a; Dial and Schwenk, 1996). Chameleons and anolines have reduced both nasal olfactory and vomeronasal senses and are heavily dependent on vision during all stages of feeding (see earlier discussion). Locomotion and exploratory behavior are typically attended by tongue flicking in most squamates, hence initial olfactory stimulation of the vomeronasal system may be unnecessary. Ambush predators have no need to trail and localize food sources, but rather assess the quality of prey items as they approach, both visually and by tongue flicking. It is well documented that many squamates, particularly scleroglossans, use chemosensory cues during foraging, primarily to locate hidden or cryptic prey. Burghardt (1970) and Halpern (1992) have reviewed much of this evidence. Bogert and Martin del Campo (1956), for example, found that Gila monsters (Varanoidea, Heloderma) could follow the scent trail of an egg dragged across the substrate and find where it was buried. In general, active foragers search for cryptic prey in leaf litter, under covering objects, or buried within the substrate by constantly probing with the snout and tongue flicking (e.g.. Fitch, 1935,1954,1958; Taylor, 1986; Vitt and Cooper, 1986; Vitt, 1991; Vitt and Blackburn, 1991). Most snakes are well known to follow scent trails left by prey species using their forked tongues and vomeronasal systems (Schwenk, 1994e). In addition to use of chemosensory cues to locate food during foraging in squamates, vomeronasal chemoreception is implicated in food/prey identification and discrimination (summarized by Cooper, 1994, 1995a). Cooper showed that herbivorous and actively foraging lizards are able to use chemical cues gathered by tongue flicking to distinguish among odors and to discriminate food from nonfood odors. These tests were based on the number of tongue flicks and attacks directed to cotton swabs saturated with test odors and hence did not involve food-specific visual cues. Although there are problems with this experimental method (Halpern, 1992; Dial and Schwenk, 1996), the results are suggestive. The ability to discriminate food odors is apparently a labile trait that is tightly associated with foraging mode; Cooper has suggested that phylogenetic transformations in foraging mode are always accompanied by a gain or loss of discriminatory ability. 3. Sensory
Integration
Although, as noted, some squamates apparently emphasize one sensory mode or another, it is clear that in nearly all species feeding behavior is a multisensory
187
undertaking. The interplay of chemical senses may occur as outlined earlier and, more importantly, these systems are integrated with vision and other sensory cues to complete a foraging and feeding sequence. For example, although prey location and identification are almost entirely visual in the extremist anolines (Curio and Mobius, 1978, and references earlier), even they are well endowed with lingual and oral taste buds to assess food palatability once ingested (Willard, 1915; Schwenk, 1985). Conversely, actively foraging taxa usually assumed to be chemosensory specialists almost certainly use visual cues at some point in a feeding sequence. Indeed, in most cases it is a visual cue, usually movement, that triggers an attack (e.g.. Fitch, 1954,1958; Burghardt, 1964; Rand et al, 1975; Boyden, 1976; Curio and Mobius, 1978; Dial, 1978; Cooper, 1981; McGovern et al, 1984; Nicoletto, 1985a,b; Vitt and Cooper, 1986; Kaufman et al, 1996; personal observation). Kaufman et al (1996) showed that visual cues were sufficient for distinguishing among prey types in highly chemosensory monitor lizards (Varanus). Among snakes, acknowledged vomeronasal specialists, some species rely on visual cues for prey identification and elicitation of a strike (e.g., Drummond, 1979). Lizards clearly modulate their attack based on visual cues (Schwenk and Throckmorton, 1989; unpublished observation). During lingual feeding attempts, tongue protrusion distance is adjusted according to prey movement or behavior, or sometimes an attack is aborted altogether if prey movement ceases. Mathias Ott (personal communication) has found that horned lizards {Phrynosoma) adjust the trajectory of the rapidly protruding tongue during a capture attempt by slightly turning the head; the behavior is so rapid that visual feedback must occur within a few milliseconds (Schwenk and Throckmorton, 1989; Schwenk, unpublished; M. Ott, personal communication). Boyden's (1976) study of butterfly predation in several populations of Ameiva exemplifies the sensory interplay undoubtedly typical of most squamates. Ameiva are highly chemosensory teiid lizards with deeply forked tongues; virtually all teiids studied are active foragers (e.g.. Fitch, 1958; Anderson and Karasov, 1981; McGovern et al, 1984; Pietruszka, 1986; Vitt, 1991; Vitt and de la Torre, 1996; Vitt et al, 1997a,c) and are capable of chemosensory trailing (Schwenk, 1994e). Nonetheless, Boyden (1976) showed that visual cues alone were used by experienced individuals to distinguish palatable from unpalatable species of butterfly. Naive lizards, however, readily attacked unpalatable species, but subsequently avoided them. Thus, initial chemosensory assessment of palatability led to subsequent visual discrimination of prey species and learned avoidance of distasteful types. Fossorial squamates usually have reduced vision
188
Kurt Schwenk
and, in some cases, the eye is lost or entirely nonfunctional as a photoreceptor (Underwood, 1970). In such species, vision obviously can play little or no role during feeding behavior and chemoreception is well developed in compensation. It is suggestive that in many fossorial species nasal olfaction seems to be relatively enhanced, as it is in the nocturnal geckos (Schwenk, 1993a); however, the precise interplay of the three chemical senses remains largely unexplored in these, as in other taxa. 4. Sensory Ecology of Sphenodon Most of what is known about the sensory ecology of feeding in tuatara is anecdotal. Tuatara are nocturnal foragers and their retinas are dense with rod-like photoreceptors (Underwood, 1970). As noted earlier, vomeronasal organs are present, but not exceptionally well developed and they do not open to the oral cavity. Not surprisingly, tuatara do not tongue flick (Walls, 1981; Gillingham et al, 1995). Their nasal, olfactory system also appears to be only moderately developed (Gabe and Saint Girons, 1976). However, they are very well supplied with lingual taste buds (Schwenk, 1986). Gillingham et al (1995) showed convincingly that visual cues alone are both necessary and sufficient during all tuatara social behaviors. Visual cues, especially movement, are also sufficient to elicit predatory attack; even inanimate objects moved within sight are attacked and bitten (Farlow, 1976; Walls, 1981). Walls (1981:92) suggested that "rounded, dark, smooth, shiny objects" were most effective in provoking attack in captive feeding experiments—a result consistent with the finding that rounded, dark, smooth, shiny beetles were the most frequent prey items in the study population! In captivity, naive tuatara were observed to capture certain, apparently noxious, prey and then reject them. Diet analysis showed that wild individuals avoided such readily available prey. It is therefore likely that gustation is used by tuatara to assess the palatability of food items and distasteful prey are rejected, subsequently discriminated visually, and avoided (Schwenk, 1986). Meyer-Rochow (1988) tested young tuatara in a darkroom. In the absence of light, or with a photographic safety light, the tuatara showed no interest in mealworms placed 2 cm in front of them, even though the mealworms were active and making scratching sounds. As soon as light was present, however, the same tuatara readily ingested the mealworms. MeyerRochow (1988) concluded that olfactory and auditory cues were insufficient to elicit feeding and that visual cues were relied upon exclusively. These observations are consistent with those of Gorniak et al. (1982) who
found that tuatara only responded to moving prey. Nonetheless, Walls (1981) noted that tuatara sometimes investigate and ingest nonmoving food items, especially petrel eggs and carrion, and believed that this behavior indicates the occasional use of olfactory cues. Farlow noted that a dead cockroach placed in a tuatara's cage was eventually eaten. Similarly, Gorniak et al. (1982:351) contended that "olfactory cues mediate the prey attack response and visual cues are used to steer the capture of prey"; however, they presented no evidence in support of this statement. In conclusion, visual cues are predominant in identifying most prey types and eliciting prey attack, but olfaction may be important for locating and identifying immobile food items. Gustation probably mediates learned food preferences. However, controlled experiments are notably lacking and these are required to determine the interplay of the senses in tuatara foraging and feeding. 5. Other Sensory
Systems
Other sensory mechanisms are less commonly employed during feeding in some squamates. In particular, tactile, vibratory cues are probably important to a variety of fossorial species, such as amphisbaenians (e.g., Gans, 1960,1969a). Gans noted that amphisbaenians moved toward insects moving at the surface. The importance of vibratory cues for locating prey was also demonstrated for the sand-swimming skink, Scincus (Hetherington, 1989). Snakes may use ground-borne vibratory cues to sense the presence of prey (Hartline, 1971). Barnett et al. (1999) recently showed that some chameleons use twig-borne vibrations as social signals. If chameleons are sensitive to such vibratory stimuli, conceivably they could use them for prey detection as well. I have virtually ignored the possible role of hearing in prey detection, but this is only because there is no information indicating it to be of widespread significance. Nonetheless, it is likely that audible cues associated with prey movement contribute to the triggering of prey attack. One species of Hemidactylus (Gekkonidae) has been shown to orient to and approach recorded cricket (prey) calls (Sakaluk and Belwood, 1984), suggesting that phonotaxis might be an important part of foraging in this species, at least. The unusual use of vocal communication by many gekkonids (Marcellini, 1977) might preadapt them for use of audition in prey detection; however, it is unknown how prevalent the behavior is in the family. The use of vibratory cues by amphisbaenians might actually be a form of low-frequency hearing rather than a strictly tactile sense. Most amphisbaenians have a unique ear morphology in which a tympanum is lacking, but an extremely long extracolumella extends anteriorly from
8. Feeding in Lepidosaurs the stapes along the mandible within a specialized zone of skin (Wever, 1978). Wever (1978:784) concluded that "the extracolumella functions in the conduction of vibrations set up in the skin of the face through the action of aerial sounds." Nonetheless, the position of the receptive area on the lower jaw and the sensitivity of the inner ear to low frequencies suggest that the amphisbaenian auditory system might actually be specialized for the reception of ground-borne, rather than aerial, vibrations.
IV. MORPHOLOGY OF THE FEEDING APPARATUS A. Skull and Mandible As reviewed in Chapter 2, the head skeleton comprises the skull, proper, as well as the mandible (lower jaw) and hyobranchial apparatus (tongue and throat skeleton). This section considers basic features of the head skeleton in tuatara and nonophidian squamates. The snake skull is treated in Chapter 9. Tables 8.1 and 8.2 list selected references for the morphology of the head skeleton in lepidosaurs. In general, lepidosaurian skulls comprise a system of bony plates and struts perforated by frequent, large vacuities. Particularly in some squamates, the skull becomes a gracile, almost delicate network of slender struts and kinetic joints. This trend reaches its maximal expression in macrostomatan snakes, but it is approached to some degree in several lizard groups. A lepidosaurian trend toward reduced cranial ossification, extensive fenestration, and cranial kinetism is in stark contrast to basal archosaurs, turtles, and many fossil reptile groups. Some of these changes relate directly to feeding system adaptation. One's impression of a typical lepidosaurian skull is of an extremely strong structure that minimizes weight and material. 1. Sphenodon Tuatara retain what is generally regarded as a primitive skull pattern (Glinther, 1867; Romer, 1956). This is most evident in the retention of a true diapsid pattern of temporal fenestration (see Chapter 2). As such, there are two, large temporal openings in the skull with a complete lower temporal arch formed by the jugal and quadratojugal bones (Fig. 8.2). The posterior rim of the lower fenestra comprises fused squamosal, quadratojugal, and quadrate bones, with the quadrate projecting ventrally below the level of the arch to articulate with the mandible (the jaw joint). In dorsal view, the temporal region is very wide due to capacious upper temporal fenestrae, but the dermal
189
roof of the braincase is narrowly constricted. Sphenodon also retains the primitive condition in having the entire anteroposterior series of medial, dermal roofing bones paired (premaxillae, nasals, frontals, parietals). A small parietal foramen is present in the midline for the parietal eye. The orbits are large. An unusual feature of the skull is the presence of a prominent, lateral notch between the maxilla and the premaxilla. The premaxillae are slightly elongate and in adults bear two, often prominent, acrodont teeth each. The effect is to create the impression of a "beak," hence the taxon name Rhynchocephalia, meaning "beak head." The posterior end of the braincase is wholly ossified and formed from an occipital series of bones surrounding the foramen magnum. Large paroccipital processes of the opisthotic bones extend laterally to articulate with the posterior rim of the lower temporal arch (at the dorsal end of the quadrate). Between these processes and the upper temporal arcades are posterior openings called the posttemporal fenestrae. Dorsally, a supraoccipital bone forms a median crest that extends toward the roofing parietal bones. It gives rise to an anteriorly directed, median process (ascending process or processus ascendens) that is "clamped" between the paired parietals. Ventrally, the basisphenoid forms the floor of the braincase. It extends anteriorly on each side to form two basipterygoid processes, which articulate with the deep, posterior wings of the pterygoid bones. Although Romer (1956) described these joints as "movable," each condylar process is held within a bony socket of the pterygoid that would seem to preclude movement here. Anteriorly, the braincase remains membranous throughout life (de Beer, 1937; Romer, 1956). Paired vomers, palatines, and pterygoids form most of the broad, bony palate that is much less fenestrated than in squamates. A derived feature of the palate is the presence of palatal teeth arrayed linearly along the lateral margin of the palatine bone parallel to the marginal, maxillary tooth row (Fig. 8.2B). During jaw closure, mandibular teeth lie within the cleft between the two upper tooth rows. As noted, the premaxillae and maxillae bear the marginal teeth (see later). Palatal fenestra are little more than longitudinal slits between palatine and maxillary bones due to the close juxtaposition of their tooth rows. The lower jaw, or mandible, comprises two halves (hemimandibles or mandibular rami) that are joined anteriorly by a loose connective tissue joint (Robinson, 1976). The looseness of this articulation is surprising given that Meckel's cartilage completely fuses anteriorly in the embryo, implying the formation of a true, symphyseal (fibrocartilage) joint (de Beer, 1937). However, Bellairs and Kamal (1981) observed late-stage
190
Kurt Schwenk TABLE 8.1
Selected References Treating the M o r p h o l o g y of the Feeding S y s t e m in Lepidosaurs Exclusive of Snakes and Chameleons"" Scleroglossa
Sphenodon
Iguania
Bellairs and Boyd (1950) Bellairs and Kamal (1981) Broom (1903,1925) de Beer (1937) Gorniak et al (1982) Giinther (1867) Howes and Swinnerton (1901) Lakjer (1927) Malan (1946) Pratt (1948) Rieppel (1993) Romer (1956) Save-Soderbergh (1947) Schauinsland (1900,1903) Verluys (1936) Werner (1962) Wettstein (1931)
Avery and Tanner (1964, 1971) Bellairs (1984) Bellairs and Boyd (1950) Bellairs and Kamal (1981) Blanc (1965) Broom (1903,1925,1935) Camp (1923) de Queiroz (1987) Engelbrecht (1951) George (1954) Greer (1989) Hallermann (1992,1994) Harris (1963) Herrel et al (1998a) lordansky (1990b, 1996) Jenkins and Tanner (1968) Kamal and Zada (1970) Kraklau (1991) Lakjer (1927) Lang (1989) Malan (1946) Oelrich (1954) Paranjape (1974) Parker(1881) Pratt (1948) Ramaswami (1946) Rieppel (1978d, 1993) Romer (1956) Stimie (1966) Throckmorton (1976) Verluys (1936) Zaluskyeffl/. (1980)
Gekkota
Scincomorpha
Anguimorpha
Bellairs (1984) Bellairs and Boyd (1950) Bellairs and Cans (1982) Bellairs and Kamal (1981) Bels et al (1993) Broom (1903,1925,1935) Camp (1923) Dalrymple (1979) de Beer (1937) El Toubi and Kamal (1959a,b) Fisher and Tanner (1970) Cans (1960,1974,1978) Greer (1989) Herrel et al (1998a) lordansky (1990b, 1996) Kamal (1965) Kesteven (1957) Kritzinger (1946) Lakjer (1927) MacLean (1974) Malan (1940,1946) Montero et al (1999) Nash and Tanner (1970) Parker(1880) Presch (1976) Rao and Ramaswami (1952) Rice (1920) Rieppel (1978d, 1981c, 1984a,c, 1992b, 1993) Romer (1956) van Pletzen (1946) Verluys (1936) Wineski and Cans (1984) Zangerl (1944)
Auffenberg (1981) Bahl (1937) Barrows and Smith (1947) Bellairs (1949,1950,1984) Bellairs and Boyd (1950) Bellairs and Kamal (1981) Bogert and Martin del Campo (1956) Broom (1925,1935) Camp (1923) Criley (1968) Frazzetta (1962) Greer (1989) lordansky (1990b, 1996) Lakjer (1927) Malan (1946) McDowell and Bogert (1954) Rieppel (1978b,d, 1983, 1984a, 1993) Rieppel and Labhardt (1979) Romer (1956) Smith (1980) Verluys (1936)
Camp (1923) Dalrymple (1979) Dessem (1985) Edmund (1969) Estes and Williams (1984) Fisher and Tanner (1970) Cans (1960,1978) Greer (1989) MacLean (1974) Mateo and Lopez-Jurado (1992,1997) Peyer(1929) Presch (1974) Rocek (1980a) Romer (1956)
Bogert and Martin del Campo (1956) Camp (1923) Cooper (1966) Edmund (1969) Estes and Williams (1984) Greer (1989) Lonnberg (1903) McDowell and Bogert (1954) Peyer(1929) Rieppel (1978f, 1979a) Rieppel and Labhardt (1979) Rocek (1980b) Romer (1956)
Skull Bauer (1990) Bellairs and Boyd (1950) Bellairs and Kamal (1981) Broom (1935) Camp (1923) Greer (1985,1989) lordansky (1990b, 1996) Kluge (1962,1987) Lakjer (1927) Malan (1946) Rieppel (1984a,b, 1992a, 1993) Romer (1956) Stephenson (1962) Underwood (1957) Verluys (1936) Webb (1951) Wellborn (1933)
Dentition Edmund (1969) Gorniak et al (1982) Harrison (1901a,b) Robinson (1976) Romer (1956)
Avery and Tanner (1964) Blanc (1965) Camp (1923) Cooper and Poole (1973) Cooper et al (1970) Edmund (1969) Estes and Williams (1984) Greer (1989) Hotton (1955) Kline (1983) Kline and Cullum (1984, 1985) Lang (1989) Montanucci (1968) Oelrich (1956) Robinson (1976) Robison and Tanner (1962) Romer (1956) Throckmorton (1976,1979)
Bauer (1990) Bauer and Russell (1990) Camp (1923) Edmund (1969) Estes and Williams (1984) Greer (1985,1989) Patchell and Shine (1986c) Rieppel (1984b) Romer (1956) Sumida and Murphy (1987)
(continues)
191
8. Feeding in Lepidosaurs TABLE 8.1 (continued) Scleroglossa Sphenodon
Iguania
Gekkota Hyobranchial
Scincomorpha
Anguimorpha
apparatus
Edgeworth (1935) Furbringer (1922) Osawa (1898) Romer (1956) Tanner and Avery (1982)
Avery and Tanner (1964, 1971) Beebe (1944) Beddard (1905) Bels (1990b, 1992) Bels et al (1994) Blanc (1965) Camp (1923) Capel-Williams and Pratten (1978) de la Cerna de Esteban (1965) de Queiroz (1987) Edgeworth (1935) Font and Rome (1990) Furbringer (1922) Gnanamuthu (1937) von Geldern (1919) Harris (1963) Jenkins and Tanner (1968) Kraklau (1991) Lang (1989) Oelrich (1954) Richter (1932) Robison and Tanner (1962) Romer (1956) Smith (1984,1988) Tanner and Avery (1982) Throckmorton et al. (1985)
Bauer (1990) Bels et al. (1994) Camp (1923) de la Cerna de Esteban (1965) Furbringer (1922) Gnanamuthu (1937) Greer (1985) Richter (1932) Rieppel (1981a) Romer (1956) Tanner and Avery (1982) Underwood (1957)
Edgeworth (1931,1935) Gorniakeffl/. (1982) Haas (1973) Kesteven (1944) Lakjer (1926) LightoUer (1939) Lubosch (1933) Ostrom (1962) Poglayen-Neuwall (1953) Wettstein (1931)
Avery and Tanner (1964, 1971) Bradley (1903) Capel-Williams and Pratten (1978) Edgeworth (1935) George (1948) Haas (1973) Herrel et al. (1998a, 1999a) lordansky (1970,1990b) Jenkins and Tanner (1968) Kesteven (1944) Kraklau (1991) Lakjer (1926) Lang (1989) Lubosch (1933) Oelrich (1954) Poglayen-Neuwall (1954) Rieppel and Gronowski (1981) Robison and Tanner (1962) Throckmorton (1976,1978)
Bradley (1903) Edgeworth (1935) Gasc (1968) Haas (1973) lordansky (1970,1990b) Kesteven (1944) Lakjer(1926) Lubosch (1933) Poglayen-Neuwall (1954) Rieppel (1984a, b) Rieppel and Gronowski (1981)
Beebe (1945) Bels et al. (1994) Camp (1923) de la Cerna de Esteban (1959,1965) Edgeworth (1935) Cans (1978) Fisher and Tanner (1970) Furbringer (1922) Gnanamuthu (1937) Harris (1985) MacLean (1974) Montero et al. (1999) Nash and Tanner (1970) Richter (1932) Rieppel (1981a) Romer (1956) Smith (1984) Tanner and Avery (1982)
Bels et al. (1994,1995) Camp (1923) Furbringer (1922) Gnanamuthu (1937) McDowell and Bogert (1954) Richter (1932) Rieppel (1981a) Romer (1956) Smith (1986) Sondhi (1958b) Tanner and Avery (1982)
Bradley (1903) Edgeworth (1935) Fisher and Tanner (1970) Cans et al. (1985) Haas (1973) Herrel effl/. (1996,1998a, 1999b) lordansky (1970,1990b) Kesteven (1944,1957) Lakjer(1926) LightoUer (1939) Nash and Tanner (1970) Poglayen-Neuwall (1954) Rieppel (1979b, 1980, 1981c, 1984a) Rieppel and Gronowski (1981) Throckmorton (1982) Wineski and Cans (1984)
Bradley (1903) Edgeworth (1935) Haas (1973) lordansky (1970,1990b) Kesteven (1944) Lakjer (1926) LightoUer (1939) Poglayen-Neuwall (1954) Rieppel (1978a, 1984a) Rieppel and Gronowski (1981) Smith (1980,1982,1986)
Jaw musculature
(continues)
192
Kurt Schwenk
TABLE 8.1 (continued) Scleroglossa Sphenodon
Iguania
Gekkota
Scincomorpha
Anguimorpha
Bayer(1899) Bels et al (1993) Bendz (1840) Bogert (1964) Camp (1923) Chieleffl/. (1992) Corning (1895) de la Cerna de Esteban (1959,1965) Dornesco and Andrei (1966) Duges (1827) Edgeworth (1935) Ferdinand (1884) Gabe and Saint Girons (1969) Cans (1978) Gnanamuthu (1937) Harris (1985) Iwasaki and Kobayashi (1992) Iwasaki and Miyata (1985) Kallius (1901) Kochva (1978) MacLean (1974) Minot (1880) Presch (1971) Renous (1977) Richter (1932) Schwenk (1985,1988,1994, 1995; 2000a) Sewertzoff (1929) Smith (1984) Taib and Jarrar (1986) Tanner and Avery (1982) von Seiller (1891,1892) Wagner and Schwenk (2000) Zanno (1974)
Camp (1923) Ferdinand (1884) Gabe and Saint Girons (1969) Gnanamuthu (1937) Kallius (1901) Kochva (1978) McDowell (1972) McDowell and Bogert (1954) Richter (1932) Schwenk (1985,1988, 1994,1995, 2000a) Sewertzoff (1929) Smith (1986) Smith and Mackay (1990) Sondhi (1958a,b) Tanner and Avery (1982) Toubeau et al (1994) von Seiller (1891) Wagner and Schwenk (2000)
Tongue Edgeworth (1935) Gabe and Saint Girons (1964,1969) Kochva (1978) Osawa (1897) Rieppel (1978e) Schwenk (1986,1988) Sewertzoff (1929) Tanner and Avery (1982)
Bendz (1840) Camp (1923) de la Cerna de Esteban (1965) Delheusydfl/. (1994) Edgeworth (1935) Ferdinand (1884) Gabe and Saint Girons (1969) Gandolfi, H. (1908) Gnanamuthu (1937) Herrel et al (1998c) Kochva (1978) Kraklau (1991) McDowell (1972) Oelrich (1954) Rabinowitz and Tandler (1986,1991) Richter (1932) Schwenk (1985,1988, 1994,1995, 2000a) Sewertzoff (1929) Smith (1984,1988) Taib and Jarrar (1985a,b) Tanner and Avery (1982) Wagner and Schwenk (2000) Willard (1915)
Camp (1923) de la Cerna de Esteban (1965) Edgeworth (1935) Gabe and Saint Girons (1969) Gnanamuthu (1937) Greer (1985) Iwasaki (1990) Kochva (1978) Ping (1932) Richter (1932) Schwenk (1985,1988, 1994,1995,2000a) Sewertzoff (1929) Tanner and Avery (1982) Underwood (1957) Wagner and Schwenk (2000) Zavattari (1909)
^See Chapter 9 and Table 8.3, respectively. For convenience, references for Dibamidae and Xantusiidae are included under Gekkota (as per Lee, 1998), and those for Amphisbaenia under Scincomorpha (as per Schwenk, 1988); otherwise the taxonomic content of each column follows Estes et al (1988). The primary literature is heavily emphasized, but important reviews or secondary sources are included. Major surveys, summaries, or overviews are shown in bold type.
regression of Meckel's cartilage in Lacerta, suggesting the likelihood of late-stage remodeling of the mandibular joint. Robinson (1976) suggested that the looseness of this joint permits rotation of each hemimandible and deformation of mandibular shape that allows propalineal movement of the lower tooth rows between the dual, parallel tooth rows of the upper jaw (see Section V). The only dentigerous element in each hemimandible is the dentary bone, which is also the largest and most anterior element. A series of smaller bones
form its posterior end. Notable among these is a dorsally projecting coronoid bone that serves as the attachment site for the adductor mandibulae musculature and an articular bone (fused with the prearticular) that articulates with the quadrate of the upper jaw to form the jaw joint. The articular facet is shallow and somewhat elongated anteroposteriorly, thus permitting anteroposterior sliding of the mandible on the quadrates (Robinson, 1976). A very small retroarticular process of the articular extends posterior to the jaw joint. On the
193
8. F e e d i n g in L e p i d o s a u r s TABLE 8.2 Selected References Treating Form and Function of the Feeding Apparatus in C h a m e l e o n s Morphology
Functional inference
Functional analysis
Bell (1987,1989) Briicke (1852) Edgeworth (1935) Duvernoy (1836b) Fineman (1940,1943) Fiirbringer (1922) Cans (1967) Gnanamuthu (1930,1937) Houston (1828) lordansky (1973) Kathariner (1895) Lubosch (1932,1933) Minot (1880) Mivart (1870) Owen (1866) Perrault (1676) Rice (1973) Rieppel (1981b, 1987) Rieppel and Crumly (1997) Sewertzoff (1923) van Leeuwen (1997) Wainwright and Bennett (1992a)
Bell (1989) Briicke (1852) Cuvier (1805) Dewevre (1895) Dumeril (1836) Duvernoy (1836a,b) Cans (1967) Gnanamuthu (1930,1937) Houston (1828) Kathariner (1895) Owen (1866) Perrault (1676) Schwenk (1983) van Leeuwen (1997)
Altevogt (1977) Altevogt and Altevogt (1954) Bell (1990) Dischner (1958) So et ah (1992) Schwenk and Bell (1988) Wainwright effl/. (1991) Wainwright and Bennett (1992a,b) Zoond (1933)
medial surface of the mandible, an open, Meckelian sulcus is apparent (Romer, 1956). 2.
Squamata
Squamates have introduced one very important novelty in skull form relative to Sphenodon and related, extinct taxa: the diapsid condition has been modified so that the lower temporal arch is absent due to loss of the quadratojugal bone and the quadrate process of the jugal (Figs. 8.3-8.6). Furthermore, the posterior margin of the lower temporal fenestra is formed entirely from the quadrate bone, which articulates somewhat loosely with the temporal arch above. In most cases, it also forms a loose joint with the pterygoid wing on its medial surface. A functional consequence of these modifications is that the quadrate is potentially free to rotate anteroposteriorly, a condition known as streptostyly (Versluys, 1912; Robinson, 1967). I emphasize "potentially" because actual streptostylic motion of the quadrate during feeding has been functionally documented only rarely (see later). Streptostyly (in the sense of potential quadrate mobility) is a diagnostic character of Squamata. Robinson (1976) thought that quadrate mobility was the initial source of selection pressure for reduction of the lower temporal bar, but Rieppel and Gronowski (1981) argued that it was expansion and thickening of a superficial adductor muscle (see later) that drove the loss.
In most squamates, the quadrate is emarginated posterolaterally and a prominent anterolateral rim supports the anterior margin of the tympanum. A minute, slender stapes runs from the tympanum across the anatomical space of the middle ear to the braincase (Fig. 8.4). Some scleroglossan taxa have no tympanum, but retain the middle ear cavity and a stapes, which attaches to the skin via a cartilaginous extracolumella. Due to expansion of the braincase in some burrowing forms, the middle ear space is virtually obliterated (Fig. 8.6). In the snout, paired septomaxillae are sometimes visible in dorsal view within the external nares (nostrils), most notably in varanids (Fig. 8.4). The septomaxilla forms part of the floor of the nasal cavity and, importantly, the roof of the skeletal capsule containing the vomeronasal organ. The complete isolation of these organs from the nasal chamber is unique to squamates (see earlier discussion). a. Iguania In general, iguanians resemble Sphenodon in skull form (Fig. 8.3). They tend to have large upper temporal fenestrae and thus are similarly broad across the temporal region. Unlike Sphenodon, however, the premaxillae, frontals, and parietals are virtually always fused into single, median bones; the premaxillae are paired in a few iguanids (Romer, 1956). The nasals remain paired, except in a few chamaeleonids (Rieppel and
194
Kurt Schwenk
FIGURE 8.2. Skull of a tuatara, Sphenodon punctatus, in dorsal (A), ventral (B), lateral (C), and posterior (D) views, a, angular; art, articular; bo, basioccipital; d, dentary; ect, ectopterygoid; eo, exoccipital; ept, epipterygoid; f, frontal; j , jugal; m, maxilla; n, nasal; opis, opisthotic; p, parietal; pal, palatine; pf, postfrontal; pm, premaxilla; po, postorbital; prf, prefrontal; pro, prootic; ps, parasphenoid; pt, pterygoid; ptf, posttemporal fossa; q, quadrate; qf, quadrate foramen; qj, quadratojugal; sa, surangular; so, supraoccipital; sq, squamosal. From Vertebrate Paleontology by Carroll © 1998 by W. H. Freeman and Company. Used with permission.
Crumly, 1997). The orbits are large to accommodate well-developed eyes. A "parietal" foramen is present, except in some chamaeleonids (Rieppel, 1981b); however, its position is variable. It is most often within the frontoparietal suture (Fig. 8.3), but is sometimes completely within the frontal or parietal bones (e.g., Estes et al, 1988; Lang, 1989). The snout is usually broad and short in iguanians. In dorsal view it is most often blunt and rounded, but it is sometimes quite triangular and pointed anteriorly (particularly in some agamids). Some taxa, notably the sceloporine iguanids ('Thrynosomatids"), or sand lizards, have flattened, wedgelike snouts with countersunk lower jaws to facilitate head-first burial in loose sand. In most chamaeleonids and some basciliscine igua-
nids, the parietal region of the skull forms a welldeveloped median crest (Fig. 8.7). The crest is formed from a blade of the parietal bone that extends posterodorsally, and in chameleons and most of the basciliscines, the jaw adductor musculature takes its origin along its length (e.g., Schwenk, 1980; Rieppel, 1981b, 1987; Rieppel and Crumly, 1997; Lang, 1989). The parietals also have ventral extensions that meet the supraoccipital. In some taxa, lateral flanges of the parietal extend partly or completely over the upper temporal fenestrae, forming a cranial "casque." This is often continued anteriorly by lateral extensions of the frontal, prefrontals, and other dermal roofing bones. Cranial casques are also common among anoloid iguanids ("Polychrotidae"). Typically in casque-headed forms.
8. Feeding in Lepidosaurs
195
F I G U R E 8.3. Skull of an iguanian lizard, Brachylophus vitiensis (Iguanidae), in dorsal (A), ventral (B), posterior (C), and lateral (D) views. (E and F) One ramus of the mandible in lateral (labial) and medial (lingual) views, respectively, aiaf, anterior inferior alveolar foramen; amf, anterior mylohyoid foramen; an, angular; ap, angular process; ar, articular; bo, basioccipital; bs, basi(para)sphenoid; cor, coronoid; den, dentary; ect, ectopterygoid; eo, exoccipital-opisthotic; ept, epipterygoid; fr, frontal; ju, jugal; la, lacrimal; mf, mental foramina; mx, maxilla; na, nasal; oc, occipital condyle; pal, palatine; par, parietal; pmf, posterior mylohyoid foramen; pmx, premaxilla; pre, prearticular; prf, prefrontal; pro, prootic; ps, parasphenoid process; ptf, postfrontal; pto, postorbital; ptr, pterygoid; q, quadrate; rap, retroarticular process; slf, supralabial foramina; smx, septomaxilla; soc, supraoccipital; sp, splenial; sq, squamosal; st, supratemporal; sur, surangular; vo, vomer. From de Queiroz (1987), with permission.
there is extensive co-ossification of the skin with the skull roof. Chameleons, in particular, show highly "sculpted'' dermal roofing bones indicative of this condition. Some chameleons have additional cranial ornamentation in the form of horns and other projections from the snout, cranium, and supraorbital region. Chamaeleo hifidis, for example, has paired, blade-like projections of the prefrontal and maxillary bones extending anteriorly from the rostrum (Rieppel and Crumly, 1997). Horns are primarily keratinous projections of the integument with a shorter, bony core projecting from the parietal and/or premaxilla. They are sometimes sexually dimorphic (e.g., de Witte, 1965) and used in male-male combat (e.g.. Bustard, 1958; Parcher, 1974), but most types of chamaeleonid cranial ornamentation are implicated in species recognition (Rand, 1961). Some anoline iguanids and agamid lizards also have elaborate, sometimes bizarre, cephalic ornaments, but these are mostly integumental. As in Sphenodon, the braincase is ossified posteriorly
and membranous anteriorly. Posttemporal fenestrae are well developed, and a median ascending process (usually cartilaginous) of the supraoccipital penetrates the parietal bone. Strong paroccipital processes extend laterally from the braincase and articulate with the temporal arches near to the quadrate's dorsal articulation (Fig. 8.3C). Basipterygoid processes of the basisphenoid are pronounced and extend anterolaterally to the medial, concave surfaces of the pterygoid wings (Fig. 8.3B). An articular cartilage is often present here. Although widely reputed to be a sliding joint in conjunction with skull kinesis (see later), in most iguanians the processes abut the pterygoid in such a way that anterior movement on the pterygoids would be impossible (Schwenk, unpublished observations). The palate is generalized and comprises, from front to back, paired vomers, palatines, ectopterygoids, and pterygoids. The premaxilla and maxillae contain the marginal teeth and contribute palatal shelves. Palatal teeth are often present on the pterygoid bones. Well-
n^DowcuL F I G U R E 8.4. Skull of a scleroglossan lizard, Varanus varius (Varanidae), in ventral (A), dorsal (B), and lateral (C) views. (D and E) One ramus of the mandible in lateral (labial) and medial (lingual) views, respectively. Note the laterally constricted temporal fenestrae and the mandibular joint posterior to the dentary evident in (D). The dentition is pleurodont and the teeth are blade-like, ang, angular; art, articular; boc, basioccipital; bsp, basisphenoid; cor, coronoid; den, dentary; ecp, ectopterygoid; epp, epipterygoid; exo, exoccipital; fro, frontal; jug, jugal; lac, lacrimal; max, maxilla; nas, nasal; pa, prearticular; pal, palatine; pap, palpebral; par, parietal; pof + poo, fused postfrontal and postorbital; prf, prefrontal; prm, premaxilla; pro, prootic; ptd, pterygoid; qut, quadrate; sep, septomaxilla; spl, splenial; squ, squamosal; sta, stapes; suo, supraoccipital; sur, surangular; tab, supratemporal; vom, vomer. From McDowell and Bogert (1954), courtesy of the American Museum of Natural History.
8. F e e d i n g in L e p i d o s a u r s
197
F I G U R E 8.5. Skulls of two scleroglossan lizards showing loss of the temporal fenestrae. Both species exhibit typical, unicuspid, pleurodont dentition. (A) Coleonyx variegatus (Gekkonidae). (B) Cordylus giganteus (Cordylidae). In Coleonyx the temporal fenestra has been eliminated by loss of the upper temporal arcade, whereas in Cordylus it has been lost through overgrowth of dermal roofing bones. Abbreviations as in Fig. 8.4. From McDowell and Bogert (1954), courtesy of the American Museum of Natural History.
developed palatal fenestrae are found between the palatines medially and the maxillae laterally. Anteriorly, another pair of narrow fenestrae lie between the vomers medially and the maxillae laterally. The vomeronasal ducts open through these anteriorly and the internal nares (choanae) posteriorly. The continuity of these openings in iguanians and some other lizards is considered primitive and is called the "paleochoanate" condition (Lakjer, 1927; Bellairs and Boyd, 1950). Posteriorly, a wide interpterygoid vacuity is pierced in the midline by a slender parasphenoid process projecting forward from the basisphenoid. The iguanian mandible is similar to that of Sphenodon with several exceptions (Figs. 8.3E and 8.3F). The mandibular joint is much tighter and typically forms a fibrocartilage symphysis. In most taxa, Meckel's cartilage persists within a mandibular canal throughout life and is continuous anteriorly at the symphysis, but it is lost in some chameleons (Bellairs, 1984). A large splenial bone is present, sometimes extending quite far anteriorly along the medial surface of the dentary and
covering the Meckelian canal so that no sulcus remains. The coronoid bone is variously developed, but is often much more pronounced than in Sphenodon. The articular facet of the jaw joint is generally more restrictive than in Sphenodon and does not permit translational movement of the mandible. Finally, the retroarticular process is also highly variable. It is small to moderate in length in iguanids, virtually absent in chamaeleonids, and often exceptionally long in agamids (Schwenk and P. Hall, unpublished results). b. Scleroglossa Skull form among scleroglossans is diverse (Figs. 8.4-8.6). It conforms generally to the pattern established previously, but certain variations are notable. Major departures include diversity in the form of the temporal region and radical modification of the skull in numerous fossorial taxa. In the latter types, the skull has been transformed into a solid, unitary structure capable of withstanding the reaction forces obtaining from head-first locomotion through the substrate.
198
Kurt Schwenk
ot-oc
pm
purposes of burrowing (Cans, 1974). Some amphisbaenians have vertically keeled snouts to enhance the displacement of soil during ramming and burrowing, whereas others have dorsoventrally compressed, wedge-like snouts for the same purpose. Amphisbaenians also have unusually complex, deeply interdigitating sutural patterns, reflecting enhanced strength and rigidity of the skull in these forms (e.g., Zangerl, 1944; Cans, 1960) (Fig. 8.6). Most important, many scleroglossan taxa have reduced or lost the upper temporal fenestrae. When present, the fenestrae are narrower than in iguanians so the temporal region of the scleroglossan skull is relatively less broad (Fig. 8.4). Loss of the fenestrae results either from obliteration by overgrowth of the surrounding bones or by loss of the arch itself (Romer,
FIGURE 8.6. Amphisbaenian skulls. (A) Amphisbaena alba (Amphisbaenidae) in lateral view. (B and C) Pachycalamus brevis (Trogonophidae) in dorsal and ventral views, respectively. Note loss of the temporal fenestrae and expansion of the braincase. Sutures are often elaborate to strengthen the skull for burrowing. Amphisbaena has large, pleurodont teeth, but the trogonophid has acrodont teeth with dental occlusion. The single, median tooth at the anterior end of the tooth row (C) is a unique feature of the amphisbaenian dentition. (A) from Rieppel (1979), (B) and (C) from Cans (1960), courtesy of the American Museum of Natural History.
including head ramming to widen or create tunnels (Cans, 1961, 1969a, 1974; Wake, 1993). With the exception of burrowing forms, however, scleroglossan skulls are typically less robust and more kinetic than in iguanians. The medial series of dermal roofing bones are variably paired or fused among scleroglossans. The premaxillae, which are fused in nearly all lizards, are paired in young individuals and adults of some scleroglossans (e.g., some gekkotans, Kluge, 1987; some skinks, Romer, 1956), and the nasal bones, usually paired, are sometimes fused (e.g., some Varanus, Greer, 1989). Skull form and snout shape are highly variable. Gekkonids, for example, typically have broad, flat skulls with relatively short snouts (Fig. 8.5A), whereas varanids often have pointed, elongate snouts that are wedge-like in profile (Fig. 8.4). Fossorial forms tend to have relatively small heads and skulls that are somewhat elongate, probably to reduce head diameter for
FIGURE 8.7. Skull and jaw musculature of a chameleon, Bradypodion pumilus. Note the parietal crest typical of many cham.eleons and some iguanid lizards, amam, adductor mandibulae medialis; amep (3a-3c), adductor mandibulae externus profundus, heads 3a to 3c; ames (lb), adductor mandibulae superficialis, head lb; bo.ap, bodenaponeurosis; dm, depressor mandibulae; lao, levator anguli oris; map, adductor posterior; m.pt, pterygoideus; ps.p, pseudotemporalis profundus; ps.s, pseudotemporalis superficialis; r.pl, rictal plate {Mundplatt). From Rieppel (1981b), with permission.
8. Feeding in Lepidosaurs 1956). Overgrowth occurs by posterior expansion of the postfrontals, medial expansion of bones forming the arch (postorbital and squamosal), fusion of the squamosal to the parietal, or some combination of these (Fig. 8.5B). Romer (1956) noted that overgrowth of the upper temporal fenestrae occurs in taxa with well-developed osteoderms and other integumental ossifications, such as the supraorbital plates found in some cordylids (e.g., van Pletzen, 1946), to which he attributed a causal basis. Fiowever, given that the function of the temporal fenestrae is poorly understood (Chapter 2), the reason for this interesting correlation remains unknown. Finally, some scleroglossan taxa have lost the upper fenestra through loss of the arch so that the widest part of the skull is across the orbits with the temporal region constricted to the width of the braincase (in small burrowing forms the posterior end of the braincase is often laterally expanded). This condition can be seen in gekkotans (Fig. 8.5A), the varanoids Heloderma and Lanthanotus, and virtually all fossorial taxa (e.g., Dibamidae, Amphisbaenia, and the small, burrowing skinks; Fig. 8.6). In such forms the jaw adductor musculature originates entirely on the side and roof of the braincase and, to some extent, from the overlying skin (e.g., Rieppel, 1979b, 1981c, 1984a,b). Ossification of the braincase is more extensive in scleroglossans than in iguanians, primarily through expansion of the prootics anteriorly along the lateral walls (Romer, 1956). In varanoids (McDowell and Bogert, 1954) and gekkotans (McDowell and Bogert, 1954; Kluge, 1987), ventral downgrowths of the frontal bones surround the olfactory bulbs of the forebrain. In fossorial forms, the braincase is further ossified and solidified, completely so in amphisbaenians (Romer, 1956; Cans, 1960), as in snakes. In these species the braincase is often "swollen" so that it closely approaches the quadrates (Fig. 8.6). As such, the paraoccipital processes are very short or absent. In most scleroglossans, however, they are well developed, as are the basipterygoid processes of the basisphenoid. The latter articulate with the pterygoid bones within medial, longitudinal grooves. The nature of this joint and the presence of a cartilaginous meniscus suggest that some sliding motion occurs here during kinetic movements of the skull in most species. The palate is generally similar to the condition described for Iguania, with some variations. The palatal fenestrae vary in size and are virtually absent in some species. In many scleroglossans, the anterior openings of the vomeronasal ducts are separated from the more posterior choanae, a condition termed "neochoanate" (Lakjer, 1927; Bellairs and Boyd, 1950). In such forms the vomeronasal openings are little more than tiny slits between the vomers and the palatal shelves of the max-
199
illae. The choanae may also be narrow. Surprisingly, a number of lizards have evolved a secondary palate in which the airway is partially or completely separated from the oral cavity by ventromedial extensions of palatal bones. In some scincids a secondary palate is merely suggested, but in others it is complete and forms a "tunnel" by downgrowths of the palatine and pterygoid bones, which extend medially over the choanae (Goppert, 1903; Greer, 1989). This would act to channel airflow posteriorly to the oropharynx in the vicinity of the glottis. Secondary palates are also found in gymnopthalmids (so-called "microteiids"; Presch, 1976), dibamids (Greer, 1985), and some amphisbaenians (Zangerl, 1944; Kritzinger, 1946; Gans, 1978). The functional significance of the secondary palate in lizards is unknown; it may be more important in strengthening the facial skeleton than in respiratory function (see discussion of the secondary palate in mammals. Chapter 2), a hypothesis consistent with its occurrence in several lineages of small burrowers. The mandible is similar to the condition described for iguanians (Figs. 8.4 and 8.5). The retroarticular process is usually moderately developed, but is sometimes reduced to a nubbin or somewhat elongated. The coronoid process is variable in size, and in some taxa the splenial extends very far anteriorly along the medial surface of each hemimandible, almost reaching the symphysis. One striking variation of the mandible is evident in varanoid lizards {Heloderma, Lanthanotus, Varanus, and the extinct mosasaurs) and snakes: each hemimandible is divided into anterior and posterior moieties that are separated by a movable, fibrous joint (McDowell and Bogert, 1954; Lee et ah, 1999) (Fig. 8.4). In Lanthanotus, Varanus, and some snakes, the joint is almost vertical, forming a simple hinge that permits mediolateral bending of the mandible and probably some torsion as well. In Varanus the mandibular joint is loose and strictly fibrous (Bellairs, 1984), as it appears to be in Lanthanotus as well (Schwenk, unpublished results). 3. Morphological Basis of Cranial Kinesis Cranial kinesis in lizards has been reviewed by Frazzetta (1962, 1983), Condon (1987), Rieppel (1993), Smith (1993), and Arnold (1998). It is important to distinguish between the inference of cranial kinesis based on skull anatomy and the actual demonstration of kinetic movements determined by functional analysis (Kardong et ah, 1997). Fiere I limit discussion to the anatomy of cranial kinesis. Its function and relevance to feeding are treated later (see Section V). Cranial kinesis refers to any intracranial mobility of the skull and therefore excludes mandibular and
200
Kurt Schwenk
hyobranchial movement. The concept derives from the early anatomical studies of Versluys (1912; reviewed in 1936) and others (e.g., Bradley, 1903; Lakjer, 1924) who noted that certain joints within the lizard skull appear to allow movement between bones. This has been confirmed by manual manipulation of fresh specimens (e.g., Frazzetta, 1962) and histological analysis of some putatively kinetic joints (e.g., van Pletzen, 1946; Condon, 1998; Schwenk, unpublished results). Versluys (1912) identified three ''types" of cranial kinesis in lizards: streptostyly refers to mobility of the quadrate bone (see earlier discussion), metakinesis refers to mobility of the dermatocranium on the braincase, and mesokinesis refers to mobility of the upper jaw or snout relative to the posterior, parietal portion of the skull. Skulls that exhibit both meta- and mesokinesis are said to be amphikinetic. Based on anatomical considerations alone, Frazzetta (1962) regarded most lizard skulls as amphikinetic. Although streptostyly is, in a general sense, a type of cranial kinesis, in most usage "cranial kinesis" refers to kinetic movements of the skull exclusive of streptostyly, which is usually discussed separately. In streptostyly, the quadrate moves on its dorsal articulation with the squamosal. This motion is primarily anteroposterior, but some lateral flexion might be possible in some species (e.g., Condon, 1987). Taxa vary considerably on the potential extent of streptostylic motion, depending on the nature of the quadratesquamosal joint (lordansky, 1990). In nearly all squamates the quadrate also articulates on its medial surface with the posterior "wing" of the pterygoid bone. This joint is also loose, but is variable in the degree of movement it allows [e.g., compare Ctenosaura, an iguanian, with very limited movement (Oelrich, 1956) and Cordylus, a scleroglossan, with significant movement (van Pletzen, 1946)]. Metakinesis implies flexion around a transverse axis defined by the lateral articulations of the paroccipital processes with the upper temporal arches. A median joint between the supraoccipital and parietal(s) also joins the braincase to the dermal skull roof. Metakinetic flexure, if it occurs at all, must be fairly complex due to the nature of these joints. Usually the axis of rotation is posterior to the medial supraoccipital-parietal joint, implying that sliding must occur at the latter point (Frazzetta, 1962), but this interpretation is problematic. In most taxa, a cartilaginous peg (ascending process) of the supraoccipital projects into a socket of the parietal bone; however, in the few taxa examined histologically, the joint would seem to permit little, if any, movement due to the nature of the collagenous fibers connecting the bones, hooking of the process, or, in some cases, actual fusion of the elements (Condon, 1998; Schwenk, unpublished
results). In mesokinesis, the principal locus of flexure is at the frontoparietal suture in the skull roof. In some species, the highly kinetic nature of this joint is indicated by significant erosion of the bones along the ventral side of the suture (Schwenk, unpublished observations). Flexure here, however, requires concomitant flexure of the palate and sliding of the pterygoids on the basipterygoid processes. Posterior movement of the pterygoids during ventroflexion of the snout on the mesokinetic joint would tend to push the quadrates backward, depending on the tightness of this joint (see earlier discussion). In addition, mesokinetic flexure requires bending in the bones of the upper temporal arcade (e.g., postorbital, prefrontal, squamosal) lateral to the frontoparietal suture. This movement is potentially distributed along the arcade in the connective tissue articulations among bones (e.g., Arnold, 1998). Finally, some flexion of the palatal bones (both dorsoventral and lateral spreading) is also implicit. Experimental evidence for streptostyly and mesokinesis in some taxa is relatively strong, but direct evidence of metakinesis is lacking. Most of the kinetic joints of the skull are fibrous. However, the basipterygoid joint and the dorsal articulation of the quadrate bone with the paroccipital process in some species are synovial (van Pletzen, 1946; Rieppel, 1978d, 1993). The quadrate-pterygoid joint sometimes develops a cartilaginous meniscus (Oelrich, 1956) or a synovial-like space (van Pletzen, 1946). The degree of movement permitted by the latter articulation varies considerably among species (e.g., Throckmorton, 1976). The potential movements of the amphikinetic lizard skull were modeled two dimensionally by Frazzetta (1962) as a quadric crank-chain system (see Fig. 2.6 in Chapter 2). This model divides the skull into several rigid, mechanically distinct moieties that are interconnected by four joints. The model demonstrates particularly how cranial kinesis permits elevation and depression of the snout (muzzle unit) relative to the braincase. In effect, the dentigerous upper jaw can be elevated and depressed as well as the lower jaw. The model also shows how the skull can be somewhat flattened by dorsoflexion of the snout relative to the rest position (see Arnold, 1998). As noted, the specifics of some of these movements are in question. In particular, the model implies that quadrate movement is coupled to snout flexion, but experimental studies suggest that these movements can be decoupled (see later). It is increasingly evident that there is significant variation in the presence, degree, and nature of skull kinesis in lizards (e.g., Rieppel, 1993; Arnold, 1998; Schwenk, unpublished results), thus, Frazzetta's (1962) model is not universally applicable. Other models are possible (e.g..
8. Feeding in Lepidosaurs lordansky, 1996). In any case, there is no doubt that streptostyly and dorsoventral movements of the snout do occur in some lizards (see later). B. Dentition The dentigerous bones of the upper jaw are the premaxilla(e) and maxillae, and in the lower jaw, the dentaries. In addition, there are small palatal teeth in some species, typically a short, somewhat transverse row of teeth across the pterygoid in squamates and a well-developed longitudinal row along the palatine in Sphenodon (see earlier discussion). Occasionally, palatine teeth occur in squamates as well. Lepidosaurs exhibit two principal types of marginal teeth: acrodont and pleurodont (see Fig. 2.7 in Chapter 2). Acrodont teeth are ankylosed directly to the apical surface of the jaw (Figs. 8.2, 8.6C, and 8.8), whereas pleurodont teeth lie along the medial side of the tooth-bearing bone, forming a rib-like pattern (Figs. 8.3-8.5). Acrodont teeth are added posteriorly to the tooth row during growth and are worn throughout life, but are not replaced. In contrast, pleurodont teeth frequently break and, in most cases, are replaced throughout life. However, the rate of replacement sometimes slows (e.g., Mateo and Lopez-Jurado, 1997) and in some pleurodont species tooth wear is significant in older individuals (e.g., Rocek, 1980b). Acrodont teeth are characteristic of Sphenodon and the iguanian families Agamidae and Chamaeleonidae, as well as one amphisbaenian family, the Trogonophidae. All other squamates have pleurodont teeth. Curiously, the anteriormost teeth in most agamids are pleurodont and some of these are usually long and "caniniform" (e.g.. Cooper et ah, 1970). The posterior, molariform teeth in some large teiids {Dracaena, Tupinambis) have a very broad base that extends well away from the lateral margin of the dentary and mandible (e.g., Peyer, 1929; Presch, 1974; Dalrymple, 1979). These have been variously described as "subpleurodont" (e.g., Dalrymple, 1979), "subacrodont," or ''acrodont" (MacLean, 1974), but are clearly a modified type of pleurodont dentition with replacement, albeit at a reduced rate (e.g., Edmund, 1969; Dalrymple, 1979). 1. Pleurodont
Dentition
In general, upper and lower pleurodont teeth bear no special positional relationship to one another except that the lower tooth row generally fits inside the upper toothrow. As such, there is no functional requirement to maintain precise alignment between upper and lower teeth as there is in mammals (Chapters
201
2 and 13). Frequent breakage and replacement of teeth along the tooth row result in a slight, but constant shifting of functional tooth position, but this has no functional consequence. Teeth are replaced in maturational waves along the tooth row called Zahnreihen (Edmund, 1960, 1969). There are two distinct types of tooth replacement in pleurodont squamates: the "iguanid method" and the "varanid method" (reviewed by Rieppel, 1978f). In the iguanid method, a replacement tooth grows into the pulp cavity of the functional tooth from below as the latter is resorbed. In the varanid method, a replacement tooth grows in an interdental position posterior to the functional tooth and no resorption is evident. The iguanid pattern is typical of most pleurodont squamates and is believed to be primitive, whereas the varanid pattern is found only in varanoid lizards and snakes. An intermediate type of tooth replacement occurs in basal anguimorphans, teiids, and some amphisbaenians. Rieppel (1978f) argued that the similarity between snakes and varanoids is superficial and suggested that similar tooth replacement methods have evolved independently in several groups. Crown shape is highly variable in pleurodont squamates, although the "typical" tooth is unicuspid and peg-like to sharp. It has been suggested that there is a fairly tight correlation between diet and crown shape in lizards (e.g., Hotton, 1955; Montanucci, 1968); however, these studies did not take into account the strong phylogenetic component of tooth crown variation. On a broad scale there appears to be little diet-related variation in crown form. The vast majority of pleurodont squamates have numerous, relatively small, unicuspid to tricuspid teeth. These tooth forms are associated with a variety of invertebrate prey types as well as some percentage of plant food (see Section III,A). However, invertebrate prey types are extremely diverse in size, hardness, and behavior, and no clear adaptive patterns emerge. Two dietary types, however, seem to promote strong adaptive modification in the teeth. In many herbivorous lizards the teeth are laterally compressed and multicuspate. The number of cusps and therefore the width of the spatulate teeth roughly correspond to the degree of herbivory (Hotton, 1955; Montanucci, 1968). Throckmorton (1976) noted the such teeth lie at a slight angle relative to the axis of the jaw and each tooth slightly overlaps the succeeding one. With the jaws closed, upper and lower tend to alternate in line. This pattern results in a clean and continuous bite to remove a section of leaf or other vegetable matter, as opposed to a series of individual punctures. This is especially evident in iguanine iguanids (Figs. 8.3D-8.3F), but has also evolved
202
Kurt Schwenk
independently in other taxa that include a high proportion of plant matter in their diets. Second, species that consume hard food items (durophagy), such as shelled molluscs, tend to form molariform teeth at the back of the tooth row (Peyer, 1929; Dalrymple, 1979; Rieppel, 1979a; Rieppel and Labhardt, 1979; Edmund, 1969; Estes and Williams, 1984). These are typically very broad based with blunted cusps and are especially well developed in several large teiids (especially the caimen lizard, Dracaena guianensis) and the Nile monitor, Varanus niloticus, which feed frequently on hard-shelled, aquatic molluscs and crustaceans. Limited dentitional adaptation to diet may have occurred in some clades. Sumida and Murphy (1987) and Bauer and Russell (1990), for example, suggested a correlation between tooth crown form and diet in geckos. However, Sumida and Murphy (1987) offered no dietary data and their conclusions were purely speculative. Bauer and Russell (1990) argued that several dentitional types accommodate an insectivorous diet in the genus Rhacodactylus, but large, caniniform teeth in R. auriculatus and several other species are adapted to vertebrate (and other soft-bodied) prey. Limited dietary data support their view. In some lacertid lizards there is an ontogenetic change in tooth crown form that is correlated with a dietary shift (Rocek, 1980a; Mateo and Lopez-Jurado, 1992, 1997). Juveniles have a very regular tooth row with small, tricuspid teeth and eat a relatively diverse, but mostly soft-bodied diet. In adults the tooth crowns become unicuspid and caniniform and vary in size along the tooth row. Adult prey diversity is higher and hard-bodied prey predominate. Mateo and Lopez-Jurado (1997) suggested that adult dentition is adapted to puncturing tough insect cuticle. Finally, the skink-eating specialist, Lialis (Pygopodidae), has sharp, recurved teeth that can fold posteriorly due to fibrous bases (Patchell and Shine, 1986c). These modifications parallel those found in some skink-eating snakes (Savitzky, 1981) and help prevent these notoriously hard and slippery lizards from escaping the jaws' grasp. Varanoid lizards typically have very sharp unicuspid teeth that are subcorneal to flattened and slightly recurved (e.g., McDowell and Bogert, 1954) (Fig. 8.4). These are often quite snakelike in appearance. Although such tooth form is usually attributed to carnivory, most varanids primarily consume a variety of invertebrates (Losos and Greene, 1988). The teeth of varanids that include vertebrates in their diets are particularly flattened mediolaterally (i.e., bladelike) and have minute serrations along their posterior margins, or both anterior and posterior margins (e.g., Auffenberg, 1981; Greer, 1989; Abler, 1992). Similar teeth occur in the dinosaur Tyranosaurus; mechanical tests
showed that they cut through flesh like a serrated steak knife using a "grip-and-rip" mechanism (Abler, 1992). Other functional ramifications of these teeth for Komodo monitors (and other varanids) are considered in Section VI. In the only venomous lizards, Heloderma, anterior and posterior grooves occur along the length of most teeth, apparently to enhance venom introduction. These are best developed in the mandible near to the point of venom discharge and are accompanied by ridges interpreted as "cutting edges'' (Bogert and Martin del Campo, 1956). Helodermatids also frequently consume vertebrates (Pregill et al, 1986). 2. Acrodont
Dentition
The most significant aspect of acrodont dentition is that it is not replaced during life. Rather, new teeth are added during growth while existing teeth wear down (Cooper ei al, 1970; Cooper and Poole, 1973; Robinson, 1976; Throckmorton, 1979). An important functional consequence of this is the stability of tooth position and the potential for upper and lower tooth alignment, or occlusion (as in mammals; see Chapters 2 and 13). Occlusion is characteristic of agamid lizards and, to a lesser extent, amphisbaenians, but not Sphenodon. In agamids, the lower tooth row fits tightly within the upper tooth row, with upper and lower teeth alternating (Fig. 8.8). The triangular form of agamid teeth in occlusal view and the tight fit of uppers and lowers produces a cutting mechanism that has been compared to the "pinking shears" used in sewing (Edmund, 1969). Importantly, as the teeth slide past one another they are worn so that fit is enhanced and sharp cutting edges are produced. Over time wear can be extreme so that all enamel on the abraded side is worn away, exposing the pulp cavity. The pulp cavity fills with bone or calcified tissue and the bone beneath the tooth changes from cancellous to compact (Cooper and Poole, 1973; Throckmorton, 1979). As tooth wear progresses, the teeth begin to cut into the bone of the jaw, with most wear concentrated on the lower jaw. In some old individuals of Uromastix, the teeth are completely worn away, but the apical margins of the dentaries have been worn into serrations and continue to function as tooth analogues (Cooper and Poole, 1973; Throckmorton, 1979, Robinson, 1976)! Loss of the premaxillary teeth through wear creates a beak-like appearance in some Uromastix and the bony edge of the premaxilla is used to crop vegetation (Throckmorton, 1979). Such wear is not as pronounced in most other agamids, nor in chamaeleonids. Uromastix evinces two adaptations of the enamel related to tooth wear and function. First, the enamel is relatively thicker than in other agamids. Second, it is
203
8. Feeding in Lepidosaurs
FIGURE 8.8. Acrodont dentition and dental occlusion in agamids and amphisbaenians. (A) Ctenophorus decresii (Agamidae) showing anterior, caniniform, pleurodont teeth and marginal acrodont teeth in lateral view. (B) Upper (white) and lower (black) tooth occlusion in Agama agama (Agamidae) in occlusal view. (C) Similar view in an acrodont amphisbaenian. (A) From Greer (1989), (B) from Cooper et al. (1970) with the permission of Cambridge University Press, and (C) after Cans (1974), with permission.
unique among living reptiles in having prismatic microstructure similar to that of mammals (Cooper and Poole, 1973). However, very few lizards have bee examined in this regard, so there may be others with similar enamel structure. The potential to develop prismastic enamel is considered a mammalian synapomorphy (Wood et ah, 1999), thus its occurrence in a lizard is of interest. Functionally, it suggests that the mechanical properties of Uromastix enamel are isotropic, i.e., the same in all direction, as opposed to the typically anisotropic nature of reptile enamel, which might suggest a greater complexity of forces acting on the teeth (Cooper and Poole, 1973). Robinson (1976) argued that streptostylic movements of the quadrates permit slight adjustments of lower jaw position to maintain precise occlusion in agamids. This hypothesis is supported by the observa-
tion of streptostyly during food processing in Uromastix, but not in the pleurodont Iguana (Throckmorton, 1976). Herrel and De Vree (1999) also documented streptostylic movements in Uromastix. However, streptostyly was not observed in Agama {Plocederma) (Herrel et al, 1996a). It is possible that precise occlusion is more important in the herbivorous Uromastix than in other acrodonts due to the necessity of cropping fibrous plant food (see Section V). In Sphenodon the acrodont teeth do not function as in agamids and occlusion does not occur. Rather, the lower tooth row fits precisely into the groove between parallel, upper tooth rows (see earlier discussion) (Fig. 8.2). During food processing, the lower jaw slides anteroposteriorly along the upper jaw (Gorniak et ah, 1982) so that food trapped between upper and lower teeth is effectively sheared. Propalineal movement is permitted by sliding of the mandible on the quadrate at the jaw joint and bending of the mandibular rami at the symphysis, and is not the result of streptostyly (Robinson, 1976). Based on limited fossil evidence, Robinson (1976) argued that the condition in Sphenodon is derived from an agamid-type pattern of acrodont occlusion. Amphisbaenians also exhibit limited dental occlusion (Fig. 8.8). This is true for both pleurodont and acrodont species (Gans, 1969a, 1974), but presumably greater opportunities for wear and the production of cutting facets are available to the acrodont trogonophids. Gans suggested that amphisbaenians use their occluding tooth rows to bite pieces out of large prey items. The bite is often accompanied by axial twisting to tear off chunks. As noted earlier, however, there are no dietary data to support this adaptive scenario (see Section VI).
C. Hyobranchial Apparatus 1. Hyobranchial Skeleton The lepidosaurian hyobranchial apparatus is exceptionally variable in form and has not been meaningfully reviewed since Fiirbringer (1922). I will not challenge this trend here. Clearly, this is an area badly in need of a comparative, phylogenetic synthesis. I suspect that a careful analysis of hyobranchial form, including microscopic anatomy and development, could provide a rich source of phylogenetically informative characters, not to mention a clearer understanding of hyobranchial function during feeding and other behaviors. The hyobranchial apparatus is the remnant of the ancient pharyngeal skeleton of tetrapods' piscine ancestors. Along with the primary (nondermatocranial)
204
Kurt Schwenk
jaw elements and their derivatives, the hyobranchium constitutes the splanchnocranial portion of the head skeleton (see Chapter 2). Most nonophidian lepidosaurs possess a generalized hyobranchial apparatus retaining some of the ancestral patterning of the visceral arches. In lepidosaurs, as in other tetrapods, it supports the tongue and the musculature of the throat. In most species the hyobranchium lies far anterior in the buccal floor between the mandibular rami. In Varanus it is positioned farther back in the throat, posterior to the mandible (e.g.. Smith, 1986). This shift is correlated with a specialization of neck and throat morphology related to derived respiratory and display functions (Bels et al, 1995; Owerkowicz et al, 1999,2000). The entire apparatus is free-floating within the buccal and gular musculature. It is heuristically useful to describe an "archetypal" lepidosaurian hyobranchial apparatus based on its most generalized adult form (e.g., Langebartel, 1968; Rieppel, 1981a) and its development (see especially Kallius, 1901; also El-Toubi and Kamal, 1959a,b), and then consider departures from this pattern. The "archetypal" condition is roughly characteristic of Sphenodon and most iguanid, agamid, gekkonid, lacertid, and scincid squamates (Fig. 8.9). Its presence in basal lepidosaurian clades suggests that it represents the primitive condition. The generalized lepidosaurian
F I G U R E 8.9. Generalized hyobranchial apparatus (Pogona muricatus, Agamidae) shown in ventral view relative to the mandible. Stippling indicates cartilage, ap, anterior process of basihyal; bh, basihyal; cbl, ceratobranchial 1; cb2, ceratobranchial 2; ch, ceratohyal; m, mandible. After Throckmorton et al (1985), with permission of Wiley-Liss, Inc., a subsidiary of John Wiley and Sons, Inc.
hyobranchium consists of a cartilaginous hyoid body, or basihyal which lies in the ventral midline. The basihyal has three anterior, cartilaginous projections: a median lingual (entoglossal) process and paired anterior processes. The lingual process projects anteriorly into the muscular body of the tongue for one-half to threequarters or more of its resting length (Schwenk, unpublished results). A laryngohyoid ligament runs posteriorly from its middle or anterior end to the ventral surface of the larynx (Oelrich, 1954; Schwenk, 1986, 1988; see later). The anterior processes (hypohyals) project anterolaterally and dorsally. The ends of these articulate by means of fibrous joints to large ceratohyals which extend posteriorly and dorsolaterally, wrapping around the sides of the throat. Medial to these are the first ceratobranchials, which project posteriorly from the posterolateral corners of the basihyal, distally curving dorsolaterally around the throat. They are the only hyobranchial elements to be consistently ossified and they typically join the basihyal at a synovial joint. Finally, second ceratobranchials are cartilaginous extensions of the basihyal that project posteriorly, in parallel, near the midline. They are closely apposed and elongated in species with extensible dewlaps (e.g., von Geldern, 1919; Font and Rome, 1990). The ceratohyals, first ceratobranchials and occasionally the second ceratobranchials, sometimes have small, cartilaginous distal elements appended to their tips called epihyals and epibranchials (e.g.. Tanner and Avery, 1982). The basihyal, lingual process, hypohyals and ceratohyals are derivatives of the second visceral arch. The first ceratobranchials are third arch derivatives and the second ceratobranchials are fourth arch derivatives (Kallius, 1901). The most common variant of the archetypal hyobranchial form is loss of the second ceratobranchials. The remaining ceratobranchials are assumed to be the first pair because they are usually positioned somewhat laterally on the basihyal, are ossified and articulated by synovial joint to the basihyal, and serve as the origin for the hyoglossus muscles (e.g., Langebartel, 1968). Circumstantially, it is also most likely that the most posterior, fourth arch elements would be lost given the general tendency in tetrapods to reduce and lose posterior visceral arch elements first. Beyond this, the hyobranchial form is remarkably variable, as noted. Virtually all (nonophidian) taxa retain vestiges of the basihyal and entoglossal process, but all other elements are reduced or lost in various and sundry patterns. As is typical in other cranial elements, small fossorial scleroglossans show the greatest divergence in hyobranchial form with reduction and loss of elements rampant (e.g., Rieppel, 1981a). Some (e.g., the burrowing skink, Acontius) approach the highly reduced snake condition in which the hyobranchium is little more
205
8. Feeding in Lepidosaurs than a Y-shaped element. The paired, posterior elements are presumed to be the first ceratobranchials. In teiids and gymnopthalmids the entoglossal process is detached from the basihyal (Beebe, 1945; Tanner and Avery, 1982; Harris, 1985). The basihyal retains the base of the process and is connected to the detached portion by a "tubular membrane" (Harris, 1985). In at least one small sphaerodactyline gecko {Gonatodes), the lingual process does not extend into the tongue, but into the floor of the mouth instead (Schwenk, unpublished results). Although most iguanians retain the archetypal hyobranchial form, chameleons have diverged significantly from this pattern (e.g., Fiirbringer, 1922; Gnanamuthu, 1930, 1937; Lubosch, 1932; Fineman, 1943; Tanner and Avery, 1982; Wainwright and Bennett, 1992a). This is not surprising in light of extensive modification of the hyolingual apparatus for ballistic tongue projection (see Section VI). The basihyal is reduced to little more than the posterior end of a robust lingual process. The lingual process is remarkable for its thickness, length, and cylindrical, parallel-sided form. The latter feature is particularly important to tongue function during projection (e.g., van Leeuwen, 1997; see later). The ceratohyals are lost, but the anterior processes remain as independent elements movably articulated with the basihyal. Robust first ceratobranchials, likewise, articulate with the minute basihyal. At rest they extend anterodorsally along the back of the skull, a position very different from other iguanians. There are no second ceratobranchials. 2. Hyobranchial
Musculature
As for the hyobranchial skeleton, hyobranchial musculature is complex and extremely variable. The extreme variation in this musculature makes generalizations difficult and the homologies of some muscles questionable. Excellent treatments of individual taxa include Rieppel (1978e) for Sphenodon, Oelrich (1956) for an iguanid {Ctenosaura), and Smith (1986) for Varanus. Reviews can be found in Camp (1923; superficial throat musculature only), Gnanamuthu (1937), Kesteven (1944), and Tanner and Avery (1982); however, evolutionary patterns remain virtually unexplored. Hyobranchial musculature is divisible into intrinsic and extrinsic components, and the latter into prehyoid and posthyoid muscles. All muscles are paired. Intrinsic muscles are those that run from one element of the hyobranchium to another. This musculature is limited in most taxa, but in Varanus and other taxa with elaborate hyobranchia it may be more extensive and complex. The principal intrinsic muscle is the ceratohyoideus (branchiohyoideus), which generally runs between the second ceratobranchial and the ceratohyal. In some
species (notably iguanines) it runs between the two ceratobranchials (Avery and Tanner, 1971). According to Tanner and Avery (1982), a stout ceratohyoideus runs from the basihyal anterodorsally to the anterior process (hypohyal), but other references do not mention this. The ceratohyoideus is innervated by the glossopharyngeal nerve (c. n. IX) (Oelrich, 1956). The principal extrinsic, prehyoid muscle is the mandibulohyoideus (geniohyoideus). There is some confusion about its name. Early authors tended to use the name geniohyoideus, which is taken from mammalian anatomy, but most recent workers prefer mandibulohyoideus, which more accurately describes the muscle in lepidosaurs. Some, however, have described both muscles in lepidosaurs (e.g., Sondhi, 1958b; Tanner and Avery, 1982). Usually, "geniohyoideus" is reserved for the most medial, longitudinal portion of the muscle. In any case, the mandibulohyoideus represents a broad series of slips originating along the posteromedial margin of the mandible and running posteriorly to the hyobranchium where fibers insert primarily onto the first ceratobranchial, but also on the ceratohyal and the second ceratobranchial. Along the way these sheetlike muscle slips often interdigitate with transverse fibers of the intermandibularis muscle running between the mandibular rami (e.g., Oelrich, 1956; see later). Fiber orientation suggests that this muscle protracts the hyobranchium (and the tongue) and abducts the hyobranchial horns. It is innervated by the hypoglossal nerve (c. n. XII), as are the lingual muscles (Oelrich, 1956). Two large posthyoid muscles attach the hyobranchial apparatus to the sternum and shoulder girdle. The sternohyoideus runs near the midline from the anterior margin of the sternum to the first ceratobranchial (e.g., Oelrich, 1954; Rieppel, 1978e). In some species, part of it originates from the clavicle and insertion can occur on the basihyal and/or ceratohyal (Tanner and Avery, 1982). A separate sternothyroideus is recognized by Tanner and Avery (1982), who suggest that in some taxa it is distinct from the sternohyoideus, but in others it is variously fused. Finally, an omohyoideus originates on the clavicle, interclavicle, and/or sternum and curves anteromedially to insert on the first or second ceratobranchia and/or the basihyal. In some species it has two heads that join anteriorly before inserting onto the hyobranchium. These muscles are innervated by motor fibers of the first spinal nerve (Oelrich, 1956). D . Jaw Musculature The jaw musculature is an ill-defined set of muscles that act directly on the jaws or which potentially affect jaw movements. These muscles can be placed into four general groups based on their innervation (and
206
Kurt Schwenk TABLE 8.3 Muscles Related to Feeding Function in Lepidosaurs" Muscle group (innervation)
Connections
Muscle
Superficial throat (c. n. V, VII)
Intermandibularis ant. Intermandibularis post. Constrictor colli
Superficial transverse fibers linking the mandibular rami and forming a sling around the throat beneath the hyobranchium
Hyobranchial: intrinsic (c. n. IX)
Ceratohyoideus
links second ceratobranchial and ceratohyal, or both ceratobranchials
Hyobranchial: prehyoid (c. n. XII)
Mandibulohyoideus
Mandible to first ceratobranchial (± ceratohyal and second ceratobranchial)
Hyobranchial: posthyoid (1st spinal n.)
Stemohyoideus (+ stemothyroideus) Omohyoideus
Sternum (± clavicle) to first ceratobranchial (± basihyal, ceratohyal)
Constrictor internus dorsalis (c. n. V)
Levator pterygoideus Protractor pterygoideus Levator bulbi
Skull roof (parietal) to pterygoid Ventrolateral braincase to posterior pterygoid Braincase or skull roof to lower eyelid (or pterygoid)
Adductor mandibulae externus (c. n. V)
Levator anguli oris A. m. e. superficialis A. m. e. medialis
Upper temporal arch (lower in Sphenodon) to rictal plate at corner of mouth Upper temporal arch (± quadrate) to lateral surface of mandible Lateral surface of parietal and posterior margin of upper temporal fenestra to lateral surface of basal aponeurosis attaching to coronoid bone of mandible Posterior wall of upper temporal fenestra to posteromedial surface of basal aponeurosis
A. m. e. profundus
Clavicle, interclavicle (± sternum) to first or second ceratobranchial (± basihyal)
Adductor mandibulae internus (c. n. V)
Pseudotemporalis
Adductor mandibulae posterior (c. n. V)
A. m. posterior
Anteromedial margin of upper temporal fenestra and epipterygoid bone to basal aponeurosis and coronoid bone of mandible Ectopterygoid, pterygoid and quadrate to posterior end of mandible + retroarticular process + lateral surface of mandible Quadrate to medial surface of mandible
Lingual: "extrinsic" (c. n. XII)
Genioglossus Hyoglossus
Mandible near symphysis to tongue First ceratobranchial to tongue + intrinsic fibers in tongue
Lingual: "intrinsic" (c. n. XII)
Verticalis Transversalis Longitudinalis Circular
Unpaired: floor of tongue beneath lingual process to dorsal transverse septum From midline laterally above dorsal transverse septum At base of papillae dorsal to transversalis Surround hyoglossus bundles; confluence of transversalis and verticalis fibers
Pterygoideus
"^See text for description, and text and Tables 8.1 and 8.2 for references. All muscles are paired, except the intrinsic lingual muscle, verticalis, which lies in the midline of the tongue. Terminology primarily follows Haas (1973), with additional reference to Oelrich (1956) and Schwenk (1986).
segmental derivation) (Table 8.3): (a) the trigeminal group (c. n. V, first visceral arch) includes the adductor musculature, the constrictor internus dorsalis muscles associated with cranial kinesis, and the anterior intermandibularis muscle of the throat; (b) the facial group (c. n. VII, second visceral arch) includes the depressor mandibulae muscle and the posterior intermandibularis and constrictor colli throat muscles; (c) the hypoglossal group (c. n. XII, hypobranchial muscles), including the lingual and prehyoid muscles; and (d) a posterior group (spinal nerves) comprising the posthyoid muscles. Here I limit discussion to muscles that act more or less directly on the jaws—the trigeminal nerve musculature associated with jaw closure and cranial kinesis and the facial nerve musculature, which
serves to open the jaws and constrict the pharynx. Preand posthyoid muscles of the hyobranchium were described earlier and the lingual musculature is described later. The cervical system is also of potential importance during some feeding functions, but is not treated here. Cervical musculature was reviewed, to some extent, by Versluys (1898) and its condition in helodermatids was described by Herrel and De Vree (1999b), who correctly pointed out its importance in generating inertial feeding thrusts in some squamate taxa (see later). The cervical epaxial muscles are also undoubtedly important in elevating the cranium at the atlanto-occipital (neck) joint, both to raise the head and to elevate the upper jaw during mouth opening (see Section V).
207
8. F e e d i n g in L e p i d o s a u r s
Haas (1973) and Gomes (1974) provided benchmark syntheses of lepidosaurian jaw musculature. Some more recent literature is listed in Table 8.1. These studies provide ample information on the extensive taxonomic variation evident in the jaw muscles, but such detail is largely beyond the scope of this review. Rather, I attempt to distill the features of the system to a form salient to interpreting feeding function. I follow primarily the terminology of Haas (1973), which is based largely on the work of Lakjer (1926). Haas (1973) provided an extensive list of muscle synonymies. 1, Muscles Innervated by the Trigeminal Nerve The traditional basis for recognizing and naming different trigeminal muscles is on the basis of their position relative to the three principal rami of the trigeminal nerve: medial opthalmic ramus, middle maxillary ramus, and lateral mandibular ramus (Lakjer, 1926; Save-Soderbergh, 1945; Haas, 1973). Nonetheless, there is significant variation in the extent to which these muscle masses are separated from one another and the extent to which they are internally subdivided. In many cases, muscles are separate at their origins but fuse distally near their insertions. Thus, patterns of splitting and fusion in the trigeminal muscles often obscure homologies of individual muscles, as they do elsewhere in the body. Furthermore, variation in the form and extent of the upper temporal arcade and fenestra among squamate taxa is accompanied by differences in muscle origin. Most of the major trigeminal adductor muscles attach to an extensive insertional tendon called the bodenaponeurosis (Fig. 8.7). The bodenaponeurosis inserts onto the coronoid bone and its vicinity on the mandible. In most taxa it extends as a sheet some distance into the muscle mass, variously separating the muscles. Table 8.3 lists the principal lepidosaurian jaw muscles based on Haas (1973). These groups are briefly described here and illustrated, in part, in Figs. 8.7 and 8.10. a. Constrictor Internus Dorsalis Group These muscles lie very deep within the skull, medial to the opthalmic ramus of the trigeminal nerve. Two are implicated in kinetic movements of the skull (e.g., Frazzetta, 1962; lordansky, 1970; Haas, 1973). A protractor pterygoideus muscle originates on the prootic and basisphenoid, including the basipterygoid process, and runs posteriorly and ventrolaterally to insert along the posterior end of the pterygoid wing (lamina). The levator pterygoideus muscle runs vertically from the ventral surface of the parietal bone, along the epipterygoid, to the pterygoid wing lateral to the protractor. These muscles are widely presunied to participate in
adms
pte
FIGURE 8.10. Superficial jaw musculature in a generalized lizard, Ctenosaura (Iguanidae). adms, adductor mandibulae externus superficialis; admm, adductor mandibulae externus medialis; ang, angular; ar, articular; cor, coronoid; dmd, depressor mandibulae; es, extracolumella (extrastapes); ju, jugal; po, postorbital; pst, pseudotemporalis; pte, pterygoideus; q, quadrate; rp, retroarticular process; sq, squamosal; sur, surangular; ty, tympanum. From Bellairs (1970).
dorsoflexion of the snout (e.g., Frazzetta, 1962; lordansky, 1970), although Smith and Hylander (1985) disputed this interpretation and suggested that they serve to resist movement at the basipterygoid joint. Ventroflexion is putatively affected by muscles of the adductor complex (e.g., Borsuk-Bialynicka, 1985). A third muscle in this group, the levator hulhi, is associated with the eyeball in most lepidosaurs, but in amphisbaenians and snakes it inserts onto the dorsal surfaces of palatine and pterygoid bones to become a functional retractor pterygoideus, in opposition to the protractor (Haas, 1973). Rieppel (1981b) noted that the levator is absent in chameleons, but that the protractor is well developed in some, with a strong head to the quadrate bone. He suggested that the latter muscle functions as a protractor quadrati, a hypothesis consistent with the fact that chameleons are akinetic but streptostylic. Due to the presence of a complete lower temporal arch, Sphenodon is presumably akinetic; however, it has been suggested that juveniles with incomplete ossification might evince some kinesis (e.g., Frazzetta, 1983). Ostrom (1962) found that the levator and protractor pterygoideus muscles were variably developed in adults and completely absent in some individuals. He suggested that individual variation reflected variation in the degree of cranial kinesis, speculating that dietary differences among individuals or populations might underlie the difference. However, there is no support for this supposition, nor any functional evidence of skull kinesis in adult tuatara (Gorniak et ah, 1982).
208
Kurt Schwenk
b. Adductor Mandibulae Externus The external muscle group lies between or lateral to the mandibular and maxillary branches of the trigeminal nerve. It is divided into superficial, medial, and lateral portions. The superficial part is, itself, usually subdivided into at least two parts: a levator anguli oris and an adductor mandibulae externus superficialis, sensu stricto. The levator anguli oris originates on the lower temporal arcade in Sphenodon, but in squamates it has shifted to the upper arch, or from the dorsolateral fascia overlying more medial muscles in taxa with the upper arch reduced. It inserts into the tough connective tissue at the corner of the mouth called the rictal plate, or Mundplatt. The superficialis lies more or less posterior to the levator and arises primarily from a quadrate aponeurosis, running anteroventrally (slightly posteroventrally in Sphenodon) to the lateral surface of the mandible behind the coronoid bone, with some fibers occasionally inserting onto a lateral portion of the bodenaponeur osis. The adductor mandibulae externus medialis is often the largest of the temporal adductor muscles, nearly filling the temporal fossa in some taxa. It is highly variable in form due, in part, to variation in the upper temporal arcade. In general, it arises from the posteromedial portion of the upper temporal arch, or the side of the braincase, and lies lateral to the bodenaponeurosis into which it most often inserts. Occasionally it attaches along the mandible posterior to the coronoid. The adductor mandibulae externus profundus lies deep to the medialis alongside the braincase. It inserts onto the medial surface of the bodenaponeurosis, which separates it from the medialis. Some fibers take origin from the anterior surface of the quadrate and bypass the bodenaponeurosis, inserting directly into a trough in the surangular bone of the mandible (posterior to the coronoid). In some cases, fibers of the profundus arise from the posterior margin of the temporal arcade and bulge posteriorly through the posttemporal fenestra. c. Adductor Mandibulae Internus The two muscles of this group are anatomically distinct, but share innervation by the mandibular ramus of the trigeminal. The pseudotemporalis has two parts in most squamates (superficialis and profundus: lateral and medial, respectively). The superficialis is often a large muscle that lies in the anteromedial portion of the temporal fossa. It arises from the anterior portion of the parietal bone and the top of the epipterygoid and runs quite vertically to the coronoid bone and the medial surface of the bodenaponeurosis, in some cases. It is lost in gekkotans (Haas, 1973). The profundus por-
tion lies lateral to the epipterygoid bone and levator pterygoideus muscle. It originates on the ventral end of the epipterygoid, fans out as it runs posteroventrally, and has a fleshy insertion on the coronoid bone. The two parts of the pseudotemporalis are sometimes fused at the point of insertion (e.g., Oelrich, 1956). The pterygoideus (pterygomandibularis) is a complex muscle that is the largest jaw adductor in some species. It runs laterally, posteriorly, and ventrally from the ectopterygoid and pterygoid bones to the posterior end of the mandible, including the retroarticular process, which it wraps in a kind of sling. The muscular "bulb" it forms around the posterior end of the mandible is sometimes referred to as the "masticatory cushion." It is often sexually dimorphic and hypertrophied in males, giving them a characteristic "jowly" appearance. The skin overlying the muscle is sometimes associated with enlarged scales (e.g.. Iguana) or seasonally bright coloration (e.g., Eumeces), suggesting that hypertrophy of these muscles may be related to display behavior. However, there may be mechanical and ecological consequences of sexual dimorphism for feeding function as well (Herrel et ah, 1996b; Lappin and Swinney, 1999). The pterygoideus is also usually divided into superficialis (lateralis) and profundus (medialis) portions. The latter lies more dorsally and its fibers are shorter and more transverse. The agamid, Uromastix, is unique in its possession of a third pterygoideus slip (e.g., Haas, 1973; Throckmorton, 1976; Herrel et al, 1999a). The pterygoideus is complexly pinnate with a central tendon. It is also functionally complex because it acts on the mandible both anterior and posterior to the jaw joint. Its position suggests that it can function as a jaw opener as well as a jaw closer (e.g., Oelrich, 1956; lordansky, 1970; Throckmorton, 1976; Schwenk and Hall, unpublished observations) and Throckmorton (1978,1980) showed that the pterygoideus profundus is indeed active during jaw opening in Uromastix; however, it may be most important in mandibular protraction (Throckmorton, 1976, 1978). Herrel et al (1997a) and Gorniak et al (1982) also found differential activity in the pterygoideus muscle in Agama and Sphenodon, respectively. Differential fiber recruitment in portions of a single, complex muscle, as evident in the lepidosaurian pterygoideus (and other muscles), is usually viewed as a mammalian trait (see Chapter 2). d. Adductor Mandibulae Posterior This is a relatively small muscle lying deep to the adductor mandibulae externus profundus. However, it takes its origin from the quadrate bone (and, in some cases, the pterygoid and paroccipital process) and runs anteroventrally to insert on the posteromedial surface of the mandible.
8. Feeding in Lepidosaurs e. Intermandibularis Anterior A more or less continuous series of transverse fibers cover the floor of the mouth and pharynx. The most anterior of these thin muscle sheets is the intermandibularis anterior (mylohyoideus). Note that the name "intermandibularis'' is preferred over "mylohyoideus/' the latter being derived from mammalian anatomy. It runs across the floor of the buccal cavity between the mandibular rami just beneath the skin. It is somewhat arbitrary to describe this muscle separate from its serially homologous portions just posterior (described in the next section), but due to its anterior position it is derived from the first visceral arch and shares its trigeminal innervation with the adductor and internus dorsalis musculature. Contraction of the intermandibularis is presumed to elevate the floor of the mouth. 2. Muscles Innervated by the Facial Nerve a. Intermandibularis Posterior This muscle lies immediately posterior to the intermandibularis anterior and is usually separated from it by an interdigitating slip of the mandibulohyoideus (Camp, 1923; see earlier discussion). b. Constrictor Colli Posterior to the mandible and the intermandibularis, the constrictor (sphincter) colli muscle comprises a third set of transverse fibers, forming a superficial sling beneath the hyobranchium and around the pharynx. Fibers arise from the neck near the dorsal midline immediately posterior to the head. Contraction of this muscle elevates the hyobranchium and constricts the pharynx which may be especially important during swallowing. c. Depressor Mandibulae This muscle depresses the mandible and is presumed to be the principal jaw opener. However, the pre- and posthyoid muscles may contribute to jaw opening in many cases and the pterygoideus muscle is also implicated in some taxa (see earlier discussion). Like the constrictor colli muscle, which often overlies it, the depressor mandibulae is a superficial neck muscle. It originates on the posterior margin of the parietal bone and upper temporal arch at the back of the skull, and from the fascia overlying the neck at the dorsal midline. It typically forms a fan-shaped muscle that runs ventrally to insert on the retroarticular process of the mandible. Haas (1973) considered the posterior, cervical portion a separate muscle, the cervicomandibularis. The cranial part of the depressor is clearly divided in some species, notably in agamids with well-
209
developed retroarticular processes, in which there is an accessory head of the depressor running from the lateral surface of the temporal arcade. Gorniak et al. (1982) identified anterior and posterior portions of the depressor in Sphenodon as indicated by different fiber lengths and orientations, but not a septal plane. Nonetheless, each portion exhibited distinct activity patterns (see later). 3. Structural and Functional Aspects of the Adductor Musculature In general, the adductor muscles of lepidosaurs are highly pinnate. Presumably this reflects the need to generate as much jaw closing force as possible given limited cranial space. By packaging more fibers in a given volume, pinnate muscles have a larger physiological cross section than parallel-fibered muscles and therefore achieve greater contractile force (Cans et al., 1985). However, muscle pinnation comes at the cost of excursion distance, i.e., the distance through which a muscle can shorten. Lizards typically exhibit reasonably large gape angles (see later), so excursion limitations imposed by highly pinnate muscles are potentially serious. Gronowski and Rieppel (1981) pointed out that there is usually a transition in adductor form from back to front that should mitigate this problem. In comparison to other tetrapods, the lepidosaurian adductor muscle mass is exceptionally subdivided. The functional, evolutionary basis for this subdivision is unclear. It is tempting to relate it to streptostyly and cranial kinesis, which endow the lepidosaurian skull with more degrees of freedom than most tetrapod skulls, thus requiring a highly differentiated musculature to control (and resist) its varied movements. However, this hypothesis is weakened by two facts: Sphenodon (and presumably ancestral lepidosaurs) lack streptostyly and kinesis but share the complex musculature, and most of the main adductors insert onto a single, large, basal aponeurosis (the bodenaponeurosis) that connects them to the mandible. Presumably this shared attachment would have the effect of summing their separate forces into a single resultant vector. Thus, anatomical complexity is not apparently mirrored in the complexity of action. However, despite a shared attachment site, the individuated muscles differ in fiber length, degree of pinnation, and angle of insertion onto the bodenaponeurosis. They therefore vary in the degree to which they are stretched during jaw opening and in their line of action at different gape angles. Anterior adductors are longest, have the most parallel fiber arrangement, and are stretched the most during jaw opening, whereas posterior adductors are the most pinnate, with the shortest fibers, and are
210
Kurt S c h w e n k
stretched the least during jaw opening. Thus there is a correspondence between effective excursion distance and degree of stretch during jaw opening such that sarcomeres of all muscles remain within their lengthtension optima. Thus, maintaining an effective bite across a wide range of gapes may be the most compelling explanation for adductor subdivision in lepidosaurs (Rieppel and Gronowski, 1981; Gans et al, 1985). A related factor is that because individual adductors differ in the histochemistry and contractile characteristics of their constituent fibers, different muscles, or portions of muscles, can be recruited differentially depending on gape angle and the type of contraction needed (see later; Throckmorton and Saubert, 1983; Herrel et ah, 1999a). Finally, it is noteworthy in this context that most snakes, with their highly kinetic skulls and jaws, lack a bodenaponeurosis (Haas, 1973), thus the "degrees of freedom" argument may be most applicable to them, if not to lizards. 4. Jaw Muscle
Histochemistry
Although squamate skeletal muscle fiber histochemistry has been well studied, most studies have considered locomotory muscles (e.g., Putnam et ah, 1980; Gleeson, 1983,1985; Gleeson and Harrison, 1986; Gleeson et al, 1980; Young et al, 1990; Mirwald and Perry, 1991). Jaw muscles have been histochemically characterized in only three lizard species (Throckmorton and Saubert, 1982; Herrel et al, 1999a). In general, three fiber types are typical of squamates: fast-twitch glycolytic (FG); fast-twitch oxidative or oxidative-glycolytic (FOG); and slow-twitch fibers (Gleeson and Johnston, 1987). Slow fibers are often identified as "tonic" fibers because they are hard to distinguish histochemically (Gleeson, 1983), but Gleeson and Johnston (1987) found that, in squamates, such muscles corresponded to slow twitch rather than true tonic fibers. FG and FOG are rapidly contracting fibers that vary in the degree to which they depend on aerobic respiration. FG fibers are very fast contracting, but fatigue rapidly. FOG fibers can sustain longer bouts of activity. Slow fibers are oxidative, can sustain long contractions, and are slow to fatigue, but they are also slow to contract. Glycolytic muscles are usually associated with rapid, powerful movements of short duration, whereas oxidative fibers are associated with sustained activity, support, and postural maintenance. Whole muscles vary in the percentage of fibers contributed by each fiber type and in their distribution within the muscle; in some muscles, fiber types are scattered or mixed throughout, but in others they are segregated into distinct "compartments." Table 8.4 summarizes known fiber type composition
TABLE 8.4 Histochemical Profile of Jaw Muscles in Lizards^ Muscle^
Species
MLAO
Tn
MAMES
Tn
MAMESA
Ua
MAMESP
Ps Ua
MAMEM
Tn Ps Ua
MAMEP MPsS
MPs? MPtS (lat) MPtP (med)
Ps Ua
Tonic FO/FOG
+ + + +
+
-
+
+ + + + + + /+ + + +
FG
Mixed
Comp
+ +
+
-
-
+
+ +
?
?
?
?
-
+
-
?
?
-
-
-
+
?
?
-
?
?
-
+
-
-
-
?
?
-
+
?
?
-
Tn Ps Ua
-
Tn Ps
+
-
+ + + +
Tn Ps
+ /+
+ /+
Tn Ps Ua
-
-
+
-
+ + +
+ + +
-
?
?
+ +
-
-
+ +
+
-
-
+
+ +
+
-
MPtE
Ua
MAMP
Tn Ps Ua
MDM
Tn Ps Ua
+ +
-
+
-
+
+ + +
-
~
?
-
?
+
-
?
?
^Muscles with a single fiber type are homogeneous. In muscles with more than one type, fibers are either mixed (Mixed) or grouped into discrete compartments (Comp). FG, fast twitch, glycolytic; FOG, fast twitch, oxidative glycolytic; Ps, Plecoderma (Agama) stellio (Agamidae); Tonic, tonic, oxidative (or possibly slow twitch, oxidative); Tn, Tupinambis nigropunctatus (Teiidae); Ua, Uromastix aegyptius (Agamidae). Data for Tupinambis from Throckmorton and Saubert (1982); data for both agamids are from Herrel et al. (1999a). ^Muscle names in listed order: MLAO, levator anguli oris; MAMES, adductor mandibulae externus superficialis (unspecified); MAMESA, a. m. e. superficialis anterior; MAMESP, a. m. e. superficialis posterior; MAMEM, a. m. e. medialis; MAMEP, a. m. e. profundus; MPsS, pseudotemporalis superficialis; MPsP, pseudotemporalis profundus; MPtS, pterygoideus superficialis (or lateralis); MPtP, pterygoideus profundus (or medialis); MPtE, pterygoideus externus; MAMP, adductor mandibulae posterior; MDM, depressor mandibulae.
of lizard jaw muscles. In most respects the jaw muscles are histochemically similar to other squamate muscles. All three fiber types have been identified and exist in varying proportions in different muscles (note that numbering of fiber "types 1, 2, and 3 " does not correspond between Throckmorton and Saubert (1982) and Herrel et al (1999a). However, Throckmorton and
8. Feeding in Lepidosaurs Saubert (1982) found that their "tonic'' fibers were less oxidative than in lizard locomotory muscles and that their fast-twitch oxidative fibers lacked a glycolytic capacity, so they classified them FO rather than FOG. Some muscles are composed of a single fiber type, whereas others have two or three fiber types that are either mixed or compartmentalized. Both studies found that compartmentalized muscles are divided into an inner oxidative and an outer glycolytic region. In Tupinambis the glycolytic regions comprise only FG fibers, but the oxidative compartments are a mixture of two or three fiber types (Throckmorton and Saubert, 1982). Herrel et ah (1999a) found a similar pattern in two agamid species with the exception that no more than two fiber types were ever found within a single muscle. Interestingly, no FG fibers were identified in any Uromastix muscles. A perusal of Table 8.4 reveals a high degree of variation in muscle fiber composition both among species and among muscles, rendering generalizations impossible. One notable similarity is that the pseudotemporalis superficialis comprises a single fiber type in all three species (FOG in the agamids and FG in the teiid). This muscle is also one of the most anterior adductors with long, nearly parallel fibers that insert almost perpendicularly onto the mandible in most species (see earlier discussion). The predominance of fast-twitch fibers with high glycolytic capacity in this muscle suggests its importance in generating rapid, powerful jaw closure with the mouth fully open. This muscle may be responsible for the initiation of fast close during a gape cycle (see Section V). In any case, the high degree of histochemical variation among muscles suggests the possibility of differential fiber recruitment during a feeding sequence, supporting the functional specialization hypothesis of adductor subdivision discussed earlier (Throckmorton and Saubert, 1982). E. Tongue Schwenk's (1988) study is the most phylogenetically comprehensive overview of lepidosaurian tongue morphology to date, but its focus is on phylogenetically informative characters, thus it does not review some aspects of tongue form. Comparative studies more limited in taxonomic scope include Sewertzoff (1929), Gnanamuthu (1937), de la Cerna de Esteban (1965) and Tanner and Avery (1982). A reasonably full set of references is listed in Tables 8.1 and 8.2, but key sources include Sphenodon (Schwenk, 1986); Squamata: Iguanidae (Oelrich, 1956; McDowell, 1972; Schwenk, 1988; Delheusy et al, 1994); Agamidae (Gandolfi, 1908; Sewertzoff, 1929; Gnanamuthu, 1937; Schwenk, 1988; Smith, 1988; Herrel et al, 1998c, 1999a); Chamaeleoni-
211
dae (Gnanamuthu, 1930, 1937; Lubosch, 1932; Bell, 1989); Amphisbaenia (de la Cerna de Esteban, 1959; Schwenk, 1988); Gekkota (Zavattari, 1909; Sewertzoff, 1929; Gnanamuthu, 1937; Ping, 1931; Schwenk, 1988; Schwenk and Rehorek, in preparation); Scincomorpha (Sewertzoff, 1929; Gnanamuthu, 1937; de la Cerna de Esteban, 1965; Schwenk, 1988); and Anguimorpha (Sewertzoff, 1929; Sondhi, 1958b; McDowell, 1972; Smith, 1986; Schwenk, 1988; Smith and MacKay, 1990; Tobeau et al, 1994). Due to radical variation in superficial form (Fig. 8.11), the lepidosaurian tongue has provided a historically important source of characters for systematic studies (Camp, 1923; Schwenk, 1988, and references therein). Although highly variable overall, tongue form is nonetheless conservative within higher taxa so that families tend to be relatively uniform and well differentiated one from another. This pattern of variation is significant for feeding studies because it indicates that tongue evolution is not tied to ecological radiations within groups, but rather reflects deeper-level divergences. As such, tongue anatomy is practically uncorrected with diet (Schwenk, 1988), a surprising pattern given its centrality to feeding function (some exceptions are noted later). Lingual diversity also reflects the tongue's essential role in vomeronasal chemoreception. These patterns are discussed further in Section VII. 1. Superficial Form In most taxa the tongue at rest fills the oral cavity (exceptions are the varanids and snakes). It lies within a depression in the floor of the mouth that is bordered laterally by mucosal ridges (sublingual plicae) containing the sublingual salivary glands. These converge anteriorly and the tongue tip usually rests on or between them. Sublingual plicae are extremely well developed in agamids so that the tongue appears to lie within a deep well (Schwenk, unpublished observations), but they are lacking in most gekkotans (Schwenk, 1988; Filoramo and Schwenk, 1998, in preparation). In most iguanians and gekkotans, the tongue is relatively broad and short with a wide, rounded apex and wide base that extends posteriorly on each side of the larynx as two lobes called the posterior limbs (McDowell, 1972) (Figs. 8.11d, 8.11f, and S.llg). The posterior limbs are lost in chamaeleonids, some teiids, and some varanoids, usually in conjunction with development of a lingual sheath. In many squamates the foretongue tapers significantly so that the tongue is triangular in shape. This is especially true in autarchoglossans where the foretongue is often very slender. In many scincomorphans the hindtongue is reduced as well. In
212
Kurt Schwenk
FIGURE 8.11. Superficial form of the tongue in squamates (dorsal view), (a) Xantusia (Xantusiidae), (b) Abronia (Anguidae), (c) Podarcis (Lacertidae), (d) Coleonyx (Gekkonidae), (e) Varanus (Varanidae, (f) Gonocephalus (Agamidae), (g) Crotaphytus (Iguanidae), (h) Cnemidophorus (Teiidae), (i) Cordylus (Cordylidae), and (j) Dasia (Scincidae). From Schwenk (1995b), with permission.
varanids, the tongue is extremely narrow and parallelsided (Fig. S.lle). It is withdrawn into a sheath anterior to the larynx and is superficially similar to snake tongues, but is histologically distinct (Schwenk, 1988; Smith and MacKay, 1990). In snakes the lingual sheath and overlying larynx are situated far forward in the floor of the mouth so that the tongue at rest is completely covered except, perhaps, for the very tips. The anterior position of the larynx and glottis in snakes presumably permits respiration during the sometimes slow process of consuming very large prey. In contrast, the sheath and larynx in varanids lie posteriorly in a more typically lepidosaurian position, and at rest a substantial part of the tongue is evident lying in the floor of the mouth. It is convenient to divide the tongue into anterior and posterior portions referred to as the "foretongue" and "hindtongue," respectively, due to the potentially different morphological and functional attributes of each. However, the shape and extent of the tongue's mucosal surface provide only a partial picture of its
full extent. Some lingual musculature extends posteriorly beneath the oral mucosa to the hyobranchium, and virtually all other oral landmarks, such as the position of the larynx or the insertion point of the genioglossus muscle, potentially vary among species. Therefore, foretongue and hindtongue are descriptive terms that do not necessarily imply precise homologies among taxa. To a first approximation they represent comparable regions in most species, but in taxa with highly modified tongues, such as snakes, varanids, and chamaeleonids, comparisons are most problematic. The tongue tip is highly variable in form (Schwenk, 1988,1994e, 1995). It is rounded and uncleft in Sphenodon, some chameleons (Schwenk, unpublished results), and reportedly in Dibamidae (Greer, 1985; however, unpublished data suggest a slight notch in Dibamus). In all other squamates it is variously bifurcate: in iguanians, gekkotans, and some scincomorphans it is slightly notched; in most scincomorphans it is significantly cleft, and in the lacertoids, especially teiids, it is forked. The tongue is deeply forked in amphisbaenians and angui-
8. F e e d i n g in L e p i d o s a u r s
morphans show a morphocline from modestly cleft to deeply forked (McDowell and Bogert, 1954; Schwenk, 1994e). The presence of a forked tongue is tightly correlated with the ability to follow scent trails using the vomeronasal system mediated by tongue flicking (Schwenk, 1994e). The tongue tips in most species are underlain by thickened, keratinous pads (ventral pallets) that in notch-tongued forms rest on top of the sublingual plicae (Fig. 8.12B). In most fork-tongued forms the pallets curl around to the dorsal surface and are drawn out into long tines (McDowell, 1972) (Fig. S.llh). In varanoids the ventral pallets are not evident, al-
213
though they occur in basal anguimorphans. There is often a striking, male-female correspondence between the form of the tongue tip and the mucosal surface of the palate around the vomeronasal fenestrae. The tongue is free anteriorly and along its lateral margins. The posterior limbs are sometimes slightly undercut so that they are partially free of the buccal floor as well. The foretongue is attached to the floor of the mouth by a frenulum, which contains the genioglossus muscles (Fig. 8.12B). The point of attachment of these muscles determines the extent of the free part of the tongue anteriorly. However, this is deceptive because the foretongue is nearly always capable of some hydrostatic elongation, the extent of which is determined by histological features. Thus the degree of lingual liberation from the buccal floor is only roughly correlated with protrusibility in life. 2. Surf ace Morphology
B
F I G U R E 8.12. (A) Transverse section through the foretongue of an iguanid lizard {Stenocercus sp.) showing the deep, muscular corpus of the tongue typical of iguanian lizards and its crown of long, glandular, filamentous papillae. The curved arrows indicate the dorsal transverse septum, and the small, straight arrow points to the origin of the laryngohyoid ligament dorsal to the lingual process. (B) Midsagittal section through the tongue of an agamid Hzard {Agama hispida) showing the kinked apex of the lingual process within the tongue and the extent of the laryngohyoid ligament. Note the reticular papillae and the penetration of lingual glands into the tongue musculature of the foretongue. The tongue tip on each side is underlain by a thick, lightly keratinized epithelium forming the ventral pallet. Small, straight arrows indicate the laryngohyoid ligament (which passes out of the plane of section posteriorly). The dark, longitudinal line demarcating the dorsal extent of the verticalis fibers is the dorsal transverse septum, g, genioglossus; h, hyoglossus; 1, larynx; m, mandible; p, lingual process of hyobranchium; t, transversalis; v, verticalis.
In all lepidosaurs except snakes and varanids (and the foretongue of Lanthanotus) the tongue's dorsum is papillose. Basal snakes retain some papillae along the lateral margins of the tongue (McDowell, 1972). Schwenk (1984,1988) identified several types of papillae, but the homology of similar papillary forms in different taxa is not certain. The following descriptions are based primarily on Schwenk (1988) and extensive unpublished data. Sphenodon and iguanids are characterized by long, filamentous papillae on the foretongue that are densely covered by a simple epitheliumi of columnar mucous cells, except at the tips (Figs. 8.12A and 8.13A). In general the papillae at the tongue tip are low and squat, but they rapidly become long and slender across the foretongue before shortening and thickening again toward the hindtongue and the posterior limbs. The posterior limbs often have large, conical papillae that point posteriorly. The papillary apices vary among taxa and across the tongue's surface (Fig. 8.14). Anteriorly they are covered by a stratified, squamous epithelium perforated by numerous taste buds (Schwenk, 1985). On the foretongue where they are longest, papillae are often capped by a stratified squamous epithelium, but in many iguanids, especially anoloids, sceloporines, and some tropidurines (Schwenk, 1984,1988, unpublished results), as well as some agamids (Herrel et ah, 1998c), the apex of each papilla is covered by a bizarre, pseudostratified epithelium in which the nucleated cell bodies hang free, tethered to the papillary apex by a slender stalk! Each so-called "plumose'' papilla (Rabinowitz and Tandler, 1986; "arborate" papillae of Schwenk, 1984) ends in a tuft of these free-floating cells (Figs. 8.13A and 8.14B). The distribution of plumose
Kurt Schwenk
their lengths to create a spongelike form (Fig. 8.12B). The intervening crypts are lined by a glandular epithelium and the papillary apices are most often squamous, but are dramatically plumose in anoles and some agamines. Reticular papillae are restricted to a narrow zone of the foretongue in anolines, but they often extend well onto the hindtongue in agamids. Variation in papillary height along the length of the tongue is similar to the iguanid condition. Stout, conical papillae on the posterior limbs are often particularly well developed in agamids. Uniquely in some agamids, glandular crypts penetrate deeply into the musculature of the foretongue (Gandolfi, 1908; Gnanamuthu, 1937; Smith, 1988; Schwenk, unpublished results) (Fig. 8.12B). There is taxonomic variation in this trait and in the particular muscles invaded, but too few species
F I G U R E 8.13. (A) Parasagittal section through the foretongue of an iguanid lizard {Sceloporus occidentalis), anterior to the left. Note the long, filamentous papillae crowned with plumose cells. Papillae are longest in the contact zone (cz). (B) Parasagittal section through the tongue of a gecko {Gonatodes antillensis), anterior to the right, showing the low-profile, scale-like papillae and the intrinsic origin of some hyoglossus fibers (arrows). Sublingual salivary glands are evident beneath the tongue to the right, cz, contact zone; G, genioglossus; H, hyoglossus.
papillae on the tongue corresponds to the area of prey contact during lingual ingestion (see later). Posteriorly the epithelium reverts to a stratified squamous type or sometimes an unusual, cuboidal type, as in Sphenodon (Schwenk, 1986). The lingual epithelium is never keratinized, except very lightly, in some cases, on the ventral pallets. The "typical" papilla contains collagenous connective tissue fibers that run its length, muscle fibers, and a vascular loop; however, some filamentous papillae are so slender they appear to contain nothing more than collagen fibers (although some vascular supply would seem to be necessary). Posteriorly, the papillae become increasingly inclined toward the pharynx, particularly on the posterior limbs, although this is not always evident. Agamids, chamaeleonids, and anoline iguanids possess a unique form of reticular papillae in which it appears that filamentous papillae are anastomosed along
F I G U R E 8.14. Scanning electron micrographs of the lingual surface in two iguanid lizards. (A) Papillary apices on the hindtongue of Sceloporus occidentalis. Note the relatively smooth epithelium at the ends of some papillae and the transition to plumose cells in others. Scale bar: 34 /xm. (B) Papillary apices in contact zone of the foretongue in Callisaurus draconoides showing extensive development of plumose cells. Individual papillae are difficult to discern. Anterior is to the top. Scale bar: 19 [xrcv.
8. Feeding in Lepidosaurs have been examined to deduce phylogenetic patterns. Gekkotans and xantusiids have unique, short, peglike papillae across the broad, spatulate surface of the foretongue (Figs. 8.11a and 8.lid). These are nonglandular, filled by large vascular sinuses and are very lightly keratinized. They form an exceptionally smooth pad that demarcates precisely the part of the tongue used in eye wiping (Schwenk et ah, manuscript in preparation). Posteriorly there is a sharp transition to long, glandular, filamentous papillae (although these remain quite short in diminutive gekkonids). On the posterior limbs these fuse progressively into transverse lamellae called lingual plicae. In xantusiids, the transition is from the peg-like papillae directly to lingual plicae so that virtually the entire hindtongue is plicate. Although the anterior papillae are perpendicular (i.e., normal to the tongue surface), the posterior papillae and plicae become imbricate posteriorly. Dibamids have a unique surface form of narrow transverse ridges that are slightly imbricate, but these
fMf
T^f}
M
215
are quite different from the lingual plicae of other taxa. The epithelium itself is aglandular, but deep glandular crypts penetrate into the tongue and open to the surface in the grooves between ridges, almost creating the appearance of reticular papillae. Scincomorphans and amphisbaenians never possess filamentous papillae, although in some cordylids the papillae can become quite long. Scincomorphan papillae typically have broad bases and flat apical surfaces that are imbricate posteriorly, most often taking the form of scales that are superficially similar to integumental scales (Figs. 8.15A and 8.16). On the lateral margins of the foretongue and across the hindtongue the scales often fuse into lingual plicae. The ventral surface of the foretongue in teiids is also plicate (Harris, 1985). Amphisbaenians lack plicae and the tongue is entirely scaled. Glandular epithelia are restricted to the crypts between scales and plicae in the hindtongue (nicely shown by von Seiller, 1892), but the amphisbaenian tongue is entirely aglandular, a condition found elsewhere only in snakes and varanids, which lack lingual papillae. The apical epithelium is a broad, flat expanse of stratified squamous cells that is lightly keratinized. The posterior margins of lingual scales are smooth, as in teiids (Fig. 8.16B), or scalloped, as in scincids and cordylids (and sometimes lacertids; e.g., Iwasaki and Miyata, 1985) (Fig. 8.16A). In cordylids the marginal projections are sometimes long and finger like. Anquimorphans are characterized by a bipartite, or diploglossan tongue (e.g., McDowell and Bogert, 1954; McDowell, 1972). The foretongue is often described as "retractile" within the hindtongue, but this is not accurate. Rather, the foretongue is specialized for hydrostatic elongation while the hindtongue is not (Schwenk,
B
FIGURE 8.15. Transverse sections through the foretongues of two scleroglossan species. (A) Cnemidophorus tigris showing lowprofile, scale-like papillae. Note the smooth, flat surface and the lack of a glandular epithelium between papillae. (B) Varanus indicus showing the absence of papillae and extreme reduction of the tongue. The hyoglossus muscle bundles are surrounded by a thick circular fiber system (arrows) implicated in hydrostatic elongation of the tongue. Compare to the iguanid shown in Fig. 8.12A. DL, dorsal longitudinal muscle fibers; H, hyoglossus.
FIGURE 8.16. Scanning electron micrographs of scale-like papillae in two scleroglossan species. (A) Scincella sp. (Scincidae) showing the broad, flat, smooth surface of a single papilla. Scale bar: 18 /nm. (B) Similar papillae in a teiid, Cnemidophorus tigris. Scale bar: 91 jmrn.. Anterior is to the top in both figures.
216
Kurt Schwenk
unpublished observations). As such, the foretongue is capable of considerable hydrostatic length change by the modulation of tongue diameter. Thus elongation and shortening are localized within the foretongue and there is no actual retraction, or telescoping, of the foretongue into the hindtongue. However, there is both an obvious break point in papillary form on the dorsal surface and a zone of epithelial folding laterally and ventrally to accommodate the length change, thus the foretongue appears superficially to be in a retracted state at rest. In other scleroglossan taxa, hydrostatic length change is more evenly distributed along the length of the tongue without a sharp transition zone. A consequence of the bipartite tongue form is a transition in papillary form along the length of the tongue. The foretongue is always relatively smooth, with very low-profile papillae or none at all (as in Lanthanotus and varanids). These are flat and squat anteriorly, but become progressively peg-like posteriorly. They are aglandular and have a lightly keratinized, stratified squamous epithelium. Proceeding posteriorly, they become slightly pointed and slightly imbricate. At the foretongue-hindtongue junction, there is a sudden transition to tall, filamentous, highly glandular papillae. These are typically capped by a sharply pointed, posteriorly directed epithelial apex. Posterior papillae are broad and conical. Thus, the bipartite tongue is divided into a highly extensible foretongue used exclusively for tongue flicking and vomeronasal chemoreception, and a papillose, glandular hindtongue used to support and manipulate prey items in the mouth (see Section VII,D). Although this division of labor is typical of most scleroglossans, its functional and morphological sharpness is unique among squamates. In varanids, the hindtongue is putatively lost and modified into the lingual sheath (McDowell, 1972; Schwenk, 1988). The remaining tongue is devoid of papillae and is covered by a thin, lightly keratinized epithelium (Filoramo and Schwenk, manuscript in preparation) (Figs. 8.11e and 8.15B). The histochemistry of the lingual glandular epithelium in lepidosaurs has been studied most extensively by Gabe and Saint Girons (1969) and more recently by Taib and Jarrar (1985a,b, 1986). Iwasaki (1991) and Rabinowitz and Tandler (1991) provided data on the ultrastructure of lingual glands with some indication of secretory product. In general, lingual mucocytes are serous, mucous, or bipartite (mucoserous or seromucous) (Gabe and Saint Girons, 1969). Mucous cells vary in their chemical constituents (Taib and Jarrar, 1985a,b, 1986). Only agamids and iguanids have purely serous cells on the foretongue (Gabe and Saint Girons, 1969; Rabinowitz and Tandler, 1991) and in Sphenodon they are mucoserous. Because these are all lingual feeders, it is tempting to relate the copious secretion of serous
fluids to the biomechanics of lingual prey adhesion (see later). Unfortunately, the functional attributes of the different secretory products are completely unknown. Furthermore, some scleroglossans also have mucoserous cells and the lingual feeding chamaeleonids have seromucous secretions. All we can say is that all lingual feeders have some serous secretion on the foretongue and none exhibit pure, mucus secretion here, as found commonly in scleroglossans. In most taxa the histochemical profile of secretory cells varies along the length of the tongue with the hindtongue being predominantly mucus secreting. 3. Connective Tissue
Organization
As described previously, in most taxa the muscular corpus of the tongue is penetrated to a varying extent by the lingual process, a hyaline cartilage rod continuous with the basihyal posteriorly. The anterior extent of the lingual process is greater in iguanians than in scleroglossans. In most species it is attached to the cricoid cartilage of the larynx by a laryngohyoid ligament (Oelrich, 1956; Schwenk, 1988) (Fig. 8.12). This ligament arises from near the anterior end of the lingual process in most iguanians, but farther posterior in others. In the former case it is often associated with a dorsoventral "kink" in the tip of the lingual process (Schwenk, unpublished observations) (Fig. 8.12B). The ligament runs posterodorsally through the verticalis musculature (later) to insert into the connective tissue matrix beneath the larynx in the midline. In some scleroglossans the posterior end of the ligament branches before it attaches to the larynx (Schwenk, 1988). Beneath its mucosal surface, the tongue is covered by a thin "tunic" of connective tissue fibers that appear to be entirely collagenous. There is so far no evidence for organized fiber winding (e.g., cross-helical arrays) around the tongue as is often found in muscular hydrostats, but this has not been examined explicitly. Smith (1986) implied helical winding of superficial fibers in Varanus (where one might most expect it), but did not provide supporting data. In a mammal (opossum), fibers of the lingual tunic form a "feltwork" that is highly elastic (Schwenk et ah, manuscript in preparation), and presumably this is the case in most lepidosaurs as well. The viscoelastic nature of the tunic allows stretching during elongation and shape change and may help restore the tongue to its resting conformation by elastic recoil. Various septae partition the tongue musculature internally. In Sphenodon the central core of the tongue is occupied by the verticalis musculature and is divided by a median septum (Schwenk, 1986). This is separated into dorsal and ventral parts where the lingual process penetrates the tongue. In all squamates there is partial
8. Feeding in Lepidosaurs or complete loss of this septum. It is complete in the posteriormost portion of the verticalis in some agamids (Gnanamuthu, 1937; Schwenk, 1988; Smith, 1988) and nearly complete in Varanus (Smith, 1986). In most squamates, however, the verticalis is undivided and its fibers cross the midline. There is usually a tiny, dorsal component of the median septum above the verticalis that serves as the point of origin for the transversalis fibers. In Sphenodon and most squamates, a dorsal transverse septum runs across the top of the tongue above the hyoglossus bundles and the verticalis (Fig. 8.12A). Longitudinalis and transversalis fibers run above it. In most squamates this septum is very short and, together with the dorsal bit of median septum noted earlier, it forms little more than an inverted "T.'' The transverse septum is absent in gekkotans, anguimorphans, some scincomorphans, and Dibamus. A ventral transverse septum is evident in Sphenodon, but not in squamates. 4.
Musculature
All lepidosaurs possess a highly mobile, muscular tongue, a trait they share with mammals and a few other tetrapods (Chapter 2). The functional complexity of lingual movement and shape change is underlain by extreme complexity in muscle fiber architecture. Recall that the tongue is essentially a solid mass of muscle without bones and joints to direct its movements. Some of these movements are affected extrinsically by muscles that literally pull the tongue forward or backward, but some movement, especially shape and length change, is generated intrinsically by means of a muscular hydrostatic mechanism (Kier and Smith, 1985; Smith and Kier, 1989; see Chapter 2). Muscular hydrostats usually have complex muscle architecture, the function of which is not necessarily intuitive given a traditional notion of musculoskeletal function. In lepidosaur tongues, muscle fiber architecture is sometimes so complex it defies precise description and authors sometimes disagree in their accounts. Fiber systems often interlace in intricate ways so that it is difficult or impossible to characterize discrete, individual muscles within the tongue. Thus most published descriptions of tongue muscular anatomy represent rather egregious simplifications. Tetrapod tongue muscles are classified as either ''extrinsic" or "intrinsic." The former are said to originate outside the tongue and to insert within it, whereas the latter are said to lie completely within the tongue. Several authors have suggested that this distinction is not always clear within the lepidosaurian tongue (e.g., Sondhi, 1958a,b; Smith, 1988). Schwenk (2000a) showed that nominal extrinsic muscles often have a large intrinsic component and that nominal intrinsic
217
muscles sometimes have an extrinsic component. He suggested that it was more accurate to talk about "fiber systems" within the tongue than discrete muscles. The complexity of fiber systems within the tongue manifests its functional nature as a muscular hydrostat and suggests that the tongue is best regarded as a "functional unit" not easily atomized into separate parts (Schwenk, 2000b). Nonetheless, for organizational clarity and historical continuity, I maintain the traditional distinction between extrinsic and intrinsic muscles here, but qualify muscle descriptions with more accurate accounts of their constituent fibers where known and appropriate. a. Extrinsic Muscles There are two, paired extrinsic muscles in lepidosaurian tongues. The genioglossus runs posteriorly from the mandible near the symphysis to insert on the ventral and lateral sides of the tongue (Figs. 8.12 and 8.13B). It is regarded as a protractor of the tongue (but see later). The genioglossus lies within the "frenulum" that attaches the tongue to the floor of the mouth. It is usually divided into two portions: a genioglossus medialis lies near the midline and inserts into the ventral side of the tongue. Its anteriormost fibers usually curl sharply anterior at the point of entry and run some distance forward along the ventrolateral surface of the tongue. In Sphenodon they reach nearly to the tongue tip and are undoubtedly an important tongue retractor (Schwenk, 1986). A genioglossus lateralis inserts into the tongue farther posterior. Its fibers run posterodorsally, forming the lateral wall of the hindtongue. In scincids and cordylids, some fibers of the lateralis also turn anteriorly and run into the foretongue lateral to the hyoglossus bundle (Schwenk, 1988). In scincids these fibers form a discrete bundle, but in cordylids they are diffuse. Fibers of the lateralis sometimes insert into the floor of the mouth as well (e.g., Schwenk, 1986; Bell, 1989). A genioglossus internus is found in agamids (Smith, 1988) and anoline iguanids (Schwenk and E. Williams, unpublished results). It lies medially and inserts into the tongue anteriorly along the ventral midline between the hyoglossus bundles. The genioglossus varies among taxa in its posterior point of attachment (e.g.. Smith, 1984). Generally, the farther posterior it inserts the greater the amount of foretongue extension possible. Scleroglossans usually exhibit relatively more posterior genioglossus attachment than iguanians due to their heavier reliance on extensive tongue protrusion for chemosensory tongue flicking. However, in the myrmecophagous iguanian genus Phrynosoma, the genioglossus lateralis remains detached from the side of the tongue until it reaches the posterior limbs (Schwenk, 1994c; in preparation). The hindtongue is therefore narrow in comparison
218
Kurt Schwenk
to related taxa, and genioglossus lateralis muscles form large ridges alongside the tongue until they join with it posteriorly. Presumably this is an adaptation for increasing tongue protrusion distance for lingual feeding—Phrynosoma has one of the longest relative protrusion distances measured in an iguanian during ingestion (Schwenk and Throckmorton, 1989). In Varanus, the genioglossus medialis attaches far back on the tongue at its origin on the hyobranchium (Gnanamuthu, 1937; McDowell, 1972; Smith, 1986). The genioglossus lateralis is reduced to a tiny slip that inserts into the lingual sheath. In chameleons, medial fibers of the genioglossus insert into the buccal floor ventral and lateral to the tongue (medialis?) and small lateral slips insert into the mucosa overlying the back of the tongue (lateralis?) (Gnanamuthu, 1930; Bell, 1989), which, in effect, constitutes a lingual sheath. The hyoglossus muscle arises as a broad sheet along the first ceratobranchial of the hyobranchium and runs forward into the corpus of the tongue alongside the lingual process (Figs. 8.12A, 8.13B, and 8.15). It is the principal tongue retractor. Within the tongue the paired hyoglossus muscles form two cylindrical or subcylindrical columns of mostly longitudinal fibers running nearly to the tongue tip. They constitute the bulk of the tongue's volume. For most of their extent within the tongue they are separated medially by the lingual process and the intervening fibers of the verticalis system (below) (Figs. 8.12A and 8.15). Each bundle is usually very deep, occupying most of the tongue's vertical height, although in some taxa there is a reasonably thick layer of intrinsic musculature dorsally between it and the base of the papillae. Sondhi (1958a,b) suggested that hyoglossus fibers in Varanus turned and twisted within the tongue to form the various intrinsic muscles, although Smith (1986) did not agree. She found that hyoglossus fibers are oblique rather than longitudinal and suggested that they form a spiral along the bundle. Schwenk (2000a) showed that in most squamate taxa each hyoglossus bundle is clearly subdivided into two or more zones that differ in fiber orientation and that the pattern of this subdivision changes along the length of the tongue. Subdivision is least evident anteriorly. Sagittal sections reveal that, minimally, the lizard hyoglossus comprises an extrinsic component that arises from the hyobranchium and an intrinsic component that arises within the hindtongue from the dorsal transverse septum (Fig. 8.13B). Intrinsic fibers loop anteroventrally, becoming confluent with extrinsic fibers anteriorly as they curve anterodorsally into the foretongue (Schwenk, 2000a). Where they join the fibers are essentially longitudinal, but posteriorly their orientation is variable, accounting for the internal subdivision.
Gekkotans share a striking pattern of hyoglossus form in which the paired bundles split once or several times to produce multiple, longitudinal bundles in the foretongue (Schwenk, 1988). These are visible externally as longitudinal ridges along the tongue's ventral surface and presumably are related to hydrostatic control of the broad, spatulate foretongue during eyewiping behavior (Schwenk et al, manuscript in preparation; see discussion of lingual papillae). Limited subdivision of the hyoglossus bundles occurs in some anguids as well (Edgeworth, 1935; Smith and MacKay, 1990). In Dibamus (Schwenk, 1988) and two of three species of amphisbaenian studied {Amphishaena, Anops; de la Cerna de Esteban, 1959), there is a medial bundle of the hyoglossus that runs along with the lingual process {hyoglossus medialis and perientoglossus, respectively). There are slight differences in the anatomy of these muscles so their homology is uncertain (Schwenk, 1988). In chameleons the hyoglossus is greatly elongated to accommodate the travel of the tongue during projection and at rest it is kinked into accordion-like folds. It attaches to the lateral surface of the accelerator muscle (see later), but a histologically distinct portion extends anteriorly to insert beneath the glandular surface, which is folded and invaginated at rest [Lubosch, 1932; Bell, 1989; Schwenk, unpublished results; Gnanamuthu (1930) and Bell (1989) considered the latter part a separate muscle, the longitudinalis linguae]. In the foretongue of some agamids, the hyoglossus muscle is perforated by glandular crypts that penetrate from the surface (see earlier discussion). Although not a lingual muscle strictly speaking, the mandibulohyoideus is closely allied with the tongue physically and developmentally. It forms longitudinal bundles or sheets beneath and lateral to the genioglossus muscles. As noted earlier in the description of hyobranchial musculature, in most squamates it comprises multiple, sheet-like slips that arise along the mandibular rami and run posteromedially to the hyobranchium. However, it is also true that there is usually a particularly well-developed medial pair of bundles with longitudinal fibers that arise from the mandible near the symphysis and therefore most closely approximate the mammalian notion of a true geniohyoideus muscle. h. Intrinsic Muscles The verticalis fiber system lies in the midline of the tongue surrounding the lingual process of the hyobranchium and completely separating the hyoglossus bundles in squamates (Figs. 8.12 and 8.15). Verticalis fibers arise from the lingual tunic at the floor of the tongue and insert dorsally into the transverse septum. In Sphenodon, however, the hyoglossus bundles are
219
8. Feeding in Lepidosaurs quite triangular in transverse section and throughout most of the tongue their upper corners meet in the midline above the verticalis, separating it from the transverse septum (Schwenk, 1986). Verticalis fibers insert, instead, on the median septum. This is also the case in the "ring" portion of the verticalis in some agamids, which retains a complete median septum. In gekkotans and anguids with multiple hyoglossus bundles in the foretongue, there is no discrete verticalis. Rather, vertical fibers run from the dorsum and form a loop or sling around each hyoglossus bundle. In most squamates, verticalis fibers take a curving path from the ventral surface on one side, around the lingual process and vertically on the opposite side. Thus fibers from both sides cross and interlace. In some (agamine) agamids, the posterior end of the verticalis is hypertrophied into a muscular sleeve that completely surrounds the lingual process. These fibers run circumferentially from ventral median septum to dorsal median septum, forming a sphincter-like arrangement, hence this portion of the verticalis has been called the "ring muscle" (e.g., Gandolfi, 1908; Smith, 1988). In chameleons the verticalis is extremely modified into a muscular tube that surrounds the entire length of the robust lingual process at rest. It provides the propulsive thrust for ballistic tongue projection (see later) and thus is called the accelerator muscle. Much has been made of the agamine ring muscle's putative similarity to the chamaeleonid accelerator muscle. However, this similarity is purely superficial: the striking feature of the chameleon accelerator is that its fibers are radial rather than circumferential, i.e., it is not a sphincter. They take a curved path from inner to outer membrane and fiber layers alternate in handedness (e.g., Gans, 1967; Bell, 1989; see van Leeuwen, 1997). In fact, the chameleon accelerator is far more similar in this regard to the iguanid condition in which verticalis fibers are radial and curve in opposite directions around the lingual process (Schwenk, unpublished results) (Fig. 8.12A). Other than expansion of verticalis mass ventral to the lingual process to create a tubular form, there is no detailed similarity between agamid and chamaeleonid conditions. Transversalis fibers run from the midline laterally across the dorsum of the tongue (Fig. 8.12A). Some fibers terminate in the dorsolateral margins of the tongue and lingual papillae, but in most taxa a ventral group of fibers loops ventrally along the perimeter of the hyoglossus bundle. In Sphenodon these insert into the ventral transverse septum, but in squamates they join verticalis fibers at the base of the tongue to form a continuous ring around the hyoglossus. In many iguanians this ring is not complete throughout the length of the tongue and, when present, it can be very thin
(e.g.. Smith, 1988). However, in scleroglossans it is often a thick band of muscle. The confluence of transversalis and verticalis fibers around the hyoglossus bundle, particularly in those taxa lacking a transverse septum, has led to the description of this fiber system as a circular muscle (e.g.. Smith, 1984, 1986). Oelrich (1956) described a ventral transverse muscle, but this is simply part of the circular system. In most lepidosaurs the longitudinalis forms a layer of longitudinal fibers running the length of the tongue beneath its dorsal surface, but it is best developed in the foretongue (Fig. 8.15A). In some taxa it is diffuse, but in others it comprises a series of discrete bundles. These are well developed in the foretongues of gekkotans and exceptionally so in the varanoid, Lanthanotus (Schwenk, unpublished results). Some longitudinalis fibers insert into the papillae where in transverse section they are easily mistaken as intrinsic papillary fibers (e.g., de la Cerna de Esteban, 1959, 1965). The sling-like fibers of the verticalis/circular muscle in some scleroglossans also extend into the papillae, as do transversalis fibers. It remains possible, however, that there are some vertical, intrinsic papillary fibers distinct from other fiber systems. 5. Relationship
to the Hyobranchial
Apparatus
The question of tongue mobility relative to the hyobranchium is relevant to our interpretation of hyolingual function during feeding. It is generally assumed that the lingual process is surrounded by a fluid-filled, synovial-type chamber that permits the tongue to slide freely along its length (e.g.. Smith, 1984,1988; Schwenk and Bell, 1988; Herrel et ah, 1995, 1998c). This is certainly true for chameleons in which the tongue is projected off the lingual process, but its relative degree of development among squamates has not been assessed. In fact, anatomical evidence suggests that significant lingual translation along the lingual process might not be possible in many squamates, particularly iguanians (Schwenk, unpublished observations). First, along most or part of the length of the lingual process there is no evidence for a fluid-filled cavity. This is especially evident posteriorly (Fig. 8.12B). The base of the tongue near to the basihyal is tightly adherent to the lingual process in some, if not most, lepidosaurs. There is no anatomical space between the tongue and the base of the lingual process and collagenous fibers of the lingual muscle endomycium are continuous with the perichondrium of the basihyal and lingual process here. Anteriorly, some space usually develops, but it is not clear that it is always fluid filled. Often connective tissue fibers cross between the tongue to the process, suggesting that the intervening space is an artifact of
220
Kurt Schwenk
shrinkage. Second, the intimate connection between the laryngohyoid ligament and the tongue would seem to prevent sliding of the tongue past the ligament (Fig. 8.12B). Thus, for the tongue to slide anteriorly on the lingual process it would either have to stretch the ligament significantly or pull the larynx forward along with it. Our films of Sphenodon show clearly that the larynx advances and retracts in lockstep with the tongue (Schwenk et ah, manuscript in preparation; see also Fig. 4 in Gorniak et ah, 1982). We have also observed this in Pogona, an agamid lizard (Schwenk and Throckmorton, unpublished results). Delheusy and Bels (1992) and Delheusy et ah (1994) speculated that the ligament is stretched during tongue protrusion and that its elastic recoil helps to retract the tongue, but this is not supported by the kinematics of the tongue (see Section V) or our film observations. It is unknown how extensible the laryngohyoid ligament is. Preliminary histological and histochemical analysis indicates that it is composed entirely of longitudinally arranged collagenous fibers, suggesting that it is virtually inextensible (Schwenk, unpublished results). Finally, in many iguanians there is a pronounced, hook-like flexure at the tip of the lingual process where the laryngohyoid ligament attaches, as noted earlier (Fig. 8.12B). Although the foretongue curls around the end of the process during lingual prey capture (see Section V), the arrangement would not seem to permit significant translation of the tongue along the process. V. FEEDING FUNCTION A. Overview of Feeding Nonophidian lepidosaurs can be characterized as hyolingual feeders because they manipulate and swallow food primarily by means of patterned movements of the tongue and hyobranchial apparatus. Of course the jaws are an essential part of the feeding apparatus as well, particularly during prey capture and processing. The participation of each morphological unit during a complete feeding sequence varies among taxa. Only in monitor lizards (Varanus) is the tongue largely excluded from participation in feeding, thus ' they are inertial, rather than hyolingual, feeders (see later; Bramble and Wake, 1985). In the following discussion the word "prey" is used for convenience to describe all food items, including plant material. Specific, food-related differences are noted where appropriate. Also note that references to 'Targe" or "massive" food items are always based on food size relative to lepidosaur size and not on absolute size. A selected bibliography of studies providing functional data on feeding in different clades is given in Table 8.5.
All prey capture in lepidosaurs involves the tongue or jaws, thus a food item must be approached very closely before feeding can commence. Chameleons are an exception and indeed the projectile tongue is an adaptation to circumvent this very problem. In most lepidosaurs, close approach is possible because the food item (a) is immobile (e.g., a plant or neonate animal); (b) is initially hidden, then uncovered by an active forager; (c) is in the open and overtaken by a short, rapid dash; (d) is in the open and is approached slowly and stealthily; or (e) unwittingly approaches an immobile lepidosaur waiting in ambush. Once in proximity, food items may be assessed for appropriateness (see Section III,C). This is accomplished visually and chemically with tongue flicking and, in experienced individuals, probably involves learned recognition of appropriate food types. Once a food item is deemed acceptable, a feeding sequence is initiated. There is likely to be strong selection on prey capture performance in taxa feeding on active prey because an initial miss usually results in loss of the prey item. With very few exceptions, ectotherraic lepidosaurs are physiologically unequipped for lengthy pursuits and rely, instead, on quick, glycolytically fueled strikes and rapid processing (e.g., Pough and Andrews, 1985; Andrews and Bertram, 1997). During a feeding sequence, most species are, themselves, vulnerable to predation, thus there is, presumably, also pressure to reduce handling time. Complete feeding bouts, from capture to swallowing, are typically very brief, on the order of several seconds to half a minute. There is usually a refractory period following each completed feeding bout during which accessible prey will not be taken, but this is modulated according to satiety and can be brief in hungry individuals (personal observation). A complete feeding sequence comprises a remarkably complex series of events that in most species occurs very rapidly. A prey item must be located and captured. If it poses a threat or might escape, it is killed or incapacited. If it cannot be swallowed immediately it must be reduced or processed until it can be. It must then be moved through the oral cavity and into the throat to pass on to the gut for digestion. This series of events is orchestrated by a series of cyclical, patterned movements involving primarily the jaws, tongue, and hyobranchial apparatus. In order to make sense of the behavior it is necessary to partition it into discrete stages corresponding to the different functional tasks required of the feeding apparatus from start to finish. This facilitates analysis and desciption and makes comparisons among species possible. The discrete functional events that occur during a single feeding event are identified in the following feeding stages: prey capture, ingestion, processing, intraoral transport, pharyngeal
221
8. Feeding in Lepidosaurs TABLE 8.5
Selected References Treating Feeding Function and Cranial Kinesis in Lepidosaurs
Exclusive of Snakes and Chameleons'^ Scleroglossa Sphenodon
Iguania
Gekkota
Scincomorpha
Anguimorpha
Cranial kinesis
Gorniak et al (1982) Frazzetta (1983) Ostrom (1962) Versluys (1912, 1936)
Arnold (1998) Borsuk-Bialynicka (1985) Frazzetta (1962) Hallermann (1992) Herrel and De Vree (1999) lordansky (1966,1973,1990a, 1996) Rieppel (1978d) Throckmorton and Clarke (1981) Versluys (1912,1936)
Arnold (1998) Borsuk-Bialynicka (1985) Frazzetta (1962) lordansky (1966, 1996) Patchell and Shine (1986a) Versluys (1912, 1936)
Arnold (1998) Borsuk-Bialynicka (1985) Bradley (1903) De Vree and Cans (1987) Frazzetta (1962) lordansky (1966,1990a, 1996) Kritzinger (1946) MacLean (1974) Rieppel (1978d) van Pletzen (1946) Versluys (1912,1936)
Arnold (1998) Boltt and Ewer (1964) Borsuk-Bialynicka (1985) Condon (1987) Frazzetta (1962,1983) lordansky (1966,1996) Rieppel (1978a,d, 1979a) Smith (1980) Smith and Highlander (1985) Versluys (1912,1936)
Feeding function
Bramble and Wake (1985) Gorniak effl/. (1982)
Bels (1990a) Bels and Baltus (1988) Bels and Goosse (1989,1990) Bels et al. (1994) Bramble and Wake (1985) Delheusy and Bels (1992) Delheusyeffl/. (1994) Kraklau (1991) Herrel and De Vree (1999a) Herrel et al. (1995,1996a, 1997a, 1998a,b) Lappin (1999) Schwenk and Throckmorton (1989) Smith (1984) Throckmorton (1976,1978,1980) Throckmorton and Clarke (1981)
Bels et al. (1994) Bramble and Wake (1985) Delheusy et al. (1995) Patchell and Shine (1986a,c)
Bels and Goosse (1990) Bels et al. (1994) Bramble and Wake (1985) Dalrymple (1979) Elias et al. (2000) Cans et al. (1985) Cans and De Vree (1986) Goosse and Bels (1992) Herrel et al. (1996b, 1998a,b, 1999b) MacLean (1974) McBrayer and White (manuscript) Smith (1984) Smith et al. (1999) Urbani and Bels (1995)
Bels et al. (1994) Bramble and Wake (1985) Condon (1987) Elias et al. (2000) Frazzetta (1983) Herrel et al. (1997b) Smith (1982,1986)
''See Chapter 9 and Table 8.3, respectively. Studies listed here are primarily functional analyses of living animals; however, some significant morphological studies with functional inferences are also included, particularly for cranial kinesis. The primary Hterature is heavily emphasized, but important reviews or secondary sources are included (in bold). The taxonomic content of columns follows Table 8.1.
packing, and pharyngeal emptying. These are reviewed here and in Chapter 2. Not every stage occurs in every species or in every feeding bout, nor is each stage necessarily discrete—some are combined and others overlap. Fiowever, they represent a heuristically important starting point because each stage has a different functional outcome and therefore makes different mechanical demands on the structural elements of the system. Thus, to elucidate the functional bases of the extensive morphological diversity reviewed earlier, it is necessary to consider the varying role of the structural elements in each feeding stage (Schwenk and Throckmorton, 1989). The term "ingestion" is sometimes used in the literature generically to refer to the entire process of feeding.
from start to finish, with "swallowing" often used in a similar way in reference to snakes. In the context of tetrapod feeding, because these terms have specific, technical meanings as defined later and in Chapter 2, such usage is undesirable. Feeding stages are characterized by kinematic patterns of jaw, tongue, and hyobranchial movement (e.g.. Bramble and Wake, 1985). The cyclical nature of these movements is consistent with the notion of motor control by a central pattern generator (CPG) (see Chapters 2 and 13). I note anecdotally that the refractory period following a feeding sequence, mentioned earlier, gives the impression of a neural "switch" or transition period during which the CPG-controlled behavior of feeding is turned off and a more alert or conscious state
222
Kurt Schwenk SOI
SO 'i"'|
1
' "1
soil
FO
FC SC
»'i'"
44 degrees
• ^^ • ••
<
LU Q.
< CO
1 • 1 •
[•
^
A^ \ ••^•••» '* i 1 •* 1 1 •••: : •••* • •
•
1 1 I
ingestion
',
• *•
• ^
^^^^^m^i^^^mm^.
open
•• • • • • • « • ,• >•«••••• • ,•* ,
•• , • • •
• ••
w»
••-
y**
^•*
• close
transport
protrude
,.—•••—T^».
••••-"^
•' >^*••-
LU
=)
CD
Z
O
; ; I
rvpv nontari —__ K^Jt iia\~ti ""^•~—~
1 fjtvj
full contact —
III
0.0
1
0.1
Ill
0.2
',
i i 1 1
> ••
—»
^ ^!^'
; ; —r^j • iiiii
0.3
i
0.4
EXTRA-ORAL
—« ««'**^_
^w-~
• *• •• «•
•
INTRA-ORAL
t
^
retract III
0.5
iiii
0.6
0.7
0.8
0.9
1
1
1.0
I.Hill
1.1
1.2
1.3
1.4
^
1.5
TIME (SEC) F I G U R E 8.17. Jaw and tongue kinematics during lingual ingestion and intraoral transport in Sphenodon punctatus. Solid vertical lines demarcate the phases of each gape cycle. A discrete SO II is only evident in the transport cycle. Note that the tongue makes contact with the prey item over an extended period of time as the prey is pushed into the papillary surface. As the prey was pinned to the substrate, it pushed the mandible down, as indicated by a momentary increase in gape angle (at second dotted line). A stationary phase intervenes between gape cycles. Based on high-speed (300 fps) 16-mm film (Schwenk, Frazzetta and Jenkins, in preparation). FC, fastclose; FO, fastopen; SC-PS, slow close-power stroke; SO, slowopen.
is initiated to control the more plastic behavior preceding a strike. However, CPG control of cyclicity does not imply immutability of the resulting kinematic patterns, which are found to vary among species, among stages, and within a stage according to food type and state of reduction. As such, there is circumstantial evidence for the modulation of feeding kinematics based on sensory feedback (see Chapter 2). During each feeding stage the jaws are repeatedly opened and closed. A single open-close sequence is known as a gape cycle and each feeding stage comprises one to many gape cycles. A covcvpXeie feeding sequence, or hout, includes several feeding stages, from ingestion to swallowing. A "model gape cycle" was proposed by Bramble and Wake (1985) based largely on kinematic patterns evident in lepidosaurs and other nonmammalian tetrapods (see Fig. 2.14 and discussion in Chapter 2). Its generality remains in dispute, but it at least provides a starting point for comparing gape cycles among species and among feeding stages. A single
gape cycle potentially comprises several phases determined by changes in the velocity of the jaws as they open and close (Fig. 8.17): slow open I (SO I), slow open II (SO II), fast open (FO), fast close (FC) and slow close (SC) or slow close power stroke (SC-PS). There is sometimes an intervening period between cycles known as a stationary phase. Gape cycles are reviewed in Chapter 2 and further discussed in Section VII,C.
B.
Feeding Stages
1. Prey Capture and Ingestion Prey capture is the apprehension and subjugation of a prey item and ingestion is its movement from the environment into the oral cavity (Chapter 2). In nearly all nonophidian lepidosaurs, prey capture (or prehension) and ingestion are accomplished by the mouth and combined into a single stage so that capture
8. Feeding in Lepidosaurs involves both the apprehension of a prey item and its immediate delivery into the mouth. Rare exceptions include the Komodo monitor {V. komodoensis) and, occasionally, amphisbaenians, whose unusual feeding behaviors are discussed in Section VI. Once captured, small prey are either transported and swallowed immediately or are killed by biting, shaking, or crushing against the substrate. In some venomous snakes, subjugation is further decoupled from prey capture. Prey are bitten, envenomated, and released to be located later after they are dead. Only then does ingestion begin. Snakes differ from other lepidosaurs in most aspects of their feeding and the traditional feeding stages used here do not clearly apply (see Chapter 9). A fundamental dichotomy is evident in the pattern
F I G U R E 8.18. Consecutive cine frames (48 fps) of a small agamid {Phrynocephalus helioscopus) during lingual ingestion. Note the conformation of the tongue during protrusion and the initial bite following retraction of the mealworm into the mouth.
223
of ingestion of small food items (see Sections IV,E and VII,A): Sphenodon and all iguanians use the tongue as a prehensile organ to apprehend small prey and draw it into the mouth, whereas all scleroglossans use jaws and teeth (Schwenk, 1988; Schwenk and Throckmorton, 1989; Wagner and Schwenk, 1999) (Fig. 2.16 in Chapter 2). As prey size increases the dichotomy fades and jaw prehension is almost universally employed, although taxonomic differences remain. A few scleroglossan species use a type of lingual prehension of small prey items in certain circumstances. These exceptions are discussed later. Lingual ingestion involves protraction and protrusion of the tongue concomitant with hyobranchial protraction, with the tongue tip ventrally curled so that the dorsal, papillary surface of the tongue is presented toward the prey item (Figs. 8.18 and 8.19). Tongueprey contact usually occurs on the anterior third of the tongue where papillae are longest in iguanians (Fig. 8.13A). One advantage of lingual prehension is that the tongue is protruded at the same time the head advances toward the prey item, thus their approach velocities are summed and prehension is more rapid than would be possible with the jaws alone. Tongue-prey contact can be relatively light, but most often involves a forceful impact that pushes the prey item against the substrate, fitting the tongue to the prey surface (see later). Adhesion is remarkably effective and involves a combination of interlocking and wet adhesion; in chameleons it may also involve suction (Schwenk, 1983; see Section VI,A). Retraction occurs almost
F I G U R E 8.19. Lingual prehension in an agamid Hzard (Pogona barbata) based on a 35-mm photograph. Note how the dorsal, papillose surface is curled around the end of the protruded, muscular tongue. The right anterior process of the basihyal is evident posteriorly, bulging as the tongue is protracted. The larynx moves forward with the tongue, presumably pulled by the laryngohyoid ligament. Sublingual plicae containing salivary glands form the sides of a well in which the tongue sits at rest.
224
Kurt Schwenk
FIGURE 8.20. Jaw prehension in a scleroglossan lizard (Varanus niloticus). Prey capture is with the tips of the jaws (more pronounced with smaller prey). Note complete retraction of the tongue. The apparatus on the lizard's head is a goniometer designed to measure mesokinetic flexion and the arrow in the upper left indicates the position of the gage (see Fig. 8.25). From Condon (1987), Exp. Biol 47, 73-87, © Springer-Verlag, with permission.
immediately following prey contact and is extremely rapid. The prey item comes to lie within the mouth on the surface of the tongue or between the tooth rows for transport or swallowing. Jaw prehension of prey involves grasping of the prey in the jaws with the anterior teeth (Fig. 8.20). Sometimes the side of the mouth is used, but grasping usually occurs with the jaw tips. Prey are moved into the mouth for further processing by the tongue or an inertial toss. Larger prey are usually pinned or dragged against the substrate or are subdued by violent shaking before subsequent feeding stages commence. 2. Processing Processing refers to the mechanical reduction or modification of food within the mouth before it is swallowed (Chapter 2). If a prey item is very small it is often transported and swallowed directly without processing. Conversely, large prey consumed by inertial feed-
ers (e.g., Varanus) are often oriented in the mouth and swallowed whole, without processing (although usually after subjugation or killing by biting or suffocating). However, in most cases some processing follows immediately upon ingestion. After ingestion a food item often comes to lie across a tooth row where it can be bitten immediately (Fig. 8.18), but sometimes it must be positioned between the teeth or shifted to a more advantageous bite point. Such manipulative cycles are usually mediated by lingual movement, including tongue twisting, but they are sometimes accomplished inertially with a lateral jerk of the head. Repositioning is also used to orient large prey items head first so that limbs, scales, or fur fold posteriorly, facilitating prey movement into the esophagus. Reduction usually involves repeated bites so that the teeth (or occasionally the palate) pierce, crush, or shear the prey item as it is coated in saliva, rendering it softened but usually whole. Most lepidosaurs are well supplied with oral salivary glands that produce a variety of secretory products (Gabe and Saint Girons, 1969; Kochva, 1978), but there is no indication thus far that these compounds initiate chemical digestion or do anything more than lubricate the bolus for swallowing (with the exception of venom glands in Heloderma, which are used primarily for defense). Such chewing behavior is called puncture-crushing in contrast to mammalian-type mastication in which a food item is reduced to tiny particles. As such, comminution of food is not a necessary outcome of most lepidosaurian chewing, although it sometimes occurs, whereas it is the raison d'etre of most mammalian mastication. This difference has a profound effect on the form and evolution of the feeding system in the two groups. Transverse movements of the mandible relative to the upper jaw and a "masticatory orbit" (e.g., Hiiemae et al, 1978; Chapter 13) are virtually never observed in nonophidian lepidosaurs and remain unique features of the mammalian masticatory mechanism (e.g., Davis, 1961; Throckmorton, 1980). Despite the absence of a masticatory orbit as in mammals, it is an unsung fact that many, if not most, lepidosaurs chew on one side at a time and often alternate between sides. Side switching is particularly evident in Sphenodon (Gorniak et al, 1982; Schwenk et al, manuscript in preparation) and iguanians, especially acrodonts (Schwenk and Throckmorton, 1989; personal observation), but it is also observed in scleroglossans (e.g.. Smith, 1984). Chewing asymmetry and the intraoral manipulation it requires have been regarded as exclusively mammalian traits associated with the presence of mobile, muscular cheeks and lips, and the absence of a lingual process within the tongue (e.g., Davis, 1961), thus its ubiquitous occurrence in
8. Feeding in Lepidosaurs lepidosaurs has significance for interpreting the evolution of mammalian mastication (see Throckmorton, 1976; Crompton, 1989). Chewing asymmetry also has implications for muscle activity patterns on active versus balancing sides, mandibular bending, the transmission of forces from one side to the other, the nature of joint reaction forces, and the position of the bite point (e.g., Druzinsky and Greaves, 1979; Gorniak et ah, 1982; Greaves, 1995). These factors have only rarely been considered in lepidosaur feeding. Chewing and intraoral transport cycles tend to be kinematically similar and are often not separated in functional analyses. The basis for such similarity is discussed in the following section. 3. Intraoral
Transport
Intraoral transport is the movement of food through the oral cavity to the pharynx for swallowing (Chapter 2). In many ways, intraoral transport is the "purest'' form of gape cycle in which the cyclical, coordinated movements of the jaws, tongue and hyobranchium are most apparent, unsullied by the mechanistic vagaries of prehension and reduction. Possibly for these reasons. Bramble and Wake (1985) based their model gape cycle on intraoral transport and suggested that it might represent an ancestral tetrapod pattern (see Chapter 2). However, Bramble and Wake (1985) also included "chewing'' or processing cycles here as well. In the vast majority of lepidosaurs, intraoral transport is accomplished by cyclical movements of the tongue and hyobranchial apparatus in coordination with the jaws (Bramble and Wake, 1985). Hyolingual transport involves anterior movement of the tongue and hyobranchial apparatus underneath the food item during SO I while it is held against the roof of the mouth. The tongue is usually deformed at the end this process (SO II) so that the bolus comes to lie within a depression in its surface, or the tongue is elevated in front of the bolus. Although the prey item is often oriented transversely across the tooth row in chewing cycles, in pure transport cycles it is oriented longitudinally on the tongue to clear the teeth and corners of the mouth (e.g., Delheusy and Bels, 1992). SO II may also be used to integrate sensory information on food position and condition to modulate the next cycle (Bramble and Wake, 1985). As the jaws are parted during FO, the bolus is freed from contact with the palate or teeth, and the hyobranchium and tongue rapidly shift posteriorly so that at jaw closure (FC and SC) the prey item has come to lie farther back in the mouth where it is once again fixed against the palate by the tongue. The tongue repeats its anterior movement beneath the prey item in preparation for the next cycle. According to
225
Bramble and Wake (1985), the presence of SC, especially with high levels of adductor activity indicating a power stroke (SC-PS), is contingent on the teeth engaging the prey item at the end of FC. They thus describe a chewing cycle as construed here, rather than a "pure" transport cycle. Indeed, an SC phase is variably present during putative transport cycles (see later). However, an SC phase is sometimes apparent even when contact of the marginal teeth with the prey item does not occur. This may represent controlled slowing of jaw closure or contact of the prey item with the palate as the jaws engage. If the latter, the presence of an SC phase during intraoral transport should be related to bolus size and condition, implying that SC should be more evident with large prey items and in earlier cycles. In some taxa, head jerking and prey inertia are substituted for hyolingual movement (Gans, 1969b). During inertial transport the prey item is momentarily released while the animal rapidly shifts its head anteriorly. The prey item thus comes to lie farther back in the mouth as the jaws close. For smaller prey, head and jaw movements are imparted to the food item itself so that it moves back relative to the ground, whereas a larger prey item remains more or less stationary while the head moves forward over it. Among lepidosaurs, only varanids are obligate inertial feeders due to extreme reduction of the tongue for chemoreception and its limited participation in feeding (e.g.. Smith, 1986; Condon, 1987; Elias et al, 2000). Other large, carnivorous lizards (e.g., Tupinambis) frequently employ inertial feeding as well (e.g., MacLean, 1974; Smith, 1984; McBrayer and White, manuscript in preparation), but these also use hyolingual transport some or most of the time (e.g., MacLean, 1974; personal observation). Inertial transport is occasionally used by other taxa, particularly when feeding on relatively large prey items (e.g., Sphenodon, see later), but small species, such as gymnopthalmids, eating relatively small prey also use it (e.g., MacLean, 1974). It is possible that the frequency of inertial feeding by lepidosaurs has been overestimated in the literature. This is because virtually all functional analyses have been undertaken in the laboratory with captive animals often fed unnatural diets (see Chapter 1). Mice are frequently used in feeding trials with larger lepidosaur species, but such vertebrate prey are rarely eaten in the wild (see Section III,A). The vast majority of prey items for such species are small invertebrates that are more likely to be manipulated with hyolingual movements. Intraoral transport and chewing cycles often overlap temporally and are difficult to distinguish kinematically (Bels et al, 1994; see later). In most taxa, they grade one into the other so that many gape cycles are a combination of the two. A pure chewing cycle would
226
Kurt Schwenk
involve no posterior movement of the prey item during the gape cycle and a pure transport cycle would involve no crushing of the prey by the teeth. However, in most cases, chewing cycles include hyolingual movement of the prey item to place it between the teeth for crushing at the end of each cycle. There is cyclical anteroposterior movement, but no net posterior translation, or some posterior movement of the prey item along the tooth row occurs between bites. Thus jaw and hyolingual movements during chewing cycles are likely to mimic transport cycles kinematically in most taxa. For these reasons, Schwenk and Throckmorton (1989) followed Bramble and Wake in suggesting that processing should be considered a type, or subset, of intraoral transport, a conclusion in which Bels et al (1994) concurred. However, functional work has shown that there may be important quantitative (e.g., Herrel 1997a; Smith, 1982; Cans et al, 1985; Cans and De Vree, 1986) and qualitative (e.g.. Smith, 1982,1984; Condon, 1987; So et al, 1992; Herrel et al, 1997b) differences between the two cycle types (see later). Given that chewing is a facultative behavior and that chewing and transport have different mechanical outcomes and can be kinematically decoupled (as in chameleons), it seems worth preserving the distinction between chewing and intraoral transport stages here, and in future functional analyses, insofar as possible. 4. Pharyngeal Emptying
(Swallowing)
Swallowing refers to the movement of food into the esophagus, where peristalsis takes over the process of transporting the bolus to the gut (see Chapter 2). In lepidosaurs, swallowing potentially comprises two separate substages: pharyngeal packing and pharyngeal compression. Although these stages serve the same function as deglutition in mammals, swallowing is so different mechanistically in the two groups that it is unlikely to be homologous in any meaningful sense (see Smith, 1992; Bramble and Wake, 1985; see Chapters 2 and 13). Pharyngeal packing occurs universally among nonophidian lepidosaurs, but not pharyngeal compression, which is most common in scleroglossans. Occurrence of the latter stage in some species may be facultative, presumably depending on bolus size and characteristics (personal observation). It is often difficult to observe, sometimes only evident as a single, brief compressive cycle without an accompanying gape cycle. Because of confusion about the mechanics and nomenclature of swallowing in lepidosaurs, it is usually not clear whether the failure to mention pharyngeal compression in literature accounts of swallowing reflects its actual absence or merely the failure to observe or document it. To complicate matters further.
terminal intraoral transport cycles sometimes grade into pharyngeal packing cycles so that kinematic distinctions between transport and swallowing stages can be blurred during the transition (see later). In part because of these ambiguities, the literature on lepidosaurian swallowing is inconsistent and sometimes muddled. Some authors do not consider pharyngeal compression at all, note it briefly in passing, or consider it "aberrant" (Herrel et al, 1997b:387). Bramble and Wake (1985), who set the tone for nearly all lepidosaurian feeding studies to date, considered pharyngeal compression to be exceptional and limited to specialized inertial feeders, such as varanids and large teiids. In some studies, the discussion of pharyngeal packing is included under intraoral transport, and a separate stage known as "cleaning" is identified, although cleaning is actually a part of pharyngeal packing (see later). The literature is also inconsistent in its application of technical terms. For example. Smith (1984) equated "swallowing" with pharyngeal compression and considered pharyngeal packing a separate, preswallowing stage, whereas Herrel et al (1999b) equated "swallowing" with pharyngeal packing and considered pharyngeal compression a postswallowing stage! Some studies (e.g., Urbani and Bels, 1995) erroneously apply the term "deglutition" to swallowing behavior in lizards, usually in the context of pharyngeal packing [see Smith (1994) and Chapter 2]. These problems make extraction of data for a comparative synthesis difficult. I have proposed a standard nomenclature here that should facilitate comparative studies. Future functional analyses should take care to watch for pharyngeal compression following pharyngeal packing and to distinguish these two components of swallowing when they occur. a. Pharyngeal Packing Pharyngeal packing is characteristic of lepidosaurs and possibly some turtles (Chapter 2). During this stage the bolus is pushed into the pharynx and the anterior part of the esophagus (Smith, 1984, 1992; Bramble and Wake, 1989). This is almost always accomplished by the tongue and may be the principal function of the tongue's posterior limbs, which serve to tamp the bolus into the pharynx. Typically (but not always) during pharyngeal packing cycles the tongue is protruded out of the mouth tip first and often appears to "lick" the labial scales and snout. A "cleaning" or "lip-licking" function is often attributed to this behavior (e.g., Throckmorton, 1980; Cans et al, 1985; Bels and Baltus, 1988; Goosse and Bels, 1992), but in most cases such tongue protrusion results from lingual positioning for asymmetrical tamping by the posterior limbs
227
8. Feeding in Lepidosaurs and probably has nothing to do with cleaning the labial scales (Smith, 1984; personal observation). The extent of tongue protrusion varies depending on the size of the bolus and its position relative to the esophagus. Packing cycles with tongue protrusion push the bolus farther back into the pharynx than cycles in which protrusion does not occur (Delheusy and Bels, 1992). Although tip-first tongue protrusion clearly distinguishes some pharyngeal packing cycles from transport cycles, early packing cycles can blend temporally and kinematically with preceding intraoral transport cycles. Both transport and packing involve hyolingual movement of the bolus posteriorly, but they differ in the position of the bolus relative to the tongue; in the former it lies on the tongue, whereas in the latter it is posterior to it. Other kinematic variables distinguish them as well, but the two phases may be largely overlapping in quantitative attributes, at least in some taxa (e.g., Herrel et al, 1995). Smith (1984), however, found the cycle types to be "distinct" in their patterns of tongue-jaw coordination and the shapes of the orbits described by tongue and hyobranchium. In packing, the hindtongue moves upward and backward during SO rather than upward and forward, and the tongue orbits become elongate and largely anteroposterior. Gape profiles become small and spiked, losing their differentiation into discrete phases. The extent to which these disagreements reflect phylogenetic differences, experimental conditions, or criteria for identifying cycle types in the first place remains to be determined. The extent to which pharyngeal packing moves food into the esophagus must be variable. Smith (1984) found that during pharyngeal packing, food collected in the pharynx only and pharyngeal compression was necessary to squeeze the bolus into the esophagus (swallowing sensu stricto). In contrast, Herrel et ah (1996a) observed that packing moved food into the esophagus as well. Because pharyngeal compression often does not occur, pharyngeal packing must be sufficient in many cases to move food fully into the esophagus to initiate peristaltic transport. Undoubtedly, bolus size and condition influence swallowing behavior, as well as taxonomic differences.
tive in taxa with posterior limbs of the tongue reduced or missing (see Section VII,B). Lepidosaurs and all other nonmammalian vertebrates lack the pharyngeal musculature characteristic of mammals (Smith, 1992; see Chapters 2 and 13) and therefore cannot compress the pharyngeal cavity internally to squeeze the bolus into the esophagus. Thus, if pharyngeal packing fails to tamp food far enough into the esophagus to initiate peristalsis, lepidosaurs resort to external compression of the pharynx by means of cervical flexure and/or contraction of the constrictor colli muscle. The constrictor colli forms a sling around the pharynx and its contraction elevates the hyobranchium and constricts the gular region. Smith (1984) noted that in order for compression to succeed, the bolus must lie behind the basihyal so that it is squeezed posteriorly and not returned into the mouth. It is conceivable that anterior and posterior intermandibularis muscles join the constrictor colli during swallowing so that their serial contraction creates a compressive wave along the pharynx from front to back, squeezing it like a tube of toothpaste, but this is purely speculative. If the bolus is posterior enough, simple constriction of the pharynx would suffice. Neck flexure also is used to compress the pharynx in some species, alone or in addition to constrictor colli constriction. Its pattern and extent vary among taxa, but it can involve ventral bending (head tucking), lateral bending, or both. In short-necked forms the bending appears to be simple flexure at the atlanto-occipital joint, but in elongate forms, lateral neck bending is sinuous and sinusoidal, i.e., it appears that a propagated wave is generated rather than simple flexure (personal observation). Sinusoidal movements continue into the trunk in some species. It is possible that internal concertina movements also aid in pharyngeal emptying in long-necked forms, especially anguimorphans, as they do in some snakes (Kley and Brainerd, 1996; N. Kley, personal communication). Such movements involve sinusoidal bending of the body axis internally, including vertebral column and esophagus, independent of the outer body wall (see suggestive Xray photo of Varanus, Fig. 89 in Greer, 1989).
h. Pharyngeal Compression Accumulating evidence indicates that pharyngeal compression is widespread among lepidosaurs (and other reptiles) and is not restricted to inertial feeders, as suggested by Bramble and Wake (1985). It occurs in both iguanian and scleroglossan squamates and although undescribed, it may be present in Sphenodon as well. It is more common or pronounced in scleroglossans, which accords well with their generally reduced hindtongues—pharyngeal packing may be less effec-
C. Feeding in
Sphenodon
1. Ingestion Lingual ingestion in tuatara was noted anecdotally by earlier authors (e.g., BuUer, 1878; Dawbin, 1962; Farlow, 1975) and has been analyzed by Gorniak et al (1982) and Schwenk et al. (manuscript in preparation). Gorniak et al. (1982) found that lingual prey prehension was used when feeding on small prey (insects).
228
Kurt Schwenk
but that large prey items (mice) were grasped directly by the jaws. Ingestion is usually initiated by prey movement (see earlier discussion). Movement elicits head cocking and visual fixation of the prey item. The head is moved toward the prey, usually forward and downward toward the substrate, but prey suspended by a thread above the head are also easily taken (Schwenk et ah, manuscript in preparation). As the head moves closer, the jaws begin to part as the tongue is protracted. Cineradiography shows that tongue protraction is caused by hyobranchial protraction (Gorniak et ah, 1982; Schwenk et al, manuscript in preparation). There is no evidence of hydrostatic elongation or sliding of the tongue along the lingual process. The tongue is slightly elevated initially and extended tip first. The initial tip-first orientation of the tongue differs from iguanians (see later). Immediately upon crossing the mandibular symphysis the tongue tip curls ventrally and is fixed to the mandible so that as the tongue continues to be pushed out of the mouth by the hyobranchium it becomes arched with its dorsal surface presented anteriorly toward the prey item. This kinematic pattern is consistent with the fact that anterior genioglossus medialis fibers run into the foretongue to the tongue tip; contraction of these fibers, or simply their inextensibility, would force the tongue tip downward and anchor it to the mandible (Schwenk, 1986; see Section V,F). As the tongue is protruded the larynx advances correspondingly (Fig. 4 in Gorniak et ah, 1982). This suggests that the laryngohyoid ligament is inelastic and when the tongue is protracted it pulls the larynx with it (see earlier discussion). Prey contact occurs on the dorsal surface of the foretongue, which is now directed anteroventrally. The zone of contact corresponds to the region of the tongue where filamentous papillae are longest (Schwenk, 1986). Differences in papillary height and orientation across the width of the tongue create a modest lingual sulcus in the midline. The sulcus may enhance the number of papillae making contact with the prey item. The prey item is usually only lightly contacted by the tongue surface, but this is sufficient to cause adhesion. In some cases, however, contact is forceful enough to push the mandible slightly downward (Fig. 8.17). Tongue-prey contact triggers tongue retraction, but the jaws continue to open as the tongue and adherent prey are rapidly withdrawn into the mouth. As tongue and prey clear the mandible, the jaws are snapped close. Tongue retraction is coupled to hyobranchial retraction. Ingestion of mice is accomplished by biting and lifting with the anterior teeth. According to Gorniak et ah (1982:337), "The tongue plays no role during capture nor is it protruded." Unfortunately, Gorniak and col-
leagues do not indicate the nature of tongue movement and conformation within the mouth during jaw prehension. It is possible either that the tongue is fully retracted to avoid prey contact, as in scleroglossans, or that it curls and contacts the prey item within the mouth concomitant with jaw prehension, as in iguanians (Schwenk and Throckmorton, 1989). Feeding studies on intermediate prey sizes are necessary. These would help determine whether the transition from tongue to jaw prehension is a graded response, as in iguanians, or a threshold response. Lingual ingestion gape cycles are highly variable in duration and kinematic pattern in a single individual (Schwenk et ah, manuscript in preparation). Most cycles roughly conform to the Bramble-Wake model and are similar to the iguanian pattern (Fig. 8.17). The tongue and hyobranchium are protracted during SO I and SO II, then retracted during FO and FC. SO II is characterized by a very high gape angle, perhaps more so than in iguanians (Schwenk and Throckmorton, 1989; see later), presumably to allow clearance for the tongue and prey as they are retracted. Thus the difference in gape between SO II and maximum gape at the end of FO is usually quite small. Herrel et ah (1995) suggested that Sphenodon exhibits no distinct SO, but this is not accurate. There is a tendency for SO I and SO II to be poorly differentiated and for SO II to blur into FO so that a distinct "plateau phase" (Schwenk and Throckmorton, 1989) during SO II is less frequently evident in Sphenodon than in iguanians. Nevertheless, an SO phase is evident in many ingestion cycles (Schwenk et ah, manuscript in preparation). A typical ingestion sequence is slow relative to most squamates (0.5 to 1.0 sec); however, tuatara are relatively large and this may be an effect of body size. Although most investigators dutifully report quantitative data on cycle duration and make comparisons among taxa, scaling effects on tetrapod-feeding kinematics are virtually unstudied. Absolute differences in cycle duration among taxa may have little biological significance. Gorniak et ah (1982) provided electromyographic (EMG) data for feeding in Sphenodon, but unfortunately did not sample hyolingual muscles. During ingestion the jaws are opened by the bilateral activity of anterior and posterior portions of the depressor mandibulae; however, the anterior part is most active earlier and the posterior part later. Both are silenced as jaw closing begins and the adductors become active, although there is some overlap. The anterior portion of the superficial external adductor is first active, followed by its posterior portion, the medialis, the pseudotemporalis superficialis, and several parts of the complex pterygoideus. Adductor activity is maximal when the teeth contact
229
8. Feeding in Lepidosaurs the prey item and crush it. They are then silenced during a stationary phase when the mouth opens slightly, probably from recoil of the prey item between the teeth. 2. Processing Tuatara use a unique form of mandibular translation to shear their food rather than the typical puncture-crushing of lizards (Gorniak et al, 1982). After two or three manipulative or killing bites, an insect is moved by the tongue to the corner of the mouth on one side where it is chewed for two to four cycles. The prey item remains fixed relative to the upper tooth row during chewing. It is then moved by the tongue and placed between upper and lower teeth of the same or opposite side and the chewing cycles are repeated. Multiple chewing clusters are repeated (approximately 15 times) until the insect is "reduced to a saliva-covered, mushy mass" (Gorniak et al, 1982:333). Each reduction cycle includes a period of mandibular translation along the articular surface of the fixed quadrate so that the lower teeth slide between the two parallel, upper tooth rows (Fig. 8.2) (see Section IV,B), shearing the prey as well as crushing it. The gape profile is typical with all phases usually evident. However, the jaws usually continue to open during SO II so that there is usually no distinct "plateau phase." SO involves simple mandibular depression. During FO, however, the mandible slides posteriorly along the quadrate-articular joint. The jaws close during FC, but there is no mandibular translation. The prey item is crushed during SC-PS as the jaws are pulled into full closure. At the end of SC-PS, a distinct "shearing phase" occurs during which the mandible is translated anteriorly. Larger prey items are slightly rolled during translation. Shearing ends when the mandible is slightly depressed at the start of a stationary or resting phase. Typically, a resting or stationary phase is present between each chewing cycle. These vary in time, but can be as long as a second. After three to five cycles of chewing, insects are shifted laterally to the opposite tooth row with twisting movements of the tongue. As chewing progresses, cycle durations decrease. Chewing cycles blend into intraoral transport. When feeding on mice, Sphenodon follows ingestion with several crushing or killing bites. Inertial transport is used to position the mouse transversely across the tooth rows of both sides at the back of the mouth for the symmetrical killing bites. Lateral jerks of the head are then used to position the mouse inertially between the teeth on one side and a series of four to seven chewing cycles begin to reduce the mouth. Initially, side
shifts and prey manipulation are inertially based, but as chewing proceeds, the tongue becomes involved in manipulatory movements and reduction cycles resemble those for insects. This probably represents a transition from processing to intraoral transport (see later). If limbs or other parts of the mouse come to lie outside the tooth rows following ingestion and manipulation, these are bitten off and not consumed. EMG recordings (Gorniak et al, 1982) showed that initial jaw closure during chewing is driven by the adductor mandibulae externus superficialis, externus profundus, and pseudotemporalis superficialis. Crushing during SC is caused by the addition of the pseudotemporalis and the pterygoideus. The shearing phase is driven by a portion of the pterygoideus whose fibers insert on the mandible with a large anterior component so that they act as a protractor, as well as an adductor, of the mandible. Chewing is highly asymmetric in Sphenodon with frequent side switching. Working and balancing side muscle activity varied from bite to bite during reduction, apparently in response to prey position and texture. 3. Intraoral
Transport
It is difficult to separate chewing from transport cycles in Sphenodon as both occur simultaneously. As reduction proceeds, the bolus is moved farther and farther back in the mouth. Chewing cycles then grade into pure intraoral transport cycles that Gorniak et al (1982) referred to as "terminal movements." However, because these cycles seem to grade quickly into pharyngeal packing cycles, the distinctions are difficult to make, at least based on the descriptions of Gorniak et al (1982). Further analysis of our cineradiographic films may help clarify differences among cycle types (Schwenk et al, manuscript in preparation). Transport is hyolingual with high amplitude anteroposterior movements of the hyoid (Schwenk et al, manuscript in preparation; see Fig. 11.23 in Chapter 11). Transport gape profiles are typical, with welldefined phases (Fig. 8.17). The bolus is released from the marginal teeth and held centrally on the tongue. During SO the tongue and hyobranchium move slowly forward, then rapidly so during FO. At FC the tongue and hyoid reverse direction and the tongue's dorsal surface is arched so that it scrapes the palate as it is rapidly retracted, pushing the bolus behind it toward the pharynx. Palate scraping does not occur with mice, presumably because they did not tend to adhere to the roof of the mouth due to their greater mass. Gape angles tend to be larger during transport cycles than during reduction, and mandibular translation is minimal (Gorniak et al, 1982). Muscle activity levels are
230
Kurt Schwenk
reduced during transport; however, EMG patterns are similar to reduction cycles with the exception of the pterygoideus muscle. The pterygoideus exhibits a great deal of regional heterogeneity, with some parts active during jaw opening. 4.
Swallowing
a. Pharyngeal Packing As noted earlier, Gorniak et al (1982) did not distinguish between late intraoral transport cycles and pharyngeal packing. However, during "terminal movements" the tongue tip is often slightly protruded between the anterior teeth prior to retraction. Gorniak ei al. (1982) interpreted this behavior as a method of "cleaning" the anterior portion of the tongue, but more likely it reflects the larger anteroposterior excursions typical of pharyngeal packing cycles in other lepidosaurs (see later). It seems to be a way of positioning the bolus as far back on the tongue as possible so that the posterior limbs of the tongue can be used to tamp the bolus into the pharynx. h. Pharyngeal Compression Unfortunately, Gorniak et al. (1982) did not describe compression cycles in Sphenodon, nor do our films continue long enough into the feeding sequence to include this stage (Schwenk et al, manuscript in preparation). Gorniak et al. (1982) implied that swallowing is accomplished by the "terminal movements" described earlier, interpreted here as intraoral transport and pharyngeal packing cycles based on comparison to other lepidosaurs (see later). Our cineradiographic films indicate that the bolus may not clear the pharynx during these cycles, thus it is possible that tuatara use pharyngeal compression and possibly cervical flexure to push the bolus fully into the esophagus to complete swallowing, but this remains undetermined. Knowledge of the presence or absence of a pharyngeal compression stage in Sphenodon would help clarify the evolution of swallowing mechanisms in lepidosaurs. D . Feeding in Iguania 1. Ingestion Lingual ingestion in chameleons is treated separately (see Section VI), but is mentioned briefly here in comparison to other iguanians. This section focuses on the more generalized iguanian families, Iguanidae and Agamidae. Lingual ingestion in these taxa was noted anecdotally in a number of studies (e.g., Abel, 1952; Cooper et al, 1970; lordansky, 1973; Smith, 1984; Frazzetta, 1986), but its universality in Iguania was pointed out by Schwenk (1988) and Schwenk and Throckmor-
ton (1989), who showed that lingual ingestion of small prey uniquely characterizes this monophyletic taxon in contrast to its jaw-feeding sister taxon, Scleroglossa. Functional treatments or complete descriptions of lingual ingestion are available for the following taxa: Iguanidae, Anolis (Bels and Baltus, 1989; Bels, 1990a; Bels and Goosse, 1990); Dipsosaurus (Schwenk and Throckmorton, 1989); Iguana (Throckmorton, 1976; Schwenk and Throckmorton, 1989); Opiums (Delheusy and Bels, 1992); Phrynosoma (Schwenk and Throckmorton, 1989); Sauromalus (Schwenk and Throckmorton, 1989); Agamidae, Agama (including Plocederma) (Kraklau, 1991; Herrel et al, 1995); Phrynocephalus (Schwenk and Bell, 1988; Schwenk and Throckmorton, 1989); Pogona (Schwenk and Throckmorton, 1989); and Uromastix (Throckmorton, 1976; Schwenk and Throckmorton, 1989; Herrel and De Vree, 1999a). As noted earlier, lingual prehension is limited to relatively small food items, whereas larger prey are grasped in the jaws (Smith, 1984; Schwenk and Throckmorton, 1989), as in Sphenodon. Nonetheless, only one study has provided any functional data for iguanian jaw prehension (Bels and Goosse, 1990). Bels and Goosse (1990) asserted that the small iguanid, Anolis carolinensis, used jaw prehension for blowfly larvae, but they did not document the behavior photographically. It is possible that 1-cm larvae were too large for lingual prehension by the small lizards, but it is also possible that poor film resolution prevented its detection. Some lingual feeding sequences involve minimal tongue protrusion and close approximation of the jaws before tongue-prey contact (particularly when prey are relatively large; see later). In small taxa, such as A. carolinensis, this is very difficult to see, even in good films (personal observation). Furthermore, studies of other Anolis species (Bels and Baltus, 1989; Bels, 1990), as well as my own observations of A. carolinensis (unpublished), confirm the use of lingual prehension. In any case, the gape profile given for this sequence was typical of lingual ingestion sequences in other taxa. The ambiguity of this case only highlights the dearth of functional data for jaw prehension in iguanians. Cropping of plants by herbivorous species may be a form of jaw prehension in some cases, but it virtually always involves initial or simultaneous lingual prehension (Schwenk and Throckmorton, 1989). This is described further later. Jaw prehension may occur in other unusual circumstances. For example, streamside basilisk lizards (Basiliscus) have been observed to capture fish under water after a lunge from their perch (Echelle and Echelle, 1972). Although underwater lingual prehension does occur in some salamanders (Schwenk and Wake, 1988, 1993; S. Deban and J. Larsen, personal communication), it is more likely that the
8. F e e d i n g in L e p i d o s a u r s
basilisks use their jaws for prehension given the mechanics of lingual prehension (see later). Montanucci (1989) observed that some individuals of Phrynosoma solare, an extreme ant specialist, eventually "learned" to use the jaws to capture unnaturally large prey, but only after repeated attempts with the tongue failed. Schwenk and Throckmorton (1989) suggested that prey size-dependent differences in iguanian prehension mode are quantitative and not qualitative. As such, tongue-prey contact always occurs, but the distance the tongue is protracted and protruded is modulated according to prey size. As prey size increases, the tongue is protruded less and less until at large prey sizes tongue-prey contact occurs within the margins of the jaws at the same time as jaw-prey contact. This scenario is supported by Throckmorton's (1976) observations of Uromastix and my own observations of Gambelia (Iguanidae) feeding on lizards (unpublished results). It also suggests the way jaw prehension may have evolved in Scleroglossa from lingual-feeding ancestors (Wagner and Schwenk, 2000; see Section VII,A). Due to the lack of functional data on jaw prehension in iguanians, the following account is limited to lingual ingestion (references given earlier). The kinematics of lingual ingestion are very similar to those described for Sphenodon, with a few exceptions. Most lizards are more likely to charge active prey than are tuatara. Prey motion typically alerts the lizard and triggers monocular fixation of the prey so that the head is often first tilted to one side. A feeding sequence begins with orientation of the snout toward the prey item (and presumably binocular fixation in most species) and approach. If the prey item is already within range, the strike usually involves forward rotation of the body over the forelimbs with the head moving down and forward toward the prey. Sometimes tongue flicking and further observation precede the start of feeding, but often it is initiated immediately. As such, the jaws may begin to part and the tongue protruded as the lizard approaches, but before it is within striking range (e.g., Kraklau, 1991). If the prey item ceases motion or moves away, the lizard may remain with its tongue partially protruded for some time before the feeding attempt is completed or aborted; sometimes the tongue moves in and out while the prey moves in and out of range (Schwenk and Throckmorton, 1989). Variation in this phase of a feeding sequence is responsible for extensive variance in ingestion cycle duration. As protraction begins, the tongue tip is immediately curled ventrally and the dorsal surface arched (Fig. 8.18). This is in contrast to Sphenodon in which the tongue is pointed during the earliest stages of protraction. As it passes the mandibular symphysis, the
231
tongue tip is more or less anchored, and as protrusion continues the arched dorsal surface of the tongue extends outside of the mouth and is presented anteroventrally (Figs. 8.18 and 8.19). Schwenk and Throckmorton (1989) found a taxonomic difference between iguanids and agamids in the orientation of the tongue tip during protrusion, but additional study has not supported a consistent dichtomy. Tongue-prey contact occurs at maximum tongue extension on the anterior third of the tongue's dorsal surface (Figs. 8.19 and 8.21). The contact zone corresponds to the region of greatest papillary length and epithelial rugosity (Figs. 8.13A and 8.14B). As contact is made the lizard usually continues to advance so that the prey item is pinned to the substrate. This behavior not only immobilizes the prey, but importantly, it forces the prey item into the deep, papillary cushion of the foretongue so that the lingual surface is literally "formed" to the prey item (Figs. 8.21A and 8.21B). This maximizes the surface area of contact and may help promote adhesion. It might further help to absorb impact energy so that the prey item is not pushed away, giving the tongue time to form an adhesive bond. The mechanism of lingual adhesion and prehension is developed in detail in Section V,F. Following upon tongue-prey contact, the jaws open rapidly, largely by means of cranial elevation at the atlanto-occipital joint, and the tongue is retracted. As soon as the tongue and adherent prey cross the tip of
FIGURE 8.21. Lingual prey capture in a horned lizard, Phrynosoma cornutum (Iguanidae). (A and B) A small cricket is first hit with the tongue and then pinned to the substrate. The cricket's body is pushed into the papillary cushion of the tongue's contact zone, maximizing the surface area of contact and ensuring interlocking of the tongue with the prey surface. (C and D) Sequential frames of retraction in another feeding bout. Note how the tongue's papillary surface is straightened and rolled around the end of the tongue like a conveyor belt. The cricket is "flipped" rapidly into the mouth on the tongue's dorsal surface. Both sequences filmed at 250 fps (4 msec between frames).
232
Kurt Schwenk
the mandible the jaws snap closed. High-speed films of Phrynosoma reveal that, during tongue withdrawal, its dorsal papillary surface is quickly retracted by "rolling" around the stiff, protracted column of lingual muscle like a conveyor-belt (Schwenk, manuscript in preparation). The prey item is "flipped" around the end of the tongue and into the mouth, its retraction accelerated by summing the independent velocities of the papillary surface and whole-tongue retraction (Fig. 8.21C and 8.21D). Bell (1989) showed that the papillary surface is moved independently around the muscular column of the accelerator muscle in chameleons, and films of other iguanians suggest a similar pattern during retraction, hence this may be a general, iguanian trait. In herbivorous species that feed by biting pieces off of whole plants (e.g.. Iguana and Uromastix), lingual ingestion succeeds initially only in getting a part of the plant into the mouth. Once there it must be cropped by the teeth (Throckmorton, 1976, 1978; Schwenk and Throckmorton, 1989). Ingestion may require several cycles to get a sufficient portion of the plant into the mouth before it is cropped (Throckmorton, 1976,1978). Multiple ingestion cycles are unique to this situation. However, one might more accurately regard cycles subsequent to the first as part of intraoral transport, and cropping as a type of reduction. In some cases, a whole plant that is fixed in place can be regarded as a "large" food item that necessitates jaw prehension for cropping. For example, the Galapagos land iguana {Conolophus subcristatus) uses both lingual and jaw prehension when feeding on cactus pads. Pads are usually scraped clean of spines and tough outer cuticle first, then bitten (see Section V,D,2). Often the tongue draws the pad in for cropping by the teeth, but sometimes it is bitten directly (H. L. Snell, personal communication). Either way, initial cropping may not completely separate a piece from the cactus. In this case the exposed cactus flesh is scraped off the tougher, outer cuticle by the teeth. The related marine iguana {Amblyrhynchus cristatus) may use jaw prehension predominantly. It feeds underwater or in the intertidal zone by grasping algae in its jaws and cropping it with a twist and jerk of the head (Carpenter, 1966). This cropping action may be enhanced by an extremely foreshortened facial skeleton and procumbent (protruding), spatulate teeth (personal observation). However, it is possible that the tongue initially contacts the algae to draw it into the mouth for cropping (see later). Because feeding frequently occurs under water, lingual prehension is probably minimally important. Nonetheless, semidomesticated individuals were said to feed terrestrially on crickets (K. Angermeyer, in Carpenter, 1966) and it seems likely that lingual protrusion would be used then, but this is unknown.
Four possible mechanisms of tongue protrusion have been proposed (Smith, 1984, 1988; Schwenk and Bell, 1988; Schwenk and Throckmorton, 1989): (1) The tongue is "pushed" out of the mouth by protraction of the hyobranchium, primarily through action of the mandibulohyoideus (geniohyoideus) muscles. (2) The tongue is "pulled" anteriorly by the genioglossus muscle. This would either slide the tongue along the lingual process or pull the hyobranchium simultaneously due to a tongue-hyobranchial linkage and inextensibility of the tongue. (3) Intrinsic verticalis contraction reduces the diameter of the central lumen around the lingual process, exerting pressure on the incompressible fluid there and causing the tongue to slide forward on the tapered process. (4) The tongue is lengthened hydrostatically by a reduction in its diameter. Clearly, these mechanisms are not mutually exclusive and they may combine to produce various tongue movements. Furthermore, they may vary in their importance along the length of the tongue. For example, the hindtongue may be more tightly coupled to hyobranchial movement than the foretongue, which might use hydrostatic lengthening to a greater extent (see Section VII,D). It is reasonably well established that lapping and tongue flicking involve primarily hydrostatic elongation of the foretongue in all squamates and that lingual movements during ingestion, intraoral transport, and swallowing in iguanians are coupled to hyobranchial movement, but the precise mechanism of tongue protrusion during lingual ingestion is problematic. The anatomy of the tongue-hyobranchial connection suggests that independent movement of the tongue should be limited in most iguanians (see earlier discussion). Schwenk and Throckmorton (1989) presented circumstantial cinegraphic data for Pogona (Agamidae) suggesting that tongue protrusion is coupled to hyoid protraction (model 1), as it is in Sphenodon and during intraoral transport in iguanians. In support of this, Herrel et al. (1995) showed that the mandibulohyoideus muscle is active during lingual protrusion in Agama and that the sternohyoideus is active during retraction, suggesting that tongue movement is coupled to hyobranchial movement. Films of lingual feeding in the iguanid Phrynosoma also indicate coupled tongue and hyobranchial movement during retraction (Schwenk, manuscript in preparation). However, Herrel et al. (1995) also found that the genioglossus and posterior verticalis ("ring" muscle) are active during lingual protraction in Agama with peak activities at maximum protrusion, thus indicating models 2 and 3. Together, data for Agama suggest that tongue protrusion occurs through a combination of models 1,2, and 3. Thus, current data suggest a combination of niechanisms for lingual protrusion during ingestion in iguanians, but so
233
8. F e e d i n g in L e p i d o s a u r s
few species have been adequately studied that a great deal of functional diversity remains possible. Tongue protrusion occurs during SO (Figs. 8.22 and 8.23). SO II typically occurs at higher gape angles than in subsequent feeding stages, probably to accommodate the protruded tongue (Schwenk and Throckmorton, 1989). The difference in gape angle between SO II and maximum gape at the end of FO is correspondingly
r
f
1
smaller than in other cycles. Herrel et al. (1996a) showed that ingestion in Agama is quantitatively distinct from transport and "swallowing" (pharyngeal packing) cycles, differing in timing, duration, and maximal excursion of jaws and tongue; however, Schwenk and Throckmorton (1989), Kraklau (1991), and Delheusy and Bels (1992) found significant overlap in the kinematics of ingestion and transport or chewing cycles
•!
i
INGESTION open ant
I
close post
•••
Gape
•••* 1
AAA^^
[.••
• .•*
•
A
Tongue
•
• •
• 1
depress Hyoid i
$
[..••
1
•
•••• '
1 1
1
•
•
• •• • •
1
1
1
1
1
1
•
*
t
1 •
•• •
• •
..
f
[.«• r
.»** A*"
11
A
A1
AAA ••
•
% A
A
A
A
•• • • ' j
elev
1
• ••
up
..••*
Head-Y down
I
TRANSPORT 2
TRANSPORT 1
• • •
*
•
,*••**
•*
1• • ' •
1 1
••
•••••••••••,^
•• >•
••
.••' . . . . . • •
•••• ••
post
••
I
•• \
•
Head-X
•
ant
•
• •
1*****1
•
•
•
•
• ••
• •
1
•
•
•
11
1
1
1
1
1
1
100
1
1
1
1L
1
•
1
•
<
• ••
• i
1
1
*
1
1
4
•
•
••
A
1
1
1
200
300
1
1I
400
MILLISECONDS F I G U R E 8.22. Kinematics of lingual ingestion and transport in an iguanid lizard, Phrynosoma cornutum. Solid vertical lines indicate maximum and minimum gape angles, and the dotted vertical line indicates moment of tongue-prey contact. Note variation in the gape profiles of the two intraoral transport cycles. The first transport cycle following prey capture is often aberrant. During prey capture the head moves forward and down toward the prey item. Note that the head continues to move forward after tongue-prey contact, pinning the prey item to the substrate before it is retracted.
234
Kurt Schwenk
protrude
B
F I G U R E 8.23. Jaw and tongue kinematics during lingual prey capture in three iguanian lizards. In these plots the gape profile is shown inverted to facilitate comparison. In Chamaeleo the tongue is projected off the hyobranchium, but the timing of this behavior is similar to tongue protrusion in the generalized taxa, occurring at the end of SO II. The slight depression in gape at the end of SO II is unique to chameleons. (A) Dipsosaurus dorsalis (Iguanidae). Data from Schwenk and Throckmorton (1989). (B) Pogona barbata (Agamidae). Schwenk and Throckmorton, unpublished data. (C) Chamaeleo zeylanicus (Chamaeleonidae). Based on Bell (1990).
in several iguanids and agamids, including another species of Agama. Because prey size and type varied among these studies, it is not possible to separate treatment from species effects at this time. If the early part of ingestion is prolonged due to prey pursuit or lizard uncertainty (see earlier discussion), this is evident as a prolonged SO II phase. The distinction between SO I and SO II, and between the latter and FO, is sometimes vague so that jaw opening is more or less continuous before it is curtailed abruptly by FC (e.g., Bels, 1990, for Anolis). However, most ingestion gape cycles conform to the standard model with a pronounced "plateau phase" (Schwenk and Throckmorton, 1989) preceding rapid FO and FC phases (Figs. 8.22 and 8.23). SC or SCPS is generally short or sometimes absent depending on whether the prey comes to lie between the tooth rows or not (e.g., Herrel and De Vree, 1999a). There may be a short stationary phase before processing and transport cycles are initiated. Maximum gape angles vary radically within and among species (23° to 48°) without any clear pattern. Throckmorton (1976) showed that the lower jaw is protracted during tongue protrusion by means of quadrate rotation (streptostyly) in Uromastix, but not in Iguana. The gape cycle during ingestion in chameleons is remarkably similar to that described earlier for generalized iguanians, despite the addition of a ballistic projection interval to tongue protrusion (e.g., Bels and Baltus, 1987; Bell, 1990) (Fig. 8.23). The kinematics of ingestion in chameleons supports the notion that the projection interval has been inserted at the end of the SO phase in a generalized, ancestral ingestion cycle (Bramble and Wake, 1985; Schwenk and Bell, 1988; Schwenk and Throckmorton, 1989; see Section VI,A). EMG data for lingual ingestion are available only for Agama (Plocederma) stellio (Herrel et al, 1995). They sampled a variety of adductor and hyolingual musculature and found a mostly predictable pattern of activity. Low-level depressor mandibulae and pterygoideus profundus (lateralis) activity accompanies mouth opening during SO. Although the pterygoideus is usually regarded as an adductor, anatomical and functional data indicate that it may also function as a mandibular depressor and protractor (see discussion in Section IV,D). FO is initiated by a burst of depressor activity that is exactly coincident with tongue-prey contact, as the pterygoideus becomes nearly silent. Contraction of the spinalis capitis, a dorsal cervical muscle, also contributes to FO by raising the upper jaw through dorsal rotation of the skull at the neck. Pterygoideus activity spikes again during FC, along with the external and posterior adductors, as the depressor fades out. Pterygoideus activity fades out, in turn, during SC while the other adductors have a second, but lower burst of activity. The genioglossus muscle and
8. Feeding in Lepidosaurs the intrinsic "ring" muscle (posterior verticalis) are active throughout tongue protrusion during SO. As noted previously, this activity pattern is consistent with theoretical models of tongue protrusion in agamids. One noteworthy result of the EMG analysis of Herrel et al. (1995) is that the hyoglossus (tongue) muscle begins high levels of activity early in SO while the tongue is being protruded and continues throughout the gape cycle. Anatomically, the hyoglossus appears to be the principal lingual retractor (see earlier discussion), thus its activity during protrusion, with onset preceding significant genioglossus and verticalis activity, is unexpected. Herrel et al. (1995) suggested that hyoglossus activity causes hyobranchial protraction, but this could only occur if the tongue is held by antagonistic action of the genioglossus, which initially it is not. It is possible that genioglossus activity begins earlier in a part of the muscle not sampled (Herrel and colleagues did not specify whether the electrode was placed in the medialis or lateralis portion of the genioglossus). Continued activity of the hyoglossus during jaw closing is consistent with its putative role as a lingual retractor. A more plausible function of hyoglossus activity is provided by Herrel et al/s (1997a) study of intraoral transport and swallowing in the same species. They suggested that hyoglossus contraction during protrusion causes the tongue to shorten and "bulge." Such bulging, in conjunction with fixation of the tongue tip at the mandible by the genioglossus medialis (and internus in agamids) might contribute to the characteristic curled form of the tongue during prehension (Figs. 8.18 and 8.19). Indeed, lingual bulging is evident in the earliest stages of a prehension cycle. In any case, interpreting the action of lingual muscle activity is problematic because of the hydrostatic nature of the tongue (Schwenk, 2000a). Some activity might serve to stiffen the tongue or to deform it in ways that are not intuitive based on traditional musculoskeletal mechanics. 2. Processing Instances of preingestion processing are extremely rare in iguanians, as in all lepidosaurs. As noted earlier, the Galapagos land iguana (Conolophus) savors cactus pads, which it scrapes clean of spines and tough outer cuticle with the forelimbs before ingestion (Carpenter, 1969; Vagvolgyi and Vagvolgyi, 1978). In old accounts, the behavior of tongue-flicking prey before ingestion was often misinterpreted as preingestion lubrication! Food processing most often involves chewing with the marginal teeth following ingestion, but in many instances prey are not reduced at all before transport and swallowing. This happens when prey are very small (e.g., ants eaten by horned lizards; Schwenk,
235
manuscript in preparation) or in herbivorous species that crop a mouth-sized portion of a plant during an initial ingestion cycle (e.g., Uromastix, Throckmorton, 1976; Sauromalus, Schwenk, unpublished observations). Iguanians specialized for vertebrate prey swallow it whole using a combination of inertial and hyolingual transport with minimal processing (personal observation). However, killing or crushing bites might be considered a kind of processing. For example, leopard lizards (Gamhelia) manipulate lizard prey until they are held by the neck and anterior torso transversely across the mouth. They are then killed by a series of small amplitude, crushing bites that probably cause death by interfering with cardiac function (Lappin, 1999). Killing bites and similar behaviors are hard to categorize. They usually follow ingestion but precede chewing in the strict sense, hence they might best be considered a separate class of manipulative cycle rather than processing per se. In most pleurodont and some acrodont species, chewing takes the form of simple puncture-crushing in which the food item is repeatedly crushed between upper and lower teeth with simple, vertical movements of the jaws. Throckmorton (1976, 1980) suggested the possibility of axial rotation of the mandibular rami during chewing in Uromastix, but this is unsubstantiated [Bradley (1903) suggested a similar movement in lizards, generally]. Asymmetrical chewing, a mobile joint between mandibular rami, and streptostyly all introduce the possibility of more complex chewing movements, but these have not been documented. Bels and Goosse (1989) speculated that lingual retraction of the food item during jaw closure might promote a shearing action by sliding it against the approaching teeth just before crushing. In agamids and chamaeleonids, tooth function during chewing may be more complex than simple puncture-crushing. This is because their acrodont dentition permits occlusion, i.e., the precise alignment of upper and lower teeth (see Section IV,B). In occlusal view, most acrodont teeth appear as a series of triangles, corner to corner, in a saw-tooth pattern. With jaws closed, the triangular upper and lower fit exactly between one another (Fig. 8.8B). The fit is so tight that wear facts develop, sharpening the edges. Such occlusion introduces a shearing action that may be more effective in reducing food, possible producing a greater degree of comminution than in pleurodonts. However, this has not been demonstrated. My own casual observations suggest no greater degree of comminution in acrodonts, but possibly the formation of a softer bolus than in iguanids feeding on identical prey. Functional analyses of chewing are available for the following taxa: Iguanidae: Anolis (Bels and Baltus, 1988, 1989; Bels and Goose, 1989); Ctenosaura (Smith,
236
Kurt Schwenk
1984); Iguana (Throckmorton, 1976); Opiums (Delheusy and Bels, 1992); Agamidae: Agama (including Plocederma) (Kraklau, 1991; 1997a, 1998b); Pogona (Throckmorton and Clarke, 1981); Uromastix (Throckmorton, 1976,1978,1980; Herrel et al, 1998b); and Chamaeleonidae: Chamaeleo (So et ah, 1992). Schwenk and Throckmorton (1989), Herrel et al (1996a), and Herrel and De Vree (1999) did not distinguish between chewing and intraoral transport cycles sensu stricto in the several iguanid and agamid species they studied. According to Herrel et al. (1997a), in Agama these cycles differ mainly in the intensity of muscle activity rather than patterns of onset and offset, or other kinematic variables, and this may be a general phenomenon (but see Section V,B,2). In general, chewing cycles are kinematically similar to intraoral transport cycles (e.g., Herrel et al, 1996a); however, the single quantitative study of a chameleon found them to be distinct (So et al, 1992). Gape profiles are variable, as always, but typically show a pronounced SO interval that is usually, but not always, divisible into SO I and SO II phases. Surprisingly, an SC phase was not observed in chewing Anolis (Bels and Baltus, 1989; Bels and Goose, 1989). Smith (1984:129) found that during intraoral transport the tongue and hyobranchium moved anteriorly beneath the prey item in SO and were retracted in FO, in conformity with the model gape cycle, but during bite (chewing) cycles (interspersed between transport cycles), the tongue continued to move forward during FO "to place the food between the teeth in readiness for tooth contact in fast closing." Similar patterns were found by Bels and Goose (1989) in Anolis and Herrel and De Vree (1999a) in Uromastix, but Delheusy and Bels (1994) and Herrel et al. (1996a) found no difference in tongue movement between chewing and transport cycles for Oplurus and Agama, respectively, with retraction beginning at the start of FO. Based on superficial indications of hyobranchial position, Kraklau (1991) suggested that there was little hyobranchial movement associated with chewing cycles, but significant anteroposterior excursion during intraoral transport. In Chamaeleo, the principal difference between chewing and transport cycles is that the hyobranchium remains more or less stationary during the former, but follows a typical pattern of movement in the latter (So et al, 1992). However, So et al. (1992) noted that despite the stationary hyobranchium during chewing, prey items were manipulated by an independently mobile tongue (see Section VI). In general, the relationship between hyolingual movement and gape phase in different feeding stages is more variable than generally acknowledged, thus taxonomic differences are difficult to interpret. This issue is discussed further in Section VII. In any case, FC is virtually always accompanied by hyolingual retraction and
placement of the food item between the teeth. SC-PS begins with tooth-food contact as the prey item is crushed between upper and lower tooth rows. The number of chewing cycles is extremely variable and depends on species, prey type, and prey size (e.g., Bels and Baltus, 1988). Chewing cycles sometimes occur in clusters and sometimes are interspersed between transport cycles, but usually they occur early in a feeding sequence and are followed by an increasing number of transport cycles. In Chamaeleo, chewing cycles precede transport cycles (So et ah, 1992; Schwenk et al., manuscript in preparation). Maximum gape angles for reduction cycles are comparable to ingestion and transport with typical values in the range of 20° to 35°; however. So et al. (1992) found that maximum gape was greater in chewing than in transport in Chamaeleo (they did not specify a value). Feeding on large prey items incurs longer cycle times, but not larger excursion distances, in Agama (Herrel et al, 1996a); however, chewing cycles were not explicitly distinguished from transport cycles in this study. Muscle activity patterns during chewing cycles in Agama (Herrel et al, 1997a) are described later along with intraoral transport. Herrel et al. (1998a) modeled bite forces in two agamids, Agama (Plocederma) and Uromastix, and estimated bite forces of 6 to 10 and 8 to 14 newtons, respectively. Based on this and one other comparison of two scleroglossan species (see later), Herrel et al. (1998a) suggested that herbivorous species generate greater bite forces than related insectivores, but the study is not adequately controlled for phylogenetic effects and this conclusion is premature. Herrel et al. (1998b) modeled bite forces in the same taxa and concluded that the temporal ligaments serve to stabilize the quadrate for increasing bite force and that the jugomandibular modification of the ligament is particularly effective in maintaining joint alignment for resisting especially powerful bite forces in akinetic forms. There may be ontogenetic changes in the biting mechanism. Capel-Williams and Pratten (1978) found that allometric changes in jaw configuration during growth in Agama bibroni lead to a rapid bite that is two and a half times more powerful in adults than in juveniles. They suggested that these mechanical differences limit diet choice in juveniles and underlie an ontogenetic shift in diet. Rieppel and Labhardt (1979) found a similar ontogenetic transition in mandibular mechanics and diet in Varanus niloticus. 3. Intraoral
Transport
Functional analyses of intraoral transport are available for the following iguanian taxa: Iguanidae: Anolis (Bels and Baltus, 1988, 1989; Bels and Goosse, 1989);
8. Feeding in Lepidosaurs Ctenosaura (Smith, 1984); Dipsosaurus (Schwenk and Throckmorton, 1989); Iguana (Throckmorton, 1976; Schwenk and Throckmorton, 1989); Opiums (Delheusy and Bels, 1992); Phrynosoma (Schwenk and Throckmorton, 1989; Schwenk, manuscript in preparation); Sauromalus (Schwenk and Throckmorton, 1989); Agamidae: Agama (including Plocederma) (Herrel et ah, 1996a, 1997a); Phrynocephalus (Schwenk and Throckmorton, 1989); Pogona (Schwenk and Throckmorton, 1989); Uromastix (Throckmorton, 1976,1978,1980; Throckmorton and Clarke, 1981; Herrel and De Vree, 1999a); and Chamaeleonidae: Chamaeleo (So et ah, 1992). The account of a generalized, hyolingual transport cycle given earlier serves as an appropriate description for intraoral transport in most iguanians (Fig. 8.22), although Herrel et al (1996a) noted that the relative lengths of SO I and SO II phases are not always consistent with the Bramble and Wake (1985) model. The principal difference between ingestion and transport cycles may be that SO typically (but not always) occurs at a lower gape angle during transport because there is no need to accommodate a protruded tongue (Schwenk and Throckmorton, 1989). Smith (1984) showed that tongue and hyobranchial movements are tightly coupled during transport in Ctenosaura, as they are in Sphenodon. Kraklau (1991) observed significant hyobranchial excursions during transport in Agama. Delheusy and Bels (1992) found that the duration of the stationary phase was longer between transport cycles than chewing cycles and that the SO phase was proportionately longer. The latter observation is consistent with the suggestion that the tongue is formed to the prey item during transport SO II (Bramble and Wake, 1985). Light and cineradiographic films confirm this hypothesis, revealing considerable lingual shape change during SO that fits the lingual surface to the bolus in order to hold it for transport (Schwenk, unpublished observations). Bramble and Wake (1985) noted the presence of muscle fibers within the lingual papillae and suggested that active papillary movement might contribute to this process, but this is unconfirmed. Smith (1984) observed localized changes in tongue length such that the anterior part of the tongue elongated underneath the bolus and then contracted, followed by expansion of the posterior end relative to the hyobranchium. The posterior end then shortens toward the basihyal during retraction. The tongue usually arches dorsally and scrapes the palate in the early part of hyolingual retraction. The bolus is generally held on the posterior part of the tongue with the anterior part arched in front of it; however, sometimes the bolus simply adheres to the anterior part of the tongue as it is retracted. In Chamaeleo the dorsal surface of the glandular portion of the tongue forms a sigmoid curve with the bolus cupped in a posterior depression
237
(Schwenk et ah, manuscript in preparation). In general, retraction is accompanied by considerable ventral movement of the tongue and hyobranchium. The ventral component of retraction is extremely pronounced in some taxa, such as Chamaeleo (Schwenk et ah, manuscript in preparation) and Phrynosoma (Schwenk, manuscript in preparation). At the end of retraction in these taxa the gular region forms a deeply distended pocket that is slowly elevated during hyolingual protraction of the next cycle. Ventral movement during retraction clears the bolus from any intraoral contact so that it can be moved far back into the pharynx. Herrel et ah (1996a) found that transport cycles were longer in duration in Agama feeding on large prey items but that there was no significant increase in maximal excursion distances of jaws and tongue. Food size and type also affect the number of transport cycles required to move the bolus into the pharynx (e.g., Bels and Baltus, 1988). Although it is likely that quantitative differences in transport cycles are related to differences in prey mass and the process of fitting the tongue to the bolus during SO II and the efficacy of the resulting adhesion (Bramble and Wake, 1985), this remains speculative. For example, Bels and Baltus (1988) found that some fruits took fewer cycles to transport than insects and related this to the lack of lingual adhesion due to expressed juices, but in fact, one would expect lowered adhesion to result in more transport cycles to compensate for the smaller posterior distance traveled by the bolus with each cycle. Alternatively, the fruit might have tended to liquefy and more or less flow into the pharynx, reducing the number of transport cycles (Bels and Baltus, 1988). Herrel and De Vree (1999a) found only minor kinematic differences in Uromastix feeding on two very different food types (endive leaves and locusts), but Herrel et ah (1996) found longer cycle times in Agama feeding on larger prey, which they interpreted as being consistent with the Bramble and Wake (1985) hypothesis. However, increased cycle times while feeding on more massive prey might be a general scaling effect on the entire feeding apparatus rather than a specific consequence of tongue fitting and adhesion. Although Herrel and De Vree (1999a) did not distinguish chewing from transport cycles in Uromastix, they seemed to imply that, in all such cycles, tongue protraction continues through FO and that retraction does not begin until FC (see Section V,D,2). Bels and Goosse (1989) found the same pattern in Anolis. This pattern is similar to biting cycles described by Smith (1984), but is inconsistent with the model gape cycle and kinematic patterns during intraoral transport in other taxa. As noted previously, there is considerable variation in these kinematic relationships with no functional or phylogenetic pattern discernible thus far (see Section VII).
238
Kurt Schwenk
Herrel et al (1997a) provided EMG data for transport cycles in Agama. Muscle activity patterns were similar to "crushing" (chewing) cycles except that in the latter, muscle activity levels were significantly higher, particularly in the anterior part of the adductor mandibulae externus superficialis. During SO, as the tongue is protracted, virtually all lingual muscles are active (genioglossus medialis and lateralis, hyoglossus, posterior verticalis). Although activity of the genioglossus and verticalis muscles is consonant with models of tongue protrusion, activity of the hyoglossus muscle, a putative retractor, is unexpected. This pattern is discussed in Section V,D,1. Herrel et al. (1997a: 112) suggested that hyoglossus activity shortens the tongue and in conjunction with protraction "causes the tongue to bulge, thus pushing it against the prey item." They noted that Anolis exhibited the same activity pattern (J. Cleuren and F. De Vree, in Herrel et al, 1997a). Although the "bulging" hypothesis is consistent with tongue deformation during ingestion (see earlier discussion), its timing during transport is more critical. Initially the tongue is "slid" underneath the bolus and bulging could not occur until the end of SO II when the tongue is fitted to the food item and readied for retraction. The mandibulohyoideus is active, presumably to protract the hyobranchium, and the stemohyoideus is active at very low levels, possibly to keep the hyobranchium in tension during protraction for better control and to initiate a rapid reversal at FO for hyolingual retraction. Herrel et al. (1997a) also noted low levels of activity in several adductors during SO II. Mouth opening is primarily caused by mandibular depression with contraction of the depressor mandibulae, but at FO a burst of activity in the dorsal cervical muscles (spinalis capitis) causes elevation of the cranium as well. The adductors (adductor mandibulae externus superficialis anterior and externus medialis, adductor mandibulae posterior, pseudotemporalis, ptergoideus superficialis and profundus) are active almost simultaneously at the beginning of jaw closing, although peak activity in the adductor mandibulae externus and pseudotemporalis lags a few milliseconds behind. Several of the adductors show biphasic activity, especially the adductor mandibulae posterior and the pterygoideus superficialis, with the second burst associated with SC. A true power stroke (SC-PS) only occurs during chewing cycles. In chewing cycles, adductor activity becomes more synchronous, and the high levels of activity tend to blot out the biphasic pattern, except in the pterygoideus superficialis. Activity of the hyobranchial retractors (stemohyoideus and omohyoideus) begins just prior to FO and, along with a burst of activity in the hyoglossus, causes rapid retraction of the tongue and hyobranchial apparatus.
Limited inertial transport has been observed in several iguanians. Smith (1984) noted that Ctenosaura used inertial transport for large or heavy food items, as did Throckmorton (1976) for Iguana and Uromastix, but large prey or heavy items are not typical of the natural diet in these species. In contrast, leopard lizards (Gatnbelia) routinely consume extremely large lizard prey and one might expect inertial transport to predominate. However, Lappin (personal communication) observed that, "Inertial transport may play a minor role, but there is a lot of hyolingual activity, even with the largest lizard prey" (see Lappin, 1999). Delheusy and Bels (1992) found that Opiums used an inertial transport cycle immediately following ingestion to position the prey item posteriorly at the corner of the mouth for chewing. Subsequent chewing and transport cycles were all hyolingual. Manipulation, rather than transport per se, may be the most typical use of inertial movements in iguanians. 4.
Swallowing
As noted previously, swallowing in lepidosaurs potentially consists of two kinematic stages, pharyngeal packing and pharyngeal compression, but in some cases the former is sufficient to engage peristalsis and compression is unnecessary to complete the act of swallowing. Our films suggest that when feeding on small, naturalistic prey items, pharyngeal compression is rare in iguanians and when it occurs it is very subtle and brief (Schwenk, unpublished results; Schwenk and Throckmorton, unpublished results). Iguanians seem to rely primarily on pharyngeal packing for swallowing with pharyngeal compression more common in scleroglossans. a. Pharyngeal Packing Pharyngeal packing is frequently described in the literature as "swallowing," "cleaning," and "lip licking," as discussed earlier. In some studies it is clear that descriptions of terminal transport cycles are actually transitional packing cycles. Sometimes packing and cleaning are described separately, but again, the distinction seems to relate to the changing kinematics of pharyngeal packing during a sequence of cycles, with "cleaning" cycles the last of the series before pharyngeal compression, if it occurs. Consequently, my delimitation of pharyngeal packing (based partly on my own observations) does not always agree precisely with descriptions in the literature. In any case, as has been noted several times already, feeding stages tend to blend into one another functionally and kinematically.
8. Feeding in Lepidosaurs Functional accounts of pharyngeal packing are limited to the following taxa: Iguanidae: Anolis (Bels and Baltus, 1988); Ctenosaura (Smith, 1984); Opiums (Delheusy and Bels, 1992); Agamidae: Agama (including Plocederma) (Fierrel et ah, 1996a, 1997a); and Uromastix (Throckmorton, 1980; Herrel and De Vree, 1999a). Smith (1984) first described pharyngeal packing in Ctenosaura. She found that packing cycles differed from intraoral transport in having lower amplitude gapes and different tongue movements due to a shift of bolus position from the dorsal surface of the tongue to mostly behind it. During SO the tongue is positioned in front of and above the bolus. With the bolus so positioned, the tongue is able to push the bolus posterior to the basihyal so that pharyngeal compression will squeeze it into the esophagus and not back into the mouth. The posterior end of the tongue, presumably the posterior limbs, is used to "tamp" the bolus into the pharynx and/or the esophagus. This usually involves high amplitude anteroposterior excursion of the hyolingual apparatus and often includes tongue protrusion during the protraction phase, especially in terminal cycles. The tongue protrudes tip first, either straight ahead or to the side, and often 'Ticks" the labial scales from one side to the other. For example, after consumption of a cactus pad, a land iguana {Conolophus) initiated "a long series of gulping movements . . . at least 43 in number. After each gulp the iguana extended its tongue slightly and licked its labials" (Vagvolgyi and Vagvolgyi (1978:162). I interpret the asymmetry of these movements to indicate asymmetrical use of the posterior limbs during the retraction phase—possibly one limb is used at a time. The protrusion may help maximize the anterior excursion of the tongue and hyobranchium to position the bolus behind the tongue more effectively and also help orient the hindtongue and posterior limbs side to side. Herrel et al. (1996a) suggested that initial packing cycles move the bolus into the esophagus and that in later cycles the tongue moves anteriorly beyond the bolus, "bulges," and then pushes the bolus farther into the esophagus. Delheusy and Bels (1992) noted that packing cycles in which the tongue was protruded were more effective at moving the bolus posteriorly than cycles without protrusion. They speculated that intrinsic shape changes in the tongue might promote the storage of elastic strain energy in the laryngohyoid ligament such that the hindtongue would "snap" back to accelerate the bolus posteriorly (also see Bels et ah, 1994), but the nature of tongue deformation does not suggest the ligament is being stretched and therefore does not support their hypothesis. The hypothesis also assumes that the larynx remains fixed so that the ligament is stretched when the tongue is protracted, but our films
239
show that this does not occur during ingestion in Pogona (Schwenk and Throckmorton, unpublished observations) nor in Sphenodon (see earlier discussion; also see description of the ligament in Section IV,E,2), suggesting that the tongue and the larynx are coupled anatomically and kinematically. However, it is unknown if laryngeal coupling via the laryngohyoid ligament maintains for other feeding stages and in other taxa. Herrel et al. (1996a) found the kinematics of pharyngeal packing to be roughly similar to transport cycles in Agama, but with a shorter FO phase, a lower maximum gape angle, and the absence of an SC-PS phase. They also noted that the distinction between SO I and SO II is weak or absent. They attributed this to the posterior position of the bolus, which implies that it does not require tongue fitting, the putative mechanistic basis of a pronounced SO II (Bramble and Wake, 1985). However, packing cycles are kinematically very distinct in Uromastix (Herrel and De Vree, 1999a). In comparison to transport cycles, packing cycles have smaller gape angles with virtually no contribution of cranial elevation to FO, a longer SO phase, and greater tongue excursions (Fig. 8.24). During retraction, "the posterior edge of the tongue" is used to push the bolus posteriorly. As in transport in this species (see earlier discussion), the tongue continues to move anteriorly during FO with maximum tongue protraction corresponding closely with maximum gape. There is a tendency for tongue retraction to be faster during pharyngeal packing. Delheusy and Bels (1992) also noted that terminal packing cycles with tongue protrusion had smaller gape amplitudes than earlier cycles, as did Smith (1984). Muscle activity patterns during pharyngeal packing are distinctly different from intraoral transport in Agama with differences concentrated in the adductor musculature (Herrel et ah, 1997a). Small differences occur in the relative timing of the hyolingual muscles, but the pattern is generally similar to transport (see earlier discussion). In late packing cycles, depressor mandibulae and spinalis capitis activity is often absent. Given the reduced gape angles in these later cycles it is possible that gravity and mechanical displacement by the bulging tongue during retraction are sufficient to depress the mandible. At jaw closing, the adductors become active as in transport, but only the "deeper parts" of the adductor mandibulae externus, adductor mandibulae posterior, and the pterygoideus remain active through the cycle. Hyolingual retraction is driven by activity of the sternohyoideus and hyoglossus. Herrel et ah (1997a) noted that activity patterns were highly variable among packing cycles. This presumably reflects the rapidly changing position and condition of the bolus as it is tamped, in contract to
240
Kurt Schwenk
o < Q LU
<
UJ
o z <
CO Q LU GL
<
TIME (SEC) F I G U R E 8.24. Jaw kinematics during intraoral transport and pharyngeal packing. (Top) Uromastix acanthinurus (Iguania, Agamidae). Based on Herrel and De Vree (1999a). (Bottom) Tiliqua rugosa (Scleroglossa, Scincidae). Based on Herrel et al. (1999b).
the more stereotypical movements of chewing and transport. h. Pharyngeal Compression Accounts of pharyngeal compression in iguanians are restricted to the following references: Iguanidae: Ctenosaura (Smith, 1984); Gambelia (Lappin, 1999); and Agamidae: Agama (Herrel et al, 1996a). Schwenk (unpublished results) and Schwenk and Throckmorton (unpublished results) have film data for several other iguanids and agamids. Ctenosaura elevates the head and uses neck bending rather than hyobranchial elevation for pharyngeal compression, but only for large food items. The natural diet of Ctenosaura would rarely include large prey items, hence this behavior may be aberrant. This is in contrast to a scleroglossan, Tupinambis, which relies heavily on hyobranchial elevation for swallowing, presumably by constrictor colli contraction (Smith, 1984; see later). Gambelia eating large lizard prey uses lateral bending of the neck to force it farther posterior once the prey item is partially engulfed (Lappin, 1999). Her-
rel et al (1996a:1734) did not recognize pharyngeal compression as a separate component of swallowing in Agama, noting only that, "Once inside the esophagus, constriction of the throat region pushes the prey further down." Our films of various iguanians eating small insect prey suggest that pharyngeal compression of any kind is rare, occurring sporadically and unpredictably after pharyngeal packing, and is usually limited to a single, small compressive cycle evident as elevation of the hyobranchium. However, we did not test the animals with larger prey, although Iguana swallowing bulky lettuce leaves did not compress the pharynx.
E. Feeding in Scleroglossa 1. Ingestion Ingestion has been poorly studied in scleroglossans. This account is based on the following references, as well as my own observations: Gekkota: Lialis (Patchell and Shine, 1986; Murray et al, 1991); Eublepharus (Delheusy et al, 1995); Scincomorpha: Tiliqua {Trachydosau-
241
8. F e e d i n g in L e p i d o s a u r s
rus) (Gans et al, 1985); Lacerta (Goosse and Bels, 1992; Urbani and Bels, 1995); Zonosaurus (Urbani and Bels, 1995); Tupinambis (McBrayer and White, manuscript); Anguimorpha: Elgaria (Gerrhonotus) (Frazzetta, 1983); Heloderma (Herrel et al, 1997b); and Varanus, (Condon, 1987; see also Section VI,C). Amphisbaenians are treated in Section VI,B. Scleroglossans almost universally use the jaws and teeth as organs of prehension during ingestion of all prey items (Schwenk and Throckmorton, 1989). There are several reported cases of lingual ingestion among scleroglossan species, however. These are discussed separately, after a description of jaw prehension. Scleroglossans are usually alerted to the presence of a prey item by visual cues, especially movement, as in iguanians. However, many scleroglossan species use chemosensory searching to locate hidden prey and either expose it or flush it from cover (see Section III). Once spotted, vision seems to guide the strike, which may be curtailed if the prey item ceases motion. Monocular fixation of prey is especially evident in anguimorphans who have a characteristic (but not universal) strike behavior (personal observation): the front part of the body is raised on the forelimbs while the head is tilted down and to the side and axially rotated so that one eye looks down toward the prey. The strike is initiated with a downward and sideways sweep while the front end of the body is twisted and bent along its longitudinal axis, pinning the prey to the substrate with the tips of the jaws. In extreme cases the head is virtually vertical at prey contact with its dorsum directed laterally relative to the initial axis of the body. Because the rear part of the body maintains the original axis, the front end is both bent and twisted. This behavior is undoubtedly facilitated by the typically elongate body form of most anguimorphans, including an elongate neck in some. Tail vibration or wriggling is observed particularly in anguids and gekkonids, but also in other scleroglossans during prey capture (e.g., Loveridge, 1953; Murray et al, 1991; Perez-Mellado, 1994; personal observation). In the pygopodid, Lialis, tail movement was only observed after skink prey escaped an initial capture attempt (Murray et al, 1991). It may act as a lure or to distract the prey from the predator's presence (e.g.. Carpenter and Ferguson, 1977). It often occurs when a prey item "disappears'' from view by becoming motionless (personal observation) or after a failed strike (Loveridge, 1953; Murray et al, 1991). Given that tail movement only seems to occur when the lizard is in a state of high arousal, it probably represents a type of displacement behavior (Radcliffe et al, 1980). Of course, lure, distraction, and displacement are not mutually exclusive hypotheses.
In most scleroglossans, the approach is similar to that described for iguanians with the exception that the tongue is not protruded as the jaws begin to part. In fact, the tongue is usually seen to be withdrawn out of the way (e.g., Frazzetta, 1983; personal observation). Frazzetta (1983) pointed out that in the anguid, Elgaria, the tip of the mandible always contacts the prey item first with the upper jaw closing on it a moment later. He suggested that the lizard aligns its lower jaw with the prey item rather than attempting the complex task of having both upper and lower jaws align simultaneously while moving forward. Once contacted, a rapid snap of the jaws seizes the prey item. Importantly, mesokinetic upper jaw elevation is evident immediately before prehension in kinetic forms (Fig. 8.25) (Frazzetta, 1983; Condon, 1987; personal observation). The prey item is usually captured near the tips of the jaws yy
[iiniiniiiiniiiiuiiiiiiiiiiiiiiiiinii 5° D°
4
df vf
-I 33
HORZ.
f-
up down
I l l l l l l H l l l l l l l i m i l I I I M I M i n I MM Ml I
B
I i I I I N N M U M M M N N M N N N M M M M Ml
_ i 5° GAPE
HORZ.
VERT.
r;;
J
t
ant.
It "^ -J T
IIIIIIIMIIIIIIIIIMMMMMMMMIMMIII
down
FIGURE 8.25. Kinematics of jaw prehension (A) and inertial transport (B) in a monitor lizard, Varanus niloticus. The upper plot in both shows mesokinetic flexion of the snout. Note that dorsoflexion above the rest point only occurs during the strike. Jaw-prey contact occurs during ventroflexion as the snout passes the rest point (arrow). This pattern is consistent with the hypothesis that mesokinesis is an adaptation for jaw prehension in scleroglossan lizards. All subsequent mesokinetic movement occurs with the snout in net ventroflexion. Not all inertial cycles are accompanied by mesokinesis. After Condon (1987), Exp. Biol 47, 73-87, © Springer-Verlag, with permission.
242
Kurt Schwenk
(Fig. 8.20). Rapid jaw closure is accompanied by mesokinetic ventroflexion of the snout beyond the rest position and the upper jaw is held in ventroflexion throughout processing and intraoral transport (see later). In addition to mesokinetic elevation of the upper jaw before capture, head elevation at the neck contributes to upper jaw elevation. In forms such as the scincid Tiliqua, the head begins close to the ground due to short limbs. Larger prey are sometimes seized by raising the head up on the forelimbs and rotating it axially so that the sides of the jaws are used for prehension rather than the tips (Gans et al, 1985). Jaw-feeding scleroglossans that feed on vertebrates or other large prey often direct the initial bite to the head or front part of the body (e.g., Murray et ah, 1991). Presumably this facilitates head-first orientation of the prey for transport and swallowing. Scleroglossans differ from iguanians in the form of a typical ingestion gape cycle in that discrete SO and FO stages rarely occur (Fig. 8.25 and 8.26). The profile is usually sharply peaked or rounded, like a bell curve, but only opening and closing phases are evident. Such gape profiles are not surprising considering that the tongue (and hyobranchium) is more or less fixed in place during ingestion. In contrast, lingual prehension (when it occurs) and intraoral transport cycles, which include anteroposterior hyolingual movements, usually appear more like the model gape cycle with typical SO and FO phases (see later). It is sometimes noted that ingestion requires more than one gape cycle in scleroglossans (e.g.. Goose and Bels, 1992; McBrayer and White, manuscript); however, this confusing observation has not been elaborated. Based on descriptive data available and my own observations, additional cycles beyond the first are almost certainly types of manipulative or processing cycles. For example, the prey item is sometimes seized by the tips of the jaws, but is then moved back quickly to
0.4
the tooth rows for a killing bite, or the prey item is shaken and crushed against the substrate to subdue it. If the prey item is initially dropped or missed, then additional capture/ingestion cycles might occur, but each attempt represents a single cycle of ingestion that potentially begins a feeding sequence, not multiple ingestion cycles. Preingestion processing cycles are also possible (see later). However, once held by the jaws after an initial seizure, all subsequent cycles are postingestion and should be characterized as such. The single, possible exception is the case of some herbivorous iguanians noted earlier. Smith (1982) and McBrayer and White (manuscript) have provided the only available data on muscle activity patterns during ingestion in scleroglossans {Varanus and Tupinambis, respectively); however. Smith (1982) did not clearly differentiate between ingestion cycles and subsequent bite (processing) cycles. In Tupinambis, ingestion sequences showed either simultaneous activity of the three adductors measured (externus superficialis, pseudotemporalis superficialis, and pterygoideus superficialis) or activity of the external adductor alone, depending on the apparent force of the bite. Ongoing studies by Anthony Herrel and colleagues may help mitigate the dearth of EMG data for scleroglossan ingestion (Herrel, personal communication). Prey capture by the jaws in most scleroglossans can be characterized most often as a rapid, pincer-like behavior. The challenge to the predator is to catch a potentially active prey item before it can dart away and without knocking it away with the jaws during the attempt. Cranial kinesis, particularly elevation and depression of the upper jaw at the mesokinetic axis, is likely to be an adaptation to enhance capture success with the jaws (Frazzetta, 1983). The issue of cranial kinesis is treated in more detail later. Prey capture, however, need not be rapid. In species that feed on stationary food items, such as Heloderma eating eggs (Herrel
0.0 TIME (SEC)
F I G U R E 8.26. Kinematics of jaw prehension in two scleroglossan lizards feeding on crickets. (Left) Lacerta viridis (Lacertidae) and (right) Zonosaurus laticaudatus (Cordylidae). Based on Urbani and Bels (1995).
8. Feeding in Lepidosaurs et al, 1997b) or Tiliqua eating snails (Gans et al, 1985), it is slow and deliberate. It is interesting to note in this context that in neither of these examples was cranial kinesis observed and in both cases the skull is probably anatomically akinetic (Gans et al, 1985; De Vree and Gans, 1987; Herrel et al, 1997b; personal observation). These observations are consistent with the notion of cranial kinesis as an adaptation for rapid jaw prehension of active prey (see later). In several scleroglossan species the tongue is involved in prey capture: Scincidae, Tiliqua scincoides (Smith et al, 1999); Tiliqua (Trachydosaurus) rugosa (Gans et al, 1985; A. Herrel, personal communication); Cordylidae, Zonosaurus laticaudatus (Urbani and Bels, 1995); and Phelsuma (lineatus?) (J. Gauthier, personal communication). In all cases, lingual prehension is used in addition to jaw prehension in the same individual and is apparently employed for certain prey/ food types. In the case of Zonosaurus and T. rugosa feeding on snails, the use of the tongue during ingestion is distinctive and unusual, but does not constitute lingual ingestion in the strict (iguanian) sense. In Zonosaurus, the prey item was contacted by the tongue as the jaws simultaneously moved over it. Thus, Urbani and Bels (1995:284) concluded that "the mechanism of prehension in cordylids seems to be based on jaw use," although the tongue is sometimes used to hold the prey item in place. Zonosaurus lingual ingestion also differed from iguanians in that tongue-prey contact occurred after the jaws began to close, as opposed to the end of SO. In T. rugosa, the tongue was generally used only after an initial bite with the jaws failed to apprehend a snail. In a second attempt using lingual ingestion, the snail was contacted with the tongue and dragged a short distance across the substrate toward the mouth as the jaws continued to move toward the prey item and grasp it. In neither case was the tongue used to apprehend and lift the prey item into the mouth, as in iguanians. Iguanian-type lingual prehension sometimes occurs in blue-tongued skinks (T. scincoides) feeding on mealworms (Smith et al, 1999). The skinks use both jaws and tongue to ingest mealworms, but the jaws exclusively for crickets, a behavioral shift consistent with Urbani and Bel's (1995) observations of Zonosaurus. In Phelsuma, however, lingual ingestion was observed with crickets, but other prey types were not tested (J. Gauthier, personal communication). Urbani and Bels (1995) related the use of tongue versus jaws to prey size [as suggested by Smith (1984), and Schwenk and Throckmorton (1989)], but Smith et al (1999) thought that jaw prehension might be employed for more active prey. Gape cycle kinematics for the two ingestion types
243
are statistically indistinguishable in Zonosaurus (Urbani and Bels, 1995), but in Tiliqua the gape cycles are qualitatively different—a typical, scleroglossan profile with a spiked peak, lacking a distinct SO, is evident during jaw prehension, but the gape profile has welldifferentiated SO I and SO II phases during lingual ingestion (Smith et al, 1999). The latter pattern is similar to iguanians and consistent with the model gape cycle. It may be related to hyolingual excursion and tongue fitting during lingual ingestion, but one interesting observation confounds this conclusion. Smith et al (1999) noted that such iguanian-type gape profiles characterized the majority of lingual ingestion cycles, but only when the mandible contacted the substrate just before tongue prehension. In two cases the mandible did not contact the substrate and a distinct SO II phase was absent, suggesting that the SO II phase was an artifact of mandible-substrate contact and not a consequence of hyolingual protraction. In iguanians, the presence of SO II is uncorrected with substrate contact, which, in any case, rarely occurs (personal observation). Although highly preliminary, this suggests that similar kinematic patterns during lingual ingestion in iguanians and scleroglossans may be achieved in different ways, further muddying the issue of kinematic homology, but supporting the argument that lingual ingestion in scleroglossans is secondarily derived, i.e., a reversal rather than a retention of the ancestral pattern (see Section VII,A). It is also possible that lingual ingestion in scleroglossans is accomplished without significant hyobranchial protraction, in contrast to Sphenodon and iguanians, perhaps using a combination of whole tongue translation relative to a fixed hyobranchium and hydrostatic elongation (see earlier discussion). In order to compare mechanisms of ingestion in iguanians and scleroglossans, it is essential to clarify the relationship between tongue and hyobranchial movement in all feeding stages. 2. Processing Processing in scleroglossans is very similar to the pattern described for pleurodont iguanians and consists primarily of puncture-crushing with the marginal teeth. Preingestion processing is unusual. The Mexican beaded lizard {Heloderma horridum) uses its slightly procumbent, sharp, anterior teeth to puncture calcareous eggs several times before ingestion with the jaws (Herrel et al, 1997b). Puncturing significantly reduces the strength of the egg, facilitating subsequent intraoral crushing. Komodo monitors (Varanus komodoensis) occasionally use the forelimbs and claws to scrape the hair from rotting carcasses before feeding (Auffenberg, 1981).
244
Kurt Schwenk
Although comminution of food during chewing is not typical of scleroglossans, nor other lepidosaurs (see earlier discussion), prey are sometimes crushed or fragmented by shaking or dragging against the substrate. Fitch's (1958:32) account of this behavior in a small teiid (Cnemidophorus) is evocative: ''As compared with some others, these lizards are dainty feeders, which spend a long time killing, and worrying their prey, and sometimes reducing it to small morsels, instead of gulping down relatively large objects." Fitch (1954:118) also observed a small skink (Eumeces) systematically dismembering a large cricket by repeated biting and shaking until most appendages were removed. Dismemberment was followed by repeated chewing cycles until "the body was softened." Powerful killing bites are sometimes observed immediately following ingestion of large prey, as well as manipulative cycles to place the prey item between the tooth rows or to orient it head first in the mouth. Large monitors and teiids feeding on mice often crush the prey against the ground or shake it violently (e.g.. Smith, 1982; Condon, 1987; McBrayer and White, manuscript). Small Eumeces grasp crickets at the thorax and then work their way to the head while chewing rapidly (Webb, 1949). Large prey are "crushed in the jaws and battered on the ground" (Fitch, 1954:121). The unusual pygopodid, Lialis, seizes its large, skink prey at the level of, or anterior to, the pectoral girdle, often shaking it violently and pushing it into the substrate (Murray et ah, 1991). This is followed by repeated biting, putatively to suffocate the skink, but probably causing rapid death by cardiac arrest. The skink is nearly always swallowed head first and is manipulated by an unusual, probably unique, pattern of lateral jaw movements to reorient the skink in the mouth before transport (Patchell and Shine, 1986a). The mechanical basis of such asymmetrical jaw motion is unknown.
but it is presumably associated with the highly kinetic skulls of these unusual pygopodids. Functional accounts of food reduction in scleroglossans are available for the following taxa: Gekkota: Eublepharis (Delheusy et ah, 1995) and Gekko (Andrews and Bertram, 1997); Scincomorpha: Lacerta (Goosse and Bels, 1992; Urbani and Bels, 1995); Tiliqua (Cans et al, 1985; Gans and De Vree, 1986) [the Herrel et al, (1999b) study of Tiliqua and Corucia does not differentiate between chewing and transport cycles, although it is clear from their description that many nominal "transport" cycles were, in fact, chewing]; Tupinambis (Smith, 1984; McBrayer and White, manuscript); Anguimorpha: Heloderma (Herrel et ah, 1997b); and Varanus (Smith, 1982; Condon, 1987). As in iguanians, chewing cycles usually occur clustered together, but sometimes they are interspersed with transport cycles. It appears that species using hyolingual transport tend to cluster their chewing cycles early in the feeding sequence followed by intraoral transport cycles (e.g., Goosse and Bels, 1992) and that inertially feeding species are more likely to intersperse transport cycles with biting cycles (e.g.. Smith, 1982; Condon, 1987). Bites in the latter case are not really equivalent to cyclical chewing, although they do seem to serve a processing function. Because cyclical chewing and hyolingual transport cycles are associated with similar types of movement in the jaws, tongue, and hyobranchial apparatus, these cycles tend to grade one into another and are sometimes not distinguished in the literature (see earlier discussion). The kinematics of chewing are very similar to the general pattern observed in iguanians. Usually all phases are clearly distinguishable in the gape cycle (Fig. 8.27), although in inertially feeding Varanus with interspersed bites, the gape profile tends to have a poorly differentiated SO phase and a large spike for
LU —I
o
LU Q-
<
F I G U R E 8.27. Kinematics of chewing in a scleroglossan lizard, Zonosaurus laticaudatus (Cordylidae). Based on Urbani and Bels (1995).
8. Feeding in Lepidosaurs FO/FC (Fig. 8.25) (Smith, 1982; Condon, 1987). In general, the number of chewing cycles is correlated with prey size (e.g., Goosse and Bels, 1992; Andrews and Bertram, 1997) and type (e.g., Andrews and Bertram, 1997; Herrel et ah, 1999b). Andrews and Bertram (1997) showed that the force and mechanical work required to crush prey increase much faster with mass for crickets than for lepidopteran larvae, presumably because of a tougher cuticle in the former. As noted earlier, side switching during chewing is common in scleroglossans, but perhaps not as frequent as in iguanians. Tongue movements during chewing and intraoral transport are similar and described for the latter cycle type (see later). Muscle activity patterns during repeated, cyclical chewing in generalized scleroglossans are available only for the related scincids, Tiliqua rugosa and Corucia zebrata (Herrel et al, 1999b). There is little muscle activity during jaw opening until bilateral bursts in the depressor mandibulae and the spinalis capitis depress the mandible and raise the cranium at FO. These become silent and the jaw closers (adductor mandibulae externus superficialis anterior and posterior, a. m. e. medialis, adductor mandibulae posterior, pseudotemporalis) become active bilaterally and simultaneously active during FC, followed by a second short burst (during SC-PS?). The jaw openers also become active again during SC-PS. Muscle activity patterns change quantitatively, and sometimes qualitatively, with food type. Feeding on hard-shelled snails elicits a specialized, crushing bite in T. rugosa (see later), but most dietary differences relate to differences in duration and recruitment level of the jaw muscles. In Corucia, for example, jaw adductor recruitment and duration are highest when feeding on leaves of endive, which Herrel et al (1999b) related to its "toughness." Although, as these authors argued, the mechanical attributes of different food types undoubtedly affect processing in lizards, leaves and other soft plant parts do not seem to require much reduction and are usually swallowed with little or no processing (e.g., Throckmorton, 1976; see Section VII,E), thus there may be alternative explanations. For example, muscle activity patterns might relate more to the need to crop leafy material so that it can be transported into the pharynx than to its toughness. Leaves often become caught in the corner of the mouth during chewing and transport, causing exaggerated gape profiles, extreme tongue movements, and an increase in the number of cycles (Schwenk and Throckmorton, unpublished results). These cycles might very well be associated with higher recruitment levels and longer duration of adductor muscle activity even though they involve no chewing, per se. EMG patterns have been documented for some spe-
245
cialized types of processing, including inertial feeding bites in Varanus (Smith, 1982) and Tupinambis (McBrayer and White, manuscript) and the process of crushing snail and egg shells in T. rugosa and Heloderma horridum, respectively (Cans et al, 1985; Cans and De Vree, 1986; Herrel et al, 1997b). Smith (1982) showed that EMG patterns in Varanus varied between bites in a feeding sequence, but that variation was primarily limited to whether a muscle was active or silent. McBrayer and White (manuscript) found a similar pattern in Tupinambis. In both taxa, peak adductor activity occurred more or less simultaneously during jaw closure, although Smith (1982) also found significant pterygoideus profundus and limited adductor mandibulae activity during the opening phase in Varanus. She also documented bilateral asymmetry in muscle activity patterns, but unfortunately did not correlate this information with the side on which the animal was biting. In Tupinambis, the number of adductor muscles recruited was modulated according to bite force (McBrayer and White, manuscript). Both studies were conducted using mice, some very large relative to lizard size, hence these patterns might not be representative of more typical feeding bouts on small invertebrates. Shingleback skinks (T. rugosa) feed on snails that are difficult to crush. After a snail has been ingested and positioned between the tooth rows (see earlier discussion), shinglebacks perform a single crushing bite to fracture the shell and liberate its contents (Gans et al, 1985; Gans and De Vree, 1986). Adductor muscle organization indicates that they are most effective at wide gape angles (Wineski and Gans, 1984; Gans et al, 1985). The crushing cycle is of longer duration than other bites and exhibits an unusual pulsatile activity synchronously in all adductor muscles. Gans and colleagues showed that the pulsatile pattern results in muscular tetanus that significantly increases bite force over nontetanic contractions. They also showed that bilateral, tetanic adductor contraction is necessary to crush the shell and that the temporal pattern of force application ensures fracture and crushing of the shell to expel its contents, rather than simple puncture. Bite force is also increased by bracing the mandible against the substrate during the crushing bite. After a crushing bite, most of the shell fragments are pushed out of the mouth by the tongue (Gans et al, 1985). Pulsatile, presuniably tetanic, contractions have also been observed in Gecko eating large insects with heavy cuticles (De Vree, unpublished, in Gans and De Vree, 1986) and in fish crushing seeds and other hard food items between the pharyngeal jaws (Irish, 1983; F. Irish and E. Brainerd, unpublished observations). Irish (1983) suggested the possibility that pulsed adductor activity might induce fatigue damage in hard objects that
246
Kurt Schwenk
would cause them to fail at lower loads than if a single contractile force was applied. Herrel et ah (1997b) observed that egg crushing in the oophagous H. horridum occurs between the palate and the floor of the mouth, not between the marginal tooth rows. This apparently ensures that the liquid contents of the egg remain within the mouth; however, it comes at the cost of significant compliance in the floor of the mouth, reducing the effectiveness of a bite. This occurs partly because crushing takes place with the jaws nearly closed and the adductor muscles outside of their optimal length-tension relationship. Rupture of the egg membrane with preingestion piercing cycles (see earlier discussion) reduces the force necessary to rupture the egg fully so that it can be crushed within the mouth without the benefit of marginal teeth and a wide gape angle. During crushing, the adductor mandibulae extemus superficialis and pseudotemporalis, but not the pterygoideus profundus, are simultaneously active bilaterally. Bite force has been measured directly in only one lepidosaur species, Tupinambis teguixin, a large teiid (McBrayer and White, manuscript). The lizards were induced to bite an apparatus containing a calibrated force transducer, with measured forces ranging from 1.4 to 10.5 newtons. Given the variability in motivation and bite point, this level of variation is not surprising. It is also difficult to relate these results to actual bite forces. Bite forces have also been estimated by modeling the jaw and adductor system: 2.0 newtons in Gecko gekko (Andrews and Bertram, 1997), 25 to 42 newtons in Tiliqua scincoides, 34 to 59 newtons in C. zebmta (Herrel et al, 1998a), 0.58 to 2.24 newtons in Podarcis (Herrel et al, 1996b), and 31 to 87 newtons in H. horridum (Herrel et al, 1997b). Andrews and Bertram (1997) modeled bite force at the tip of the mandible and apparently on one side only, whereas Herrel and colleagues calculated forces under several different conditions. Herrel et al (1997b) showed that the calculated bite force in H. horridum was always greater than the average force needed to crush an eggs. Bite forces indicated for Tupinambis and Gecko are probably underestimates because only bites at the tips of the jaws were considered, as opposed to typical chewing or crushing bites farther back on the tooth row, which incur greater mechanical advantage. McBrayer and White (manuscript) noted that mouse skulls were "audibly crushed" in the jaws of Tupinambis. Given that Herrel et al (1997b) found that it took an average of 27 newtons to crush a quail egg and that a mouse skull is probably stronger than an egg, it is likely that these large lizards are capable of producing larger bite forces than measured. Nonetheless, Andrews and Bertram (1997) found that 2 newtons was precisely the amount of force needed to crush a large cricket to full occlusion.
Andrews and Bertram (1997) suggested that the mechanical attributes of different prey types might determine how a prey item is processed. They found that the work invested in processing a putatively difficult prey type (cricket) changed with prey mass, but did not change with increasing size of soft-bodied larvae. The energetically costly crickets were chewed by a gecko at a slower rate than larvae. They speculated that anaerobic glycogen depletion in the adductor muscles could represent an organ-level constraint on the types of prey chosen and the way in which they are processed. 3. Intraoral
Transport
Functional data for intraoral transport in scleroglossans are available for the following taxa: Gekkota: Eublepharus (Delheusy et al, 1995); Lialis (Patchell and Shine, 1986a); Scincomorpha: Corucia (Herrel et al, 1999b); Lacerta (Goosse and Bels, 1992; Urbani and Bels, 1995); Tiliqua (Herrel et al, 1999b); Tupinambis (Smith, 1984); Zonosaurus (Urbani and Bels, 1995); Anguimorpha: Heloderma (Herrel et al, 1997b); and Varanus (Smith, 1982,1986; Condon, 1987). Perhaps the most striking feature of hyolingual transport in scleroglossans is how similar it is to iguanians, despite radically different tongue and hyobranchial form (Smith, 1984; Urbani and Bels, 1995; Herrel et al, 1999b). As noted previously, this result is not surprising given that the mechanical roles of the jaws, tongue, and hyobranchium during transport are comparable in both groups. It follows that the kinematics will be similar. Most scleroglossan lingual modifications are unrelated to postingestion feeding, and the tongue retains its role in moving the bolus back into the pharynx. Thus, all scleroglossans except varanids retain a papillose "frictional surface" (McDowell, 1972) on all or most of the tongue that serves to interlock with the bolus during hyolingual transport. Despite similarities, intraoral transport in scleroglossans differs from iguanians in subtle, but noteworthy ways. The differences seem to relate to modifications of the foretongue associated with enhanced elongation and protrusion, primarily for chemosensory tongue flicking (see Section VII). During cyclical transport cycles the coupling of the tongue and the hyobranchium is somewhat looser than in iguanians (e.g.. Smith, 1984). Anteroposterior tongue excursions can be more extensive than the hyobranchium and presumably result from tongue sliding on the lingual process, intrinsic elongation, or both. In all scleroglossans the foretongue, and sometimes the hindtongue, is narrow side to side, and/or thin dorsoventrally (spatulate), and lacks the deep, high-profile papillae of iguanians. The papillary surface is often flat and hard (Figs. 8.15 and 8.16). Therefore, the tongue's ability to
8. Feeding in Lepidosaurs form itself to the bolus during retraction is compromised. Some scleroglossans overcome this difficulty by resorting to dramatic contortions of the foretongue to create an insurmountable elevation in front of the bolus that pushes the bolus back. Sometimes this takes the form of ribbon-like rippling in the foretongue surface. We have observed this behavior in Tiliqua, Tupinambis, Diploglossus, and Heloderma (Schwenk, unpublished results; Schwenk and Throckmorton, unpublished results), although it is probably quite widespread. In general, length and conformation changes in the tongue, especially the foretongue, seem to be more pronounced during transport and other feeding stages in scleroglossans. In most taxa the gape cycle during transport is typical and conforms well to the model cycle (Fig. 8.24). Both SO I and II are usually, but not always, present, as is an SC phase. Hyobranchial retraction begins at the start of FO, as expected. The peak in gape afforded by FO allows disengagement of the bolus with the teeth or palate. Thus, relatively high maximum gapes are maintained during transport cycles, although these tend to decline as the bolus is moved into the pharynx and packing cycles begin (e.g., Goosse and Bels, 1992). The number of transport cycles (e.g., Urbani and Bels, 1995; Herrel et al., 1999b), as well as their kinematics and motor patterns (Herrel et al, 1999b), are affected by food type. Cycle number undoubtedly relates to bolus size and shape following reduction, which is a direct consequence of food type, but it might also be affected by surface qualities of the food that affect the tongue's ability to move it. Herrel et al. (1999b) provided EMG data for "intraoral transport" in Tiliqua and Corucia, but their data do not distinguish between chewing and pure transport cycles and probably represent primarily chewing. These data are discussed in the previous section. However, the combination of transport function with chewing in these lizards makes the results relevant here. Based on their description, it is clear that chewing cycles grade into pure transport cycles and these, in turn, grade into pharyngeal packing cycles (Fig. 8.24). Among lepidosaurs, only varanids are obligate inertial feeders. The use of inertial transport is correlated with reduction of the tongue, freeing of its anterior end, and the loss of its frictional surface (e.g., McDowell, 1972; Smith, 1986). A causal link between the two is suggested by the fact that large teiids show a parallel trend in tongue form and frequently use inertial transport. However, these species retain lingual papillae and often use hyolingual transport as well (e.g., MacLean, 1974; Smith, 1984). Among the teiids, Tupinambis may be the most committed to inertial transport (the large taxon Callipistes is another possible candidate). According to MacLean (1974:196), "the long narrow tongue
247
often hangs limply out of the angle of the mouth during feeding." However, Smith (1984) observed that hyolingual transport cycles usually followed after a few inertial cycles. Our films of Tupinambis show that both inertial and hyolingual transport cycles are common (Schwenk and Throckmorton, unpublished results). Other large teiids, such as Dracaena, rely primarily on hyolingual transport (see later), but a small, microteiid (Gymnopthalmid) species (Neusticurus) was observed to use inertial feeding when feeding on small prey items (MacLean, 1974). It is very likely that the balance between inertial and hyolingual transport in these species is determined by prey size and type, but these factors have not been controlled in studies thus far. Scleroglossans, as a group, seem to employ inertial transport much more frequently than iguanians, possibly for the same reasons that they are more likely to use pharyngeal compression (see later), but the lack of controlled, comparative data make this speculative. The kinematics of inertial transport were described earlier (see Section V,B), but a few details can be added. In Varanus, prey capture with the jaws is followed by a series of manipulative or repositioning bites, killing or crushing bites, head lifts, inertial thrusts and "toothclearing" cycles (Smith, 1982, 1986; Condon, 1987). Five to 12 thrusts are sufficient to move a large prey item into the pharynx (Smith, 1986). Inertial thrusts are characterized by relatively simple gape cycles that look like simple spikes with intervening periods of gradual SO (Fig. 8.25) (Smith, 1982; Condon, 1987). Maximum gape ranges from 18° to 25° (Smith, 1982). Head movement is more or less up and back during jaw opening and down and forward during jaw closing. During the entire process, the snout remains ventroflexed at the mesokinetic axis relative to the rest position (Condon, 1987). However, many inertial transport cycles and some manipulative cycles are associated with brief moments of dorsoflexion, but never beyond the rest point. Condon (1987) showed convincingly that these are active upper jaw movements and not merely the result of prey impact. The snout gradually resets to the rest position during swallowing, following transport. Streptostylic movements of the quadrates also accompany inertial transport (Smith, 1982). These usually involve slow protraction and rapid retraction associated with each inertial thrust, but the pattern and timing of movement relative to the gape cycle vary among cycles. Smith (1982) found that adductor muscle activity patterns are far more stereotyped during inertial transport cycles than during biting. In addition, inertial thrusts are unusual in having the jaw adductors (adductor mandibulae externus, pterygoideus, pseudotemporalis) active during both jaw opening and jaw closing. Smith (1982) suggested that this activity
248
Kurt Schwenk
pattern permits an extremely rapid (<5 msec) reversal from jaw opening to jaw closing during inertial thrusts, possibly to ensure that active prey do not escape. An alternative, entirely speculative explanation is that adductor activity during jaw opening is associated with mesokinetic movements of the snout. Although the constrictor internus dorsalis muscles dorsoflex the snout, the adductors are the principal ventroflexors (see Section IV,D). It may be necessary to keep the two sets of muscles in tension to coordinate the rapid, small-scale movements of the upper jaw during inertial thrusts. Hyolingual muscles are also active during inertial thrusts, including simultaneous activity of hyobranchial protractors and retractors (Smith, 1986). Smith argued that such simultaneous contraction would function either to open the gular passage or to depress the lower jaw, depending on which portion of the mandibulohyoideus (a protractor) is active. The ability of Varanus to recruit axial and hypoglossal muscles for jaw opening might help the relatively feeble depressor mandibulae in overcoming the antagonistic forces of the adductors during jaw opening. A novel form of transport has been described in the large teiid, Dracaena guianensis, which specializes on hard-shelled molluscs. According to Conant (1955, in Dalrymple, 1979), after jaw prehension of a snail, the head is elevated while the jaws open and the snail rolls to the back of the tooth row where it is more effectively crushed between the massive, molariform teeth. However, lingual movements are used to manipulate the soft parts of the snail and to expel pieces of shell once it is crushed (MacLean, 1974; Dalrymple, 1979). 4.
Swallowing
a. Pharyngeal Packing Pharyngeal packing has been described under various names in the following taxa: Gekkonidae: Euhlepharus (Delheusy et ah, 1995); Scincidae: Corucia and Tiliqua (Herrel ei ah, 1999b); Cordylidae: Zonosaurus (Urbani and Bels, 1995); Lacertidae: Lacerta (Goosse and Bels, 1992; Urbani and Bels, 1995); Teiidae: Tupinambis (Smith, 1984); Helodermatidae: Heloderma (Herrel et al, 1997b); and Varanidae: Varanus (Smith, 1986). In scleroglossans that use hyolingual movements for transport, pharyngeal packing cycles are very similar to those described for iguanians. Packing cycles typically have small amplitude gapes (6° to 8°) with a simplified profile (Fig. 8.24). The gape profile lacks a divided SO phase and an FO phase is either not evident or reduced to a low spike at the end of a gradual or rounded SO. Tongue protrusion distance is variable and often involves "licking" movements across the labial scales. Observations of this behavior in Heloderma
suspectum suggest that the tongue twists and contorts dramatically along its length during the protrusion/ licking phase, probably to position the posterior limbs relative to the bolus (personal observation) The relationship between tongue movement and the gape cycle is unknown or difficult to discern in most studies. In the scincids Corucia and Tiliqua, tongue protrusion and retraction are exactly coincident with jaw opening and closing, respectively (e.g., Herrel et al, 1999b). The number of packing cycles necessary ranges from a few to many and obviously depends on the size, position, and condition of the bolus. It may also depend on whether pharyngeal compression follows. Given the lack of food contact with the teeth during packing cycles, it is not surprising that muscle recruitment is much lower than during transport/chewing cycles, although activity patterns are roughly similar [Herrel et ah (1999b) for the scincids Tiliqua and Corucia]. Both superficial and deep parts of the pterygoideus have a long, low-amplitude burst of activity early in jaw opening, followed by the depressor mandibulae and spinalis capitis. In accordance with reduced gape amplitudes, the latter muscle is active at lower levels than during previous stages and ceases activity early. During jaw closing the adductor mandibulae externus complex (mainly superficialis) becomes active along with the pterygoideus (mainly profundus). The pseudotemporalis and adductor mandibulae posterior are silent. Smith (1984) noted that, in contrast to transport cycles, posthyoid muscles are active during the initial stages of jaw opening. Smith suggested that this activity pulls the tongue and hyoid up and back in this phase, but according to most accounts the tongue is moving forward at this time. It may be that the foretongue continues forward while the hindtongue moves up and back. More cineradiographic data are needed to elucidate tongue movements during this stage. In scleroglossans with reduced tongues, pharyngeal packing is accomplished with little or no lingual involvement {Tupinambis and Varanus, respectively; Smith, 1984, 1986). In Tupinambis, packing cycles are characterized by large dorsoventral excursions of the hyobranchium. Smith (1984) suggested that tamping of the bolus by the posterior end of the tongue is aided by elevation of the floor of the mouth once the bolus is posterior to the basihyal. Musculoskeletal modifications evident in the hyobranchium might enhance this function. In Varanus, the hyobranchium has entirely taken over the role of the tongue in pharyngeal packing (Smith, 1986). During this phase, cyclical movements of the hyobranchium move the bolus from the rear of the oral cavity into the gular area. These movements are accompanied by small inertial thrusts and, in later cycles, by tongue protrusion and "lip licking." The lat-
8. Feeding in Lepidosaurs ter observation is very interesting given that the protrusion phase is interpreted as a mechanism for positioning the tongue in front of the bolus in order to push it back. If the hyobranchium has taken over this function mechanically, why does the tongue protrude at all? It suggests that kinematic and muscle activity patterns are conserved in this derived taxon, despite loss of the tongue's direct function in packing. As such, tongue protrusion represents an epiphenomenon associated with the mechanics of moving the hyobranchium, which, ancestrally, were coupled. b. Pharyngeal Compression Pharyngeal compression has been noted more frequently in scleroglossans than in iguanians, but still not very often: Pygopodidae: Lialis (Patchell and Shine, 1986a); Scincidae: Corucia, and Tiliqua (Herrel et al, 1999b); Teiidae: Tupinambis (Smith, 1984); Helodermatidae: Heloderma (Herrel et al, 1997b); and Varanidae: Varanus (Smith, 1986). With his usual clarity. Fitch (1954:118-119) described pharyngeal compression in a small skink (Eumeces) feeding on a large cricket in the wild: "Then, although the cricket was of body diameter almost as great as the skink itself, the lizard swallowed it head first, engulfing it with violent gulping movements. After the front end of the prey had entered the gullet, muscles of the throat and neck were brought into play in forcing it farther down. Swallowing movements were snake-like, the lizard turning its head at right angles to the body to squeeze the morsel down." According to Smith (1984, 1986), pharyngeal compression moves the food from the gular region (posterior pharynx) into the esophagus where peristalsis takes over {Tupinambis and Varanus), but according to Herrel et al. (1999b), compression cycles "always occur after swallowing, i.e., after passage of food from the oral cavity to the esophagus" {Tiliqua and Corucia). As discussed previously (Section V,B), there is bound to be variation among species and among feeding bouts in the degree to which food is moved into the esophagus during pharyngeal packing. However, if the bolus is fully within the esophagus, as suggested by Herrel et al. (1999b), peristalsis should be all that is necessary to move it to the gut, suggesting that some of the bolus remains within the pharynx and that compression is necessary to make the final push. This issue needs to be addressed with additional cineradiographic data. Pharyngeal compression can take the form of pharyngeal constriction, neck bending (which can be dorsoventral or lateral), or both. Fitch's (1954) description of Eumeces exemplifies lateral neck bending associated in this species with consumption of large prey. In other scincids {Corucia and Tiliqua), pharyngeal com-
249
pression is evident as head/neck extension concomitant with pharyngeal constriction, presumably by constrictor colli contraction (Herrel et al, 1999b). In these species there are one or two compression cycles after pharyngeal packing and each of these is preceded by a very small open-close gape cycle. Jaw openers and closers are mildly active during the initial gape cycle, but the compressive phase is accompanied by prolonged and intense activity in the jaw openers, which may be repeated and followed by a short burst in the jaw adductors. Constrictor colli and intermandibularis muscles were not recorded. Activity of the jaw openers during pharyngeal compression (with the mouth closed) is difficult to interpret, but may be related to an antagonistic, stiffening, or bracing function. Herrel et al (1997b) noted several compression cycles in H. horridum following egg eating, which they described as "aberrant" without explanation. Pharyngeal compression occurs facultatively in the related H. suspectum and can include both lateral neck bends and pharyngeal constriction (Schwenk, unpublished observations). Patchell and Shine (1986) observed contractions of the throat musculature during swallowing in Lialis. In inertially feeding taxa such as Tupinambis and Varanus, pharyngeal compression is a distinct and sometimes dramatic stage following packing (Smith, 1984, 1986). In both taxa the hyobranchium is drawn anteriorly and then dorsally and posteriorly, constricting the pharynx and forcing the bolus from the pharynx into the esophagus. The tongue remains more or less in place and moves along with the hyobranchium. Compression is more pronounced in Varanus than in Tupinambis and is accompanied by side-to-side neck bending (Smith, 1986).
F. Biomechanics of Lingual Prey Capture 1. Mechanism of Tongue Protrusion Only a small portion of the iguanian foretongue is free from attachment to the floor of the mouth. The attachment is by means of the genioglossus muscles, anterior fibers of which turn within the tongue and run anteriorly toward the tip. Nonetheless, extraoral tongue protrusion is possible. In fact, two kinematically distinct forms of protrusion are exhibited: tip first (as during chemosensory tongue flicking, lapping, and "lip licking") and ventrally curled, dorsum first (as during lingual ingestion). Although iguanians do not extend the tongue as far nor move it as complexly as most scleroglossans, they are nonetheless capable of significant extraoral excursions during tip-first protrusion [e.g., see Fig. 1 in Schwenk (1995b) for tongue
250
Kurt Schwenk
flicking in an iguanid]. During these behaviors the foretongue is apparently unconstrained by its ventral connections. This is possible for two reasons: (1) the foretongue is capable of some hydrostatic lengthening so that it elongates anterior to the point of attachment and (2) anterior genioglossus fibers are elastic and stretch during tip-first extension. Tip-first protrusion during lapping and tongue flicking is effected with little or no hyobranchial protraction (Smith, 1984; Schwenk, unpublished observations). In contrast, during lingual ingestion the tongue tip is ventrally curled and fixed to the mandible as the remainder of the tongue is protruded over it, primarily by hyobranchial protraction (see earlier discussion). This tongue conformation is probably caused by an absence of hydrostatic elongation and activity of anterior genioglossus fibers, which in the contracted state are inelastic and serve to tether the tongue tip to the mandible. Thus, as the tongue is pushed out of the mouth by hyobranchial protraction, it is forced to "roll" around the symphysis, presenting its dorsal, papillose surface toward the prey item (Figs. 8.18, 8.19, and 8.21). High-speed films of Phrynosotna and other iguanians clearly show the nature of this genioglossus tether during protrusion (Schwenk, unpublished observations). Although the genioglossus is assumed to be a tongue protractor, with the tongue in a protruded conformation, anterior fibers are positioned to act as retractors of the foretongue. The tongue in Sphenodon is attached to the floor of the mouth for nearly its entire length, leading Gabe and Saint Girons (1964) to suggest that it is nonprotrusible. However, they did not anticipate the kind of whole-tongue protrusion characteristic of lingual prehension. Only the very tip of the tongue is free and it is this bit that is seen to extend tip first at the start of an ingestion sequence, in contrast to iguanians (see Section V,C). As the tongue passes over the mandible, however, the tip is pulled downward and protrusion continues as described earlier. The only other time the tongue is protruded tip first is during pharyngeal packing in which the anteriormost end of the tongue is passed between upper and lower teeth (see earlier discussion). Tuatara do not tongue flick (Schwenk, 1986) and, to my knowledge, lapping behavior has not been described. Circular muscles around the hyoglossus bundles are incomplete and anterior genioglossus fibers form a thick bundle that run to the very tip of the tongue (Schwenk, 1986). In sum, the tongue tip is more fully tethered to the mandible than in iguanians and there is no evidence that the foretongue is, or can be, lengthened hydrostatically. Thus, it is unlikely that Sphenodon is capable of significant extraoral, tip-first protrusion.
It appears that tip-first and dorsum-first tongue protrusion represent two, fundamentally different motor patterns (see also Herrel et al, 1998c). In the former, the hyobranchium participates little in tongue protrusion, the foretongue is lengthened hydrostatically, and anterior genioglossus fibers are relaxed; in the latter, protrusion is effected by hyobranchial protraction and (putatively) contraction of anterior genioglossus fibers. The integrated nature of these behaviors and their phylogenetic distribution has interesting evolutionary implications (Section VII). Here it is sufficient to note that lingual ingestion employs a distinct type of tongue protrusion that has implications for the biomechanics of prehension. 2. Mechanism of Adhesion The force of lingual prehension has been measured only in chameleons (see Section VI,A), but an anecdote serves to illustrate the efficacy of lingual adhesion in a generalized iguanian. While filming a tiny agamid (Phrynocephalus helioscopus) in my laboratory, problems with depth of field made it necessary to restrict feeding sequences to a very small zone. In one attempt I glued a mealworm to the substrate to control the point of ingestion, fully expecting it to break free at the moment of prehension. Instead, when the lizard struck the mealworm with the tongue and attempted to retract it, the mealworm remained fixed in place and the lizard was pulled toward it rapidly. As the tongue continued to retract, the lizard became completely airborne and "flipped" over the mealworm onto its back! Only then did the tongue break free of the mealworm. I interpret this incident to indicate that this small lizard is capable of generating a pulling force with a small area of its tongue equivalent to its own body weight (approximately 4.5 g)—probably much more if one considers the frictional and inertial forces as well. Lingual adhesion is typically attributed to the "stickiness" of mucus coating the tongue. This facile explanation undoubtedly oversimplifies the actual mechanism about which some reasonable inferences are possible. Emerson and Diehl (1980) listed five potential types of adhesion in reference to toe-pad sticking in treefrogs: interlocking, friction, dry adhesion, suction, and wet adhesion (including both stefan adhesion and capillarity, or surface tension). Friction and dry adhesion can be eliminated on theoretical grounds at the outset (see Emerson and Diehl, 1980). There is no evidence of suction except in chameleons (see later). The filamentous form of the papillae, the elaboration of their epithelial apices, and the fact that the prey item is pushed into the papillae suggest that interlocking is important. The high density of mucocytes in the con-
8. Feeding in Lepidosaurs tact zone, the obvious coating of the tongue with their serous and mucous secretions, and the ability of many iguanians to apprehend very smooth food items (waxy leaves and featureless insect larvae with smooth cuticles) suggest that wet adhesion is also important. Thus, there is strong evidence that a combination of interlocking and wet adhesion is responsible for tongueprey adhesion (Bramble and Wake, 1985). Theoretical and circumstantial evidence points to contributions of both stefan adhesion and capillarity in the mechanism of wet adhesion, with the latter predominant. Capillarity depends on a thin layer of fluid between tongue and prey surfaces. The adhesive force stems from the surface tension of the meniscus at the edge of the contact area. If the intervening fluid extends beyond the contact area, a meniscus will not form and capillarity cannot function. Thus Emerson and Diehl (1980) found that frogs sticking to a glass plate immediately separated when placed in water (removing the meniscus). Similarly, Green (1981) found that by reducing surface tension with detergent added to the fluid layer, the bond was broken. In a typical case of lingual prehension, the tongue's fluid coat would form a meniscus against the surface of the prey item at the edge of the lingual contact zone. However, when feeding on food items with a wet surface, a meniscus cannot form. When horned lizards {Phrynosoma) were fed earthworms instead of crickets or mealworms, their capture rate dropped precipitously—the tongue simply would not stick to the worms (Schwenk, unpublished observations). Marx (1955) similarly observed that chameleons were unable to apprehend earthworms with the tongue. The implication is that the fluid layer between tongue and prey surface in the contact zone was continuous with the fluid covering the worm, hence no meniscus could form and therefore no adhesion, supporting the hypothesis of capillarity. Earthworms also lack surface features for interlocking, but other featureless but dry prey items, such as mealworms, are apprehended without difficulty. The viscoelasticity of mucus may play an important role in lingual adhesion. The biomechanics of lepidosaurian lingual secretions are utterly unknown, but we do know that they are both serous and mucous (see Section IV,E). Serous fluids are more "watery" and presumably spread more rapidly ("wetting") than mucous secretions. However, the latter are more viscous and potentially "stickier." Mucus, in general, has unusual, viscoelastic properties, i.e., sometimes it behaves like a liquid, sometimes more like an elastic solid, depending on the conditions under which it is strained (e.g., Wainwright et ah, 1976; Alexander, 1983; Denny, 1984; Vincent, 1990). Most important, its characteristics, including its "stickiness," change depend-
251
ing on the way it is strained. Imagine two surfaces with a thin layer of fluid between them. If one plate is pulled directly away from the other, placing the fluid in tension, it will be very hard to move, whereas if one plate is slid along the other, shearing the fluid, then it will move more easily. With a more viscous fluid such as mucus, the adhesive force is greater, particularly in tension (Alexander, 1992). Second, if mucus is strained slowly and persistently, it flows and behaves like a liquid (i.e., it is not "sticky"). However, if it is strained rapidly and briefly, it becomes "stiff," behaving like an elastic solid (i.e., "sticky") (e.g., Alexander, 1983; Denny, 1984). In sum, the adhesive bond between two surfaces (tongue and prey) with an intervening layer of mucus will be very strong if they are pulled apart rapidly, but relatively weak if they are pulled apart slowly or sheared. The rate/time-dependent components of this adhesion are independent of meniscus formation and thus do not reflect capillarity. Their dependence on fluid viscosity and viscoelasticity suggests a possible role for stefan adhesion. 3. A Model of Lingual Prehension It is now possible to propose a general model of lingual prehension that incorporates facts about the anatomy of the tongue, the kinematics of ingestion, and the theory of tongue-prey adhesion outlined in the previous section. Components of the model are summarized in Fig. 8.28. Work in our laboratory is endeavoring to test the model by verifying its predictions. 1. Compared to jaw prehension, lingual ingestion might theoretically facilitate capture of active prey in two subtle ways: (i) The speed of capture is increased because the velocity of tongue protrusion is summed with the velocity of the jaws' approach (whole body plus head rotation) toward the prey item. The extent and speed of tongue protrusion might be further augmented by simultaneous streptostylic protraction of the quadrates, (ii) The likelihood of prey escape is decreased because the prey animal is deceived into attempting escape too late. The prey estimates its distance to the lizard based on the looming form of the head (especially the eyes) and is alerted to escape accordingly. However, the tongue (and therefore the point of prehension) moves in advance of the head and arrives at the prey item a few milliseconds before the jaws. Those few milliseconds might be enough to allow prehension before the prey escapes. Thus, the tongue not only moves faster toward the prey item than the jaws, but prey escape is slightly delayed, allowing more time for capture. 2. When the tongue is protruded during an ingestion cycle, its dorsal surface is arched, as described
252
Kurt Schwenk
B
D F I G U R E 8.28. A model of lingual prehension in iguanian lizards. See text for explanation. Based on films of Dipsosaurus dorsalis (Iguanidae).
earlier. The muscular corpus of the tongue takes on a stiff, cylindrical form while the overlying mucosal surface of high-profile papillae is curled around its apex. Protraction and sliding of the papillary surface around the end of the protruded tongue may, to some extent, be independent of whole-tongue protraction (see earlier discussion). This action not only presents the dorsal surface of the tongue anteriorly, but stretches it, placing it in tension so that the papillae of the contact zone are erected and pointed directly toward the prey item (Fig. 8.28A). 3. The contact zone corresponds approximately to the anterior third of the resting tongue, exclusive of the notched tip. Papillae in this region are the longest and thinnest on the tongue and, except for the very tip, are covered by a dense layer of mucocytes (Figs. 8.12 and 8.13). The hexagonal shape of these cells in tangential section indicates that they are packed as tightly as possible (Schwenk, 1986, unpublished observations). Plumose cells of the papillary apices, when present, are most elaborated here (Figs. 8.13A and 8.14B). Each papilla contains numerous collagenous fibers that run its length.
4. At tongue-prey impact, (Fig. 8.28B), several things happen: (i) Papillae contact the prey item apex first and crumple. This absorbs impact energy, making the prey item less likely to bounce away, (ii) Compression and bending of the papillae causes expression of mucous and serous secretions in the contact zone, (iii) In most cases the tongue continues to move toward the prey item, pinning it against the substrate. The prey item is pushed deeper into the papillae of the impact zone, "forming" the tongue around the prey item and maximizing the surface area of tongue-prey contact (Figs. 8.21A and 8.21B). A median sulcus in the contact zone may also increase the number of papillae and the total surface area of the tongue that contact the prey surface. A sulcus is anatomically fixed in Sphenodon (Schwenk, 1986) and appears facultatively in some iguanians during ingestion (e.g., Delheusy and Bels, 1992). Interlocking with the prey surface occurs on three structural levels: papillae become entangled in any extremities (e.g., limbs, antennae, hairs), papillary apices interdigitate with prey surface features, and individual plumose cells interlock with the prey surface at an even finer scale. At the end of SO II the tongue-prey bond is formed. Gravity plays no role in this process—all forces holding the prey on the tongue stem from its adhesive bond. Several factors now come into play. For prey items with extremities and surface irregularities (most insects), significant interlocking will occur. However, interlocking will be less effective when the retracting force is normal to the prey surface. The efficacy of interlocking then will depend on the degree of entanglement (as opposed to simple interdigitation) between lingual and prey surfaces. Filamentous papillae and plumose cells are likely to provide a high level of entanglement which at the cellular level (e.g., plumose cell vs insect hair) might provide ample adhesive force even when the tongue is pulled directly away from the prey surface. A possible mechanical analogy for the interlocking of tongue and prey surfaces is the hook-andloop design of Velcro. Wet adhesion also contributes to the tongue-prey bond. On the smooth and sometimes waxy surfaces of leaves, fruit, and some insect cuticles, interlocking will not be effective and wet adhesion alone must suffice. The serous component of lingual mucocyte secretion, characteristic of iguanian lizards (see Section IV,E), might be important here. The watery serous fluid would flow better across relatively high-energy, nonpolar surfaces such as insect and plant cuticle so that the contact zone is wetted more rapidly. The adhesive force of capillarity depends on the formation of a meniscus, which could occur at two levels: around the perimeter of the contact zone, as a whole, and/or
253
8. Feeding in Lepidosaurs around multiple, individual points of tongue-prey contact, each potentially as small as the apex of a single papilla. The contribution of stefan adhesion will increase with increasing viscosity of the lingual secretions. Although serous secretions are always present and sometimes predominate in the contact zone of iguanians, the foretongue is also bathed in more viscous mucous fluids secreted by the hindtongue and the sublingual salivary glands. The viscoelastic qualities of mucus may be critically important during retraction (next). 5. The tongue and adherent prey are retracted with remarkable rapidity during FO-FC—the entire retraction phase lasts less than 20 msec in some cases (Schwenk and Throckmorton, 1989; Delheusy and Bels, 1992). Thus, the tongue accelerates a prey item from a standstill to high velocity within a few milliseconds and the adhesive bond is stressed almost instantaneously by inertial reaction forces tending to separate the prey from the tongue (Fig. 8.28C). As the tongue is retracted, its papillary surface is uncurled, flattened, and withdrawn into the mouth (Figs. 8.21C and 8.21D). Although the prey item starts by being suspended from the apex of the tongue, it ends with the flat surface of the tongue beneath it. During this reorientation, inertial reaction forces tending to dislodge the prey item shift from normal to the tongue's surface to tangential. This shift would seem to promote failure of the adhesive bond because, as noted earlier, the fluid layer responsible for wet adhesion is more likely to fail in shear than in tension (i.e., less stress is required to cause separation). However, shear stresses may, in fact, be negligible. Because of the flexible, filamentous form of the papillae, they instantly align themselves with the inertial stresses tending to separate the prey from the tongue, no matter what the position of the tongue or the prey item during the retraction phase. As such, the multiple contact points between tongue and prey surfaces remain in constant tension. In effect, the papillary system is a mechanism for translating shear into tensile stress. Finally, the speed and brevity of retraction suggest that during this critical phase of prey capture the lingual mucus becomes stiff, or sticky, due to its viscoelastic nature. The rapidity of retraction might actually enhance, rather than weaken, the adhesion of the prey to the tongue. 6. Given the apparent efficacy of the adhesive mechanism, a second mechanical challenge is confronted once the prey item is held within the mouth on the surface of the tongue—how is it removed? Delheusy and Bels (1992) suggested that an initial bite following retraction of the prey into the mouth serves to free it from the adhesive bond. In addition, three factors should
mitigate adhesion at this point: (i) When the tongue is moved relative to a fixed prey item, as during SO of a chewing or transport cycle, it moves slowly. The prey item is held against the teeth or palate and the tongue is protracted and elongated beneath it. At low strain rates the mucus should behave as a liquid lubricant rather than as a stiff glue. The relative slowness of tongue protraction during SO is frequently noted in the literature, (ii) When held in the mouth and in most cases chewed, the surface of the bolus becomes thoroughly wetted. A continuous coating of fluid across the bolus surface will prevent the formation of a meniscus, disrupting capillarity. Capillarity may be the single most important component of lingual adhesion, hence this may be an especially important factor, (iii) Interlocking is promoted during capture by forcibly pinning the prey item against the substrate. During intraoral support, the degree of interlocking can be controlled by the extent to which the prey item is pressed against the palate. Finally, when the bolus needs to be moved posteriorly by the tongue during intraoral transport, tongue retraction is rapid so that the lingual mucus would once again behave as a glue to hold the bolus on the tongue's surface. In most cases, however, the tongue is deformed to cup or push the bolus posteriorly so that significant adhesion would not be necessary (Fig. 8.28D). 4. Comparative Implications
of the Model
The niodel posits passive orientation of filamentous papillae in parallel with inertial reaction forces to maintain tensile stresses and mitigate shear throughout retraction of the prey item on the tongue. This aspect of the model is based primarily on the morphology of the tongue in Sphenodon and iguanid lizards, which possess filamentous papillae. Anoline iguanids and agamids have reticular papillae in the contact zone, although in many species these grade into filamentous papillae posteriorly. Reticular papillae are joined at points along their lengths so that they could not individually bend and change direction to align with separation stresses. Although the feltwork pattern of papillary interconnections and their intrinsic collagen fibers suggests that stresses in any direction are easily accommodated, the partially continuous tongue surface should be more subject to shear separation than a surface of filamentous papillae. Nonetheless, interlocking is maintained by peg-like extensions of papillary apices at the surface and in many species by plumose cells in the contact zone (Herrel et ah, 1998c; Schwenk, unpublished observations). It is conceivable that the free-floating plumose cells could perform a similar function of stress translation at the
254
Kurt Schwenk
cellular level. Agamids seem as effective as iguanids in lingual prey prehension, but controlled tests of performance have never been made, thus it is impossible to assess the impact, if any, of these morphological differences on tongue-prey adhesion. The model implies that prehensile performance will vary according to food and lingual surface features. The case of wet prey surfaces was discussed earlier. Herrel and De Vree (1999a) noted that Uromastix requires more attempts to capture a locust than Agama (Plocederma). They attributed this performance difference to the presence of plumose papillae in the contact zone of the latter and their absence in the former. This interpretation is consistent with an interlocking function of plumose cells with insect prey surfaces. Uromastix is herbivorous and feeds almost entirely on plant matter in the wild (citations in Herrel and De Vree, 1999a). The lack of plumose papillae and a generally more featureless tongue surface might facilitate wet adhesion by capillarity on smooth or waxy plant surfaces, but limit the effectiveness of adhesion when feeding on insects because the lack of interlocking would make shear failure more likely during retraction. 5. Lingual Prehension in Scleroglossans According to the model, scleroglossan lingual prehension faces several challenges: (1) The foretongue is relatively freer of the hyobranchium and genioglossus attachment and is therefore more weakly supported and apparently less well suited to assume the necessary conformation during protrusion. (2) The foretongue is usually narrower and has less surface area available for the contact zone. (3) The tongue lacks high-profile papillae for (i) shock absorption—prey are more likely to bounce off the tongue; (ii) fitting of the tongue surface to the prey item to maximize area of contact; (iii) interlocking—no entanglement is possible; and (iv) stress translation—prey can be more easily sheared from the surface during rapid retraction. (4) The foretongue lacks lingual epithelial glands, and papillary surfaces are devoid of surface elaborations and often keratinized—wetting and interlocking capabilities are reduced. The model, therefore, predicts that scleroglossans should be less effective at lingual prehension than iguanians. This is supported indirectly by the fact that among thousands of scleroglossan species, only a few are known occasionally to employ any form of lingual prehension. In two of three documented cases, the prediction is directly supported {Zonosaurus and T. rugosa; see earlier discussion)—the tongue is used only to hold or stabilize the prey item on the substrate while the jaws are used for the actual prehension. Only one
species (T. scincoides) is known to exhibit effective lingual prehension kinematically similar to iguanians (Smith et ah, 1999; see earlier discussion). However, the crickets used in the study were very light relative to lizard mass and were not a strong test of prehensile performance. Indeed, when related T. rugosa uses its tongue (morphologically almost identical to T. scincoides; Schwenk, unpublished observations) to pull heavier, smooth-surfaced snails into the mouth, it does so by drawing them slowly a short distance across the substrate and closing the jaws around them, thus implying a weak adhesive bond. Tiliqua belongs to a putative clade of skinks known as the "Egernia group," also including genera Cyclodomorphus, Egernia, and Corucia (Greer, 1979,1989; Shea, 1990). These skinks share an unusual tongue morphology that might make them more effective at lingual prehension than other scleroglossans (Schwenk, unpublished results; I have not observed Cyclodomorphus, but these species are sometimes placed in Tiliqua and are likely to be similar): (1) large absolute size correlated with large body size (Cyclodomorphus species tend toward a gracile body form and more slender head); (2) unusually broad foretongue; (3) genioglossus lateralis fibers that turn as they enter the side of the tongue and run anteriorly toward the tip as longitudinal bundles lateral to the hyoglossus (shared with other skinks; Schwenk, 1988); and (4) bizarre lingual defensive displays involving erection and protrusion of the tongue with wide gape in most Tiliqua and some Cyclodomorphus (Greer, 1989); Corucia do something similar (personal observation). Large size and a broad foretongue imply that a large surface area can be brought to bear during lingual prehension. The lingual defense behavior is difficult to interpret except that it suggests both unusual motor patterns and the ability to extend and support the tongue beyond the mandibular symphysis using a mechanism that does not appear to involve hydrostatic elongation. As such, it might be mediated by hyobranchial protraction and support and is therefore more similar to iguanian ingestion than to tongue flicking or lapping (see earlier discussion). The model implies that, all things being equal, lingual prehension might be slightly more effective at capturing active insect prey than jaw prehension. Smith et al.'s (1999) observations are consistent with this prediction. Most scleroglossans using jaw prehension have other adaptations to facilitate prey capture, especially mesokinesis (see later), but Tiliqua lacks kinesis (De Vree and Gans, 1987) and is notoriously slow moving (Greer, 1989). Species of Tiliqua are omnivorous (Greer, 1989) and most of their food types do not require particular speed or proficiency to ingest. Gans et ah (1985), for example, described the remarkable
8. Feeding in Lepidosaurs inefficiency of ingestion in T. rugosa. The ability to resort to lingual prehension for certain, more active prey types might be especially valuable for these slowmoving lizards. In any case, given a broad enough foretongue (as in most gekkotans and some scincomorphans), lingual prehension should be possible using wet adhesion alone. However, in most scleroglossans, the foretongue is sufficiently modified that this is not possible. Skinks of the Egernia group have the breadth of foretongue and other features to make lingual prehension by wet adhesion practicable. Some cordylids and most geckos have a similarly broad foretongue and it is in these families that two other known cases of scleroglossan lingual prehension occur. Controlled performance experiments must be conducted to test the model adequately. Perhaps the most significant obstacle to scleroglossan lingual ingestion is that the appropriate motor pattern has to be reacquired (see Section VII). G. Function of Cranial Kinesis The literature dealing with the functional significance of streptostyly and cranial kinesis is badly in need of critical review. It is plagued, particularly, by the lack of a comparative, phylogenetic approach (Schwenk, 1994b). Such a review is not possible here, but some preliminary observations are warranted. As noted previously, cranial kinesis sensu lato includes any intracranial skull movement, including movement of the quadrates (streptostyly), but cranial kinesis in the strict sense refers to intracranial movement exclusive of streptostyly. The latter meaning is used here for clarity. 1,
Streptostyly
Six principal functions for streptostyly have been proposed: (1) It has no function, per se, but is mechanically coupled to snout elevation and depression during cranial kinesis (e.g., Frazzetta, 1962). (2) It permits anteroposterior movements of the mandible for proper tooth alignment during occlusion in acrodont agamids (Robinson, 1976). (3) It increases the mechanical advantage of jaw adductors (external adductors: Gingerich, 1971; Rieppel, 1978; pterygoideus: Smith, 1980). Gingerich also thought that it reduces joint reaction forces. (4) It facilitates protrusion and positioning of the tongue and lower jaw during lingual prey capture in iguanians (Frazzetta, 1986). (5) It facilitates intraoral transport (Robinson, 1967; Throckmorton and Clarke, 1981). (6) It promotes shearing of prey with the teeth by propalineal movements of the mandible (De Vree and Cans, 1994). It is first of all clear that these hypotheses are not mu-
255
tually exclusive. Only the first has been falsified. Since Frazzetta (1962) proposed his quadric-crank model of cranial kinesis, several studies have shown that streptostylic movements of the quadrates frequently occur independent of mesokinesis (e.g., Throckmorton, 1976; Smith, 1980, 1982; Throckmorton and Clarke, 1981; Herrel and De Vree, 1999a). In particular, streptostyly occurs in some iguanians (notably Uromastix), which lack any other type of skull kinesis (see later). The same condition would have applied to the earliest squamates (e.g., Robinson, 1967; Rieppel, 1978; Frazzetta, 1986; lordansky, 1990a). Thus, the origin and adaptive significance of streptostyly, if any, must be explained independent of cranial kinesis in the strict sense. However, it is possible that streptostyly took on other functional roles in scleroglossan taxa after the origin of cranial kinesis. Conversely, functional streptostyly is sometimes absent in taxa reputed to be streptostylic based on anatomy (e.g.. Iguana, Throckmorton, 1976; Agama, Herrel et ah, 1996a; Tiliqua, and Corucia, Herrel et ah, 1999b). This highlights the fact that there is a vast degree of variation in the anatomy of the quadrate-pterygoid joint and its potential for movement (Throckmorton, 1976; lordansky, 1996; Herrel et al, 1998b; see Section IV). Workers sometimes disagree even on the potential for quadrate movement. Bradley (1903), for example, noted that quadrates are fixed in chameleons, but based on manipulations, Frazzetta (1986) suggested that they are capable of forward rotation. In Uromastix, the quadrates rotate forward, protracting the lower jaw, during mouth opening and retract during mouth closing (Throckmorton, 1976; Herrel and De Vree, 1999a). A similar pattern occurred in the agamid, Pogona (Amphibolurus) (Throckmorton and Clarke, 1981), but no streptostyly was observed in Agama (Herrel et al, 1996a). The pattern of quadrate movement is similar during all feeding stages in Uromastix. Notably, the quadrates are stationary during SC-PS (Herrel and De Vree, 1999a). This is consistent with Robinson's (1976) suggestion that streptostyly serves to align upper and lower teeth for occlusion during a bite. It would be necessary for mandibular movement to stop as the teeth engage due to the tightness of their fit. As discussed previously (see Section IV,B), wear evidence indicates that such occlusion is especially important in Uromastix, as compared to other agamids. Throckmorton (1976) observed that Uromastix was capable of cropping thick pieces of sweet potato in a single bite, whereas the pleurodont Iguana had to shake the same food item violently to crop it. Shear does not occur in Uromastix duxin^ propalineal movements of the lower jaw as suggested by De Vree and Gans (1994) and Herrel and De Vree
256
Kurt Schwenk
(1999a), but occurs between alternating upper and lower teeth at the moment of occlusion (see earlier discussion). Thus, the specialized dental-shearing mechanism of Uromastix should require initial alignment of upper and lower teeth by means of streptostyly followed by stabilization during tooth engagement. This is exactly the pattern observed (Herrel and De Vree, 1999a). Similar streptostylic movements during swallowing might be necessitated by tooth engagement during jaw closure irrespective of the presence of food between the tooth rows. The same pattern of streptostyly during ingestion (Throckmorton, 1976) is also consistent with Frazzetta's (1986) hypothesis (No. 4, earlier), but its continuation throughout the feeding sequence and its absence during ingestion in Iguana mitigate against its importance during prey capture. Studies of streptostyly in Varanus are contradictory. Boltt and Ewer (1964) and Rieppel (1978a, 1979a) found that quadrate movement was similar to the pattern described for Uromastix and linked with mesokinetic flexion of the snout, as predicted by the quadric-crank model, whereas Smith (1982) found complex and variable patterns of quadrate movement during inertial thrusts that were uncorrected with upper jaw kinesis. De Vree and Gans (1989) reported a pattern of streptostyly and kinesis identical to the first varanid pattern. MacLean (1974:197) observed "rocking of the quadrates and advance of the mandibles, independent of the skull" during intraoral transport in a gymnopthalmid. Reconsidering the six hypotheses in the order stated previously: (1) There is no necessary link between streptostyly and dorsoventral flexion of the snout, but the coupling may be important in some species. (2) Limited data suggest that, at least in the specialized Uromastix, streptostyly is important to control occlusion of the acrodont dentition, but this cannot be a universal explanation for its presence. (3) Streptostyly may very well increase the mechanical advantage of various adductor muscles, although Herrel et al. (1998b) disputed the validity of Gingerich's (1971) model. Only modeling approaches have addressed this issue so far. (4) Its role during lingual ingestion is unclear and not well supported for reasons stated earlier. (5) Streptostyly is unlikely to be important in intraoral transport because this is accomplished entirely by hyolingual retraction independent of jaw movement (Herrel and De Vree, 1999a) and would be irrelevant during inertial transport. (6) Although the shearing mechanism of Uromastix necessitates fixation of the mandible and quadrates during tooth engagement, it is possible that in pleurodont species some shearing action might occur as a result of quadrate retraction during biting, as in Sphenodon. Bels and Goosse (1989) suggested a similar function for lingual translation of the bolus against the
tooth rows during a bite. There are no data to support or refute this hypothesis except to note that chewing in most squamates does not usually slice the bolus into parts (as it does in Sphenodon), but rather softens it by means of puncture-crushing (see earlier discussion). Significant shearing seems to occur only during cropping and biting in acrodonts and amphisbaenians, and by vigorous head shaking in all taxa, including pleurodonts. Before these matters can be clarified, investigators should consider patterns of streptostyly in the comparative contexts of phylogeny, feeding stage, and food type. Only when enough comparative data are available for each of these confounding variables is there hope of general understanding. 2.
Mesokinesis
Many hypotheses for the functional significance of cranial kinesis have been suggested. Reviews are provided by Condon (1987) and Arnold (1998). This section focuses on the phylogenetic distribution of cranial kinesis and its implication for the functional morphology of ingestion. It does not address the question of meso- versus metakinesis, nor the validity of the quadric-crank model. As noted previously, variation in the anatomy of the quadrate-pterygoid joint and various functional studies suggest that snout flexion can be independent of both quadrate rotation and metakinesis. In any case, I suggest that elevation and depression of the snout at the mesokinetic axis are functionally most relevant to squamate feeding. Whether or not minute metakinetic flexure occurs as well does not affect the following arguments. Frazzetta (1962,1983) argued that kinesis allows the control of upper and lower jaw alignment during prey capture so that jaw closure and prehension will be fast and precise, and prey will not be knocked away. The upper jaw closes on the prey immediately after initial contact with the lower jaw. Upper jaw flexion allows an instantaneous momentum reversal in the upper jaw as it changes from opening to closing, without requiring head flexure at the neck. This putatively streamlines the gape cycle during the strike so that all activity is confined to a single, integrated functional unit (the cranial system), whereas jaw prehension with an akinetic skull would require coordinated activity of the cervical system as well. While coordination of two functional systems is, of course, possible (and evident, for example, during most chewing behavior), the kinetic system enhances both the precision and the rapidity of a prehensile bite, which has obvious benefits for feeding on volatile prey. In a clever analogy, Frazzetta (1983) showed that if one catches a moving object between thumb and index finger (analogues for lower
257
8. Feeding in Lepidosaurs and upper jaws, respectively), one does not fix the index finger and move the thumb alone, but rather, both fingers are first separated (opened) and then closed simultaneously to capture the object between the tips of the fingers, i.e., the "upper jaw" is raised as the "lower jaw" is depressed, then both reverse direction to close on the moving object. The limited functional data available for cranial kinesis are perfectly consistent with this hypothesis (Frazzetta, 1983; Smith and Hylander, 1985; Condon, 1987; Schwenk, unpublished observations). These studies show that the only time the snout is dorsoflexed above the rest position is the instant before prehension (Fig. 8.25). The actual monient of prehension occurs as the snout crosses back over the rest point as the mandible is elevated. The snout continues on past the rest point into ventroflexion and remains ventroflexed throughout the remainder of the feeding sequence. Dorsoventral flexion of the snout continues to occur sporadically with inertial thrusts and other feeding behaviors, but virtually never above the rest position (in several attempts to "unhook" the teeth from the limbs of a mouse, one individual Varanus exhibited "tooth clearing" behavior, which involved some dorsoflexion beyond the rest position; Condon, 1987). Frazzetta's (1962, 1983) hypothesis, therefore, suggests that kinetic flexion of the snout is an adaptation for jaw prehension of active prey. Kinematic data available so far strongly support the hypothesis. Two additional, comparative tests of the hypothesis are possible. First, cranial kinesis (as functionally determined, not anatomically inferred) and jaw prehension should be phylogenetically correlated. Given the phylogenetic dichotomy of the prehension mode suggested by Schwenk and Throckmorton (1989) and discussed in this chapter, this implies that cranial kinesis should be present in Scleroglossa, but absent in the lingualfeeding Iguania. This is, indeed, the case (Schwenk, 1987, 1994b, in preparation; also Arnold, 1998). The pattern of distribution of kinesis conforms exactly to the prediction, lending further strong support to Frazzetta's hypothesis. Note that the absence of kinesis in a scleroglossan would not weaken the hypothesis (see next), but its demonstration in an iguanian during lingual feeding would. Only one functional study has ever suggested the possibility of kinetic flexion in the skull of an iguanian (Throckmorton and Clarke, 1981). These workers found weak evidence of slight mesokinesis during chewing, but average flexion was less than 1°, far below the accurate resolution of their cineradiographic technique (Throckmorton and Clarke, 1981; Condon, 1987) and nothing like the excursions evident in scleroglossans (12°-30°; Boltt and Ewer, 1964; Rieppel, 1979a; Frazzetta, 1983; Condon, 1987). Morpho-
logical evidence suggests that significant mesokinetic flexion, particularly ventroflexion, cannot occur in most, if not all, iguanians (Schwenk, in preparation). Second, herbivorous scleroglossan taxa, or those that feed on relatively stationary food items might be expected to lack cranial kinesis. There are too few comparative data for an adequate test, but the absence of cranial kinesis in the slow-moving, omnivorous/herbivorous scincid taxa, Tiliqua (De Vree and Cans, 1987; Herrel et ah, 1999b) and Corucia (Herrel et al, 1999b), is consistent with the hypothesis. It is interesting to note that these taxa are exceptional in resorting to lingual prehension in some circumstances (see earlier discussion). In conclusion, strong evidence suggests that mesokinesis arose in the scleroglossan clade as an adaptation to enhance the efficacy of jaw prehension. This does not preclude additional roles for kinesis in other stages of feeding, but prey capture appears to be its fundamental basis. In general, large lizards, especially those with extensive co-ossification of the head scales, exhibit less anatomical kinesis than juveniles (e.g., Hallerman, 1992) and smaller species (personal observation). It is unfortunate that most functional studies of cranial kinesis heretofore have focused on large species of Varanus. It is likely that future studies of small, insectivorous scleroglossans will more accurately reveal the kinematics, and therefore the function, of cranial kinesis in lizards. My films of a small gecko {Coleonyx variegatus), for example, reveal large raesokinetic excursions during ingestion. Such studies would do well to exploit Condon's (1987) goniometric technique, which has provided the best data so far on kinesis (see Chapter 1).
VL SPECIALIZED FEEDING SYSTEMS A. Chameleons 1. Natural
History
Chameleons are unique among lepidosaurs in their use of tongue projection to capture prey. This behavior is distinguished from typical iguanian lingual ingestion by its interjection of a ballistic (i.e., free-floating) stage at the end of SO II such that the tongue is "launched" from the lingual process as a projectile, continuing on toward the prey by virtue of its own inertia. In this sense, tongue "projection" is distinct from tongue "protrusion" and the two behaviors should not be conflated (Chapter 2). Tongue projection is universally present within Chamaeleonidae (sensu stricto) and is functional in neonates at birth (Bustard, 1965; personal observation). Variation in the chamaeleonid feeding system is virtually unstudied and the projectile
258
Kurt Schwenk
mechanism is assumed to be identical in all species. There are, however, differences in superficial tongue form (Schwenk, unpublished observations), but these may have no impact on function. Maximum projection distances range from one to two body lengths (personal observation), but each projection is modulated according to visual assessment of prey distance (see discussion of chameleon vision). Chameleons are opportunistic predators feeding on a large variety of invertebrate and vertebrate animals (e.g., Loveridge, 1953; Bourgat, 1970; Burrage, 1973). None are herbivorous, and the occasional bit of plant matter is undoubtedly ingested incidental to prey capture, although Bourgat (1972) thought it might be intentionally consumed. A single, large species (C. melleri) has been shown to consume prey as small as ants and as large as adult birds (e.g., Loveridge, 1953; Pitman, 1958, 1962). Birds probably represent, proportionately and absolutely, the largest prey items apprehended with the tongue by any lizard. Tongue projection is used for capture of all prey items and there is no apparent shift to jaw prehension with increasing prey size as in other iguanians. Thus, more than any other lizard, chameleons are obligate tongue feeders and, in most cases, are unwilling or incapable of using the jaws for prey capture. When confronted by a prey item directly in front of it, a C. jacksoni backed up along its perch to project its tongue rather than capture the insect in its jaws (personal observation). Burrage (1973:75) noted that Bradypodion (Chamaeleo) pumilus was similarly committed to tongue projection. However, he also suggested that C. namaquensis sometimes pursued escaping prey and "snapped up the victim in their jaws in the normal saurian manner." C. namaquensis is one of only a few desert chameleons and is the only species that is truly terrestrial, thus it might be expected to show unusual behavior. Marx (1955) also reported a rare case of jaw prehension in captive C. oshaugnessyi, which used the sides of its jaws to grasp earthworms after repeated attempts at lingual prehension had failed (see Section V,F,2). Chameleons also use the jaws for ritualized (and completely ineffectual) biting during agonistic behavior (e.g.. Bustard, 1965; Parcher, 1974). Chameleons occupy narrow perches in complex, three-dimensional environments in which rapid pursuit of prey is not possible. This is true even for small, semiterrestrial species (e.g., Brookesia), possibly representing the primitive condition (Rieppel and Crumly, 1997; C. Crumly, personal communication), which climb through the irregularities of the ground-level environment and the lower reaches of small plants. Virtually all aspects of chameleon morphology and behavior seem to be predicated on life among precarious perches—prehensile tail (e.g., Zippel et ah, 1999),
zygodactylus feet (e.g., Hurle et al, 1987), turreted eyes (see earlier discussion), cryptic coloration, color change and emphasis on visual communication (e.g., Parcher, 1974; Dodd, 1981; Okelo, 1986), slow and deliberate locomotion (Peterson, 1984), muscle physiology (Abu-Ghalyun et ah, 1988), and rocking during locomotion for background matching (e.g.. Bustard, 1965; Fleishman, 1992). Thus, it is likely that chameleons evolved ballistic tongue projection as a means of capturing active prey without the need for lunge and chase—the extreme expression of an ambush strategy derived from a lingual feeding ancestor. 2. Specializations
of the Hyolingual
Apparatus
The chameleon hyolingual apparatus is highly specialized, but nonetheless derivable from a generalized agamine condition (Schwenk, 1983,1988; unpublished data; Smith, 1988). There are several key novelties of relevance here. 1. The verticalis is hypertrophied and modified to form a muscular "tube" around the lingual process; in chameleons it is known as the "accelerator muscle." It forms the great bulk of the tongue and is accommodated within the mouth by a deep pocket of distensible skin between the mandibular rami. 2. The lingual process is robust and parallel sided, except for the very tip, which is tapered. This is in contrast to most other iguanians in which the lingual process is tapered along its entire length. It is surrounded by a cavity filled with fluid that is histochemically similar to synovial fluid (Bell, 1989). 3. A dense, two-layered tendinous sheath surrounds the lingual process, lying between it and the accelerator muscle (Zoond, 1933; Bell, 1989). It is apparently evaginated during projection (Zoond, 1933; Cans, 1967). 4. The hyoglossus muscles are greatly elongated and pleated at rest. 5. The hyoglossus muscles contain supercontracting fibers: myofilament Z discs are fenestrated so that thick (myosin) filaments can pass through them during contraction (Rice, 1973). Thus, each sarcomere can shorten to 16% of its extended length instead of the 40% typical of striated muscle. The hyoglossus muscles are therefore capable of extreme shortening. 6. At rest the dorsal, papillose surface ("membrana glandulosa") contains a deep pocket or "dimple" formed by a transverse fold that is sigmoid in sagittal section (Cans, 1967; Bell, 1989; Schwenk, unpublished results). In some species there is a posterior, longitudinal fold so that at rest the dimple is formed from confluent, perpendicular creases in the mucosal surface. Numerous plumose cells line the clefts leading into the deep glandular crypts of this region (unpublished observations). The dimple is evaginated during SO as the
8. Feeding in Lepidosaurs tongue is protruded and the glandular membrane is pulled around the end of the accelerator muscle. The portion of the surface that comes to lie at the apex of the tongue during projection corresponds to the everted inner surface of the dimple. It serves as the contact zone during prehension (Schwenk and Bell, unpublished results). 7. Extensions of the hyoglossus muscles (longitudinalis linguae adductores) run along the sides of the accelerator column and insert into the base of the dimple at rest (see Section IV). These are positioned to pull the everted contact zone back into its "dimpled" conformation during prehension. Some fibers of the hyoglossus muscles similarly insert at the surface of the tongue beneath the contact zone in other iguanians, but they do not form discrete bundles (Schwenk, 2000a; unpublished results). 8. The base of the tongue, corresponding topologically to the posterior end of the accelerator muscle or the posterior end of the dorsal, glandular region ("lingual pad" of Bell, 1989), is completely free of the basihyal. Direct attachment of the tongue to the hyobranchial apparatus is restricted to the posterior origin of the hyoglossus muscles and the tendinous sheath of the lingual process, which is evaginated and stretched during projection. In all other iguanians the base of the tongue is tightly adherent to the basihyal (see Section IV). 3. Kinematics of Ingestion Chameleons rely on cryptic morphology and behavior to approach prey to within projection distance. Once in position, they brace the head and body to create a stable platform from which (1) to aim the tongue, (2) to absorb the inertial reaction forces of projection, and (3) to resist the destabilizing forces of tongue retraction. Such fixation is distinct from other iguanians in which rapid head and body movement toward the prey item are an integral part of ingestion. Despite insertion of a projectile phase at the end of SO II (Bramble and Wake, 1985; Schwenk and Bell, 1988; Schwenk and Throckmorton, 1989), the kinematics of chameleon ingestion are, otherwise, remarkably similar to generalized iguanians (Fig. 8.23) (Bell, 1990; Wainwright et al, 1991). During SO I the tongue is initially protruded, and in SO II it is held in a protracted state, cantilevered outside the mouth on the lingual process. During this sometimes prolonged "aiming" stage, accommodation cues are used to assess prey distance (see earlier discussion) and the mucosal, prehensile component of the tongue is translated anteriorly and curled around the distal end of the protruded tongue (Schwenk and Bell, 1988; Bell, 1989,1990). This process evaginates the lingual "dimple" so that its inner surface is exposed at the end of the tongue (see earlier discussion). Projection distance is precisely modu-
259
lated so that the tongue rarely travels much beyond the prey item. Tongue-prey contact occurs at the end of SO II and is closely correlated with the start of FO, as in other iguanians (Fig. 8.23). Retraction is prolongued relative to other iguanians due to the great length of tongue that must be contracted and fitted back over the lingual process. During this process, the tongue swings pendulously back and forth, potentially dislodging the chameleon from its narrow perch were it not braced, particularly when adherent prey are heavy. Finally, the tongue is retracted fully onto the lingual process and the hyobranchium is retracted by posthyoid muscles, withdrawing the tongue into the mouth as the jaws close. If the prey item is large, initial jaw closure results in a bite. Otherwise, the prey item is manipulated by the tongue and placed between the tooth rows in the next gape cycle to initiate chewing (see earlier discussion). Previous kinematic studies have identified five phases of prey capture that do not correspond to the standard gape cycle phases (Altevogt and Altevogt, 1954; Bell, 1990; Wainwright et al, 1991): (1) fixation, (2) tongue protrusion, (3) tongue projection, (4) tongue retraction, and (5) hyobranchial retraction. During fixation the turreted eyes are rotated forward as the head is oriented toward a prey item. This process is described in Section III,C,1. Ingestion begins with tongue protrusion, which occurs during SO I and SO II (Fig. 8.23). As noted earlier, protrusion may cease during the plateau phase of SO II and the tongue is held in readiness for projection. Projection is initiated at the end of SO II and is marked by a slight decrease in gape (Bels and Baltus, 1987; Bell, 1990; Wainwright et al, 1991). Tongue retraction occurs primarily during FO. Hyobranchial retraction overlaps with tongue retraction (Wainwright and Bennett, 1992a) and is therefore not a distinct phase. It begins during FO, but is most evident during FC and SC (if it occurs) and into the stationary phase. The slight gape reduction occurring with tongue projection is unique to chameleons (Fig. 8.23) (Wainwright et al, 1991). It seems to correspond to the onset of tongue retraction and may be an indirect effect of inertial changes or changes in muscle activity pattern. Wainwright et al (1991) also noted that, in contrast to agamines, head depression occurs during retraction. This corresponds to the period in which the tongue hangs vertically, often swinging pendulously as it is retracted. Head depression thus aligns the lingual process with the retracting tongue and may be as much passive as active. 4. Mechanism of Tongue Projection Over the last several hundred years the mechanism of chameleon tongue projection has attracted the
260
Kurt Schwenk
attention of many distinguished anatomists, beginning seriously with Perrault (1676). Several fanciful theories were proposed, including the action of blood (e.g., Houston, 1828) and air (e.g., Dumeril, 1836), but Cuvier (1805) and Duvernoy (1836b) correctly inferred the role of muscular contraction. Nonetheless, some workers attributed projection to a "sling-shot" action of the hyobranchium, which was thought to impart a violent inertial force to the tongue (e.g., Duvernoy, 1836b; Owen, 1866). Although he erred in his final conclusion, Owen (1866) presented a remarkably modern functional interpretation of projection that has been virtually ignored in the subsequent literature. Kathariner (1895) and Gnanamuthu (1930, 1936) reviewed some of the earlier work, but a complete history of investigation in this area has never been undertaken. Such a treatment is beyond the scope of this chapter, but Table 8.2 lists many of the relevant studies. This section focuses on our current understanding of hyolingual function during lingual ingestion in chameleons. Chameleon tongue projection depends on the interaction between accelerator muscle and lingual process (Zoond, 1933; Cans, 1967; Wainwright and Bennett, 1992b). Contraction of the accelerator results in a decrease in the diameter of its internal lumen, thereby compressing the fluid around the lingual process (Owen, 1866; Zoond, 1933; Cans, 1967; Wainwright and Bennett, 1992b; van Leeuwen, 1997). However, activity in the accelerator begins approximately 200 msec before projection actually takes place (Wainwright and Bennett, 1992a). The delay in projection onset occurs because the sides of the lingual process are parallel (untapered), thus the compressive force of the accelerator acting against it is normal (perpendicular) to its surface and has no forward component that would tend to move it off the process (in contrast to other iguanians). However, the accelerator is a membranebound muscular hydrostat (see Chapter 2) and during this period it narrows in diameter and lengthens anteriorly (Owen, 1866; Zoond, 1933; Wainwright and Bennett, 1992b; Schwenk et aL, manuscript in preparation). Activity in the accelerator ceases just before the onset of projection and begins again immediately after (Wainwright and Bennett, 1992a), but maximal compressive force is maintained throughout (Wainwright and Bennett, 1992b). Eventually, the anterior end of the lengthening accelerator muscle extends over the tapered tip of the lingual process at which point a forward component of the compressive force is introduced rather suddenly (Wainwright and Bennett, 1992b). This force causes the accelerator to slide along the process so that its entire length eventually passes over the tapered tip. The synovial-like lubricant surrounding the process reduces the frictional resistance of sliding (Cans, 1967; Bell, 1989). As it passes over the
tip, the accelerator rapidly accelerates forward. Acceleration may be enhanced at this point by the concomitant reduction in the diameter of the internal lumen, which is predicted to increase the compressive force exerted against the tip (van Leeuwen, 1997). The rapid forward acceleration of the tongue causes it to project off the lingual process, continuing toward the prey item by virtue of its own inertia. The accelerator remains elongated throughout the ballistic phase and begins to shorten during retraction, returning slowly to resting length during subsequent cycles (Schwenk et ah, manuscript in preparation). The hyoglossus muscles become active immediately after projection and serve to retract the tongue (Wainwright and Bennett, 1992a). The rapidity of projection suggested to many investigators that the projectile apparatus was somehow "preloaded" and that a "trigger" mechanism was necessary to release it. Although several mechanisms were proposed, the most plausible was that the accelerator was held back by activity of the hyoglossus muscles, which then relaxed to trigger projection (e.g., Zoond, 1933; Altevogt and Altevogt, 1954). Wainwright and Bennett's (1992a,b) studies falsified this specific hypothesis, but supported the notion of a preloading mechanism. They showed that projection is delayed due to the parallel-sided lingual process, until the accelerator can produce maximal tension. Because the base of the accelerator is prevented from moving posteriorly, hydrostatic elongation during preloading activity moves the tongue anteriorly along the process, extending the propulsive part of the accelerator over the tapered tip where projection (translation along and off the process) is initiated. In experimental manipulations, the longer the tapered region, the sooner the tongue begins to project, but the weaker its thrust. A short taper produces a more sudden and more forceful projection, but with a longer onset delay (Wainwright and Bennett, 1992b). Van Leeuwen (1997) showed that the curved, radial muscle fiber architecture of the accelerator satisfies the needs of dense fiber-packing, high internal pressures and evenly distributed pressure, without the problems of a circular, sphincter-like muscle. The normal stress created by the latter fiber arrangement would be high enough to cause binding and would interfere with sliding along the lingual process. Alternation of "handedness" in sequential layers of curved, radial fibers along its length (Gnanamuthu, 1937; Cans, 1967; Bell, 1989) prevents torsion of the accelerator during elongation. 5. Mechanism of Prehension The model of tongue-prey adhesion outlined earlier for generalized iguanians applies in most respects
261
8. Feeding in Lepidosaurs to chameleons, with some additions. As in other iguanians, the adhesion is promoted by a combination of wet-adhesion (capillarity) and interlocking. Owen (1866:437) noted that the prehensile part of the tongue is "bedewed by adhesive secretion" and indeed, the surface within the dimple (corresponding to the contact zone) is very glandular, comprising reticular papillae virtually identical to the agamine condition. It is noteworthy that the papillae become longest and most filamentous (i.e., least reticular) directly in the center of the dimple/contact zone (Schwenk, unpublished observations; see photos in Bell, 1989), which might aid in stress translation and tension resistance, as per the model. Plumose cells also abound here, presumably to enhance interlocking. Chameleons deviate from the generalized model by actively and dramatically changing the shape of the prehensile surface of the tongue during prey capture. As noted previously, the contact zone is in-folded at rest, forming a deep pocket lined by the prehensile mucosa. The fold creates upper and lower "lobes" in the glandular surface that Owen (1866) compared to the prehensile end of an elephant's trunk. This pocket is everted at the apex of the tongue prior to projection and oriented toward the prey item. During the ballistic phase, the contact zone begins to reinvaginate just before contact (Schwenk and Bell, unpublished results), presumably by action of the anterior portions of the hyoglossus muscles (longitudinales linguae adductores), which insert at the base of the dimple (Schwenk, 1983; see earlier discussion). Prey contact occurs within this zone, and invagination proceeds rapidly so that the prey item is drawn deeply into the tongue, virtually surrounded by the surfaces of the contact zone (Schwenk and Bell, unpublished results; see figures in Dischner, 1958). Invagination continues even if the prey item is missed. The shape of the contact zone conforms closely to the shape of the prey surface. On a regular surface it gives the strong impression of a suction cup. The possibility of suction is supported by the close conformity of the tongue surface to the prey surface, the potential for marginal sealing with tongue secretions, and the potential to create negative pressure within the diniple by retraction of the hyoglossus/longitudinal adductors (Schwenk, 1983). Although the action of suction in prehension is purely speculative, the great mass of some prey taken by chameleons (see earlier discussion) and the fact that any tendency of the prey item to become dislodged will be aggravated by gravity and inertia during the pendulous phase of tongue retraction suggest that interlocking and capillarity may not always be enough. A small chameleon of unknown weight (C. montium) was shown to produce a pulling force of up to 43 g against a mealworm attached to a scale (Dischner,
1958). The worm was pulled deeply into the retracted dimple and, when measured, was oriented end first into the tongue, i.e., parallel to the direction of pulling, making the tongue's prehensile tenacity all the more remarkable. 6.
Evolution
As radical as the chamaeleonid lingual projection system is, it is clearly derived from an agamine-type ancestral condition (Schwenk, 1986; Schwenk and Bell, 1988; Smith, 1988). Abel (1952) and lordansky (1973) noted the similarity of agamine tongue prehension to the chameleon condition. Schwenk and Bell (1988) showed that some agamine lizards exhibit a kinematic phase intermediate between typical lingual protraction and ballistic projection. This phase corresponds to the aiming or tongue protrusion stage of chameleon feeding during which the tongue is cantilevered outside the mouth on the protracted lingual process and the papillose, prehensile surface is translated forward around the apex of tongue. Agamines sometimes maintain this conformation briefly while in pursuit of prey, although they never hold it as long as chameleons often do. As discussed earlier, virtually all novel aspects of the chameleon system relate to the insertion of a ballistic projection phase between protrusion and retraction. Nonetheless, many aspects of the derived chameleon morphology are evident in agamines, although sometimes in a less developed state. The mucosal, papillary surface, for example, is virtually identical in chameleons and some agamines (Schwenk, unpublished results), and some agamines form a facultative lingual "dimple" during retraction of prey that is identical in conformation to the chameleon condition (Schwenk and Throckmorton, unpublished results). Although Agamidae may be paraphyletic relative to Chamaeleonidae, there is little doubt that these are closely related taxa, thus the similarities between them are almost certainly shared derived features. The derivation of the chamaeleonid system from agamines is also supported by similarities in muscle activity pattern (Herrel et ah, 1995).
B. Amphisbaenians 1. Feeding Function Although almost certainly scleroglossans, amphisbaenians have been regarded traditionally as a separate taxon (see Section II). Because of this taxonomic status, their highly specialized cranial form, and the unfortunate dearth of data on feeding in the group, I consider them separately here so as to summarize in one place our current knowledge of feeding in the
262
Kurt Schwenk
group. Gans (1966,1969a, 1974) provided the only published descriptions of (unspecified) amphisbaenians feeding, but Nate Kley (unpublished results) has supplemented these accounts with observations of feeding in several pleurodont species {Amphisbaena fenestrata, Blanus strauchii, and Geocalamus acutus). Gans' feeding trials used insects and vertebrate prey ("reptiles" and "almost fully grown mice;" Gans, 1966) and Kley's observations were of feeding on mealworms. Gans noted that amphisbaenians feeding on the surface make a short lunge toward the prey item with mouth agape. After jaw prehension, small prey are pulled back into a burrow hole. If the prey item is large it is sometimes dismembered by axial spinning. The jaw joint lies below the line of the tooth row, a highly unusual condition in squamates, and Gans suggested that this imparts a posteriorly directed force that pushes the prey between the teeth, locking it in place. A small prey item sensed within the burrow is pinned against the wall with a curve of the body and is held in place with successive loops while the amphisbaenian backs up to bring its mouth alongside. Prehension is then with the sides of the jaws against the tunnel wall. Large prey on the surface are killed by biting (Gans, 1966) and are reduced incrementally by subsequent bites. When eventually small enough, the carcass is pulled below ground. Kley confirmed most of Gans' observations for the taxa he studied. He likewise found that prey placed on the substrate are quickly sensed by an amphisbaenian, which emerges from its burrow to feed on the surface (see Section III,C,5). He noted that some burrow entrances are funnel shaped, probably as a simple consequence of repeated emergence, but suggesting the possibility that they could function as a "pitfall trap." Strikes involve a "lunge" straight ahead to grasp the prey item in the jaws or a "jerk to the side" to pin it against the substrate. If after pinning, the mealworm is next to the jaws, it is ingested immediately; if posterior to the jaws, the amphisbaenian backs up until it can grasp the mealworm. Occasionally the prey item is rubbed against the substrate, possibly to subdue or reduce it. No obvious chewing or processing cycles were evident, but the initial prehension is a powerful, crushing bite. No inertial thrusts were observed, and all intraoral transport cycles appeared to be hyolingual. Swallowing behavior was not observed, usually because the amphisbaenian withdrew into its burrow. Gans' and Kley's accounts of feeding in amphisbaenians are remarkably consistent. Although extremely divergent in cranial morphology, the hyolingual component of the amphisbaenian feeding apparatus is not much different from other scleroglossans, thus basic patterns of feeding are conserved. However, amphis-
baenians depart from the typical scleroglossan pattern in two ways: (1) Prey capture and ingestion are sometimes decoupled, as when the side of the head or a body loop is used to pin the prey item against the substrate until the jaws can be positioned for grasping (see also Section VI,C). (2) The dental battery and jaw apparatus are specialized for powerful, shearing bites that allow amphisbaenians to excise and ingest parts of larger prey. Gans observed the removal of "cookie-cutter" portions from large vertebrate prey and Kley noted that neonate mice left on the surface had limbs neatly excised. Both derived features of amphisbaenian feeding are interpretable as adaptive consequences of limblessness and fossoriality. 2. Ecological and Evolutionary Aspects of Feeding Gans (1969a, 1974) believed that amphisbaenians are specialized for preying on large vertebrate prey, which they eat with relish in the laboratory. In contrast, taking an ecological approach, Andrews et al. (1987) concluded that fossorial squamates, in general, are forced to consume many small prey items (rather than a few large ones) due to constraints on predation and the feeding apparatus in an underground habitat. Although equally plausible, these hypotheses are in direct opposition. Gans' conclusion stemmed from functional inferences based on anatomy, laboratory observations of captive animals, and the assumption that amphisbaenians typically feed on the surface. Andrews et al. (1987) based their conclusion on observations of prey handling in two species of skink, one terrestrial and one fossorial, and the assumption that fossorial species typically feed underground. Overall, amphisbaenian diets include a diverse array of small invertebrates with a preponderance of termites, ants, and insect larvae (see Section III,A). Sand and grit in amphisbaenian stomachs led Cabrera and Merlini (1990) to suggest that feeding occurs underground, within the litter, or beneath logs. The association of some amphisbaenians with ant and termite nests further suggests underground feeding. Despite the examination of hundreds of amphisbaenians stomachs, evidence for only four small vertebrate prey items has been found (see Section III,A). Thus, natural history data strongly support the Andrews et al (1987) contention. However, the natural diet of so few amphisbaenian taxa is known that it remains possible that some species are vertebrate specialists. Nonetheless, the imposing jaw mechanism and dental battery (including occlusion) shared by all amphisbaenians are not easily accounted for by dietary data. It is possible that jaw size and maximum gape are so limited by the constraints on head form imposed by fossoriality and
8. Feeding in Lepidosaurs head-first burrowing, as suggested by Andrews et al. (1987), that amphisbaenians must compensate with a highly effective biting mechanism. Although their typical prey do not appear to be especially hard to kill or process, the diets of more species need to be considered before conclusions are drawn. C. K o m o d o Monitor Although monitor lizards (Varanidae) are widely reputed to be large prey specialists, this erroneous belief seems to have been promulgated by the incautious and inaccurate inference of behavior and ecology from morphology (Greene, 1982; see Chapter 1 for discussion). In fact, most varanids consume many small prey items rather than a few large ones (Shine, 1986; Losos and Greene, 1988; James et al, 1992). Many monitors do eat the occasional large food item and a few regularly do so, but overall the varanid craniocervical system is specialized for rapid and accurate jaw prehension of relatively small, fast-moving prey and its subsequent manipulation within the mouth without benefit of the tongue. Among varanids, only the komodo monitor {Varanus komodoensis, ''ora") is highly specialized for feeding on very large prey. The consumption of prey too large to swallow whole is exceptional among lepidosaurs and most nonmammalian tetrapods. Juvenile oras are mostly arboreal and consume a typical lizard diet of relatively small invertebrates and vertebrates, but as they grow they become fully terrestrial and shift to a diet predominating in vertebrates, including very large species such as wild boar, deer, and water buffalo (Auffenberg, 1981). Although widely known as scavengers, adults are also predators of live prey. The large mammal species consumed are all late introductions to the islands inhabited by oras, thus Auffenberg (1981) speculated that these largest of lizards (up to 3 m total length) evolved to feed upon the pygmy elephants that were once abundant there (see Diamond, 1987). The following account of ora feeding is based on Auffenberg (1978,1981). 1. Capture and
Subjugation
While scavenging, a carcass is simply approached and ingestion begun, but one consequence of feeding on very large, living prey is that prey capture is decoupled from ingestion. Thus, before an ora can begin a feeding sequence it must subdue its victim and reduce it to appropriately sized pieces for swallowing. Auffenberg (1978:317) noted that oras (and some other varanids) "are the only living reptiles, other than turtles, that cut their food into sections before swallowing it . . . y. komodoensis is the most adept at this pro-
263
cess.'' Attacks are undertaken either from ambush or after stealthy approach. Favored prey items are small deer {Cervus timorensis), which are often located while asleep in the brush. In any case, the lizard must be within 1 m of its prey before it can lunge. An impending attack is often indicated by a lateral, sigmoid curve in the neck, as in some snakes (Chapter 9). Initial bites are directed to a variety of locations, including the back, flanks, legs, and belly. Very large prey animals are often bitten first on the legs and crippled. Somewhat smaller animals are twisted to the ground or lifted bodily, thrown down, and pinned. Repeated bites inflict gaping wounds and massive hemorrhaging. As soon as the animal is safely immobilized it is eviscerated and the intestines are consumed. Auffenberg suggested that evisceration increases the rate of bleeding and causes immediate physiological shock and rapid death. Biting is often accompanied by violent head shaking, as is typical of many scleroglossans. Once the prey is incapacited, portions of it are removed and ingested with the jaws. Virtually the entire animal is eventually consumed, including hooves and bones. Feeding often occurs in groups. 2. Ingestion The pieces of flesh removed from a dead or incapacitated animal and eventually swallowed whole are sometimes "astonishingly large" (Auffenberg, 1981: 210). Ingestion is facilitated by several adaptations of the skull and teeth. The snout is unusually broad in oras, as compared to most varanids in which it is typically narrow and pincer-like (see earlier discussion). The maxillary tooth rows are convex in dorsal and lateral views. The teeth are laterally compressed, recurved and blade-like, with small serrations along the posterior and apical anterior margins (see Section IV,B; Auffenberg, 1981; Abler, 1992). They lie at a slight angle relative to the jaw margin so that the leading edge is somewhat medial to the trailing edge. The carcass is bitten and the head is jerked back on one side so that the ipsilateral maxillary teeth rotate in an arc around a pivot provided by the tallest teeth on the contralateral side. Because of the broad, laterally convex shape of the snout, each tooth on the cutting side follows in the path of the one behind it so that a single cut is made, which grows progressively deeper with each passing tooth. Jerks are alternated left and right so that a very large piece of flesh is eventually removed. Abler (1992) showed that the marginal serrations of the teeth facilitate a "grip-and-rip" mechanism of cutting analogous to a serrated steak knife. Their much greater development on posterior edges is consistent with a cutting function during backward head jerk. This cut-
264
Kurt Schwenk
ting mechanism also serves during the initial attack to open large wounds that bleed profusely. The mouths of ora harbor large populations of bacteria that sometimes cause serious infection and death following a bite. Although their bites are widely believed to be especially toxic (Auffenberg, 1981), it is not clear that ora bites are any more infectious than those of other vertebrates. However, long-term captives have a depauperate bacterial fauna, and it appears that consumption of rotting carcasses is necessary to maintain large and diverse populations, thus ora bites may be particularly infectious by virtue of their scavenging habits. Auffenberg (1981) noted that in life as much as two-thirds of each tooth is covered by thick, spongy gum tissue that is pushed back during a bite. Repeated sliding of the gums against the teeth causes them to bleed and the blood to mix with saliva, forming a viscous fluid that Auffenberg speculated might serve as a bacterial culture medium. Abler (1992) suggested that tiny bits of rotting flesh caught in the serrations might inoculate the mouth with fresh bacterial cultures. In any case, bites incurred by prey animals during unsuccessful predation attempts might lead to infection, incapacity, and death, making food more readily available at a later time.
3. Other Feeding Stages and Cranial Kinesis Oras typically bolt their food in large chunks with no indication of significant processing. As noted previously, the forelimbs are occasionally used before ingestion to remove hair from rotting carcasses. As in other varanids, food items are moved through the oral cavity by inertial transport. Auffenberg (1981) thought that inertial thrusts were more rapid than in most lizards. The jaws were heard to snap together with each thrust when oras fed on small-diameter food items. Auffenberg also observed that very large items were sometimes pushed against the substrate to force it back into the throat. This may constitute a novel form of transport in lizards analogous to a similar process in constricting snakes (see Chapter 9). Auffenberg did not describe swallowing behavior specifically, but noted that large portions are swallowed whole. Oras salivate copiously when feeding, which may help lubricate the passage of food items through the pharynx and into the gut. Cranial kinesis has not been documented in V. komodoensis, but Auffenberg (1981) thought that the mesokinetic hinge remains functional in adults, despite earlier claims that it does not. He claimed to witness a slight flattening of the front part of the head during feeding and dorsoflexion of the snout as the jaws open.
D . Snakes Feeding in snakes is explicated in Chapter 9, but it is worth noting here that the specialized systems of snakes are derived from those of ancestral, scleroglossan lizards whose relatives live today. As diverse as they are, snake-feeding systems stemmed from ancestors that were specialized jaw feeders with highly kinetic skulls and slender, bifid tongues modified for hydrostatic elongation and chemosensory tongue flicking. Just as ballistic tongue projection in chameleons is, in some sense, the "logical" extension of a lingual prehension system, ophidian-feeding specializations can be viewed as extreme elaborations of an ancestral system already committed to jaw prehension, intracranial mobility, and vomeronasal chemoreception. It is possible that early specialization of snakes for jaw prehension and loss of the hyolingual transport mechanism associated with extreme tongue reduction committed them to animal diets and precluded the evolution of herbivorous forms (see Section III,A). Varanid lizards share a similar tongue form and loss of hyolingual transport and are similarly committed to animal food—with one exception. V. olivaceus is herbivorous (Auffenberg, 1988), but it is telling that it eats solid, compact fruit and not, for example, leaves. The fruits are swallowed whole and can be manipulated like animal prey using inertial transport, unlike other types of plant food (Auffenberg, 1988). Thus even the exception of V. olivaceus lends credence to the notion that the absence of a hyolingual transport mechanism acts as a constraint on the evolution of herbivory.
VII. THE EVOLUTION OF FEEDING IN LEPIDOSAURS Despite a recent surge of interest in the functional morphology of lepidosaur feeding, our sample of wellstudied taxa remains nearly as depauperate in phylogenetic breadth as it did in 1985 when Bramble and Wake published their seminal synthesis. Reviews since then have reiterated the need for a broader phylogenetic sample to discern patterns and test hypotheses (e.g., De Vree and Cans, 1989; Schwenk and Throckmorton, 1989; Bels et al, 1994), but despite movement in this direction, large gaps in our knowledge persist. To some extent this is because recent studies have tended to focus on many of the same taxa for which we already have data (e.g., Varanus, Tupinambis, Tiliqua, Iguana, and Uromastix), adding greatly to the depth of our knowledge, but not its breadth. These large, tractable species are understandably attractive for laboratory study, but unfortunately, they are atypical in
265
8. F e e d i n g in L e p i d o s a u r s
squamate phylogeny and Figure 8.30 summarizes major events in the volution of lepidosaurian feeding.
many aspects of morphology and diet and may not be representative of the vast majority of lizards. It is the functional diversity of small, insectivorous lizards that is most in need of exploration. It is in such species, for example, that the true significance of streptostyly and cranial kinesis may be revealed. Given the paucity of comparative functional data, such evolutionary conclusions as are possible remain tentative. Their generality is best tested by the study of additional species from a greater diversity of clades. Figure 8.29 shows the distribution of feeding-related characters across the
1 <&
•CHEMOSENSORY TYPE' TONGUE •BIPARTITE TYPE' TONGUE 'COMPROMISE TYPE' TONGUE 'FEEDING TYPE' TONGUE PHAR. COMPRESSION COMMON PHARYNGEAL PACKING INERTIAL TRANSPORT COMMON HYOLINGUAL TRANSPORT MESOKINESIS JAW - PREHENSION TONGUE - PREHENSION
n
D D
•
m m D
•
D D
•
Q> CO
•o
'c to
_5)
D D D
•D •D • D D
A. Evolution of Ingestion M o d e Within Squamata, two fundamentally different modes of ingestion are evident: lingual and jaw prehension of prey. Which was the ancestral method of prey capture? Historical questions in functional morphology, like other historical questions, must be approached through standard methods of phylogenetic character
c o
"6
a>
E
E
^ D D D
o
O
QL
D D D
D D
D D
C3
•D •D •n
D
o
CO
• •D •n
D
D
• • •
•D • •D • • •C
•D • •D • • •D
0 (0
sQ. O
0) to
5
c E >> G
• • • • •m •
•D D• B • m •D • •H B• m •D
D D
D
D D
n n
"t:
3
X
n n
CO
D D
CO •D
CO
C
S
D D
• •D • B • m • H D a • mm
o O
D D
•D • • n • • u m D E • G
'3
q3
•D O
< D
c X
I
D
D
D
D H B
D
c
c
CO
3 D
•D •D D• •
• • •D • D m B • m g• • • • D• D a
c
CO
•D
•g
CO
3
CO
CD CO
a>
B c
K
(0 T3 C
1 •D •D >
CO
JO in
Q.
<E D D
n D D • D D n n ffl • •D D m m B D D • D mD • m D • D m • •D D• D D E?3
n absent • present B unknown / unstudied
disintegration of the ESC origin / integration of the lingual prehension ESC FIGURE 8.29. Distribution of phenotypic traits related to feeding across squamate phylogeny. The phylogeny is the same as Fig. 8.1 A except that Amphisbaenia has been added insertae sedis. Note that many taxa have not been studied at all. Taxa for which there are data are inadequately sampled, sometimes represented by a single species. Phenotypic ''traits'' listed on the left are not intended to represent "characters" in the strict sense. Rather, they summarize gross phenotypic characteristics, or suites of characters, discussed in the text. The half white and half stippled box (Scincidae/mesokinesis) indicates that mesokinesis has been shown to be absent in one, possibly atypical, species, but mesokinesis may be present in other scincids. White boxes with small black squares in their upper corners (tongue prehension) indicate those scleroglossan families in which one or two species have been shown to use some form of lingual prehension, although the primitive ingestion mode for the family in each case is jaw prehension. Bold lines indicate the distribution of lingual prehension on the phylogeny, inferred to represent a phenotypically stable functional unit known as an "evolutionarily stable configuration" (ESC). See text for explanation.
266
Kurt Schwenk (D CO T3
0) CO
]D
I
0) to T3
E m
D)
<
0)
O
a>
(U (0
E
to
x: O
CO
-o "c
E
(U (0 •D
ao O)
>» £L
(D CO
5
(U
H
H. o c F >» C!)
CO
o
CO _i
T3
CO CO
•D
X
CO
.^ •§ o
c
CO
c CO
CO
CO
0)
CO
(55
(O (D
c
CD
CD
CO
loss of hyoiingual transport obligate inertial feeding loss of hyobranchial participation in feeding loss of hyoiingual transport obligate inertial feeding
acrodont teeth dental occlusion
bipartite tongue
ballistic tongue projection
inertial feeding common acrodont teeth propalineal jaw movement shearing mechanism
pleurodont teeth streptostyly
jaw - prehension foretongue-hyobranchium uncoupled foretongue modification mesokinesis pharyngeal compression
ANCESTRAL CONDITION: tongue - prehension foretongue-hyobranchium coupled hyoiingual transport pharyngeal packing akinetic skull pleurodont or acrodont teeth? F I G U R E 8.30. Major events in the evolution of feeding systems in lepidosaurian reptiles. Based on data shown in Fig. 8.29 and other information discussed in the text. Due to inadequate sampling of most taxa, some inferences remain tentative.
analysis (e.g., Lauder, 1981,1990). Outgroup comparison is widely acknowledged to be the best method for determining character state polarities (Watrous and Wheeler, 1981; Forey et al, 1992), i.e., establishing primitive (ancestral or plesiomorphic) vs derived (apomorphic) states. Schwenk (1987, 1988) and Schwenk and Throckmorton (1989) pointed out that lingual prehension of small prey is restricted to Iguania, whereas Scleroglossa universally employs jaw prehension (Fig. 8.29). Using Sphenodon as an outgroup, the criterion of parsimony argues unambiguously that lingual prehension is the ancestral ingestion mode for squamates (one rather than two evolutionary transformations in prehension mode). Lingual prey capture in Sphenodon is nearly identical to that seen in Iguania, with only minor differences (see earlier discussion), so there is little doubt about the homology of ingestion mechanisms in the two taxa. Thus, the commitment of scleroglossans to jaw prehension of small prey is a derived condition in Squamata (and the ancestral ingestion mode for all scleroglossans). As discussed previously, it has been shown that a
few scleroglossan species also employ a form of lingual prehension. Does this change our conclusion about the polarity of ingestion modes, i.e., is it possible that jaw prehension is primitive for squamates? The cases of lingual prehension among scleroglossans do not affect the conclusion that lingual prehension is primitive in squamates for the following reasons. First, each lingual-feeding scleroglossan species represents an isolated case within its genus or family and the tongue prehension behavior is not shared by closely related species (so far as we know). Therefore, it is clear in each case that tongue prehension is secondarily derived and that jaw prehension is the ancestral ingestion mode for the genus or family to which each species belongs. Second, with the exception of T. scincoides, lingual prehension in scleroglossans lacks "detailed similarity" to the iguanian condition and therefore fails the most important criterion of homology. In other words, lingual feeding in scleroglossans is probably not homologous to lingual feeding in iguanians (and Sphenodon), but was reinvented from a jaw feeding ancestor. In T. scincoides, it was shown earlier that kinematic similarity
8. Feeding in Lepidosaurs with iguanian lingual prehension is achieved by different means, also hinting at the lack of homology. Thus, both parsimony and homology arguments show that lingual prehension in scleroglossan squamates is secondarily derived and that similarity to the iguanian condition represents homoplasy, not symplesiomorphy. The evidence, therefore, strongly supports the conclusion that jaw prehension is the ancestral ingestion mode in Scleroglossa, but this pattern is consistent with either polarity hypothesis. Thus, the parsimony arguments presented earlier are unaffected by these data, and the conclusion that lingual prehension is the ancestral ingestion mode in squamates, based on outgroup analysis, is supported. To summarize, ancestral lepidosaurs captured small prey with their tongues. Early in squamate evolution, however, a cladistic bifurcation led to two major lineages, Iguania and Scleroglossa. Iguanians retained the ancestral ingestion mode. Within this lineage the lingual prehension mechanism was modified in small ways, and in chameleons the novel element of ballistic projection was introduced, but all systems have remained within a phenotypic space circumscribed by the lingual prehension mechanism. Scleroglossans, however, departed from the ancestral pattern very early in their history and evolved a novel method of capturing small prey with the jaws (Figs. 8.29 and 8.30). Inferences about the evolutionary processes underlying this pattern are presented in the final section of the chapter. B. Post-Ingestion Feeding Stages Despite differences in ingestion mode, lepidosaurs are remarkably consistent in the kinematics and muscle activity patterns of subsequent feeding stages (Smith, 1984; Bramble and Wake, 1985; Schwenk and Throckmorton, 1989; Bels et al, 1994; Herrel et ah, 1997a). This apparently stems from the common use of a homologous, fundamentally similar hyolingual-feeding mechanism (Bramble and Wake, 1985). If similarity in lepidosaurian-feeding function is imposed by the mechanics of the hyolingual system, it is not surprising to see the greatest deviation from the common pattern in cases where the system is circumvented. This is most apparent, for example, in jaw prehension and in inertial transport. It may also help explain the extensive remodeling of the feeding apparatus evident in snakes, as compared to other lepidosaurs. Extreme tongue and hyobranchial reduction early in snake ancestry (presumably related to chemoreception) deprived snakes of the ancestral hyolingual transport system, but at the same time released them from its mechanical constraints. Many snake-feeding specializations, including
267
the mobile jaws of scolecophidians and asymmetrical skull kinesis in alethinophidians, reflect novel solutions to the problem of prey transport and swallowing. Varanids, with superficially similar reduced tongues, might be expected to show comparable departures, but varanids differ fundamentally from snakes in their retention of a robust hyobranchial apparatus. Thus varanids deviate from the common pattern in emphasizing the jaws for prey capture and inertial transport, but pharyngeal packing and swallowing remain essentially "lizard-like" due to their continued reliance on the hyobranchium during these feeding stages. Despite general similarities, scleroglossans depart to some extent from the basal lepidosaurian pattern in the way they swallow. Iguanians rarely employ pharyngeal compression after pharyngeal packing, and when they do it is very brief and barely evident. In contrast, pharyngeal compression is commonly observed in many scleroglossan species. Given that the posterior limbs of the tongue are used to tamp food into the esophagus during pharyngeal packing, it follows that in taxa with reduced hindtongues the efficacy of pharyngeal packing is also reduced, for which they compensate by compression of the pharynx (Herrel et ah, 1999b). Iguanians, with well-developed posterior limbs, are therefore able to complete swallowing (movement of food into the esophagus) with little or no need for pharyngeal compression, whereas scleroglossans with reduced hindtongues must complete swallowing with a compressive stage. It seems generally true that scleroglossans with the most reduced hindtongues, such as varanids and teiids, exhibit the most dramatic pharyngeal compression. Other factors might contribute to the increased use of pharyngeal compression by scleroglossans, notably body and neck elongation. Body elongation and even limblessness are hallmarks of scleroglossan, but not iguanian evolution (e.g.. Camp, 1923). Neck elongation might render pharyngeal packing with the tongue less effective due to the greater length of the pharynx. At the same time it would make the constrictor colli more effective in pharyngeal compression due to its enlargement and freedom from the mandible posteriorly. Elongation would also enhance the ability to bend the head and, in extreme cases, to permit the formation of a propagated wave. Once again, snakes might represent an extreme manifestation of this tendency; sinusoidal body waves and internal concertina flexion are interpretable as derived forms of pharyngeal compression that assist in the final phases of swallowing (see Chapters 2 and 9). Flowever, the fact that short-necked scleroglossans, such as Tiliqua, typically use pharyngeal compression as well (Herrel et ah, 1999b) suggests that tongue reduction alone may be sufficient to
268
Kurt Schwenk
explain the predominance of pharyngeal compression among scleroglossans. C. Evolution of the Gape Cycle Based on a limited data base. Bramble and Wake (1985) postulated that tetrapods share a fundamental pattern of feeding behavior involving coordinated movements of the jaws, tongue and hyobranchial apparatus underlain by similar muscle activity patterns. They summarized this pattern of relationships in a generalized model cycle showing predicted movements of various elements relative to a standard gape cycle exhibiting discrete SO I, SO II, FO, FC, and SC-PS phases (see earlier discussion and Chapter 2). Studies of feeding in lepidosaurians since 1985 have tended to support the notion of a fundamental feeding pattern (e.g., Schwenk and Throckmorton, 1989; Herrel et al, 1997a, 1999b; Herrel and De Vree, 1999a). Nonetheless, some lepidosaurian studies have found deviations from the relationships predicted by the BrambleWake model, raising the question of its generality (e.g.. Smith, 1984; Delheusy and Bels, 1992; Bels et al, 1994; Herrel et al, 1996a; Herrel and De Vree, 1999a). Qualitative departures from the model are usually found in the shape of the gape profile or the timing of tongue movement relative to the jaws. Whether one believes that lepidosaurian data support the Bramble-Wake model or refute it depends on one's expectations for such a model. If taken as a literal, point-by-point prediction for all feeding cycles in all lepidosaurs, then clearly the model is rejected. However, if one's expectation is for the model to serve as a generally predictive, heuristic set of organizational rules, then the model is strongly supported by lepidosaurian studies. I take the latter view. The model was based on intraoral transport and chewing stages because these comprise rhythmic, hyolingually mediated cycles that are most plausibly related to control by a central pattern generator. It is therefore not surprising that most kinematic deviations from the model have been observed during ingestion and swallowing stages, and in cases where the hyolingual mechanism is not employed—jaw prehension and inertial feeding, for example (see earlier discussion). Paradoxically, these exceptions offer the strongest support of all because their very deviance is a predictable outcome of the model. For example, the model gape cycle predicts a characteristic sequence of discrete SO I, SO II, FO, FC, and SC-PS phases due to the coordination of the tongue and hyobranchium with the jaws. Ingestion in scleroglossans predictably deviates because during jaw prehension the jaw-tongue linkage postulated by the model is broken. It is therefore not surprising that
scleroglossan ingestion cycles are variable, often exhibiting spiked or bell-shaped gape profiles without discrete phases. Inertial feeding, likewise, breaks the tongue-jaw couple and shows variable gape patterns. More problematic is that some cycles in which the jaw-tongue linkage is maintained fail to exhibit a complete set of discrete phases. Most often this results from a blurring of SO I and SO II phases into a single SO phase and, occasionally, the loss of a discrete SO phase altogether. However, such "deviant" cycles usually precede, follow, or are interspersed with more typical (i.e., "model-like") cycles. Given the extreme extent of modulation possible during SO (a major prediction of the model), such variation in gape profiles is to be expected as bolus position and condition changes with each gape cycle. Absolute adherence to the model is an unreasonable expectation for a biological system. Critics might argue that too liberal an acceptance of variation renders the model unfalsifiable (see Smith, 1994). While this is true, the fact remains that the kinematics and muscle activity patterns of feeding in lepidosaurs conform to the predictions of the BrambleWake model often enough to demonstrate its merit. Even as we amass the inevitable exceptions to its predicted patterns, the model should continue to serve as a useful guide because in most cases it effectively points to the potential causes of deviation. Fifteen years after Bramble and Wake (1985), lepidosaurian gape cycles can be characterized in the following way: 1. Rhythmic intraoral transport and chewing cycles most often conform to the model gape pattern of discrete phases (SO I, SO II, FO, FC, SC-PS). When deviations occur they usually take one of the following forms: (i) SO I and SO II vary in their relative lengths; (ii) SO I and II sometimes merge into a single SO phase, i.e., there is no "plateau" in the gape profile; (iii) occasionally, SO is absent and the gape profile is spiked or bell shaped; (iv) SC is sometimes absent and may not include a PS component; (v) coordination of hyolingual movement sometimes varies between transport and chewing cycles, with chewing cycles tending to deviate from the model; and (vi) inertial transport cycles deviate from the model because the hyolingual apparatus does not participate, thereby breaking the j a w tongue linkage (a postulate of the model; see earlier discussion). 2. Lingual ingestion cycles in Sphenodon and Iguania typically conform to the model and are similar to transport cycles with the following exceptions: (i) hyolingual protraction during SO is more extreme and carries the tongue outside the mouth; and (ii) SO II often occurs at a higher percentage of the maximum
8. Feeding in Lepidosaurs gape angle than in transport, presumably to allow passage of the prey item past the jaws into the mouth. In T. scincoides, the only scleroglossan so far described to use iguanian-like lingual ingestion, the gape cycle is similar, although it may be arrived at by a different mechanism. 3. Jaw ingestion cycles in scleroglossans deviate from the model because the hyolingual apparatus does not participate, thereby breaking the jaw-tongue linkage. The gape cycle is typically spiked or bell shaped. Jaw ingestion cycles in iguanians are too little known to draw any conclusions. However, based on the model and Schwenk and Throckmorton's (1989) finding that iguanian jaw prehension actually represents truncated lingual prehension, iguanian jaw prehension should be similar to lingual ingestion cycles, i.e., conform to the model, with the expectation that deviations, if they occur, will be found in SO II. 4. Swallowing cycles are at present too poorly differentiated from transport cycles in most studies to make supportable generalizations. Pharyngeal packing cycles are expected to conform most often to the model; however, the introduction of tip-first, extraoral tongue protrusion and side-to-side asymmetry during this stage suggests the potential for significant deviation, particularly in later cycles. Pharyngeal compression is not expected to conform to the model cycle because it breaks the jaw-tongue linkage on which the model is predicated. The limited data available support this conclusion. Finally, the dependence of swallowing behavior on bolus characteristics (food type, size, and condition) makes a high degree of variation in swallowing cycles more likely. D . Tongue Evolution In general there is a clear distinction between iguanian and scleroglossan squamates in many aspects of tongue form (see Schwenk, 1988,1993,1995; Schwenk and Throckmorton, 1989; Wagner and Schwenk, 2000). The distinction is particularly pronounced in the foretongue. Sphenodon is similar to iguanians in most respects. Except for the tongue tip, iguanians are characterized by long, slender "high-profile'' papillae that are densely glandular. In contrast, scleroglossans always have "low-profile" papillae on the foretongue or sometimes no papillae at all, and the foretongue is always aglandular (Schwenk, 1984, 1988). Iguanians often show epithelial elaborations at the apex of each papilla that increase the tongue's rugosity (e.g., plumose cells), but in scleroglossans the epithelial surface of the foretongue is smooth, firm, and lightly keratinized. The iguanian foretongue is usually broad and deep (thick), but in scleroglossans it is always reduced in
269
one or both dimensions. It is typically tapered and in some cases is very slender. In iguanians the tongue tip is never more than slightly notched, but in scleroglossans (with the exception of dibamids), particularly autarchoglossans, it is more deeply cleft and in some taxa it is forked and attenuate. In iguanians the lingual process is typically robust, extends most of the tongue's length, and is often "kinked" at its apex, but in scleroglossans it is usually slender, rarely extends more than half the tongue's resting length, and, in some cases, is detached from the basihyal or fails to penetrate the tongue at all. The laryngohyoid ligament usually attaches to the lingual process near the tongue tip in iguanians, but attaches farther back in scleroglossans. In most iguanians the genioglossus muscles insert into the tongue anteriorly, but in scleroglossans they insert relatively farther back, sometimes extremely so. Altogether, the proportion of the foretongue that is free of the floor of the mouth and the lingual process (and therefore the hyobranchium) are much greater in scleroglossans. Circular muscles around the hyoglossus bundles, implicated in hydrostatic elongation of the tongue (Smith, 1984; Smith and Kier, 1989), are weakly developed and sometimes incomplete in iguanians, but in scleroglossans they are well developed. These statements are generalizations and the exceptions are certainly worth exploring, but for the most part they accurately characterize each group. This morphological dichotomy is mirrored precisely by the functional dichotomy discussed earlier: iguanians use the foretongue as a prehensile organ to capture relatively small prey items (and sometimes large prey as well), whereas scleroglossans (with a very few exceptions) use the jaws and teeth for prehension of virtually identical prey types. In iguanians, vomeronasal chemoreception is slightly or moderately well developed (Schwenk, 1993,1995b) and tongue flicks are limited to short, simple extensions (Gove, 1979; Bels et al, 1994; Herrel et al, 1998c). In scleroglossans, vomeronasal chemoreception is relatively more highly developed, sometimes extremely so (e.g., Schwenk, 1994e). Tongue-flick protrusion distances are typically large and flicks are often kinematically complex, including rapid, multiple oscillations in some taxa (Gove, 1979; Bels et al, 1994). In iguanians, most tongue flicks contact the substrate, but in scleroglossans, tongue flicks are frequently directed into the air to sample volatile chemicals. Thus, it is reasonable to interpret many features of the iguanian tongue in light of its role in prey prehension and scleroglossan departures from this primarily as specializations related to enhanced performance of tongue flicking and vomeronasal chemoreception (see later). My interpretation of these patterns is that there is
270
Kurt Schwenk
a functional-morphological distinction between hindtongue and foretongue in most lepidosaurs such that the hindtongue is coupled anatomically and functionally to the hyobranchium whereas the foretongue is relatively uncoupled. Taxa vary in the proportion of the foretongue that is free of hindtongue coupling. In iguanians, relatively little of the foretongue is free, whereas a much larger part is free in scleroglossans. Lingual movement that is independent of the hyobranchium is concentrated in the foretongue and is primarily hydrostatically generated. It may involve some translation along the lingual process, particularly in iguanians where the process extends for most of the tongue's length, although it primarily results from intrinsic, hydrostatic elongation with concomitant reduction in foretongue diameter. Although scleroglossans are more specialized for foretongue mobility, iguanians are also capable of limited hydrostatic elongation of the anteriormost portion of the foretongue, which is evident when the tongue is protruded tip first, as during tongue flicking and lapping. Cineradiography indicates that the hyobranchium is not protracted during these behaviors in iguanians (Smith, 1984; Schwenk, unpublished observations) and flattening of the tongue is observed (see Fig. 1 in Schwenk, 1995b), as predicted by the hydrostatic model of tongue elongation (Kier and Smith, 1985). In some iguanians, notably most agamine agamids and the iguanid genus Phrynosoma, the tongue is capable of remarkable extraoral protrusion distances during ingestion (e.g., Schwenk and Bell, 1988). However, most of this protrusion is coupled to hyobranchial protraction and protrusion of the lingual process, as during the "aiming" stage of chameleon feeding. Given the extensive attachment of the tongue to the hyobranchium, this kind of protrusion requires the laryngohyoid ligament to be stretched or the larynx to be protracted along with the tongue and hyobranchium. The latter is clearly the case in Sphenodon and some iguanians, but the situation may vary among species (see Section IV). The analysis just described has identified two dualities related to the lepidosaurian tongue: the morphological duality of foretongue and hindtongue and the functional duality of feeding and chemoreception. General patterns or "strategies" of tongue evolution reflect an interplay between these two dualities. In this context, four types of lepidosaurian tongue can be identified (Fig. 8.29). 1. Feeding type {Sphenodon and Iguania): This type represents a near total commitment of tongue form to feeding function. The tongue is broad, deep, and muscular and is covered with high-profile, prehensile pa-
pillae. The tip is unnotched or slightly notched. Most of the tongue is coupled anatomically and functionally to the hyobranchium. The posterior limbs are robust. In effect, the foretongue-hindtongue duality is minimal. Although iguanians possess a vomeronasal system, which they stimulate through tongue flicking, vomeronasal function and evolution are constrained by the lingual feeding system (see Section VII,F). The anteriormost portion of the foretongue is modified in iguanians to permit limited hydrostatic elongation and tip-first protrusion, but the minimal commitment of iguanian tongue form to chemosensory function is indicated by its extreme similarity to the condition in Sphenodon, which lacks the behavior of tongue flicking altogether. 2. Compromise type (Gekkota and Scincomorpha, including Amphisbaenia): This type represents a compromise between feeding and vomeronasal function. The foretongue no longer participates in lingual prehension and is highly modified for greater tongueflicking performance (and eye-wiping in gekkotans), but the tongue remains important in hyolingual transport and swallowing, functions served especially by the hindtongue. Although the foretongue-hindtongue duality is evident functionally, morphologically there is a continuum between them with the foretongue most modified and free of the hyobranchium, and the hindtongue tending to retain hyobranchial coupling and other plesiomorphic attributes, especially in gekkotans. The tongue tip ranges from notched to deeply forked. The posterior limbs are usually well developed, but in some derived forms they are reduced or lost. Both foretongue and hindtongue remain papillose, but papillae are low profile, especially on the foretongue. Because the tongue retains its ancestral function during postingestion feeding stages, the kinematics of these stages also tend to retain the primitive pattern. 3. Bipartite or diploglossan type (Anguimorpha except varanids and snakes): In this type both functional and morphological dualities are extremely developed and clearly evident. The tongue is literally divided into an anterior portion devoted to chemoreception and a posterior portion devoted to feeding function. There is a sharp transition between hindtongue and foretongue evident as a crease or "retractile" zone and marked by a sudden, dramatic change in papillary height and glandularity. The foretongue is slender, smooth, histologically specialized for hydrostatic elongation, and entirely free of the hyobranchium. The papillae are low profile (in Lanthanotus they are lost), aglandular, and lightly keratinized, and the tongue tip is deeply cleft or forked. The foretongue is functionally devoted to tongue flicking and chemoreception with minimal
8. Feeding in Lepidosaurs participation in feeding function at any stage. The vomeronasal system is highly developed in these taxa. In contrast, the hindtongue essentially retains the plesiomorphic condition with long, glandular papillae and a tight coupling to the hyobranchium. It participates fully in transport and swallowing stages of feeding. Thus, postingestion-feeding kinematics resemble the primitive condition due to retention of the hyolingual transport system despite radical modification of the foretongue. 4. Chemosensory type (Varanidae and snakes): In these taxa the tongue's feeding function is lost not only in ingestion, but in postingestion stages as well. Therefore, the tongue is almost entirely committed to tongue flicking and chemosensory function. The ancestral function of the hindtongue in feeding is lost and it has been modified into a part of the tongue sheath. The entire oral portion of the tongue is free of the floor of the mouth and the hyobranchium. What is left is essentially a greatly expanded foretongue devoid of papillae and deeply forked. The kinematics of varanid swallowing retain some similarity to the ancestral pattern due to continued participation of the hyobranchium, if not the tongue, but in snakes the extreme reduction of the hyobranchial apparatus is associated with complete remodeling of the feeding system and novel kinematic patterns.
E. Dietary Specialization 1. Diet vs Phenotype This topic is deserving of a detailed analysis beyond the scope of this chapter. Several general points need to be emphasized, however. Despite the widespread assumption to the contrary, there is no necessary relationship between dietary specialization and phenotypic specialization in the lepidosaurian-feeding apparatus (Greene, 1982; Schwenk, 1988). A narrow or specialized diet leads to phenotypic specialization in some taxa, but not in others. This is a critical observation because it implies that, with only a few exceptions (see Section IV,B), we cannot, with any confidence, infer lepidosaur diet from morphology (see Chapter 1). This is clearly illustrated by the example of the amphisbaenian-feeding system. Despite mechanical analyses and laboratory observations strongly suggesting adaptive specialization of the system for killing and reducing large, vertebrate prey, studies of natural diet have shown that, if anything, amphisbaenians consume a disproportionate number of small invertebrates. Armed with this dietary data it may be possible to devise testable, or a least plausible, explanatory hypotheses for the origin of the unique amphisbaenian
271
morphology (see earlier discussion), but the lesson is clear: there can be no facile connection of form to diet. The fundamental importance of natural history data to functional morphology is indicated (see Chapter 1). If diet cannot be inferred from morphology, neither can morphology be predicted from diet. The case of myrmecophagy (ant feeding) illustrates this point well. Most species of the North American iguanid genus, Phrynosoma, consume mostly ants. Most share a bizarre phenotype distinct from related iguanids: wide, dorsoventrally compressed body, splayed limbs, broad head, abbreviated snout, occipital spines, relatively slow movement, long activity period, variable body temperature, and a very large stomach relative to body size (Pianka and Parker, 1975). They are some of the only lizards known to show diet-based tongue modification (Schwenk and Sherbrooke, manuscript in preparation). Pianka and Parker (1975) related all of these traits to a diet of ants: (i) ants are small and chitinous, thus many must be eaten for adequate nutrition; (ii) this requires a large stomach to store the ants; (iii) a large stomach requires a broad, tank-like body; (iv) the need to eat many small prey requires a long activity period to extend foraging time; (v) a long activity period exposes the lizard to more predators and the tanklike body makes rapid escape behavior impossible, therefore cryptic behavior and body spines are necessary for defense; and (vi) long periods of foraging in the open require relaxed thermoregulation, resulting in a high variance in body temperature. Pianka and Parker (1975:156) concluded: "Thus, Phrynosoma platyrhinos, and perhaps other members of the genus Phrynosoma, seem to be characterized by a unique constellation of anatomical, behavioral, physiological and ecological adaptations that facilitate efficient exploitation of ants as a food source and set the horned lizards apart from most other species of lizards." There is no doubt that Phrynosoma represents a phenotypically specialized lizard and that many of its putative adaptations are related to its myrmecophagous diet. Pianka and Parker's (1975) arguments are so plausible that they are sometimes viewed as the inevitable consequence of ant specialization in lizards. Remarkable phenotypic convergence in Moloch horridus (Agamidae), Phrynosoma's ecological counterpart in Australia (Pianka and Pianka, 1970), has bolstered the popular view that this is what ant-eating lizards must look like. However, Pianka and Parker (1975), themselves, were at pains to point out that Moloch differs from Phrynosoma in several key features and that their "integrated view of Phrynosoma ecology clearly does not apply in general to all ant-eating lizards." A brief survey of myrmecophagy in lizards shows convincingly that Pianka and Parker's (1975) "inte-
272
Kurt Schwenk
grated view" not only does not apply to all ant-eating lizards, it may not apply to any other ant-eating lizards, apart from Moloch, to some extent. Table 6 lists lizard and amphisbaenian species in which 30% or more of the diet is ants; some species consume almost nothing but ants. None show phenotypic specialization comparable to Phrynosoma and all are more similar to their nonmyrmecophagous relatives than to other ant-eaters (some traits, such as length of activity period and variance in body temperature, are unknown for most). As such, they retain a generalized phenotype and do not show specialization in the feeding apparatus or otherwise for extreme myrmecophagy. Table 8.6 also lists termite-feeding species because termites are similar to ants as a prey type, and in mammalian biology ant and termite feeding are often grouped together under the rubric of myrmecophagy (see Chapter 15). There are no species showing obvious phenotypic specialization for feeding on termites, with the possible exception of amphisbaenians. Seasonal or ontogenetic stenophagy, or geographic variation in diet, may account for the failure of most ant and termite eaters to evolve phenotypic specializations (see Section I1I,A), however several of the species listed in Table 8.6 are well known to consume virtually nothing but ants or termites. Furthermore, the Pianka and Parker (1975) scenario might apply only to open habitat and desert species or it might depend on the social structure and foraging behavior of the particular ant or termite species preyed upon. However, these points only reinforce the conclusion that the "integrated view" is case specific and therefore has little explanatory value outside its specific realm. It cannot serve as a model of phenotypic specialization for myrmecophagy in lizards. 2. What Is a Specialized Diet? The amphisbaenian example discussed previously called attention to the importance of natural history data to functional analysis. Functional analyses should be cognizant of natural diet, the manner in which food is procured in the field, and the mechanical tasks relevant to animal performance implied by these. For example, Herrel and De Vree (1999a) thought that "reduction of particle size is of prime importance" for herbivorous lizards, but in fact particle size reduction is usually minimal in herbivorous species because dental adaptations allow them to crop mouth-sized portions of food during ingestion that are rapidly transported and swallowed with little or no processing (e.g., Throckmorton, 1976). The assumption that particle size reduction is a necessary part of herbivory stems from a mammalian bias and the failure to appreciate the significance of cropping behavior during ingestion in
many herbivorous lizards. In other words, the mechanical tasks actually required of the feeding system in a folivorous lizard are very different from the tasks that might be assumed to be important. Most of the apparent dental specialization in Uromastix, including dental occlusion and the development of wear facets, are better interpreted as adaptations for initial cropping function than for particle size reduction during chewing. Given that mechanical attributes of food affect the behavior of the feeding system (e.g., Bels and Baltus, 1988; Herrel et ah, 1999b), it is critical to duplicate or least, approximate, natural food type, form, and presentation in order to reveal the functional linkages underlying the evolutionary relationship between phenotype and diet. Extraordinarily few functional studies of lepidosaur feeding have attempted to do this. Another point related to natural diet and the mechanics of feeding is the perception of what constitutes dietary specialization in the first place. Herbivory, once again, serves to make the point. There is a long tradition in the herpetological literature of regarding herbivorous lizards as dietary specialists, apparently because most lizards eat a variety of invertebrates or are, to some extent, omnivorous. Yet, as pointed out previously in Section III,A, "herbivory" implicates a diversity of potential food types, including leaves, stems, shoots, flowers, seeds, pollen, nectar, and fruits of various kinds, each potentially requiring different abilities to handle. A lizard that consumes several or all of these plant parts may be just as much a dietary generalist as an omnivorous species that eats both animal and plant food, even though the herbivore consumes food items from a more restricted taxonomic group. In other words, the traditional notion of dietary specialization is based on taxonomic restriction of the foods taken, not the diversity of mechanical tasks required to eat them. This is a strictly ecological notion of dietary specialization that in many cases may not be relevant to the question of dietary specialization in a functional sense. It is the latter type of dietary specialization that relates directly to the evolutionary question of phenotypic specialization in the feeding apparatus. As such, it is possible that the taxonomically disparate food types of fallen fruit and earthworms present a common challenge to a lizard because they are both wet, compliant foods requiring similar mechanisms for ingestion and processing, whereas a slug and a snail represent radically different food types because the latter is shelled, despite the fact that both are gastropod molluscs. Clearly, organisms adapt to the mechanical tasks required to eat, not to food taxon per se, yet we persist in basing our identification of dietary specialists and generalists on taxon-based food categories. This is reasonable only to the extent that a given taxonomic group shares a common set of mechanical
273
8. F e e d i n g in L e p i d o s a u r s TABLE 8.6
A Partial List of Ant- and Termite-Eating Squamates^ Termites
Ants Species Iguanidae Anolis aeneus A. bonairensis^ A. oculatus Liolaemus monticola^ Sceloporus graciosus S. occidentalis S. olivaceus S. undulatus Tropidurusflaviceps^ T. hispidus T. (Plica) umbra^ Uma paraphygas U. scoparia
Stamps et ah (1981) T. Schoener (personal communication) Bullock etfl/. (1993) Jaksic et al (1979) Rose (1976) Rose (1976) Kennedy (1956) Hotton (1955) Vitt and Zani (1996) Vitt et al. (1996b) Witt etal {1997b) Gadsden and Palacios-Orona (1977) Pianka (1986)
Agamidae Agama hispida Ctenophorus fordi C. isolepis C. scutulatus Draco maximus ^ D. melanopogon D. obscurus D. quinquefasciatus^ D. volans^
Pianka (1986) Pianka (1986) Pianka (1971,1986) Pianka (1986) Inger(1983) Inger (1983) Inger(1983) Inger(1983) Auffenberg (1980)
Gekkonidae Cosymbotus platyurus
Auffenberg (1980)
Scincidae Apterygodon vittatus^ Egernia inornata Leiolopisma tricolor
Mori et al (1995) Pianka (1986) Bauer and De Vaney (1987)
Lacertidae Acanthodactylus erythrurus
S. Busack (personal communication)
Amphisbaenia Amphisbaena alba Monopeltis sphenorhynchus
Colli and Zamboni (1999) Broadleyetfl/. (1976)
Source
Species
Source
Iguanidae Enyalius leechii Tropidurus hispidus
Vitt effl/. (1996a) Vitt gffl/. (1996b)
Agamidae Agama impalearis Caimanops amphiboluroides ^ Draco obscurus D. melanopogon
Znari and Nagy (1997) Pianka (1986) Inger (1983) Inger (1983)
Gekkonidae Chondrodactylus angulifer Colopus wahlbergi Diplodactylus conspicullatus^ D. elderi D. pulcher^ Gehyra variegata Hemidactylus frenatus Pachydactylus bibroni P. capensis Ptenopus garrulus Rhynchoedura ornata^
Pianka (1986) Pianka (1986) Pianka and Pianka (1976) Pianka (1986) Pianka and Pianka (1976) Pianka (1986) Auffenberg (1980) Pianka (1986) Pianka (1986) Pianka (1986) Pianka and Pianka (1976)
Scincidae Cryptoblepharus boutonii Ctenotus ariadnae C. atlas C. calurus C. dux C. grandis C. helenae C. leonhardii C. pantherinus ^ C. quattuordecimlineatus C. schomburgkii Egernia depressa E. striata Lerista bipes L. mueleri Mabuya spilogaster M. frenata M. variegata Menetia greyii Morethia butler Tiliqua branchialis Typhlosaurus gariepensis^ T. lineatus^
Auffenberg (1980) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Twigg et al (1996) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Vrcibradic and Rocha (1998) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986) Pianka (1986); Brain (1959)
Lacertidae Eremias lineo-ocellata E. lugubris^ E. namaquensis Ichnotropis squamulosa Meroles suborbitalis
Pianka Pianka Pianka Pianka Pianka
Teiidae Cnemidophorus uniparens
Eifler and Eifler (1998)
Amphisbaenia Amphisbaena darwinii^ A. mertensii Cercolophia roberti Dalophia pistillum Monopeltis anchietae M. capensis M. leonhardi
Cabrera and Merlini (1990) Cruz Neto and Abe (1993) Cruz Neto and Abe (1993) Broadley^ffl/. (1976) Broadleyefd. (1976) Broadley^tfl/. (1976) Broadley^ffl/. (1976)
'^Only species in which 30% or more of the diet consists of each prey type are listed. ^Species in which ants or termites compose virtually the entire diet.
(1986) (1986) (1986) (1986) (1986)
274
Kurt Schwenk
attributes as food. This is clearly not the case for herbivory. Should we expect similar feeding adaptations in a folivorous iguana, a frugivorous varanid, and a nectivorous gecko? Dietary specialization cannot be assumed because a given species feeds upon food types belonging to a single taxonomic group. Rather, dietary specialists are those taxa that consume a set of food items presenting the same mechanical challenge, regardless of the food's taxonomic affinity. As such, herbivores are no more likely to be dietary specialists than insectivores or omnivores. As for herbivory, ''myrmecophagy" does not necessarily imply dietary specialization in a functional sense. Specialization is assumed because ants are viewed as uniformly small, chitinous prey lacking in nutritive value and defended by formic acid, biting, or stinging. In fact, formicids are highly variable in size, form, biochemical composition, social structure, foraging, and defensive behavior. For example, a single Moloch horridus consumes as many as 2500 tiny, innocuous ants {Iridomyrmex) at a time, each as small as 0.002 cc in volume (Pianka and Pianka, 1970), whereas a Phrynosoma platyrhinos typically eats fewer (<100), much larger (-^1 cm), highly venomous ants {Pogonomyrmex; approximately 1 cm long) (Pianka and Parker, 1975; Schmidt and Schmidt, 1989). Is feeding on tiny, harmless ants comparable to feeding on large, noxious ants? Comparing the extremes, a small lizard eating a large ant does not face the same mechanical challenge as a large lizard eating a small ant, thus a diet of "ants" does not necessarily indicate commonality in feeding mode. Dietary specialization in this case might be expressed more accurately as "feeding on many small prey" vs "feeding on few large prey." It is likely that some species of ants are comparable to other types of insect as prey and require only a standard feeding apparatus for their consumption, whereas others, such as the highly toxic and pugnacious species of Pogonomyrmex fed upon by Phrynosoma, depart from typical insect prey and require specializations for their effective exploitation (e.g., blood factors to detoxify the venom; Schmidt and Schmidt, 1989; Schmidt et ah, 1989). In summary, phenotype and diet show no predictable association among lepidosaurs. The rules governing the relationship between ecological and phenotypic specialization remain unknown. The prediction of diet from form and form from diet is undertaken at considerable risk. Most important, the use of taxonomic categories to imply dietary specialization is based on ecological principles relating trophic levels and is of little relevance to functional and evolutionary analyses of feeding. Phenotypic specialization of the feeding apparatus, if it occurs, depends on the mechanical challenges posed by a given food type, not its
taxonomic affiliation. Therefore, the assumption by morphologists of dietary specialization or limitation based on traditional ecological-taxonomic categories, such as "herbivory" and "myrmecophagy," maybe unfounded. Morphologists should devise new, functional, or mechanistic conceptions of "dietary specialization" if they are to relate diet to the functional morphology of feeding and therefore phenotypic evolution. The distinction between functional and trophic specialization does not imply that ecological issues are irrelevant to the question of phenotypic specialization in the feeding apparatus. It is possible that the ultimate question of why some species become phenotypically specialized relates to differences in the ecological milieau, such as resource overlap and competition (e.g., Greene, 1982; Pianka, 1986). At a proximate level, however, it is the "fit" between phenotype and food type that determines the performance of the feeding system. If nothing else, this review of lepidosaur feeding and diet suggests that a given lepidosaur-feeding system can accommodate a large variety of food types without an apparent sacrifice in performance. This may explain, in part, why the feeding apparatus of most species manifests phylogenetic more than ecological effects. The stability of feeding systems regardless of ecology is discussed further in the next section. F. Feeding Systems, Functional Units, and Evolutionary Constraint Organisms comprise many different phenotypic traits or characters that in an idealized sense are each capable of evolving independently In reality we recognize that suites of characters are often associated by genetic, developmental, and functional interactions. The extent to which these associations affect evolutionary trajectories of individual characters and therefore the evolution of organismal phenotype depends on the nature and magnitude of the interactions. Functional associations among characters are apparent in the operation of various body systems. Systems comprising tightly integrated sets of functionally interrelated characters are often identified by morphologists as "functional units" (Schwenk, 2000b). Different types of functional units have been described, but morphologists are most often concerned with dynamically interacting sets of characters that work together in a coordinated way to produce a particular functional output. Wagner and Schwenk (2000) developed a model of such a functional unit called an "evolutionarily stable configuration" (ESC). The model was developed around the example of lepidosaurian-feeding systems and is briefly summarized here to support the interpretation of feeding system evolution that follows.
8. Feeding in Lepidosaurs An important property of an ESC is that the functional interactions among characters within it impose strong, "internal selection" on character phenotypes. Internal selection arises from the need for functionally interacting characters to "intermesh" in a particular way so that the web of causality among characters in the system remains unbroken and system functionality is maintained. In other words, as phenotypic variants arise among ESC characters by random mutation, they are "tested" within the organism by system functionality. Variants that diminish or disrupt system performance are likely to have fitness consequences and will be selected against. Only those that maintain or improve system functionality will persist in a population. As such, internal selection is an intrinsic attribute of the organism stemming from system-level function that is not directly dependent on the particular habitat or environment occupied by the organism. The relative independence of internal selection from the environment distinguishes it from the traditional. Darwinian notion of selection ("external selection"), which is, by definition, environment dependent. Internal and external selection are complementary components of natural selection (Whyte, 1965; Dullemeijer, 1980; Arthur, 1997; Schwenk and Wagner, 2000). As a consequence of internal selection on its component characters, phenotypic evolution of the functionally integrated system (the ESC) is constrained, as implicit in most uses of the term "functional constraint" (e.g.. Hall, 1996). Because of the relative independence of internal selection from the environment, ESCs are likely to persist in a lineage for a long time, i.e., remain phenotypically stable through environmental transitions, cladogenesis, and adaptive radiation. Phenotypic evolution of the ESC can occur throughout, but only within a limited phenotypic space circumscribed by the thresholds of functional interactions within the system. In other words, component characters can evolve as long as system-level performance is maintained or improved. One implication of this process is that ESC systems become increasingly self-stabilized as they add new characters and hone their functional interactions. A long-term ESC is likely to be represented in a comparative phylogenetic analysis as "variations on a theme," i.e., a cluster of phenotypically similar systems all serving the same fundamental function. Disruption and loss of an ESC are certainly possible, but are expected to be relatively infrequent (see later). Wagner and Schwenk (2000) proposed that the iguanian lingual prehension mechanism is an ESC. As such, it represents a system of tightly integrated characters whose coordinated interactions perform the function of prey capture with the tongue. Because of the complex interplay among characters during a success-
275
ful prehension sequence, internal selection on components of the lingual feeding system is likely to be strong. For example, a mutation leading to a nonpapillose foretongue would be selected against because it would disrupt the adhesive mechanism essential for prey capture. Mutations that change central pattern generation so that the coordination of jaw and hyobranchial movement is disrupted would similarly be selected against. Foretongue form should be especially subject to strong (internal) stabilizing selection due to its central role in prey prehension. Consequently, the system overall is functionally constrained and phenotypically stabilized. Phylogenetic analysis supports the contention that the lingual feeding system has remained phenotypically stable (Schwenk, 1995a; Wagner and Schwenk, 2000). Outgroup analysis shows that lingual prehension was the ancestral feeding mode for squamates (see earlier discussion), and iguanians have maintained a fundamentally similar system throughout their long history and extensive diversification (Fig. 8.29). Despite radiation into nearly every conceivable lizard niche, occupation of nearly every conceivable habitat and the adoption of nearly every conceivable diet, a lingual prehension mechanism has been maintained. Only chameleons have departed more than trivially from the ancestral pattern, but the insertion of a ballistic projection phase into the prehension sequence has not changed the fundamental phenotype of the feeding system (see earlier discussion; also Wagner and Schwenk, 2000). The evidence for functional constraint imposed by internal selection of the lingual prehension mechanism is bolstered by the likelihood that throughout iguanian history there must have been constant selection for improvement of the vomeronasal chemosensory system. This is inferred indirectly from two, complementary lines of evidence: (1) Iguanians possess vomeronasal organs, which they stimulate by tongue flicking, as do all squamates, but they have never evolved more than kinematically simple tongue flicks and modest vomeronasal function. (2) The absence of lingual prehension in the iguanian sister taxon, Scleroglossa, is correlated with superior vomeronasal function and more complex tongue flicking behavior, as well as a suite of changes in the morphology of the foretongue (see earlier discussion). The implication of this pattern is that the acquisition of jaw prehension in scleroglossan ancestors, for whatever reason, released the tongue from the constraint of the lingual prehension mechanism so that it could respond to other selection pressures. How and why did the transformation from tongue to jaw prehension occur? Although relatively rare, dissolution of ESCs must sometimes occur. Wagner and
276
Kurt Schwenk
Schwenk (2000) outlined a series of conditions under which disruption of the ESC would be expected. Two of these are relevant here: a key environmental shift and duplication of the ESC function by another set of characters. They offered a speculative scenario to show that, despite the highly integrated nature of the ancestral lingual prehension system, the transition to jaw prehension might have been relatively straightforward if conditions were precisely right. The key may have been the fact that jaw prehension is not entirely novel in scleroglossans, as it also occurs in iguanians (and Sphenodon) when prey size is large (Schwenk and Throckmorton, 1989). As previously discussed, tongue protrusion distance is modulated according to prey size so that when feeding on large prey, tongue-prey contact occurs at the jaw margins or within the mouth at the same time the jaws and teeth engage the prey. If one postulates a simple environmental shift leading to decreased availability of sniall prey (or an increase in the availability of easily captured, large prey), then selection pressures serving to maintain the lingual prehension apparatus would dissipate and be replaced by selection to increase the efficacy of jaw prehension. Under these particular conditions the function of the lingual prehension mechanism would have been, in effect, duplicated by the jaw apparatus. As with gene duplication, functional redundancy releases the original system from its functional constraints so that it is free to evolve along a new trajectory. In this environment, early scleroglossans might have made the transition to obligate jaw prehension quite rapidly, given that the essential motor patterns and morphology were already in place. The rapidity of this evolutionary transformation would have been enhanced by simultaneous selection tending to modify tongue morphology in ways that interfere with adhesion and prehension (see model proposed earlier). Concomitant selection promoting improvement of the vomeronasal chemosensory system through specialization of the tongue for tongue flicking would have had this effect because a tongue form optimized for tongue flicking is incompatible with a form optimized for prehension (Wagner and Schwenk, 2000; also see earlier discussion). Thus, as jaw prehension replaced lingual prehension in scleroglossan ancestors, the functional constraint imposed on the foretongue by the necessity of lingual prehension was broken, allowing the foretongue to be modified for tongue flicking and other purposes. The price of this transition, however, was the loss of lingual prehension as a viable option, leading to obligate jaw prehension early in scleroglossan history (Figs. 8.29 and 8.30). Although for obvious reasons I have focused on the tongue, the transition to jaw prehension in scleroglos-
sans released a cascade of other phenotypic changes due to disintegration of the lingual prehension ESC. Enhancement of vomeronasal function is also correlated with modifications in neuroanatomy and profound changes in behavior (e.g., Schwenk, 1993b). For example, the evolution of efficient chemical trail following and a tendency for scleroglossans to evolve active foraging modes (see earlier discussion) followed upon the evolution of forked tongue tips and modifications to the brain's amygdaloid nuclei (Schwenk, 1993b, 1994e, 1995b; Martinez-Marcos et al, 1999). Mesokinesis evolved to enhance the efficacy of jaw prehension (see earlier discussion) and, for similar reasons, scleroglossans may be more likely than iguanians to evolve relatively longer, more slender snouts (Schwenk, 1987). These patterns highlight one other component of evolutionary constraint—the notion of trade-offs. The participation of the tongue in several functions, including food prehension, food transport, food swallowing, and chemosensory tongue flicking, leads to a composite of selection pressures tending to "push" tongue form toward optimization of each. The realized phenotype represents the summation or balance of these pressures. It has been argued that lingual prehension, obviously of fundamental importance to survival, is so tightly integrated a function that the internal selection for prehensile foretongue form overwhelms other selection pressure in Iguania. The tongue's role in intraoral transport and, to a lesser extent, swallowing are compatible with its prehensile phenotype, but optimization for tongue flicking is not. Thus, in Iguania, the balance between feeding and chemosensory function is heavily weighted toward feeding. This accounts for the "feeding type" tongue identified previously. This trade-off has the indirect effect, however, of constraining vomeronasal evolution due to the tongue's role in stimulating the vomeronasal organs. Indeed, the functional constraint of lingual prehension was first suggested by the pattern of chemosensory character states across the squamate phylogeny—iguanians retain plesiomorphic states whereas scleroglossans exhibit variable, usually derived states (Schwenk, 1993b, 1995a). Thus a functional constraint in one system (the feeding system) has imposed an evolutionary constraint on another system (the vomeronasal system) through common use of a single organ, the tongue. Given the scenario just described, it is surprising that some scleroglossans have been able to reevolve lingual prehension at all. Wagner and Schwenk (2000) thought that modifications to the scleroglossan tongue should make lingual prehension mechanistically unfeasible given the model of lingual adhesion previously developed. However, Smith et al/s (1999) study
8. Feeding in Lepidosaurs showed that true lingual prehension is possible in at least one species (T. scincoides). As noted, however, Tiliqua and the Egernia group skinks generally might have been preadapted for the reacquisition of lingual prehension. In any case, the model of lingual prehension presented in this chapter suggests that lingual adhesion in Tiliqua is unlikely to be as effective as in iguanians, but this hypothesis remains to be tested. In conclusion, the hyolingual transport system is the central organizing principle of most lepidosaurianfeeding systems and the basis of their fundamental similarity. Ancestrally, the hyolingual system participated in all stages of feeding, including prey capture and swallowing, and this condition has been retained in iguanian squamates. In scleroglossan squamates, the dependence of feeding function on the hyolingual system was incrementally eroded, first by substitution of the jaws for the tongue as the organs of prehension and later by the substitution of pharyngeal compression and inertial transport mechanisms for swallowing and intraoral transport, respectively, in some lineages. These trends achieve their most extreme expression in snakes with the complete loss of hyolingually mediated feeding. Although we might expect feeding systems to be highly sensitive to diet or other ecological variables, this is mostly not the case. Virtually identical systems characterize large phylogenetic groups, irrespective of food type. Even trophically specialized taxa have failed to modify basic aspects of the system. Transitions between feeding modes, as between lingual and tongue prehension, have occurred only rarely, reflecting significant organizational shifts at deep phylogenetic nodes. The evolutionary stability manifested by these systems reflects both their ability to accommodate a diversity of food types and the operation of constraints acting to preserve functional integration. For lepidosaur feeding systems, it is not what is eaten that matters, but how it is eaten.
VIIL FUTURE DIRECTIONS Despite tremendous progress in our understanding of lepidosaur feeding since 1985, this review has pointed to a number of problem areas. We are in the precarious position of having enough data to discern patterns and generate hypotheses, but not enough data in most cases to test the hypotheses. Most of the conclusions presented herein are based on inadequate samples and require additional testing before their generality can be accepted. Some areas particularly worthy of further investigation are suggested.
277
1. Jaw-prehension in iguanians. Although lingual prehension is reasonably well understood, jaw prehension among iguanians is virtually unstudied. Schwenk and Throckmorton (1989) presented a hypothesis of tongue-jaw modulation based on prey size. This hypothesis needs to be tested and the kinematics of jaw prehension described for a variety of species. Is there always a continuum of tongue protrusion distance or is the switch to jaw prehension sometimes a threshold response? Are some species less likely to switch to jaw prehension than others, i.e., are some species obligate tongue feeders? The most phenotypically derived taxa (e.g., Phyrnosoma and Chamaeleonidae) may be the most committed to lingual prehension. 2. More scleroglossans. Although there is likely to be some interesting diversity left to discover among iguanians, the functional morphology of feeding in the diverse and specious scleroglossan clade remains mostly unexplored. Small, generalized taxa from diverse lineages are particularly in need of study—most families are virtually unsampled. Although additional studies of well-known taxa are unlikely to advance the front of knowledge very far, if undertaken they should at least consider diversity within the group. For example, small varanid species (e.g., V. gilleni and V. prasinus) or the frugivore, V. olivaceus, would be of much greater interest than another study of Varanus exanthematicus. 3. Scaling effects. Scaling effects of body size on feeding kinematics are virtually unstudied in lepidosaurs. Comparisons of timing variables among taxa have little meaning unless these effects are accounted for. 4. More cineradiography. More cineradiographic studies of diverse species are required to characterize adequately the relationship between tongue and hyobranchial movement in different feeding stages, different phases of the gape cycle, different parts of the tongue, and different lingual behaviors. The mechanisms of tongue protrusion and tongue movement, generally, are in need of elucidation. Is the foretonguehindtongue duality proposed here valid? 5. Effects of food type. The effects of food type on feeding performance and kinematics need to be explored. Carefully controlled experiments are required to dissect out the confounding effects of food size, mass, shape, texture, wetness, and activity level. In particular, investigators should pay more attention to natural diet and duplicate this as closely as possible in laboratory analyses. The "switch" between hyolingual and inertial transport should be investigated using naturalistic prey types of different sizes in different taxa. 6. Effects of experience. As a corollary to the effects of food type, the effect of previous experience with different food types should be considered in comparative.
278
Kurt Schwenk
quantitative studies of feeding. Individual familiarity with different foods, or lack thereof, might confound interpretation of interspecific difference in kinematic variables. 7. Studies of cranial kinesis in diverse taxa. Additional studies of cranial kinesis using diverse, small, generalized taxa (see No. 2) are needed. Condon's (1987) goniometric technique should be employed more widely and studies should include all feeding stages. The hypothesis proposed here that cranial (meso-)kinesis is restricted to scleroglossans using jaw prehension needs to be further tested. 8. Streptostyly and chewing in acrodonts. The Robinson (1976) hypothesis linking streptostyly to dental occlusion in acrodonts needs to be tested further. 9. Swallowing. Swallowing is now the most poorly understood feeding stage in lepidosaurs. Additional data are needed to clarify the distinction between pharyngeal packing and pharyngeal compression, and the relationship of these behaviors to bolus position in the pharynx and esophagus. The functional morphology of "labial licking" and its association with pharyngeal packing needs to be better characterized, as do the phylogenetic, morphological, and ecological correlates of pharyngeal compression. Acknowledgments I am very grateful to colleagues who offered critical comment on various parts of the manuscript: David Cundall, Leo Fleishman, Harry Greene, Nate Kley, Matthias Ott, and C. B. Wood. Nate Kley, Kris Lappin, and Howard Snell shared unpublished observations of feeding in amphisbaenians, Gambelia and Conolophus, respectively. Anthony Herrel, Ken Kardong, Lance McBrayer, and Tammy Smith provided copies of manuscripts submitted or in press. Ken Kardong and Tammy Smith allowed me to view their films of T. scincoides using lingual prehension and discussed their findings with me. The help, comments, and discussion of all these people are gratefully acknowledged. Finally, I thank my long-suffering collaborators on sundry lizard-feeding projects, who have been forced to adapt to my snail-like pace: Aaron Bauer, Douglas Bell, Tom Frazzetta, Farish Jenkins, Jr., Susan Rehorek, Judy Sheen, and Wade Sherbrooke. The University of Connecticut Research Foundation and the National Science Foundation (NSF IBN-9601173) supported empirical work and made preparation of the manuscript possible.
References Abel, E. (1952) Zur Biologie von Agama agilis Ol. und Agama ruderata Ol. Zool. Anz. 149:125-133. Abu-Ghalyun, Y. L., L. Greenwald, T. E. Hetherington, and A. S. Gaunt (1988) The physiological basis of slow locomotion in chameleons. J. Exp. Zool. 245:225-231. Alexander, R. M. (1983) Animal Mechanics. Blackwell Scientific, Oxford. Alexander, R. M. (1992) Exploring Biomechanics. Scientific American Library, Freeman, New York.
Altevogt, R. (1977) Chamaeleo jacksonii (Chamaeleonidae)—Beutefang. Publ. Inst. Wissenschaftlichen Film, Gottingen. Wissenschaftlichen Filmen Sek. Biol. Ser. 10, No. 49:3-12. Altevogt, R., and R. Altevogt (1954) Studien zur Kinematik der Chamaleonzunge. Zeit. vergl. Physiol. 36:66-77. Anderson, R. A., and W. H. Karasov (1981) Contrasts in energy intake and expenditure in sit-and-wait and widely foraging lizards. Oecologia 49:67-72. Andrews, C., and J. E. A. Bertram (1997) Mechanical work as a determinant of prey-handling behavior in the tokay gecko {Gekko gecko). Physiol. Zool. 70:193-201. Andrews, R. M., F. H. Rough, A. CoUazo, and A. de Queiroz (1987) The ecological cost of morphological specialization: feeding by a fossorial lizard. Oecologia 73:139-145. Armstrong, J. A., Gamble, H. J., and Goldby, F. (1953) Observations on the olfactory apparatus and the telencephalon of Anolis, a microsmatic lizard. J. Anat. 87:288-307. Arnold, E. N. (1998) Cranial kinesis in lizards: variations, uses, and origins. Pp. 323-357. In: Evolutionary Biology, Vol. 30. M. K. Hecht, R. J. Maclntyre, and M. T. Clegg (eds.). Plenum Press, New York. Arnold, S. J. (1981) Behavioral variation in natural populations. II. The inheritance of a feeding response in crosses between geographic races of the garter snake, Thamnophis elegans. Evolution 35:510-515. Arthur, W. (1997) The Origin of Animal Body Plans: A Study in Evolutionary Developmental Biology. Cambridge Univ. Press, Cambridge. Auffenberg, W. (1978) Social and feeding behavior in Varanus komodoensis. Pp. 301-331. In: Behavior and Neurology of Lizards: An Interdisciplinary Colloquium. N. Greenberg and P. D. MacLean (eds.). U. S. Dept. of Health, Education, and Welfare; National Institute of Mental Health, Rockville, MD. Auffenberg, W. (1980) The herpetofauna of Komodo, with notes on adjacent areas. Bull. Florida State Mus. Biol. Sci. 25:39-156. Auffenberg, W. (1981) The Behavioral Ecology of the Komodo Monitor. Univ. of Florida Press, Gainesville, FL. Auffenberg, W. (1988) Gray's Monitor Lizard. Univ. of Florida Press, Gainesville, FL. Avery, D. G., and W. W. Tanner (1964) The osteology and myology of the head and thorax regions of the obesus group of the genus Sauromalus Dumeril (Iguanidae). Brigham Young Univ. Sci. Bull. Biol. Ser. 5(3): 1-30. Avery, D. G., and W. W. Tanner (1971) Evolution of the iguanine lizards (Sauria, Iguanidae) as determined by osteological and myological characters. Brigham Young Univ. Sci. Bull. Biol. Ser. 12(3): 1-79. Bahl, K. N. (1937) Skull of Varanus monitor (Linn.). Indian Mus. Calcutta Rec. 39:133-174. Barnett, K. E., R. B. Cocroft, and L. J. Fleishman (1999) Possible communication by substrate vibration in a chameleon. Copiea 1999: 225-228. Barrows, S., and H. M. Smith (1947) The skeleton of the lizard Xenosaurus grandis (Gray). Univ. Kansas Sci Bull. 31:227-275. Bauer, A. M. (1990) Phylogenetic systematics and biogeography of the Carphodactylini (Reptilia: Gekkonidae). Bonner Zool. Monogr. No. 30:1-218. Bauer, A. M., and K. D. DeVaney (1987) Comparative aspects of diet and habitat in some New Caledonian lizards. Amph.-Rept. 8: 349-364. Bauer, A. M., and A. P. Russell (1990) Dentitional diversity in Rhacodactylus (Reptilia: Gekkonidae). Mem. Queensland Mus. 29:311321. Bauer, A. M., A. P. Russell, and B. D. Edgar (1989) Utilization of the termite Hodotermes mossambicus (Hagen) by gekkonid lizards near Keetmanshoop, South West Africa. S. Afr. J. Zool. 24:239-243.
8. F e e d i n g in L e p i d o s a u r s Bayer, R (1899) Bemerkungen zur Entwicklung der Eidechsenzunge. Morph.Jb. 27:712-716. Beddard, F. E. (1905) A contribution to the anatomy of the frilled lizard {Chlamydosaurus kingi) and some other Agamidae. Proc. ZooLSoc.Lond. 1:9-22. Beebe, W. (1944) Field notes on the lizards of Kartabo, British Guiana, and Caripito, Venezuela. 2. Iguanidae. Zoologica 24:195216. Beebe, W. (1945) Field notes on the lizards of Kartabo, British Guiana, and Caripito, Venezuela. 3. Teiidae, Amphisbaenidae and Scincidae. Zoologica 30:7-32. Bell, D. A. (1987) Identification of perikarya in the chameleon tongue. Pp. 63-66. In: Proceedings of the 4th Ordinary General Meeting of the Societas Europaea Herpetologica. J. J. van Gelder, H. Strijbosch, and P. J. M. Bergers (eds.). Nijmegen. Bell, D. A. (1989) Functional anatomy of the chameleon tongue. Zool. Jb.Anat. 119:313-336. Bell, D. A. (1990) Kinematics of prey capture in the chameleon. Zool. Jb. Physiol. 94:247-260. Bellairs, A. d'A. (1949) Observations on the snout of Varanus, and a comparison with that of other Hzards and snakes. J. Anat. 83: 116-147. Bellairs, A. d'A. (1950) Observations on the cranial anatomy of Anniella, and a comparison with that of other burrowing Hzards. Proc. Zool. Soc. Lond. 119:887-904. Bellairs, A. d'A. (1970) The Life of Reptiles, Vol. 1. Universe Books, New York. Bellairs, A. d'A. (1984) Closing address: with comments on the organ of Jacobson and the evolution of Squamata, and on the intermandibular connection in Squamata. Pp. 665-683. In: The Structure, Development and Evolution of Reptiles. M. W. J. Ferguson (ed.). Zoological Society of London and Academic Press, London (Symp. Zool. Soc. Lond. No. 52). Bellairs, A. d'A., and J. D. Boyd (1950) The lachrymal apparatus in lizards and snakes. II. The anterior part of the lachrymal duct and its relationship with the palate and with the nasal and vomeronasal organs. Proc. Zool. Soc. Lond. 120:269-310. Bellairs, A.d'A., and C. Gans (1982) A reinterpretation of the amphisbaenian orbitosphenoid. Nature 302:243-244. Bellairs, A. d'A., and A. M. Kamal (1981) The chondrocranium and development of the skull in recent reptiles. Pp. 1-263. In: Biology of the Reptilia, Vol. 11. C. Gans and T. S. Parsons (eds.). Academic Press, London. Bels, V L. (1990a) Quantitative analysis of prey-capture kinematics in Anolis equestris (Reptilia: Iguanidae). Can. J. Zool. 68:21922198. Bels, V. L. (1990b) The mechanism of dewlap extension in Anolis carolinensis (Reptilia: Iguanidae) with histological analysis of the hyoid apparatus. J. Morph. 206:225-244. Bels, V. L. (1992) Functional analysis of the ritualized behavioural motor pattern in lizards: evolution of behaviour and the concept of ritualization. Zool. Jb. Anat. 122:141-159. Bels, V. L., and I. Baltus (1987) First analysis of the feeding sequence of Chameleo dilepis. Proc. 4th Ord. Gen. Meet. Soc. Eur. Herp. 4: 67-70. Bels, V. L., and I. Baltus (1988) The influence of food items on the feeding cycle in Anolis equestris (Reptilia: Iguanidae). Copeia 1988:479-481. Bels, V L., and I. Baltus (1989) First analysis of feeding in Anolis lizards. Pp. 141-145. In: Trends in Vertebrate Morphology (Forschritte der Zoologie, Vol. 35). H. Splechtna and H. Hilgers (eds.). Gustav Fisher, Stuttgart. Bels, V. L., and V. Goosse (1989) A first report of relative movements within the hyoid apparatus during feeding in Anolis equestris (Reptilia: Iguanidae). Experientia 45:1088-1091.
279
Bels, V. L., and V. Goosse (1990) Comparative kinematic analysis of prey capture in Anolis carolinensis (Iguania) and Lacerta viridis (Scleroglossa). J. Exp. Zool. 255:120-124. Bels, V. L., V. Goosse, and K. V. Kardong (1993) Kinematic analysis of drinking by the lacertid lizard, Lacerta viridis (Squamata, Scleroglossa). J. Zool. Lond. 229:659-682. Bels, V. L., M. Chardon, and K. Kardong (1994) Biomechanics of the hyolingual system in Squamata. Pp. 197-240. In: Biomechanics of Feeding in Vertebrates. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Adv. Comp. Environ. Physiol., Vol. 18. Springer-Verlag, Berlin. Bels, V. L., J.-P Gasc, V Goosse, S. Renous, and R. Vernet (1995) Functional analysis of the throat display in the sand goanna Varanus griseus (Reptilia: Squamata: Varanidae). J. Zool. Lond. 235:95116. Bendz, H. (1840) Bidrag til den sammenlignende Anatomie af Nervus Glossopharyngeus, Vagus, Accessorius willisii og Hypoglossus hos Reptilierne. Vid. Sel. natnrvid. og mathem. Afh. 10:115152. Benes, E. S. (1969) Behavioral evidence for color discrimination by the whiptail lizard, Cnemidophorus tigris. Copeia 1969:707-722. Bjorndal, K. A. (1997) Fermentation in reptiles and amphibians. Pp. 199-230. In: Gastrointestinal Microbiology, Vol. 1. R. I. Mackie and B. A. White (eds.). Chapman and Hall, New York. Blanc, C.-P. (1965) fitude sur les Iguanidae de Madagascar. I. Le squelette de Chalarodon madagascariensis Peters, 1854. Mem. Mus. Nat. Hist. Nat. Ser. A Zool. 33:93-146. Bogert, C. M. (1964) Amphisbaenids are a taxonomic enigma. Nat. Hist. 73:17-24. Bogert, C. M., and R. Martin del Campo (1956) The Gila monster and its allies: the relationships, habits, and behavior of the lizards of the family Helodermatidae. Bull. Am. Mus. Nat. Hist. 109:1238. Boltt, R. E., and R. F Ewer (1964) The functional anatomy of the head of the puff adder, Bitis arietans (Merr.) J. Morph. 114:83-106. Borsuk-Bialynicka, M. (1985) Comments on the kinetics of the lizard skull. Pp. 135-137. In: Functional Morphology in Vertebrates (Fortschr. Zool. Vol. 30). H.-R. Duncker and G. Fleischer (eds.). Gustav Fischer, Stuttgart. Bourgat, R. M. (1970) Recherches ecologiques et biologiques sur le Chamaeleo pardalis Cuvier 1829 de File de la Reunion et de Madagascar. Bull. Soc. Zool. France 95:259-269. Bourgat, R. M. (1972) Nourriture du cameleon: regime alimentaire et capture des proies. Ceylon J. Sci. (Bio. Sci.) 10:1-5. Boyden, T. C. (1976) Butterfly palatability and mimicry: experiments with Ameiva lizards. Evolution 30:73-81. Bradley, O. S. (1903) The muscles of mastication and the movements of the skull in Lacertilia. Zool. Jahrb. Abth. Anat. Ont. 18:475486. Brain, C. K. (1959) Note on the breeding of the South African skink, Typhlosaurus lineatus Boulenger. Copeia 1959:70-71. Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. R Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Broadley, D. G., C. Gans, and J. Visser (1976) Studies on amphisbaenians (Amphisbaena, Reptilia) 6. The genera Monopeltis and Dalophia in southern Africa. Bull. Am. Mus. Nat. Hist. 157:313-485. Broom, R. (1903) On the development of the pterygo-quadrate arch in the Lacertilia. J. Anat. 37:107-111. Broom, R. (1925) On the origin of lizards. Proc. Zool. Soc. Lond. 1925 (1):1-16. Broom, R. (1935) On the structure of the temporal region in lizard skulls. Ann. Transvaal Mus. 18:13-22. Brown, R. P., and V. Perez-Mellado (1994) Ecological energetics and
280
Kurt Schwenk
food acquisition in dense Menorcan islet populations of the lizard Podarcis lilfordi Funct. Ecol. 8:427-434. Briicke, E. (1852) Ueber die Zunge der Chamaleonen. Sitzungsber. Kaiserlich. Akad. Wiss. Math.-Naturwis. CI. (Vienna) 8:65-70. BuUer, W. L. (1878) Further notes on the habits of the tuatara lizard. Trans. Proc. New Zeal. Inst. 1878:349-351. Bullock, D. J., H. M. Jury, and P. G. H. Evans (1993) Foraging ecology in the lizard Anolis oculatus (Iguanidae) from Dominica, West Indies. J. Zool. Lond. 230:19-30. Burghardt, G. M. (1964) Effects of prey size and movement on the feeding behavior of the lizards Anolis carolinensis and Eumecesfasciatus. Copeia 1964:576-578. Burghardt, G. M. (1970) Chemical perception in reptiles. Pp. 241308. In: Communication by Chemical Signals. J. W. Johnston, Jr., D. G. Moulton, and A. Turk (eds.). Appleton-Century-Crofts, New York. Burghardt, G. M. (1980) Behavioral and stimulus correlates of vomeronasal functioning in reptiles: feeding, grouping, sex, and tongue use. Pp. 275-301. In: Chemical Signals: Vertebrates and Aquatic Invertebrates. D. Miiller-Schwarze and R. M. Silverstein (eds.). Plenum Press, New York. Burrage, B. R. (1973) Comparative ecology and behaviour of Chamaeleo pumilus pumilus (Gmelin) and C. namaquensis A. Smith (Sauria: Chamaeleonidae). Ann. S. Afr. Mus. 61:1-158. Bustard, H. R. (1958) Use of horns by Chamaeleo jacksoni. Br. J. Herp. 2:105-107. Bustard, H. R. (1965) Observations on the life history and behavior of Chamaeleo hohnelii (Steindachner). Copeia 1965:401-410. Cabrera, M. R., and H. O. Merlini (1990) The diet oi Amphisbaena darwinii heterozonata in Argentina. Herp. Rev. 21:53-54. Caldwell, M. W. (1999) Squamate phylogeny and the relationships of snakes and mosasauroids. Zool. J. Linn. Soc. 125:115-147. Caldwell, M. W., and M. S. Y Lee (1997) A snake with legs from the marine Cretaceous of the Middle East. Nature 386:705-709. Camp, C. L. (1923) Classification of the lizards. Bull. Am. Mus. Nat. Hist. 48:289-481. Capel-Williams, G., and D. Pratten (1978) The diet of adult and juvenile Agama bibroni (Reptilia: Lacertae) and a study of the jaw mechanism in the two age groups. J. Zool. Lond. 185:309-318. Carpenter, C. C. (1966) The marine iguana of the Galapagos Islands, its behavior and ecology. Proc. CaHf. Acad. Sci. 34:329-376. Carpenter, C. C. (1969) Behavioral and ecological notes on the Galapagos land iguanas. Herpetologica 25:155-164. Carpenter, C. C , and G. W. Ferguson (1977) Variation and evolution of stereotyped behavior in reptiles. Pp. 335-554. In: Biology of the Reptilia, Vol. 7. A. C. Gans and D. W. Tinkle (eds.). Academic Press, New York. Carroll, R. L. (1988a) Late Paleozoic and Early Mesozoic lepidosauromorphs and thier relation to lizard ancestry. Pp. 99-118. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Carroll, R. L. (1988b) Vertebrate Paleontology and Evolution. Freeman, New York. Chiel, H. J., R Crago, J. M. Mansour, and K. Hathi (1992) Biomechanics of a muscular hydrostat: a model of lapping by a reptilian tongue. Biol. Cybem. 67:403-415. Colli, G. R., and D. S. Zamboni (1999) Ecology of the worm-lizard Amphisbaena alba in the Cerrado of central Brazil. Copeia 1999: 733-742. Conant, R., and J. T. Collins (1998) Reptiles and Amphibians: Eastern Central North America (Peterson Field Guide). 3rd ed., expanded. Houghton Mifflin, Boston. Condon, K. (1987) A kinematic analysis of mesokinesis in the Nile monitor (Varanus niloticus). Exp. Biol. 47:73-87. Condon, K. (1998) Morphological basis for cranial kinesis in the Nile monitor, Varanus niloticus. Am. Zool. 38(5): 201 A.
Cooper, J. S. (1966) Tooth replacement in the slow worm {Anguisfragilis). J. Zool. Lond. 150:235-248. Cooper, J. S., and D. R G. Poole (1973) The dentition and dental tissues of the agamid lizard, Uromastyx. J. Zool. Lond. 169:85-100. Cooper, J. S., D. R G. Poole, and R. Lawson (1970) The dentition of agamid lizards with special reference to tooth replacement. J. Zool. Lond. 162:85-98. Cooper, W. E., Jr. (1981) Visual guidance of predatory attack by a scincid lizard, Eumeces laticeps. Anim. Behav. 29:1127-1136. Cooper, W. E., Jr. (1994) Prey chemical discrimination, foraging mode, and phylogeny. Pp. 95-116. In: Lizard Ecology: Historical and Experimental Perspectives. L. J. Vitt and E. R. Pianka (eds.). Princeton Univ. Press, Princeton, NJ. Cooper, W. E., Jr. (1995a) Foraging mode, prey chemical discrimination, and phylogeny in lizards. Anim. Behav. 50:973-985. Cooper, W. E., Jr. (1995b) Prey chemical discrimination and foraging mode in gekkonoid lizards. Herp. Monogr. 9:120-129. Cooper, W. E. Jr., and N. Greenberg (1992) Reptilian coloration and behavior. Pp. 298-422. In: Biology of the Reptilia, Vol. 18. C. Gans and D. Crews (eds.). Univ. of Chicago Press, Chicago. Corning, H. K. (1895) Ueber die Entwickelung der Zungenmusculatur bei Reptilien. Verhandl. Anat. Gesell. 9:165-175. Cowles, R. B., and R. L. Phelan (1958) Olfaction in rattlesnakes. Copeia 1958:77-83. Criley, B. B. (1968) The cranial osteology of gerrhonotiform lizards. Am. Midi. Nat. 80:199-219. Crompton, A. W. (1989) The evolution of mammalian mastication. Pp. 23-40. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, Chichester. Crook, J. M., and T. S. Parsons (1980) Visceral anatomy of the Amphisbaenia. J. Morph. 163:99-133. Cruz Neto, A. P., and A. S. Abe (1993) Diet composition of two syntopic species of Neotropical amphisbaenians, Cercolophia roberti and Amphisbaena mertensii. J. Herp. 27:239-240. Curio, E., and H. Mobius (1978) Versuche zum Nachweis eines Riechvermogens von Anolis I. lineatopus (Rept., Iguanidae). Zeit. Tierpsychol. 47:281-292. Cusumano, M., and R. Powell (1991) A note on the diet of Amphisbaena gonavensis in the Dominican Republic. Amph.-Rept. 12: 350-352. Cuvier, G. (1805) Legons d'Anatomic Comparee, Vol. 3. Baudouin, Imprimeur de ITnstitut, Paris (1969 facsimile reprint by Culture et Civilisation, Brussels, Belgium). Dalrymple, G. H. (1979) On the jaw mechanism of the snail-crushing lizards, Dracaena Daudin 1802 (Reptilia, Lacertilia, Teiidae). J. Herp. 13:303-311. Daugherty, C. H., A. Cree, J. M. Hay, and M. B. Thompson (1990) Neglected taxonomy and continuing extinctions of tuatara {Sphenodon). Nature 347:177-179. Davis, D. D. (1961) Origin of mammalian feeding mechanisms. Am. Zool. 1:229-234. de Beer, G. R. (1937) The Development of the Vertebrate Skull. Oxford Univ. Press, Oxford, de la Cema de Esteban, C. (1959) Anatomia microscopica de la lengua de Amphisbaena vermicularis darwin [sic] (D. y B.). Actas y Trabajos del Primer Congreso Sudamericano de Zoologia 5:143-150. de la Cerna de Esteban, C. (1965) Anatomia microscopica comparada de la lengua de algunos saurios Argentines. Anais do 2nd Congresso Latinoamericano de Zoologia 2:235-245 Delheusy, V, and V. L. Bels (1992) Kinematics of feeding behaviour in Oplurus cuvieri (Reptilia: Iguanidae). J. Exp. Biol. 170:155-186. Delheusy, V, C. Brillet, and V. L. Bels (1995) Etude cinematique de la prise de nourriture chez Eublepharis macularius (Reptilia, Gekkonidae) et comparaison au sein des geckos. Amph.-Rept. 15: 185-201.
8. Feeding in Lepidosaurs Delheusy, V., G. Toubeau, and V. Bels (1994) Tongue structure and function in Opiums cuvieri (Reptlia: Iguanidae). Anat. Rec. 238: 263-276. Denny, M. W. (1984) Mechanical properties of pedal mucus and their consequences for gastropod structure and performance. Am. Zool. 24:23-36. de Queiroz, K. (1987) Phylogenetic systematics of iguanine lizards: a comparative osteological study. Univ. Calif. Publ. Zool. 118:1203. de Queiroz, K. (1996) A phylogenetic approach to biological nomenclature as an alternative to the Linnaean systems in current use. In: Proceedings of a Mini-Symposium on Biological Nomenclature in the 21st Century. J. L. Reveal (ed.). Univ. of Maryland, College Park, (www.life.umd.edubees96sym.html). de Queiroz, K. (1997) The Linnaean hierarchy and the evolutionization of taxonomy, with emphasis on the problem of nomenclature. Aliso 15:125-144. de Queiroz, K., and J. Gauthier (1992) Phylogenetic taxonomy. Arm. Rev. Ecol. Syst. 23:449-480. de Queiroz, K., and J. Gauthier (1994) Toward a phylogenetic system of biological nomenclature. Trends Ecol. Evol. 9:27-31. Dessem, D. (1985) Ontogenetic changes in the dentition and diet of Tupinambis (Lacertilia: Teiidae). Copeia 1985:245-247. De Vree, R, and C. Gans (1987) Kinetic movements in the skull of adult Trachydosaurus rugosus. Anat. Histol. Embryol. 16:206-209. De Vree, R, and C. Gans (1989) Functional morphology of the feeding mechanisms in lower tetrapods. Pp. 115-127. In: Trends in Vertebrate Morphology. H. Splechtna and H. Hilgers (eds.). Gustav Fischer, Stuttgart (Fortschritte der Zoologie, Vol. 35). De Vree, R, and C. Gans (1994) Reeding in tetrapods. Pp. 93-118. In: Biomechanics of Feeding in Vertebrates. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Adv. Comp. Environ. Physiol., Vol. 18. Springer-Verlag, Berlin. Dewevre, Dr. (1895) Le mecanisme de la projection de la lengue chez le cameleon. J. Anat. Physiol. 31:343-360. de Witte, G.-R (1965) Les cameleons de TAfrique centrale (Republique Democratique du Congo, Republique du Rwanda et Royaume du Burundi). Arm. Mus. Roy. Afr. Cent. (Belgique) Ser. 8, Sci. Zool. No. 142:1-215. Dial, B. E. (1978) Aspects of the behavioral ecology of two Chihuahuan Desert geckos (Reptilia, Lacertilia, Gekkonidae). 12:209216. Dial, B. E., and K. Schwenk. (1996) Olfaction and predator detection in Coleonyx brevis (Squamata: Eublepharidae) with comments on the functional significance of buccal pulsing in geckos. J. Exp. Zool. 276:415-424. Diamond, J. M. (1987) Did Komodo dragons evolve to eat pygmy elephants? Nature 326:832. Dischner, H. (1958) Zur Wirkungsweise der Zunge beim Chamaleon. NaturVolk 88:320-324. Dodd, C. K., Jr. (1981) Infrared reflectance in chameleons (Chamaeleonidae) from Kenya. Biotropica 13:161-164. Dornesco, G. T., and D. Andrei (1966) Les glandes buccales (salivaires) des Sauriens. Anat. Anz. 118:7-26. Drummond, H. M. (1979) Stimulus control of amphibious predation in the northern water snake {Nerodia s. sipedon). Zeit. Tierpsychol. 50:18-44. Druzinsky, R. E., and W. S. Greaves (1979) A model to explain the posterior limit of the bite point in reptiles. J. Morph. 160:165-168. Duffield, G. A., and C. M. Bull (1998) Seasonal and ontogenetic chagnes in the diet of the Australian skink Egernia stokesii. Herpetologica 54:414-419. Duges, A. (1827) Recherches anatomiques et physiologiques sur la Deglutition dans les Reptiles. Ann. Sci. Natur. 12:337-395. Dullemeijer, P. (1980) Functional morphology and evolutionary biology. Acta Biotheoretica 29:151-250.
281
Dumeril, C. (1836) Sur les mouvements de la langue chez les Cameleons. C. R. Acad. Sci. Paris 2:228-230. Duvernoy, G. L. (1836a) Sur les mouvements de la langue du Cameleon. C. R. Acad. Sci. Paris 2:349-351. Duvernoy, G. L. (1836b) Memoire sur quelques particularites des organes de la deglutition de la classe des Oiseaux et des Reptiles, pour servir de suite a un premier memoire sur la langue. C. R. Acad. Sci. Paris 2 (Suppl.):l-24. Echelle, A. A., and A. R Echelle (1972) Observations of fish-eating behavior and maintenance behavior in two species of Basiliscus. Copeia 1972:387-389. Edgeworth, R H. (1931) On the development of the external ocular, masticatory and hyoid muscles of Sphenodon punctatus. Proc. Zool. Soc. Lond. 1931:803-808,817-818. Edgeworth, R H. (1935) The Cranial Muscles of Vertebrates. Cambridge Univ. Press, Cambridge. Edmund, A. G. (1960) Tooth replacement phenomena in the lower vertebrates. Roy. Ont. Mus. Life Sci. Div. Contr. No. 52:1-190. Edmund, A. G. (1969) Dentition. Pp. 117-200. In: Biology of the Reptilia, Vol. 1. C. Gans, A. d'A. Bellairs, and T. S. Parsons (eds.). Academic Press, New York. Eifler, D. A., and M. A. Eifler (1998) Roraging behavior and spacing patterns of the lizard Cnemidophorus uniparens. J. Herp. 32: 24-33. Elias, J. A., L. D. McBrayer, and S. M. Reilly (2000) Prey transport kinematics in Tupinambis teguixin and Varanus exanthematicus: Conservation of feeding behavior in 'chemosensory-tongued' lizards. J. Exp. Biol. 203:791-801. Ellingson, J. M., L. J. Fleishman, and E. R. Loew (1995) Visual pigments and spectral sensitivity of the diurnal gecko Gonatodes albigularis. J. Comp. Physiol. 177:559-567. El Toubi, M. R., and A. M. Kamal (1959a) The development of the skull of Chalcides ocellatus. I. The development of the chondrocranium. J. Morph. 104:269-306. El Toubi, M. R., and A. M. Kamal (1959b) The development of the skull of Chalcides ocellatus. II. The fully formed chondrocranim and the osteocranium of a late embryo. J. Morph. 105:55-104. Emerson, S. B., and D. Diehl (1980) Toe pad morphology and mechanisms of sticking in frogs. Biol. J. Linn. Soc. 13:199-216. Engelbrecht, D. van Z. (1951) Contributions to the cranial morphology of the chamaeleon Microsaura pumila Daudin. Ann. Univ. Stellenbosch 27A:1-31. Estes, R. (1983) The fossil record and early distribution of lizards. Pp. 365-398. In: Advances in Herpetology and Evolutionary Biology: Essays in Honor of Ernest E. Williams. A. G. J. Rhodin and K. Miyata (eds.). Museum of Comparative Zoology, Cambridge, MA. Estes, R., and E. E. Williams (1984) Ontogenetic variation in the molariform teeth of lizards. J. Vert. Paleon.4:96-107. Estes, R., K. de Queiroz, and J. Gauthier (1988) Phylogenetic relationships within Squamata. Pp. 119-281. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Etheridge and de Queiroz (1988) A phylogeny of Iguanidae. Pp. 283367. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Farlow, J. O. (1976) Observations on a captive tuatara {Sphenodon punctatum). J. Herp. 9:353-355. Ferdinand, L. Prinz von Bayern (1884) Anatomic der Zunge. Fine vergleichend-anatomische Studie. Literarisch-Artistische Anstalt (Theodor Riedel), Munich. Filoramo, N., and K. Schwenk (1998) Morphological evidence for variation in the mechanism of chemical delivery to the VNO in squamate reptiles. Am. Zool. 38:107A. Fineman, G. (1940) Zur entwicklungsgeschichte des kopfskelets bei Chamaeleon bitaeniatus ellioti. Gegenbaurs Jb. 85:91-114. Fineman, G. (1943) Die Entwicklung des Zungenbeines bei Chamae-
282
Kurt Schwenk
leon bitaeniatus Ellioti. Kungl. Fysiografisk Sallskapets I Lund Forhandlingar (Proc. Roy. Physiogr. Soc. Lund) 13:78-90. Fisher, D. L., and W. W. Tanner (1968) Osteological and mylogical [sic] comparisons of the head and thorax regions of Cnemidophorus tigris septentrionalis Burger and Ameiva undulata parva Barbour and Noble (Family Teiidae). Brigham Young Univ. Sci. Bull. Biol. Ser. 11(1):1-41. Fitch, H. S. (1935) Natural history of the alligator lizards. Trans. Acad. Sci. St. Louis 29:1-38. Fitch, H. S. (1954) Life history and ecology of the five-lined skink Eumeces fasciatus. Univ. Kansas Publ. Mus. Nat. Hist. 8:1-156. Fitch, H. S. (1958) Natural history of the six-lined racerunner (Cnemidophorus sexlineatus). Univ. Kansas Publ. Mus. Nat. Hist. 11: 11-62. Fite, K. v., and B. C. Lister (1981) Bifoveal vision in Anolis lizards. Brain Behav. Evol. 19:144-154. Flanders, M. (1985) Visually guided head movement in the African chameleon. Vision Res. 25:935-942. Fleishman, L. J. (1986) Motion detection in the presence and absence of background motion in an Anolis lizard. J. Comp. Physiol. A 159:711-720. Fleishman, L. J. (1988) Sensory and environmental influences on display form in Anolis auratus, a grass anole from Panama. Behav. Ecol. Sociobiol. 22:309-316. Fleishman, L. J. (1992) The influence of the sensory system and the environment on motion patterns in the visual displays of anoline lizards and other vertebrates. Am. Nat. (Suppl.) 139:S36-S61. Fleishman, L. J., E. R. Loew, and M. Leal (1993) Ultraviolet vision in lizards. Nature 365:397. Fleishman, L. J., W. J. McClintock, R. B. D'Eath, D. H. Brainard, and J. A. Endler (1998) Colour perception and the use of video playback experiments in animal behaviour. Anim. Behav. 56:10351040. Font, E. (1996) Los sentidos quimicos de los reptiles. Un enfoques etologico. Pp. 197-259 (+ refs.). In: Etologia, Psicologia Comparada y Comportamiento Animal. F Colmenares (ed.). Editorial Sintesis, Madrid. Font, E., and L. C. Rome (1990) Functional morphology of dewlap extension in the lizard Anolis equestris (Iguanidae). J. Morph. 206: 245-258. Forey, P. L., C. J. Humphries, I. L. Kitching, R. W. Scofland, D. J. Siebert, and D. M. Williams (1992) Cladistics: A Practical Course in Systematics. Clarendon Press, Oxford. Forstner, M. R. J., S. K. Davis, and E. Arevalo (1995) Support for the hypothesis of anguimorph ancestry for the suborder Serpentes from phylogenetic analysis of mitochondrial DNA sequences. Mol. Phylog. Evol. 4:93-102. Frazzetta, T. H. (1962) A functional consideration of cranial kinesis in lizards. J. Morph. I l l :287-320. Frazzetta, T. H. (1983) Adaptation and function of cranial kinesis in reptiles: a time-motion analysis of feeding in alligator lizards. Pp. 222-244. In: Advances in Herpetology and Evolutionary Biology: Essays in Honor of Ernest E. Williams. A. G. J. Rhodin and K. Miyata (eds.). Museum of Comparative Zoology, Cambridge, MA. Frazzetta, T. H. (1986) The origin of amphikinesis in lizards. A problem in functional morphology and the evolution of adaptive systems. Pp. 419-461. In: Evolutionary Biology, Vol. 20. M. K. Hecht, B. Wallace, and G. T. Prance (eds.). Plenum, New York. Frost, D. R., and R. Etheridge (1989) A phylogenetic analysis and taxonomy of iguianian lizards (Reptilia: Squamata). Univ. Kansas Mus. Nat. Hist. Misc. Publ. No. 81:1-65. Fiirbringer, M. (1922) Das Zungenbein der Wirbeltiere insbesondere der Reptilien und Vogel. Abh. der Heidelberger Akad., math.n a t u r w . K l . i l : 1-164. Gabe, M., and H. Saint Girons (1964) Contribution a E'Histologie de
Sphenodon punctatus Gray. Editions du Centre National de la Recherche Scientifique, Paris. Gabe, M., and H. Saint Girons (1969) Donnees histologiques sur les glandes salivaires des lepidosauriens. Mem. Mus. Natn. Hist. Nat. Paris A58:1-112. Gabe, M., and H. Saint Girons (1976) Contribution a la morphologie con\paree des fosses nasales et de leurs annexes chez les lepidosoriens. Mem. Mus. Natn. Hist. Nat. Paris A98:1-87. Gadsden, H., and L. E. Palacios-Orona (1997) Seasonal dietary patterns of the Mexican fringe-toed lizard {Uma paraphygas). J. Herp. 31:1-9. Gandolfi, H. (1908) Die Zunge der Agamidae und Iguanidae. Zool. Anz. 32:569-580. Gans, C. (1960) Studies on amphisbaenids (Amphisbaenia, Reptiliia). I. A taxonomic revision of the Trogonophinae, and a functional interpretation of the amphisbaenid adaptive pattern. Bull. Am. Mus. Nat. Hist. 119:133-204. Gans, C. (1966) Studies on amphisbaenids (Amphisbaenia, Reptilia). 3. The small species from southern South America commonly identified as Amphisbaena darwini. Bull. Am. Mus. Nat. Hist. 134: 185-260. Gans, C. (1969a) Amphisbaenians: reptiles specialized for a burrowing existence. Endeavour 28:146-151. Gans, C. (1969b) Comments on inertial feeding. Copeia 1969:855857. Gans, C. (1969c) Functional components versus mechanical units in descriptive morphology. J. Morph. 128:365-368. Gans, C. (1974) Biomechanics: An Approach to Vertebrate Biology. Lippincott, Philadelphia. Gans, C. (1975) Tetrapod limblessness: evolution and functional corollaries. Am. Zool. 15:455-467. Gans, C. (1978) The characteristics and affinities of the Amphisbaenia. Trans. Zool. Soc. Lond. 34:347-416. Gans, C , and F. De Vree (1986) Shingle-back lizards crush snail shells using temporal summation (tetanus) to increase the force of the adductor muscles. Experientia 42:387-389. Gans, C , F. De Vree, and D. Carrier (1985) Usage pattern of the complex masticatory muscles in the shingleback lizard, Trachydosaurus rugosus: a model for muscle placement. Am. J. Anat. 173: 219-240. Gasc, J.-P (1968) Contribution a Lost^ologie et a Is myologie de Dibamus novaeguineae Gray (Sauria, Reptilia). Discussion systematique. Ann. Sci. Nat. Zool. 10:127-150. Gauthier et al. (1988) A phylogenetic analysis of Lepidosauromorpha. Pp. 99-118. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Gehlbach, F R., J. R Watkins, II, and J. C. Kroll (1971) Pheromone trail-following studies of typhlopid, leptotyphlopid, and colubrid snakes. Behaviour 40:282-294. George, J. C. (1948) The muscular system of Uromastix hardwickii (Gray). J. Univ. Bombay 17:1-23. George, J. C. (1954) On the cranial osteology of Uromastyx hardwickii (Gray). Anim. Morph. Physiol. 1:23-29. Gil, M. J., F. Guerrero, and V. Perez-Mellado (1993) Observations on morphometries and ecology in Blanus cinereus (Reptilia: Amphisbaenia). J. Herp. 27:205-209. Gillingham, J. C , C. Carmichael, and T. Miller (1995) Social behavior of the tuatara, Sphenodon punctatus. Herp. Monogr. 9:5-16. Gingerich, P. D. (1971) Functional significance of mandibular translation in vertebrate jaw mechanics. Postilla 152:1-10. Gleeson, T. T. (1983) A histochemical and enzymatic study of the muscle fiber types in the water monitor, Varanus salvator. J. Exp. Zool. 227:191-207. Gleeson, T. T. (1985) Glycogen synthesis from lactate in skeletal
8. Feeding in Lepidosaurs muscle of the lizard Dipsosaurus dorsalis. J. Comp. Physiol. B 156: 277-283. Gleeson, T. T., and J. M. Harrison (1986) Reptilian skeletal muscle: fiber-type composition and enzymatic profile in the lizard. Iguana iguana. Copeia 1986:324-332. Gleeson, T. T., and I. A. Johnston (1987) Reptilian skeletal muscle: contractile properties of identified, single fast-twitch and slow fibers from the lizard Dipsosaurus dorsalis. J. Exp. Zool. 242: 283-290. Gleeson, T. T., R. W. Putnam, and A. R Bennett (1980) Histochemical, enzymatic, and contractile properties of skeletal muscle fibers in the lizard Dipsosaurus dorsalis. J. Exp. Zool. 214:293-302. Gnanamuthu, C. P. (1930) The anatomy and mechanism of the tongue of Chamaeleon carcaratus (Merrem). Proc. Zool. Soc. Lond. No. 31:467-486. Gnanamuthu, C. P. (1937) Comparative study of the hyoid and tongue of some typical genera of reptiles. Proc. Zool. Soc. Lond. Ser.B 1937:1-63 Goldsmith, T. H. (1990) Optimization, constraint and history in the evolution of eyes. Quart. Rev. Biol. 65:281-322. Gomes, N. M. B. (1974) Anatomie comparee de la musculature trig e m i n a l des lacertiliens. Mem. Mus. Natn. Hist. Nat. Paris A90: 1-107. Goosse, v., and V. Bels (1992) Kinematic and functional analysis of feeding behaviour in Lacerta viridis (Reptilia: Lacertidae). Zool. Jb. Anat. 122:187-202. Goppert, E. (1903) Ueber die Bedeutung der Zunge fiir die Entstehung des sekundaren Gaumens. Verhandl. Anat. Gessellsch. 17: 75-83. Gorniak, G. C., H. I. Rosenberg, and C. Gans (1982) Mastication in the tuatara, Sphenodon punctatus (Reptilia: Rhynchocephalia): structure and activity of the motor system. J. Morph. 171:321-353. Gove, D. (1979) A comparative study of snake and lizard tongue flicking with an evolutionary hypothesis. Z. Tierpsychol. 51:58-76. Greaves, W. S. (1995) Functional predictions from theoretical models of the skull and jaws in reptiles and mammals. Pp. 99-115. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Green, D. M. (1981) Adhesion and the toe-pads of treefrogs. Copeia 1981:790-796. Greene, H. W. (1982) Dietary and phenotypic diversity in lizards: why are some organisms specialized? Pp. 107-128. In: Environmental Adaptation and Evolution. D. Mossakowski and G. Roth (eds.). Gustav Fischer, Stuttgart. Greene, H. W. (1983) Dietary correlates of the origin and radiation of snakes. Am. Zool. 23:431-441. Greene, H. W. (1997) Snakes: The Evolution of Mystery in Nature. Univ. of California Press, Berkeley, CA. Greer, A. E. (1979) A phylogenetic subdivision of Australian skinks. Rec. Aust. Mus. 32:339-371. Greer, A. E. (1985) The relationships of the lizard genera Anelytropsis and Dibamus. J. Herp. 19:116-156. Greer, A. E. (1989) The Biology and Evolution of Australian Lizards. Surrey Beatty and Sons, Chipping Norton, Australia. Giinther, A. (1867) Contribution to the anatomy of Hatteria {Rhychocephalus, Owen). Phil. Trans. Roy. Soc. Lond. B 157:595-629. Haas, G. (1937) The structure of the nasal cavity in Chamaeleo chameleon (Linnaeus). J. Morph. 61:433-451. Haas, G. (1947) Jacobson's organ in the chameleon. J. Morph. 81:195207. Haas, G. (1960) On the trigeminus muscles of the lizards Xenosaurus grandis and Shinisaurus crocodilurus. Am. Mus. Nov. No. 2017: 1-54. Haas, G. (1973) Muscles of the jaws and associated structures in the Rhynchocephalia and Squamata. Pp. 285-490. In: Biology of the
283
Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Hall, B. K. (1996) Bauplane, phylotypic stages, and constraint. Why there are so few types of animals. Pp. 215-261. In: Evolutionary Biology, Vol. 29. M. K. Hecht, R. J. Macintyre, and M. T. Clegg (eds.). Plenum Press, New York. Hallermann, J. (1992) Morphological significance of the orbitotemporal region in amphikinetic skulls of juvenile iguanians (Squamata). Zool. Jb. Anat. 122:203-206. Hallermann, J. (1994) Zur Morphologie der Ethmoidalregion der Iguania (Squamata): eine vergleichend-anatomischeUntersuchung. Bonner Zool. Monogr. Nr. 35:1-135. Hallermann, J. (1998) The ethmoidal region of Dibamus taylori (Squamata: Dibamidae), with a phylogenetic hypothesis on dibamid relationships within Squamata. Zool. J. Linn. Soc. 122:385-426. Halpern, M. (1992) Nasal chemical senses in reptiles: structure and function. Pp. 423-523. In: Biology of the Reptilia, Vol. 18. C. Gans and D. Crews (eds.). Univ. of Chicago Press, Chicago. Halpern, M., J. Halpern, E. Erichsen, and S. Borghjid (1997) The role of nasal chemical senses in garter snake response to airborne odor cues from prey. J. Comp. Psych. I l l : 251-260. Harris, D. M. (1985) Infralingual plicae: support for Boulenger's Teiidae (Sauria). Copeia 1985:560-565. Harris, V. A. (1963) The Anatomy of the Rainbow Lizard. Hutchinson Tropical Monographs, Hutchinson and Co., London. Harrison, H. S. (1901a) Hatteria punctata, its dentition and its incubation period. Anat. Anz. 20:145-158. Harrison, H. S. (1901b) The development and succession of teeth in Hatteria punctata. Quart. J. Microsc. Sci. 44:161-213. Hartline, P. H. (1971) Physiological basis for detection of sound and vibration in snakes. J. Exp. Biol. 54:349-371. Hedges, S. B., and L. L. Poling (1999) A molecular phytogeny of reptiles. Science 283:998-1001. Herrel, A. J., and F. De Vree (1999a) Kinematics of intraoral transport and swallowing in the herbivorous lizard Uromastix acanthinurus. J. Exp. Biol. 202:1127-1137. Herrel, A. J., and F. De Vree (1999b) The cervical musculature in helodermatid lizards. Belg. J. Zool. 129:175-186. Herrel, A., J. Cleuren, and F. De Vree (1995) Prey capture in the lizard Agama stellio. J. Morph. 224:313-329. Herrel, A., J. Cleuren, and F. De Vree (1996a) Kinematics of feeding in the lizard Agama stellio. J. Exp. Biol. 199:1727-1742. Herrel, A., R. Van Damme, and F. De Vree (1996b) Sexual dimorphism of head size in Podarcis hispanica atrata: testing the dietary divergence hypothesis by bite force analysis. Neth. J. Zool. 46: 253-262. Herrel, A., J. Cleuren, and F. De Vree (1997a) Quantitative analysis of jaw and hyolingual muscle activity during feeding in the lizard Agama stellio. J. Exp. Biol. 200:101-115. Herrel, A., I. Wauters, P. Aerts, and F. De Vree (1997b) The mechanics of ovophagy in the beaded lizard {Heloderma horridum). J. Herp. 31:383-393. Herrel, A., P. Aerts, and F. De Vree (1998a) Ecomorphology of the lizard feeding apparatus: a modelling approach. Neth. J. Zool. 48:1-25. Herrel, A., P. Aerts, and F. De Vree (1998b) Static biting in lizards: functional morphology of the temporal ligaments. J. Zool. Lond. 244:135-143. Herrel, A., J.-P. Timmermans, and F. De Vree (1998c) Tongue flicking in agamid lizards: morphology, kinematics, and muscle activity patterns. Anat. Rec. 252:102-116. Herrel, A., P. Aerts, J. Fret, and F. De Vree (1999a) Morphology of the feeding system in agamid lizards: ecological correlates. Anat. Rec. 254:496-507. Herrel, P., M. Vestappen, and F. De Vree (1999b) Modulatory com-
284
Kurt Schwenk
plexity of the feeding repertoire in scincid lizards. J. Comp. Physiol. A 184:501-518. Hetherington, T. E. (1989) Use of vibratory cues for detection of insect prey by the sandswiming lizard Scincus scincus. Anim. Behav. 37:290-297. Hiiemae, K. M., A. J. Thexton, and A. W. Crompton (1978) Intraoral food transport: the fundamental mechanism of feeding. Pp. 181208. In: Muscle Adaptation in the Craniofacial Region. Monograph No. 8, Craniofacial Growth Series. D. Carlson and J. MacNamara (eds.). Univ. of Michigan, Ann Arbor, MI. Hoffstetter, R. (1968) Review of "A contribution to the classification of snakes," by Garth Underwood. Copeia 1968:201-213. Hoppe, G. (1934) Das Geruchsorgan von Hatteria punctata. Zeit. Anat. 102:434-461. Hotton, N. (1955) A survey of adaptive relationships of dentition to diet in the North American Iguanidae. Am. Midi. Nat. 53:88-114. Houston, J. (1828) On the structure and mechanism of the tongue of the chameleon. Trans. Roy. Irish Acad. 15:177-201. Howes, G. B., and H. H. Swinnerton (1901) On the development of the skeleton of the tuatara, Sphenodon punctatus; with remarks on the egg, on the hatching and on the hatched young. Trans. Zool. Soc.Lond. 16:1-86. Huey, R. B., and E. R. Pianka (1981) Ecological consequences of foraging mode. Ecology 62:991-999. Hurle, J. M., V. Garcia-Martinez, Y. Ganan, V. Climent, and M. Blasco (1987) Morphogenesis of the prehensile autopodium in the common chameleon {Chamaeleo chamaeleo). J. Morph. 194:187-194. Inger, R. F. (1983) Morphological and ecological variation in the flying lizards (genus Draco). Fieldiana Zool. N. S., No. 18:1-35. lordansky, N. N. (1966) Cranial kinesis in lizards: contribution to the problem of the adaptive significance of skull kinesis. Zoologicheskii Zhurnal 45:1398-1410 (in Russion; 1968 English translation by the Smithsonian Herpetological Information Services). lordansky, N. N. (1970) Structure and biomechanical analysis of functions of the jaw muscles in the lizards. Anat. Anz. 127:383-413. lordansky, N. N. (1973) Morphological and functional patterns of the jaw apparatus and cranial kinetism in chameleons. Zoologicheskii Zhurnal 52:891-901 (in Russian with English summary and many figures). lordansky, N. N. (1990a) Evolution of cranial kinesis in lower tetrapods. Neth. J. Zool. 40:32-54. lordansky, N. N. (1990b) Functional Anatomy and Evolution of the Jaw Apparatus of the Amphibians and Reptiles. Mockba ''Hayka" [in Russian, with many illustrations]. lordansky, N. N. (1996) The temporal ligaments and their bearing on cranial kinesis in lizards. J. Zool. Lond. 239:167-175. Irish, R J. (1983) The dynamics of seed-eating in the characiform fish Colossoma. Am. Zool. 23:1027. Iverson, J. B. (1982) Adaptations to herbivory in iguanine lizards. Pp. 60-76. In: Iguanas of the World: Their Behavior, Ecology, and Conservation. G. M. Burghardt and A. S. Rand (eds.). Noyes, Park Ridge, NJ. Iwasaki, S. (1990) Fine structure of the dorsal lingual epithelium of the lizard, Gekko japonicus (Lacertilia, Gekkonidae). Am. J. Anat. 187:12-20. Iwasaki, S., and K. Kobayashi (1992) Fine structure of the dorsal lingual epithelium of the Japanese lizard, Takydromus tachydromoides. Acta Anat. Nipponica 67:214-225. Iwasaki, S., and K. Miyata (1985) Scanning electron microscopy of the lingual dorsal surface of the Japanese lizard, Takydromus tachydromoides. Okajimas Folia Anat. Jap. 62:15-26. Jackson, D. R., and S. R. Telford, Jr. (1975) Food habits and predatory role of the Japanese lacertid Takydromus tachydromoides. Copeia 1975:343-351. Jaksic, F. M., and K. Schwenk (1983) Natural history observations on
Liolaemus magellanicus, the southernmost lizard in the world. Herpetologica 39:457-461. Jaksic, F. M., E. R. Fuentes, and J. L. Yanez (1979) Two types of adaptation of vertebrate predators to their prey. Arch. Biol. Med. Exper12:143-152. James, C. D. (1991) Temporal variation in diets and trophic partitioning by coexisting lizards (Ctenotus: Scincidae) in central Australia. Oecologia 85:553-561. James, C. D., and J. B. Losos (1991) Diet and reproductive biology of the Australian sand-swimming lizards, Eremiascincus (Scincidae). Wildl. Res. 18:641-654. James, CD., J. B. Losos, and D. R. King (1992) Reproductive biology and diets of goannas (Reptilia: Varanidae) from Australia. J. Herp. 26:128-136. Jenkins, R. L., and W. W. Tanner (1968) Osteology and myology of Phrynosoma p. platyrhinos Girard and Phrynosoma d. hernandesi Girard. Brigham Young Univ. Sci. Bull. Biol. Ser. 9(4): 1-34. Jollie, M. T. (1960) The head skeleton of the lizard. Acta Zool. 4 1 : 1 - 64. Kamal, A. M. (1965) The cranial osteology of adult Chalcides ocellatus. Anat. Anz. 117:338-370. Kamal, A. M., and S. K. Zada (1970) The phylogenetic position of the family Agamidae in the light of the study of the chondrocranium. Zool. Anz. 184:327-335. Kallius, E. (1901) Beitrage zur Entwickelung der Zunge. I. Teil. Amphibien und Reptilien. Anatomischen Institut zu Gottingen, Anatomischen Heften 52/53:532-760. Kardong, K. V., T. L. Kiene, and V. Bels (1997) Evolution of trophic systems in squamates. Neth. J. Zool. 47:411-427. Kathariner, L. (1895) Anatomie und Mechanismus der Zunge der Vermilinguier. Jena. Zeit. Naturwis. 29:247-268. Kaufman, J. D., G. M. Burghardt, and J. A. Phillips (1996) Sensory cues and foraging decisions in a large carnivorous lizard, Varanus albigularis. Anim. Behav. 52:727-736. Kennedy, J. P. (1956) Food habits of the rusty lizard, Sceloporus olivaceus Smith. Texas J. Sci. 8:328-349. Kesteven, H. L. (1944) The evolution of the skull and the cephalic muscles: A comparative study of their development and adult morphology. III. The Sauria (Reptilia). Aust. Mus. Mem. 8, No. 3: 237-269. Kesteven, H. L. (1957) Noes on the skull and cephalic muscles of the Amphisbaenia. Proc. Linn. Soc. N. S. W. (Aust.) 82:109-116. Kier, W. M., and K. K. Smith (1985) Tongues, tentacles and trunks: the biomechanics of movement in muscular-hydrostats. Zool. J. Linn. Soc. 83:307-324. Kirmse, W. (1988) Foveal and ambient visuomotor control in chameleons (Squamata): experimental results and comparative review. Zool. Jb. Physiol. 92:341-350. Kley, N. J., and E. L. Brainerd (1996). Internal concertina swallowing: a critical component of alethinophidian feeding systems. Am. Zool. 35:81 A. Kline, L. W. (1983) The tooth replacement phenomenon and growth in the green iguana. Iguana iguana. Experientia 39:595-596. Kline, L. W., and D. Cullum (1984) A long term study of the tooth replacement phenomenon in the young green iguana. Iguana iguana. J. Herp. 18:176-185. Kline, L. W., and D. Cullum (1985) Tooth replacement and growth in the yoimg green iguana. Iguana iguana. J. Morph. 186:265-269. Kluge, A. G. (1962) Comparative osteology of the eublepharid lizard genus Coleonyx Gray. J. Morph. 110:299-332. Kluge, A. G. (1987) Cladistic relationships in the Gekkonoidea. Misc. Publ. Mus. Zool. Univ. Michigan No. 173:1-54. Kochva, E. (1978) Oral glands of Reptilia. Pp. 43-161. In: Biology of the Reptilia, Vol. 8. C. Gans and K. A. Cans (eds.). Academic Press, New York. Kraklau, D. M. (1991) Kinematics of prey capture and chewing in
8. Feeding in Lepidosaurs the lizard Agatna agama (Squamata: Agamidae). J. Morph. 210: 195-212. Kritzinger, C. C. (1946) The cranial anatomy and kinesis of the South African amphisbaenid Monopeltis capensis Smith. S. Afr. J. Sci. 42: 175-204. Kumazawa, Y., and M. Nishida (1999) Complete mitochondrial DNA sequences of the green turtle and blue-tailed mole skink: statistical evidence for archosaurian affinity of turtles. Mol. Biol. Evol. 16:784-792. Lakjer, T. (1926) Studien iiber die Trigeminus-versorgte Kaumuskulatur der Sauropsiden. C. A. Rietzel, Copenhagen. Lakjer, T. (1927) Studien iiber die Gaumenregion bei Sauriern im Vergleich mit Anamniern und primitiven Sauropsiden. Zool. Jahrb. Anat. 49:57-356. Land, M. R (1995) Fast-focus telephoto eye. Nature 373:658-659. Lang, M. (1989) Phylogenetic and biogeographic patterns of basiliscine iguanians (Reptilia: Squamata: "Iguanidae")- Borm. Zool. Monogr.No.28:l-172. Langebartel, D. A. (1968) The Hyoid and Its Associated Musculature in Snakes. Illinois Biological Monographs No. 38. Univ. of Illinois, Urbana, IL. Lappin, A. K. (1999) Ecomorphological Evolution of the Feeding Biology of Crotaphytid Lizards. Unpublished Ph.D. dissertation, Univ. of California, Berkeley, CA. Lappin, A. K., and E. J. Swinney (1999) Sexual dimorphism as it relates to natural history of leopard lizards (Crotaphytidae: Gambelia). Copeia 1999:649-660. Lauder, G. V. (1981) Form and function: structural analysis in evolutionary morphology. Paleobiology 7:430-442. Lauder, G. V. (1983a) Functional and morphological bases of trophic specialization in sunfishes (Teleostei, Centrarchidae). J. Morph. 178:1-21. Lauder, G. V. (1983b) Neuromuscular patterns and the origin of trophic specialization in fishes. Science 219:1235-1237. Lauder, G. V. (1990) Functional morphology and systematics: studying functional patterns in an historical context. Annu. Rev. Ecol. Syst. 21:317-340. Lee, M. S. Y. (1997) The phytogeny of varanoid lizards and the affinities of snakes. Phil. Trans. Roy. Soc. Lond. B 352:53-91. Lee, M. S. Y. (1998). Convergent evolution and character correlation in burrowing reptiles: towards a resolution of squamate relationships. Biol. J. Linn. Soc. 65:369-453. Lee, M. S. Y. (manuscript) Soft anatomy, squamate phytogeny and diffuse homoplasy. Lee, M. S. Y, and M. W. Caldwell (1998) Anatomy and relationships of Pachyrhachis prohlematicus, a primitive snake with hindlimbs. Phil. Trans. Roy. Soc. Lond. B 353:1521-1552. Lee, M. S. Y, G. L. Bell, Jr., and M. W. Caldwell (1999) The origin of snake feeding. Nature 400:655-659. Lightoller, G. H. S. (1939) Probable homologues: a study of the comparative anatomy of the mandibular and hyoid arches and their musculature. I. Comparative myology. Trans. Zool. Soc. Lond. 24: 349-445. Lonnberg, E. (1903) On the adaptations to a moUuscivorous diet in Varanus niloticus. Ark. Zool. 1:65-83. Lopez, P., J. Martin, and A. Salvador (1991) Diet selection by the amphisbaenian Blanus cinereus. Herpetologica 47:210-218. Loveridge, A. (1953) Zoological results of a fifth expedition to East Africa. III. Reptiles from Nyasaland and Tete. Bull. Mus. Comp. Zool. 110:143-322. Losos, J. B., and H. W. Greene (1988) Ecological and evolutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35:379407. Lubosch, W. (1932) Bemerkungen iiber die Zungenmuskulatur des Chamaleons. Gegenbaurs Morph. Jahrb. 71:158-170.
285
Lubosch, W. (1933) Untersuchungen iiber die Visceral muskulatur der Sauropsiden. (Der Untersuchungen iiber die Kaumuskulatur der Wirbeltiere 3. Teil). Gegenbaurs Morph. Jb. 72:584-666. MacLean, W. P. (1974) Feeding and locomotor mechanisms of teiid lizards: functional morphology and evolution. Pap. Avulsos Zool. (Sao Paulo) 27:179-213. Macey, J. R., and A. Verma (1997) Homology in phylogenetic analysis: alignment of transfer RNA genes and the phylogenetic position of snakes. Mol. Phylog. Evol. 7:272-279. Magnusson, W. E., L. Junqueira de Paiva, R. Moreira de Rocha, C. R. Franke, L. A. Kasper, and A. P. Lima (1985) The correlates of foraging mode in a community of Brazilian lizards. Herpetologica 41:324-332. Makaretz, M., and R. L. Levine (1980) A light microscopic study of the bifoveate retina in the lizard Anolis carolinensis: general observations and convergence ratios. Vision Res. 20:679-686. Malan, M. E. (1940) Cranial anatomy of the genus Gerrhosaurus. S. Afr. J. Sci. 37:192-217. Malan, M. E. (1946) Contributions to the comparative anatomy of the nasal capsule and the organ of Jacobson of the Lacertilia. Ann. Univ. Stellenbosch 24:69 -137. Marcellini, D. (1977) Acoustic and visual display behavior of gekkonid lizards. Am. Zool. 17:251-260. Martinez-Marcos, A., E. Lanuza, and M. Halpern (1999) Organization of the ophidian amygdala: chemosensory pathways to the hypothalamus. J. Comp. Neurol. 412:51-68. Marx, H. (1955) Notes on the behavior of some Madagascar chameleons (Reptilia, Chamaeleontidae). Nat. Hist. Misc. Chicago Acad. Sci. No. 149:1-3. Mateo, J., and L. F Lopez-Jurado (1992) Study of dentition in lizards from Gran Canaria Island (Canary Islands) and its ecological and evolutionary significance. Biol. J. Linn. Soc. 46:39-48. Mateo, J., and L. F. Lopez-Jurado (1997) Dental ontogeny in Lacerta lepida (Sauria, Lacertidae) and its relationship to diet. Copeia 1997:461-463. McBee, R. H., and V. H. McBee (1982) The hindgut fermentation in the green iguana, Iguana iguana. Pp. 77-83. In: Iguanas of the World. Their Behavior, Ecology, and Conservation. G. M. Burghardt and A. S. Rand (eds.). Noyes, Park Ridge, NJ. McBrayer, L. D., and T. D. White (manuscript) Bite force, behavior and electromyography in the teiid lizard, Tupinambis teguixin. Submitted for publication. McDowell, S. B. (1972) The evolution of the tongue of snakes, and its bearing on snake origins. Pp. 191-273. In: Evolutionary Biology, Vol. 6. T. Dobzhansky, M. K. Hecht, and W. C. Steere (eds.). Appleton-Century-Crofts, New York. McDowell, S. B., Jr., and C. M. Bogert (1954) The systematic position of Lanthanotus and the affinities of the anguinomorphan lizards. Bull. Am. Mus. Nat. Hist. 105:1-142. McGovern, G. M., J. C. Mitchell, and C. B. Knisley (1984) Field experiments on prey selection by the whiptail lizard, Cnemidophorus inornatus, in Arizona. J. Herp. 18:347-349. Meyer-Rochow, V. B. (1988) Behaviour of young tuatara (Sphenodon punctatus) in total darkness. Tuatara 30:36-38. Millikan, R. G. (1984) Language, Thought, and Other Biological Categories. MIT Press, Cambridge, Mass. Minot, C. S. (1880) Studies on the tongue of reptiles and birds. Anniv. Mem. Boston Soc. Nat. Hist. 1880:3-20. Mirwald, M., and S. F. Perry (1991) Muscle fiber types in ventilatory and locomotor muscles of the tokay, Gekko gecko: a histochemical study. Comp. Biochem. Physiol. A 98:407-411. Mivart, St. G. (1870) On the myology of Chamaeleon parsonii. Proc. Zool. Soc. Lond. No. 57:850-890. Montanucci, R. R. (1968) Comparative dentition in four iguanid lizards. Herpetologica 24:305-315.
286
Kurt Schwenk
Montanucci, R. R. (1989) The relationship of morphology to diet in the horned lizard genus Phrynosoma. Herpetologica 45:208-216. Montero, R., C. Gans, and M. L. Lions (1999) Embryonic development of the skeleton of Amphisbaena darwini heterozonata (Squamata: Amphisbaenidae). J. Morph. 239:1-25. Mori, A., K. Araya, and T. Hikida (1995) Biology of the poorly known Bornean lizard, Apterygodon vittatus (Squamata: Scincidae): an arboreal ant-eater. Herp. Nat. Hist. 3:1-14. Murray, B. A., S. D. Bradshaw, and D. H. Edward (1991) Feeding behavior and the occurrence of caudal luring in Burton's pygopodid Lialis burtonis (Sauria: Pygopodidae). Copeia 1991:509-516. Nagy, K. A., R. B. Huey, and A. F. Bennett (1984) Field energetics and foraging mode of Kalahari lacertid lizards. Ecology 65:588596. Nash, D. F., and W. W. Tanner (1970) A comparative study of the head and thoracic osteology and myology of the skinks Eumeces gilberti Van Denburgh and Eumeces skiltonianus (Baird and Girard). Brigham Young Univ. Sci. Bull. Biol. Ser. 12(2): 1-32. Nicoletto, R F. (1985a) The relative roles of vision and olfaction in prey detection by the ground skink, Scincella lateralis. J. Herp. 19: 411-415. Nicoletto, P. F. (1985b) The roles of vision and the chemical senses in predatory behavior of the skink, Scincella lateralis. J. Herp. 19: 487-491. Oelrich, T. M. (1954) The anatomy of the head of Ctenosaura pectinata (Iguanidae). Misc. Publ. Mus. Zool. Univ. Michigan No. 94. Okelo, O. (1986) Neuroendocrine control of physiological color change in Chameleo gracilis. Gen. Comp. Endocrinol. 64: 305-311. Osawa, G. (1897) Beitrage zur Lehre von den Eingeweiden der Hatteria punctata. Arch. Mikrosk. Anat. 49:113-226. Osawa, G. (1898) Beitrage zur anatomie der Hatteria punctata. Arch. Mikrosk. Anat. 54:481-676. Ostrom, J. H. (1962) On the constrictor dorsalis muscles of Sphenodon. Copeia 1962:732-735. Ott, M., and F. Schaeffel (1995) A negatively powered lens in the chameleon. Nature 373:692-694. Ott, M., F. Schaeffel, and W. Kirmse (1998) Binocular vision and accommodation in prey-catching chameleons. J. Comp. Physiol. A 182:319-330. Owen, R. (1866) On the Anatomy of Vertebrates. Vol. 1. Longmans, Green, and Co., London. Owerkowicz, T., C. G. Farmer, J. W. Hicks, and E. Brainerd (1999) Contribution of gular pumping to lung ventilation in monitor lizards. Science 284:1661-1663 Owerkowicz, T., E. Brainerd, and D. Carrier (2000) Electromyographic pattern of the gular pump in monitor lizards. Bull. Mus. Comp. Zool. (in press). Paranjape, S. Y. (1974) The Anatomy of the Garden Lizard. Univ. of Poona Zoological Monogr., Poona Univ. Press, Poona, India. Parcher, S. R. (1974) Observations on the natural histories of six Malagasy Chamaeleontidae. Zeit. Tierpsych. 34:500-523. Parker, W. K. (1880) On the structure and development of the skull in the Lacertilia. L On the skull of the common lizards (Lacerta agilis, L. viridis, and Zootoca vivipara). Phil. Trans. Roy. Soc. 170: 595-640. Parker, W. K. (1881) On the structure of the skull in the chameleons. Trans Zool. Soc. Lond. 11:77-105. Parker, W. S., and E. R. Pianka (1976) Ecological observations on the leopard lizard {Crotaphytus wislizeni) in different parts of its range. Herpetologica 32:95-114. Parsons, T. S. (1970) The nose and Jacoboson's organ. Pp. 99-191. In: Biology of the Reptilia, Vol. 2. C. Gans and T. S. Parsons (eds.). Academic Press, New York.
Patchell, F. C , and R. Shine (1986a) Feeding mechanisms in pygopodid lizards: how can Lialis swallow such large prey? J. Herp. 20:59-64. Patchell, F. C , and R. Shine (1986b) Food habits and reproductive biology of the Australian legless lizards (Pygopodidae). Copeia 1986:30-39. Patchell, F. C , and R. Shine (1986c) Hinged teeth for hard-bodied prey: a case of convergent evolution between snakes and legless lizards. J. Zool. Lond. 208:269-275. Perez-Mellado, V. (1994) Tarentola mauritanica (Moorish gecko). Behav. Herp. Rev. 25:68-69. Perez-Mellado, V, and J. L. Casas (1997) Pollination by a lizard on a Mediterranean island. Copeia 1997:593-595. Perez-Mellado, V, and C. Corti (1993) Dietary adaptations and herbivory in lacertid lizards of the genus Podarcis from western Mediterranean islands (Reptilia: Sauria). Bonn. Zool. Beitr. 44: 193-220. Perrault, C. (1676) Memoires pour servir a I'histoire naturelle des animaux. Academie Royale des sciences de Paris. [1688 English translation by Alexander Pitfield, Memoirs for a natural history of animals containing the anatomical descriptions of several creatures dissected by the Royal Academy of Sciences at Paris. Joseph Streater London]. Perry, G. (1999) The evolution of search modes: ecological versus phylogenetic perspectives. Am. Nat. 153:98-109. Perry, G., and M. Brandeis (1992) Variation in stomach contents of the gecko Ptyodactylus hasselquistii guttatus in relation to sex, age, season and locality. Amph.-Rept. 13:275-282. Perry, G., I. Lampl, A. Lerner, D. Rothenstein, E. Shani, N. Sivan, and Y. L. Werner (1990) Foraging mode in lacertid lizards: variation and correlates. Amph.-Rept. 11:373-384. Persons, M. H., L. J. Fleishman, M. A. Frye, and M. E. Stimphil (1999) Sensory response patterns and the evolution of visual-signal design in anoline lizards. J. Comp. Physiol. A. 184:585-607. Peterson, E. H. (1992) Retinal structure. Pp. 1-135. In: Biology of the Reptilia, Vol. 17. C. Gans and P. S. Ulinski (eds.). Univ. of Chicago Press, Chicago. Peterson, J. A. (1984) The locomotion of Chamaeleo (Reptilia: Sauria) with particular reference to the forelimb. J. Zool. Lond. 202:1-42. Peyer, B. (1929) Das Gebiss von Varanus niloticus L. und von Dracaena guianensis Daud. Rev. Suisse Zool. 36:71-102. Pianka, E. R. (1966) Convexity, desert lizards, and spatial heterogeneity. Ecology 47:1055-1059. Pianka, E. R. (1971) Ecology of the agamid lizard Amphibolurus isolepis in Western Australia. Copeia 1971:527-536. Pianka, E. R. (1986) Ecology and Natural History of Desert Lizards. Princeton Univ. Press, Princeton, NJ. Pianka, E. R., and W. S. Parker (1975) Ecology of horned lizards: a review with special reference to Phrynosoma platyrhinos. Copeia 1975:141-162. Pianka, E. R., and H. D. Pianka (1970) The ecology of Moloch horridus (Lacertilia: Agamidae) in Western Australia. Copeia 1970: 90-103. Pianka, E. R., and H. D. Pianka (1976) Comparative ecology of twelve species of nocturnal lizards (Gekkonidae) in the Western Australian desert. Copeia 1976:125-142. Pietruszka, R. D. (1986) Search tactics of desert lizards: how polarized are they? Anim. Behav. 34:1742-1758. Ping, C. (1932) Anatomy of the tongue and its adjacent muscles of Hemidactylus bowringii. Contr. Biol. Lab. Sci. Soc. China 8:97-111. Pitman, C. R. S. (1958) Snake and lizard predators of birds. Bull. Br. Ornith. Club 78:120-124. Pitman, C. R. S. (1962) More snake and lizard predators of birds. Bull. Br. Ornith. Club 82:45-55.
8. F e e d i n g in L e p i d o s a u r s Poglayen-Neuwall, J. (1953) Untersuchungen iiber die Trigeminusmuskulatur von Hatteria. Zeit. wissensch. Zool. 157:57-76. Poglayen-Neuwall, J. (1954) Die Kiefermuskulatur der Eidechsenund ihre Innervation. Zeit. Wissensch. Zool. 158:79-132. Pough, F. H. (1973) Lizard energetics and diet. Ecology 54:837-844. Pough, F. H. (1983) Feeding mechanisms, body size, and the ecology and evolution of snakes. Introduction to the symposium. Am. Zool. 23:339-342. Pough, F. H., and R. M. Andrews (1985) Energy costs of subduing and swallowing prey for a lizard. Ecology 66:1525-1533. Pough, F H., R. M. Andrews, J. E. Cadle, M. L. Crump, A. H. Savitzky, and K. D. Wells (1998) Herpetology. Prentice Hall, Upper Saddle River, NJ. Pratt, C. W. M. (1948) The morphology of the ethmoidal region of Sphenodon and lizards. Proc. Zool. Soc. Lond. 118:171-201. Pregill, G. K., J. A. Gauthier, and H. W. Greene (1986) The evolution of helodermatid lizards, with description of a new taxon and an overview of Varanoidea. Trans. San Diego Soc. Nat. Hist. 21: 167-202. Presch, W. (1971) Tongue structure of the teiid lizard genera Ameiva and Cnemidophorus with a reallocation of Ameiva vanzoi. J. Herp. 5:183-185. Presch, W. (1974) A survey of the dentition of the macroteiid lizards (Teiidae: Lacertilia). Herpetologica 30:344-349. Presch, W. (1976) Secondary palate formation in microteiid lizards (Teiidae: Lacertilia). Bull. S. Calif. Acad. Sci. 75:281-283. Putnam, R. W., T. T. Gleeson, and A. F Bennett (1980) Histochemical determination of the fiber composition of locomotory muscles in a lizard, Dipsosaurus dorsalis. J. Exp. Zool. 214:303-309. Rabinowitz, T., and B. Tandler (1986) Papillary morphology of the tongue of the American chameleon: Anolis carolinensis. Anat. Rec. 216:483-489. Rabinowitz, T., and B. Tandler (1991) Ultrastructure of lingual salivary glands in the American chameleon: Anolis carolinensis. Anat. Rec. 229:489-494. Radcliffe, C. W., D. Chiszar, and H. M. Smith (1980) Prey induced caudal movements in Boa constrictor with comments on the evolution of caudal luring. Bull. Maryland Herp. Soc. 16:19-22. Ramaswami, L. S. (1946) The chondrocranium of Calotes versicolor (Daud.) with a description of the osteocranium of a just-hatched young. Quart. J. Microsc. Sci. 87:237-297. Rand, A. S. (1961) A suggested function of the ornamentation of east African forest chameleons. Copeia 1961:411-414. Rand, A. S., B. A. Dugan, H. Monteza, and D. Vianda (1990) The diet of a generalized folivore: Iguana iguana in Panama. J. Herp. 24: 211-214. Rand, A. S., G. C. Gorman, and W. M. Rand (1975) Natural history, behavior, and ecology of Anolis agassizi. Pp. 27-38. In: The Biological Investigation ofMalpelo Island, Colombia. J. B. Graham (ed.). Smithsonian Contributions to Zoology No. 176, Smithsonian Inst. Press, Washington, DC. Rao, M. K. M., and L. S. Ramaswami (1952) The fully formed chondrocranium of Mabuya with an account of the adult osteocranium. Acta Zool. Stockholm 33:209-275. Regal, P. J. (1978) Behavioral differences between reptiles and mammals: an analysis of activity and mental capabiHties. Pp. 183202. In: Behavior and Neurology of Lizards. N. Greenberg and P. D. MacLean (eds.). U. S. Dept. of Health, Education and Welfare Publ. No. (ADM) 77-491, Natl. Inst. Mental Health, Bethesda, MD. Regal, P. J. (1983) The adaptive zone and the behavior of lizards. Pp. 105-118. In: Lizard Ecology: Studies of a Model Organism. R. B. Huey, E. R. Pianka, and T. W. Schoener (eds.). Harvard Univ. Press, Cambridge, MA.
287
Renous, S. (1977) Musculature of the buccal floor of Bipes canaliculatus (Reptilia: Amphisbaenia). Copeia 1977:464-471. Reperant, J., J.-P. Rio, R. Ward, S. Hergueta, D. Miceli, and M. Lemire (1992) Comparative analysis of the primary visual sytem of reptiles. Pp. 175-240. In: Biology of the Reptilia, Vol. 17. C. Cans and P. S. Ulinski (eds.). Univ. of Chicago Press, Chicago. Richter, H. (1932) Das Zungenbein und seine Muskulatur bei den Lacertilia vera. Jena. Zeit. Naturwis. 66:395-480. Rice, E. L. (1920) The development of the skull in the skink, Eumeces quinquelineatus L. J. Morph. 34:119-216. Rice, M. J. (1973) Supercontracting striated muscle in a vertebrate. Nature 243:238-240. Rieppel, O. (1978a) Streptostyly and muscle function in lizards. Experientia 34:776-777. Rieppel, O. (1978a) The braincase of Anniella pulchra Gray (Lacertilia: Anniellidae). Rev. Suisse Zool. 85:617-624. Rieppel, O. (1978c) The evolution of the naso-frontal joint in snakes and its bearing on snake origins. Zeit. Zool. Syst. Evolut.-Forsch. 16:14-27. Rieppel, O. (1978d) The phylogeny of cranial kinesis in lower vertebrates, with special reference to the Lacertilia. N. Jb. Geol. Palaont.Abh. 156:353-370. Rieppel, O. (1978e) The throat musculature of Sphenodon, with comments on the primitive character states of the throat muscles in lizards. Anat. Anz. 144:429-440. Rieppel, O. (1978f) Tooth replacement in anguinomorph lizards. Zoomorphologie 91:77-90. Rieppel, O. (1979a) A functional interpretation of the varanid dentition. Gegenbaurs Morphol. Jahrb. 125:797-817. Rieppel, O. (1979b) The external jaw adductor of amphisbaenids (Reptilia: Amphisbaenia). Rev. Suisse Zool. 86:867-876. Rieppel, O. (1980) The trigeminal jaw adductor musculature of Tupinambis, with comments on the phylogenetic relationships of the Teiidae (Reptilia, Lacertiha). Zool. J. Linn. Soc. 69:1-29. Rieppel, O. (1981a) The hyobranchial skeleton in some little known lizards and snakes. J. Herp. 15:433-440. Rieppel, O. (1981b) The skull and jaw adductor musculature in chameleons. Rev. Suisse Zool. 88:433-445. Rieppel, O. (1981c) The skull and the jaw adductor musculature in some burrowing scincomorph lizards of the genera Acontias, Typhlosaurus and Feylinia. J. Zool. Lond. 195:493-528. Rieppel, O. (1983) A comparison of the skull oi Lanthanotus borneensis (Reptilia: Varanoidea) with the skull of the primitive snakes. Zeit. Zool. Syst. Evolut.-Forsch. 21:142-153. Rieppel, O. (1984a) Miniaturization of the lizard skull: its functional and evolutionary implications. Pp. 503-520. In: The Structure, Development and Evolution of Reptiles. M. W. J. Ferguson (ed.). Zoological Society of London and Academic Press, London (Symp. Zool. Soc. Lond. No. 52). Rieppel, O. (1984b) The cranial morphology of the fossorial lizard genus Dibamus with a consideration of its phylogenetic signficance. J. Zool. Lond. 204:289-327. Rieppel, O. (1984c) The upper temporal arcade of lizards: an ontogenetic problem. Rev. Suisse Zool. 91:475-482. Rieppel, O. (1987) The phylogenetic relationships within the Chamaeleonidae, with comments on some aspects of cladistic analysis. Zool. J. Linn. Soc. 89:41-62. Rieppel, O. (1988) A review of the origin of snakes. Pp. 37-130. In: Evolutionary Biology, Vol. 22. M. K. Hecht, B. Wallace, and G. T. Prance (eds.). Plenum Press, New York. Rieppel, O. (1992a) Studies on skeleton formation in reptiles. I. The postembryonic development of the skeleton in Cyrtodactylus pubisulcus (Reptilia: Gekkonidae). J. Zool. Lond. £27:87-100. Rieppel, O. (1992b) Studies on skeleton formation in reptiles. III.
288
Kurt Schwenk
Patterns of ossification in the skeleton of Lacerta vivipara Jacquin (Reptilia, Squamata). Fieldiana, ZooL N.S. No. 68:1-25. Rieppel, O. (1993) Patterns of diversity in the reptilian skull. Pp. 344390. In: The Skull Vol. 2. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Rieppel, O., and C. Crumly (1997) Paedomorphosis and skull structure in Malagasy chamaeleons (Reptilia: Chamaeleonidae). J. Zool. Lond. 243:351-380. Rieppel, O., and R. W. Gronowski (1981) The loss of the lower temporal arcade in diapsid reptiles. Zool. J. Linn. Soc. 72:203217. Rieppel, O., and L. Labhardt (1979) Mandibular mechanics in Varanus niloticus (Reptilia: Lacertilia). Herpetologica 35:158-163. Riley, J., J. M. Winch, A. F. Stimson, and R. D. Pope (1986) The association of Amphisbaena alba (Reptilia: Amphisbaenia) with the leaf-cutting ant Atta cephalotes in Trinidad. J. Nat. Hist. 20:459470. Rissing, S. W. (1981) Prey preferences in the desert homed lizard: influence of prey foraging method and aggressive behavior. Ecology 62:1031-1040. Robinson, P L. (1967) The evolution of the Lacertilia. Pp. 395-407. In: Problemes Actuels de Paleontologie (Evolution des Vertebres). Colloques Intemationaux du CNRS, No. 163, Paris. Robinson, P. L. (1976) How Sphenodon and Uromastyx grow their teeth and use them. Pp. 43-64. In: Morphology and Biology of Reptiles. Symp. Linn. Soc. No. 3. A. d' A. Bellairs and C. B. Cox (eds.). Academic Press, London. Robison, W. C , Jr., and W. W. Tanner (1962) A comparative study of the species of the genus Crotaphytus Holbrook (Iguanidae). Brigham Young Univ. Sci. Bull. Biol. Ser. 2(1):1-31. Rocek, Z. (1980a) Intraspecific and ontogenetic variation of the dentition in the green lizard Lacerta viridis (Reptilia, Squamata). Vest, cs. Spolec. zool. 44:272-277. Rocek, Z. (1980b) The dentition of the European glass lizard Ophisaurus apodus (Pallas, 1775) (Reptilia, Sauria: Anguidae), with notes on the pattern of tooth replacement. Amph.-Rept. 1:19-27, Rocha, C. F. D. (1989) Diet of a tropical lizard {Liolaemus lutzae) of southeastern Brazil. J. Herp. 23:292-294. Rocha, C. F. D. (1996) Seasonal shift in lizard diet: the seasonality in food resources affecting the diet of Liolaemus lutzae (Tropiduridae) Ciencia e Cultura 48:264-269. Rocha, C. F. D. (1998) Ontogenetic shift in the rate of plant consumption in a tropical lizard [Liolaemus lutzae). J. Herp. 32:274-279. Romer, A. S. (1956) Osteology of the Reptiles. Univ. of Chicago Press, Chicago. Rose, B. R. (1976) Dietary overlap of Sceloporus occidentalis and S. graciosus. Copeia 1976:818-820. Sakaluk, S. K., and J. J. Belwood (1984) Gecko phonotaxis to cricket calling song: a case of satellite predation. Anim. Behav. 32:659662. Save-Soderbergh, G. (1945) Notes on the trigeminal musculature in nonmammalian tetrapods. Nova Acta Reg. Soc. Sci. Upsaliensis 13:1-59. Save-Soderbergh, G. (1947) Notes on the brain-case in Sphenodon and certain Lacertilia. ZooL Bidrag Uppsala 25:489-516. Savitzky, A. H. (1981) Hinged teeth in snakes: an adaptation for swallowing hard-bodied prey. Science 212:346-349. Savitzky, A. H. (1983) Coadapted character complexes among snakes: fossoriality, piscivory, and durophagy. Am. Zool. 23:397-409. Schall, J. J., and S. Ressel (1991) Toxic plant compoimds and the diet of the predominantly herbivorous whiptail lizard, Cnemidophorus arubensis. Copeia 1991:111-119. Schauinsland, H. (1900) Weitere Beitrage zur Entwicklungsgeschichte der Hatteria. Skelettsystem, schalleitender Apparat, Hirnnerven etc. Arch. Mikr. Anat. 56:747-767.
Schauinsland, H. (1903) Beitrage zur Entwicklungsgeschichte und Anatomie der Wirbeltiere. I. Sphenodon, Callorhynchus, Chamaleo. Zoologica (Stuttgart) 16:1-198. Schmidt, K. P, and R. R Inger (1957) Living Reptiles of the World. Hanover House, Garden City, NY. Schmidt, P. J., and J. O. Schmidt (1989) Harvester ants and horned lizards predator prey interactions. Pp. 25-50. In: Special Biotic Relationships in the Arid Southwest. J. O. Schmidt (ed.). Univ. of New Mexico Press, Albuquerque, NM. Schmidt, P J., W. J. Sherbrooke, and J. O. Schmidt (1989) The detoxification of ant (Pogonmyrmex) venom by a blood factor in horned lizards (Phrynosoma). Copeia 1989:603-607. Schoener, T. W. (1971) Theory of feeding strategies. Annu. Rev. Ecol. Syst. 2:369-404. Schoener, T. W., J. B. Slade, and C. H. Stinson (1982) Diet and sexual dimorphism in the very catholic lizard genus, Leiocephalus of the Bahamas. Oecologia 53:160-169. Schwenk, K. (1980) Functional morphology of cranial crests in lizards. Am. Zool. 20:783. Schwenk, K. (1983) Functional morphology and evolution of the chameleon tongue tip. Am. Zool. 23:1028. Schwenk, K. (1984) Evolutionary Morphology of the Lepidosaur Tongue. Unpublished Ph.D. dissertation, Univ. of California, Berkeley, CA. Schwenk, K. (1985) Occurrence, distribution and functional significance of taste buds in lizards. Copeia 1985:91-101. Schwenk, K. (1986) Morphology of the tongue in the tuatara, Sphenodon punctatus (Reptilia: Lepidosauria), with comments on function and phylogeny. J. Morph. 188:129-156. Schwenk, K. (1987) Evolutionary determinants of cranial form and function in lizards. Am. Zool. 27:105A. Schwenk, K. (1988) Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny. Pp. 569-598. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford University Press, Stanford, CA. Schwenk, K. (1993a) Are geckos olfactory specialists? J. Zool. Lond. 229:289-302. Schwenk, K. (1993b) The evolution of chemoreception in squamate reptiles: a phylogenetic approach. Brain Behav. Evol. 41:124-137. Schwenk, K. (1994a) Comparative biology and the importance of cladistic classification: a case study from the sensory biology of squamate reptiles. Biol. J. Linn. Soc. 52:69-82. Schwenk, K. (1994b) Craniology: getting a head. Science 263:17791780. Schwenk, K. (1994c) Form, function and evolution of lingual feeding in ant-eating lizards, genus Phrynosoma (Squamata, Iguania). J. Morph. 220:392. Schwenk, K. (1994d) Systematics and subjectivity: the phylogeny and classification of iguanian lizards reconsidered. Herp. Rev. 25:53-57. Schwenk, K. (1994e) Why snakes have forked tongues. Science 263: 1573-1577. Schwenk, K. (1995a) A utilitarian approach to evolutionary constraint. Zoology 98:251-262. Schwenk, K. (1995b) Of tongues and noses: chemoreception in lizards and snakes. Tr. Ecol Evol. 10:7-12. Schwenk, K. (2000a) Extrinsic vs intrinsic lingual muscles: a false dichotomy? Bull. Mus. Comp. Zool. (in press). Schwenk, K. (2000b) Functional units and their evolution. In: The Character Concept in Evolutionary Biology. G. P. Wagner (ed.). Academic Press, San Diego (in press). Schwenk, K., and D. A. Bell (1986) Chameleon-like tongue protrusion in an agamid lizard. Am. Zool. 26:65A. Schwenk, K., and D. A. Bell (1988) A cryptic intermediate in the evolution of chameleon tongue projection. Experientia 44:697-700. Schwenk, K., and G. S. Throckmorton (1989) Functional and evolu-
8. Feeding in Lepidosaurs tionary morphology of lingual feeding in squamate reptiles: phylogenetics and kinematics. J. Zool. Lond. 219:153-175. Schwenk, K., and D. B. Wake (1988) Medium-independent feeding in a plethodontid salamander: tongue projection and prey capture underwater. Am. Zool. 28:115A. Schwenk, K., and D. B. Wake (1993) Prey processing in Leurognathus marmoratus and the evolution of form and function in desmognathine salamanders (Plethodontidae). Biol. J. Linn. Soc. 49:141162. Schwenk, K., E. Brainerd, and A. W. Crompton (manuscript) Biomechanics of the mammal tongue: role of the lingual tunic. Sewertzoff, S. A. (1923) Die Entwicklungsgeschichte der Zunge des Chamaeleo bilineatus. Rev. Zool. Russe 3:263-283. Sewertzoff, S. A. (1929) Zur Entwicklungsgeschichte der Zunge bei den Reptilien. Acta Zool. 10:231-341. Shaffer, D. T., Jr., and W. G. Whitford (1981) Behavioral responses of a predator, the round-tailed horned lizard, Phrynosoma modestum and its prey, honey pot ants, Myrmecocystus spp. Am. Midi. Nat. 105:209-216. Shafir, S., and J. Roughgarden (1998) Testing predictions of foraging theory for a sit-and-wait forager, Anolis gingivinus. Behav. Ecol. 9:74-84. Shea, G. M. (1990) The genera Tiliqua and Cydodomorphus (Lacertilia: Scincidae): generic diagnoses and systematic relationships. Mem. Queensland Mus. 29:495-519. Shine, R. (1986) Food habits, habitats and reproductive biology of four sympatric species of varanid lizards in tropical Australia. Herpetologica 42:346-360. Sillman, A. J. (1973) Avian vision. Pp. 349-387. In: Avian Biology, Vol. 3. D. S. Earner, J. R. King, and K. C. Parkes (eds.). Academic Press, New York. Smith, K. K. (1980) Mechanical significance of streptostyly in lizards. Nature 283:778-779. Smith, K. K. (1982) An electromyographic study of the function of the jaw adducting muscles in Varanus exanthematicus (Varanidae). J. Morph. 173:137-158. Smith, K. K. (1984) The use of the tongue and hyoid apparatus during feeding in lizards (Ctenosaura similis and Tupinambis nigropunctatus). J. Zool. Lond. 202:115-143. Smith, K. K. (1986) Morphology and function of the tongue and hyoid apparatus in Varanus (Varanidae, Lacertilia). J. Morph. 187: 261-287. Smith, K. K. (1988) Form and function of the tongue in agamid lizards with comments on phylogenetic significance. J. Morph. 196: 157-171. Smith, K. K. (1992) The evolution of the mammalian pharynx. Zool. J. Linn. Soc. 104:313-349. Smith, K. K. (1993) The form of the feeding apparatus in terrestrial vertebrates: studies of adaptation and constraint. Pp. 150-196. In: The Skull Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Smith, K. K. (1994) Are neuromuscular systems conserved in evolution? Brain Behav. Evol. 43:293-305. Smith, K. K., and W. L. Hylander (1985) Strain gage measurement of mesokinetic movement in Varanus exanthematicus. J. Exp. Biol. 114:53-70. Smith, K. K., and W. M. Kier (1989) Trunks, tongues, and tentacles: moving with skeletons of muscle. Am. Sci. 77:29-35. Smith, K. K., and K. A. MacKay (1990) The morphology of the intrinsic tongue musculature of snakes (Reptilia, Ophidia): functional and phylogenetic implications. J. Morph. 205:307-324. Smith, T. L., K. V. Kardong, and V. L. Bels (1999) Prey capture behavior in the blue-tongued skink, Tiliqua scincoides. J. Herp. 33:362369. So, K.-K. J., P C. Wainwright, and A. R Bennett (1992) Kinematics of
289
prey processing in Chamaeleo jacksonii: conservation of function with morphological specialization. J. Zool. Lond. 226:47-64. Solessio, E., and G. A. Engbretson (1993) Antagonistic chromatic mechanisms in photoreceptors of the parietal eye of lizards. Nature 364:442 - 445. Sondhi, K. C. (1958a) On a hitherto unknown disposition of the intrinsic musculature of the reptilian tongue. Pp. 422-424. In: Proceedings of the XVth International Congress of Zoology, Sect. 5, Vertebrate Zoology. Sondhi, K. C. (1958b) The hyoid and associated structures in some Indian reptiles. Ann. Zool. Acad. Zool. (Lond.) 2:155-239. Stamps, J., S. Tanaka, and V V. Krishnan (1981) The relationship between selectivity and food abundance in a juvenile lizard. Ecology 62:1079-1092. Stephenson, N. G. (1962) The comparative morphology of the head skeleton, girdles and hind limbs in the Pygopodidae. Zool. J. Linn. Soc. 44:627-644. Stimie, M. (1966) The cranial anatomy of the iguanid lizard Anolis carolinensis (Cuvier). Ann. Univ. Stellenbosch41A:239-268. Sumida, S. S., and R. W. Murphy (1987) Form and function of the tooth crown structure in gekkonid lizards (Reptilia, Squamata, Gekkonidae). Can. J. Zool. 65:2886-2892. Swanson, P. L. (1950) The iguana. Iguana iguana iguana. Herpetologica 6:187-193. Taib, N. T., and B. M. Jarrar (1985a) Histochemical analysis of mucosubstances in the lingual salivary glands of the lizard Agama blandfordi (Agamidae, Reptilia). Sudan J. Sci. 1:97-101. Taib, N. T., and B. M. Jarrar (1985b) Histochemical studies on the lingual salivary glands of the spiny-tailed lizard Uromastyx microlepis (Blandford). Bull. Inst. Zool. Acad. Sinica 24:203-212. Taib, N. T., and B. M. Jarrar (1986) The histochemistry of the lingual salivary glands of the lizard Acanthodactylus schmidti (Wiegmann) (Reptilia, Lacertilia, Lascertidae). Bull. Maryland Herp. Soc. 22: 27-36. Tanner, W. W., and D. F. Avery (1982) Buccal floor of reptiles, a summary. Great Basin Nat. 42:273-349. Tanner, W. W., and J. E. Krogh (1974) Ecology of the leopard lizard, Crotaphytus wislizeni at the Nevada Test Site, Nye County, Nevada. Herpetologica 30:63-72. Taylor, J. A. (1986) Food and foraging behaviour of the lizard, Ctenotus taeniolatus. Aust. J. Ecol. 11:49-54. Throckmorton, G. S. (1976) Oral food processing in two herbivorous hzards. Iguana iguana (Iguanidae) and Uromastix aegyptius (Agamidae). J. Morph. 148:363-390. Throckmorton, G. S. (1978) Action of the pterygoideus muscle during feeding in the lizard Uromastix aegyptius (Agamidae). Anat. Rec. 190:217-222. Throckmorton, G. S. (1979) The effect of wear on the cheek teeth and associated dental tissues of the lizard Uromastix aegyptius (Agamidae). J. Morph. 160:195-208. Throckmorton, G. S. (1980) The chewing cycle in the herbivorous lizard Uromastix aegyptius (Agamidae). Arch. Oral Biol. 25:225-233. Throckmorton, G. S. (1982) Histochemical properties of some jaw muscles of the lizard Tupinambis nigropunctatus (Teiidae). Anat. Rec. 203:345-352. Throckmorton, G. S., and L. K. Clarke (1981) Intracranial joint movements in the agamid lizard Amphibolurus barbatus. J. Exp. Zool. 216:25-35. Throckmorton, G. S., J. de Bavay, W. Chaffey, B. Merrotsy, S. Noske, and R. Noske (1985) The mechanism of frill erection in the bearded dragon Amphibolurus barbatus with comments on the jacky lizard A. muricatus (Agamidae). J. Morph. 183:285-292. Throckmorton, G. S., J. A. Hopson, and P. Parks (1981) A redescription of Toxolophosaurus cloudi Olson, a lower Cretaceous herbivorous sphenodontid reptile. J. Paleon. 55:586-597.
290
Kurt Schwenk
Tollestrup, K. (1979) The ecology, social structure, and foraging behavior of two closely related species of leopard lizards, Gambelia silus and Gambelia wislizenii. Unpubl. doctoral dissertation, Univ. of California, Berkeley, CA. Toubeau, G., C. Cotman, and V. Bels (1994) Morphological and kinematic study of the tongue and buccal cavity in the lizard Anguis fragilis (Reptilia: Anguidae). Anat. Rec. 240:423-433. Troyer, K. (1984) Diet selection and digestion in Iguana iguana: the importance of age and nutrient requirements. Oecologia 61:201207. Troyer, K. (1991) Role of microbial celulose degradation in reptile nutrition. Pp. 311-325. In: Biosynthesis and Biodegradation of Cellulose. C. H. Haigler and R H. Weimer (eds.). Dekker, New York. Tucker, D. (1971) Nonolfactory responses from the nasal cavity: Jacobson's organ and the trigeminal system. Pp. 151-181. In: Handbook of Sensory Physiology, Vol. IV. L. M. Beidler (ed.). Springer Verlag, Berlin. Twigg, L. E., R. A. How, R. L. Hatherly, and J. Dell (1996) Comparison of the diet of three sympatric species of Ctenotus skinks. J. Herp. 30:561-566. Underwood, G. (1957) On lizards of the family Pygopodidae: a contribution to the morphology and phytogeny of the Squamata. J. Morph. 100:207-268. Underwood, G. (1970) The eye. Pp. 1-97. In: Biology of the Reptilia, Vol. 2. C. Gans and T. S. Parsons (eds.). Academic Press, New York. Urbani J.-M., and V. L. Bels (1995) Feeding behaviour in two scleroglossan lizards: Lacerta viridis (Lacertidae) and Zonosaurus laticaudatus (Cordylidae). J. Zool. Lond. 236:265-290. Vagvolgyi, J., and A. E. Vabvolgyi (1978) Feeding habits of the Galapagos land iguana. Copeia 1978:162-163. van Leeuwen, J. L. (1997) Why the chameleon has spiral-shaped muscle fibres in its tongue. Phil. Trans. R. Soc. Lond. B. 352:573589. van Pletzen, R. (1946) The cranial morphology of Cordylus with special reference to the cranial kinesis. Ann. Univ. Stellenbosch 24: 41-68. Versluys, J. (1898) Die mittlere und aussere Ohrspare der Lacertilia und Rhynchocephalia. Zool. Jahrb. Anat. 12:161-406. Versluys, J. (1912) Das Streptostylie-Problem und die Bewegungen im Schadel bei Sauropsiden. Zool. Jahrb. Suppl. 15 (Festschrift fiir J. W. Spengel, Bd. 2):545-716. Versluys, J. (1936) Kranium und Visceralskelett der Sauropsiden. 1. Reptilien. Pp. 699-808. In: Handbuch der Vergleichenden Anatomie der Wirbeltiere. Vol. 4. L. Bolk, E. Goppert, E. Kallius, and W. Lubosch (eds.). Urban and Schwarzenberg, Berlin (1967 reprint, A. Asher and Co., Amsterdam). Vincent, J. (1990) Structural Biomaterials, rev. ed. Princeton Univ. Press, Princeton, NJ. Vitt, L. J. (1991) Ecology and life history of the wide-foraging lizard Kentropyx calcarata (Teiidae) in Amazonian Brazil. Can. J. Zool. 69: 791-799. Vitt, L. J., and D. G. Blackburn (1991) Ecology and life history of the viviparous lizard Mabuya bistriata (Scincidae) in the Brazilian Amazon. Copeia 1991:916-927. Vitt, L. J., and W. E. Cooper, Jr. (1986) Foraging and diet of a diurnal predator {Eumeces laticeps) feeding on hidden prey. J. Herp. 20: 408-415. Vitt, L. J., and S. de la Torre (1996) Guia para la Investigacion de Las Lagartijas de Cuyabeno (A Research Guide to the Lizards ofCuyabeno). Pontificia Universidad Catolica del Ecuador, Quito. Vitt, L. J., and C. Morato de Carvalho (1995) Niche partitioning in a tropical wet season: lizards in the Lavrado area of northern Brazil. Copeia 1995:305-329. Vitt, L. J., and P. A. Zani (1996) Ecology of the elusive tropical lizard
Tropidurus [^Uracentron] flaviceps (Tropiduridae) in lowland rain forest of Ecuador. Herpetologica 52:121-132. Vitt, L. J., R. C. van Loben Sets, and R. D. Ohmart (1981) Ecological relationships among arboreal desert lizards. Ecology 62:398-410. Vitt, L. J., T. C. S. Avila-Pires, and P A. Zani (1996a) Observations on the ecology of the rare Amazonian lizard, Enyalius leechii (Polychrotidae). Herp. Nat. Hist. 4:77-82. Vitt, L. J., P A. Zani, and J. P Caldwell (1996b) Behavioural ecology of Tropidurus hispidus on isolated rock outcrops in Amazonia. J. Trop.Ecol. 12:81-101. Vitt, L. J., P. A. Zani, J. P. Caldwell, M. Carmozina de Araujo, and W. E. Magnusson (1997a) Ecology of whiptail lizards {Cnemidophoru) in the Amazon region of Brazil. Copeia 1997:745-757. Vitt, L. J., R A. Zani, and T. C. S. Avila-Pires (1997b) Ecology of the arboreal tropidurid lizard Tropidurus (=Plica) umbra in the Amazon region. Can. J. Zool. 7b: 1876-1882. Vitt, L. J., P A. Zani, and A. C. Marinho Lima (1997c) Heliotherms in tropical rain forest: the ecology of Kentropys calcarata (Teiidae) and Mabuya nigropunctata (Scincidae) in the Curua-Una of Brazil. J. Trop.Ecol. 13:199-220. von Geldem, C. E. (1919) Mechanism in the production of the throatfan in the Florida chameleon, Anolis carolinensis. Proc. Calif. Acad. Sci. 9:313-329. von Seiller, R. F. (1891) Ueber die Zungendriisen von Anguis, Pseudopus und Lacerta. Arch. Microsk. Anat. 38:177-264. von Seiller, R. R (1892) Die Zungendrusen von Lacerta. Pp. 250-258. In: Festschrift zum Siebenzigsten Geburtstage Rudolf Leuckarts. Vrcibradic, D., and C. F. D. Rocha (1998) The ecology of the skink Mabuya frenata in an area of rock outcrops in southeastern Brazil. J. Herp. 32:229-237. Wagner, G. P., and K. Schwenk (2000) Evolutionarily stable configurations: functional integration and the evolution of phenotypic stability. Pp. 155-217. In: Evolutionary Biology, Vol. 31. M. K. Hecht, R. J. Maclntyre, and M. T. Clegg (eds.). Kluwer Academic/ Plenum Publishers, New York. Wainwright, P. C , and A. F. Bennett (1992a) The mechanism of tongue projection in chameleons. I. Electromyographic tests of functional hypotheses. J. Exp. Biol. 168:1-21. Wainwright, P C , and A. R Bennett (1992b) The mechanism of tongue projection in chameleons. II. Role of shape change in a muscular hydrostat. J. Exp. Biol. 168:23-40. Wainwright, P C , D. M. Kraklau, and A. R Bennett (1991) Kinematics of tongue projection in Chamaeleo oustaleti. J. Exp. Biol. 159: 109-133. Wainwright, S. A., W. D. Biggs, J. D. Currey, and J. M. Gosline (1976) Mechanical Design in Organisms. Edward Arnold, London. Wake, M. H. (1993) The skull as a locomotor organ. Pp. 197-240. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). Univ. of Chicago Press, Chicago. Waldschmidt, S. R., S. M. Jones, and W. P Porter (1987) Reptilia. Pp. 553-619. In: Animal Energetics, Vol. 2. T. J. Pandian, and R J. Vernberg (eds.) Academic Press, San Diego. Walls, G. L. (1942) The Vertebrate Eye and Its Adaptive Radiation. Bull. Cranbrook Inst. Sci. No. 19. Bloomfield Hills, Michigan. Walls, G. L. (1961) The evolutionary history of eye movements. Vision Res. 2:69-80. Walls, G. Y. (1981) Feeding ecology of the tuatara (Sphenodon punctatus) on Stephens Island, Cook Strait. N. Z. J. Ecol. 4:89-97. Watkins, G. G. (1997) Inter-sexual signalling and the functions of female coloration in the tropidurid lizard Microlophus occipitalis. Anim. Behav. 53:843-852. Watrous, L. E., and Q. D. Wheeler (1981) The outgroup comparison method of character analysis. Syst. Zool. 30:1-11. Webb, G. R. (1949) Notes on the coition and feeding of the blue-tailed skink. Copeia 1949:294.
8. F e e d i n g in L e p i d o s a u r s Webb, J. K., and R. Shine (1992) To find an ant: trail-following in Australian blindsnakes (Typhlopidae). Anim. Behav. 43:941-948. Webb, M. (1951) The cranial anatomy of the South African geckoes Palmatogecko rangei (Andersson) and Oedura Karroica (Hewitt). Ann. Univ. Stellenbosch27A: 131-165. Wellborn, V. (1933) Vergleichende osteologishe Untersuchungen an Geckoniden, Eublephariden und Uroplatiden. Sber. Ges. Naturf. Freunde Berlin 1933:126-199 (1997 English translation by A. P. Russell, A. M. Bauer, and A. Deufel, Herp. Trans. No. 1, Breck Bartholomew, Bibliomania, Logan, Utah). Werner, G. (1962) Das Cranium der Briickenechse, Sphenodon punctatus Gray, von 58 mm Gesamtlange. Zeit. Anat. Entwickl. 123: 323-368. Werner, Y. L., S. Okada, H. Ota, G. Perry, and S. Tokunaga (1997) Varied and fluctuating foraging modes in nocturnal lizards of the family Gekkonidae. Asiatic Herp. Res. 7:153-165. Wettstein, O. (1931) Rhynchocephalia. Pp. 1-128. In: Handbuch der Zoologie, Vol. 7, part 1. W. Kiikenthal and T. Krumbach (eds.). de Gruyter, Berlin. Wever, E. G. (1978) The Reptile Ear. Its Structure and Function. Princeton Univ. Press, Princeton, NJ. Whitaker, A. H. (1968) The lizards of the Poor Knights Islands, New Zealand. N. Z. J. Sci. 11:623-651. Whitaker, A. H. (1987) The roles of lizards in New Zealand plant reproductive strategies. N. Z. J. Bot. 25:315-328. White, L. R., R. Powell, J. S. Parmerlee, Jr., A. Lathrop, and D. D. Smith (1992) Food habits of three syntopic reptiles from the Barahona Peninsula, Hispaniola. J. Herp. 26:518-520. Whitford, W. G., and M. Bryant (1979) Behavior of a predator and its prey: the homed lizard {Phrynosoma cornutum) and harvester ants {Pogonomyrmex spp.). Ecology 60:686-694. Whiting, M. J., and J. M. Greeff (1997) Facultative frugivory in the cape flat lizard, Platysaurus capensis (Sauria: Cordylidae). Copeia 1997:811-818. Whyte, L. L. (1965) Internal Factors in Evolution. George Braziller, New York.
291
Willard, W. A. (1915) The cranial nerves of Anolis carolinensis. Bull. Mus. Comp. Zool. 59:15-116. Willson, M. F , C. Sabag, J. Figueroa, J. J. Armesto, and M. Caviedes (1996) Seed dispersal by lizards in Chilean rainforest. Rev. Chilena Hist. Nat. 69:339-342. Wilson, K. J., and A. K. Lee (1974) Energy expenditure of a large herbivorous lizard. Copeia 1974:338-348. Wineski, L., and C. Gans (1984) Morphological basis of the feeding mechanics in the shingle-back lizard Trachydosaurus rugosus (Scincidae, Reptilia). J. Morph. 181:271-295. Wood, C. B., E. R. Dumont, and A. W. Crompton (1999) New studies of enamel microstructure in Mesozoic mammals: a review of enamel prisms as a mammalian synapomorphy. J. Mamm. Evol. 6: 177-213 Yoimg, B. A., D. K. Magon, and G. E. Goslow, Jr. (1990) Lengthtension and histochemical properties of select shoulder muscles of the savannah monitor lizard (Varanus exanthematicus): implications for function and evolution. J. Exp. Zool. 256:63-74. Zalusky, S. B., A. J. Gaudin, and J. R. Swanson (1980) A comparative study of cranial osteology in the North American sand lizards, genus Uma (Reptilia: Iguanidae). Copeia 1980:296-310. Zangerl, R. (1944) Contribution to the osteology of the skull of the Amphisbaenidae. Am. Midi. Nat. 31:417-454. Zanno, C. (1974) Osservazioni su cellule pigmentate nella lingua de Eacerta podargis sicula. Anat. Anz. 136:286-291. Zavattari, E. (1909) Ricerche suUa muscolatura della lingua dei Geconidi. Atti R. Accad. Sci. Torino 44:282-290. Zippel, K. C , R. E. Glor, and J. E. A. Bertram (1999) On caudal prehensility and phylogenetic constraint in lizards: the influence of ancestral anatomy on function in Corucia and Furcifer. J. Morph. 239:143-155. Znari, M., and K. A. Nagy (1997) Field metabolic rate and water flux in free-living Bibron's agama {Agama impalearis, Boettger, 1874) in Morocco. Herpetologica 53:81-88. Zoond, A. (1933) The mechanism of projection of the chameleon's tongue. J. Exp. Biol. 10:174-185.
This Page Intentionally Left Blank
C H A P T E R
9 Feeding in Snakes DAVID CUNDALL Department of Biological Sciences Lehigh University Bethlehem, Pennsylvania 18015
HARRY W. GREENE Department of Ecology and Evolutionary Biology Cornell University Ithaca, New York
reptiles only Serpentes has achieved substantial adaptive radiation and high species richness. The more than 2500 species of living snakes inhabit most temperate and tropical land masses, and they often are prominent predators in terrestrial, arboreal, fossorial, aquatic, and even marine faunas. Snakes eat prey as different as onycophorans, fish eggs, centipedes, cormorants, and porcupines; many species commonly consume individual items weighing 20% of their own mass, and some venomous species occasionally subdue and eat prey that exceed their own mass by as much as 50% (Greene, 1984, 1992). Such extraordinary feeding biology has important implications for ecological energetics and life history (e.g., Secor and Nagy, 1994), and snakes are increasingly prominent as case studies in functional and evolutionary morphology (e.g.. Cans, 1961; Frazzetta, 1970a; Cundall, 1987; Kardong and Lavin-Murcio, 1993). This chapter first briefly surveys snake diversity and then examines in some detail the functional and morphological aspects of capturing, swallowing, and processing prey that generally characterize relatively derived subgroups; it only touches on sensory aspects of feeding, but for a review see Ford and Burghardt (1993). We consider briefly how prey size, prey shape, and other handling characteristics are related to evolutionary diversification of form and function in these animals. Finally, we use outgroup analysis to reconstruct major historical transformations in the feeding biology of snakes, focusing on novel components in the feeding mechanism that characterize major clades within the group as well as subsequent modifications.
I. INTRODUCTION A. Phylogenetic Synopsis of The Snake Feeding Apparatus B. Feeding Terminology as It Applies to Snakes IL FORM AND FUNCTION A. Ingestion-Prey Capture B. Prey Restraint C. Prey Manipulation D. Intraoral Transport E. Swallowing F. Digestion and Defecation III. PERFORMANCE AND SIZE IV. EVOLUTION A. Ecological Patterns B. Historical Patterns V. CONCLUDING REMARKS References
L INTRODUCTION Body elongation and limblessness have evolved numerous times within Tetrapoda, typically associated with aquatic, fossorial, crevice dwelling, or grassswimming lifestyles (Cans, 1986). Some lineages of secondarily elongate vertebrates (e.g., limbless skinks) have solved the concomitant problem of reduction in size of the feeding apparatus by eating many tiny items, whereas others (e.g., some caecilians) shear ingestible chunks out of large prey. Many advanced snakes achieved a third solution by radically restructuring their heads and feeding infrequently on large items; perhaps not coincidentally, among limbless squamate
FEEDING (K. SchwenKed.)
293
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
294
David Cundall and Harry W. Greene A. Phylogenetic Synopsis of the Snake Feeding Apparatus
Caldwell and Lee (1997, see also Lee and Caldwell, 1998; Lee, 1998) have argued that Pachyrachis, a marine mosasaur-like anguimorph from the Cretaceous of Israel, is the sister taxon of Serpentes. They therefore concluded that the earliest snakes might have been marine rather than terrestrial. Regardless of whether their phylogenetic hypothesis is correct, Caldwell and Lee (1997) failed to clarify two points. First, either a marine or a terrestrial origin of Serpentes would entail two evolutionary shifts on their preferred tree (one gain and one loss versus two gains), and in this sense both scenarios are equally probable; whether terrestrial to marine and marine to terrestrial are equally likely transitions remains to be explored in a broader context (cf. Lee and Shine, 1998). Second, their view follows an apomorphy-based taxon definition, i.e., they defined "snakes" as all members of the clade identified by characters linking Pachyrachis and living snakes; alternatively, we could define "snakes" as a more inclusive clade by adding one or more lineages basal to Pachyrachis or as a less inclusive clade if some fossil is found whose characteristics place it closer than Pachyrachis along the branch leading to Serpentes. We think that compelling general arguments (de Queiroz and Gauthier, 1994) and widespread implicit usage favor a crown clade definition, namely Serpentes (informally "snakes" or "serpents") includes the common ancestor of living snakes and all of its extinct and extant descendants. Irrespective of the relationship of crown group snakes to Pachyrachis and other fossil taxa, their closest living relatives are other Anguimorpha; moreover, among living squamates, snakes are probably related most closely to varanids, including Lanthanotus (Estes et al, 1988; Forstner et al, 1995; Lee, 1998; Schwenk, 1988). Accordingly, our analysis of the evolution of structure, function, and behavioral ecology emphasizes living snakes and their immediate extant outgroups, namely monitor lizards (Varanidae), alligator lizards (Anguidae), and their close relatives. There is now widespread agreement that living snakes are divided into two major clades, Scolecophidia and Alethinophidia (Cadle, 1988; Rieppel, 1988; Kluge, 1991; Cundall et al, 1993; Pough et al, 1998; see Fig. 9.1), and that these clades represent different approaches to limbless existence. The three families of scolecophidians (about 320 described species) presumably coevolved with their prey (small soil invertebrates, primarily ants and termites) that they find by extensive chemosensory searching (e.g., Gehlbach et al, 1971; Webb and Shine, 1992,1993a,b; Gaulke, 1995). Despite
extraordinary similarity in external form, size, and general feeding habits, these families encompass two widely divergent feeding mechanisms based on profound differences in the feeding apparatus (Haas, 1973), reflected in Fig. 9.2. Typhlopids (five genera, >200 species) use rapid protraction and retraction of the maxillary teeth to rake prey into their mouths (Thomas, 1985; lordansky, 1997; Kley, 1998), whereas leptotyphlopids (two genera, 80 species) use some form of mandibularbased manipulation and transport of prey. Typhlopids, like most alethinophidian snakes, appear to ingest prey whole. Leptotyphlopids ingest some prey whole, squeeze out the abdominal contents, and discard the exoskeletons of other arthropod prey (Smith, 1957) or rub off the heads and consume only the abdomens of termites. Feeding in anomalepidids (four genera, 15 species) is undescribed but they are presumed to use both maxillae and mandibles and to eat prey similar in size, form, and number to those exploited by other scolecophidians. In all scolecophidians, the braincase appears relatively rigid and the snout is a bulbous shell of bone attached at its periphery to the rim of the braincase. In all, the mouth is subterminal, the quadrates angle anteriorly, and the mandibles are short. All have splint-like pterygoids, which, in anomalepidids and typhlopids, are surrounded by robust protractor muscles. The palatines are small, laterally directed bones that attach to a pedestal-shaped maxilla. From the maxilla extends a large maxillary retractor muscle in typhlopids that may represent part of the pterygoideus fused to the levator pterygoidei (Haas, 1973). Anomalepidids retain separate levator pterygoidei and pterygoideus muscles, the latter having a dorsal head that attaches the maxilla to the otic region dorsal to the quadrate and probably functions like the retractor maxillae of typhlopids. A large prefrontomandibular ligament (lordansky, 1997) limits both maxillary and mandibular movements in typhlopids (see later). In leptotyphlopids, the maxilla is relatively immobile and toothless but the mandible is short, robust, and toothed. Teeth are arranged transversely on the dentigerous bones (dentary for leptotyphlopids; maxilla for typhlopids; anomalepidids typically have so few teeth that their arrangement is equivocal). Alethinophidians ("true" or "typical" snakes; about 2200 species) apparently coevolved with a much broader range of mostly terrestrial prey (Greene, 1997). The most basal, and in many respects, most primitive living alethinophidians are Malaysian and Indonesian anomochilids (one genus, two species) and the specialized, burrowing Sri Lankan and Indian uropeltids (nine genera, >45 species). Both clades share with scolecophidians an edentulous palate but otherwise show
295
9. Feeding in Snakes Leptotyphlopidae
£
Scolecophidia
Anomalepididae Typhlopidae
Anomochilidae Uropeltidae Cylindrophiidae Aniliidae Xenopeltidae Ijoxocemidae
Pythonidae Alethinophidia
Boidae Tropidophiidae Bolyeriidae Acrochordidae
MacrostDmata
Viperidae Atractaspididae El^idae
Colubroidea
Colubridae FIGURE 9.1. Cladogram of relationships among living snakes based on morphological (Rieppel, 1988; Kluge, 1991; Cundall et al, 1993) and molecular analyses (Cadle, 1994; Cadle et ah, 1990; Dessauer et ah, 1987; Heise et ah, 1995). This arrangement differs slightly from some other recent phytogenies (e.g., Pough et ah, 1998) in treating basal alethinophidians, as well as boids and pythonids, as separate lineages.
alethinophidian affinities (Rieppel, 1977b; Cundall and Rossman, 1993; Cundall et al, 1993). Their snouts are anchored to the middle of the braincase (not its rim), their maxillae have longitudinal rather than transverse tooth rows, and their mandibles have long dentaries (Fig. 9.2). More importantly, their upper jaws are anchored to their snouts by close attachments of the maxillae to the premaxilla and to the pterygoid (Fig. 9.3) and by tight connections between the palatine and the vomer. The latter feature is shared with southeast Asian Cylindrophis (Cundall, 1995) and the neotropical Anz7/ws (Rieppel, 1977a,b; 1978a). These latter two semifossorial clades have toothed palatines and pterygoids, and robust mandibles. Cylindrophis has a mobile intramandibular joint but a relatively inelastic intermandibular connection, features thought to be shared by Anilius, Anomochilus, and uropeltids. Whereas uropeltids (Wall,
1921; Rajendran, 1985), and we suspect Anomochilus, eat primarily earthworms or other small soil invertebrates, the larger Cylindrophis and Anilius take elongate vertebrate prey, which are sometimes constricted (Greene, 1983). We treat anomochilids, uropeltids, cylindrophiids, and aniliids as independent lineages (Cundall et ah, 1993; Fig. 9.1); a more conservative arrangement unites all Asian taxa in the Uropeltidae (Pough e^ a/., 1998). Two relict taxa, the southern Asian Xenopeltis (two species) and the middle American Loxocemus bicolor, are currently thought to be basal to all the remaining alethinophidians (Kluge, 1991; Cundall et al, 1993). Xenopeltis eats lizards and small mammals (Savitzky, 1981; Greene, unpublished observation) and Loxocemus is an active forager exploiting lizards and their eggs and hatchlings as well as small mammals (Greene, 1983;
296
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
Anguimorpha Dibamidae Leptotyphlopidae Anomalepididae Typhlopidae Anomochilidae Uropeltidae Cylindrophiidae Aniliidae Xenopeltidae Loxocemidae Pythonidae Boinae Erycinae Bolyeriidae Tropidophiidae Acrochordidae Colubroidea
F I G U R E 9.2. Phylogenetic diversity of skull form. Supratemporal, quadrate, and mandible are in gray. Critical changes include the gradual rearward migration of the quadrate, enlargement and free extension of the supratemporal (pythonids and colubroids), and concomitant lengthening of the mandible. Taxa illustrated are, from the top, Dibamus novaeguineae (after Greer, 1985), Leptotyphlops macrolepis, Typhlops punctatus (both after Groombridge, in Parker and Grandison, 1977), Anomochilus weheri, Uropeltis ocellatus, Cylindrophis ruffus, Loxocemus bicolor (last after McDowell, 1975), Python molurus, Acrochordus javanicus, Dispholidus typus (Colubridae), and Bitis arietans (Viperidae).
Mora, 1991). Hence, these species appear to be the most primitive living snakes to consume relatively bulky prey, although both also appear to be gape limited. Reports on their basic skull structure, myology, and tooth form (Haas, 1955; McDowell, 1975; Savitzky, 1981, 1983; Zaher, 1994b; Frazzetta, 1999) have yet to be augmented by careful functional-morphological studies. Although considered by some (Rieppel, 1988; Pough et al, 1998) to be macrostomate (large-mouthed) snakes, known functional and structural traits suggest that they are basal to Macrostomata. We therefore regard their inclusion within Macrostomata as equivocal and choose to exclude them until they are known to possess critical behavioral (and presumably structural) macrostomatan apomorphies associated with wide spreading of the anterior tips of the mandibles. All remaining alethinophidian snakes are contained within the monophyletic clade Macrostomata. Synapo-
morphies characterizing Macrostomata all correlate with increased gape size (Rieppel, 1988) and include increased length of the mandible and suspensorium (Fig. 9.2) and modifications of the intermandibular soft tissues to allow stretching (Groombridge, 1979a; Bellairs, 1984; Young, 1998). Boas and sand boas (Boidae: two subfamilies, seven genera, 50 species) and pythons (Pythonidae: eight genera, 27 species) display radical innovations in the feeding apparatus while retaining a host of plesiomorphic features (Frazzetta, 1966; Rieppel, 1988; Kluge, 1991,1993a,b). This bizarre mosaic of traits leaves their relationship to other snakes open to question, although the structure and behavior of their feeding apparatus qualify most of them as macrostomatan. Modifications of the trunk correlate with increased speed and distance of strikes; combined with the enlargement of the more caudal axial musculoskeletal components, they make possible the ambush
297
9. F e e d i n g in Snakes
po /^^zs
-SZSZS/
F I G U R E 9.3. Upper jaw form and pterygoideus muscle origins in four functional types in dorsal view (anterior toward the top), (a) Cylindrophis rujfus, a basal-constricting alethinophidian with palatal teeth; (b) Python sp., a basal macrostomate; (c) Agkistrodon piscivorus, a venomous viperid colubroid; and (d) Nerodia sp., a nonvenomous colubrid colubroid. Ligamentous connections to the braincase (po, postorbitomaxillary ligament, bp, basipterygoid ligament), snout (pm, premaxillomaxillary ligament, vp, vomeropalatine ligament), and quadrate (qp, quadratopterygoid ligament, variably present in both basal and derived clades) in basal taxa are shorter and tighter. Separation of the pterygoideus into principal (pg) and accessory (pg,a) heads and migration of the origin of the principal head to the distal end of the ectopterygoid in colubroids (c,d) changed the mechanics of the upper jaw radically (see transport). Dotted lines (pf) show the approximate axis of prefrontal bone articulation with the palatomaxillary apparatus, and arrows denote possible directions of morphological transformation.
298
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
capture and killing by constriction of relatively large prey (Greene, 1983; Slip and Shine, 1988; Rodriguez-Roh\esetalA999). Bold and pythonid heads display a number of innovations associated with feeding. Specialized trigeminal innervation of the labial skin provides infrared reception that gains directional acuity in some taxa through invagination of receptor regions (Barrett et al, 1970; Maderson, 1970; de Cock Buning, 1983). The skull is marked by loss of contact between the palatine and the vomer and elongation of the supratemporal to carry the quadrate behind the skull in most derived members of these clades (Rieppel, 1978b; Tokar, 1989; Kluge, 1991, 1993a,b). The mandible is relatively long, although the increase in length occurs on both sides of the angle of the mouth. Briefly, this means that the snout is relatively longer than in more basal alethinophidians, creating an increase in preorbital jaw length, and increased suspensorial length increases the length of the region of the mandible lying behind the orbit (Fig. 9.2). Most importantly, the soft tissues between the mandibles are reorganized to permit wide separation of the mandibular tips (Fig. 9.4). Extraordinary gape diameters are seen first in boids and pythonids and have only rarely been equaled in more derived clades. The West Indian and Central American tropidophiids [two to four genera, 18-21 species, although some evidence supports inclusion of Exiliboa and Ungaliophis in the Boidae (Zaher, 1994a; Pough et al, 1998)] and two species of bolyeriids now restricted to Round Island off the coast of Mauritius appear to be boa-like relicts based on bones and muscles of the head. Trunk anatomy in both cases is more equivocal (Gasc, 1981). Bolyeriid genera share a unique intramaxillary joint (Frazzetta, 1970b), which in Casarea allows the anterior upper jaw to bend ventrally, a feat that may aid in trapping skinks in their mouths (Cundall and Irish, 1989). Three species of Australasian filesnakes (Acrochordidae) comprise the living sister group of all advanced snakes (Colubroidea) and are generally considered an aberrant clade of highly specialized piscivores (Savitzky, 1983; Shine, 1986). Unlike more basal snakes, but like many colubrids, they have long quadrates (Fig. 9.2) that are angled caudally and, as a result, mandibles that are much longer than their braincase (Hoffstetter and Gayrard, 1965). They exhibit a variety of anatomical peculiarities (e.g., tubercular scales, plate-like supratemporal bone, potentially functional mesokinetic joint, absence of contact between snout bones and braincase) whose homologies remain obscure (McDowell, 1979, 1987; Pough efd., 1998). The vast majority of snakes, in terms of diversity of species and numbers of individuals in n\ost communities, are colubroids. Phylogenetic relationships among
a
F I G U R E 9.4. Simplified model of intermandibular muscles in basal alethinophidians and pythonids (a-d) and in boids and all other macrostomatans (e and f) in ventral view. The longitudinal arrangement of anterior intermandibular muscles (e.g., ima, intermandibularis anterior; tb, transversus branchialis) potentially allows spreading of the anterior tips of the mandibles (Md) through deformations suggested in (c) and (d). The posterior intermandibular muscle (imp) lies dorsal to the ceratomandibularis (cer, part of the neurocostomandibularis of colubroids; Langebartel, 1968; Groombridge, 1979a). Gape size in basal alethinophidians is apparently constrained by limited elasticity of the interramal pad [irp, containing elements described by Young (1998) as lateral and perpendicular bands and including, in some taxa, Meckel's and symphyseal cartilages] and possibly by attachments of the skin to the mandibles and interramal pad, not by muscle arrangements. Rearrangement of the anterior intermandibular muscles in boids and other macrostomatans ties mandibular separation more closely to the stretch capabilities of the intermandibularis anterior, pars anterior (imaa), and pars posterior (imap), as well as the passive stretch potential of intermandibular connective tissues. Posterior intermandibular muscles in many colubroids lie dorsal and ventral to the neurocostomandibularis, but match the position shown in (c) and (d). els, constrictor of lateral sublingual gland; Is, lateral sublingual gland.
colubroids are incompletely resolved and controversial (e.g., Cadle, 1988,1994; Knight and Mindell, 1993; Heise et al, 1995; Dowling et al, 1996; Kraus and Brown, 1998).
299
9. Feeding in Snakes Molecular and some morphological evidence implies that venomous snakes, in particular viperids, are basal to all other colubroids (Fig. 9.1). Colubroidea includes all venomous snakes of the world. Early in the evolution of colubroids, serous cells of a posterior oral (dental) gland acquired the capacity to secrete toxins as well as enzymes (Kochva, 1987; Kardong, 1996). Ducts from this gland(s) drained into dental clefts in the oral mucosa and the evolution of envenomation began (Savitzky, 1980; Kardong, 1979, 1981, 1982a; McDowell, 1986; Underwood, 1997). However, the structural and functional properties of the basic colubroid jaw apparatus had enormous evolutionary potential for exploiting diverse prey even in the absence of envenomation (Cundall, 1983). Whereas no living colubrids have tubular fangs, many have grooved (canaliculate) teeth, and in most of these, the grooved teeth are enlarged and lie at the rear end of the maxilla (Anthony, 1955; Jackson and Fritts, 1995, 1996; Young and Kardong, 1996). Viperidae (28 genera, about 230 species) is perhaps the sister group to the other colubroid clades, including atractaspidids (1-15 genera, 9-65 species depending on which features one chooses to stress), elapids (62 genera, about 270 species), and colubrids (about 300 genera, 1700 species) (McDowell, 1987; Underwood and Kochva, 1993; Zaher, 1994b; Cadle, 1994; Greene, 1997; Pough et ah, 1998). Most viperids, elapids, and Atractaspis are distinguished by maxillae in which anterior tooth generations are oriented longitudinally in the head rather than obliquely (i.e., new teeth arise directly posterior to the functional generation and migrate anteriorly; figuratively, this could be arrived at by bending the anterior maxilla medially about 45°). In viperids (Fig. 9.2) and Atractaspis, the maxilla is greatly shortened and hinged to the prefrontal. Its development does not reveal homologies, although telescoping of the ancestral maxillary field seems most likely (A. Savitzky, 1992). Shortening of the palatine and elongation of the ectopterygoid in viperids (Fig. 9.3) provide a lever system that allows extensive rotation of the maxilla around its prefrontal hinge, and reduction or loss of the anteromedial wing of the prefrontal may release this bone from structural constraints on rotation imposed by adjacent snout elements (A. Savitzky, 1992). In many elapids, conversely, the palatine is elongate, as in many colubrids, but the ectopterygoid has a much reduced distal fork or notch as compared to most colubrids (Lombard et al, 1986). The maxilla is quite variable within elapids but generally much less mobile than in viperids, and both its rostral and caudal ends curve medially (Fairley, 1929; Haas, 1973). Molecular analyses (Keogh, 1998) support McDowell's (1970) structural-functional division of elapids into basal palatine
erectors (New and Old World cobras and kraits) and derived palatine draggers (sea snakes and AustraloPapuan terrestrial elapids). Whereas the rotatable maxillae of viperids circumvents some constraints of fitting the fangs into the anterior head, the more elongate, less rotatable maxillae of elapids leaves their fangs closely tied to the height of the anterior head, hence, shorter (Bogert, 1943). Inclusive views of the family Atractaspididae (McDowell, 1986; Underwood and Kochva, 1993) make it surprisingly diverse in cephalic anatomy, ranging from possibly nonvenomous taxa {Aparallactus modestus) to those with grooved rear fangs (e.g., Amblyodipsas, Polemon, Xenocalamus), one with short, fixed tubular front fangs {Homoroselaps), and still others with long, movable tubular front fangs (Atractaspis). Along with this diversity in tooth morphology and arrangement are palatomaxillary linkage and muscle attachment patterns that cover the logical range of nnorphological intermediates between a basal macrostomatan apparatus and a derived colubroid condition (McDowell, 1986). Whether these morphological intermediates in any way reflect historical patterns remains unclear, as does the extent to which morphology reflects function. Colubrids encompass the remaining living snakes and although many appear similar in their basic anatomy, the musculoskeletal system varies widely (Haas, 1973). Despite efforts to resolve colubrid phylogeny (e.g., Cadle, 1984a,b,c; 1988; 1994; Heise et al, 1995; Dowling et ah, 1996), evolutionary patterns within the group remain controversial. Feeding mechanics have been described in detail for surprisingly few colubrid taxa (Cans, 1952; Albright and Nelson, 1959a,b; Cundall and Cans, 1979; Kardong, 1986b). Table 9.1 summarizes the distribution of empirical observations on feeding mechanics among snakes. Most of our information comes from a few colubrids, viperids, and pythonids. No functional details exist for atractaspidids or elapids and surprisingly few for boines or erycines. Among basal alethinophidians, only cylindrophiids have been studied in any depth, and among scolecophidians, only a few typhlopids and leptotyphlopids. B. Feeding Terminology as It A p p l i e s to Snakes Terms widely used for various phases of vertebrate feeding are mammalocentric and some can be applied to snakes only by merciless stretching of the original definition (see Chapter 2). Hence, we may refer to prey capture in snakes as ingestion, although in most cases the prey item does not come to lie entirely within the oral cavity. For most snakes, the food (prey) occupies a
300
D a v i d C u n d a l l a n d H a r r y W. G r e e n e TABLE 9.1
Phylogenetic Distribution of Functional Data o n Feeding a m o n g Snakes Ingestion''
Family
1
Leptotyphlopidae
X^
2
3
4
5
6 X'^
Transport''
7
8
9
10
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
Anomalepididae Typhlopidae
X
X
X
X
X
X
X
X
Pythonidae
X
X
X
X
X
Boidae
X
X
X
X
X
Anomochilidae Uropeltidae Cylindrophiidae Aniliidae Xenopeltidae Loxocemidae
Tropidophiidae Bolyeriidae Acrochordidae Viperidae
X
X
X
X
X
X
X
X
X
X
Atractaspididae Elapidae
X
Colubridae
X
X
X
X
X
''I, direct observation; 2, detailed mechanical analysis based on anatomy; 3, light cine records at slow framing rates (16-36 fps); 4, high-speed light cine records at 64-1000 fps; 5, video records at 30-60 fps; 6, high-speed video records, usually 120-250 fps; 7, radiographic records; 8, EMG records; 9, synchronized EMC and behavioral records; 10, displacement or pressure sensor records. ^X, published data. '^x, studies in progress with preliminary reports. ^Records of spitting, not actual ingestion, in spitting cobras.
point in space that becomes invaded by the snake's head. Invasion of the prey space by the snake's head may be a rapid event, although the process is usually keyed to the potential speed of movement of the prey. After attaching its head to the prey, some snakes use one of various possible restraint mechanisms to immobilize or kill the prey. Other species simply begin transporting the living prey item through the oral cavity. In many snake species that kill prey prior to transport, prey are often released. Release of prey may occur immediately after an envenomating bite but before prey death in venomous species, or shortly following death of the prey in most constricting species. Those snakes that release prey ''reingest" dead prey by simply opening their mouth, usually by bilaterally synchronous depression of the mandibles. "Reingestion" is often followed by a longer kinematic phase usually referred to as manipulation which serves to orient the prey for intraoral transport. We include prey manipulation as a category of intraoral transport and not as a type of ingestion because the kinematic patterns used in manipu-
lation appear to be modifications of unilateral transport kinematics. Once the jaws are attached to the prey item (ingestion), the process of transporting it through the oral cavity (intraoral transport) begins. Snakes envelop the whole prey with their heads. Because many snakes prey on animals larger than their own heads, this process is often prolonged and typically requires mechanisms very different from those employed in simply attaching the head to the prey in the first place. Furthermore, because the prey is often longer than the snake's head, it can occupy the environment, oral cavity, and esophagus simultaneously. Hence, transport and swallowing can occur simultaneously and result from the same set of kinematic events. In our account, we equate prey capture with ingestion and consider ingestion mechanics to be of two basic types: slow or fast. Slow systems involve movements of the head within the range of normal locomotor movements. Fast systems involve specialized rapid movements of the head atypical of all locomotor modes except slide-pushing (Gans, 1986). Both slow and fast
301
9. Feeding in Snakes systems have been commonly referred to as strikes, with slow systems sometimes being distinguished as lunges. We consider prey restraint as a type of food processing or reduction and include under restraint all mechanisms that snakes use to limit a prey's ability to escape transport and swallowing. Such mechanisms include using (a) jaws and teeth to damage prey tissues mechanically, (b) venoms to damage prey tissues chemically, and (c) constricting body coils to restrain prey locomotion or ventilation mechanically. A few snakes reduce prey before swallowing, and venomous snakes introduce both toxins and digestive enzymes into prey that achieve digestive goals similar to chewing in mammals. In snakes that do not release prey after ingestion, prey restraint is often combined with jaw movements that manipulate the prey to orient it for transport. Snakes that release a prey item after ingestion and restraint (strike followed by envenomation or constriction) must "reingest" and, often, manipulate a dead or dying prey item. Intraoral transport in snakes has been generally regarded as a specialized product of a highly kinetic jaw apparatus that allows independent anteroposterior movements of the right and left jaws. It has been clear for some time, however, that this simplification hides critical evolutionary transformations in both the form and the function of the head. We recognize three fundamentally different transport modes in snakes: one involving use of the lateral jaw elements only, one employing both lateral and medial jaw elements, and a third dependent primarily on medial upper jaw elements. Whether these categories deserve formal distinction or not, it is hoped that their use here will encourage broader investigation of the problem. Finally, swallowing has received little attention in snakes because the process of transporting prey through the oral cavity often achieves much of swallowing as well. The notion that distinct swallowing events occur and influence transport has suggested reinterpretations of feeding mechanics and energetics.
IL FORM A N D FUNCTION Most of what we know about how snakes eat is based on species that are relatively large, easily maintained in captivity, and readily recorded. The vast majority of living species do not fit these criteria and therefore the function of their feeding apparatus has never been studied. However, the anatomy of the head, particularly the morphology of the skull (no comprehensive review, although one is now in progress; Cundall and Irish, manuscript in preparation) and some
groups of head muscles (Haas, 1973), is known for representatives of most major clades. Comparing this information with what we know of diet in snakes (e.g.. Shine, 1991c; Greene, 1997) suggests that prey diversity may loosely reflect the potential range of basic structural-functional systems in the clade. A. Ingestion-Prey Capture Boas and pythons catch prey by lightning fast strikes, followed by rapid coiling and constriction. Many vipers ambush prey detected by visual or thermal cues, then release the prey after injection of venom that immobilizes and kills it with varying speed (Hayes and Duvall, 1991). Both of these capture strategies depend on remote sensing of prey through light (visual or infrared) or vibration receptors (de Cock Buning, 1983; Hayes and Duvall, 1991). Other snakes appear to catch prey encountered during foraging movements and identified by vomerolfactory messages (e.g., Burghardt, 1969; Teather, 1991; Schwenk, 1994). In these cases, capture is achieved by crawling up to the prey and either lunging the remaining distance or simply opening the mouth as the prey is reached. Most detailed analyses of prey capture have concentrated on striking. A few species combine luring—usually caudal luring—with striking (Pough, 1988), and unrelated aquatic snakes use novel but convergent lateral sweeping movements to catch free-swimming fish (Radcliffe and Chiszar, 1980; B. Savitzky, 1992; Braun and Cundall, 1995; Alfaro, 1998; Mori, 1998). Prey capable of inflicting damaging or fatal wounds to snakes are usually taken with highly specialized, stereotypic behaviors, many of which are documented by Greene (1997). There are few detailed descriptions of snakes catching prey in the field. Hence natural foraging behavior, in the sense of the behavior leading up to and including prey capture, is poorly known (cf. Arnold, 1993). Most kinematic details of prey capture stem from laboratory observations of captive animals, fed, in many cases, prey they rarely or never eat in the field. In the field, snakes have been seen in positions interpreted as ambush or foraging postures (e.g., Greene and Santana, 1983; Reinert et al, 1984; Slip and Shine, 1988; Greene, 1992) but actual prey capture is rarely witnessed or, when seen, rarely recorded in a manner allowing kinematic analysis. Much of the foraging literature actually describes gut contents gained from preserved specimens (e.g., Godley, 1980; Mushinsky et al, 1982; Greene, 1983; Shine, 1986) or from forcing captured snakes to regurgitate their last meal (e.g., Arnold, 1992; Greene, 1997), not the process by which the contents of the gut came to be there. For the relatively few snakes recorded catching prey in the field—such as
302
David Cundall and Harry W. Greene
Epicrates angulifer (Hardy, 1957) and E. inornatus (Rodriguez and Reagan, 1984) catching bats; some Thamnophis (Drummond, 1983), Nerodia (Evans, 1942; Brow^n, 1958), Agkistrodon (Bothner, 1974; B. Savitzky, 1992), and various marine snakes (Kropach, 1975; Voris et ah, 1978; Heatwole, 1987) catching fish—details of the capture process are invariably limited. Even accurately documenting prey selection in snakes is more challenging than previously recognized (e.g., Arnold, 1992; Rodriguez-Robles, 1998; Shine et al, 1998). The low frequency of feeding events in many of the species most easily radiotracked (Reinert et al, 1984; Greene and Santana, 1983; Slip and Shine, 1988; Shine, 1991c; Greene, 1997) and the equipment requirements for gaining good behavioral records have discouraged field studies of feeding. Hence, we know little about how most snakes find or catch prey except by inference and extrapolations from the behavior of captive snakes. Given the limits of the morphological apparatus available for prey capture, it is not surprising that snakes catch prey by simply thrusting their opening mouths at the prey. The thrust is largely a function of the anterior trunk. Its execution, however, is coordinated with movements of the jaws, and in some species, this coordination occurs in very short time intervals over appreciable distances. In other species, these time intervals seem longer and the distance travelled is shorter. In other words, snakes use both short and slow lunges or long and fast strikes for catching prey. The extremes of these two patterns correlate loosely with phylogeny and with the nature of prey selected. We postulate that slow patterns are plesiomorphic for snakes and occur in most basal clades of alethinophidians and in all scolecophidians. Many macrostomates also use slow patterns which may be behavioral apomorphies associated with particular kinds of prey. These behaviors are matched in some cases to structural modifications that converge on morphological features of basal clades. Because slow patterns are apparently uninteresting to functional morphologists, they have never been described in any detail. We treat them here because they exist and because we think that the evolution of prey capture in snakes cannot be understood without establishing the plesiomorphic condition. 1. Slow
Systems
As we view them, species characterized by slow systems propel their heads short distances at relatively slow speed. However, movements of the jaws may approximate in speed jaw movements seen in fast systems. At a superficial level, the differences are easily seen with the naked eye and underlie use of the term lunge for the slower approach to prey. A key difference
may be that slow ingestion lacks a definable precapture posture. Basal alethinophidians ingest prey with short lunges but relatively rapid jaw movements. Cundall (1995) claimed that Cylindrophis achieved gapes up to 90° during a short lunge, but this figure reflects measurements made at the edge of mouth, not at the quadrate mandibular joint. Maximum angular rotation at the quadrate mandibular joint during lunging in Cylindrophis is closer to 45-55°, and movements of the upper jaw are minimal. We assume that anomochilids, uropeltids, and aniliids behave much like cylindrophiids and use relatively simple opening and closing of the lower jaws during prey capture. Film records of prey capture in Cylindrophis show only slight elevation of the snout and limited lateral displacement of the maxillae. Protraction of the upper jaws is limited by short premaxillarymaxillary ligaments, and mobility of the snout is limited by relatively tight attachment of the nasals to the frontals (Rieppel, 1977a, 1978a). Movement of the ventral snout elements relative to the nasals in basal alethinophidians (snout shifting of Cundall, 1995; see later under lateromedial transport) provides the basis for the limited movements of the upper jaws. In\mobile or slowly moving prey require little speed to be successfully captured, and the macrostomate species that feed on them appear to ingest them in much the same manner as does Cylindrophis. Finding eggs of fishes and amphibians must require extensive foraging, but eating them would appear to require few modifications. However, the relatively sniall colubrid snakes in the genus Dasypeltis show extraordinary specializations for eating bird eggs (Gans, 1952), including a slow ingestive technique that consists of slowly pushing the jaws over the egg with the aid of unilateral jaw motions that grade immediately into those used for transporting the egg through the oral cavity. Structural modifications in Dasypeltis all focus on increasing gape size and retaining control over the ventral skin despite its distance from the tips of the mandibles as ingestion and transport progress. A number of aquatic species in unrelated clades (various hydrophiine clapids, Kropach, 1974; Voris et al, 1978; Heatwole, 1987; homalopsire colubrids, Jayne et ah, 1988; some species of the natricine colubrid genera Nerodia and Thatnnophis, Drummond, 1983; Halloy and Burghardt, 1990; Alfaro, 1998; viperid Agkistrodon; B. Savitzky, 1992) catch free-swimming fish by holding the mouth open and sweeping the head sideways through the water. The sweep typically falls within the speed of normal swimming movements. This strategy works remarkably well when used in schools of fish or in drying ponds with high fish densities. Gape size is independent of fish size within a range of fish sizes, suggesting
9. F e e d i n g in Snakes
that the posture used for open-mouthed sweeping might minimize drag (Braun and Cundall, 1995). Tests on paraffin models of different head shapes and postures showed that holding the mouth open to the degree used in fishing produced drags that were not only lower than lunging forward with the mouth open (not surprising) but also lower than lateral sweeping with the mouth closed (Braun, personal communication). Some smaller aquatic piscivorus snakes (small Thamnophis, Nerodia, Alfaro, 1998; Erpeton, Smith et al, 1998) appear to strike forward in water, and in Erpeton, water may collect in an expanded buccal cavity. Among small colubrid species that may have secondarily acquired slow capture behavior are the many species of Atractus. These species all appear to eat earthworms, presumably in the same manner, but show extraordinary variation in the cervicomandibularis muscle (Fig. 9.5) with an as yet unexplained functional relevance (F. J. Irish, personal communication).
303
We assume that many other colubrid snakes use slow capture techniques on both elusive and nonelusive prey. Because species that use fast capture techniques may use slow capture facultatively for particular kinds of prey or in particular environments, it is of interest to determine which species are structurally constrained to use slow capture techniques and to test empirically how structural features act to limit speed of capture. Scolecophidian snakes do not fit neatly into prey capture categories defined by specializations of the trunk. In typhlopids, for example, ingestion involves rapid acquisition of many small prey. The actual process of ingestion observed visually (Thomas, 1985; R. Thomas, T. J. Gush, and A. Diaz, personal communication) and recorded with videofluorography and both regularand high-speed video (Kley, 1998; personal communication) appears to result from the same maxillary motions used to transport prey. During maxillary motion, prey items located at the edge of the mouth sometimes
FIGURE 9.5. Evolutionary modification of the cervicomandibularis muscle (cm) in four fossorial, wormeating xenodontine colubrid species: (a) Geophis blanchardi, (b) Geophis ruthveni, (c) Atractus elaps, and (d) Atractus duboisi. Id, longissimus dorsi; ncm, neurocostomandibularis vertebral head; rcb, retractor costae biceps (iliocostalis); ssp, spinalis-semispinalis; te, tendon of insertion of the cervicomandibularis. Illustration by R J. Irish.
304
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
show irregular, jerky motions and then suddenly move into and through the oral cavity. As a prey item passes into the mouth the head is moved slightly until the edge of the mouth contacts another prey item. Maxillary motion may cease between prey items or continue if prey density is high (Kley, unpublished data). Detailed models of jaw mechanics in typhlopids (lordansky, 1981, 1997) suggest that the large prefrontomandibular ligament (Fig. 9.6) not only limits depression of the quadrate mandibular joint but, when stretched by maxillary protraction, also depresses the Maea Moem
mandible. lordansky (1997) suggested that typhlopids protract the upper jaw after depressing the lower jaw, performing cyclical movements of both upper and lower jaws. However, Kley's data show that typhlopids may simply leave the lower jaw partially depressed and use rapid unilateral or bilateral fluttering of the maxillae to both ingest and transport small prey, or use synchronized mandibular and maxillary movements to ingest larger prey. As the maxillae are protracted the mandibles are adducted, possibly to prevent the maxillae from sweeping the prey out of the mouth. Typhlops platycephalus can ingest hundreds of prey in 10-min periods and numerous prey per second (R. Thomas, T. Gush, and A. Diaz, personal communication). Kley's high-speed video records suggest slightly alternating maxillary movements (maxillae are protracted and retracted slightly out of phase) in which maxillary teeth are swept from lateral to medial as well as rostral to caudal. The speed and duration of jaw movements imply structural and physiological properties of the enlarged pterygoid protractor and maxillary retractor (Haas, 1930; lordansky, 1997) potentially different from most head muscles of snakes. In fact, there is little about typhlopid feeding that is similar to that of any other snake clade. Ingestion in leptotyphlopids and anomalepidids is assumed to differ from that of typhlopids (Smith, 1957; Greene, 1997; personal observation). Leptotyphlops uses rapid retracting and ratcheting movements of the dentaries to haul prey into the oral cavity; dentary movements suggest a complex recruitment of geniohyoideus, mandibular adductor, and intermandibular muscles (Kley, 1998; personal communication). Ingestion may also involve mucosal deformations produced by muscle X, a slip of the geniohyoideus attaching to the dorsal and lateral surfaces of the caudal, pouch-like oral mucosa (Groombridge, 1979b). 2. Fast Systems
FIGURE 9.6. Muscles and ligaments in the head of Typhlops lumbricalis shown after removal of (a) the skin and subcutaneous muscles; (b) prefrontomandibular (jugomandibular) ligament (Ijm), depressor mandibulae (Mdm), intermandibularis posterior (Mimp), and constrictor colli (Mcc); (c) adductor extemus medius (Maem), adductor externus posterior (Maep), intermandibularis anterior (Mima), cervicoquadratus (Mcq), and all subhyoidean muscles; and (d) lower jaw, all mandibular adductors, and all trunk muscles. Maea, adductor externus anterior (superficialis); Mpp, protractor pterygoidei; Mps, pseudotemporalis; Mpta, pterygoideus anterior; Mrm, retractor maxillae. From lordansky (1997), Folium Publishing Co., Moscow, Russia, with permission.
Larger macrostomate snakes that exploit endothermic prey use movements of the trunk that are not only much faster than normal locomotor movements, but that are often preceded by complex preparatory postural changes (Frazzetta, 1966; Kardong, 1982b; Kardong and Bels, 1998). The final capture movements are commonly referred to as strikes, a term unfortunately used for many different kinds of behaviors, including prey capture in lizards (e.g., O'Cormell and Formanowicz, 1998). Snake strikes are often discussed under two assumptions—that they are the most common form of prey capture in snakes and that they are the plesiomorphic mode of prey capture for the clade. We think both assumptions are wrong.
9. F e e d i n g in Snakes
The fastest capture strategies in snakes are found among boids, pythonids, viperids, and some colubrids. But how fast is fast? Van Riper (1954) found the velocity of the head of Crotalus viridis in midstrike to vary from 1.6 to 3.5 m-sec~^ and Greenwald (1974) measured similarly variable maximum velocities [1.22 (at 18°C)2.85 m-sec~^] in Pituophis. These seem both highly variable and not very fast until one reflects that, at the higher velocity, a snake would strike a foot in a tenth of a second. These velocities for a viperid and a constricting colubrid are an order of magnitude faster than velocities recorded for thamnophiine strikes at fish (0.03-0.3 m-sec^^: Alfaro, 1998) and, despite differences in size, support our distinction of slow and fast systems. Two kinds of strikes have been defined in the literature, called stabbing and biting strikes (Van Riper, 1950, 1953, 1955; Klauber, 1956) or, more recently, defensive and predatory strikes (Kardong, 1986a). In defensive strikes, snakes typically use a large gape achieved by flexing the braincase dorsally on the neck as well as depressing the mandibles (Kardong, 1986a). If the jaws actually hit the snake's aggressor, they are often withdrawn without biting. In predatory strikes of booids and viperids, gape is usually smaller (35-80°), the mandible usually contacts the prey first, and the strike ends in a bite (Frazzetta, 1966; Kardong, 1986a; Kardong et al, 1986; Janoo and Gasc, 1992; Kardong and Bels, 1998; Cundall and Deufel, 1999; Deufel and Cundall, 1999). Kardong and Bels (1998) found that C. viridis drives the head to the prey by rapidly straightening all curves in the anterior trunk simultaneously (gate model) or by driving part of the trunk around a postural curve (tractor-tread model), a phenomenon that looks more like moving a curve caudally (Fig. 9.7). Acceleration of points on the trunk begins anteriorly and progresses caudally in both gate and tractor-tread strike types, and hence anterior points achieve higher velocities than more caudal points. In most strikes a third or less of the trunk was recruited to move the head, with the remainder of the trunk remaining essentially stationary, its inertia serving as the launching platform for the head. That rapidly striking snakes can use one of a number of kinematic variants to strike is an unexpected finding considering the complexity of the musculature of the anterior trunk (Gasc, 1981), the presumed neural coordination required to modulate the behavior, and the speed of the process. However, observations of corrections to the strike trajectory made by Frazzetta (1966) and used in a dramatic but fictional account of a Porthidium nasutum strike (Greene, 1997) have not been supported by studies of striking in either booids
305
Postural curve F I G U R E 9.7. Gate (A) and tractor tread (B) models of trunk kinematics during striking. (A) Angles at all ''joints'' between illustrated links increased simultaneously. (B) A single curve is moved rapidly caudally on the body while the trunk anterior to the curve is straightened. From Kardong and Bels (1998), Company of Biologists Ltd., with permission.
(Cundall and Deufel, 1999) or viperids (Kardong and Bels, 1998). Gans (1986) suggested that concertina kinematics might be the most plesiomorphic locomotor mode in snakes. Greene (1997) postulated that striking is derived from concertina movement, and aspects of the gate model of Kardong and Bels (1998) indeed fit this pattern. We now need the kinds of empirical data provided by Jayne (1988a,b) for locomotion to sort out the processes by which the trunk drives the head during a strike. With respect to the behavior of the jaws during a strike, anecdotal accounts emphasize extraordinary gape (to 180°) and the speed of head movement (Klauber, 1956; Murphy and Henderson, 1997). Highspeed films (400-1000 fps) have resolved many details (Frazzetta, 1966; Kardong, 1974; Janoo and Gasc, 1992; Kardong and Bels, 1998). Kardong and Bels (1998) saw little variability: one out of 21 strikes missed the prey initially and resulted in two corrective jaw movements after the prey had been reached. Janoo and Gasc (1992) graphed a flawed strike in Bitis gabonica in which the snake missed the prey, partially closed its mouth, and then used two successive reopenings (both of which also missed the prey) within 0.1 sec before finally closing the mouth. In boids, however, precise kinematics vary from strike to strike and partly determine which jaw (upper or lower) hits the prey first (Cundall and Deufel, 1999). The following description is based partly on existing literature and partly on unpublished data collected by Cundall and Deufel for a colubrid {Elaphe ohsoleta), three species of Crotalus, and nine species of booids, seven of the latter covered in Cundall and Deufel (1999). Viperids and booids share similar predatory strike
306
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
kinematics to the point of prey contact. In the majority of strikes, the braincase is aimed at the dorsal surface of the prey with varying dorsal flexion (10-40°). The mandibles are depressed, but usually only 45 - 60° relative to the braincase axis. Hence, gape values at prey contact are typically well below 90° if measured as the angle between the mandible and an axis parallel to the braincase at the quadrate mandibular joint, not as the angle formed between the edges of the upper and lower jaws as has been done most frequently. In about a third of analyzed booid strikes, gape averaged 35°. These strikes are referred to as driving scissors strikes (DSC strikes; Cundall and Deufel, 1999) because the prey is wedged between upper and lower jaws, which tend to hit the prey simultaneously (Fig. 9.8). In the majority of strikes, however, gape angles are larger and the mandibles hit the prey first (Janoo and Gasc, 1992; Kardong and Bels, 1998; referred to as MAN strikes by Cundall and Deufel, 1999). The braincase travels over the prey and is brought down on the prey's surface as the bite phase of the strike begins. Booids, and presumably most snakes, rarely hit prey with the upper jaw first (PMX strikes of Cundall and Deufel, 1999). Kinematics underlying such a strike pattern may, however, be characteristic of defensive strikes (Kardong, 1986a). Generally, the upper jaws are protracted as the man-
dibles are depressed. Upper jaw protraction drives the anterior maxilla of booids (and probably colubrids) laterally and dorsally and results in anterior rotation of the fangs in viperids. Protraction is a more complex process than has been previously appreciated, however. Protraction, activated by the dorsal constrictor muscles, pulls the pterygoids anteriorly and dorsally. The palatines are pulled anteriorly, and the palatopterygoid joint is usually pulled laterally and dorsally by the pterygoid levator. As protraction of the upper jaws continues, the palatines push the snout dorsally. In boids and pythonids, and possibly in many colubrids, the palatine teeth may contact the prey before the maxillary teeth because the palatines usually lie ventral to the anterior maxillae as the upper jaw contacts the prey (Cundall and Deufel, 1999). Predatory striking has a number of possible consequences in snakes. In many constrictors, the strike ends in a bite and immediate coiling of the anterior trunk as the snake applies constricting coils. In booids (and some colubrid constrictors), the head usually flexes ventrally to initiate the first coil (Greene and Burghardt, 1978), a movement probably facilitated in MAN strikes if the snake's braincase retains momentum as mandibular movement is impeded by the inertia of the prey (Fig. 9.8). DSC strikes invariably result in slight
F I G U R E 9.8. Models of strike kinematics: (a-c) a DSC strike showing orientations of the snout (sn), braincase (br), supratemporal (link from 3 to 5), quadrate (link from 5 to 6), mandible, and maxilla immediately before contact, at contact, and immediately postcontact with the prey. As shown, if the prey exceeds the gape, the teeth slide over the prey until movements of the snake or the prey align forces acting on the teeth with the long axes of the teeth, (d-f) The same elements during a MAN strike. 1, combined prokinetic joint between snout and braincase and dorsal prefrontal-frontal joint; 2, prefrontal-maxillary joint; 4, joint between neck and braincase; qml, quadratomaxillary joint.
307
9. Feeding in Snakes delays in initiation of constriction, possibly because the snake must expend muscle effort to flex the braincase ventrally (Cundall and Deufel, 1999) or laterally (Greenwald, 1978). Rapid formation of the first coil is therefore enhanced by the continued momentum of the snake's braincase after contact, and most constrictors may therefore depend on their mandibles, suspensoria, and associated adductor muscles to absorb the shock of impact with the prey (Cundall, 1998). Viperid predatory strikes end with rapid deceleration of the head either just prior to contact with the prey (Kardong and Bels, 1998) or immediately following contact, with the braincase swinging around the decelerating mandible as in boids and pythonids (Cundall and Deufel, 1999). Biting is accompanied by rapid closure of the jaws and retraction of the maxilla, readily visualized if the fang fails to penetrate the prey. It therefore appears that the bite is a response to contact with the prey (Kardong and Bels, 1998) as it is in the slower openmouth sweeping of aquatic snakes. The bite in elapids and viperids is usually accompanied by venom injection, a process considered later under prey restraint. Video records of striking in Agkistrodon imply that the bite can be modulated and the mouth kept open even after mandibular contact with the prey (Cundall and Deufel, personal observation). How this is done remains a mystery. In viperids, protraction not only rotates the maxillae anteriorly (Kardong, 1974), but swings them laterally, with the result that left and right fang tips may be separated by twice their resting distance when they contact the prey. The extraordinary spread reached by the fangs occasionally results in the failure of one fang to penetrate the body of a smaller prey item, a phenomenon noted in Crotalus over a century ago by Mitchell (1861), more recently in Bitis (Janoo and Case, 1992) and Crotalus (Kardong and Bels, 1998), and assumed to occur in some elapids (Acanthophis, Notechis) in which the fangs are oriented lateroventrally instead of simply ventrally when fully protracted (Fairley, 1929). Some functional implications are considered in Section II,B,3. One of the problems faced by many ambush predators such as viperids is that larger prey may approach without warning and with such speed that little time is available for strike preparation. We know of no studies that have considered this problem directly, nor the problem of maintaining optimal muscle response times. Do ambush predators really remain motionless for days on end or do they engage in exercise periods to maintain muscle tone and reaction times? One of the potential solutions to this problem is to entice prey to approach, and a number of juvenile viperids, boids, elapids, and a few colubrids use caudal luring of prey (Pough, 1988; Sazima and Puorto, 1993). Luring involves a strategy dif-
ferent from typical ambush strikes in that the snake activates the lure and strikes near the caudal end of its own body—in other words, the entire behavior is under some control of the snake. In snakes that use caudal luring, both the color and the shape of the tail tip may be modified, but the effectiveness of these changes for luring remains controversial and difficult to test (Schuett et al, 1984; Sisk and Jackson, 1997). B. Prey Restraint Contact of the snake's head with the prey item initiates either immediate transport or some type of restraint reaction to the prey's efforts to escape. For smaller prey that are transported while still alive, activities prior to transport are typically referred to as prey manipulation, but that necessarily includes some restraining mechanism to prevent the prey from escaping. For larger or potentially dangerous items, restraint usually involves mechanisms for immobilizing the prey before transport. Many of the specialized features of the jaws of snakes are associated with restraint, some of which have been reviewed by Savitzky (1983). 1. Blunt Trauma and Mechanical
Reduction
We suspect that the most primitive restraint pattern is holding prey in the jaws until their struggling diminishes due to fatigue or trauma caused by lacerations and internal bleeding. This strategy, which depends on the capacity of the adductor musculature of the snake's jaws to resist fatigue longer than the locomotor muscles of the prey, may occur in basal alethinophidian snakes and, in modified form, in a variety of nonvenomous colubrids. In the latter taxa (e.g.. Coluber, Masticophis, Ruben, 1977a,b; some Thamnophis, Gregory et al., 1980), prey are often held in the mouth but are partially restrained by a coil of the body pressing them against an object in the environment. Another behavior, used by some colubrids such as Masticophis, Coluber, and Drymarchon, is to whip the prey from side to side while holding it firmly in the jaws, again often hitting the prey against objects in the environment, a behavior that would appear to risk damage to the snake's head but is surprisingly effective at immobilizing small mammal prey. Although snakes generally do not chew or mechanically reduce their food prior to ingestion, many species manipulate prey items and a few even radically modify the external characteristics of prey. Kinematic patterns associated with chewing in other tetrapods do appear in crude form in some snakes, as successive advance-close movements of the jaws on one side, while the other side maintains a holding grip on the prey. Successive jaw
308
David Cundall and Harry W. Greene
movements might crush prey and/or slice its skin, both of which would increase the penetration of digestive enzymes and hence serve the same function as chewing. Heterodon platirhinos uses successive retractions of the maxilla on one side to slice open the skin of toads (Kroll, 1976; Cundall, 1983). Similar motions are used by species of Oligodon (Coleman et ah, 1993) and, presumably, Prosymna (Broadley, 1979) to open the eggs of other squamates. In an extraordinary behavioral convergence based on an entirely different morphology, some species of the Australian elapid Simoselaps use enlarged rear dentary teeth to slice open squamate eggs (Shine, 1984; Scanlon and Shine, 1988). Most snakes swallow relatively large items head first, thereby folding surface features back rather than working the jaws against appendages, fur, etc. (Greene, 1976; Kardong, 1986b). Jaw movements leading to headfirst transport might incidentally repeatedly puncture the prey's integument and thus introduce oral secretions. Tail-first transport usually involves relatively small prey (e.g., Erythrolamprus, Greene, 1976; Charina bottae, Rodriguez-Robles et al, 1999), but Regina alleni swallows odonate larvae head first and crayfish tail first, thereby avoiding defensive claw attacks by the latter (Franz, 1977; Godley, 1980). Snakes of diverse families often pull immobilized prey a short distance prior to ingestion (e.g., Micrurus fulvius, Greene, 1984), perhaps aligning the item for easier swallowing; some boas and pythons substantially stretch mammals by holding them in a constricting coil and pulling the foreparts with their jaws (personal observation). A few species actively modify the external characteristics of prey prior to ingestion [e.g., removal of crab appendages by Fordonia leucohalia, Shine (1991c); pulling snails from their shells by Storeria, Rossman and Myer (1990)]. The adductor musculature and slender form of the mandible in most snakes obviate most types of prey reduction requiring large forces. However, a few snakes are known to crush prey either before or during transport. The homalopsine colubrid Fordonia, which feeds almost entirely on crabs, may pull some or all of the legs off crabs prior to ingestion and not only punctures the carapace, but may apply forces large enough to crack it entirely during transport (Savitzky, 1983; Jayne and Voris, unpublished data). Limited available data suggest power strokes during adduction that may have some similarities to temporal summation patterns used by other squamates that feed on hard but brittle prey (e.g., Trachydosaurus, Gans and de Vree, 1986). In Cylindrophis (Cundall, 1995) and possibly many other snakes with robust jaws and adductor musculature (Greene, 1997), forces generated by restraining bites may be strong enough to cause widespread hemorrhaging and
mechanical damage to subdermal muscular and vascular tissues. As yet, neither the kinematics nor the mechanics of crushing have been examined in detail. 2.
Constriction
Many species in most living alethinophidian families (including Cylindrophiidae, Loxocemidae, Xenopeltidae, Boidae, Pythonidae, Tropidophiidae, Bolyeriidae, Acrochordidae, Colubridae, and Elapidae) are known to constrict (Greene and Burghardt, 1978; Greene, 1994). The conservation of this complex behavior was one of the first concrete examples of behavioral homology above the family level, but the origin of constricting remains obscure. Although Frazzetta (1970a) interpreted Dinilysia as a nonconstrictor and no descriptions exist of constriction in Anilius, the mere existence of robust jaws is no proof of nonconstriction, as shown by the use of a constriction pattern in Cylindrophis similar to that seen in xenopeltids, boids, pythonids, and acrochordids (Greene and Burghardt, 1978). Constriction is not known to occur in scolecophidians nor in lower alethinophidians selecting invertebrate prey—notably uropeltids and probably Anomochilus. Hence, one of the critical novelties characterizing the ancestor of cylindrophiids and all remaining alethinophidians was constriction, and its appearance coincided with the exploitation of an increasing diversity of vertebrate prey (Greene, 1994). The anatomy of the anterior trunk is known in detail for remarkably few snakes (d'Alton, 1838; McKay, 1889; Gasc, 1981; Pregill, 1977) and how these elements behave during constriction remains obscure. Constriction may well include at least two qualitatively different behaviors: restraining constriction and killing constriction. The former is used by snakes that feed on elongate prey such as eels and seems mainly an effort to prevent the prey from escaping during handling. The latter is used by many snakes that eat endotherms. Constrictors appear to have shorter epaxial muscle segments and more vertebrae (Ruben, 1977a,b), but the relationship between niuscle length and constriction remains problematic (Jayne, 1982), as does the longitudinal homogeneity of muscle organization in snakes. It is possible that muscle lengths may vary at different positions on the trunk and, further, that constrictors use only part of the trunk in active constriction. Many constrictors use only the anterior third to half of the trunk for constriction except following the capture of very large prey. Certain constricting patterns are used by most basal alethinophidians above uropeltids, including basal macrostomates (Greene and Burghardt, 1978). These involve ventral bending of the trunk immediately behind
9. Feeding in Snakes the head followed by a twist and then horizontal coiling of the anterior trunk with the ventral surface of the snake facing either toward or away from the snake's head. This pattern is favored by MAN striking kinematics. The use in colubrids of diverse alternate constricting patterns that do not involve an initial twist and may involve lateral rather than ventral bending of the anterior trunk may follow different strike patterns, but this has yet to be examined carefully. 3. Envenomation and Venom Metering Although observations on the methods venomous snakes use to inject venom date back centuries, surprisingly little has been published recently. Envenomation occurs during the bite phase of the strike or during prolonged "chewing" actions in many rear-fanged species. The bite phase of rapid viperid and elapid strikes is relatively prolonged in Crotalus, even when prey are released. Kardong and Bels (1998) found that the bite lasted three to five times as long as the time taken to reach the prey (up to 0.2 sec), and tended to be shorter for larger snakes and for bites of larger prey. During the bite all of the muscles of the head became maximally active (Kardong et al, 1986). Temporal analysis of muscle activity patterns by Kardong et al. (1986) showed that most kinematic events of the strike were initiated before the muscles became electrically active, a finding we regard as unlikely. In Crotalus and Bitis, failure to embed a fang may result in unilateral rapid opening and closing of the jaws, sometimes involving several rapid protractions of the unembedded fang (Janoo and Gasc, 1992; Kardong and Bels, 1998). In about half the records of Cundall and Deufel (unpublished data) in which one fang initially missed the prey, the fang sheath on the unembedded fang still extended to the fang tip during withdrawal of the head from the prey. There is no evidence that venom was released from the unembedded fang, suggesting that viperids may be able to control right and left fang use and venom injection rapidly and independently. Behavioral records suggest that correcting fang placement and venom injection are both under feedback control from receptors in the mouth (Kardong and Bels, 1998), possibly associated with the fang sheath. Numerous measures of venom injected into prey following a single strike suggest that some elapids (Morrison et al, 1983a,b) and viperids (Gennaro et al, 1961; Hayes, 1995; Hayes et al, 1995) regulate injection, introducing larger quantities of venom into larger prey. Although evidence remains equivocal and may apply to only some or a few venomous species (e.g., AUon and Kochva, 1974), enough evidence exists for the regula-
309
tion of bite mechanics to suggest that regulation of venom injection is within the limits of potential motor control. Venomous snakes vary widely in the form and histology of the venom gland (Kochva, 1978), organization of extrinsic venom gland muscles (Haas, 1973), and structure and relationships of the fangs and supporting maxillae (McDowell, 1968, 1986, 1987). Feeding mechanics and fang use have been examined carefully in about 1% of the 500 species of front-fanged elapids and viperids and, with the exception of Fairley's (1929) examinations of fresh, dead elapids, all these studies concern viperids (Mitchell, 1861; Van Riper, 1950, 1953, 1955; Kardong, 1975; Kardong et al, 1986; Janoo and Gasc, 1992; Kardong and Bels, 1998). Considerable hydrostatic pressure can be generated at the discharge orifice of the fang as is obvious from the behavior of elapid spitters (Rasmussen et al, 1995), many of which show reductions in the length of the discharge orifice (Wiister and Thorpe, 1992). Although we still do not know how tooth shape relates to spitting performance, performance varies among species (Rasmussen et al, 1995), with few spitting venom further than a few meters. Some species eject venom as a spray, others as a stream. Head movement in space can be quite variable among spitting species: some spitting while lunging toward an aggressor, with others spitting while lunging away. Spitting may be accompanied by a hiss; forceful exhalation may help propel venom but is not needed for spitting and tends to disperse the venom stream (Rasmussen et al, 1995). Although spitting is rarely if ever used for catching prey, aspects of the fluid mechanics of venom injection might be most easily approached by studying spitting. Tubular fangs in viperids (Klauber, 1956) and elapids (Bogert, 1943) act as injection needles. The needle is fed from a nonmuscularized reservoir at the terminus of a duct from a muscularized venom gland (Hager, 1906). The moderately high pressures at the discharge orifice of the fang (Kardong and Lavin-Murcio, 1993) presumably stem from venom gland compression. Some venomous snakes produce venom that is moderately viscous (e.g., Dendroaspis; Broadley, 1990); pumping it rapidly through fang canals should require considerable hydrostatic pressure. Thin tendons extend from the pterygoideus muscle into the fang sheaths of both elapids and viperids (e.g., Mitchell, 1861; Hager, 1906; Phisalix, 1922; Kardong, 1973). The arrangement of tendons would cause the anterior, lateral, and medial edges of the sheath to be pulled taut against the fang surface during contraction of the pterygoideus at the beginning of the bite phase (but the only published EMG results suggest that the pterygoideus becomes
310
David Cundall and Harry W. Greene
maximally active about 0.3-0.4 sec after the bite begins; Kardong et al, 1986). In both viperids and elapids, the venom gland itself is surrounded by the adductor musculature, and in most species, one of the external adductors (superficialis in elapids, profundus in viperids— terminology of Haas, 1973) is modified as a gland compressor. During the bite phase of the strike, as the fang penetrates the prey surface and the fang sheath is pushed dorsally, venom squeezed into the sheath chamber anterior to the opening of the venom canal is forced down the venom canal, presumably by pressure generated by compression of the venom gland by the compressor glandulae and compression of the fang sheath chamber by the pterygoideus. The behavior of individual fangs during the strike, as discussed earlier, should reveal a great deal about the use of the fang sheath in regulation of venom flow. Rear-fanged snakes have a variety of fang structures that have been reexamined by Jackson and Fritts (1995, 1996) and Young and Kardong (1996). Enlarged maxillary teeth occur in many colubrid taxa, and in many of these the fangs are associated with serous glands that secrete toxins (McKinstry, 1983; Jansen, 1987; Weinstein and Kardong, 1994; Underwood, 1997). Kardong and Lavin-Murcio (1993) measured capillary pressure in the fang canal of the rear-fanged colubrid Boiga irregularis and found it to be very low, although Jansen and Foehring (1983) found rapid venom flow (they did not measure pressure) following stimulation of the superficial external adductor in Thamnophis elegans. Open fangs beg the question of how, or even if, venom is "injected'' by rear-fanged snakes. Kardong and LavinMurcio (1993) concluded, as have many previous workers, that the low-pressure systems of rear-fanged snakes probably function in part by laceration of tissues and the introduction of venom through open wounds. The question of whether venom serves in restraining or immobilizing prey is more difficult to determine. Boiga irregularis kills mammal prey by constriction before they succumb to envenomation (Rochelle and Kardong, 1993) and other venomous snakes also use constriction (Shine and Schwaner, 1985). However, rear-fanged species have been observed to immobilize vertebrate and invertebrate prey solely by envenomation before transport (Greene, personal observation). Venoms from some rear-fanged snakes facilitate digestion (e.g., Jansen, 1987; Rodriguez-Robles and Thomas, 1992). Kardong (1996) has questioned the assumption that envenomation occurs in most colubrids with rear fangs and enlarged Duvernoy's glands, but acknowledges that some rear-fanged snakes have injecton abilities that far exceed those implied by Kardong and Lavin-Murcio (1993). Human fatalities from the bites of rear-fanged snakes (Minton, 1990), including the deaths
of two prominent herpetologists (Greene, 1997), attest to the effectiveness of their envenomating system (see also Rodriguez-Robles, 1994). We clearly remain ignorant of the most critical functional aspects of rearfanged systems. C. Prey Manipulation Although most snakes use obvious and sometimes spectacular jaw motions to orient prey after ingestion, these movements have received no formal analysis. They tend to be highly variable and directly related to the orientation and nature of the particular prey item. A number of prey manipulation movements deserve closer study, however. Many aquatic snakes catch prey at midbody and pass them laterally through the oral cavity to reach an end. This movement is most effective if the palatomaxillary arches move laterally with limited advance, a feat accomplished by some snakes (Nerodia, Farancia; personal observation). Although manipulatory movements are invariably unilateral in nature, there is good reason to think that precise palatomaxillary movements are modulated differently from patterns used during intraoral transport. D . Intraoral Transport The vast majority of living snakes differ from other tetrapods in their ability to transport entire prey through their oral cavities using alternate movements of the left and right jaw elements. It has only recently become clear that many of the structural modifications underlying this extraordinary behavior evolved within snakes and that some elements of transport kinematics may be intimately tied to prey capture kinematics. Of the numerous possible approaches to analyzing transport in snakes, we arrange our treatment around structural modifications that best fit current notions of the phylogeny of crown clade snakes. The evolutionary implications of our approach are detailed in Section V. 1. Lateral Jaw Transport Living basal snakes (scolecophidians plus anomochilid and most uropeltid alethinophidians) lack teeth on their medial upper jaw elements (palatines and pterygoids). Given that these clades also have narrow, bifid, scleroglossan tongues, they are structurally constrained to using lateral jaw elements to transport prey through the oral cavity. However, it is extremely misleading to equate scolecophidian lateral jaw transport systems with the alethinophidian condition, despite their shared dependence on the lateral jaws. In scolecophidians, as noted previously, either the maxilla
9. Feeding in Snakes (typhlopids and presumably anomalepidids) or the mandible (leptotyphlopids) is modified for raking in small prey, in both cases with pronounced anteroposterior movements. The same kinematic patterns appear to be used for both ingestion and transport. In all scolecophidians the right and left dentaries are tightly attached and movements of the mandibles are synchronous (Haas, 1964; 1968; List, 1966; Groombridge, 1979b; Kley, 1998). In leptotyphlopids this results in the bilateral transport of prey with the dentaries alone functioning as a single median unit to grasp and ingest prey. Kley's (1998, personal communication) records show caudal flexing of the mandibular symphysis during retraction, a process that would increase the efficiency of the system by carrying prey further caudally in the oral cavity. How the dentary teeth are disengaged and the prey retained in the mouth remain obscure; as noted previously, muscle X may help and suction might play a role in leptotyphlopid ingestion and transport of some prey types (Groombridge, 1979b). There is no evidence yet of a transport process distinct from ingestion (Kley, 1998). Typhlopids effectively sweep prey into the oral cavity with their rapidly fluttering maxillae, and the same process presumably drives prey into the esophagus. There is no evidence of a change in kinematics following ingestion, and because multiple prey items are sometimes ingested simultaneously or in quick succession, there is little opportunity to partition ingestion and transport (Thomas, 1985; R. Thomas, T. J. Gush, and R. Diaz, personal communication; Kley, 1998, and personal communication). How one species of typhlopid {Acutotyphlops subocularis) transports worms remains a puzzle; the vast majority of prey recovered from stomachs are small, intact arthropods (Webb and Shine, 1993a). Transport in anomalepidids has yet to be described, as is true of anomochilids and uropeltids (Table 9.1). In the latter two alethinophidian clades, however, the maxillae are longer, the dentaries toothed, and the snout slightly prokinetic and, presumably, rhinokinetic in the sense that ventral snout elements may move slightly relative to the nasals (Cundall and Shardo, 1995). It is therefore possible that limited unilateral maxillary motion occurs in these clades as a result of a process described in Cylindrophis as snout shifting (Cundall, 1995) in which the pterygoid, palatine, and vomer-septomaxilla (ventral snout) complex move together, the movement of the palatine generating lateral and rostral motion of the maxilla. In Anomochilus the ectopterygoid is a splint of bone floating in the ectopterygo-maxillary ligament. The ligament and enclosed ectopterygoid presumably cannot transmit compressive forces from the pterygoid to the maxilla, but may
311
serve to retract the rear end of the maxilla during adduction (Cundall and Rossman, 1993). We suspect that if Anomochilus or uropeltids have any mobility of the upper jaws, the maxillae function much as in a hypothetical primitive stage of unilateral transport (Figs. 9.9a and 9.9b). In this model, jaw advance during transport depends on lateral rotation of the whole head augmented by lateral movement of the caudal end of the maxilla away from the central axis of the braincase. This system retains tight connections between the upper jaw and the snout but might show an early stage of snout shifting in which maxillary abduction occurred through palatomaxillary displacements rather than through the pterygoid-ectopterygoid-maxillary linkage. The putative basal position of anomochilids and uropeltids among alethinophidians is not yet well tested. Aspects of their anatomy, in particular the absence of palatopterygoid teeth and reduced kinesis of the upper jaws, were formerly interpreted as being derived secondarily in response to fossoriality. On both counts, however, the evidence now seems to favor plesiomorphy as the explanation for these conditions. With regard to the absence of palatopterygoid teeth, this is the general condition among scleroglossan lizards and departures from this condition are seen only as the presence of small teeth unlike lateral teeth (Estes et al, 1988), as evident in anguids and the few uropeltids that have teeth on the medial upper jaw elements {Melanophidium punctatum, Rieppel, 1977b; Platyplectrurus, Smith, 1943). In the absence of any living scleroglossan lizard with palatopterygoid teeth forming a continuous longitudinal row of teeth similar in shape and size to maxillary and dentary teeth, it is possible that palatopterygoid teeth of most alethinophidians are neomorphic and not homologous to the palatal teeth of lizards. Restricted kinesis of the upper jaws (Fig. 9.10) is seen in all currently accepted basal alethinophidian clades. Most of these clades appear to be fossorial. However, there are a number of fossorial colubroid genera and none show the structural patterns associated with reduced palatomaxillary mobility. Of the various possible evolutionary scenarios, one is that basal alethinophidians retained a plesiomorphic pattern of upper jaw kinesis, an interpretation most consistent with current phylogenies. 2. Lateromedial Jaw Transport The appearance of teeth on the medial upper jaw bones (palatine and pterygoid) in all clades above uropeltids was accompanied by elongation of the anterior palatine and its extension over the ventral surface of the vomeronasal capsule. The origin of teeth on the
a
"35P^^^
^w^^
F I G U R E 9.9. Simple model of prey diameter constraints acting in a generalized squamate skull assumed to be transporting an elongate, cylindrical prey item without use of the tongue. Inertial transport would initially involve rapid side-to-side movements of the snout around the posterior bite point of the jaws (shown by the small circle). As prey diameter increases (b), progression of the head relative to the prey diminishes as a function of constraints on the lateral edge of the jaw limiting lateral flexion of the head. Prey larger in diameter (c) either could not be transported at all or could be transported more efficiently if the lateral edge of the mouth could move laterally, a feature of streptostylic quadrate suspension. Lateral streptostylic movements are limited by connections between the anterior palate and the quadrate through a medial linkage composed of the pterygoid, palatine, vomer, and premaxilla and/or a lateral linkage formed by the pterygoid, ectopterygoid, maxilla, and premaxilla. As prey diameter increases further (d), spreading of the quadrates must increase. This may occur through increased flexion at palatal-quadrate cormections or through anterior separation of upper jaw bones from the snout and braincase. Various scleroglossan clades, including basal alethinophidians, have exploited solutions (a-c). Solution (d) and its attendant morphological innovations are seen only in macrostomatans and possibly, in modified form, in loxocemids and xenopeltids.
9. Feeding in Snakes
FIGURE 9.10. Dorsal or ventral view of the upper jaws of alethinophidians basal to colubroids (a) and of colubroids (b). Rest position is shown unshaded in (a) and in dashed outline in (b); position following protraction by the protractor and levator pterygoidei muscles is shaded in both (a) and (b). The arrow F indicates the approximate direction of force generated by contraction of the pterygoideus in each system, (a) The entire upper jaw might be retracted except that it is tightly bound to the snout through the palatine (PI). Regardless, both the maxilla (Mx) and the palatopterygoid bar (PI and Pt) will tend to act together because of the shortness of the ectopterygoid link (ec) between the pterygoid and the maxilla, (b) Contraction of the pterygoideus will pull the ectopterygo-maxillary joint caudally and cause the palatopterygoid joint to flip medially (unstippled solid outline), but only if the rear end of the pterygoid is prevented from moving caudally (as shown here) by the continued contraction of protractor muscles.
palatopterygoid bar creates an interesting problem in transport mechanics because these teeth become the part of the skull in most intimate contact with the prey. As a result, intraoral transport becomes dependent on movements of one palatopterygoid bar relative to the other. In cylindrophiids, these movements are limited and unilateral transport is slow. However, when a prey item reaches the esophagus in C. rujfus, it is somehow gripped with the esophagus following longitudinal compression of the anterior trunk by very tight bends in the vertebral column. The snake then releases the prey with both right and left jaws and straightens the trunk, thrusting the head forward over the prey. Transport of elongate prey using trunk compression and bilaterally symmetric jaw movements can be very rapid (Cundall, 1995). Lateromedial transport may be characteristic of all clades above uropeltids but basal to colubroids. In other words, it spans the macrostomatan transition. In all of these clades the ectopterygoid is relatively short
313
and the pterygoideus muscle arises from the proximal ectopterygoid and lateral edge of the pterygoid (Frazzetta, 1966; Cundall and Greene, 1982; Cundall, 1983; McDowell, 1986,1987; see Figs. 9.3 and 9.10). However, transport becomes more efficient in some taxa because the choanal process of the palatine either loses close contact with the vomer (many boids and pythonids) or becomes thin and pliable (xenopeltids) and the joints at both ends of the prefrontal become looser or more mobile. In most boids and pythonids, the palatine ends well short of the anterior end of the maxilla. The anterior maxilla may be moved laterally and dorsally, and is only loosely connected to the premaxilla. However, the effective advance of the maxilla is limited by its caudal attachments, and as a result, the caudal end of the maxilla closely tracks the movements of the palatopterygoid bar (Frazzetta, 1966). Given variations in the attachment of the palatine choanal process to the vomer, and of the pterygoid to the sphenoid in boids and pythonids (Kluge, 1991, 1993a,b), movements of the palatopterygoid bar probably vary accordingly but remain to be analyzed. In many boas and pythons, the intermandibular soft tissues (Young, 1998) are capable of extraordinary stretch (Fig. 9.11). Whereas in pythons this may be aided by the longitudinal arrangement of the anterior intermandibular muscles (Fig. 9.4), boas have a typical transverse arrangement of anterior intermandibular muscles (Groombridge, 1979a) but still have considerable stretch capability—estimated from diameters of prey that have been eaten to be in the range of 20-30 times the resting length of these tissues, possibly more (Fig. 9.12; see also Cogger and Zweifel, 1997, p. 178). Hence, the macrostomate qualities of these snakes com^e in large measure from as yet unexplained aspects of soft tissue organization and only to a relatively minor extent from features of the skeleton. Most boas and pythons eat prey relatively large in diameter (it must be long enough to extend beyond the jaws to be constricted) and the prey is usually contacted by the entire palatomaxillary arch during transport. EMG records show that Python reticulatus juveniles use a combination of unilateral and asymmetric bilateral activity of muscles when swallowing small prey (Kardong and Berkhoudt, 1998). The patterns are similar to previous findings for colubroids (Cundall and Cans, 1979; Cundall, 1983). Python uses bilaterally symmetrical jaw movements in the final phase of transport, usually combined with elevation of the entire head, possibly to allow gravity to help tranport. The basic pattern appears similar to that seen in Cylindrophis (Cundall, 1995). Whether bilateral jaw movements in Python stem from bilaterally symmetric muscle activity remains unclear. Kardong and Berkhoudt (1998) listed three phases of
314
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
Qu
St
cur during this phase of transport (Kardong and Berkhoudt, 1998). Previous discussion of the differences between lateromedial transporters and medial transporters (later) has been limited to a few anatomical descriptions (Cundall and Greene, 1982; Cundall, 1983; McDowell, 1986,
a
F I G U R E 9.11. Anterolateral view of a pythonid skull to show elements of the head that define potential gape size, as a function of prey diameter (Dp). The length of the mandible (Lm) is the primary skeletal determinant of gape size, and this length is correlated with lengths of the supratemporal (St) and quadrate (Qu). Rotations at the quadrate-supratemoral and dentary-compound (Dt, Cp) joint allow the skeleton to conform more closely to prey shape. The stretch capability of intermandibular soft tissues (Sim) provides the other major determinant of gape size.
Opening for Python (slow opening I, fast opening, and slow opening II) and the usual two phases of closing (fast closing and slow closing) that characterize most tetrapod jaw cycles during transport. Their interpretation comes from measurements of gape made using a magnetic sensor fixed to the braincase and a magnet glued to the skin between the mandibles. Cundall and Deufel's (unpublished data) video records of transport in various species of boids and pythonids suggest that transport is different in these clades. Although unilateral jaw advances occur, after the approximate middle of a mammalian prey reaches the mouth, boids and pythonids appear to use relatively rapid alternating rightleft advances in which the palatomaxillary teeth may not even contact the prey but the mandibles are brought forward around the side of the prey. Jaw movements are coupled to sinuous movements of the vertebral column similar to those described by Kley and Brainerd (1996) and Moon and Kley (1997) in a few colubrid genera. Bilaterally synchronous upper jaw advance may also oc-
F I G U R E 9.12. Models of cross sections through braincases of a basal alethinophidian (a) and a macrostomatan (b), same size. By increasing suspensorial (Qu) and mandibular (Md) lengths, macrostomatans have achieved a larger gape (left side). In addition, the increased stretch potential of intermandibular soft tissues further increases the potential prey cross-sectional area (right side). Numbers indicate relative size of areas enclosed by semicircles.
9. Feeding in Snakes 1987) and one functional treatment (Cundall, 1995). Kardong and Berkhoudt (1998) found only broad similarity between Python and other snakes. As noted previously, a number of differences may exist. It is hoped that future work on clades proposed to use lateromedial jaw transport will focus on variables that might provide direct tests of how these systems differ from the medial jaw transport mechanisms of colubroids. Such variables include precise movement of the maxilla relative to the palatine and pterygoid, precise timing of dorsal constrictor and ptergoideus activity, and measures of the performance of the system. 3. Medial Jaw Transport Medial jaw transport looks superficially similar to lateromedial jaw transport. However, it occurs only in clades for which structural reorganization of the upper jaws allows functional decoupling of the medial and lateral upper jaws (Cundall, 1983; McDowell, 1986). Alternating advances of the right and left jaws produce a motion that Boltt and Ewer (1964) termed the "pterygoid walk." The process had been accurately described by Gadow (1901) who noted, as did Boltt and Ewer, that snakes probably use their head muscles to crawl over larger prey rather than pulling prey into their mouths. Empirical evidence for this claim did not arrive until 1979 (Cundall and Cans) and little has been added since (Cundall, 1983, 1987). The transport process involves repeated use of a basic pattern of muscle activity and bone movements now described for 6 of the approximately 2100 species of colubroids [the viperids Crotalus durissus, Kardong et al. (1986) and Agkistrodon piscivorus, Cundall (1983); the natricine colubrids Nerodia fasciata and N. rhomhifer, Cundall (1983); Cundall and Cans (1979); the xenodontine colubrid Heterodon platirhinos, Cundall (1983) and the colubrine colubrid Elaphe ohsoleta, Cundall (1983); Table 9.1], but observations of feeding in other colubroid species suggest that many conform to the mechanism reviewed by Cundall (1987). An abbreviated account is given later with emphasis on how medial transport differs from lateromedial and lateral mechanisms. Medial transport is achieved by alternating advances of the medial upper jaw bones. Palatine and pterygoid movements are carried to the maxilla through the ectopterygoid link and are correlated with abduction and adduction movements of the mandible. The process looks as though the snake is alternately opening and closing its right and left jaws, but the upper jaws actually move forward and laterally, with mandibular movements timed to the release, advance, and catch of the palatopterygoid tooth row on the prey surface. Initial movements of the advance phase often occur
315
before there is evidence of muscle activity and appear to result from elastic recoil or gravity. One or more of the dorsal constrictor protractor series then become active, followed by the depressor mandibulae as the upper jaw is drawn forward and laterally and the lower jaw is depressed and protracted (Cundall and Cans, 1979; Cundall, 1983; Kardong et al, 1986). In most medial transporters, jaw movement increases in velocity at the end of advance and some of the mandibular adductors frequently become active, although their activity is not reflected kinematically (Figs. 9.15 and 9.16). Because the prey usually lies on top of the anterior mandible during ingestion, the mandible behaves much like a crutch on which the braincase and upper jaw are swung upward and forward during most of the advance phase. Very large prey come to lie medial and ventral to all of the mandible during much of the transport process. This may not be intuitively obvious but is readily seen in a few spectacular photographs [see Arnold (1983, p. 348) for a homolopsine example; Bauchot (1994, p. 110) or Cogger and Zweifel (1997, p. 178) for a lateromedial transporting pythonid example]. The end of advance [termed fast opening by Cundall and Cans (1979)] is marked not only by rapid protraction of the active upper jaw but also by protraction of the braincase relative to the holding side jaws. This motion is produced by various muscles (Figs. 9.13 and 9.14), including the neurocostomandibularis, cervicomandibularis, retractor palatini (pterygoidei), and some holding side mandibular adductors. Anterior trunk muscles also play a role at this time in many colubroids (Moon and Kley, 1997; Moon, 1998), drawing the neck over the prey and probably also stabilizing braincase orientation on the anterior trunk. Popular notions that the supratemporal, quadrate, and mandible are all drawn ventrally during advance to enlarge the gape have yet to be recorded for any medial transporter. What usually happens is that the quadrate mandibular joint is protracted in approximate line with the caudal end of the pterygoid to which it is closely attached in many colubroids. The net result is that the quadrate actually rises at its dorsal end along with the rear end of the supratemporal. Transporting very large diameter prey apparently depends (good empirical data are lacking) on sliding both the mandible and the palatopterygoid bar forward without separating either from the prey surface. What little evidence exists suggests that even for very large prey, the quadrate mandibular joint is pulled laterally but not ventrally. At the end of the advance phase, a burst of activity of the pterygoideus muscle initiates the fast closing phase [Cundall and Cans (1979), equivalent to late advance and early closing phases of Kardong et al. (1986)]. Contraction of the pterygoideus pulls the maxilla caudally
316
D a v i d C u n d a l l a n d H a r r y W. G r e e n e PF
FR
PA
AES
AEM
DM
CM
NCM
Cerv. Quad. FIGURE 9.13. Lateral view of the superficial head muscles of Agkistrodon piscivorus (Viperidae). AEM, adductor externus medius (AEP of Zaher, 1994b); AEP, adductor externus profundus (AEM of Zaher, 1994b); AES, adductor externus superficialis; Cerv. Quad., cervicoquadratus; CGI, compressor glandulae; CM, cervicomandibularis; DM, depressor mandibulae; FR, frontal; MX, maxilla; NCM, neurocostomandibularis; PA, parietal; PF, prefrontal; Pt, pterygoideus. In this and Fig. 9.14, open and closed circles on the mucles show positions of electrode tips yielding electromyographic data from 13 different snakes. Open circles represent tips deeply implanted, and closed circles show sites of tips near the surface of the muscle.
(F in Fig. 9.10b) and the quadrate mandibular joint rostrally. Cessation or reduction of activity in the levator pterygoidei but continued activity of the protractor
pterygoidei often marks this short phase as w^ell. Combined muscle activity patterns (Fig. 9.16) usually flip the palatopterygoid joint medially and ventrally onto
FIGURE 9.14. Ventral views of (a) ventral constrictors and (b) dorsal constrictors of Agkistrodon piscivorus (Viperidae). Abbreviations as for Fig. 9.13, including APP, adductor posterior profundus; CC, constrictor colli; lAA, intermandibularis anterior, pars anterior; lAP, intermandibularis anterior, pars posterior; IPA, intermandibularis posterior, pars anterior; IPP, intermandibularis posterior, pars posterior; PT, pterygoid bone; PtA, ptergoideus accessorius.
317
9. Feeding in Snakes the prey surface while the mandible is drawn forward to the maximum extent. Mandibular adductor activity then rises rapidly, lifting the mandible and clamping the jaws onto the prey near their points of maximum advancement relative to the braincase (Figs. 9.15 and 9.16).
For smaller prey, continued slow closure of the active side drags that side of the prey into the mouth. For larger prey, tooth penetration simply locks the active side jaws on the prey surface. If they appear to retract, they often move only the skin of the prey. Effective ''re-
el
W X y
Z
FIGURE 9.15. EMG records for an Agkistrodon piscivorus (Cundall, unpublished data) during intraoral transport of a mouse (top, a-d) and a fish (bottom, a - d ) showing modulation of transport kinematics and muscle activity in response to prey type. Tracings from cine records synchronized to EMG records (synchronizing photocell trace at bottom of EMG record above time trace set at 1-sec intervals) show the kinematic profile of the snake's head at the times shown by equivalent vertical lines on EMG traces. Both sets of recordings were made on the same day within an hour of each other. Tracings for transport of the mouse were made from cine of a mirror image, whereas those for fish were from a direct view; hence, both show right-side movements. Behavioral phases shown are (a) beginning of advance, (b) end of fast opening, (c) end of fast closing, and (d) end of slow closing. Lines w, X, y, and z are not illustrated but show the same behavioral points for the left side. Muscle records for the bottom are in the same order as the top and include R AES, right adductor externus superficialis; R CM, right cervicomandibularis; R AEP, right adductor externus profundus (medius of Zaher, 1994b); L DM, left depressor mandibulae; L NCM, left neurocostomandibularis; L CGI, left compressor glandulae. Note reduced fang erection and adductor activity, reduced duration of the hold phase between right and left cycles, changed pattern of NCM activity, and increased duration of slow closing (c-d) during transport of the fish.
318
D a v i d C u n d a l l a n d H a r r y W. G r e e n e 2
.
3
,
4
.
5
.
6
.
b
1
FIGURE 9.16. Generalized EMG patterns for medial transport [upper half of each trace based on data of Cundall and Gans (1979) and Cundall (1983); lower half of each trace (where present) based on data of Kardong et al (1986)]. Available data for lateromedial transport in Python (Kardong and Berkhoudt, 1998) suggest that patterns for AES, AEM, AEP, Pt, and DM are similar to those of Cundall and Gans (1979) and Cundall (1983) for medial transport, (a) Includes muscles innervated by V3 and (b) muscles innervated by V4, VII, and anterior spinal nerves. Activity patterns variably present (only some individuals or during only part of transport) are shown in dashed outline. AEM, adductor extemus medius [using Zaher (1994b) designation]; AEP, adductor externus profundus [also after Zaher (1994b)]; AES, adductor externus superficialis; CGI, compressor glandulae; CM, cervicomandibularis; CQ, cervicoquadratus; DM, depressor mandibulae; lAP, intermandibularis anterior, pars posterior; IPA, intermandibularis posterior, pars anterior; IPP, intermandibularis posterior, pars posterior; LP, levator pterygoidei; NCM, neurocostomandibularis; PP, protractor pterygoidei; PQ, protractor quadrati; PS, pseudotemporalis; Pt, pterygoideus; RP, retractor pterygoidei.
traction" of the active side usually does not occur until the opposite side is protracted and the braincase is carried forw^ard as part of that cycle. The slow closing phase is often marl^ed by activity in some ipsilateral mandibular adductors (Cundall, 1983; Cundall and Gans, 1979; Kardong et ah, 1986) as well as the neurocostomandibularis and sometimes the cervicomandibularis (Figs. 9.13-9.16). Activity of the latter muscles usually becomes stronger toward the end of prey transport and often shifts in timing to occur during both ipsilateral closing and contralateral advance (Kardong et al, 1986; also see Fig. 9.15, bottom). Our functional divisions of transport are clade-based characterizations. The critical difference between lateromedial transporters vs medial transporters is the behavior of the palatopterygoid bar relative to the maxilla.
In colubroids, elongation of the ectopterygoid, and migration of the pterygoideus origin to the distal end of the ectopterygoid and caudal end of the maxilla (Figs. 9.3c, 9.3d and 9.10b), produces a complex array of potential palatomaxillary movements during medial transport (Cundall and Greene, 1982; Cundall, 1983). These movements differ most from lateromedial transport during the shift from fast open to fast close. In lateromedial transporters, as the pterygoideus fires, both the maxilla and the palatopterygoid bar tend to move in unison and there is little or no differential movement of the two tooth rows in the longitudinal plane (Fig. 9.10a). In medial transporters, firing of the pterygoideus is often matched by continued firing of the ipsilateral protractor pterygoidei but cessation of levator pterygoidei activity. The caudal end of the maxilla and associated
9. Feeding in Snakes ectopterygoid are yanked caudally, but the pterygoid does not move backward. Instead, the palatopterygoid joint moves medially and ventrally, dropping the pterygoid tooth row onto the prey surface at its point of maximum protraction. The complex fast-closing behavior of medial transporters solves a significant functional problem—how to attach the teeth of the active side on the prey when the jaws are at the point of maximum advance. Inasmuch as this point lies far removed from the resting position for all moving parts, any relaxation of the muscles that produce advance would lead to instant elastic recoil. This may partly explain the continued firing of some or all of the protractor muscles and the strong firing of the pterygoideus, but not many of the other adductor muscles, at the beginning of fast closing (Cundall, 1983). While the active side is advancing and closing, the other side of the head remains clamped onto the prey. This usually involves activity of some or all of the mandibular adductors on the clamping jaw, typically at lower amplitude and often skewed to the advance and fast-closing phases of the advancing jaws (Cundall, 1983; Kardong et al, 1986). These patterns of muscle activity are reminiscent of slow motion chewing in mammals with working and balancing side activity patterns (e.g., Hiiemae and Crompton, 1985). 4. Mandibular
Transport
Boltt and Ewer (1964) and Cundall (1983,1987) suggested that the mandible does little during intraoral transport, a view tempered by Kardong (1986b) who pointed out that the mandible in many cases is not simply passively dragged along but is actively brought over the prey, possibly helping to shape it. When feeding on smaller prey, the mandible appears to play a major role in transport in most snakes, although no one has yet successfully partitioned mandibular and palatomaxillary contributions to transport. However, among the few natural deaths of snakes recorded in the field (not due to freezing or predation), one was of a starving Bothrops asper whose only visible musculoskeletal damage was a fractured mandible (Greene and Hardy, 1989). Asian pareatines {Aplopeltura, Pareas, Brongersma, 1956, 1958) and neotropical dipsadine xenodontines {Dipsas gaigea, Harris and Simmons, 1967; Sibynomorphus neuwiedi, Laporta-Ferreira et al, 1986,1988; Dipsas indica, Sazima, 1989) use a snail extraction method based entirely on mandibular transport, a process that could be termed the mandibular walk. Whether the mandible moves independent of the palatomaxillary arch is unclear. Sazima (1989) described mandibular movements during feeding on both slugs and snails.
319
with the mandible protracting and retracting up to onethird of its length, movements that must occur partially independent of the palatomaxillary arch. In some species of Dipsas, Pareas, Aplopeltura, and Sibynomorphus there are either no anterior palatomaxillary teeth or no anterior maxilla and the remaining palatomaxillary teeth are reduced (Brongersma, 1956,1958; Peters, 1960; Kofron, 1985). The upper jaw passes over the shell to serve as a brace during extraction and the maxillary teeth apparently rotate inwardly (Cans, 1972). Snail extractors and slug eaters have extremely long, comb-like teeth on one or more of the dentigerous bones—on the dentary in mandibular-based snail extractors and on both maxilla and dentary in slug eating species of Dipsas, Sibon, and the unusual North American Contia tenuis (Zweifel, 1954). Slugs and snails produce copious quantities of mucus when attacked, potentially immobilizing small predators (Arnold, 1982), and snails have been known to pull a snake's head (Duberria lutrix) into the shell, suffocating the snake (FitzSimons, 1962)! Hence, a strategy in which the upper jaw is lifted free of the snail body to pass over the edge of the shell while elongate mandibular teeth remain embedded in the snail appears to be an improvement, despite the location of the glottis on the lower jaw. It remains to be demonstrated how the relatively fragile mandible of the specialized snail eaters actually functions and what patterns of muscle activity are used to extract snails. Two other colubrid genera, the African lamprophiine, Duberria lutrix (FitzSimons, 1962) and the North American thamnophiines Storeria dekayi and S. occipitomaculata (Rossman and Myer, 1990), also extract snails. Exactly how Duberria does so is unclear, although its method, as noted earlier, is not always successful. Storeria uses transport typical of natricines but prey manipulation prior to transport is unusual for its duration and tonicity. Storeria maintains torsion on the snail body until the columellar muscle fatigues, a process that may take 10 min or more. The snail is then pulled from its shell, typically with both upper and lower jaws clamped onto the body. The snail shell must be wedged against the substrate for this technique to work (Rossman and Myer, 1990), a limitation not shared by dipsadine snail extractors. Not all colubroids use pterygoid or mandibular walks to transport prey. The exceptions pose fascinating questions about the coopting of motor patterns for novel kinematic strategies. In the African egg-eating snakes, Dasypeltis scabra, for example, eggs are transported essentially by pushing the head over the egg (Cans, 1952,1974). In these snakes, teeth are lost except for a few small teeth on the posterior ends of the maxilla and dentary. The mucosal connection between the edge of the dentary and the lower lip is highly folded and the
320
David Cundall and Harry W. Greene
intermandibular muscles are arranged more longitudinally and are quite long. Whereas the skin forms an extraordinary arrangement of elastic components along the lateral side of the neck, there is no mental groove and the chin shields and lower labials are connected tightly to make a relatively stiff, shovel-like blade between the tips of the dentaries. This stiffened skin area is shoved under the egg and curls around it slightly as the dentaries reach the widest point. The key to swallowing eggs in Dasypeltis is thus to separate the skin of the chin from the constraining skeleton and drag it gradually around the egg while the rest of the skull is pushed over the egg by the trunk. Gans (1952) described a rigid connection between the quadrate and the supratemporal but a mobile joint between supratemporal and braincase, allowing the quadrate mandibular joint to move laterally and ventrally, enlarging the gape. However, in the only film available to us and in most pictures of egg eating by Dasypeltis, there is little evidence of ventral movement of the quadrate mandibular joint. Alternating movements of the palatomaxillary arches are not used during transport—instead, the entire upper jaw simply slides over the egg, usually by being pushed forward and downward. A critical element of the process is the ability of the snake to curve around the exposed part of the egg after the mandibles reach the midpoint—a feat accomplished very rapidly by the stiff ventral blade that is somehow moved anteriorly far removed from the tips of the mandibles and then lifted at its front end (Gans, 1952). Eating of bird eggs in other colubrids is a process very different from that in Dasypeltis. Elaphe climacophora transports eggs using typical jaw advance mechanisms, and thus transporting eggs does not necessarily require extraordinary morphological or functional innovations (Gans and Oshima, 1952). However, the presence of elongate or otherwise modified hypapophyses on anterior vertebrae in some species that eat bird eggs indicates that a critical limiting factor is an ability to break the shell inside the snake (MuUin, 1997), a factor apparently of less importance for species that feed on squamate eggs with less mineralized shells (e.g., Oligodon, Coleman et ah, 1993). 5. Suction
Mechanisms
In scolecophidians the geniomucosalis muscle (which extends fron\ the mandible caudally to attach over much of the ventral oral mucosa) might create suction (Groombridge, 1979b), as suggested by Smith (1957) for Leptotyphlops goudotii. This is not impossible despite terrestrial conditions. The mouth and pharyngeal cavities of scolecophidians are small enough that insect bodies or pupae might fill them in the same man-
ner that a plum or grape can be held in human lips. Creating suction behind the prey might rapidly draw it in to the point where peristalsis could take over. In Leptotyphlops, a second muscle (muscle X) extends from the anterior supracostal muscles to the dorsal oral mucosa. The arrangement of these muscles suggests the possibility of moving prey by active protraction and retraction of the oral mucosa or by suction (Groombridge, 1979b). Kley's (1998) video data, however, show a jaw-based ratchet system in the taxa he examined. The marine elapids Aipysurus eydouxi and Emydocephalus, species known to feed almost exclusively on fish eggs, have a geniomucosalis muscle (possibly a dorsal division of the geniotrachealis) similar in most respects to that of scolecophidians (McCarthy, 1987). The behavior of these snakes while foraging combined with their oral morphology suggests that they might suck fish eggs out of nest burrows. Laticauda colubrina, another aquatic elapid, presents a very different kind of transport problem that is mirrored to varying degrees by some other snakes that eat either fish or worms. Laticauda first envenomates an eel and then releases it and waits until it is motionless before beginning swallowing. Radcliffe and Chiszar (1980) saw no detectable palatomaxillary or mandibular movements, but transport occurred with incredible rapidity. The mechanism remains an enigma but may involve esophageal peristalsis like that during worm eating in some caecilians (Bemis et ah, 1983). E. S w a l l o w i n g Although swallowing tends to follow transport in most vertebrates (some teleosts may be exceptions), these processes occur together in most snakes. Depending on the clade, the tongue may play a role in both functions (Bramble and Wake, 1985; Hiiemae and Crompton, 1985). In most alethinophidians and a few other elongate limbless clades, a prey item enters the esophagus while the mouth continues to transport it. The tongue plays no role in either function. In many alethinophidians the anterior trunk begins undulatory movements characteristic of posttransport swallowing when the prey reaches the esophagus. Kley and Brainerd (1996), using videofluorography, and Moon and Kley (1997) and Moon (1998), using video and EMG analysis, showed that the concertina-like trunk movements seen in Pituophis aid in transport, and this is likely to apply to most snakes showing this kind of behavior. One of the morphological features of snakes that underlies all of their feeding biology is the anatomy of the trunk between the head and the stomach (Fig. 9.17). Gans (1961) pointed out the importance of the loss of
321
9. Feeding in Snakes
{Python molurus, and Thamnophis sirtalis) Hoxc-6 (which specifies the boundary between cervical and thoracic regions) and Hoxc-S (typically expressed only in the posterior thorax) are expressed right behind the head (Cohn and Tickle, 1999; A. C. Burke, personal communication). Morphological arguments in favor of this interpretation are many, but those with important functional correlates are that ribs, hypaxial muscles, and the pleuroperitoneal cavity all extend to the head. Combined with the loss (not reduction) of the pectoral girdle and median thoracic skeleton (sternum), the anterior snake trunk has no skeletal constraints on body circumference, a feature augmented in some snakes by reorganization of the anterior keratinized skin. The folded area of the hinge region between the lateral scales of the anterior trunk is increased so that the maximum stretched circumference is equivalent to or greater than that of the caudal trunk, a feature obvious during swallowing of large diameter prey (type III or IV; see Section IV,A). F. Digestion and Defecation
F I G U R E 9.17. Cross-sectional view of the anterior snake trunk showing (a) absence of ventral skeletal elements and (b) with prey in esophagus to show functional implications. Critical elements in the organization of the anterior trunk are modifications of the skin to provide increased potential stretch and apparently highly stretchable skeletal muscles connecting the ribs to the skin externally and wrapping the coelom (transverse and oblique abdominals) internally. Note that with prey in the esophagus, most of the snake's skeleton (black) and its axial musculature (gray) lie on top of the prey. Esophagus shown with folded wall in (a) with hollow trachea and solid tongue below it, both of which are flattened in (b). Stippling collectively represents the dermis and the cutaneous and deep hypaxial muscles.
the pectoral limb girdle and snakes have long been assumed to represent the evolutionary product of limb loss. It now seems more likely that snakes are the result of changes in axial segmental patterning in which the neck and anterior thorax regions are absent, possibly as a result of the anterior expression of Hox genes associated with that region of the thorax lacking connections to sternal elements. In the two snakes analyzed
Snakes encompass a variety of foraging strategies that are presumably correlated with prey selection. Some species that feed very irregularly on large prey have a remarkable digestive physiology. Exploiting very large prey necessitates efficient digestion over a short time to reduce the probability of putrefaction of the prey. Between feedings, however, energetic costs of body maintenance are reduced by "turning off" the intestine, reducing nutrient use by tissues, which regress in size and cell function (Secor and Diamond, 1997). This "turning off" is an active regulatory process that immediately follows completion of digestion and is matched by a similar regulatory activation immediately following prey ingestion (Secor, 1995; Secor and Diamond, 1995; Secor et al, 1994). The up-regulation of digestion in Python is accompanied by increases in metabolic rate and tissue genesis that rank among the highest recorded for any vertebrate. Similar physiological modifications probably characterize other snakes that exploit very large prey infrequently (Secor and Diamond, 1998). Hence, sit-and-wait, or ambush predation, which is associated with large gape size and extraordinary prey restraint mechanisms (envenomation or constriction), is also coupled to regulatory cycling of the gut. Ambush predation in some ground-dwelling viperids and booids appears to be loosely correlated with an unusual pattern of defecation in which feces may be retained either until the next feeding or, in some cases, for long periods that seem independent of feeding schedules (Lillywhite et ah, 1998). Because closely related
322
David Cundall and Harry W. Greene
arboreal species show much shorter fecal passage times, Lillywhite et al. (1998) suggested that fecal retention has little relation to digestive physiology but may benefit terrestrial species by increasing the inertia of the rear part of the trunk that serves as the base from which the head and anterior trunk are launched during the strike. Fecal retention may also produce changes in posterior body shape and mass similar to those characteristic of gravid females and perhaps be associated with anatomical features of the caudal trunk similar to those underlying sexual dimorphism in some snakes. Whether fecal retention and other aspects of feeding share any structural correlates with reproduction will only be shown by careful comparisons among trunk regions.
III. PERFORMANCE A N D SIZE Being able to eat large prey and also being limited to eating these prey whole has provided a set of simplifying conditions that make for ripe ecological picking. There is clearly a relationship between the size of a snake's head and the size of prey it exploits—the larger the snake, the larger the prey (Forsman, 1991; Shine, 1991a; Forsman and Lindell, 1993). Flowever, this relationship could have various allometric properties (Rossman, 1980) that remain largely unexplored for most snakes. Head size is also sexually dimorphic in some snakes (e.g., Camilleri and Shine, 1990; Shine, 1991b) and correlates with diet in some of the cases examined carefully. Arnold (1993) reviewed many of the studies relating prey size to snake size and found that there is a general trend among alethinophidian snakes for prey size to roughly track predator size. This is true also during ontogeny, which means that some (possibly most) species show marked shifts in prey preference during ontogeny (e.g., Mushinsky et al, 1982; Plummer and Goy, 1984) and drop smaller items from their diet as they grow. Although energetic costs of capture and transport in Thamnophis elegans (Feder and Arnold, 1982) are very low in comparison to the energetic gains from normal salamander prey, ingestion times increase dramatically as relative prey diameter increases and the snake reaches the so-called breaking point of its potential gape size (Pough and Groves, 1983; Shine, 1991a). B. E. Dial (personal communication) found that energetic costs of transport in Pituophis remained relatively similar through a large range of prey sizes but increased dramatically as prey mass and diameter neared estimated maximum gape size. This relationship may apply to most snakes, suggesting that handling costs are minimal for most prey within broad relative size ranges. Arnold's (1993; see also Shine, 1991a) hypothe-
ses concerning why some snake species drop small prey from their diets (hard-to-find, hard-to-catch, and hardto-eat prey, low-energy or marginal-energy prey) remain largely untested. Observation of captive booids shows that catching and killing prey becomes increasingly problematic as relative prey size diminishes (personal observation, but see Shine, 1991a), although detailed quantitative analyses of this relationship remain to be done. Many mammals or other active, potentially dangerous prey too small to constrict may thus be removed from the diets of constrictors, and perhaps constrictors are more limited by small prey sizes than any other group of snakes.
IV. EVOLUTION A. Ecological Patterns Snake diets collectively encompass an amazing range of animals, including annelids, crustaceans, arthropods, molluscs, and all major vertebrate groups. Ontogenetic shifts often complicate assessments of diet specialization (see Chapter 8). Most such changes commonly involve predation by small individuals on frogs, lizards, and other ectotherms and a shift in adults to an exclusive diet of rodents and other endotherms (e.g., Greene, 1989a; 1992). Nevertheless, particular diets characterize individual snakes, populations, species, and higher taxa along a spectrum ranging from extremely broad to very narrow. At the generalist extreme, one individual Drymarchon corais (Colubridae) contained two frogs, three small turtles, and three mice, whereas another had eaten a toad, a hatchling tortoise, and two snakes (Ruthven, 1912; Mount, 1975); the diet for just one population of that species thus might encompass dozens of prey species with diverse handling characteristics. At the specialist extreme, all Dasypeltis (Colubridae) eat only bird eggs (Gans, 1974) and each species of Xenocalatnus (Atractaspididae) feeds on a single species of amphisbaenian (Broadley, 1963). At least under some circumstances, narrow diets should favor specific alternative phenotypes because of the demands of finding and eating particular prey types. This section reviews some of the better documented associations between prey characteristics and particular feeding mechanisms. These specializations are placed in an historical evolutionary context in the next section. Early discussion of snake feeding often stressed predation on relatively 'Targe" animals, but failed to specify that an item might be large in either or both of two ways. Each measure of relative prey size potentially affects feeding performance and is substantially correlated with diversification in snake biology, and together
9. Feeding in Snakes the two factors define four extremes (Greene, 1983, 1997). Type I prey are relatively small in mass and diameter, regardless of shape, such that they require modifications neither for high potential handling costs nor for large gape. Type II prey are elongate, often limbless, and relatively heavy, thus increasing potential handling cost because, all else being equal, a relatively heavy adversary is more difficult to subdue than a lighter one. Because their relative diameter is small, type II prey do not require greatly increased gape (Fig. 9.18). Type III prey are fusiform or roughly spherical as well as relatively heavy so they necessitate compensation for both high handling cost and increased gape (Fig. 9.19). Type IV prey weigh substantially less than predicted by their maximum diameter, either because their cross-sectional shape is not circular (e.g., many fishes) or they possess protruding body parts (e.g..
323
wings of birds and bats) so they require large gape but not specialized immobilization mechanisms. Living anguimorphs, including elongate limbless forms, typically feed frequently on relatively tiny type I prey [e.g., varanoids (Pregill et al, 1986; Losos and Greene, 1988); Anniella (Miller, 1944)]. As predicted, many snakes consuming types II (e.g., eels) and III prey (e.g., some lizards, rodents) are constrictors, venomous, or both; because heavy items are nutritionally better than small items, predators on types II and III prey feed infrequently compared to those species taking types I and IV. Snakes that eat earthworms (e.g., Carphophis and Geophis) are typically nonconstrictors and nonvenomous, whereas those taking elongate vertebrates either constrict (e.g., Cylindrophis and Lampropeltis getula) or are venomous (e.g., various fossorial elapids), suggesting that among type II prey there are differences
FIGURE 9.18. Striped crayfish snake {Regina alleni, Colubridae; 42 g) and a harlequin coral snake {Micrurus fulvius, Elapidae; 60 g, total length 73 cm) that had consumed it in nature (both now CU13430 in the Cornell University Vertebrate Collection). This relatively heavy, elongate prey item (type lib, prey/predator mass ratio ca. 0.70) was presumably subdued by neurotoxic venom. The only evidence of digestion is at the prey's anterior end, which had been lodged in the pyloric end of the stomach.
324
David Cundall and Harry W. Greene
^iiiiii.
FIGURE 9.19. California ground squirrel {Spermophilus heecheyi, 195 g) and a western rattlesnake {Crotalus viridis, Viperidae; 392 g, total length 87 cm) that had consumed it (both now MVZ 5326, in the Museum of Vertebrate Zoology, Berkeley). This relatively heavy, bulky item (type III, prey/predator mass ratio ca. 0.50) illustrates the rapid digestive action of venom; the rodent's snout is intact, despite having been lodged near the pyloric end of the stomach, whereas the abdominal body wall is open, exposing the viscera (see also Thomas and Pough, 1979).
between annelids (type Ila) and vertebrates (type lib) in their mass-specific struggling abilities. Distinctive functional demands imposed by prey mass and prey shape might elucidate the ecological roles of major venom types; regardless of shape, heavy prey (types II and III) presumably are subdued by im-
mobilizing toxins, but snakes that take heavy bulky items (type III) may also profit from deeply injected digestive components. Snake venoms are complex mixtures that vary extensively among and within higher taxa, but short-fanged elapids tend to have neurotoxic venoms and long-fanged viperids usually have tissuedestructive venoms (Greene, 1997, and references therein). In accordance with predictions from prey mass and prey shape, at least some basal and morphologically primitive elapids feed mainly on other snakes (type lib. Fig. 9.18; Greene, 1984; Keogh and Smith, 1996), whereas most viperids eat mammals (type III, Fig. 9.19; Greene, 1992). Taxa for which types III and IV predominate in the diet should require adaptations for increased gape compared to those that eat types I and 11. Based on morphological surveys in Marx and Rabb [1972; see Greene (1997) for dietary summaries and other references], snakes that eat fish regularly (e.g., Gmyia), toads {Heterodon), birds (e.g., some Boiga), and rodents (e.g., viperids, the elapid Acanthophis) indeed have longer quadrates and/or mandibles than species feeding on small and/or elongate items. Conversely, relatively very short quadrates are found in snakes that eat eels and elongate amphibians (e.g., Farancia), worms (e.g., Carphophis and Geophis), centipedes (e.g, Tantilla), and other snakes (e.g., Miodon and Micrurus). Beyond matters of relative mass, struggling ability, and shape, various surface and internal features of particular prey might affect capture, ingestion, and processing by snakes. Snakes that specialize on mucuscovered invertebrates [e.g., goo-eating dipsadines (Savitzky, 1983; Salmao and Laporta-Ferreira, 1994; Sazima, 1989)] have long needle-like teeth, whereas those that cope with the hard chitonous exoskeletons of arthropods have short, sometime spatulate teeth [e.g., sonorine colubrids (Savitzky, 1983)]. The only snakes that eat relatively large, incompressible bird eggs have numerous specializations for gape and for internally crushing the shells (Gans, 1974; Savitzky, 1983; see earlier discussion), whereas species that feed on softer, compressible eggs of squamates slice those items with blade-like dentition [e.g., Prosymna (Broadley, 1979); Simoselaps (ScanIon and Shine, 1988)]. The slippery scales of skinks pose special problems for gripping and capture, which have been circumvented in Xenopeltis and several groups of colubrids (e.g., Psammodynastes) by ratchet-like or tonglike grasping dentition (Savitzky, 1981, 1983; Greene, 1989b) and by a combination of snout depression and tooth row modifications in bolyeriids (Cundall and Irish, 1989). The skin secretions of some amphibians are so toxic that only snakes with specialized resistance can eat them without dying [e.g., populations of Taricha (newt)-resistant Thamnophis sirtalis (Brodie and Brodie,
9. Feeding in Snakes 1990); Heterodon and Hemachatus that eat mainly toads (Macdonald, 1974)]. The extent to which locomotor specializations for particular macrohabitats might have influenced and been affected by feeding biology are thus far poorly explored in snakes. Fossorial snakes sometimes have consolidated skulls and other modifications that apparently facilitate burrowing and of necessity reduce the potential for increased gape (Wake, 1993); often those same species eat type II prey as well, such that particular character states, such as a short quadrate bone, might be functionally related to feeding, locomotion, or both. Some arboreal species might be constrained by their slender build to take lighter items and/or process food more rapidly than terrestrial species (Lillywhite and Henderson, 1993). B. Historical Patterns Combining general patterns of prey selection explored earlier with current hypotheses of snake relationships (Fig. 9.1), interesting correlations emerge suggesting sequential appearance of apomorphies characterizing higher alethinophidians. The basal split of scolecophidian and alethinophidian clades correlates with exploitation of two abundant but different invertebrate prey resources, short type I prey (such as termites and possibly other short invertebrates, like pupae and larvae of various soft-bodied soil arthropods) by all but one scolecophidian species (Webb and Shine, 1993), and long type Ila prey (earthworms) by basal alethinophidians (uropeltids, possibly Anomochilus). If the ancestor of snakes was small and fossorial, this is not an unreasonable scenario, assuming that both invertebrate clades were abundant members of Mesozoic terrestrial faunas. Available evidence suggests that ants were not (Holldobler and Wilson, 1990), but other arthropod communities could have sufficed. Despite few fossil records of terrestrial oligochaetes, they are assumed to have been widely distributed and prevalent since the Paleozoic (e.g., Bouche, 1983). The oldest fossil that is clearly a snake and provides evidence of skull structure remains Dinilysia, a Cretaceous alethinophidian that provides no clues about earlier ancestral feeding mechanisms (Estes et ah, 1970; Frazzetta, 1970a). We assume that early snakes were gape limited predators like living scolecophidians and basal alethinophidians, and thus the evolution of feeding in snakes commences with a hypothetical ancestor to these two very divergent clades. Phylogenetic analysis suggests that this ancestor had no palatal teeth, limited palatal mobility focused at the caudal end, a short quadrate liberated at its ventral end and liberated dorsally by loss of the overlying bones, a loss seen otherwise only in fos-
325
sorial dibamids and amphisbaenians (Cundall, 1998). We also assume the ancestor had a flexible but nonstretchable mandibular symphysis like that of scleroglossan lizards, allowing flexure but little or no separation of the mandibular tips. Snout-braincase joints probably remained lizard-like with most attachments occurring at the periphery. Loss of the posterior maxillary attachments through reduction of the postorbital and postfrontal bones is also assumed to have occurred in the fossorial common ancestor inasmuch as this is a common reduction in all fossorial squamates. Combined with reductions of attachment surfaces between the maxilla and prefrontal, the caudal ends of the maxillae may have become spreadable. The three scolecophidian families represent highly derived solutions to the exploitation of small prey (type I). From both anatomical and functional perspectives, these families are dissimilar to each other and to all other snakes. Their sister clade status arises from synapomorphies of head and body structure largely unrelated to feeding (Haas, 1964, 1968; List, 1966; Kluge, 1991; Cundall et ah, 1993), an unusual occurrence among snakes for which the trophic apparatus defines most major clades. The rapid (2-4 Hz) jaw movements of typhlopids and leptotyphlopids (Kley, 1998, personal communication) are so unlike feeding mechanics in alethinophidian snakes that points of comparison are hard to find. The evolution of the alethinophidian feeding apparatus is keyed to changes in the snout, palate, suspensorium, and intermandibular soft tissues. Anomochilids and uropeltids show loosening of the ventral snout but tight attachment of the toothless palatine to the vomer. Their quadrates are short and slightly anteriorly directed, with the dorsal edge attached to the lateral wall of the otic capsule and the posterodorsal corner connected to the stapes. Uropeltids (and probably anomochilids) are active burrowers, feeding exclusively on earthworms (type Ila) and other soil invertebrates. Jaw kinesis is anticipated to be limited to lateral movements of the posterior maxilla, with transport depending on rapid side-to-side movements of the head (Rajendran, 1985). Cylindrophiids and aniliids exploit elongate, heavy vertebrate prey (type lib). Constriction appears in cylindrophiids along with strongly developed palatal teeth and increased palatomaxillary mobility based on snout shifting. The suspensorium and mandibles remain little changed. Flexibility of the intramandibular joint may have increased, but in essence it represents a plesiomorphic feature typical of a variety of scleroglossans. A general trend in these clades is an increase in overall size and probably reduced fossoriality, or restriction to less compacted soils.
326
David Cundall and Harry W. Greene
Xenopeltids and loxocemids are the first clades to show liberation of the caudal end of the supratemporal. However, the supratemporal is still relatively short and provides little effective increase in suspensorial or mandibular length. Anterior intermandibular tissues may show increased elasticity and the palate is liberated through the flexibility of palatal attachment to the vomer and loosening of the snout. Significantly, both clades exploit prey of various shapes and sizes, including small endotherms (type III). Moreover, Xenopeltis has hinged teeth (Savitzky, 1981), a dental specialization that reappears independently in a variety of colubroid clades. Basal macrostomatans encounter prey by active searching for good ambush sites and kill their prey by constriction. Data from boas, pythons, and other living basal taxa imply that Macrostomata is the first point in snake evolution for which diverse bulky prey (types III and IV) form a regular part of the diet, and for which diverse locomotor modes and macrohabitat preferences become possible (Greene, 1983, 1997; Murphy and Henderson, 1997; Rodriguez-Robles et al, 1999). Functional diversification of the feeding apparatus is modest in basal macrostomatans, compared to Colubroidea, but does include enlarged anterior teeth in many taxa (Kluge, 1991, 1993b) and the divided maxillae of bolyeriids (Cundall and Irish, 1986,1989). Maxillary liberation coincident with the origin of the Colubroidea evidently provided a greatly increased potential for diverse feeding adaptations. Only colubroids exhibit diverse morphological specializations for prey types that pose particular handling difficulties; with rare exceptions, only among advanced snakes are found dental and other modifications associated with overcoming the surface characteristics of soft-bodied molluscs (Laporta-Ferreira et ah, 1986, 1988; Sazima, 1989; Rossman and Myer, 1990), hard-bodied skinks (Savitzky, 1983; Cadle, 1999), bird eggs (Cans, 1952), or crabs (Savitzky, 1983). Such items are correspondingly rare or absent from the diets of more basal snakes. Venoms have apparently arisen only with or subsequent to the origin of colubroids and, the legendary feats of pythons notwithstanding, only certain venomous viperids and elapids take prey much exceeeding their own mass (Greene, 1984,1992). Decoupling the maxilla and palatopterygoid bar allowed the colubroid maxilla to respond to selection pressures associated with prey capture while the palatopterygoid bar maintained transport functions. In combination with the accumulated arsenal of prey handling tools inherited from their ancestors (large gape, shock absorbing suspensorium), the elaboration of posterior dental (rictal) glands (Kochva, 1978; Underwood, 1997), and the refinement of duct and tooth conduits for
carrying venom to prey tissues, basal colubroids underwent an explosive radiation, the results of which crawl all around us. V. C O N C L U D I N G REMARKS We built our evolutionary scenario on a proverbial house of cards. We currently have little data on feeding mechanics in limbless anguimorphs and our knowledge of scolecophidian feeding remains fragmentary. The gape-limited nature of basal alethinophidian lower jaws needs thorough exploration. Current behavioral predictions need testing in the laboratory and we need new methods for examining feeding mechanics in the field. It is clear from existing data that prey capture behaviors change quickly under captive conditions and this may occur even if natural prey are offered under simulated natural conditions. To date, there are no studies on the effects of captivity on kinematic patterns of feeding in snakes. Apart from measuring the relation between field and laboratory behavior, much more can now be done in the laboratory. New video capabilities and improved sonomicrometry hardware will offer the potential of measuring precise bone displacements in three dimensions, allowing answers to innumerable questions about the exact contributions of the individual suspensorial elements and the motions of the snout and upper jaws. High-speed video makes quantitative studies of prey capture affordable and feasible, even for species notoriously difficult to observe. The increasing number of ongoing kinematic and behavioral studies bodes well for the future. To really understand feeding in these animals so different from us, we need new, imaginative approaches for studying smaller species in the field and continued collection of quantitative data on prey taken by snakes. Connections now being made among diet, habitat, population structure, social behavior, and distribution need to be tied more extensively to functional morphology and phylogeny. The vast majority of snakes are small and their success will not be explained by continued examination of only moderate-sized members of a few clades. Finally, we need to know more about the anatomy of snakes to understand the proximate basis of their feeding behavior. What structures limit gape size, how are the components of the trunk wall rearranged during swallowing, and what neurological patterns underlie motor coordination of intraoral transport striking are among the many questions needing anatomical details currently unavailable for any snake. Comparative anatomical analyses will also contribute significantly to the
327
9. Feeding in Snakes production of more robust phylogenies, a requisite for improving our understanding of the evolution of the remarkable feeding apparatus of snakes. Notes A d d e d in Proof Assuming a mosasauroid ancestry of snakes and using measurements of fossils, Lee et al (1999) suggested that mosasaurs could flex the middle of the mandible laterally ll-?>7° to increase gape size (see p. 294). They further suggested that this behavioral capability, similar to that of Cylindrophis, was the critical earliest step in the evolution of snakes. Both the phylogeny and the functional steps proposed refocused the problem on the mandible rather than on the upper jaw and built on an aquatic feeding scenario (Lee and Caldwell, 1997) developed more fully by Scanlon et al (1999). The model suggests that gape enlargement, and the ability to ingest large diameter prey, occurred prior to liberation of the upper jaw to power transport. Recent analysis of another Cretaceous snake-like fossil with more complete hind limbs {Haasiophis terrasanctus), combined with reanalysis of Pachyrachis prohlematicus, indicates that both fossils from the Middle East are macrostomate snakes (Tchernov et al, 2000). Hence, neither fossil lies basal to all living snakes, as had been suggested for Pachyrachis by Lee and colleagues. Although this latest contribution to our understanding of ancient snakes is unlikely to end the controversy over the ancestry of snakes, aquatic feeding does not seem likely to represent an ancestral condition for crown clade snakes (Greene and Cundall, 2000). High flexibility of the intramandibular joint in snake ancestors seems likely, but not necessarily because of its influence on gape size. Such flexibility could have arisen independently in several clades, and it has been lost in a number of crown clade snakes (e.g., most living booids: Frazzetta, 1966; personal observation). An example of intramandibular joint flexibility apparently unrelated to gape has recently been demonstrated in scolecophidians. High-resolution, high-speed video records show that Leptotyphlops culcis (see p. 304) ingests prey by rapidly and repeatedly retracting the mandibular interramal joint, a movement made possible by extraordinary flexibility of both the intramandibular and interramal joints (Kley and Brainerd, 1999). The rapid mandibular raking of prey by leptotyphlopids requires mechanisms quite unlike those used in maxillary transport by typhlopids, but both clades remain gape-limited, fossorial predators. Two aspects of striking in snakes have recently been experimentally tested. During defensive strikes (Zamudio et al, 2000; see p. 307, left column), Crotalus atrox
spreads its fangs on average slightly more than twice the resting distance between the fangs. This capability influences patterns of envenomation and the use of suction devices to treat envenomation. The lateral spread of maxillae in Crotalus was loosely anticipated by the structural and radiographic analysis of upper jaw movements in Crotalus ruber by DuUemeijer and Powel (1972), a reference we inadvertently omitted in our accounts of viperid feeding mechanics. In our account of fast capture systems in snakes (p. 305), we described thamnophiine aquatic strikes as slow. Alfaro (1999) has now demonstrated a range of strike behaviors in three species that feed predominantly on fish. Nerodia rhombifer uses a slow, sweeping capture strategy, but two species of Thamnophis use directed strikes. One of these fish specialists, T. couchii, reaches peak velocities equivalent to terrestrial strikes of some viperids and uses prestrike preparatory coiling, whereas the other, T. rufipunctatus, uses slower strikes and incorporates less of the trunk in prestrike preparatory coils. Striking performance may therefore be evolutionarily labile, even among closely related species. Acknowledgments We thank A. Deufel, F. Irish, N. Kley, and K. Schwenk for their helpful comments on the manuscript, N. Kley for providing data on scolecophidian feeding prior to publication, A. C. Burke for sharing her data on Hox gene expression in Thamnophis embryos, and innumerable colleagues and friends for sharing observations and experiences with us over the years, particularly J. Groves, D. Hardy, B. Jayne, K. Kardong, A. and B. Savitzky, and B. Young. Financial support for some of our research was provided by Lehigh University faculty research grants, NSF and American Philosophical Society awards to D.C., and by NSF and Lichen Foundation awards to H.W.G.
References Albright, R. G., and E. W. Nelson (1959a) Cranial kinetics of the generalized colubrid snake, Elaphe obsoleta quadrivittata. I. Descriptive morphology. J. Morphol. 105:193-239. Albright, R. G., and E. W. Nelson (1959b) Cranial kinetics of the generalized colubrid snake, Elaphe obsoleta quadrivittata. II. Functional morphology. J. Morphol. 105:241-291. Alfaro, M. E. (1998) Striking and sweeping: prey capture in thamnophiine snakes. Am. Zool. 38:11 A. Alfaro, M. (1999) Sweeping, striking, and sniping: Modes of aquatic prey capture in thamnophiine snakes. Am. Zool. 39:97A. Alton, N., and E. Kochva (1974) The quantities of venom injected into prey of different size by Vipera palaestinae in a single bite. J. Exp. Zool. 188:71-76. Anthony, J. (1955) Essai sur revolution anatomique de Lappareil venimeux des Ophidiens. Ann. Sci. Nat., Zool. 17:7-53. Arnold, S. J. (1983) Morphology, performance and fitness. Am. Zool. 23:347-361. Arnold, S. J. (1992) Behavioral variation in natural populations. VI.
328
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
Prey responses by two species of garter snakes in three regions of sympatry. Anim. Behav. 44:705-719. Arnold, S. J. (1993) Foraging theory and prey-size-predator-size relations in snakes. Pp. 87-115. In Snakes, Ecology and Behavior. R. A. Seigel and J. T. Collins (eds.). McGraw-Hill, New York. Barrett, R., P. F. A. Maderson, and R. M. Meszler (1970) The pit organs of snakes. Pp. 277-300. In: The Biology of the Reptilia, Vol. 2. C. Gans and T. S. Parsons (eds.). Academic Press, London. Bauchot, R. (1994) Snakes: A Natural History. Sterling Publ. Co., New York. Bellairs, A. d'A. (1984) Closing address, with comments on the organ of Jacobson and the evolution of Squamata, and on the intermandibular connection in Squamata. Symp. Zool. Soc. Lond. 52:665 - 683. Bemis, W. E., K. Schwenk, and M. Wake (1983) Morphology and function of the feeding apparatus in Dermophis mexicanus (Amphibia, Gymnophiona). Zool J. Linn. Soc. 77:75-96. Bogert, C. M. (1943) Dentitional phenomena in cobras and other elapids with notes on adaptive modifications of fangs. Bull. Am. Mus. Nat. Hist. 81:285-360. Boltt, R. E., and R. F. Ewer (1964) The functional anatomy of the head of the puff adder, Bitis arietans (Merr.). J. Morphol. 114:83-106. Bothner, R. C. (1974) Some observations on the feeding habits of the cottonmouth in southeastern Georgia. J. Herpetol. 8:257-258. Bouche, M. B. (1983) The establishment of earthworm communities. Pp. 431-448. In: Earthworm Ecology: from Darwin to Vermiculture. J. E. Satchel (ed.). Chapman and Hall, London. Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. R Liem, and D. B. Wake (eds.). Belknap Press, Harvard University, Cambridge, MA. Braun, T. B., and D. Cundall (1995) Hydrodynamics of fishing in snakes (Nerodia). Am. Zool. 35:105A. Broadley, D. G. (1963) Two rare fossil reptiles in southeastern Rhodesia: Predator and prey. J. Herp. Assoc. Rhodesia 20:7-8. Broadley, D. G (1979) Predation on reptile eggs by African snakes of the genus Prosymna. Herpetologica 35:338-341. Broadley, D. G (1990) FitzSimons' Snakes of Southern Africa. Jonathan Ball and Ad. Donkers, Parklands, South Africa. Brodie, E. D., Ill, and E. D. Brodie, Jr. (1990) Tetrodotoxin resistance in garter snakes: an evolutionary response of predators to dangerous prey. Evolution 44:651-659. Brongersma, L. D. (1956) The palato-maxillary arch in some Asiatic Dipsadinae (Serpentes). Proc. K. Ned. Akad. Wet. (Ser. C) 59: 439-446. Brongersma, L. D. (1958) Some features of the Dipsadinae and Pareinae (Serpentes, Colubridae). Proc. K. Ned. Akad. Wet. (Ser. C) 61:7-12. Brown, E. E. (1958) Feeding habits of the northern water snake, Matrix s. sipedon Linnaeus. Zoologica 43:55-71. Burghardt, G M. (1969) Comparative prey-attack studies in newborn snakes of the genus Thamnophis. Behaviour 23:78-114. Burghardt, G M. (1977) Learning processes in reptiles. Pp. 555-681. In: Biology of the Reptilia, Vol. 7. C. Gans and D. W. Tinkle (eds.). Academic Press, London. Cadle, J. E. (1984a) Molecular systematics of neotropical xenodontine snakes. I. South American xenodontines. Herpetologica 40,8-20. Cadle, J. E. (1984b) Molecular systematics of neotropical xenodontine snakes. II. Central American xenodontines. Herpetologica 40,21-30. Cadle, J. E. (1984c) Molecular systematics of neotropical xenodontine snakes. III. Overview of xenodontine phylogeny and the history of New World snakes. Copeia 1984:641-652. Cadle, J. E. (1988) Phylogenetic relationships among advanced snakes. Univ. California Publ. Zool. 119:1-77.
Cadle, J. E. (1992) Phylogenetic relationships among vipers: immunological evidence. Pp. 41-48. In: Biology of the Pitvipers. J. A. Campbell and E. D. Brodie, Jr. (eds.). Selva, Tyler, Texas. Cadle, J. E. (1994) The colubrid radiation in Africa (Serpentes: Colubridae): phylogenetic relationships and evolutionary patterns based on immunological data. Zool. J. Linn. Soc. 110:103-140. Cadle, J. E. (1999) Dental morphology, systematics, and phylogeny of Pseudoxyrhopus and related genera from Madagascar (Serpentes: Colubridae), with descriptions of a new species and a new genus. Bull. Mus. Comp. Zool., Harvard 155:381-443. Cadle, J. E., H. C. Dessauer, C. Gans, and D. F. Gartside (1990) Phylogenetic relationship and molecular evolution in uropeltid snakes (Serpentes: Uropeltidae): allozymes and albumin immunology Biol. J. Linn. Soc. 40:293-320. Caldwell, M. W., and M. S. Y Lee (1997) A snake with legs from the marine Cretaceous of the Middle East. Nature 386:705-709. Camilleri, C , and R. Shine (1990) Sexual dimorphism and dietary divergence: differences in trophic morphology between male and female snakes. Copeia 1990:649-658. Cogger, H., and R. Zweifel (1997) Encyclopedia of Reptiles and Amphibians, 2nd Ed. Academic Press, San Diego. Cohn, M. J., and C. Tickle (1999) Developmental basis of limblessness and axial patterning in snakes. Nature 399:474-479. Coleman, K., L. A. Rothfuss, H. Ota, and K. V. Kardong (1993) Kinematics of egg-eating by the specialized Taiwan snake Oligodonformosanus (Colubridae). J. Herpetol. 27:320-327. Cundall, D. (1983) Activity of head muscles during feeding by snakes: a comparative study. Am. Zool. 23:383-396. Cundall, D. (1987) Functional morphology Pp. 106-140. In Snakes: Ecology and Evolutionary Biology. R. A. Seigel, J. T. Collins, and S. S. Novak (eds.). Macmillan, New York. Cundall, D. (1995) The behaviour of the feeding apparatus of Cylindrophis and its bearing on the evolution of alethinophidian snakes. J. Zool. Lond. 237:353-376. Cundall, D. (1998) Squamate suspensorial evolution and feeding behavior. Am. Zool. 38:12A. Cundall, D., and C Gans (1979) Feeding in water snakes: an electromyographic study J. Exp. Zool. 209:189-208. Cundall, D., and H. W. Greene (1982) Evolution of the feeding apparatus in alethinophidian snakes. Am. Zool. 22:924. Cundall, D., and F. J. Irish (1986) Aspects of locomotor and feeding behaviour in the Round Island boa, Casarea dussumieri. Dodo, J. Jersey Wildl. Preserv. Trust 23:108-111. Cundall, D., and F. J. Irish (1989) The function of the intramaxillary joint in the Round Island boa, Casarea dussumieri. J. Zool. Lond. 217:189-207. Cundall, D., and D. A. Rossman (1993) Cephalic anatomy of the rare Indonesian snake Anomochilus weberi. Zool. J. Linn. Soc. 109: 235-273. Cundall, D., and J. D. Shardo (1995) Rhinokinetic snout of thamnophiine snakes. /. Morphol. 225:31-50. Cundall, D., and A. Deufel (1999) Striking patterns in booid snakes. Copeia 1999:868-883. Cundall, D., V. Wallach, and D. A. Rossman (1993) The systematic relationships of the snake genus Anomochilus. Zool. J. Linn. Soc. 109:275-299. d'Alton, E. (1834) Beschreibung des Muskelsystems einer Python bivittatus. Arch. Anat. Physiol. 1834:346-364,432-450,528-543. de Cock Buning, T. (1983) Thermal sensitivity as a specialization for prey capture and feeding in snakes. Am. Zool. 23:363-375. de Queiroz, K., and J. A. Gauthier (1994) Toward a phylogenetic system of biological nomenclature. Trends Ecol. Evol. 9:27-31. Dessauer, H. C , J. E. Cadle, and R. Lawson (1987) Patterns of snake evolution suggested by their proteins. Fieldiana, Zool. 34:1-34.
9. F e e d i n g in Snakes Deufel, A., and D. Cundall (1999) Do booids stab prey? Copeia 1999: 1102-1107. Dowling, H. G., C. A. Haas, S. B. Hedges, and R. Highton (1996) Snake relationships revealed by slow-evolving proteins: a preliminary survey. J. Zool. Lond. 240:1-28. Drummond, H. M. (1983) Aquatic foraging in garter snakes: a comparison of specialists and generalists. Behaviour 86:1-30. Dullemeijer, P., and G. D. E. Povel (1972) The construction for feeding in rattlesnakes. Zool. Meded. 47:561-578. Estes, R., T. H. Frazzetta, and E. E. Williams (1970) Studies on the fossil snake Dinilysia patagonica Woodward. I. Cranial morphology. Bull. Mus. Comp. Zool., Harvard 119:25-74. Estes, R., K. de Queiroz, and J. Gauthier (1988) Phylogenetic relationships within Squamata. Pp. 119-281. In: Phylogenetic Relationships of the Lizard Families: Essays Commemorating Charles L. Camp. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Evans, P. D. (1942) A method of fishing used by water snakes. Chicago Nat. 5:53-55. Fairley, N. H. (1929) The dentition and biting mechanism of Australian snakes. Med. J. Aust. 1:313-327. Feder, M. E., and S. J. Arnold (1982) Anaerobic metabolism and behavior during predatory encounters between snakes {Thamnophis elegans) and salamanders {Plethodon jordani). Oecologia (Berlin) 53:93-97. FitzSimons, V. F. M. (1962) Snakes of Southern Africa. Macdonald, London. Ford, N. B., and G. M. Burghardt (1993) Perceptual mechanisms and the behavioral ecology of snakes. Pp. 117-164. In: Snakes: Ecology and Behavior. R. A. Seigel and J. T. Collins (eds.). McGraw-Hill, New York. Forsman, A. (1991) Adaptive variation in head size in Vipera berus L. populations. Biol. J. Linn. Soc. 43:281-296. Forsman, A., and L. E. Lindell (1993) The advantage of a big head: swallowing performance in adders, Vipera berus. Funct. Ecol. 7: 183-189. Forstner, M. R. J., S. K. Davis, and E. Arevalo (1995) Support for the hypothesis of anguimorph ancestry for the suborder Serpentes from phylogenetic analysis of mitochondrial DNA sequences. Mol. Phyl. Evol. 4:93-102. Franz, R. (1977) Observations of the food, feeding behavior, and parasites of the striped swamp snake, Regina alleni. Herpetologica 33:91-94. Frazzetta, T. H. (1966) Studies on the morphology and function of the skull in the Boidae (Serpentes). 2. Morphology and function of the jaw apparatus in Python sebae and Python molurus. J. Morphol. 118:217-296. Frazzetta, T. H. (1970a) Studies on the fossil snake Dinilysia patagonica Woodward. 11. Jaw machinery in the earliest snakes. Forma etfunctio3:205-221. Frazzetta, T. H. (1970b) From hopeful monsters to bolyerine snakes? Am. Nat. 104:55-72. Frazzetta, T. H. (1999) Adaptations and significance of the cranial feeding apparatus of the sunbeam snake (Xenopeltis unicolor). 1. Anatomy of the skull. J. Morphol. 239:27-43. Gadow, H. (1901) Amphibia and Reptiles. Cambridge Natural History, Macmillan, London. Gans, C. (1952) The functional morphology of the egg-eating adaptations in the snake genus Dasypeltis. Zoologica 37:209-244. Gans, C. (1961) The feeding mechanism of snakes and its possible evolution. Am. Zool. 1:217-227. Gans, C. (1972) Feeding in Dipsas indica and Dunn's paradox. Am. ZooL 12:730. Gans, C. (1973) Uropeltid snakes—survivors in a changing world. Endeavor 32:60-65.
329
Gans, C. (1974) Biomechanics: An Approach to Vertebrate Biology. Lippincott, Philadelphia. Gans, C. (1986) Locomotion of limbless vertebrates: pattern and evolution. Herpetologica 42:33-46. Gans, C , and M. Oshima (1952) Adaptations for egg eating in the snake Elaphe climacophora (boie). Am. Mus. Novitates 1571:1-16. Gans, C , and F. de Vree (1986) Shingle-back lizards crush snail shells using temporal summation (tetanus) to increase the force of the adductor muscles. Experientia 42:387-388. Gasc, J.-P (1981) Axial musculature. Pp. 355-435. In: Biology of the Reptilia, Vol. 11. C. Gans and T. S. Parsons (eds.). Academic Press, London. Gaulke, M. (1995) Observations on arboreality in a Philippine blind snake. Asian Herp. Res. 6:45-48. Gehlbach, F R., J. R. Watkins III, and J. C. Kroll (1971) Pheromone trail-following studies of typhlopid, leptotyphlopid, and colubrid snakes. Behaviour 40:282-294. Gennaro, J. F, R. S. Leopold, and T. W. Merriam (1961) Observations on the actual quantity of venom introduced by several species of crotalid snakes in their bites. Anat. Rec. 139:303. Godley, J. S. (1980) Foraging ecology of the striped swamp snake, Regina alleni, in southern Florida. Ecol. Monogr. 50:411-436. Greene, H. W. (1976) Scale overlap, a directional sign stimulus for prey ingestion by ophiophagous snakes. Z. Tierpsych. 41: 113-120. Greene, H. W. (1983) Dietary correlates of the origin and radiation of snakes. Am. Zool. 23:431-441. Greene, H. W. (1984) Feeding behavior and diet of the eastern coral snake, Micrurus fluvius. Spec. Publ. Univ. Kansas Mus. Nat. Hist. 10:147-162. Greene, H. W. (1989a) Ecological, evolutionary, and conservation implications of feeding biology in Old World cat snakes, genus Boiga (Colubridae). Proc. California Acad. Sci. 46:193-207. Greene, H. W. (1989b) Defensive behavior and feeding biology of the Asian mock viper, Psammodynastes pulverulentus (Colubridae), a specialized predator on scincid lizards. Chinese Herp. Res. 2:21-32. Greene, H. W. (1992) The ecological and behavioral context of pitviper evolution. Pp. 107-117. In: Biology of the Pitvipers. J. A. Campbell and E. D. Brodie, Jr. (eds.). Selva, Tyler, TX. Greene, H. W. (1994) Homology and behavioral repertoires. Pp. 369391. In: Homology, the Hierarchical Basis of Comparative Biology. B. K. Hall (ed.). Academic Press, San Diego. Greene, H. W. (1997) Snakes: The Evolution of Mystery in Nature. Univ. California Press, Berkeley, CA. Greene, H. W., and G. M. Burghardt (1978) Behavior and phylogeny: constriction in ancient and modern snakes. Science 200:74-77. Greene, H. W., and M. A. Santana (1983) Field studies of hunting behavior by bushmasters. Am. Zool. 23:897. Greene, H. W., and D. Cundall (2000) Perspectives: Evolutionary biology: Limbless tetrapods and snakes with limbs. Science 287: 1939-1941. Greenwald, O. E. (1974) Thermal dependence of striking and prey capture by gopher snakes. Copeia 1974:148. Greenwald, O. E. (1978) Kinematics and time relations of prey capture by gopher snakes. Copeia 1978:263-268. Greer, A. E. (1985) The relationships of the lizard genera Anelytropsis and Dibamus. J. Herpetol. 19:116-156. Gregory, P. T, J. M. Macartney, and D. H. Rivard (1980) Small mammal predation and prey handling behavior by the garter snake Thamnophis elegans. Herpetologica 36:87-93. Groombridge, B. C. (1979a) Comments on the intermandibular muscles of snakes. J. Nat. Hist. 132:477- 498. Groombridge, B. C. (1979b) A previously unreported throat muscle
330
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
in the Scolecophidia (Reptilia, Serpentes), with comments on other scolecophidian throat muscles. J. Nat. Hist. 13:661-680. Groombridge, B. C. (1986) Phyletic relationships among viperine snakes. Pp. 219-222. In: Studies in Herpetology. Z. Rocek (ed.). Prague, Czech Republic. Haas, G. (1930) Uber das Kopfskelett und die Kaumuskulatur der Typhlopiden und Glauconiiden. Zool. Jahrb., Abt. Anat. 52:1-94. Haas, G. (1955) The systematic postion of Loxocemus bicolor Cope (Ophidia). Am. Mus. Novitates 1748:1-8. Haas, G (1964) Anatomical observations on the head of Liotyphlops albirostris (Typhlopidae, Ophidia). Acta Zool. 45:1-62. Haas, G (1968) Anatomical observations on the head of Anomalepis aspinosus (Typhlopidae, Ophidia). Acta Zool. 49:63-139. Haas, G (1973) Muscles of the jaws and associated structures in the Rhynchocephalia and Squamata. Pp. 285-490. In: Biology of the Reptilia, Vol. 4. C. Gans and T. S. Parsons (eds.). Academic Press, London. Hager, P.-K. (1906) Die Kiefermuskeln der Schlangen und ihre Beziehungen zu den Speicheldriisen. Zool. Jahrb., Abt. Anat. 22: 173-224. Halloy, M., and G M. Burghardt (1990) Ontogeny of fish capture and ingestion in four species of garter snakes (Thamnophis). Behaviour 112:299-318. Hardy, J. D. (1957) Bat predation by the Cuban boa, Epicrates anguli/er.Copeia 1957:151-152. Harris, H. S., Jr., and R. S. Simmons (1967) Another Dipsas gaigeae (Oliver) from Colima, Mexico. Herpetologica 23:234-235. Hayes, W. K. (1995) Venom metering by juvenile prairie rattlesnakes, Crotalus v. viridis: effects of prey size and experience. Anim. Behav. 50:33-40. Hayes, W. K., and Duvall, D. (1991) A field study of prairie rattlesnake predatory strikes. Herpetologica 47:78-81. Hayes, W. K., P Lavin-Murcio, and K. V. Kardong (1995) Northern Pacific rattlesnakes {Crotalus viridis oreganus) meter venom when feeding on prey of different sizes. Copeia 1995:337-343. Heatwole, H. (1987) Sea Snakes. New South Wales Univ. Press, Kensington, NSW, Australia. Heise, P J., L. R. Maxson, H. G Dowling, and S. B. Hedges (1995) Higher-level snake phylogeny inferred from mitochondrial DNA sequences of 12S rRNA and 16S rRNA genes. Mol. Biol. Evol. 12: 259-265. Hiiemae, K., and A. W. Crompton (1985) Mastication, food transport, and swallowing. Pp. 262-290. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Belknap Press, Harvard University, Cambridge, MA. Hoffstetter, R., and Y. Gayrard (1965) Observations sur I'osteologie et la classification des Acrochordidae (Serpentes). Bull. Mus. Natl. Hist. Nat., Paris (Ser. 2) 36:677-696. Holldobler, B., and E. O. Wilson (1990) The Ants. Belknap Press, Harvard University, Cambridge, MA. lordansky, N. N. (1978) On the origin of snakes. Zool. Zh. 57:888898 (in Russian). lordansky, N. N. (1981) Cranial kineticism and functions of the jaw muscles in Typhlops vermicularis: a contribution to the problem of origin of the Typhlopidae. Zool. Zh. 40:97-108 (in Russian). lordansky, N. N. (1997) Jaw apparatus and feeding mechanics of Typhlops (Opidia, Typhlopidae), a reconsideration. Russian J. Herpetol. 4:120-127. Jackson, K., and T. H. Fritts (1995) Evidence from tooth surface morphology for a posterior maxillary origin of the proteroglyph fang. Amphibia-Reptilia 16:273-288. Jackson, K., and T. H. Fritts (1996) Observations of a grooved anterior fang in Psammodynastes pulverulentus: does the mock viper resemble a protoelapid? J. Herpetol. 30:128-131.
Janoo, A., and J.-P. Gasc (1992) High speed motion analysis of the predatory strike and fluorographic study of oesophageal deglutition in Vipera ammodytes: more than meets the eye. AmphibiaReptilia 13:315-325. Jansen, D. W. (1987) The myonecrotic effect of Duvernoy's gland secretion of the snake Thamnophis elegans vagrans. J. Herpetol. 21: 81-83. Jansen, D. W., and R. C. Foehring (1983) The mechanism of venom secretion from Duvernoy's gland of the snake Thamnophis sirtalis. J. Morphol. 175:271-277. Jayne, B. C (1982) Comparative morphology of the semispinalisspinalis muscle of snakes and correlations with locomotion and constriction. J. Morphol. 172:83-96. Jayne, B. C. (1988a) Muscular mechanisms of snake locomotion: an electromyographic study of lateral undulation of the Florida banded water snake (Nerodia fasciata) and the yellow rat snake {Elaphe obsoleta). J. Morphol. 197:159-181. Jayne, B. C. (1988b) Muscular mechanisms of snake locomotion: an electromyographic study of the sidewinding and concertina modes of Crotalus cerastes, Nerodia fasciata and Elaphe obsoleta. J. Exp. Biol. 140:1-33. Jayne, B. C , H. K. Voris, and K. B. Heang (1988) Diet, feeding behavior, growth, and numbers of a population of Cerberus rynchops (Serpentes, Homalopsinae) in Malaysia. Fieldiana: Zool. (n.s.) 50: 1-15. Kardong, K. V. (1973) Lateral jaw and throat musculature of the cottonmouth snake, Agkistrodon piscivorus. Gegenbaurs morph. Jahrb., Leipzig 119:316-335. Kardong, K. V. (1974) Kinesis of the jaw apparatus during the strike in the cottonmouth snake, Agkistrodon piscivorus. Forma et functio 7:327-354. Kardong, K. V. (1975) Prey capture in the cottonmouth snake {Agkistrodon piscivorus). J. Herpetol. 9:169-175. Kardong, K. V. (1979) 'Trotovipers'' and the evolution of snake fangs. Evolution 33:433-443. Kardong, K. V. (1981) Evolutionary patterns in advanced snakes. Am. Zool. 20:269-282. Kardong, K. V. (1982a) The evolution of the venom apparatus in snakes from colubrids to viperids and elapids. Mem. Inst. Butantan 46:105-118. Kardong, K. V. (1982b) Comparative study of changes in prey capture behavior in the cottonmouth {Agkistrodon piscivorus) and Egyptian cobra {Naja haje). Copeia 1982:337-343. Kardong, K. V. (1986a) The predatory strike of the rattlesnake: when things go amiss. Copeia 1986:816-820. Kardong, K. V. (1986b) Kinematics of swallowing in the yellow rat snake, Elaphe obsoleta quadrivittata: a reappraisal. Jap. J. Herpetol. 11:96-109. Kardong, K. V. (1996) Snake toxins and venoms, an evolutionary perspective. Herpetologica 52:36-46. Kardong, K. V, and P. A. Lavin-Murcio (1993) Venom delivery of snakes as high-pressure and low-pressure systems. Copeia 1983: 644-650. Kardong, K. V, and V. Bels (1998) Rattlesnake strike behavior: kinematics. J. Exp. Biol. 210:837-850. Kardong, K. V, and H. Berkhoudt (1998) Intraoral transport of prey in the reticulated python: tests of a general tetrapod feeding model. Zoology (ZACS) 101:7-23. Kardong, K. V, P. Dullemeijer, and J. A. M. Fransen (1986) Feeding mechanism in the rattlesnake Crotalus durissus. AmphibiaReptilia 7:271-302. Keogh, J. S. (1998) Molecular phylogeny of elapid snakes and a consideration of their biogeographic history. Biol. J. Linn. Soc. 63: 177-203.
9. F e e d i n g in Snakes Keogh, J. S., and S. A. Smith (1996) Taxonomy and natural history of the Australian bandy-bandy snakes (Elapidae, Vermicella) with description of two new species. J. Zool. Lond. 240:677-701. Klauber, L. M. (1956) Rattlesnakes: Their Habits, Life Histories, and Influence on Mankind, Vol. 2. University of California Press, Berkeley, CA. Kley, N. (1998) Kinematics and mechanics of prey transport in scolecophidian snakes. Am. Zool. 38:202A. Kley, N., and E. L. Brainerd (1996) Internal concertina swallowing, a critical component of alethinophidian feeding systems. Am. Zool. 36:81 A. Kley, N. J., and E. L. Brainerd (1999) Feeding by mandibular raking in a snake. Nature 402:369-370. Kluge, A. G. (1991) Boine phylogeny and research cycles. Misc. Publ. Mus. Zool., Univ. Michigan 178:1-58. Kluge, A. G. (1993a) Calabaria and the phylogeny of erycine snakes. Zool. J. Linn. Soc. 107:293-351. Kluge, A. G. (1993b) Aspidites and the phylogeny of pythonine snakes. Rec. Australian Mus. Suppl. 19:1-77. Knight, A., and D. P. Mindell (1993) Substitution bias, weighting of DNA sequence evolution, and the phylogenetic position of Fea's viper. Syst. Biol. 42:18-31. Kochva, E. (1978) Oral glands of the Reptilia. Pp. 43-161. In: Biology of the Reptilia, Vol. 8. C. Gans and K. A. Gans (eds.). Academic Press, London. Kochva, E. (1987) The origin of snakes and evolution of the venom apparatus. Toxicon 25:65-106. Kofron, C. P. (1985) Systematics of the neotropical gastropod-eating snake genera, Tropidodipsas and Sibon. J. Herpetol. 19:84-92. Kraus, F., and W. M. Brown (1998) Phylogenetic relationships of colubroid snakes based on mitochondrial DNA sequences. Zool. J. Linn. Soc. 122:455-487. Kroll, J. C. (1976) Feeding adaptations of hognose snakes. Southwest. Nat. 20:537-557. Kropach, C. (1975) The yellow-bellied sea snake, Pelamis, in the eastern Pacific. Pp. 185-213. In: The Biology of Sea Snakes. W. A. Dunson (ed.). University Park Press, Baltimore, MD. Langebartel, D. A. (1968) The hyoid and its associated muscles in snakes. Illinois Biol. Monogr. 38:1-156. Laporta-Ferreira, I. L., M. G. Salomao, and P. Sawaya (1986) Biologia de Sibynomorphus (Colubridae-Dipsadinae): Reprodugao e habitos alimentares. Rev. Brasil. Biol. 46:793-799. Laporta-Ferreira, I. L., M. da G. Salomao, P. Sawaye, and G. Puorto (1988) Mecanismo de tomado de alimento por serpentes tropicals moluscofagas (Sibynomorphus neuwiedi e Sibynomorphus mikani): Adaptagoes morfofisiologicas do esqueleto cefalico. Bol. Fisiol. anim. (Sao Paulo) 12:81-88. Lee, M. S. Y. (1998) Convergent evolution and character correlation in burrowing reptiles: towards a resolution of squamate relationships. Biol. J. Linn. Soc. 65:369-453. Lee, M. S. Y., and M. W. Caldwell (1998) Anatomy and relationships of Pachyrachis problematicus, a primitive snake with hindlimbs. Phil. Trans. R. Soc. Lond. B 353:1521-1552. Lee, M. S. Y, and R. Shine (1998) Reptilian viviparity and Dollo's Law. Evolution 52:1441-1450. Lee, M. S. Y, G. L. Bell, and M. W. Caldwell (1999) The origin of snake feeding. Nature 400:655-658. Lillywhite, H. B., and R. W. Henderson (1993) Behavioral and functional ecology of arboreal snakes. Pp. 1-48. In: Snakes, Ecology and Behavior. R. A. Seigel and J. T. Collins (eds.). McGraw-Hill, New York. Lillywhite, H. B., P. E. de Delva, and B. Noonan (1998) Retention of fecal mass in snakes: is constipation adaptive. Am. Zool. 38:157A.
331
List, J. C. (1966) Comparative osteology of the snake families Typhlopidae and Leptotyphlopidae. Illinois Biol. Monographs 36:1-112. Lombard, R. E., H. Marx, and G. B. Rabb (1986) Morphometries of the ectopterygoid in advanced snakes (Colubroidea): a concordance of shape and phylogeny. Biol. J. Linn. Soc. 27:133-164. Losos, J. B., and H. W. Greene (1988) Ecological and evolutionary implications of diet in monitor lizards. Biol. J. Linn. Soc. 35: 379-407. Macdonald, H. S. (1974) Bradycardia during death-feigning of Heterodon platyrhinos Latreille (Serpentes). J. Herpetol. 8:157-164. Maderson, P. F A. (1970) The distribution of specialized labial scales in the Boidae. Pp. 301-304. In: Biology of the Reptilia, Vol. 2. C. Gans and T. S. Parsons (eds.). Academic Press, London. Marx, H., and G. B. Rabb (1972) Phyletic analysis of fifty characters of advanced snakes. Fieldiana Zool. 63:1-321. McCarthy, C. J. (1987) Adaptations of sea snakes that eat fish eggs; with a note on the throat musculature of Aipysurus eydouxi (Gray, 1849). J. Nat. Hist. 21:1119-1128. McDowell, S. B. (1968) The affinities of the snakes usually called Elaps lacteus and E. dorsalis. Zool. J. Linn. Soc. Lond. 47:561-578. McDowell, S. B. (1970) The status and relationships of the Solomon Island elapid snakes. J. Zool. Lond. 161:145-190. McDowell, S. B. (1975) A catalogue of snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Museum. II. Anilioidea and Pythoninae. J. Herpetol. 9:1-79. McDowell, S. B. (1979) A catalogue of snakes of New Guinea and the Solomons, with special reference to those in the Bernice P. Bishop Museum. III. Boinae and Acrochordoidea. J. Herpetol. 13:1-92. McDowell, S. B. (1986) The architecture of the corner of the mouth of colubroid snakes. J. Herpetol. 20:353-407. McDowell, S. B. (1987) Systematics. Pp. 3-50. In: Snakes, Ecology and Evolutionary Biology. R. A. Seigel, J. T. Collins, and S. S. Novak (eds.). Macmillan, New York. McKay, W. J. (1889) The osteology and myology of the death adder {Acanthophis antarctica). Proc. Linn. Soc. New South Wales, Sydney 14:893-986. McKinstry, D. M. (1983) Morphologic evidence of toxic saliva in colubrid snakes, a checklist of world genera. Herp. Rev. 14:12-15. Miller, C. M. (1944) Ecological relations and adaptations of the limbless lizards of the genus Anniella. Ecol. Monogr. 14:271-289. Minton, S. A. A. (1990) Venomous bites by nonvenomous snakes: an annotated bibliography of colubrid envenomation. J. Wild. Med. 1:119-127. Mitchell, S. W. (1861) Researches upon the venom of the rattlesnake. Smithson. Contrib. ECnowl. 12(6): 1-145. Moon, B. R. (1998) Structural and Functional Integration of the Snake Axial System. Unpublished Ph.D. Dissertation, Univ. Michigan, Ann Arbor, MI. Moon, B. R., and N. J. Kley (1997) Vertebral mechanics and motor control during swallowing in snakes. Am. Zool. 37:82A. Mora, J. M. (1991) Loxocemus bicolor: feeding behavior. Herp. Rev. 22:61. Mori, A. (1998) Prey-handling behavior in three species of homalopsine snakes: features associated with piscivory and Duvernoy's glands. J. Herpetol. 32:40-50. Morrison, J. J., N. T. Charles, and J. H. Pearn (1983a) The use of experimental models to study the biting habits of Australian snakes in both ''defensive'' and "hunting" bites. Toxicon Suppl. 3:305-308. Morrison, J. J., J. H. Pearn, N. T. Charles, and A. R. Coulter (1983b) Further studies on the mass of venom injected by elapid snakes. Toxicon 21:279-284. Mount, R. H. (1975) The Reptiles and Amphibians of Alabama. Auburn Univ. Press, Auburn, Alabama. MuUin, S. J. (1996) Adaptations facilitating facultative oophagy in
332
D a v i d C u n d a l l a n d H a r r y W. G r e e n e
the gray rat snake, Elaphe obsoleta spiloides. Amphibia-Reptilia 17: 387-394. Murphy, J. C , and R. W. Henderson (1997) Tales of Giant Snakes: A Historical Natural History of Anacondas and Pythons. Krieger Publishing, Malabar, FL. Mushinsky, H. R., J. J. Hebrard, and D. S. Vodopich (1982) Ontogeny of water snake foraging ecology. Ecology 63:1624-1629. O'Connell, D. J., and D. R. Formanowicz, Jr. (1998) Differential handling of dangerous and non-dangerous prey by naive and experienced Texas spotted whip tail lizards, Cnemidophorus gularis. J. Herpetol. 32:75-79. Parker, H. W., and A. G. C. Grandison (1977) Snakes: A Natural History. British Museum (Natural History), London. Peters, J. A. (1960) The snakes of the subfamily Dipsadinae. Misc. Publ. Mus. Zool. Univ Michigan 114:1-224. Phisalix, M. (1922) Animaux Venimeux et Venins, Vol. 1. Masson et Cie, Paris. Plummer, M. V., and J. M. Goy (1984) Ontogenetic dietary shift of water snakes {Nerodia rhombifera) in a fish hatchery. Copeia 1984: 550-552. Rough, F. H. (1988) Mimicry and related phenomena. Pp. 153-234. In: Biology of the Reptilia, Vol. 16. C. Gans and R. B. Huey (eds.). A. R. Liss, New York. Rough, F. H., and J. D. Groves (1983) Specializations of the body form and food habits of snakes. Am. Zool. 23:443-454. Rough, R H., R.M. Andrews, J. E. Cadle, M. L. Crump, A. H. Savitzky, and K. D. Wells (1998) Herpetology. Prentice Hall, Upper Saddle River, NJ. Pregill, G. K. (1977) Axial myology of the racer Coluber constrictor with emphasis on the neck region. Trans. San Diego Soc. Nat. Hist. 18:185-206. Pregill, G. K., J. A. Gauthier, and H. W. Greene (1986) The evolution of helodermatid squamates, with description of a new taxon and an overview of Varanoidea. Trans. San Diego Nat. Hist. Mus. 21: 167-202. Radcliffe, C W., and D. A. Chiszar (1980) A descriptive analysis of predatory behavior in the yellow-lipped sea krait {Laticauda colubrina). J. Herpetol. UA22-A2^. Rajendran, M. V. (1985) Studies in Uropeltid Snakes. Publications Division, Madurai Kamaraj University, Madurai, India. Rasmussen, S., B. Young, and H. Krimm (1995) On the "spitting" behavior in cobras (Serpentes: Elapidae). J. Zool. Lond. 237: 27-35. Reinert, H. K., D. Cundall, and L. M. Bushar (1984) Foraging behavior of the timber rattlesnake, Crotalus horridus. Copeia 1984: 976-981. Rieppel, O. (1977a) The naso-frontal joint in Anilius scytale (Linnaeus) and Cylindrophis rufus (Schiegel): Serpentes, Aniliidae. J. Nat. Hist. 11:545-553. Rieppel, O. (1977b) Studies on the skull of the Henophidia (Reptilia: Serpentes). J. Zool. Lond. 181:145-173. Rieppel, O. (1978a) The evolution of the naso-frontal joint in snakes and its bearing on snake origins. Z. Zool. Syst. Evolut.-Forsch. 16:14-27. Rieppel, O. (1978b) A functional and phylogenetic interpretation of the skull of the Erycinae (Reptilia, Serpentes). J. Zool. Lond. 186: 185-208. Rieppel, O. (1988) A review of the origin of snakes. Pp. 37-130. In: Evolutionary Biology, Vol. 22. M. K. Hecht, B. Wallace, and G T. Pance (eds.). Plenum, New York. Rochelle, M. J., and K. V. Kardong (1993) Constriction versus envenomation in prey capture by the brown tree snake, Boiga irregularis (Squamata: Colubridae). Herpetologica 49:301-304. Rodriguez, G A., and D. P. Reagan (1984) Bat predation by the Puerto Ricanboa, Epicrates inornatus. Copeia 1984:219-220.
Rodriguez-Robles, J. A. (1994) Are the Duvernoy's gland secretions of colubrid snakes venoms? J. Herpetol. 28:388-390. Rodriguez-Robles, J. (1998) Alternative perspectives on the diet of gopher snakes {Pituophis catenifer, Colubridae): literature records versus stomach contents of wild and museum specimens. Copeia 1998:463-466. Rodriguez-Robles, J. A., and R. Thomas (1992) Venom function in the Puerto Rican racer, Alsophis portoricensis (Serpentes: Colubridae). Copeia 1992:62-68. Rodriguez-Robles, J., C J. Bell, and H. W. Greene (1999) Gape size and evolution of diet in snakes: feeding ecology of erycine boas. J. Zool. Lond. 248:49-58. Rossman, C E. (1980) Ontogenetic changes in skull proportions of the diamondback water snake, Nerodia rhombifera. Herpetologica 36:42-46. Rossman, D. A., and P. A. Myer (1990) Behavioral and morphological adaptations for snail extraction in the North American brown snakes (genus Storeria). J. Herpetol. 24:434-438. Ruben, J. A. (1977a) Some correlates of cranial and cervical morphology with predatory modes in snakes. J. Morphol. 152:89100. Ruben, J. A. (1977b) Morphological correlates of predatory modes in the coachwhip (Masticophis flagellum) and rosy boa {Lichanura roseofusca). Herpetologica 33:1-6. Ruthven, A. G (1912) The amphibians and reptiles collected by the University of Michigan-Walker Expedition in southern Vera Cruz, Mexico. Zool. Jahrb. Syst. 32:295-332. Salomao, M. G , and I. L. Laporta-Ferreira (1994) The role of secretions from the supralabial, infralabial, and Duvernoy's gland of the slug-eating snake Sibynomorphus mikani (Colubridae: Dipsadinae) in the immobilization of moUuscan prey. J. Herpetol. 28: 369-371. Savitzky, A. H. (1980) The role of venom delivery strategies in snake evolution. Evolution 34:1194-1204. Savitzky, A. H. (1981) Hinged teeth in snakes: an adaptation for swallowing hard-bodied prey. Science 212:346-349. Savitzky, A. H. (1983) Coadapted character complexes among snakes: fossoriality, piscivory, and durophagy Am. Zool. 23:397-409. Savitzky, A. H. (1992) Embryonic development of the maxillary and prefrontal bones of crotaline snakes. Pp. 119-141. In: Biology of the Pitvipers. J. A. Campbell and E. D. Brodie, Jr. (eds.). Selva, Tyler, TX. Savitzky, B. A. C (1992) Laboratory studies on piscivory in an opportunistic predator, the cottonmouth, Agkistrodon piscivorus. Pp. 347-368. In: Biology of the Pitvipers. J. A. Campbell and E. D. Brodie, Jr. (eds.). Selva, Tyler, TX. Sazima, I. (1989) Feeding behavior of the snail-eating snake, Dipsas indica. J. Herpetol. 23:464-468. Sazima, I., and G Puorto (1993) Feeding technique of juvenile Tropidodryas striaticeps: probable caudal luring in a colubrid snake. Copeia 1993:222-226. Scanlon, J. D., and R. Shine (1988) Dentition and diet in snakes: adaptations to oophagy in the Australian elapid genus Simoselaps. J. Zool. Lond. 216:519-528. Scanlon, J. D., M. S. Y. Lee, M. W. Caldwell, and R. Shine (1999) The palaeoecology of the primitive snake Pachyrachis. Hist. Biol. 13: 127-152. Schuett, G W., D. L. Clark, and R Kraus (1984) Feeding mimicry in the rattlesnake Sistrurus catenatus with comments on the evolution of the rattle. Anim. Behav. 32:625-626. Schwenk, K. (1988) Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny Pp. 569-598. In: Phylogenetic Relationships of the Lizard Families: Essays Commemorating Charles L. Camp. R. Estes and G Pregill (eds.). Stanford University Press, Stanford, CA.
9. F e e d i n g in Snakes Schwenk, K. (1994) Why snakes have forked tongues. Science 263: 1573-1577. Secor, S. M. (1995) Digestive response to the first meal in hatchling Burmese pythons (Python molurus). Copeia 1995:947-954. Secor, S. M., and K. A. Nagy (1994) Bioenergetic correlates of foraging mode for the snakes Crotalus cerastes and Masticophisflagellum. Ecology 75:1600-1614. Secor, S. M., and J. Diamond (1995) Adaptive responses to feeding in Burmese pythons: pay before pumping. J. Exp. Biol. 198: 1313-1325. Secor, S. M., and J. Diamond (1997) Determinants of the postfeeding metabolic response of Burmese pythons. Python molurus. Physiol. Zool. 70:202-212. Secor, S. M., and J. Diamond (1998) A vertebrate model of extreme physiological regulation. Nature 395:659-662. Secor, S. M., E. D. Stein, and J. Diamond (1994) Rapid upregulation of snake intestine in response to feeding: a new model of intestinal adaptation. Am. J. Physiol. 266:G695-G705. Shine, R. (1984) Ecology of small fossorial Australian snakes of the genera Neelaps and Simoselaps. Univ. Kansas Mus. Nat. Hist. Spec. Publ. 10:173-183. Shine, R. (1986) Ecology of low-energy speciaHsts: food habits and reproductive biology of the Arafura filesnake (Acrochordidae). Copeia 1986:424-437. Shine, R. (1991a) Why do large snakes eat larger prey items? Funct. Ecol. 5:493-502. Shine, R. (1991b) Intersexual dietary divergence and the evolution of sexual dimorphism in snakes. Am. Nat. 138:103-122. Shine, R. (1991c) Australian Snakes: A Natural History. Reed Books, Sydney, Australia. Shine, R., and T. Schwaner (1985) Prey constriction by venomous snakes: a review, and new data on Australian species. Copeia 1985:1067-1071. Shine, R., W. R. Branch, R S. Harlow, and J. K. Webb (1998) Reproductive biology and food habits of horned adders, Bitis caudalis (Viperidae) from South Africa. Copeia 1998:391-401. Sisk, N. R., and J. F. Jackson (1997) Tests of two hypotheses for the origin of the crotaline rattle. Copeia 1997:485-495. Slip, D., and R. Shine (1988) Feeding habits of the diamond python, Morelia s. spilota: ambush predation by a boid snake. J. Herpetol. 22:323-330. Smith, H. M. (1957) Curious feeding habit of a blind snake, Leptotyphlops. Herpetologica 13:102. Smith, M. A. (1943) The Fauna of British India, Ceylon and Burma, including the Whole of the Indo-Chinese Subregion: Reptilia and Amphibia, Vol. III. Taylor and Francis, London. Smith, T L., G. D. E. Povel, and K. V. Kardong (1998) Predatory strike of the aquatic tentacled snake: getting a mouth full. Am. Zool. 38:202A. Tchernov, E., O. Rieppel, H. Zaher, M. J. Polcn, and L. J. Jacobs (2000) A new fossil snake with limbs. Science 287:2010-2012. Teather, K. L. (1991) The relative importance of visual and chemical cues for foraging in newborn blue-striped garter snakes (Thamnophis sirtalis simulus). Behaviour 117:225-261.
333
Thomas, R. (1985) Prey and prey processing in blind snakes of the genus Typhlops. Am. Zool. 25:14A. Thomas, R. G., and F H. Rough (1979) The effect of rattlesnake venom on digestion of prey. Toxicon 17:221-228. Tokar, A. A. (1989) A revision of the genus Eryx using ostelogical data. Vestnik Zool. 4:46-55 (in Russian). Underwood, G. (1997) An overview of venomous snake evolution. Symp. Zool. Soc. Lond. 70:1-13. Underwood, G., and Kochva, E. (1993) On the affinities of the burrowing asps Atractaspis (Serpentes: Atractaspididae). Zool. J. Linn. Soc. 107:3-64. Van Riper, W. (1950) How a rattlesnake strikes. Nat. Hist. 59:128-129. Van Riper, W. (1953) How a rattlesnake strikes. Sci. Am. 189 (4): 100-102. Van Riper, W. (1954) Measuring the speed of a rattlesnake's strike. Animal Kingdom 57(2): 50-53. Van Riper, W. (1955) How a rattlesnake strikes. Nat. Hist. 64:308-311. Voris, H. K., H. H. Voris, and L. B. Liat (1978) The food and feeding behavior of a marine snake, Enhydrina schistosa (Hydriphiidae). Copeia 1978:134-146. Wake, M. H. (1993) The skull as a locomotor organ. Pp. 197-240. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). University of Chicago Press, Chicago, IL. Wall, F (1921) Ophidia Taprobanica, or the Snakes of Ceylon. H. R. Cottle, Government Printer, Colombo, Ceylon (Sri Lanka). Webb, J. K., and R. Shine (1992) To find an ant: trail-following in Australian blind snakes (Typhlopidae). Anim. Behav. 43:941-948. Webb, J. K., and R. Shine (1993a) Dietary habits of Australian blindsnakes (Typhlopidae). Copeia 1993:762-770. Webb, J. K., and R. Shine (1993b) Prey-size selection, gape limitation and predator vulnerability in Australian blindsnakes (Typhlopidae). Anim. Behav. 45:1117-1126. Weinstein, S. A., and K. V. Kardong (1994) Properties of Duvernoy's gland secretions from opisthoglyphous and aglyphous colubrid snakes: a critical review. Toxicon 32:1161-1185. Wiister, W., and R. S. Thorpe (1992) Dentitional phenomena in cobras revisited: spitting and fang structure in the Asiatic species of Naja (Serpentes: Elapidae). Herpetologica 48:424-434. Young, B. A. (1998) The comparative morphology of the intermandibular connective tisssue in snakes (Reptilia: Squamata). Zool. Anz. 237:59-84. Young, B. A., and K. V. Kardong (1996) Dentitional surface features in snakes (Reptilia: Serpentes). Amphibia-Reptilia 17:261-276. Zaher, H. (1994a) Les Tropidopheoidea (Serpentes; Alethinophidia) sont-ils reellement monophyletiques? Arguments en faveur de leur polyphyletisme. C. R. Acad. Sci. Paris Life Sci. 317: 471-478. Zaher, H. (1994b) Comments on the evolution of the jaw adductor musculature of snakes. Zool. J. Linn. Soc. Ill:339-384. Zamudio, K., D. L. Hardy, Sr., M. Martins, and H. W. Greene (2000) Fang tip spread, puncture distance, and suction for snake bite. Toxicon 38:723-728. Zweifel, R. G. (1954) Adaptation to feeding in the snake Contia tenuis. Copeia 1954:299-300.
This Page Intentionally Left Blank
S E C T I O N
V
This Page Intentionally Left Blank
C H A P T E R
10 Feeding in Crocodilians JOHAN CLEUREN AND FRITS DE VREE Department of Biology University of Antwerp Antwerp, Belgium
Therefore, crocodilians and birds represent the only surviving archosaurian clades among the modern vertebrates. Crocodylia are traditionally placed into three suborders: Protosuchia, Mesosuchia, and Eusuchia. Three morphological characteristics are used to distinguish these suborders: the development of the secondary palate, the morphology of the vertebral centra, and the exclusion of the pubis from the acetabulum. The earliest forms are represented by the most primitive suborder, the Protosuchia, which flourished in the Late Triassic of South and North America, southern Africa, and Europe and in the Cretaceous of Mongolia (Steel, 1973). Mesosuchians enjoyed considerable evolutionary radiation in the Jurassic and Lower Cretaceous. Eusuchians became dominant in the Late Cretaceous and Cenozoic and include all extant crocodilians. In contrast to the recent aquatic forms, protosuchians were probably terrestrial in habitat (Hecht and Tarsitano, 1983), as indicated by their relatively long limbs, large preacetubular iliac crest, construction of the dermal armor, and, most importantly, by a locomotory system that uses dermal osteoderms and epaxial musculature to flex and extend the vertebral column (Frey, 1984,1988). Protosuchians were hardly more than 1 m long and had a rather lizard-like appearance and a short snout. However, even in this primitive suborder, crocodilian ordinal character states were probably developed to some degree. Langston (1973) summarizes 16 character states of craniological significance: (1) the depressed skull, (2) external nares in terminal position, (3) some development of a bony hard palate, (4) choanae posterior to external nares, (5) parietals united solidly along the middorsal line, (6) postorbital skull roof
I. INTRODUCTION A. Phylogenetic Relationships of Crocodylia B. Inertial Feeding 11. MORPHOLOGY A. Morphology of the Cranium and Mandible B. Morphology of the Hyolingual Apparatus and Its Associated Musculature C. Morphology of the Neck and Cervical Muscles III. FUNCTION A. General Feeding Behavior B. Feeding Stages C. Kinematics D. Role of the Hyolingual Apparatus in the Inertial Feeding Process E. Motor Patterns IV. EVOLUTION A. Diet in Relation to Skull Morphology: Long Snouted versus Short Snouted B. Skull Morphology in Relation to the Bauplan of Jaw Adductors and the Cervical Musculature References
I. I N T R O D U C T I O N A. Phylogenetic Relationships of Crocodylia 1. Relation of Crocodilians to Other Vertebrates Crocodilians are the last surviving reptilian representatives of the subclass Archosauria. The analysis of proteins, lipids, and nucleic acids and the most recent morphological studies (Rowe, 1986) agree in aligning crocodilians with birds and dinosaurs (Fig. 10.1).
FEEDING (K. Schwenk, ed.)
337
Copyright © 2000 by Academic Press. All rights of reproduction iri any form reserved.
338
Johan Cleuren and Frits De Vree
^.>^^ y
y
J- y o^^' o^ y
Alligator Lineage
Crocodyle Lineage
Gavial Lineage
Crocodylia
# ^o°
/^ ^e^
^^* ..>r
#
5i^
/
,j-*
J"
Tetrapoda F I G U R E 10.1. (A) Phylogenetic relation of the Crocodylia to the other vertebrates (after Gauthier et al, 1988). (B) Relationships among the living crocodilians and the position of Tomistoma schlegelli according to morphological (dotted line) or biochemical (dashed line) studies. Modified from Densmore and Owen (1989) and Frey et al (1989).
broad and flat producing the cranial table, (7) antorbital fenestra reduced or absent, (8) quadrate strongly inclined and bordered anteriorly by a long, slender quadratojugal, (9) palate akinetic with pterygoids and quadrates fused to the braincase, (10) pterygoids with wide and deep wings, (11) interpterygoid vacuity absent, (12) many cranial bones and articular more or less pneumatic, (13) eustachian passages more or less en-
closed in bone, (14) posttemporal fenestrae reduced, (15) squamosal, quadrate, and paroccipital process combine to form an otic meatus, and (16) mandible deepened posteriorly with retroarticular process well developed. Nine of the character states (1,2, 5, 6, 8,10, 14, 15, and 16) are foreshadowed in some thecodonts, but are never found in any extensive combination in a single noncrocodilian taxon.
339
10. F e e d i n g in Crocodilians
To the present, only 8 of the 124 described genera survive, all being members of the suborder Eusuchia (Densmore and Owen, 1989). Twenty-two extant species are currently recognized, with Crocodylus the largest genus, containing 12 living species. Caiman the second largest (2 species and 3 subspecies), followed by Alligator (2), Paleosuchus (2), Melanosuchus (1), Osteolaemus (1), Tomistoma (1), and Gavialis (1). 2. Relationships
among the Extant Eusuchia
A major problem in resolving the systematics and evolution of the eusuchian crocodilians is their tendency toward general morphological conservatism and convergence/parallelism in cranial morphology (Densmore and Owen, 1989). The morphological conservatism is explicit in the postcranial region, where few reliable characters can be used for phylogenetic studies (Sill, 1968). Therefore, most traditional assessments of crocodilian phylogeny are based on analysis of the numerous differences in head morphology and skull structure among different species. This cranial variability strongly reflects variation in ontogeny (Steel, 1973) or habitat and diet (lordansky, 1973), and therefore stresses the importance of feeding in crocodilian evolution. An example of this variation is the
Gavialis gangeticus
Tomistoma scliiegelii
specialization for ichthyophagy, which is reflected in elongation of the snout and a reduction in tooth size. Even today, long-snouted species (Fig. 10.2) with reduced teeth, which live mainly or exclusively on fish, are found among the gharials {Tomistoma schlegelii and Gavialis gangeticus) and the crocodiles {Crocodylus johnsoni, C. novaeguinea, C. cataphractus). Similar but opposite morphological adaptations toward a broadening of the snout are associated with a more general diet (large tetrapods such as reptiles, birds, and mammals as well as fish). Broad-snouted (brevirostrine) forms can be found among alligators, caimans, and true crocodiles (Fig. 10.2). These examples clearly show that similar adaptive strategies have led to convergent skull morphology and head shape in various groups of recent and fossil crocodilians (Densmore and Owen, 1989). Such convergence in character states has long been considered important phylogenetically, but only complicates the interpretation of systematic relationships and evolution in crocodilians. For example, the ecological adaptation toward a piscivorous diet and its associated morphological consequences on head shape reoccurred many times throughout crocodilian history, presumably in widely divergent lineages (Romer, 1956). Fiowever, this does not necessarily mean that taxa showing this
Crocodylus intermedius
Alligator mississippiensis
FIGURE 10.2. Dorsal (A), lateral (B), and ventral (E) views of the skull and lateral (C) and dorsal (D) views of the mandible of the gharial, Gavialis gangeticus; the false gharial, Tomistoma schlegelii; a narrow-snouted crocodile, Crocodylus intermedius; and the extreme broad-snouted American alligator. Alligator mississippiensis. Note the gradual increase of the snout width from the longirostrine Gavialis to the brevirostrine Alligator and a decrement in the size of the supratemporal fenestrae, an increasing heterodontic appearance of the dentition, an decreasing length of the mandibular symphysis, a more pronounded undulation of the jaw margins, a less distinct verticalization process of the basisphenoid, basioccipital, and the posterior end of the pterygoid, and an increase in the cranial osteodermic relief. Modified from Mook (1921b).
340
Johan Cleuren and Frits De Vree
condition share recent ancestry. The cranial similarity between Gavialidae and Teleosauridae may be interpreted to result from convergent adaptations to ichthyophagy, but is no proof for close relationships (lordansky, 1973). Studies of protein divergence (Densmore and Owen, 1989) and molecules (Poe, 1996), while in agreement with traditional interpretations of affinities between alligators and caimans, suggest that the true (Gavialis) and false {Tomistoma) gharials are more closely related to each other than to other crocodilians, and that the true crocodiles, Crocodylus, are all very close relatives that may have diverged recently (Fig. 10.1). However, contradictory to the results of the biochemical techniques, new morphological studies have shown that the braincase structure, neural pocket, air sinus systems, jaw adductor mechanism, pelvic and hindlimb morphology, and epaxial musculature of the caudal region of Gavialis gangeticus do not correspond to the rest of the living Eusuchia and therefore may be its most primitive living member (Tarsitano et ah, 1989). B. Inertial Feeding Many reptiles transport food in the oral cavity by cyclic movements of the tongue (De Vree and Gans, 1989; Schwenk and Throckmorton, 1989; Bels et al, 1994). During these transport cycles the tongue moves forward beneath the food, lifts it up, and then retracts, drawing the food toward the esophagus. This lingual transport overcomes the mass-dependent inertial resistance of the food mediated by surface-dependent bonding mechanisms, such as adhesion and interlocking (Bramble and Wake, 1985). As the weight and size of the food item increase, the surface bonding must increase disproportionately. This requires increased surface contact between the tongue and the food, which can only be achieved by enlarging the surface relief (papulation) and/or total tongue size. However, an enlargement of the tongue produces several disadvantages. Support of large, active prey by a large tongue requires an increase in gape, which in turn decreases the grip of the jaws on the prey and thus increases the chance of prey escape. A large tongue also decreases the gular opening, providing an obstacle during swallowing. Predatory reptiles, such as crocodilians, varanid lizards (Smith, 1982,1986), and snakes (Gans, 1961; Cundall and Gans, 1979; Cundall, 1983) feed on very large and heavy prey items and use an alternative method for prey transport. They do not transport the prey with the tongue but employ, instead, inertial feeding in which the inertia of the food item is utilized in shifting the prey toward the back of the oral cavity (Gans, 1969).
Bramble and Wake (1985) note that "inertial feeding is a facultative behavior for most tetrapods that use it" and state that snakes and varanid lizards are perhaps the only tetrapods in which inertial feeding has become obligatory. Obviously, crocodilians also belong to this latter group (Cleuren and De Vree, 1992) as the wide and flat crocodilian tongue lacks the ability to transport prey. Thus, the group of obligatory users of inertial feeding consists of species with a tongue that is too simple/primitive to be suitable for true lingual transport (crocodilians) or is highly specialized for chemosensory behavior (varanids and snakes).
IL MORPHOLOGY A. Morphology of the Cranium and Mandible 1. Cranial
Osteology
a. Skull The skull of many crocodilian species has been described in detail by many authors (Briihl, 1862; Miall, 1878; Mook, 1921a, c; Wermuth, 1953; lordansky, 1964). A wonderful review of the general crocodilian craniology is presented by lordansky (1973). The crocodilian skull (Fig. 10.2) conforms to the archosaurian diapsid type and is akinetic. The most notable modifications of the crocodilian skull are the formation of the cranial table, the elongation of the jaws, and the development of a secondary palate. The cranial table is formed by the flattened dorsal, postorbital part of the cranial roof. The secondary palate is formed by palatal processes of the premaxillae, maxillae, palatines, and pterygoids. This formation results in a posterior extension of the nasal passages, which terminate in secondary choanae ventral to the base of the brain case. This specialization permits crocodilians to breathe via the dorsally placed nostrils even when the mouth is holding prey under water. Pterygoids and quadrates are attached firmly to the braincase. The anterodorsal inclination of the immobile quadrates results in a posterior displacement of the retroarticular processes. The jaw margins are undulating, forming three convex and two intermediate concave arches. This pattern is more developed in brevirostrine crocodilians than in longirostrine species (Fig. 10.2); it is practically absent in T. schlegelii and G. gangeticus (lordansky, 1964). The length of the mandibular symphysis is also related positively to the length of the snout (Fig. 10.2). In G. gangeticus it reaches the level of the 23rd or 24th dentary tooth, whereas in Crocodylus niloticus it only reaches the 4th tooth.
10. Feeding in Crocodilians
341
b. Dentition
d. Cartilago Transiliens
Thecodont teeth occur on the premaxilla, maxilla, and dentary. Crocodilian teeth have conical, pointed, thick-walled crowns that are often separated from the cylindrical root by a slight constriction. The teeth are deeply embedded in the alveolar ridges. All extant crocodilian species lack palatal teeth. Most authors (Edmund, 1969; Ferguson, 1981,1984) refer to the dentition as being pseudoheterodont, although the differentiation in tooth size differs among species. Crocodilians with moderate to broad snouts show more variation than narrow-snouted species such as Gavialis and Tomistoma, which tend to have all their mature teeth more or less the same size (Fig. 10.2). Kieser et al. (1993) identified three morphogenetic zones in each of the age classes of C. niloticus: an incisor, a canine, and a molar region. They suggest that the Nile crocodile has five premaxillary incisors, followed by five canines and six or more postcanines. In the lower jaw they identified three incisors, five canines, and up to seven postcanines, and therefore concluded that the dentition of the Nile crocodile is heterodont rather than homodont. Fieterodonty is increased by the undulation of the jaws. The largest teeth of both the upper and the lower jaws are located in the central portion of the convex arches of the undulating jaw margins, whereas the smallest teeth occupy the concave arches (lordansky, 1973). The fine structure and chemical analysis of the teeth of Alligator mississippiensis are described by Sato et al (1990) and Shimada et al (1992).
The pyramidally shaped cartilago transiliens (Fig. 10.3) consists of two triangular cartilaginous disks, which are detached in the median plane and covered by a thick tendinous sheet (fibrous pillow in lordansky, 1964). It is positioned between the torus transiliens of pterygoidal flanges and the coronoid by the presence of many tendons that insert on it. On the dorsal side the cartilago is attached to the mandibular adductor tendon (stem tendon in lordansky, 1964; tendon B in Van Drongelen and DuUemeijer, 1982), the tendon of the pseudotemporal muscle, and some fibers of the m. adductor mandibulae externus profundus. The intramandibular tendon attaches to its ventral surface. The cartilago is also connected to the angular bone and to the m. pterygoideus anterior. It thus forms a connection between the adductor tendon and the lower jaw, and a junction of the vertical tendon system of the intramandibular muscle. A certain degree of freedom is permitted. The cartilago transiliens is not uniquely found in crocodilians. It forms a part of the gliding joint in the turtles CheIonia and Caretta (Schumacher, 1973).
c. Jaw Joint The mandibular joint shows a simple hinge mechanism in which only movements in the sagittal plane are allowed. Several morphological arrangements guarantee the rotational motion in the crocodilian jaw joint by preventing lateral movements of the mandible. During closing, the medial sides of the angular are guided by the pterygoid wings to ensure sagittal movements. This guiding is necessary as the medial traction component of the muscles implies that both halves of the lower jaw are drawn toward the pterygoid, which is made possible by the flexible connection of the two halves of the lower jaw at the mandibular symphysis. The fibrous pillow of the mandibular adductor tendon (stem tendon in lordansky, 1964) serves as a special buffer between the lateral edge of the pterygoidal flange and the mandible during opening and closing. This tendon, together with the cartilago transiliens (a cartilaginous disk situated between the tendon and the surangular), thus plays a special role in the guidance of the lower jaw past the pterygoid wings. The medial components of the adductor muscles also provide a firm guide in the jaw joints (Schumacher, 1973).
2. Jaw Muscles The jaw musculature of the several crocodilian species has been described extensively by many authors: A. mississippiensis (Poglayen-Neuwall, 1953; lordansky, 1964; Schumacher, 1973; Busbey, 1989), Crocodylus niloticus, C. rhomhifer, and C. porosus (Lubosch, 1914; Lakjer, 1926; Poglayen-Neuwall, 1953; lordansky, 1964; Schumacher, 1973), and Caiman crocodilus (Schumacher, 1973; Van Drongelen and DuUemeijer, 1982). The terminology of lordansky (1964) and Schumacher (1973) will be followed. The classification of the adductors corresponds to that of the crossopterygians (Luther, 1914) and depends on their relation to the N. trigeminus. Generally, three adductors are recognized: the m. adductor mandibulae externus (subdivided in a pars superficialis, a pars medialis, and a pars profundus, although not clearly separable), the m. adductor mandibulae posterior, and the m. adductor mandibulae internus (including the m. pterygoideus anterior and posterior, the m. pseudotemporalis, and the m. intramandibularis). The jaw adductors are quite uniform in the crocodilians; only minor differences are observed. For a full description of the jaw muscles, refer to the mentioned authors. The origins and insertions are summarized in Table 10.1 and Fig. 10.3. The tendinous structure and histochemical characteristics of the jaw muscles of A. mississippiensis are described by Sato et al (1992) and Shimada et al (1993). They distinguished the fiber types as red, intermediate.
342
Johan Cleuren and Frits De Vree
mandibijlae°'^ MAMP MAMEP MAMES
M. pseudotemporalis
M. pterygoideus anterior
M. intramandibularis
M. pterygoideus posterior cartilago transiliens
F I G U R E 10.3. Dorsal and lateral views of the lines of action of the jaw muscles and the position of the cartilago transiliens in Caiman crocodilus (fr, frontal; j , jugal; 1, lacrimal; MAMP, m. adductor mandibulae posterior; MAMEP, m. adductor mandibulae extemus profundus; MAMES, m. adductor mandibulae extemus superficialis; m, maxillary; n, nasal; p, parietal; pm, premaxillary; po, postorbital; prf, prefrontal; pr.r, retroarticular process; sq, squamosal; q, quadrate; qj, quadratojugal).
white, and tonic and found the highest percentage of red and intermediate fibers in the m. intramandibularis. In the superficial and medial portions of the m. depressor mandibulae and of the m. pterygoideus, all three fiber types are present in approximately equal amounts. The highest white fiber composition is found in the m. pseudotemporalis, the m. adductor mandibulae posterior, and the m. adductor mandibulae externus. B. Morphology of the Hyolingual Apparatus and Its Associated Musculature 1. Hyobranchial Apparatus and Tongue Crocodilians have a rather simple hyobranchial apparatus that consists of a hyoid body (basihyoid or corpus hyoidei) and a pair of anterior cornua (cornu bran-
chiale I). The posterior cornua (cornu branchiale II) have been fused or lost and there is no processus lingualis (processus entoglossum). The cartilaginous hyoid body (corpus hyoidei) is the most prominent part of the hyobranchium (Figs. 10.2 and 10.4). It is a broadly rectangular, ventrally convex plate that has rounded corners and resembles a widebladed shovel. The part of the hyobranchial apparatus posterior to the articulation of the cornu branchiale I is narrower than the anterior part. The trachea and the larynx are embedded in the posterior dorsal concavity of this posterior part. In older animals, ossifications are found in the posterior part of the basihyal (Fiirbringer, 1922). The slightly dorsally curved anterior edge of the basihyal is thinner than its lateral and posterior edges. The anterior portion of the basihyal bears small incisures or fenestrations, which are most obvious in older animals and are closed by a thin membrane
343
10. Feeding in Crocodilians TABLE 10.1 Origin and Insertion of the Jaw Muscles of the Crocodylia Muscle
Origin
Insertion
MDM
The posterior edge of the parietal and supraoccipital, the posterior surface of the squamosal, and the posterodorsal part of the quadrate anterior to the paroccipital process From a groove in the lateral surface of the squamosal
The dorsal surface of the retroarticular process of the mandible The anterior part of the retroarticular process
MAMES
The ventral surface of the descending processus (formed by the quadratojugal and quadrate) The lateral part of the ventral surface of the quadrate bone
The dorsolateral edge of the surangular
MAMEM
The anterior and lateral surface of the medial lamina of the CAT The lateral part of the ventral surface of the quadrate
By means of a small aponeurosis (a part of the MAT) on the cartilago transiliens The dorsomedial surface of the surangular and the angular, medial to the insertion of the MAMES
MAMEP
The ventral side of the quadrate and from between the lateral and medial lamina of the CAT
On the dorsomedial surface of the cartilago transiliens by means of the lamina lateralis of the MAT By means of an aponeurosis to the dorsal crest of the angular
MAMP
The ventral side of the descending process of the quadrate The medial part of the ventral surface of the quadrate The posterior surface of the lamina medialis of the CAT
The medial surface of the lateral crest of the angular bone, the medial side of the surangular, and on the posterior wall of the Meckelian fossa Posterior lamina (runs backward from the angular and the articular bone) of the MAT
MPST
From the laterosphenoid
On the angular bone, by means of the dorsal surface of the posterior lamina of the MAT On the dorsal surface of the cartilago transiliens
MPTA
The dorsal surface of the cartilaginous septum area between the orbit and the nasal cavity and the ventral part of the interorbital septum From the medial surface of the maxilla, the dorsolateral surface of the palatine, the anteromedial surface of the pterygoid, the descending prefrontal pillar, and the caudoventral part of the lateral surface of the basisphenoidal rostrum
On the dorsomedial surface of the angular, by means of the lamina anterior of the MAT The dorsal surface of the cartilago transiliens
MPTP
Dorsal and ventral surface of the pterygoid and three aponeuroses attached to the posterior part of the pterygoid flange
On the surangular and articular bone, by means of three lamina of the pterygoid tendon
MIM
The ventral surface of the cartilago transiliens
The lateral surface of the angular, coronoid and splenial, the dorsal surface of Meckel's cartilage, and the medial surface of the dentary
(Schumacher, 1973). They are poorly developed or sometimes absent in Caiman and Crocodylus, but rather large in Alligator (Flirbringer, 1922). In Caiman (Cleuren and De Vree, 1992) and Gavialis gangeticus (Sondhi, 1958), the rod-shaped cornu branchiale I articulates medially with the lateral margin of the basihyoid and then extends posteromedially (Fig. 10.2). In Alligator, this articulation lies more posteriorly and in Crocodylus, more anteriorly (Flirbringer, 1922). In Caiman crocodilus (Cleuren and De Vree, 1992) the posterior part of the cornu branchiale I gradually widens, flattens, and twists toward the trachea, ending in a thin, leaf-like cartilaginous epibranchial. In G. gangeticus, a ligament connects the base of the cornua branchiales I with the basihyal (Sondhi, 1958), whereas this sheath of ligament is absent in other crocodilians (Fig. 10.2). The ossification of the cornu branchiale I progresses from proximal to distal (Schu-
macher, 1973). In crocodilians, the cornua branchiales II are not separated and are represented by the posterior corners of the basihyal (Flirbringer, 1922; Gnanamuthu, 1937). As described for A. mississippiensis (Busbey, 1989) and C. crocodilus (Cleuren and De Vree, 1992), the ventral surface of the basihyal is connected to the posterior part of the tongue by a fibrous pad and thus does not support the tongue. However, the fibrous connection between the hyobranchial apparatus and the tongue will transmit forces passively between them when the hyobranchium moves. Also, the curved anterior border of the basihyal forms a buccal fold, which can contact the gular fold on the palate. The buccal fold lies in front of the gular fold with the mouth closed. Both folds form a seal between the posterior edge of the tongue and the palate and tend to prevent flooding of the esophagus and larynx.
344
Johan C l e u r e n a n d Frits D e Vree
The crocodilian tongue lacks any specific subdivision into base, body, or apex. This condition is observed in Alligator (Busbey, 1989), Crocodylus (Tanner and Avery, 1982), Gavialis (Sondhi, 1958), and C. crocodilus (Cleuren and De Vree, 1992). The wide, roughly triangular tongue forms a fibrous pad that thickens posteriorly from the tip. It is connected to the lining of the buccal floor over its total length; only its anterior tip is free. The tongue covers nearly the whole ventral surface of the oral floor, filling the space between the mandibular symphysis and the glottis. An intrinsic tongue musculature is completely absent in the crocodilian tongue (Ferguson, 1981b). The surface of the tongue in Alligator is covered by heavily keratinized, conical papillae (Shimada et ah, 1990). 2. Hyobranchial
Muscles
The hyobranchial apparatus and its associated musculature are described for Gavialis gangeticus (Fiirbringer, 1922; Sondhi, 1958), several alligatorines {A. mississippiensis, C. crocodilus, Paleosuchus palpebrosus: Fiirbringer, 1922; Schumacher, 1973; personal observations), and crocodylines {Crocodylus acutus, C. niloticus, C. palustris, and C. rhombifer: Fiirbringer, 1922; Gnanamuthu, 1937; Schumacher, 1973; personal observations). Muscles of the hyolingual apparatus (Fig. 10.4) are subdivided into four different muscle groups: the hypoglossal muscles, the hypobranchial longitudinal muscles, the glossopharyngeal muscles, and the m. intermandibularis (Schumacher, 1973). The terminology of Lubosh (1933) is followed. Hypoglossal muscles consist of hypobranchial spinal muscles (m. branchiomandibularis spinalis) and
sternohyoideus
the muscles of the tongue (m. hyoglossus and m. genioglossus). They all form a connection between parts of the hyobranchial apparatus and the lower jaw (Fig. 10.4). The m. branchiomandibularis has its origin on the basal end of the cornu branchiale I and inserts on the medial surface of the lower jaw. Together with the fibrous pad, the m. genioglossus and the m. hyoglossus constitute the main mass of the tongue; the m. hyoglossus lying ventromedially and the m. genioglossus laterally. The m. hyoglossus arises from the posterodorsal edge of the corpus hyoidei and inserts on the ventral surface of the fibrous pad of the tongue. The m. genioglossus originates from an aponeurosis from the medial surface of the mandibular symphysis of the dentary bone. The fibers of the pars medialis insert on the medioventral surface of the corpus hyoidei, and the thicker pars lateralis inserts on the lateral area of the tongue. The hypobranchial longitudinal muscles (m. coracohyoideus, m. episternobranchiotendineus, and the m. episternobranchialis) consist of long, parallel running muscle fibers that arise from the coracoid bone or the sternum and insert on the hyoid body or the cornu branchiale I (Fig. 10.4). These muscles are extrinsic tongue muscles that act as retractors of the hyoid apparatus. The m. coracohyoideus (syn.: m. omohyoideus) originates from the lateral margin of the coracoid bone, runs anteriorly paralleling the trachea, and attaches with a tendinous aponeurosis to the caudal edge of the first ceratobranchials. The m. episternobranchiotendineus (Schumacher, 1973) originates from the anteroventral surface of the sternum and runs to the medial surface and the posterodorsal margin of the splenial bone. Because the episternobranchiotendineus does not insert on the hyoid apparatus (personal ob-
branchiomandibularis visceralis hyoglossus intermandibularis genioglossus medialis
genioglossus lateralis branchiomandibularis spinalis coracohyoideus
sternomandibularis
F I G U R E 10.4. Ventral view of the lines of action of the hyolingual muscles and the position of the hyobranchial body and the first ceratabranchials (CBI) and the tongue (dotted line) in Caiman crocodilus.
10. Feeding in Crocodilians servations), its status as a hypobranchial longitudinal muscle (Schumacher, 1973) can be questioned. Lubosch (1933) refers to the anterior part of this muscle as m. tendineomandibularis. Because none of these names reflects its true topography, a new name is suggested: m. sternomandibularis (presented in Fig. 26 of Gnanamuthu, 1937). Fibers of the m. episternobranchialis (syn.: m. sternohyoideus) originate from the anteroventral and ventrolateral surface of the episternum and run anteriorly medial to the m. episternobranchiotendineus, also paralleling the trachea. They insert on the lateral surface of the hyoid body and on the medial surface of the first branchials. Glossopharyngeal muscles (m. branchiomandibularis visceralis and m. thyrohyoideus) are small muscles connecting the hyobranchial apparatus to the pharynx. The m. branchiomandibularis visceralis (Fig. 10.4; m. mandibulohyoideus in Sondhi, 1958) originates from the lateral sides of the cornu branchiale I. The fibers run anterolaterally to the ventral edge of the mandibula to insert on the ventromedial surface of the angular bone. Some fibers arise from the flap-like, cartilaginous extension of the cornu branchiale I and insert on the fascia surrounding the pharynx. The m. intermandibularis (Fig. 10.4) consists of a thin layer of fibers that arise from the dorsomedial surface of the splenial bone and from the medial surface of the anterior part of the dental bone. The fibers extend transversely medianly and insert on a medial raphe (gular septum in Sondhi, 1958). C. Morphology of the Neck and Cervical Muscles 1. Osteology of the Cervical Region The osteology of the cervical region is described extensively by Van Bemmelen (1887), Virchow (1914), Boschma (1920), Mook (1921b), Troxell (1925), Kalin (1933, 1955), Hofstetter and Case (1969), Seidel (1978), and Frey (1988). All three crocodilian subfamilies, Crocodylinae, Alligatorinae, and Gavialinae, show a homogeneous vertebral osteology (Hofstetter and Case, 1969). As in most other crocodilian vertebrae, the cervical vertebrae are procoelous, meaning that they are concave-convex with the hollow end in front (Troxell, 1925). Only the axis, the second sacral, and the first caudal vertebrae form an exception to this rule. Confusion exists whether the neck includes the first seven or nine vertebrae. Proof for the existence of seven cervical vertebrae is given by the coelom extending as far forward as the eighth vertebra. However, according to Hofstetter and Case (1969) and Frey (1988), the first nine verte-
345
brae may be called cervical for having no connection with the sternum. This latter theory will be followed here, as the first nine cervical vertebrae seem to form a functional unity. They include the pro-atlas and atlas (referred to as C-1), the axis (C-2), and the uniform third to ninth vertebrae (C-3 to C-9). The chain of cervical centra forms a curved cylinder that is concave dorsally. This curvature, known as the cervical or nuchal curvature (Seidel, 1978), is accentuated in the neck by the neural spine tips (Fig. 10.5). 2. Cervical Muscles The crocodilian epaxial muscles (Fig. 10.5) of the cervical and occipital region have been studied in two members of the Alligatorinae—A. mississippiensis (Seidel, 1978; Frey, 1988) and C. crocodilus (personal observations)—and in one crocodyline—Crocodylus niloticus (personal observations). The three epaxial subdivisions found in the thoracolumbar region of the Crocodylia are extended into the cervical and occipital region; the transversospinalis system, the longissimus system, and the iliocostalis system. The presence of these three muscle systems represents a truly primitive condition in the Crocodylia, as it is found in all living reptiles, mammals, and birds, but not in fishes and amphibians. The nomenclature of the cervical epaxial musculature is determined by the assignment of a muscle to any of these three systems and is followed by an appendix, which is related to its topography. The appendix "dorsi" is used for muscles of the thoracal region (trunk muscles). Muscles originating and inserting on the vertebrae of the neck (cervical muscles) are assigned with the appendix "cervicis," and muscles that arise from the cervical vertebrae and insert on the cranium (cervical-occipital muscles) are assigned with the appendix "capitis." The cervical musculature is covered superficially by the complex fascia of the neck and shoulder. All three epaxial systems are divided through the formation of fascial compartments. The dorsal intermuscular septum is situated between the transversospinalis and the longissimus system. It extends superficially to enclose the transversospinalis system dorsally and the longissimus system laterally. The dorsal intermuscular septum also forms a strong connection with the fascia of the skin and osteoderms of the neck. The longissimus system is fully separated from the iliocostalis system, which is enclosed laterally and ventrally by its own fascia. Because some of the iliocostal myosepta merge with the longissimus system, the distinction between these two systems is only distinct from the occiput back to the fifth cervical vertebra (Seidel, 1978). The cervical transversospinalis system is the most
346
Johan Cleuren and Frits D e Vree MTSCa
MSCaP
MLCe
MECa C3
MICCe
C2
C1
MLCaP
MTSCa T .MECaS MECaM
L
.MSCaP MLCaS MICCa
FIGURE 10.5. Lateral view of the lines of action of the cervical muscles and the position of their insertion of the occipital region of the skull in Caiman crocodilus (C1-C9, first to ninth cervical vertebra; bsphen, basisphenoid; bocc, casioccipital; co, occipital condyle; ectopt, ectopterygoid; exocc,exoccipital; MECa, m. epistroheo-capitis; MICCa, m. ilio-costalis capitis; MICCe, m. ilio-costalis cervicis; MLCaP, m. longissimus capitis profundus; MLCaS, m. longissimus capitis superficialis; MLCe, m. longissimus cervicis; MScaP, m. spinocapitis posticus; MTSCa, m. transversospinalis capitis; p, parietal; pteryg, pterygoid; q, quadrate; qj, quadratojugal; socc, supraoccipital; sq, squamosal).
differentiated and complex system and is associated with the neural spines and pre- and postzygapophyses. It is the most dorsally positioned system, bordered medially by the neural spines and ventrally by the longissimus system. The transversospinalis system is subdivided into a "cervicis'' part, inserting on the atlas (m. transversospinalis cervicis), a "capitis" part inserting on the occiput (m. transversospinalis capitis, m. spinocapitis posticus, and m. epistropheo-capitis) and several small intervertebral muscles (m. interneuralis cervicis and m. interarticularis cervicis).
The cervical longissimus system is concerned with the cervical transverse processes and forms an extension of the longissimus system of the trunk. The system consists of three muscles: one "cervicis" muscle (m. longissimus cervicis) and two "capitis" muscles (m. longissimus capitis superficialis and m. longissimus capitis profundus). The cervical iliocostalis system is associated with cervical ribs. It consists of two muscles continuing from the iliocostalis dorsi (m. iliocostalis capitis and m. iliocostalis cervicis).
347
10. Feeding in Crocodilians
In addition to the detailed description of A. tnississippiensis, Seidel (1978) investigated the neck muscles of other crocodilians, including Caiman, Melanosuchus, Osteolaemus, Crocodylus, Tomistoma, and Gavialis, and he observed a relative uniformity or regularity of the musculature. All muscles are present in all genera. The differences are of proportion and size, without any qualitative difference. The only notable features that Seidel (1978) mentioned are that long snouts seem to be correlated with elongated musculature and that the crocodiles have a reduced spinocapitis posticus and longissimus musculature. In contrast to this, the large m. spinocapitis of the flat-snouted alligator indicates the relative importance of roll and yaw muscles. The cervical musculature of C. crocodilus and C. niloticus (personal observations) strongly resembles the descriptions of A. mississippiensis (Seidel, 1978), but differs slightly from Frey (1988). Therefore, the subdivision given by Seidel (1978) is used to summarize the origins and insertions of the most important crocodilian cervical muscles in Table 10.2 and Fig. 10.5.
III. FUNCTION A. General Feeding Behavior Crocodiles widely exceed the size of all other recent reptiles. In addition to the giants Crocodylus porosus and Crocodylus niloticus, which can reach lengths up to 10 m, remarkably small forms are found in genera Paleosuchus and Osteolaemus. The adults of these species slightly exceed a body length of 1.5 m. Because the hatchlings of the larger species generally measure only 25 cm, they possess the ability to enlarge their birth length by 40 times, by far the largest increase in length of all vertebrates (Wermuth, 1953). This massive growth is accompanied by a change in diet and feeding behavior. Hatchling crocodilians subsist predominantly on aquatic and shoreline insects of many species; during their first years they progress through a phase of frog and fish eating. Only animals exceeding 2 m rely heavily on eating mammals and birds, but do not lose the ability to feed on smaller prey. This change
TABLE 10.2 Origin and Insertion of the Cervical Musculature of Caiman crocodilus" and Alligator mississippiensis^ Muscle
Origin
Insertion
m. transversospinalis capitis
Medial part: tips of the neural spines of C-9 to the axis Lateral part: fascia of the shoulder region
Tendinous to the dorsal surface of the processus postoccipitalis (suture of the supraoccipital and squamosal)
m. spino-capitis posticus
Tendinous aponeurosis from the posterolateral surface of the neural spine of the axis Dorsolateral surfaces of the neural spines of C-3 to C-7 Tips of neural spines of C-8 and C-9
By lateral tendon to the lateral edge of the exoccipital Fleshy to the tip of the processus paraoccipitalis
m. transversospinalis cervicis
Complex system of tendinous aponeurosis, which is attached to the prezygapophyses and neural spines of C-3 to C-9 through the dorsal intermuscular septum
Fleshy to the posterior surface of postzygapophyses of C-4, C-3, and atlas
m. epistropheocapitis
Anterolateral surface of the neural spine of the axis
Fleshy to the occipital surface of the supraoccipital bone and exoccipital-squamosal suture
m. longissimus cervicis
Tendinous aponeuroses that connect the dorsal intermuscular septum to the prezygapophyses The undersurfaces of prezygapophyses C-4, C-5, C-6, and C-7
By tendon on the postzygapophyse of the atlas (same aponeurosis as the first two bundles of the m. transversospinalis cervicis)
m. longissimus capitis superficialis
Lateral surface of the neural arches of C-5 to C-8
m. longissimus capitis profundus
Dorsolateral surface of the neural arches from C-6 to the atlas Transverse processes of C-7 to C-3
By a narrow, strong tendon on the lateral surface of the processus paraoccipitalis Flat aponeurosis to the edge of the basioccipital bone Fleshy to the medial surface of the basioccipital bone
m. iliocostalis cervicis
Anterolateral surface of the cervical ribs Anterior surface of the myoseptum arising from the posterior edge of the ribs
Posterior surface of the myoseptum of the next anterior rib Tendinous aponeurosis to the posterior surface of the atlas-rib
m. iliocostalis capitis
Distal half part of atlas-rib First septum of the m. iliocostalis cervicis
By transversal tendon to the ventral edge of paraoccipital process
^Author's research. ^From Seidel (1978) and Frey (1988).
348
Johan Cleuren and Frits De Vree
in diet during development is known for all crocodilian genera (Cott, 1961; Dodson, 1975; Pooley and Cans, 1976; Taylor, 1979; Schaller and Crawshaw, 1982; Davenport et al, 1990). Examination of hundreds of stomach samples from all sizes and from many species documents that the crocodilian diet not only undergoes considerable changes with size and age, but also with habitat (Pooley, 1989). In brackish water, estuaries, and lagoons, young crocodilians feed principally on insects, as well as on mud and fiddler crabs, mud prawns, shrimps, molluscs, mudskippers, and a variety of small marine fishes. This is supported by the study of Davenport et al. (1990), which showed that C. porosus feeds on a wide range of invertebrates and vertebrates, including beetles, crabs, prawns, and small mammals. Freshwater species subsist largely on tadpoles, frogs, freshwater snails, fishes, small mammals, and possibly a greater variety of insect life (Pooley, 1989). Numerous observations suggest that crocodilian species, especially the short-snouted ones, reduce very large prey items to a size convenient for further transport by jerking and twisting motions (Pooley and Gans, 1976). Commonly, large crocodilians seize some part of the prey and then rotate about their longitudinal axis. This tactic is not used by crocodilian species that feed exclusively on small prey animals, such as small mammals, fish, crabs, and prawns, as no reduction takes place with these prey items. Davenport et al (1990) observed that C. porosus, when feeding on prey caught in the water, manipulates its prey wholly in air, with the mouth being kept clear of the water. This aerial manipulation is also present in Caiman crocodilus, but only during transport and deglutition (Cleuren and De Vree, 1992). This feeding behavior is necessary because swallowing large items of food under water would involve breaching the seal between the tongue and the palate, thereby flooding the esophagus. The feeding habits of A. mississippiensis (Busbey, 1989) and C. niloticus (Pooley and Gans, 1976) are very similar to those described for C. crocodilus (Cleuren and De Vree, 1992). B. Feeding Stages The feeding behavior in the different crocodilian species is very similar. As in other lower tetrapods (De Vree and Gans, 1989, 1994), the feeding sequence is generally subdivided into three phases: ingestion, intraoral transport, and swallowing (Cleuren and De Vree, 1992). Ingestion involves capturing of the prey with sideways bites in which the head is rotated around the ver-
tical axis so that the food item is grasped by the teeth on one side of the snout. After capture of the prey, the prey is repositioned within the mouth by a series of inertial bites. A forceful bite then follows with a welldefined crushing phase in which the prey is killed and crushed. This subset of several repositioning bites, followed by a killing/crushing bite, is repeated until the prey has been killed and reduced to a size suitable for further transport. Repositioning and crushing subsequences may be interrupted for several seconds, while the prey is held between the median teeth (Cleuren and De Vree, 1992). Instead of killing by crushing and biting, crocodilians frequently carry the struggling prey to the water to hold it submerged until struggling ceases. After drowning, repositioning and crushing follow. During intraoral transport, the prey is first oriented lengthwise and is then shifted headfirst between upper and lower jaws with rapid inertial repositioning bites. The prey is then moved toward the back of the oral cavity. Once the food is well within the gular region, swallowing cycles shift the prey into the esophagus. Swallowing is not an inertial process; the jaw apparatus plays only a minor role. It involves cyclic movements of the hyobranchial apparatus that push the food item into the esophagus. Once the prey reaches the entrance of the esophagus, it is squeezed more posteriorly by compressive movements of the gular region (Cleuren and De Vree, 1992). C. Kinematics 1.
Overview
Several researchers have examined the kinematics of the inertial feeding process in crocodilians: C. crocodilus (Van Drongelen and DuUemeijer, 1982; Cleuren and De Vree, 1992) and A. mississippiensis (Busbey, 1989). The following kinematic description is based mainly on cineradiographic records of C. crocodilus (Cleuren and De Vree, 1992). They characterize four types of bite: inertial repositioning bites, inertial killing/ crushing bites, inertial transport bites, and swallowing cycles. Each bite type is characterized by a specific displacement pattern of the neck, cranium, and hyolingual apparatus and has a unique gape profile. As in other lower tetrapods (Bramble and Wake, 1985), characteristic changes in the gape profiles are used to subdivide each open-close cycle into several kinematic phases: slow opening (SO), fast opening (FO), fast closing (FC), slow closing (SC), and crushing or power close (Cr). The change in the features of these kinematic phases or their absence/presence is used for the identification of the bite types throughout the feeding
349
10. F e e d i n g in Crocodilians
sequence. Gape profiles of repositioning bites involve a FO, FC, and an unclear SC phase. Killing/crushing bites have a profile identical to that of repositioning bites, but the terminal SC and Cr phase is well defined. Transport bites begin with a SO phase, followed by FO and FC, but lack SC. Swallowing cycles show only an opening and a closing phase. Busbey (1989) uses another terminology in the subdivision of the gape profile into kinematic phases: closing, closed, open 1, static open, and open 2. All phases can probably be correlated to the phases observed in C. crocodilus (Cleuren and De Vree, 1992): open 1 and static open corresponding with the slow-open phase, open 2 with fast opening, closing with fast closing, and closed with the crushing phase. Crocodilian gape profiles lack the stereotypy observed in mammalian chewing cycles and are affected strongly by the position of the prey between the teeth. This is represented in the gradual modification of the gape profile throughout the feeding sequence and results in the presence of intermediate bite types in the transition phases from manipulating/crushing to intraoral transport and from intraoral transport to swallowing. Repositioning and killing/crushing bites occur in the first part of the feeding sequence whenever the prey is caught between the opposed tooth rows. Once the prey has been killed and oriented lengthwise, only transport cycles occur, followed by swallowing. The number of bites depends on food type and size; feeding on large prey involves a larger number of bites. In C. crocodilus, the influence of prey size is very clear for repositioning bites (10 bites for newborn mice to 100 for large juvenile mice), whereas it is less obvious for the other bite types (between 2 and 5 killing/crushing, 3 and 8 transport, and 5 and 12 swallowing bites; Cleuren and De Vree, 1992). 2. Inertial Bites Inertial feeding in C. crocodilus (Cleuren and De Vree, 1992) proceeds at a rate of approximately three to four bites per second, which means that the average bite lasts less then 300 msec. Only killing/crushing bites exceed this duration. The onset of inertial bites involves a slow elevation of the neck and cranium, with the cervical elevation being accompanied slightly later by cranial elevation. During this time the hyolingual apparatus is slowly lifting the prey dorsad (Fig. 10.6). Fast opening results from the rapid elevation of the cranium and neck and the depression of the mandible (Fig. 10.6) and is nearly always associated with a rapid sideways head movement. Further elevation of the cranium and neck dur-
Killing/crushing
Transport
Swallowing
FIGURE 10.6. Major events occurring during killing/crushing, transport, and swallowing in Caiman crocodilus. Positions of the head, hyolingual apparatus, and prey are drawn from a cineradiographic sequence taken at 50 frames per second. The first six frames (left) of the killing/crushing sequence (0-200 msec) are identical to those observed during a repositioning bite. Frames 1 to 4 for each biting mode represent positions during the slow-opening or interbite phase. Frame 5 shows maximum gape for each mode; at this point the head is maximally pulled backward. Frame 6 shows the ensuing closed jaw position. The forward thrust of the head at this time is accompanied by a depression of the cranium in killing/crushing and repositioning bites and by a further cranial elevation in transport bites. Frame 7 represents the crushing phase in killing/crushing bites and the beginning of the interbite or slow-opening phase (cf. frames 1-4) in transport and swallowing bites. The position of the prey (juvenile mouse) is represented by the stippled oval. From Cleuren and De Vree (1992), with permission.
ing fast opening retracts the head and accelerates the prey backward. Depression of the lower jaw allows the prey to be pushed rapidly upward by the hyolingual apparatus. During the following fast-closing phase, the neck and cranium are depressed abruptly and return to their starting position with a rapid reversal of the lateral movement. Depression of the neck and cranium
350
Johan Cleuren and Frits De Vree
thrusts the head forward toward the backward moving prey. The prey is grabbed by the jaws at a more posterior position when the mandible is elevated. Fast closing retracts the hyolingual apparatus rapidly ventrally. As the jaws touch the prey, its resistance decreases the closing velocity, which marks the beginning of the slow-closing phase. Fiead and neck are depressed further during slow closing. a. Repositioning and Killing/Crushing Bites Repositioning and killing/crushing bites begin by an elevation of the neck and head, during which time the hyolingual apparatus is protracted anterodorsally until it supports the prey (frames 1-4 in Fig. 10.6). This dorsal movement of the hyobranchium and of the posterior part of the tongue has a greater excursion (approximately twice as much) than that of the anterior part. This slightly elevates the tongue posterior to the prey, forming a kind of bowl surrounding it. During the following fast opening, the prey is pushed rapidly upward (frame 5 in Fig. 10.6). The elevation of the neck and head and the upward movement of the hyolingual apparatus impart a backward and upward acceleration to the prey. This disengages the prey from the teeth so that it is "floating'' backward between the jaws. The hyolingual apparatus retracts ventrally immediately after reaching its maximal dorsal position. Fast closing then starts as the neck and cranium are depressed simultaneously, thrusting the head forward in the opposite direction of the backward moving prey. During this phase the tongue and hyobranchium are retracted further posteroventrally (frames 6 and 7 in Fig. 10.6). The buccal cavity thus enlarges so that the prey can be caught in a more advantageous position. h. Transport Bites In transport bites, the slow-opening phase is obviously subdivisible into a SO I phase in which the gape increases rather fast and a SO II phase in which there is only a slight increase in gape. The tongue and hyobranchium initially move to their maximal anterior position and slightly dorsad, pushing the prey against the palate and resulting in a depression of the lower jaw, thus increasing the gape (slow open) (shown in frames 1-3 of Fig. 10.6). They then move posteriorly, shifting the prey slightly backward into the pharynx (frame 4 of Fig. 10.6). Fast opening then follows. The upward displacement of the basihyal and the posterior part of the tongue during FO is 30-50% less than that seen during repositioning and killing/crushing bites. This reflects the more posterior position of the prey as the tongue is pushing the prey against the palate. In the fast-closing phase of transport bites the cranium is lifted further while the neck is already depressed (Fig. 10.6). Half-
way through this phase, the cranium will have reached its maximal elevation before it is depressed. Further elevation of the cranium places the jaws into a maximum vertical position, thus increasing gravitational effects. These effects associated with the ventral displacement of the hyolingual apparatus facilitate transport of prey into the pharynx and toward the esophagus. In contrast to repositioning and killing/crushing bites, transport bites show greater ventral displacement of the hyolingual apparatus during fast closing (frame 6 in Fig. 10.6). This can be explained by a partially passive movement and depends on the size and form of the prey. As the lengthwise oriented prey is compressed between the palatine and the buccal floor, it deflects the hyolingual apparatus downward during jaw closure. This movement is absent in the former bite types because the prey is oriented perpendicular to the tooth rows and thus does not exert a push against the buccal floor. The gape at the beginning of the transport bite decreases gradually in subsequent transport bites. This decreasing gape reflects the further shift of the prey into the pharynx, which no longer obstructs the closing of the jaws. 3.
Swallowing
Swallowing starts as soon as the prey has reached a position in which the hyobranchial apparatus lies anterior to it; cyclic movements then push the prey into the esophagus. Swallowing cycles consist of active protraction and retraction of the hyoid apparatus. The tongue passively follows the hyobranchium movement but does not participate in swallowing. Swallowing cycles differ from inertial bites in having a longer interbite interval between two subsequent cycles. Swallowing cycles start with a forward displacement of the hyobranchial apparatus until it reaches a position anterior to the prey. The hyobranchium then moves slightly posterodorsad to reach the anterior end of the prey (shown in frames 1 and 2 of Fig. 10.6). Its rapid posteroventrad retraction forces the prey into the esophagus during the opening and closing phase (frames 3-6 of Fig. 10.6). Halfway through retraction of the hyobranchial apparatus, the jaws open slightly to facilitate the passage of the prey. The opening phase mainly reflects depression of the lower jaw accompanied by a slight depression of the head. At the end of hyobranchial retraction, the mouth is closed by elevating the mandible relative to the simultaneously elevating cranium, which pushes the prey further into the esophagus. After mouth closure, the hyobranchial apparatus restarts its forward displacement, completing its cyclic movement.
351
10. Feeding in Crocodilians D . Role of the Hyolingual Apparatus in the Inertial Feeding Process A lot of confusion has existed on the use of the hyolingual apparatus during inertial feeding and swallowing. Sewertzoff (1929) described the tongue in crocodiles as being unable to move independently because it is fixed so firmly to its base, and he believed it to be the most primitive of reptilian tongues. The morphological relation of the tongue, hyobranchial apparatus, and buccal floor led to the assumption that the crocodilian tongue can be elevated and depressed, following passively the raising and lowering of the buccal floor. Based on the form and position of the basihyal and the anterior cornua in Gavialis gangeticus, Sondhi (1958) suggested that they are mainly responsible for the dorsoventral movements of the buccal floor, whereas the function of the posterior cornua lies in their support for the hyoglossus muscles. Busbey (1989) assigned the raising of the buccal floor in A. mississippiensis to a constriction of the m. intermandibularis, consequently enabling the tongue to immobilize food items against the palate. Generally, it is assumed that the tongue cannot be protruded and thus does not aid in manipulation and anteroposterior prey transport. Busbey (1989) ascribed a rather passive role to the tongue during the transport of food through the pharynx. The hyobranchial apparatus was supposed to be specialized for sealing the pharynx, but is not supposed to support the tongue (Busbey, 1989). However, Busbey (1989) observed the hyoid cornua pressing against the skin of the throat after the prey was partially transported into the pharynx. He also noticed that the hyobranchium may move in several small orbits, or move backward during this phase, although prey transport may not be obvious. Similar observations were done during swallowing in submerged Crocodylus porosus (Davenport et al, 1990). The presence of this stage was revealed only by throat movements, as the teeth were held tightly together. Thus, although both researchers report movements of the tongue and hyobranchial apparatus during feeding in A. mississippiensis (Busbey, 1989) and C. porosus (Davenport et al., 1990), they both assumed that the dorsoventral movements of the tongue are a passive result of the movements of the buccal floor and do not play an active role in the inertial process. In contrast to all this, cineradiographic recordings of Caiman crocodilus (Cleuren and De Vree, 1992) revealed an active role for both the tongue and the hyobranchial apparatus. The tongue aids in the inertial feeding process by pushing the prey item dorsally during the upward acceleration of the craniocervical complex, just prior to its release (Fig. 10.6). This upward
acceleration and velocity of the cranium and neck must be sufficiently rapid to overcome the downwardly directed gravitational acceleration on the food object (Gans, 1969). In other inertial feeders, the upward acceleration is imparted exclusively to the food item by an upward and backward thrust of the cranium and neck. However, in crocodilians the upward motion of the hyolingual apparatus during the FO-phase assists the posterodorsal thrust of the craniocervical region. It imparts an upwardly directed acceleration to the prey and thus increases its "upward" kinetic energy. As a result, the food item is pushed higher and thus takes longer to travel up and back downward to its starting point. Crocodilians can use this additional time in shifting their head into the most advantageous position to catch the prey, facilitating food transport. Movements of hyobranchium and tongue change gradually with position of the prey relative to the hyobranchial apparatus. A major change in hyobranchial movement occurs whenever the posterior end of the prey is right above it. Thereafter, the hyobranchium shows reduced dorsal movement during the fastopening phase and increased ventral movement during the FC phase. All repositioning, killing/crushing, and the first (1-4) transport cycles occur prior to this transition, and the other transport cycles and swallowing cycles take place thereafter. During transport of the food through the pharynx, the tongue is depressed, giving the throat and floor of the mouth the appearance of a large sack, which opens up the pharynx, despite the protruding gular fold. In the meantime, lifting of the head facilitates gravitational transport (Busbey, 1989; Cleuren and De Vree, 1992). Whenever a large proportion of the prey is in the pharynx, the jaws close and the hyolingual apparatus is retracted, after it is placed in front of the prey, effectively squeezing the prey posteriorly into the esophagus. The hyolingual apparatus thus also plays a vital role during swallowing.
E. Motor Patterns 1. Jaw Musculature Activity patterns of the jaw musculature are examined in C. crocodilus (Van Drongelen and DuUemeijer, 1982; personal observation), in A. mississippiensis (Busbey, 1989), and in C. niloticus (personal observations). The following description mostly contains data from an extensive study of C. crocodilus, which covered the complete feeding process from ingestion to deglutition (Cleuren, 1996), and supported by data on C. niloticus (personal observations). The fragmentary results of other authors will also be discussed.
352
J o h a n C l e u r e n a n d Frits D e Vree
In all bites (Fig. 10.7), jaw opening is achieved by a depression of the lower jaw by contraction of the m. depressor mandibulae. Simultaneously, the upper jaw is elevated by the contraction of several dorsal cervical muscles (see later). According to Van Drongelen and DuUemeijer (1982), jaw opening in C. crocodilus is
Mean electromyograms MDM MAMP MAMES MAMEP MPTA MIM MPST
Capture
MDM MAMP MAMES MAMEP MPTA MIM MPST
Killing/crushing
MDM MAMP MAMES MAMEP MPTA MIM MPST
Repositioning
MDM MAMP MAMES
n=l
n=12
n=31 1
—
:
Transport
MAMEP MPTA MIM MPST
n=12
MDM MAMP MAMES
Swallowing
MAMEP MPTA MIM MPST
n=16 800
1000
Time (ms) FIGURE 10.7. Mean electromyograms of the jaw muscles for capture, killing/crushing, repositioning, transport, and swallowing in Caiman crocodilus. The activity of each muscle is subdivided into three bursts; however, not every bite type contains all three bursts. In the adductors, burst 1 is the activity that is usually present during fast closing. Burst 2 is high-level activity specific for the pulsatile activity during the crushing phase (capture and killing/crushing). The postburst is low-level activity, primarily seen after full closure of the jaws. The height of each block is related to the intensity of the activity; full height equals maximal activity. Time zero is determined by the maximal gape at the end of the fast-opening phase (MDM, m. depressor mandibulae; MAMP, m. adductor mandibulae posterior; MAMES, m. adductor mandibulae externus superficialis; MAMEP, m. adductor mandibulae externus profundus; MPTA, m. pterygoideus anterior; MIM, m. intramandibularis; MPST, m. pseudotemporalis).
mainly accomplished by the contraction of the cervical rauscles, as it is rarely accompanied by activity of the m. depressor mandibulae in their experiments. This finding was supported by the fact that their animals constantly kept the lower jaw in a horizontal position. However, in our experiments the depressor muscle always shows major activity, which results in fast opening of the jaws (Fig. 10.7). This rapid increase in gape is always accomplished by both a lower jaw depression (by the m. depressor mandibulae) and a cranial elevation (by the dorsal cervical muscles) as shown by the profiles of both kinematical characteristics (Fig. 4 in Cleuren and De Vree, 1992). In all examined species, the m. depressor mandibulae shows low-level activity, simultaneous with the activity in the jaw adductors during jaw closure (Fig. 10.7). This activity reaches a peak at the start of fast closing, falls off during further closing, and might peak again at jaw closure (Busbey, 1989; personal observations). Van Drongelen and DuUemeijer (1982) assigned a strain-regulating function to this activity peak during the crushing phase. However, Cleuren et ah (1995) showed activity levels of the depressor muscle going from 0 to 19%, whereas jaw adductors showed recruitment levels from 27 to 100% (measured relative to the maximal observed value per muscle). This, together with the fact that the m. depressor mandibulae only forms a small component (7.3%, see Cleuren et al, 1995) of the total physiological cross section and the proportional role of the bite force (a negative component of only 0.3% on the total bite force), makes a strain-regulating hypothesis questionable. Simulations with a static bite model (unpublished data) support this argument. Making the depressor muscle maximally active during crushing, simultaneous with all jaw adductors, results in a significant increase in joint force (10-34% for gape 0°, 10-23% for gape 10°, depending on the angle of the food reaction force), accompanied by a decrease in bite force (4% for gape 0° and 10°), and this in a phase where bite force seems to be crucial. The main differences in muscle activity can be found in the activity patterns of the jaw adductors (Fig. 10.7). In the first part of the feeding sequence, fast jaw closure is achieved by the simultaneous contraction of most jaw adductors, in which they show 10 to 30% of their maximal activity. In acquisition bites and killing/crushing bites (Fig. 10.7), this is followed by a crushing phase, which is characterized by the presence of pulsatile high-level activity of all closers (70 to 100%). Similar tetanic potentials were first described in lizard jaw muscles by Cans and De Vree (1986) in Trachydosaurus rugosus during crushing of snails. This mechanism of synchronized tetanus proves to be widely used by vertebrates in crushing hard prey.
10. Feeding in Crocodilians Toward the end of the feeding sequence, during intraoral transport, and swallowing, fewer adductors remain active during the closing of the jaws and their activity decreases gradually (Fig. 10.7). In swallowing cycles, only four jaw adductors remain active (Fig. 10.7). The duration of low-level activity (less than 10% of the maximal activity) after full jaw closure increases toward swallowing. Thus, based on activity pattern, jaw closers can be divided into two groups: group one, containing muscles that show major activity throughout the complete feeding process—the m. adductor mandibulae posterior, the deep part of the m. adductor mandibulae externus, the m. intramandibularis, and the m. pseudotemporalis. Group 2 contains the superficial part of the m. adductor mandibulae externus, and the m. pterygoideus anterior and posterior, which are only active when group one muscles show high levels of activity (Fig. 10.7). Histochemical data for the American alligator (Sato, 1992) revealed that the muscles of the first group consist of a large amount of red muscle fibers, whereas those of the second group consist of a high percentage of white fibers. As the distributions of fiber types is not homogeneous in crocodilian jaw muscles, precise knowledge of the placement of the probing electrodes is of crucial importance to reveal the relationships between activity pattern and histochemical characteristics. Van Drongelen and DuUemeijer (1982) described unusual activity patterns, which are unique in vertebrates, i.e., during prey drowning, the jaw adductors become active before jaw opening and remain active during opening and closing. During some other unspecified feeding activity, all adductors activate prior to jaw opening and remain active during opening, and the PTA, MAME, and MAMP become silent during jaw closing. None of these patterns resemble those reported by Busbey (1989) or were observed during our experiments.
2. Cervical
Musculature
Seidel (1978) included theoretical predictions on the function of the cervical musculature in his study of the axial musculature of A. mississippiensis, based on the morphological-topographical characteristics. He assumed that lateral movements of the skull and neck are caused by unilateral contractions of the ipsilateral side of certain muscles. However, electromyographical examination of the major neck muscles in C. crocodilus revealed that sideways movements are always produced by simultaneous activation of more than one muscle and by an interaction of the ipsilateral and contralateral side (personal observations).
353
All cervical muscles show extensive activity during the inertial feeding process. During swallowing, most muscles become silent. Straight lifting of the cranium during fast opening of the jaws is mainly caused by a bilateral contraction of all three muscles of the transversospinalis system: the m. transversospinalis capitis, m. spinocapitis posticus, and m. epistropheo-capitis. When head elevation is accompanied by a sideways shift, the m. transversospinalis capitis and m. iliocostalis capitis show a bilateral activity, but with the ipsilateral muscle at a higher intensity. The ipsilateral sides of the m. spinocapitis posticus and m. longissimus capitis superficialis then also show high-level activity, whereas the contralateral sides show low-level activity or are completely inactive. This bilateral difference is only obvious during large lateral head moveraents in the m. epistropheo-capitis. Simultaneous with cranial elevation, the neck is lifted by bilateral contraction of the m. transversospinalis cervicis, m. longissimus cervicis, and m. iliocostalis cervicis. Unilateral contractions of the m. iliocostalis cervicis cause lateral flexion of the neck. During intraoral transport, "cervicis" muscles seize their activity at the end of the fast-open phase, whereas "capitis" muscles remain active. This results in a continued elevation of the head and a static position of the neck in the fast-closing phase. Because the m. longissimus profundus only shows light activity during jaw closure, the downward displacement of the skull during this phase can probably be ascribed mainly to gravitational forces. The m. longissimus profundus, the only muscle positioned to function as a depressor of the neck, only shows major activity during the lifting of the neck, probably revealing a stabilizing function for the occipital joint. The fast elevation of the heavy crocodilian cranium causes immense inertial forces at the level of this joint. As manipulation of the neck of fixated specimens revealed no mechanical restriction of dorsal movement (150° backward elevation in Virchow, 1914), these forces cannot be counteracted by the presence of bony structures or ligaments. Simultaneous activation of an antagonistic muscle allows dosing of the dorsal movement, and thus minimizes inertial forces occurring at the occipital joint. The same principle is observed in contralateral muscles during sideways movements of the cranio-cervical complex. The occipital joint is thus stabilized during all head movements. In dorsad movements, inertial forces are counteracted by the most ventral neck muscle, in lateral movements by a contraction of the contralateral muscle. Most cervical muscles have multiple functions; muscles with cranial elevation as a major function also assist in the lateral head flexion or in the elevation of the neck and vice versa. Table 10.3 summarizes the role
354
Johan Cleuren and Frits De Vree
TABLE 10.3 Importance of Cervical Muscles in Elevation and Lateral Movements of the Cranium and Neck and in the Depression and Rotation of the Head Cranium Muscle
Elevation
TSC
***
SCP EC
Lateral
* ***
** *** *
*
***
Neck
DepresElevasion Rotation tion
*
***
LCe LCS
***
LCP ICCe ICC
Lateral
*
*
*
*sf*
*
cervical muscles play in the elevation and lateral movements of the neck and cranium, as well as in axial rotation and depression of the cranium. IV. EVOLUTION A. Diet in Relation to Skull Morphology: Long Snouted versus Short Snouted Among crocodilian species, variation in diet is strongly reflected in skull morphology and head shape. Many adaptations to diet have both an ecological and a biomechanical explanation. The development of an elongated snout as in ichthyophagous species, such as the gharials {Tomistotna schlegelii and Gavialis gangeticus) and the crocodiles {Crocodylus johnstoni, C. novaeguinea, and C. cataphractus), proves to be advantageous biomechanically. Long and narrow snouts offer less resistance to the water when sweeping sideways to catch fish and are also effective in probing for crabs in subterranean burrows (Pooley, 1989). However, a slender snout is too fragile to take large prey, which explains the evolution toward broad snouts in crocodilian species feeding on a more general diet. Besides a change in head shape, many other morphological characteristics of the crocodilian skull can be related to the feeding behavior, many of them incorporating engineering principles to increase the mechanical strength. In his work on inertial feeding, Gans (1969) stated that the forces required to accelerate either food or the head will tend to induce equal reaction forces on the body and tend to shift it. For floating or swimming crocodilians, these reaction forces are critical, as they must keep their body from shifting while their head manipulates the prey. According to Gans (1969), sus-
pended animals show two ways to counteract these reaction forces. They can induce equivalent but opposed forces with the appendages or they have the evolutionary option of decreasing the ratio of head mass to body mass. Terrestrial forms are able to transmit reaction forces to the substratum, utilizing the friction of the contact zones (Gans, 1969). However, in order to minimize emerging reaction forces, terrestrial species specializing in cranioinertial feeding might also be expected to show modification for the reduction of the craniocervical mass (Bramble and Wake, 1985). This tendency to very lightly built crania is demonstrated in carnivorous lizards and birds that regularly use cranioinertial feeding, such as Varanus, Tupinambis (Smith, 1982), and pigeon (Zweers, 1982; Zweers et al, 1994). At first sight, this tendency toward cranial slenderization seems to be absent in crocodilian skull. A mechanical explanation for its firmly built appearance can be found in the substantial forces that occur during the jerking and twisting manoeuvers in feeding behavior, especially seen in short-snouted crocodilian species. As this tactic is not used by the crocodilian species that feed exclusively on small prey animals, one might expect to find a lighter built cranium in exclusively ichthyophagous "long-snouted" crocodilians (Cleuren and De Vree, 1992). A first confirmation of this assumption can be found in the presence of the cranial osteodermic relief. This relief increases the mechanical strength of the flattened skull, and consequently also its resistance to fracture. Because the longitudinal crests of the osteodermic relief of short-snouted crocodilians coincide with the loads that occur in twisting of prey, this principle may apply to crocodilians (lordansky, 1973). It is further supported by the fact that crests are absent in all longsnouted crocodilians. Apparently, osteodermic relief, indeed, increases the mechanical strength of the crocodilian skull. According to Lanyon and Rubin (1985), local increases in mass can avoid points of potentially high stress. Many crests, lines, tuberosities, or local cortical thickenings in the crocodilian skull may thus be interpreted as local reactions that reduce stress concentrations. Their absence in longirostrine crocodilians supports the hypothesis that their cranium is not subjected to stresses equivalent to those in brevirostrine species (Cleuren and De Vree, 1992). Analogous to this, one would also expect that younger animals possess more lightly built skulls, as they commonly feed on relatively small animals, such as insects, fish, crabs, and small rodents. Mook (1922a) confirmed this hypothesis by stating that the skulls of young crocodilians show a relatively smiooth surface. In medium-sized skulls, the pitting is deeper and the surface rough. In old animals, the pitting and rugose
10. Feeding in Crocodilians condition of the surface of many of the bones is emphasized greatly. Other specific characters, such as oblique ridges in front of the orbits, median elevations of the snout, and facial ridges, are usually also emphasized in older animals. This ontogenetic variation in age also applies to the thickness of the bone (Mook, 1922a). Dodson (1975) registered the belief that isometry in the length of the skull with respect to body length and positive allometry of the jaws is an adaptation to everincreasing size of prey. The shape and proportions of the upper temporal fenestrae change dramatically during ontogeny and differ in long-snouted and shortsnouted species. Gavialis, for example, shows enlarged temporal fenestrae in contrast to the short-snouted Alligator. Apparently, the demand for thicker and more solid bone, and an increase in the osteodermic relief, is also related to age and may be associated with a change in diet. This hypothesis is further supported by observations of Webb and Messel (1978). They observed that Crocodylus porosus over 120 cm in total length eat more vertebrates; the change in diet is associated with broadening of the head. The secondary bony palate has considerable importance for the respiratory function, enabling the animal to breathe from the surface even when the mouth is open underwater. In addition to this function, it has an important second advantage in that it braces the long snout against heavy stresses engendered by the capture of large prey (Buffetaut, 1989). Ferguson (1981a,b) also recognized the engineering principle of tubular reinforcements in the form of the palate of the American alligator. Consistent with maintaining a light anterior snout, maxillary sinuses may serve as extra strengthening for the flat skull. B. Skull Morphology in Relation to the Bauplan of Jaw Adductors and the Cervical Musculature The evolutionary potential of feeding behavior is limited by the mechanical restrictions (physical arrangement of the muscles, tendon, bone, joint, etc.) on the capabilities of the musculature due to the morphology of the cranial and vertebral bones. With a given morphology, the crocodilian head and neck must adequately perform such varied and mechanically complex functions, such as capture of prey, manipulation and swallowing of prey, nest building, and care of the hatchlings. Solutions to all these problems require a high degree of refined adaptation. Considering the large sizes attained by some crocodilians, it will become apparent that the crocodilian neck is a highly specialized structure that meets its functional demands (Seidel, 1978).
355
The posthatchling skull undergoes a verticalization process caused by a downward growth of the basisphenoid and basioccipital (Romer, 1956) and the transformation of the diffuse sinus into a basicranial tube system (Tarsitano, 1985,1988). Among all living crocodilians, this verticalization process only differs in Gavialis gangeticus in the presence of a large, midsagittal, anterior pocket above the braincase. The verticalization of the basisphenoid also requires the verticalization of the posterior end of the pterygoid, which is likely to change the angle of force application of the m. pterygoideus and allows a larger volunie of this muscle (Tarsitano, 1985, 1989). Together with the enlarged volume in G. gangeticus, the origin of the pterygoid muscle is shifted posteriorly by a posterior elongation of the processus retroarticularis (Fig. 10.2). This elongation also lengthens the moment arm of the depressor muscle and therefore improves its force efficiency. The morphology of the m. pseudotemporalis is related to the head shape, as it has been enlarged at the expense of the m. pterygoideus anterior in longirostrine species, whereas the reverse trend is observed in brevirostrine crocodilians. This suggests quicker and stronger muscle contractions in short-snouted species (lordansky, 1964). This is demonstrated further in the G. gangeticus (Tarsitano, 1989) and Tomistoma schlegelii (Kalin, 1933), which differ from other living crocodilians in having larger, fairly vertical supratemporal fenestrae (Fig. 10.2). In Gavialis, the expanded volume of the supratemporal fenestra allows thickening of the pseudotemporalis muscle, a strategy for muscle enlargement that differs from the one observed in tomistomines (despite the presence of enlarged supratemporal fenestrae) and other crocodylines and alligatorines. In these forms the width of the skull table is reduced and the supratemporal fenestrae have moderate dimensions or may be closed entirely by an expansion of the parietal, postorbital, and squamosal (sometimes in Osteolaemus, usually in Paleosuchus; Kalin, 1933), which results in a different method of housing the m. pseudotemporalis. In most eusuchian genera, except in Crocodylus porosus (Tarsitano, 1989), the pseudotemporal muscle is elongated posteriorly as it extends posteriorly within the temporal fenestra along a pulley or trochlear surface (Lakjer, 1926; Schumacher, 1973). lordansky (1964) promoted the hypothesis that the cartilago transiliens can be used as a locking mechanism to keep the jaws open without muscle activity of the depressor muscle. He supposed that this behavior functions during thermoregulation, as it is frequently observed in crocodilians when lying on a riverbank, sunbathing with fully opened jaws. With a wide gape, the pterygoidal flanges are displaced from under the
356
Johan Cleuren and Frits De Vree
mandibular adductor tendon and are placed immediately above and in front of this tendon. By activation of the m. intramandibularis, the tendon is drawn forward, downward, and outward, together with the attached cartilago transiliens. This fixes the position of the widely opened jaws because the passage of the pterygoidal flanges under the adductor tendon will be blocked. Activation of dorsal adductor muscles will normalize the position of the tendon and thus close the jaws (lordansky, 1964). Van Drongelen and Dullemeijer (1982) extended this locking function to the closed jaw position as they observed no adductor activity during drowning of prey. This suggests that the cartilago can be manipulated in dorsal and ventral direction by the attached muscles. To test their hypothesis, they injected amalgam in the cartilago to determine its position on lateral X-ray photographs. With fully closed jaws, the cartilago is positioned caudodorsal to the edge of the torus transiliens (ventral extension of the pterygoids), whereas with fully opened jaws, it lies ventral to it. These observations, together with manipulation experiments, confirmed their hypothesis. During electromyographical experiments on C. crocodilus (personal observations), it was observed that jaws can be kept open without any muscle activity and that a change in jaw position is often preceded by activity of the m. intramandibularis. These observations further support the hypothesis. Schumacher (1973) and Ferguson (1981) suggested that the m. intramandibularis, which runs exclusively in the inferior dental canal alongside the persistent Meckel cartilage, may act as an antagonist to the mandibular adductors during jaw opening, thus preventing distortion of the cartilago transiliens and the mandibular adductor tendon. Contraction of the m. intramandibularis during adduction would stretch the fibers of the mandibular adductor muscles (m. pterygoideus anterior and m. adductor mandibulae externus profundus), thus broadening the length-tension curve of these muscles, giving them a larger range of isometric contraction. This hypothesis is supported by irregular activity in the m. intramandibularis prior to a changing gape (Busbey, 1989; personal observations). The position of the lower jaw in crocodilians is controlled by eight jaw adductors and one opener, all pulling in different directions. To allow determination of the role each muscle plays during the five observed bite types and to estimate the force that each muscle can exert, a static bite model was developed by Sinclair and Alexander (1987). Their simulation was based on the assumption that muscle forces are proportional to the physiological cross section and that all muscles are fully active simultaneously. As these conditions conflict with reality, Cleuren et ah (1995) improved the
model by using the actual recruitment levels of the jaw muscles, which were determined by a quantitative electromyographical analysis. Given a range of orientations of the food reaction force, the magnitude of the bite force and the orientation and magnitude of the joint forces were calculated. Their model showed that bite forces are largely deterniined by changes in the orientation of the muscle forces, a finding with two important biological implications. By using different compartments of complex muscles, crocodilians are able to modulate bite force extensively, and slight morphometric differences may determine a shift in the feeding ecology of closely related species. The model also showed that the different direction of pull and the modification of the force level of each individual muscle not only affected bite force but also determined the magnitude and angle of the forces occurring at the level of the jaw joint. The orientation of the joint forces always fits within the heavily ossified triangle at the level of the jaw suspension. The anteriorly pointing leg of this triangle is formed by the massive quadrate, which inclines medially. The quadratojugal and jugal form the other leg, i.e., the lower temporal bar, a strong bony strut pointing rostrally in a sagittal plane. This means that joint forces in C. crocodilus result in compressive loading of both bony legs of the triangle, irrespective of the orientation of the food reaction forces. The more they point forward, the higher the lower temporal bar will be loaded, as forward pointing food reaction forces coincide with increasing joint force magnitudes and decreasing joint force angles, which tend to come in line with the lower temporal bar. The sagittal position of the lower temporal bar ensures pure axial loading during symmetrical muscle activity (observed during holding and crushing; Cleuren et ah, 1995). In the case of the quadrate, the joint forces participate in a bending moment too. This might explain why, despite the much smaller axial loading, the quadrate appears to be stronger built than the lower temporal bar. As the orientation in which the caiman can expect and thus must also absorb joint forces is highly determined by its jaw muscle morphology, reinforcements of the skull can be limited to the essential structures and therefore minimized. This also fits into the hypothesis of Bramble and Wake (1985) that terrestrial species specializing in cranioinertial feeding are expected to show modification for the craniocervical mass in order to minimize inertial forces on the body. A study of the form of the lower jaw of C. crocodilus (Van Drongelen and Dullemeijer, 1982) provided further evidence for this hypothesis. For each food-intake action, the amount of bony material necessary to resist muscle force and the required specific shape of the mandible was calculated
10. Feeding in Crocodilians
in this study. Except for the difference in the level of the mandibular fenestrae, the integrated shape highly resembled the form of the actual mandible. As further distribution of material is impossible due to the support for dentition, the articulation, and the muscle attachments, a minimization of the required material for the same mechanical demand is only possible in the area between the joint and the dentition. This finding is evident in the position and shape of the mandibular fenestra. As the cervical muscles play a very important role in the rapid head movements during inertial feeding, they may also affect the morphology of the crocodilian skull. The presence of the many powerful neck muscles could result in the evolutionary tendency toward enlargement and reinforcement of the insertion area, and thus an increase in the size and mass of the occipital region. However, this is not observed as it would influence the inertial feeding process negatively. As the insertions of the neck muscles are nearly always tendinous, an increase in musculature is permitted without the need for an expansion of the attachment areas. The placing of these occipital insertion points as far as possible away from the occipital condyle also achieves the maximum length of lever arm. Given the length of the fibers of the cervical-occipital muscles, an increase of the moment arm results in a mechanical advantage in terms of reduced force requirements.
References Bels, V. L., M. Chardon, and K. V. Kardong (1994) Biomechanics of the hyolingual system in Squamata. Pp. 197-240. In: Biomechanics of Feeding in Vertebrates. Vol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer Verlag, Berlin. Boschma, H. (1920) Das halsskelet der Krokodile. Leiden: Tijdschrift der Nederlandsche Dierkundige Vereniging, serie 2, 18:85-123. Bramble, D. M., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge. Briihl, C. B. (1862) Das Skelett der Krokodilinen, dargestelt in 20 Tafeln (Icones ad Zootomiam illustrandam). Wilhelm Braumiiller, Wien. Buffetaut, E. (1989) Evolution. Pp. 26-41. In: Crocodiles and Alligators. C. A. Ross (ed.). Merehurst Press, London. Busbey, A. B. (1989) Form and function of the feeding apparatus of Alligator mississippiensis. J. Morphol. 202:99-127. Cleuren, J. (1996) Functionele morfologie van het craniocervicaal en hyolinguaal apparaat van Caiman crocodilus tijdens de inertiele voedselopname. Unpublished Ph.D. Dissertation, University of Antwerp (UIA), Belgium. Cleuren, J., and F. De Vree (1992) Kinematics of the jaw and hyolingual apparatus during feeding in Caiman crocodilus. J. Morphol. 212:141-154. Cleuren, J., P. Aerts, and F. De Vree (1995) Bite and joint force analysis in Caiman crocodilus. Belg. J. Zool. 125:79-94. Cott, H. B. (1961) Scientific results of an inquiry into the ecology
357
and economic status of the Nile crocodile {Crocodylus niloticus) in Uganda, northern Rhodesia. Trans. Zool. Soc. Lond. 29:211-356. Cundall, D. (1983) Activity of head muscles during feeding by snakes: a comparative study. Am. Zool. 23:383-396. Cundall, D., and C. Cans (1979) Feeding in water snakes: an electromyographic study. J. Exp. Zool. 209:189-208. Davenport, J., D. J. Grove, J. Cannon, T. R. Ellis, and R. Stables (1990) Food capture, appetite, digestion rate and efficiency in hatchling and juvenile Crocodylus porosus. J. Zool. Lond. 220:569-592. Densmore, L. D., Ill, and R. D. Owen (1989) Molecular systematics of the order Crocodilia. Am. Zool. 29:831-841. De Vree, R, and C. Cans (1989) Functional morphology of the feeding mechanisms in lower tetrapods. Pp. 115-127. In: Fortschritte der Zoologie, Vol. 35. (H. Splechtna and H. Hilgers (eds.). Gustav Fisher Verlag, New York. De Vree, R, and C. Cans (1994) Feeding in tetrapods. Pp. 93-118. In: Biomechanics of Feeding in Vertebrates, Vol. 18. (V L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Dodson, P. (1975) Functional and ecological significance of relative growth, in Alligator. J. Zool. (London) 175:315-355. Edmund, A. G. (1969) Dentition. Pp. 117-200. In: Biology of the Reptilia, Vol. 1. (C. Gans, A. d'A. Bellairs and T. S. Parsons (eds.). Academic Press, London. Ferguson, M. W. J. (1981a) The value of the American alligator. Alligator mississippiensis as a model for research in craniofacial development. J. Craniofac. Genet. Dev. Biol. 1:123-144. Ferguson, M. W. J. (1981b) The structure and development of the palate in Alligator mississippiensis. Arch. Oral. Biol 26:427-443. Ferguson, M. W. J. (1984) Craniofacial development in Alligator mississippiensis. Symp. Zool. Soc. Lond. 52:223-273. Frey, E. (1984) Aspects of the biomechanics of crocodilian terrestrial locomotion. Pp. 93-98. In: Third Symposium on Mesozoic Terrestrial Ecosystems W. E. Reif and F. Westphal (eds.). Attempto Verlag, Tubingen. Frey, E. (1988) Anatomie des Korperstammes von Alligator mississippiensis Daudin (Anatomy of the body stem of Alligator mississippiensis Daudin). Stuttgarter Beitr. Naturk. Ser. A. Nr. 424:1-106. Furbringer, M. (1922) Das Zungenbein der Wirbeltiere, insbesondere der Reptilien und Vogel. Abh. Heidelb. Akad. Wis. 11:1-164. Gans, C. (1961) The feeding mechanism of snakes and its possible evolution. Am. Zool. 1:217-227. Gans, C. (1969) Comments on inertial feeding. Copeia 4:855-857. Gans, C. (1992) Electromyography. Pp. 175-204. In: Biomechanics Structures and systems. A. A. Biewener (ed.). Oxford Univ. Press, Oxford. Gans, C , and F. De Vree (1986) Shingle-back lizards crush snails shells using temporal summation (tetanus) to increase the force of the adductor muscles. Experientia 42:387-389. Gauthier, J. A., A. G. Kluge, and T. Rowe (1988) Amniote phylogeny and the importance of fossils. Cladistics 4:105-209. Gnanamuthu, C. P. (1937) Comparative study of the hyoid and tongue of some typical genera of reptiles. Proc. Zool. Soc. Lond. 107:1-63. Hecht, M. K., and S. F. Tarsitano (1983) On the cranial morphology of the Protosuchia, Notosuchia and Eusuchia. N. Jb. Geol. Paleont.Mh. 11:657-668. Hoffstetter, R., and J. P. Gasc (1969) Vertebrae and ribs of modern Reptiles. Pp. 201-310. In: Biology of Reptilia, Vol. 1. C. Gans, A.d'A. Bellairs, and T. S. Parsons (eds.). Academic Press, London. lordansky, N. N. (1964) The jaw muscles of the crocodiles and some relating structures of the crocodilian skull. Anat. Anz. 115:256280. lordansky, N. N. (1973) The skull of the crocodiles. Pp. 201-262. In Biology of the Reptilia, Vol. 4. C. Gans (ed.). Academic Press, London.
358
J o h a n C l e u r e n a n d Frits D e Vree
Kalin, J. A. (1933) Beitrage zur vergleichenden Osteologie des Crocodilidenschadels. Zool. Jb. Abt. Anat. 57 (4): 535-714. Kalin, J. A. (1955) Crocodilia. Pp. 695-783. In: Traite de Paleontologie, Vol. 5. J. Piveteau (ed.). Masson et Cie., Paris. Kieser, J. A., C. Klapsidis, L. Law, and M. Marion (1993) Heterodonty and patterns of tooth replacement in Crocodylus niloticus. J. Morphol. 218:195-201. Lakjer, T. (1926) Studien iiher die Trigeminus-versorgte Kaumuskulatur der Sauropsiden. C. A. Reitsel, Copenhagen. Langston, W., Jr. (1973) The Crocodilian skull in historical perspective. Pp. 263-284. In: Biology of the Reptilia, Vol. 4. C. Cans (ed.). Academic Press, London. Lanyon, L. E., and C. T. Rubin (1985) Functional adaptation in skeletal structures. Pp. 1-25. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge. Lubosch, W. (1914) Zwei vorfaufige Mitteilungen iiber die Anatomie der Kaumuskeln der Krokodile. Jena. Zschr. Naturw. 51:697-706. Lubosch, W. (1933) Untersuchungen liber die Visceralmuskulatur de Sauropsiden. Morph. Jb. 72:584-666. Luther, A. (1914) tjber die vom N. trigeminus versorgte Muskulatur der Amphibien, mit einem vergleichenden Ausblick iiber den Adductor mandibulae der Gnathostomen, und einem Beitrag zum Verstandnis der Organisation der Anurenlarven. Acta Soc. sc.Fenn. 44(7): 1-115. Miall, L. C. (1878) The skull of the crocodile, a manual for students. Stud. Comp. Anat. 1:1-50. Mook, C. C. (1921a) Individual and age variations in the skulls of recent Crocodilia. Bull. Am. Mus. Nat. His. 44:51-66. Mook, C. C. (1921b) Notes on the postcranial skeleton in the Crocodilia. Bull. Am. Mus. Nat. His. 44:67-100. Mook, C. C. (1921c) Skull characters of recent Crocodilia, with notes on the affinities of the recent genera. Bull. Am. Mus. Nat. His. 44: 123-268. Poe, S. (1996) Data set incongruence and the phylogeny of crocodilians. Syst. Biol. 45:393-414. Poglayen-Neuwall, I. (1953) Untersuchungen der Kiefermuskulatuur und deren Innervation an Krokodilen. Anat. Anz. (Jena) 99:257277. Pooley, A. C , and C. Cans (1976) The nile crocodile. Sci. Am. 234: 114-124. Pooley, A. C. (1989) Food and feeding habits. Pp. 76-91. In: Crocodiles and Alligators. C. A. Ross (ed.). Merehurst Press, London. Romer, A. S. (1956) Osteology of the Reptiles. University of Chicago Press, Chicago. Rowe, T. (1986) Homology and evolution of the deep dorsal thigh musculature in birds and other Reptilia. J. Morphol. 189: 327-346. Sato, I., K. Shimada, A. Yokoi, J. C. Handal, N. Asuwa, and T. Ishii (1990) Morphology of the teeth of the American Alligator {Alligator mississippiensis): fine structure and chemistry of the enamel. J. Morphol. 205:165-172. Sato, I., K. Shimada, T. Sato, and T. Kitagawa (1992) Histochemical study of jaw muscle fibers in the American Alligator {Alligator mississippiensis). J. Morphol. 211:187-199. Schaller, G. B., and P G. Crawshaw (1982) Fishing behavior of Paraguayan Caiman {Caiman crocodilus). Copeia 1:66-72. Schumacher, G. H. (1973) The head muscles and hyolaryngeal skeleton of turtles and crocodilians. Pp. 101-199. In: Biology of the Reptilia, Vol. 4. C. Gans (ed.). Academic Press, London. Schwenk, K., and G. S. Throckmorton (1989) Functional and evolutionary morphology of lingual feeding in squamate reptiles: phylogenetics and kinematics. J. Zool. (Lond.) 219:153-175.
Seidel, R. (1978) The Somatic Musculature of the Cervical and Occipital Regions of Alligator mississippiensis. Ph.D. Dissertation, City University of New York, NY. Sewertzoff, S. A., Jr. (1929) Zur Entwicklungsgeschichte der Zunge bei den Reptilien. Acta Zool. 10:231-341. Shimada, K., I. Sato, A. Yokio, T. Kitagawa, M. Tezuka, and T. Ishii (1990) The fine structure and elemental analysis of keratinized epithelium of the filiform papillae analysis [sic] on the dorsal tongue in the American alligator {Alligator mississippiensis). Okajimas Folia Anat. Japan 66:375-392. Shimada, K., I. Sato, and H. Moriyama (1992) Morphology of the tooth of the American Alligator {Alligator mississippiensis): the fine structure and elemental analysis of the cementum. J. Morphol. 211:319-329. Shimada, K., I. Sato, and H. Ezure (1993) Morphological analysis of tendinous structure in the American alligator jaw muscles. J. Morphol. 217:171-181. Sill, W. D. (1968) The zoogeography of the Crocodilia. Copeia 1968: 76-88. Sinclair, A. G., and R. McN. Alexander (1987) Estimates of forces exerted by the jaw muscles of some reptiles. J. Zool. (London) 213: 107-115. Smith, K. K. (1982) An electromyographic study of the function of the jaw adducting muscles in Varanus exanthematicus (Varanidae). J. Morphol. 173:137-158. Smith, K. K. (1986) Morphology and function of the tongue and hyoid apparatus in Varanus (Varanidae, Lacertilia). J. Morphol. 187: 261-287. Sondhi, K. C. (1958) The hyoid and associated structures in some Indian reptiles. Ann. Zool. Agra. 2:155-240. Steel, R. (1973) Crocodilia. Handbuch der Paleontologie 16:1-116. Tanner, W. W., and D. F Avery (1982) Buccal floor of reptiles, a summary. Great Basin Nat. 42 (3): 273-349. Tarsitano, S. F (1985) Cranial metamorphosis and the origin of the Eusuchia. N. J. Geol. Palaont. 170(1): 27-41. Tarsitano, S. F, E. Frey, and J. Reiss (1989) The evolution of the Crocodilia: a conflict between morphological and biomechanical data. Am. Zool. 29:843-856. Taylor, J. A. (1979) The foods and feeding habits of subadult Crocodylus porosus Schneider in Northern Australia. Aust. Wildl. Res. 6:347-359. Troxell, E. L. (1925) Mechanics of Crocodile vertebrae. Bull. Geol. Soc. Am. 36:605-614. Van Bemmelen, J. F (1887) Beitrage zur kenntniss der Halsgegend bei Reptilien. I. Anatomischer theil. P. W. M. Trap, Amsterdam. Van Drongelen, W., and P. Dullemeijer (1982) The feeding apparatus of Caiman crocodilus, a functional-morphological study. Anat. Anz. 151:337-366. Virchow, H. (1914) Uber die AUigatorwirbelsaule. Arch. Anat. 1914: 103-142. Webb, G. J. W., and H. Messel (1978) Morphometric analysis of Crocodylus porosus from the north coast of Arnhem Land, northern Australia. Aust. J. Zool. 26:1-27. Wermuth, H. (1953) Systematik der rezenten Krokodile. Mitt. Zool. Mus. Berlin 29:375-514. Zweers, G. A. (1992) Pecking of the pigeon {Columba livia L.). Behavior 81:173-230. Zweers, G. A., H. Berkhoudt, and J. C. Vanden Berge (1994) Behavioral mechanisms of avian feeding. Pp. 241-279. In: Biomechanics of Feeding in Vertebrates, Vol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin.
C H A P T E R
11 Feeding in Paleognathous Birds CAROLE A. BONGA TOMLINSON Department of Organismic and Evolutionary Biology Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138 I. INTRODUCTION II. MATERIALS AND METHODS III. MORPHOLOGY OF THE HYOLINGUAL APPARATUS A. Neognathous Birds B. Paleognathous Birds IV. FUNCTION OF THE HYOLINGUAL APPARATUS A. Lingual Feeding in a Generalized Neognath (Meleagris gallopavo) B. Cranioinertial Feeding in Paleognaths C. Comparison of Ratite Cranioinertial and Neognathous Lingual Feeding V. EVOLUTION OF THE FEEDING SYSTEM A. Avian Phylogeny and Outgroup Choice B. Primitive Condition of the Neornithine Hyolingual Apparatus C. Changes in Feeding Function during the Theropod-Bird Transition D. Proposed Functional Evolution of Early Avian Transport Mechanisms E. Evolutionary Morphology: An Overview F. Conservation of Pattern Generation G. Phylogenetic Relationships References
1988; Padian and Chiappe, 1998; Sibley and Ahlquist, 1990). The large, sometimes giant, flightless ratites include 10 species in six extant genera {Struthio, the African ostrich; Rhea and Pterocnemia, South American rheas; Dromaius, the Australian emu; Casuarius, Nev^ Guinea and Australian cassowraries; and Apteryx, the kiw^is of New Zealand) and two extinct groups (the elephantbirds of Madagascar and Africa and the moas of New Zealand). Ratites are believed to have reached their modern pattern of distribution on southern land masses by means of vicariance and dispersal via land routes across Antarctica during the late Cretaceous and/or early Tertiary, approximately 80 to 50 million years ago (Cracraft, 1973,1974,1986,1988; van Tuinen, 1998; see also Sampson et ah, 1998). Tinamous are moderate to small-sized, volant birds (nine genera, 47 species) restricted to the Neotropics and savannas of South America. Superficially they resemble the neognathous galliforms (pheasants, fowl). Parkes and Clark (1966) proposed that a "proto-tinamou" was ancestral to all ratites and tinamous. Kurochkin (1995) listed fossil birds that he considered paleognaths that occurred worldwide in the Cretaceous and early Tertiary (see also Alvarenga, 1983; Alvarenga and Bonaparte, 1988; Houde, 1988; Houde and Haubold, 1987; Houde and Olson, 1981; Peters, 1988; Tambussi, 1995), but he included no tinamou. Feduccia (1996) stated that the earliest tinamou fossils date only from the Miocene of South America. Paleognathous birds possess a small tongue, a mostly cartilaginous hyobranchial skeleton and feed cranioinertially (Bramble and Wake, 1985; Ftirbringer, 1922; Lang, 1956; McLelland, 1979; Mtiller, 1963; Parker, 1866; Parker, 1891; Pycraft, 1900; Webb, 1957). Neognathous birds are primarily hyolingual (tonguebased) feeders. Nonetheless, cranioinertial feeding is
L INTRODUCTION Within Neornithes (modern birds), monophyly of the nominal taxa Paleognathae (ratites and tinamous; Pycraft, 1900) and Neognathae (all other modern birds, >8600 species; Sibley and Monroe, 1990) is disputed. Divergence of the two putative lineages may have occurred as long ago as 120 million years ago during the Cretaceous period (Cooper and Penny, 1997; Rambaut and Bromham, 1998; van Tuinen et ah, 1998), but there is no consensus on which group is more phenotypically primitive (see Cracraft and Mindell, 1989; Houde, FEEDING (K. Schwenk,ed.)
359
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
360
Carole A. Bonga Tomlinson
increasingly evident in neognaths as food size increases (e.g., Columba; Zweers, 1982a,b), thus all birds may be capable of using cranioinertial feeding to some extent (Bramble and Wake, 1985). Few neognathous birds are known to be obligate cranioinertial feeders (e.g., Egretta; Homberger, 1989), but all paleognathous taxa are presunied to share this behavior because the small tongue would seem to preclude lingual feeding (see Gussekloo and Zweers, 1997; Tomlinson, 1997a). Bock and Blihler (1988) proposed that tongue reduction occurred independently in at least two paleognathous groups (ostrich vs all other paleognaths) as a result of similar feeding patterns in which large food items are swallowed whole (thus reducing the functional role of the tongue). The ancestral condition for birds is assumed to be an unreduced tongue and an extensively ossified hyobranchium (Bock and Biihler, 1988). Ecological studies of paleognathous species, however, show that extant paleognaths are neither uniform in the habitats they occupy (arid grasslands and tropical forest) nor in their omnivorous diets. They choose foods of various sizes, and in many taxa, large food items are broken up by pecking before ingestion (del Hoyo et al, 1992). Furthermore, primitive character states of the avian hyolingual apparatus (tongue, hyobranchium, and musculature) remain undetermined, a problem exacerbated by the fact that the closest relevant outgroups (toothed Mesozoic birds, theropod dinosaurs, or thecodont reptiles) are extinct (see Chiappe, 1995; Cooper and Penny, 1997; Feduccia, 1994,1996; Hecht, 1985; Martin, 1983,1985,1987; Molnar, 1985; Ostrom, 1969, 1973, 1985, 1991; Padian and Chiappe, 1998; Welman, 1985). The modern avian hyobranchial apparatus differs significantly from those of other extant reptiles, yet its evolution has not been addressed. For example, the ceratohyals characteristic of nonavian reptiles (henceforth referred to simply as "reptiles" for convenience) have been lost in birds, and a novel element, the paraglossal, occurs within the tongue (Crompton, 1953). The muscular tongue of most reptiles is absent in birds and intrinsic hyolingual muscles that connect hyobranchial elements have taken the place of intrinsic lingual muscles. The hyobranchium is located in the neck region of most reptiles and retractor muscles are attached to the sternopectoral region (Busbey, 1989; Cleuren and De Vree, 1992; Delheusy et al, 1994; Kesteven, 1944; Oelrich, 1956; Schumacher, 1973; Smith, 1984, 1986; Sondhi, 1958), whereas in birds the hyobranchium is located immediately behind and beneath the mandible, and the major retractors originate on the laterocaudal surfaces of the mandible (see Baumel et ah, 1993; Bhattacharyya, 1980; Burton, 1984; Homberger, 1986; Homberger and Meyers, 1989; Zweers, 1982b). Changes in hyolingual function that accompanied
this structural transformation are also largely unknown, although the basic patterns of cranioinertial transport and swallowing are believed to be similar in birds and some extant reptiles (Smith, 1992; see Suzuki and Nomura, 1975). Nevertheless, mechanisms of intraoral transport clearly differ between birds and reptiles: reptiles possess teeth and (with the exception of crocodilians) a fleshy, muscular tongue capable of movement independent of the hyobranchium (e.g.. Smith, 1984,1988; Schwenk, 1986,1988; see Chapters 2 and 8), whereas in modern birds, teeth are absent and tongue movements depend entirely on movements of the hyobranchial skeleton (Zweers, 1974, 1982a,b). Nearly all reptiles employ some form of hyolingual feeding, with cranioinertial feeding exceptional (e.g., in crocodilians. Chapter 10; and Varanus, Chapter 8; Cans, 1969). The muscular, manipulative tongue of parrots, with highly differentiated, intrinsic hyolingual musculature, is uniquely derived and presumably associated with their ability to position seeds within the beak for husking (Homberger, 1986; see Chapter 2). Thus, the neornithine feeding apparatus is significantly different from that of other modern reptiles, and within birds there is a deep phenotypic and phylogenetic divergence between paleognathous and neognathous forms. This divergence offers an opportunity to examine evolutionary transformations in the avian feeding apparatus in the light of outgroup comparison. Although higher-level phylogenetic relationships among birds are highly contentious and largely unresolved (see Chapter 12), Fig. 11.1 presents one generally accepted phylogeny for the relationships of the paleognath taxa relative to the Neognathae. Although not all phylogenetic analyses agree with this hypothesis, it is used here because it is most consistent with form and function of the hyolingual apparatus (see later). The purpose of this chapter is twofold. First, the morphology and function of the paleognathous hyolingual apparatus are described and compared to the generalized neognathous condition. Second, these data are compared to comparable data for fossil and extant reptilian outgroups in order to determine whether the paleognathous or the neognathous condition is representative of the primitive condition for neornithine birds. The evolutionary origins of the modern avian hyolingual apparatus and of two basic types of avian feeding—ratite (obligate) cranioinertial feeding and avian lingual feeding—are discussed. IL MATERIALS A N D M E T H O D S Morphological and functional comparisons are based on dissections and cineradiographic films of
361
11. F e e d i n g in P a l e o g n a t h o u s Birds
Paleognathae
I Toothed Mesozoic Early Theropods Birds Paleognaths Emu
Other ratites
Ostrich
Rhea
1
Lithornithlds Tinamous
Neognathae
F I G U R E 11.1. Phylogenetic relationship of paleognathous birds that is most consistent with form and function of the feeding apparatus.
feeding in three ratite species (South American greater rhea, Rhea americana; Australian emu, Dromaius novaehollandiae; and African ostrich, Struthio camelus) in comparison to a neognathous species (North American wild turkey, Meleagris gallopavo). Additional anatomical data were obtained for a Chilean tinamou {Nothoprocta perdicaria). These observations were supplemented by reference to the literature wherever possible. The wild turkey is assumed to be representative of generalized, ground-pecking, lingual-feeding birds. It belongs to the order Galliformes, widely regarded as basal within the neognath clade (see Sibley and Ahlquist, 1990; Zweers, 1985, 1991a,b; Zweers et al, 1994). Morphological character state polarities were determined by comparison of data to descriptions of hyolingual structure in modern reptiles, as well as fossils of theropod dinosaurs and toothed Mesozoic birds. In addition, hyobranchial function in a lepidosaurian reptile (tuatara, Sphenodon punctatus; also see Chapter 8) is compared to bird data in an effort to discern polarities in avian functional patterns. High-speed cineradiographic films (100 fps: Struthio, Sphenodon; 200 fps: Rhea, Dromaius, Meleagris) were recorded using a Siemens cine X-ray machine with a Sirecon image intensifier attached to an Eclair GV-16 camera {Struthio and Sphenodon films were made by
earlier workers) and digitized using a Vanguard motion analyzer and custom software. Functional data for other taxa are taken from the literature on neognathous birds, reptiles, and fossil taxa, as noted. III. MORPHOLOGY OF THE HYOLINGUAL APPARATUS A general description of the neognathous hyolingual apparatus (tongue, hyobranchial skeleton, and associated musculature) given later provides a basis for comparison with the paleognathous condition that follows. Muscle terminology follows Nomina Anatomica Avium (Baumel et al, 1993) and/or Homberger and Meyers (1989) wherever possible, but modifications and descriptive clarifications are added to supplement the incomplete, inadequate, and often inaccurate literature on paleognathous taxa (Bock and Biihler, 1988; Kesteven, 1945; Webb, 1957; see Table 11.1). The terms for muscle groups (e.g., intrinsic and extrinsic hyolingual, hyolaryngeal) in neognathous taxa differ in Baumel et al (1993), Homberger and Meyers (1989), and Zweers (1982b), and here follow Baumel et al (1993), with some exceptions (see Table 11.2). Avian terminology for muscle groups, however, differs from that
362
Carole A. Bonga Tomlinson
used in reptiles and mammals, and attempts are made to clarify these differences in the text. Muscle innervation in neognathous taxa follows Bubien-Waluszewska (1981) and Kontges and Lumsden (1996) and in paleognathous taxa follows Kesteven (1945) and Webb (1957). A. Neognathous Birds The hyolingual apparatus is well documented for several lingual- and filter-feeding neognathous taxa (e.g., Bhattacharrya, 1980; Burton, 1984; Homberger and Meyers, 1989; Kallius, 1905; Zweers, 1974,1982b). Hyolingual morphology in the turkey, M. gallopavo, is used to characterize the general neognathous pattern for comparison to paleognathous species (Figs. 11.2 and 11.3). The dorsal surface of the tongue is cornified and covered by filamentous papillae (Fig. 11.2B).
The hyobranchium consists of seven elements. There are three articulated median elements: a paraglossal within the tongue, a basihyal within the tongue base, and a urohyal anteroventral to the larynx. Paired, lateral elements consist of the hyoid horns (cornua), which articulate with the lateral surface of the basihyal. Each cornu is formed by a ceratobranchial that curves around the larynx anteroventrally and laterally, articulating distally with an elongate epibranchial. The epibranchial curves upward from below the mandible to the occipital region, where it is attached to the skull by connective tissue (fascia vaginalis hyoideus; Homberger, 1986; Homberger and Meyers, 1989). In neognathous embryos, the basihyal and urohyal form a single cartilaginous anlage that later separate; the basihyal always ossifies, but the urohyal may not. The paraglossal originates as two cartilaginous "paraglossalia" that fuse and ossify last (Kallius, 1905; Fiirbringer, 1922).
TABLE 11.1 Synonymous Terms for Muscles Acting on the Hyolingual Apparatus in Paleognathous Species Described in This Study and Previous Works ^ Present study (Rhea, Struthio, and Dromaius)
Kesteven (1945) {Struthio and Dromaius)
Webb (1957) (Struthio)
Bock and Buhler (1988) (all paleognathous species)
M. constrictor colli cervicalis
Second dorsal superficial constrictor
Constrictor colli (cucullaris)
M. constrictor colli intermandibularis
Second ventral superficial constrictor, pars posterior
Constrictor colli (cucullaris)
M. intermandibularis
First ventral superficial constrictor
Mylohyoideus
M. branchiomandibularis
Hyomandibularis
Ceratomandibularis
Branchiomandibularis Genioglossus, medial slip
M. genioglossus M. genioceratohyoideus*
Geniohyoideus
M. serpihyoideus
Second ventral superficial constrictor, pars anterior
M. hyomandibularis* (== M. H. lateralis* in Dromaius) (absent in Rhea)
Genioglossus
Serpihyoideus Hyomandibularis medialis
M. hyomandibularis medialis* (absent in Rhea and Struthio)
Interhyoideus
M. cricohyoideus
? Thyro-hyoideus + ? Ceratohyoideus
M. ceratocricoideus* (absent in Rhea)
Ceratothyroideus
Genioglossus
? Stylohyoideus
' Stylohyoideus
M. basiarytaenoideus* M. ceratohyoideus
Ceratohyoideus
M. ceratoglossus
Ceratoglossus
M. hyoglossus* (sling absent in Struthio)
Hypoglossus obliquus (absent in Struthio)
''An attempt was made to standardize nomenclature in accordance with Nomina Anatomica Avium (Baumel et ah, 1993; see Table 11.2). An asterisk (*) indicates a muscle that is unique to paleognathous species. Uncertain homologies are indicated with a question mark. A blank indicates that the muscle was not described.
363
11. Feeding in Paleognathous Birds TABLE 11.2 S y n o n y m s for Avian H y o l i n g u a l M u s c l e s and M u s c l e Groups^ in Paleognathous Species, W i l d Turkey, and Other N o n s p e c i a l i z e d N e o g n a t h o u s Species^ Paleognathous species This study {Rhea, Struthio, and/or Dromaius) External hyolingual mm M. constrictor colli cervicalis M. constrictor colli intermandibularis M. Intermandibularis
Neognathous species This study (Meleagris)
Constrictor colli cervicalis Constrictor colli intermandibularis Intermandibularis
Zweers (1982b) {Columba) External hyoid mm^ Cutaneous colli
Intermandibularis ventralis caudalis Intermandibularis ventralis
Homberger and Meyers (1989) (Gallus), except as noted Superficial neck mm Constrictor colli cervicalis Gular mm Constrictor colli intermandibularis Mylohyoideus
Baumel et al. (1993) {Nomina Anatomica Avium) External hyolingual mm Constrictor colli
Constrictor colli, pars intermandibularis Intermandibularis ventralis
Extrinsic hyolingual mm Protractive extrinsic hyolingual mm Extrinsic lingual mm Extrinsic hyoid mm Branchiomandibularis Branchiomandibularis M. branchiomandibularis Branchiomandibularis Geniohyoideus (rostralis and caudalis) (anterior and posterior) (rostralis and caudalis) Genioglossus M. genioglossus Genioglossus Geniophar3mgealis (external hyoid mm) M. genioceratohyoideus* Retractive extrinsic hyolingual mm M. serpihyoideus Serpihyoideus
Serpihyoideus (external hyoid mm)
Serpihyoideus
Serpihyoideus
Stylohyoideus
Stylohyoideus
Stylohyoideus
Cricohyoideus
Cricohyoideus
M. hyomandibularis (including M. H. lat. and med.) Stylohyoideus Hyolaryngeal mm M. cricohyoideus M. ceratocricoideus* M. basiarytaenoideus* Intrinsic hyolingual mm M. ceratohyoideus M. ceratoglossus M. hyoglossus*
Cricohyoideus
Ceratohyoideus Ceratoglossus Hyoglossus obliquus Hyoglossus anterior
Extrinsic laryngeal mm Cricohyoideus
Ceratohyoideus Intrinsic hyoid mm Ceratoglossus Hyoglossus obliquus Hyoglossus anterior
Intrinsic lingual mm Ceratohyoideus
Intrinsic hyolingual mm Ceratohyoideus
Ceratoglossus
Ceratoglossus
Hypoglossus obliquus Hypoglossus anterior
Hyoglossus obliquus Hyoglossus rostralis
^Showninbold. ^An asterisk (*) indicates a muscle that is unique to paleognathous species. '^Some external hyoid muscles are listed under "extrinsic hyoid m m / '
The paraglossal remains cartilaginous throughout life in many neognaths, but ossifies in others (Homberger, 1986, 1989, 1999). In most cases the epibranchials remain cartilaginous throughout life. The paraglossal is arrow shaped in dorsal view, flattened dorsoventrally and may include cartilaginous processes (anterior and/or posterolateral), as in the turkey. It occupies most of the dorsal region of the avian tongue and is overlain by a tough, cornified, papillose epithelium (Fig. 11.2B). The remainder of the tongue consists of salivary glands and "intrinsic hyolingual muscles." Intrinsic hyolingual muscles are complex in some
neognathous taxa (e.g., parrots; Homberger, 1986), but in generalized taxa such as the wild turkey and chicken, only four muscle pairs are present (Fig. 11.3A): (1) the hyoglossus (innervated by c.n. XII) is divided into anterior and oblique segments connecting the articulated paraglossal and basihyal; the hyoglossus rostralis inserts ventrally on the anterior paraglossal process and originates on the ventral surface of the basihyal; (2) the hyoglossus obliquus originates on the ventral surface of the basihyal and inserts on the ventral surface of the paraglossal; (3) the ceratoglossus (c.n. XII) originates on the ceratobranchial rostrally and inserts by tendon on the ventral surface of
364
Carole A. Bonga Tomlinson
A.Rhea
Cb
mg
B. M e l e a g r i s
F I G U R E 11.2. Sagittal sections through the head of (A) a paleognathous species, Rhea americana, and (B) a neognathous species, Meleagris gallopavo, showing relative positions of the hyolingual apparatus and palatal bones. Superimposed on each section are the left ceratobranchial and epibranchial of the hyobranchium and the pterygoid and basipterygoid process of the skull. The position of the basipterygoid process and a mesial segment of the palatine bone are shown in relation to the pterygoid and vomer. Cartilaginous hyoid elements are stippled (note that this is the opposite of subsequent figures). Bh, basihyal; BP, basipterygoid process; Cb, ceratobranchial; Eb, epibranchial; mg, mucus glands; Pa, palatine bone; Pg, paraglossal; PR, parasphenoid rostrum; Pt, pterygoid bone; RE, rostral esophagus; sg, salivary glands; V, vomer.
the paraglossal; and (4) the ceratohyoideus (c.n. XII, VII?) originates on the medial surface of the ceratobranchial; the muscles of both sides join in a median raphe ventral to the urohyal and larynx. The floor of the mouth ventral to the free portion of the tongue contains salivary glands that lie between the intermandibularis (ventral) and the Mm. genioglossus (dorsal) (see extrinsic hyolingual muscles, later). Fibers of the intermandibularis (c.n. V) originate along the mandibular rami and join at a midventral raphe. It does not extend as far rostrocaudally as in ratites [ratites shown in Fig. 11.11; Bhattacharyya (1980) and Homberger and Meyers (1989) illustrate the intermandibularis in neognaths]. A cricohyoideus muscle (c.n. IX, X, XII), connecting the cricoid cartilage to the dorsal surface of the basihyal, is always present (Fig. 11.3B). Kinematic data suggest that muscles originating on
the mandible and inserting on the hyobranchial apparatus either protract or retract the hyobranchium relative to the mandible (see Table 11.1: "extrinsic hyolingual muscles") (Fig. 11.3C). Protractor muscles are innervated by cranial nerve XII and/or IX, whereas retractors are innervated by cranial nerve VII. The main protractors (Mm. branchiomandibularis rostralis and caudalis) originate on the mandible and run caudally to encircle the distal end of the ipsilateral hyoid horn. The genioglossus (c.n. XII, IX?) originates on the mandibular symphysis and inserts on the epithelium at the root of the tongue. This muscle is small in Meleagris and in other neognathous taxa it is either small or absent (Burton, 1984; Homberger and Meyers, 1989). Three extrinsic retractor muscles originate on the laterocaudal mandible: (1) the serpihyoideus, (2) the constrictor colli intermandibularis, and (3) the stylohyoideus. All are innervated by cranial nerve VII and have a common developmental origin with the constrictor colli cervicalis, a superficial dermal muscle that extends rostrally from the side of the upper neck and fans out in the throat region where the two sides meet in the midline (Fig. 11.3C; superficial constrictor not shown) (Kesteven, 1945; Noden, 1983a,b). Paired serpihyoideus muscles arise from the posterolateral margins of the mandible and run anteromedially to join at a midventral raphe ventral to the urohyal. The constrictor colli intermandibularis has a similar origin, with some of its fibers taking origin from the fascia overlying the serpihyoideus and the depressor mandibulae muscle (not shown) and from the tough connective tissue surrounding the external ear opening. It also inserts on a midventral raphe. Some fibers of the serpihyoideus and the constrictor colli intermandibularis overlap, with the latter more ventral. The common midventral raphe is connected to a fascial sheet that attaches to the rostroventral surfaces of the ceratobranchials and the ventral surface of the basihyal between the ceratobranchial articulations (see Homberger and Meyers, 1989). The latter two retractor muscles and their conimon raphe lie ventral to the ceratohyoideus (an intrinsic hyolingual muscle; see earlier discussion). The constrictor colli intermandibularis occurs in all reptiles and birds (Kesteven, 1944). The serpihyoideus is known only in birds and is characteristic of all known species (Kesteven, 1945). In contrast, the stylohyoideus (see later) occurs only in neognathous birds. The stylohyoideus originates on the mandible just rostral to the serpihyoideus and runs anteromedially to insert on the dorsal surface of the basihyal (Figs. 11.3A and 11.3C). It passes ventral to the branchiomandibularis caudalis and rostralis muscles and dorsal to the ceratoglossus. "Stylohyoideus" is a misno-
11. F e e d i n g in P a l e o g n a t h o u s Birds
365
F I G U R E 11.3. Hyolingual apparatus in a generalized neognathous bird, based on the wild turkey, Meleagris gallopavo (Galliformes). Stippling indicates ossification. (A) Ventral view of intrinsic hyolingual muscles and hyobranchium. (B) Dorsal view of main hyolaryngeal muscle, hyobranchium, larynx, and anterior end of the trachea. (C) Ventral view of extrinsic hyolingual muscles, hyobranchium, and mandible. Protractors are shown on the left, retractors on the right, apgp, anterior paraglossal process; Bh, basihyal; Cb, ceratobranchial; crl, cricoid (larynx); Eb, epibranchial; epr, epithelium at root of tongue; gl, glottis (larynx); MBmr, M. branchiomandibularis rostralis; MBmc, M. branciomandibularis caudahs; MCg, M. ceratoglossus; MCh, M. ceratohyoideus; MCrh, M. cricohyoideus; MGg, M. genioglossus; MHgo, M. hyoglossus obHquus; MHgr, M. hyoglossus rostralis; Mn, mandible; MSph, M. serpihyoideus; MSth, M. stylohyoideus; Pg, paraglossal; ppgp, posterior paraglossal process; tr, trachea; Uh, urohyal.
mer because birds lack a styloid process. However, it is maintained because the name has been in common use since the 19th century (refer to Table 11.2; see also Homberger, 1986; Homberger and Meyers, 1989). Other workers have noted that the stylohyoideus in neognathous birds can insert on the basihyal or the ceratobranchial (Burton, 1984). This distinction is important because muscle forces acting directly on the basihyal will have different mechanical consequences than forces acting on the ceratobranchials (see Section IV). The avian stylohyoideus as described here is known in no other vertebrate taxa and can be considered a synapomorphy of neognathous birds. 1. Summary of the Neognathous
Condition
The tongue of neognathous birds is extremely variable (McLelland, 1979), but its surface is often cornified and its length usually closely matches the length of the beak and oral cavity. In the wild turkey, filamentous papillae occur on the dorsal surface, and salivary glands occur within the base, but not the body, of the tongue.
Features of the neognath hyobranchial apparatus can be summarized as follows: (1) the paraglossal is ossified in the turkey, but it remains cartilaginous in other species (Homberger, 1986,1989,1999); (2) the basihyal is always ossified; (3) the paraglossal and basihyal meet to form a movable bony articulation; (4) the basihyal and urohyal separate during development; (5) the epibranchials are elongate and curve strongly upward; and (6) if the condition described for the chicken (Homberger and Meyers, 1989) and wild turkey (both galliforms) is representative, epibranchials connect to the occipital region by means of a complex and extensive fascia vaginalis hyoideus. Intrinsic hyolingual musculature connecting the paraglossal and basihyal is divided into hyoglossus rostralis and obliquus muscles. Extrinsic hyolingual protractor musculature (branchiomandibularis rostralis and caudalis) originates on the middle and anterior portion of the mandible and runs posteriorly to insert on the epibranchials. A third, small protractor muscle, the genioglossus, is variably present. It runs from the tip of the mandible to the base of the tongue. Extrinsic hyolingual retractor musculature consists of
366
Carole A. Bonga Tomlinson
muscles innervated by cranial nerve VII that are present in all reptiles and birds (constrictors), in all birds (serpihyoideus), or in neognathous birds alone (stylohyoideus). The stylohyoideus inserts directly onto the basihyal. B. Paleognathous Birds The small paleognathous tongue is supported by a hyobranchium that is unique among living birds—it is entirely cartilaginous except for the ceratobranchials (Figs. 11.2A, 11.4-11.6, 11.12) (Lang, 1956; MiiUer, 1963; Parker, 1866, 1891; Pycraft, 1900; Webb, 1957). Certain features of the neognathous condition described earlier are absent in all paleognathous taxa examined, including ossification of the paraglossal, basihyal, and urohyal, upwardly curved epibranchials attached to the occipital region, and a basibranchial insertion for the stylohyoideus muscle. A common paleognathous pattern occurs in extrinsic hyolingual protractor musculature that is distinct from neognathous taxa. However, variations among paleognathous taxa
Cb
F I G U R E 11.5. Dorsal view of the hyobranchium and tongue in the emu, Dromaius novaehollandiae, relative to the mandible. Note the fringed margins of the tongue. Stippling indicates ossification. Buh, basiurohyal; Cb, ceratobranchial; Eb, epibranchial; Mn, mandible; Pg, paraglossal; Tg, tongue.
occur in the form of the hyobranchium and extrinsic retractor musculature (see later). Among the four paleognathous species examined, the rhea appears to possess the most generalized condition of the paleognathous hyobranchial skeleton and displays the least complex extrinsic hyolingual retractor muscles among all birds. Thus, the condition of the hyolingual apparatus in the rhea is described first in most detail, and those features that distinguish the paleognathous condition from the basic neognathous condition are noted. Distinctive aspects of the hyolingual apparatus in the emu, ostrich, and a tinamou are then enumerated. 1. The Greater Rhea
F I G U R E 11.4. Dorsal view of the hyobranchium and tongue in the rhea, Rhea americana, relative to the larynx and mandible. Ossification occurs in the ceratobranchials only. Bh, basihyal; Cb, ceratobranchial; Eb, epibranchial; Lx, laryngeal glottis; Mn, mandible; Pg, paraglossal; Tg, tongue.
The tongue consists of a thick, rough epithelium containing mucus-secreting cells closely applied to the paraglossal; there are no salivary glands in the body of the tongue (Fig. 11.2A). The shape of the tongue mirrors the dorsoventrally flattened, arrow-like shape of the paraglossal. Three globose papillae occur at the posterolateral corners of the tongue (Fig. 11.4). In embryos, the paraglossal is the last element to form (Miiller, 1963). The basihyal is cylindrical, with its anterior end rounded. No urohyal portion projects caudal to the ceratobranchial articulations (Figs. 11.2A, 11.4, 11.7B, and 11.8B; see later). The rod-like ceratobranchials are
11. Feeding in Paleognathous Birds
367
FIGURE 11.6. Dorsal (left) and lateral (right) views of the hyobranchium and tongue in the ostrich, Struthio camelus, shown in relation to the larynx and trachea. Stippling indicates ossification; shading denotes the position and shape of the lingual pocket. Buh, basiurohyal; Cb, ceratobranchial; Eb, epibranchial; gl, glottis; Lp, lingual pocket; Ix, larynx; Pgp, paraglossalia.
the only ossified hyobranchial elements and, in resting position, lie medioventral to and parallel with the mandible. The epibranchials are short and do not extend beyond the caudal end of the mandible (Figs. 11.2A, 11.4, and 11.19C). Each cornu is attached to the vicinity of the external ear opening by means of tough connective tissue. This tissue is analogous to the fascia vaginalis hyoideus of neognaths, but it does not extend as far as the occipital region. The ventral surface of the paraglossal is covered by a tough, elastic connective tissue sheet that forms a tube-like space beneath the basihyal, completely investing the hyoglossus muscle and the basihyal. The sheet attaches to the posteroventral surface of the basihyal and the ceratobranchial-basihyal articulations
(Figs. 11.5 and 11.7B). The hyoglossus forms a muscular sling looping around the basihyal from the ventrolateral surface of the paraglossal. The hyoglossus and its connective tissue investment form the only connection between the basihyal and the paraglossal. A ceratoglossus muscle runs anteriorly from the ventrolateral surface of the ceratobranchial, over (ventral to) the connective tissue sheath covering the hyoglossus, to insert on the ventral surface of the paraglossal at two points anterior and lateral to the hyoglossus (Fig. 11.9B). According to Bock and Biihler (1988), the hyoglossus sling occurs in all tinamous and ratites, with the exception of the ostrich. An unusual connective tissue "collar" loosely encircles the anterior end of the basihyal (Figs. 11.8B and
FIGURE 11.7. Ventral view of intrinsic hyolingual muscles in ratites. (A) The emu, Dromaius; (B) the rhea, Rhea; and (C) the ostrich, Struthio. The genioglossus muscle is also shown. Stippling indicates ossification. Bh, basihyal; Buh, basiurohyal; Cb, ceratobranchial; cts, connective tissue sheath; Meg, M. ceratoglossus; Mgg, M. genioglossus; Mhg, M. hyoglossus; Pg, paraglossal; Pgp, paraglossalia.
368
Carole A. Bonga Tomlinson
MCrh MCrh
F I G U R E 11.8. Ventral view of hyolaryngeal muscles in ratites. (A) The emu, Dromaius; (B) the rhea, Rhea (caudal laryngeal papillae not shown); and (C) the ostrich, Struthio (only right paraglossalium shown). Not to scale; stippling indicates ossification. Bh, basihyal; bhs, basihyal sheath (connective tissue); Buh, basiurohyal; Cb, ceratobranchial; crl, cricoid cartilage (larynx); Eb, epibranchial; MCrh, M. cricohyoideus (right muscle only shown in A and B); Pgp, paraglossalium.
11.9B). Dorsal to the basihyal, paired, slender, straplike muscles, the basiarytaenoideus (new muscle; Table 11.1) extend caudally from the collar and attach to the epithelium of the ipsilateral arytaenoid cartilage of the larynx (Fig. 11.9B). No comparable muscles have been described in neognathous taxa. Cricohyoideus muscles insert on the rostral ceratobranchials ventrally and the mediocaudal surface of the basihyal (Figs. 11.8B and 11.9B). The ceratohyoideus is narrow and strap-like in contrast to its significant breadth in the wild turkey (Fig. 11.10; shown as if cut close to its origin). There are no salivary glands beneath the smooth
epithelium covering the floor of the mouth. The floor of the oral cavity between the mandibular rami is formed primarily by two thin muscles: the intermandibularis muscle (originating on the mandible and running medially to a midventral raphe) and a prominent, paired extrinsic protractor muscle, the genioceratohyoideus (new muscle; see Table 11.1 and later) (Figs. 11.2A, ll.lOB, and l l . l l B ) . Extrinsic protractor musculature is distinct as compared to all known neognathous taxa (Fig. 11.1 OB). The branchiomandibularis is undivided and analogous only to the "caudalis'' portion in neognathous
Buh
MCrh
F I G U R E 11.9. Dorsal view of hyolaryngeal muscles and tendons in ratites. (A) The emu, Dromaius; (B) the rhea, Rhea; and (C) the ostrich, Struthio. Not to scale; stippling indicates ossification. Bh, basihyal; bhs, basihyal (connective tissue) sheath; Buh, basiurohyal; Cb, ceratobranchial; gl, glottis; crl, cricoid cartilage (larynx); MBa, M. basiarytaenoideus; MCrh, M. cricohyoideus; Pgp, paraglossalium; TBa, basiarytaenoideus tendon; Tu, unified tendon of TBa.
369
11. F e e d i n g in P a l e o g n a t h o u s Birds
MGch MGch
MBm
MBm MSph
MSph
A
B
FIGURE 11.10. Ventral view of extrinsic hyolingual muscles in ratites. (A) The emu, Dromaius; (B) the rhea, Rhea; and (C) the ostrich, Struthio. Protractors (genioceratohyoideus and branchiomandibularis) are shown on the left of each figure, retractors are shown on the right. The genioglossus, also a protractor, arises dorsally from the genioceratohyoideus and is shown on the right of each figure. Also shown are the ceratocricohyoideus and ceratohyoideus muscles. MBm, M. branchiomandibularis; MCcr, M. ceratocricohyoideus; MCh, M. ceratohyoideus; MGch, M. genioceratohyoideus; MHm, M. hyomandibularis; MHml, M. hyomandibularis lateralis; MHmm, M. hyomandibularis medialis; MSph, M. serpihyoideus.
taxa. The genioglossus is small and arises from a much larger genioceratohyoideus muscle (new name; Table 11.1) (Fig. II.IOB), also innervated by cranial nerve XII (Webb, 1957). Both muscles were called the genioglossus by Bock and Biihler (1988) for all modern paleognaths. The strap-like genioceratohyoideus, however, originates on the mandibular symphysis and inserts on the caudoventral surface of the ceratobranchial; no analogous muscle is known in neognathous birds (Fig. 11.1 OB). The genioglossus originates from the middorsal surface of the genioceratohyoideus, runs caudally to the root of the tongue near the midline, turns laterally, and enters the tongue to insert on the posteroventral margin of the paraglossal. The genioglossus is the most ventral muscle within the tongue. Extrinsic hyolingual retractor muscles consist of the constrictor colli intermandibularis and the serpihyoideus (Figs. ll.lOB and l l . l l B ) . The constrictor colli intermandibularis originates on fascia attaching to the
caudal surface of the mandible and the external ear region, and the serpihyoideus originates on the caudalmost surface of the mandible. As in the neognathous wild turkey, both muscles insert on a midventral raphe that connects indirectly to the posteroventral surface of the basihyal and anteroventral surfaces of the ceratobranchials by means of a fascial sheet. No retractors insert on the basihyal because it is covered by the hyoglossus, the connective tissue sheath investing it and the additional "collar" of connective tissue anchoring the basiarytaenoideus muscles (see earlier discussion). Thus, the neognathous stylohyoideus muscle is absent in paleognaths. Absence of a urohyal segment in the rhea is unique among paleognaths. Miiller (1963) reported a separate globular body ("copula 11") posterior to the basihyal in the rhea embryo that is presumably homologous to the urohyal. It is apparently lost during later development, i.e., it is absent in juveniles and adults.
370
Carole A. Bonga Tomlinson
F I G U R E 11.11. Constrictor musculature in ratites. (A) The emu, Dromaius; (B) the rhea, Rhea; and (C) the ostrich, Struthio. MCci, M. constrictor colli intermandibularis; MIm, intermandibularis; Mime, M. intermandibularis, caudal portion; Mimr, M. intermandibularis, rostral portion.
2. The Emu In all paleognathous taxa other than the rhea (see earlier discussion), the basihyal and urohyal form a continuous, cartilaginous structure termed the basiurohyal (Figs. 11.4-11.8). The emu differs from the rhea additionally in lingual embellishment, hyolaryngeal connections, and aspects of the extrinsic hyolingual retractor musculature. The paraglossal, epibranchials, the fascia vaginalis hyoideus, and intrinsic and extrinsic protractor musculature are similar in the rhea and emu. Differences are described later. Protruding posterolaterally from the lateral margins of the emu tongue are five dorsoventrally flattened, curved papillae that taper to rounded points (Fig. 11.5). Basihyal-paraglossal and basihyal-laryngeal articulations are as in the rhea with some exceptions. The ceratoglossus originates more anteriorly on the ceratobranchial and inserts on the paraglossal lateral to the hyoglossus (Fig. 11.7A). In place of the basiarytaenoideus muscles, a paired tendon (T. basiarytaenoideus) originates directly on the dorsal surface of the basihyal
and a connective tissue collar is absent from the anterior end of the basihyal (Figs. 11.8A and 11.9A). An unusual hyolaryngeal muscle, the ceratocricoideus (new name; Table 11.1), originates on the ceratobranchial (medial to the posterior end of the genioceratohyoideus insertion and posterior to the ceratohyoideus insertion) and inserts on the cricoid cartilage of the larynx (Fig. 11.10A). Extrinsic protractor musculature is essentially the same as in the rhea, although the genioceratohyoideus is broader and its insertion on the ceratobranchial extends further rostrally (Fig. 11.10A). The emu possesses an extrinsic retractor in addition to the serpihyoideus and constrictor colli intermandibularis muscles (Figs. II.IOA and l l . l l A ) . The hyomandibularis (Webb, 1957) originates on the posterior end of the mandible anterior to the serpihyoideus, runs anteromedially, and divides into two portions (medialis and lateralis) that insert on the urohyal and midceratobranchial, respectively (Fig. 11.10A). Both branches pass ventral to the branchiomandibularis, but the anterior portion of the lateral branch alone passes
11. Feeding in Paleognathous Birds dorsal to the genioceratohyoideus. This configuration of the hyomandibularis has not been reported in other avian taxa. 3. The Ostrich The ostrich hyolingual apparatus differs from that of the rhea in several ways. The epibranchials are long and curve conspicuously downward, lying alongside muscles of the neck (Fig. 11.6). This condition is unique to the ostrich. The fascia vaginalis hyoideus connects the ear region to the hyoid horns at the level of the ceratobranchial-epibranchial junction. Other differences are described later. The unique tongue of the ostrich is exceedingly short and virtually fixed in position immediately anterior to the laryngeal glottis. There are no lingual papillae (Fig. 11.6). Paraglossal form is unique among birds; it comprises two narrow, wing-like cartilages called the paraglossalia (Fiirbringer, 1922; inaccurately referred to as paraglossal "processes" by Bock and Btihler, 1988). The anterodorsal surface of each paraglossalium is attached by connective tissue to the anteroventral surface of a very broad basiurohyal. The paraglossalia do not contact one another in the midline. The paraglossalia project posterolateral^ and somewhat ventrally, but the tips curve upward so that they lie in a plane dorsal to the basiurohyal. Thus, as compared to other birds, the paraglossal is displaced ventrally and laterally relative to the basihyal. Only in parrots is the paraglossal also formed by two separate elements, but these are ossified and located rostral to the basihyal (Homberger, 1986). According to Fiirbringer (1922), the ostrich paraglossalia retain the form and position of the cartilaginous anlagen (paraglossalia) of the adult paraglossal present during embryonic development in all birds (see Kallius, 1905). This suggests the possibility that the condition of the ostrich paraglossal arose through paedomorphosis (evolutionary juvenilization) (Elzanowski, 1986). The position of the paraglossalia relative to the basiurohyal in the ostrich leaves the anterior tip of the basiurohyal to form the tip of the tongue. The lingual epithelium encloses a "lingual pocket" on the dorsum of the tongue with an opening facing posteroventrally (Fig. 11.6). The pocket appears grossly to be lined with the same type of epithelium found on the tongues outer surfaces and is similar to the rhea. The structure and conformation of the lingual pocket suggest that ancestrally there was a portion of the tongue anterior to the basiurohyal that eventually folded over the base of the tongue. This scenario suggests that the lining of the pocket represents the ancestral dorsal surface of the
371
tongue and that the epithelial surface exposed dorsally in the ostrich was ancestrally the ventral surface of the tongue. The paraglossalia are located within the ventralmost portion of the tongue and extend to the ventrolateral margins of the pocket. The lingual pocket changes shape during intraoral transport (see later), apparently in response to muscular action, suggesting that the ostrich tongue is specialized for an as yet unexplained biological role (Bock and Biihler, 1988). The ostrich tongue appears to have been secondarily derived from a more general paleognathous condition as evident in the rhea. The hyoglossus muscle is absent in the ostrich. The ceratoglossus inserts on the ipsilateral paraglossalium anterodorsally, intervening between the paraglossalium and the anterior basihyal (Fig. 11.7C). The genioglossus inserts on the posterior surface of the paraglossalium (Figs. 11.7C and ll.lOC). The cricohyoideus has a broad zone of insertion on the dorso- and ventrolateral surfaces of the basihyal (Bock and Biihler, 1988). Thus, this "hyolaryngeal" muscle is located partially within the tongue in the ostrich, a condition unknown in other taxa (Figs. 11.8C and 11.9C). Extrinsic protractor musculature is essentially the same as in the other paleognaths. Insertion of the genioceratohyoideus on the ceratobranchialis occurs more anteriorly than in the rhea (Fig. 11.IOC). Regarding extrinsic retractor muscles, the serpihyoideus (Fig. 11.8D) inserts directly onto the ventral surface of the cricoid cartilage, a configuration unique among known avian taxa. The identity of the muscle was determined by its origin on the mandible—anterior to the origin of the constrictor colli intermandibularis and caudal to the origin of the hyomandibularis— and by its position ventral to the branchiomandibularis (compare Figs. 11.IOC and l l . l l C ) . An additional retractor, the hyomandibularis, originates on the posterolateral surface of the mandible anterior to the serpihyoideus and inserts on the ceratobranchial anterodorsally (Figs. II.IOC and l l . l l C ) . Along the way it passes ventral to the branchiomandibularis and dorsal to the genioceratohyoideus. Webb (1957) called this muscle the hyomandibularis medialis (see Table 11.1). Simplification of the name was justified by the fact that there is no lateral branch. Recall, however, that in the emu the hyomandibularis is divided into lateral and medial moieties (Fig. II.IOA; see earlier discussion). Bock and Biihler (1988) may have referred to the ostrich hyomandibularis as the stylohyoideus (Table 11.1). The latter name, however, is reserved for the muscle in neognathous taxa that inserts on the basihyal (Table 11.2; see earlier discussion).
372
Carole A. Bonga Tomlinson
4. The Chilean Tinamou The tongue and lingual epithelium in this small bird resemble that of the rhea, with the exception that the tongue tip ends in a rounded point and there are no lingual or laryngeal papillae present (Fig. 11.12A). The paraglossal is paper thin and its margins are scalloped. The lingual epithelium is thick relative to the paraglossal and is not scalloped, but it follows the general outline of the paraglossal (compare in Figs. 11.12A and 11.12B). The paraglossal of Nothoprocta contrasts strongly with that of another tinamou, Tinamus (Parker, 1866), which appears to resemble that of the rhea. The floor of the oropharyngeal cavity is deeply folded and forms a "shelf" that curves from the rostral end of the mandible to the posterior end of the oropharynx, around the tongue and larynx (Fig. 11.12A). In addition, a thickened epithelial fold forms another shelf between the tongue and the larynx. This shelf encloses a pocket that is lined by the oral epithelium. The pocket extends rostrally beneath the base of the tongue
but dorsal to the basiurohyal, thus it lies within the lingual base (the lingual-base pocket; Fig. 11.12A; compare Figs. 11.12A and 11.12B; see also Figs. 11.2A and 11.2B). The 'Tingual-base pocket" differs from the "lingual pocket" in the ostrich by virtue of its position, which is entirely ventral, or caudoventral, to the paraglossal and body of the tongue. A lingual-base pocket has not been previously reported in birds. Connection of the paraglossal and basiurohyal by means of connective tissue and the hyoglossus is the same as in the rhea and emu. In the region between its articulations with the ceratobranchials, the basiurohyal is partially ossified. The distal ends of the short epibranchials extend slightly posterior to the mandible, where they curve only slightly upward and attach by means of the fascia vaginalis hyoideus to the ear region (Fig. 11.12B). The short epibranchials oi Nothoprocta are in marked contrast to the elongate epibranchials depicted for Tinamus (Parker, 1866). Extrinsic hyolingual musculature is similar to that of the rhea (Fig. 11.12B). Presence or absence of the basiarytaenoideus muscles, a basihyal collar and the condition of the cricohyoideus (presumed to be present), could not be determined. 5. Summary of the Paleognathous
FIGURE 11.12. Tongue, hyolingual apparatus, and mandible in a tinamou, Nothoprocta perdicaria. (A) Dorsal view of the tongue and lingual-base pocket (see text) relative to mandible, larynx, and anterior end of the trachea; position of the pocket ventral to the tongue is denoted by a dashed line. (B) Ventral view of extrinsic hyolingual muscles. Except for the genioglossus, protractors are shown on the left of figure (with the exception of the genioglossus, which is shown on the right), retractors are shown on the right. A dashed circle denotes the attachment site of the sling-like hyoglossus muscle on the ventral surface of the paraglossal (see description of the rhea in text). Stippling indicates ossification. Buh, basiurohyal; Cb, ceratobranchial; Eb, epibranchial; crl, cricoid cartilage (larynx); sas, attachment site for muscular sling; epf, epithelial fold on floor of oral cavity; gl, glottis; Ibp, lingual-base pocket; MBm, M. branchiomandibularis; MGch, M. genioceratohyoideus; MGg, M. genioglossus; MSph, M. serpihyoideus; Pg, paraglossal; Tg, tongue.
Condition
The paleognath taxa described earlier are assumed to be representative of the group as a whole, but this is not demonstrated. The ratite cassowaries (Casuarius) and kiwis (Apteryx) and 46 additional species of tinamou remain undescribed, but based on Pycraft's (1900) description of the hyobranchium and tongue, these are unlikely to deviate significantly from the conditions described here. Because Australasian ratites are likely to resemble one another, the description of the emu condition may approximate that of Casuarius and Apteryx (Cooper et al, 1992). The lesser rhea {Pterocnemia pennata) is assumed to be similar to the greater rhea (JR. americana) described previously, and the close relationship of the tinamous (Prager and Wilson, 1976) suggests that the description for N. perdicaria is generally applicable, except where noted. The tongue in all paleognathous taxa examined here is composed of rough epithelium (in the rhea, it contains mucous cells) with few papillae, which appears to represent the ancestral paleognathous condition. The fringed lateral margins on the tongue in the emu may be autapomorphic, but conditions in other Australasian ratites are unknown. In the ostrich, the tongue appears to be secondarily reduced and a pocket (lingual pocket) forms within the lingual epithelium. The lingual pocket is uniquely derived in Struthio. A different epithelial "pocket" forms within the lingual base of
11. Feeding in Paleognathous Birds the tinamou, but these structures do not appear to be homologous. It is unknown whether the tongue-base pocket is present in other tinamou species. Globose papillae posterolateral to the tongue are found in the rhea only and may represent an autapomorphy for rheas. However, the condition in Pterocnemia is unknown. The hyobranchial skeleton in the paleognathous ratites demonstrates a consistent pattern in which the paraglossal is cartilaginous, the basihyal or basiurohyal is cartilaginous, the basiurohyal forms a single structure, the paraglossal does not join the basi(uro)hyal by means of an articulated joint, epibranchial curvature is usually downward and never upward, and the fascia vaginalis hyoideus is less extensive than in neognaths and connects the hyoid horns (ceratoepibranchials) to the ear region. Tinamous show a condition intermediate between ratites and neognaths in having more ossification in the hyobranchium and a slight upward curvature at the tips of the epibranchials. The overall similarity between ratites and tinamous, however, suggests that they descended from a common ancestral condition. The hyobranchial skeleton in the emu may represent the ancestral paleognathous condition because it contains the full complement of hyobranchial elements known to occur in the group, including the species described earlier, as well as a kiwi (Parker, 1891), a cassowary, and another tinamou (Parker, 1866). The hyobranchial skeleton in the rhea {Rhea) is virtually identical to that in the emu except that a urohyal portion is absent due to its disappearance posthatching (Mliller, 1963) (it is unknown if this is also true for the lesser rhea). The paraglossal in the tinamou, N. perdicaria, differs from the putative ancestral type in its possession of scalloped margins, but this trait may not be characteristic of all tinamous—the paraglossal in Tinamus is shown to be similar to the rhea by Parker (1966). In the ostrich, the paraglossal comprises two separate elements, the paraglossalia, which occur ventral and lateral to the basihyal, seemingly displaced from the position of the paraglossal in all other birds. It is thus autapomorphic for Struthio. Also uniquely derived in the ostrich are elongate epibranchials, which curve strongly downward. In all paleognathous taxa except the ostrich, the intrinsic hyolingual musculature is similar—an undivided hyoglossus sling ensheathed by connective tissue forms the sole connection between the paraglossal and the basihyal. This form of connection between the paraglossal and the basihyal seems to represent the ancestral paleognathous condition, with loss of the hyoglossus derived in the ostrich. Extrinsic hyolingual protractor musculature consists of the same three muscle pairs in all paleognaths
373
(branchiomandibularis, genioceratohyoideus, genioglossus), one of which (the genioceratohyoideus) is unknown in neognathous taxa. The genioglossus inserts on the paraglossal (or paraglossalium) in all paleognaths. These shared features are likely to represent the ancestral paleognathous condition. Extrinsic hyolingual retractor musculature varies, but in no instance does a retractor insert on the basihyal. In the rhea and tinamou there are no retractor muscles other than the serpihyoideus. This may represent the ancestral paleognath condition (see later). The presence of the hyomandibularis in the ostrich and its division into medial and lateral portions in the emu are derived relative to the proposed ancestral condition. Muscular and/or tendinous connections between the basihyal and the arytaenoid cartilage are unknown in other tetrapods. Thus the presence of a basiarytaenoideus muscle (or tendon) may be a synapomorphy of paleognaths, but its presence in a tinamou could not be confirn\ed in the present study. Thus a basihyalarytaenoid connection may be unique to ratites or may have been present in the common ancestor of ratites and tinamous. A ceratocricoideus muscle connecting the ceratobranchial to the cricoid cartilage is found uniquely in the emu and ostrich. The ancestral condition is clearly absence of such connections, as evident in the rhea and tinamou. IV. FUNCTION OF THE HYOLINGUAL APPARATUS The following descriptions focus primarily on movement patterns of the basi(uro)hyal (protraction, retraction, orbit) and the hyoid horns (depression, elevation) as measured relative to the mandible. As such, movements of the hyobranchium are shown independent of mandibular movement (as if the mandible was stationary). Tongue position is dependent on the position of the basihyal. Extrinsic hyolingual muscles act on the hyoid horns (cerato-epibranchials) to affect movement, but in neognathous species (such as the turkey) they can act on the basihyal directly. Synovial joints between the ceratobranchials and the basi(uro)hyal permit considerable movement, but the hyoid horns move symmetrically, each in concert with the other, suggesting bilaterally symmetric motor patterns in the hyobranchial muscles. The ceratoepibranchials act as third-class levers as they move the basi(uro)hyal. The caudal suspensorium of each epibranchial by fascia serves as a movable fulcrum (= primary fulcrum) for each lever (Fig. 11.13). Curvature of the epibranchials, location of their fulcra, and
374
Carole A. Bonga Tomlinson
Primary hyoid fulcra —
BasihyaH»ratobranchial articulation viewed from lateral —
Fascial attachment site of caudal end of epibranchial
Hyoid horn (comu; lateral view)
External ear region
Occipital region
Rostral
Rostral Loose fascia
Basihyal
Protraction
#
Late protraction
/
/ B a s i h y a l path
Basihyal orbit
Paleognath
Retraction
Neognath
F I G U R E 11.13. Effect of primary hyoid fulcra on motion of the comua (curved bars) and basihyal (black circles) during an intraoral transport cycle. Ratite cranioinertial feeding is shown on the left (with a 'loose fulcrum") and neognathous lingual transport shown on the right (with a ''firm fulcrum"). Solid bar in paleognath shows position during early protraction, in neognath during retraction. Heavy dashed bars indicate fully protracted positions. Light dashed bar shows fully retracted position in a paleognath. Arrows indicate direction of basihyal orbit.
insertion sites of protractor and retractor muscles thus circumscribe the range of possible movements potentially exhibited by the hyoid horns in response to forces exerted by extrinsic muscles originating on the mandible. One consequence of this restricted movement is that the hyoid horns are seen to depress and elevate relative to the mandible (see Figs. 11.17 and 11.21; discussed later). Two other types of hyolingual movement pattern are observed. Intrinsic hyolingual muscles flex the tongue, but flexion is more important in lingual feeding by neognathous birds than in ratite cranioinertial feeding, as noted earlier. Second, the distance between the larynx and the basi(uro)hyal fluctuates during the hyolingual cycle, i.e., during protraction this distance increases whereas during retraction it decreases. Retraction is pronounced in the wild turkey and limited in the ratites, probably reflecting the importance of this movement during hyolingual transport in neognathous feeding. Because the mandible depresses and elevates (= jaw
cycle) as the hyolingual apparatus moves, the hyolingual cycle {= protraction and retraction) is superimposed on the jaw cycle (Figs. 11.15,11.16,11.19,11.20, and 11.22). The jaw cycle is indicated by gape distance (distance between rostral tips of the upper and lower beaks) in Figures 11.15 and 11.22. Because gape distance in birds is a function of movements by both mandible and upper beak, movement of each beak tip relative to the cranium is also shown (Figs. 11.16, 11.19, and 11.20). Hyolingual position is measured from the point of articulation between the ceratobranchials and the basi(uro)hyal; protraction and retraction of the hyobranchium are measured relative to the mandible, as noted previously. Tongue movement patterns mirror those of the basi(uro)hyal. As the mandible is depressed and elevated, so too are the hyoid horns relative to the mandible. The summation of hyobranchial protraction-retraction movements and hyoid horn depression-elevation movements on the basihyal results in a regular, cyclic orbit relative to the mandible (Fig. 11.14). Distinct kinematic
11. Feeding in Paleognathous Birds
375
F I G U R E 11.14. Basihyal orbit relative to a fixed mandible during intraoral transport in neognathous (A) and paleognathous (B-D) birds. (A) Wild turkey, Meleagris; (B) emu, Dromaius; (C) rhea, Rhea; and (D) ostrich, Struthio. Position of the basihyal is indicated when the beak tips release the food (R), at maximum jaw gape (Mx), and when the mandible completes elevation (C). In the three ratites, early protraction (epr) precedes food release, and late protraction (Ipr) occurs during a jaw open-close cycle; in the turkey, all protraction takes place while the jaws are closed.
patterns involving these three movement parameters occur in lingual and cranioinertial feeding. Descriptions are based on lateral views of the head as recorded on high-speed cineradiographic film. Kinematic patterns are based on digitized points (Figs. 11.15, 11.16, 11.19, 11.20, 11.22, and 11.23), traced frame sequences (Figs. 11.13,11.17,11.18, and 11.21) or a combination of the two (Fig. 11.14). "FiyolinguaP' refers to both tongue and hyoid and therefore includes the paraglossal, whereas "hyoid" is used here to refer to the parts of the hyobranchial skeleton [the basi(uro)hyal and/or the cerato-epibranchials] acted on by extrinsic hyolingual protractor and retractor muscles originating on the mandible. A. Lingual Feeding in a Generalized Neognath (Meleagris gallopavo) Five movement parameters of lingual feeding are shown in Figs. 11.15 (small food transport) and 11.16
(large food transport)—head, food (relative to ground and passage through the oral cavity), hyobranchium, and gape. After ingestion with the tips of the beak, a small food item (pellet; Purina Turkey Chow) is transported intraorally by the tongue in stepwise fashion from beak tips to the posterior palate, or pharynx. Usually three steps are required to complete transport to the pharynx. Each step is accomplished with one complete hyolingual cycle (retraction-protraction) and one gape cycle (open-close) (see later). In the initial step, food is transported by the retracting tongue tip as the mandible is depressed and the tongue is lifted from the floor of the oral cavity. In subsequent steps, food is pushed caudally by the upwardly bulging midregion of the retracting tongue. Between each retraction phase, the mandible is held closed as the tongue is protracted beneath the food item, pushing it against the palate and completing the hyolingual cycle. The food is held in place by posteriorly protruding palatal papillae. The food becomes more rounded as
376
Carole A. Bonga Tomlinson
14H 13 O
1?
o> E 10 o
Head
k_
CD O
c (fl (fi •O
"(5 o tQ
>
8 6 4
Beginning of lingual tranport
2 H
Food dropped
11 m 11111
Gape
-0.5 H
$
1 M1111111111M1111111111111111111111111111111111111111111111111111111111111111111 n 111111111111 n n 11111 n I'
-1-5-
-2.0 H
-
-2 5-
-3.0 H j i i i i i i i i M i i i M i i i i i [ i i i i i i i i M i i i i i i i i i i | i i i i i M i i i i i i i i i i i i i | i i i i i i i i i i i i i i i i i i i i | i i i i i i i i i m i i i i i i i i | i i i i i i i i i i m i 11
Beak tips
OH CO CD
1A 2 -I
CD o c CO
Food
Retraction
Protraction
^
3 4
o
5 H
I
6 - ^ I 11 I 1111 M I M I I I M I I I I 11 M I I 11 M I 11 I M M I 11 I I I 11 I M I 11 11 I M 11 M I I 11 I 11 I I M M I I M M I I 11 11 I I I I I i I i I I I I i 1 11 I I I I I I i I I I i I 0
Time
(1 frame = 5 msec)
60
80
100
FIGURE 11.15. Kinematic profile of avian lingual transport of a small food item in the neognathous wild turkey, Meleagris gallopavo. (Top) Vertical displacement of the head and food item. (Center) Gape distance. (Bottom) Horizontal displacement of the hyobranchium and the food item relative to a fixed mandible. The food item is initially dropped, recaught, and then transported lingually.
IiIIIIr
377
11. F e e d i n g in P a l e o g n a t h o u s Birds
it mixes with saliva and is repeatedly rolled against the palatal papillae and compressed between the tongue and the palate. The transport stage ends when food reaches the pharynx, but as subsequent food items are picked up and transported, items transported earlier collect in a postoropharyngeal expansion of the esophagus {- rostral esophagus) before being swallowed. A large food item (clumped pellets) can be transported by tossing it caudally with the retracting tongue during the initial step; kinetic energy is imparted to the food by rapid lifting of the head. Fewer steps are then required for the food to reach the pharynx, but direct, inertial transport from the beak tips into the rostral esophagus does not occur (Fig. 11.16). Following numerous transport phases (up to 29 have been observed) and the collection of multiple food items in the rostral esophagus, the head is lifted, thereby straightening the neck. Several hyolingual cycles then occur, apparently while contracting the constrictor muscles, with the jaws held essentially closed and tipped upward. This final stage constitutes swallowing. During swallowing, hyolingual retraction occurs with the tongue pressed against the palate so that the last items ingested are pushed completely into the rostral esophagus. Now engorged, the rostral esophagus constricts and peristalsis moves the bolus into the gut.
Food leaves tongue Food released by beak tips
|llllllllll|llllll|||||||lllllll l|llllll llll||| Mill
0.5-i
.-
E
1.0-i
«
1.5H
0
2.OH
B w
2.5H
o
o
'.
Upper beak tips
Lower beak tips
T3 CO
o
•^ 3.5 H > 4.0-
CO
CO
n
E o
CD O
c iS T3 C
o
N O X
iiiiiniiniinii|niiiiiiii|niii
|iiiiiii^
Beak tips
0 -
. • ' • "
• • " •
1
O"^—
1 -
Food leaves tongue
2 -
Protraction
1
3 4 -
\
5 -
-
/••
\
TTI
0
"•
/
\
6 -
f^ood "••"""•.
"'' /
1 1 1 1 1• ii | i i i i i i i i i i | i i i i i i i i i i i i i i i r T i r
10
1
Hyoid
20 30 1 franne = 5 msec
Pharynx ri'i 1111111111
40
F I G U R E 11.16. Kinematic profile of transport of a large food item in the neognathous wild turkey, Meleagris gallopavo, showing combined cranioinertial and lingual transport. Coordination of jaw and hyoid cycles is not substantially altered from transport of smaller items by the tongue alone (see Fig. 11.15).
I. J aw-Hyolingual
Coordination
During intraoral transport of small food items, each hyolingual cycle coincides with one mandibular cycle in the following way: hyolingual retraction occurs during the gape cycle (open-close) and protraction occurs with the jaws held closed (Fig. 11.15). Transport of a large item differs in the rapidity of hyolingual retraction; with large food, the tongue is fully retracted at the end of mandibular depression and remains retracted until the jaws close (Fig. 11.16). Jaw and hyolingual coordination is consistent with a "basic tetrapod pattern" (Bramble and Wake, 1985) that is reported in other tetrapods (crocodiles: Busbey, 1989; Cleuren and De Vree, 1992; lizards: Smith, 1984, 1986, 1992; Schwenk and Throckmorton, 1989; mammals: Crompton, 1989; Hiiemae and Crompton, 1985; birds: Van den Heuvel, 1992; Zweers, 1982a,b; 1991a,b; Zweers et al., 1994) and is believed to be controlled by "central pattern generators" in the brain stem (see Chapters 2 and 13; Bramble, 1980; Deich and Balsam, 1994; Dellow and Lund, 1971; De Vree and Cans, 1989; Dubbeldam, 1984; Gorniak et al, 1982; Hiiemae et al, 1978; Homberger, 1988; Jean, 1984; Smith, 1992, 1994; Zeigler et al, 1975,1980,1994).
378
Carole A. Bonga Tomlinson
2. The Hyobranchial
Mechanism
Extrinsic hyolingual protractor muscles act on the epibranchials (branchiomandibularis, rostralis and caudalis) and root of the tongue (genioglossus); retractor muscles act indirectly on the hyoid horns and larynx (serpihyoideus, constrictor colli intermandibularis, constrictor colli cervicalis) and directly on the basihyal (stylohyoideus). Movably anchored by the caudal suspensoria ("primary fulcra")/ the hyoid horns continuously depress and elevate relative to the mandible. Hyoid horn depression and elevation account for most lingual retraction and protraction, respectively. The basihyal follows an elongate, somewhat elliptical orbit moving anterodorsally and posteroventrally relative to the mandible (Fig. 11.14A). This orbit also reflects epibranchial bending (extended during hyoid horn elevation, flexed during depression). The system just described results in continuous, rapid retraction and protraction of the neognathous hyoid. However, the stylohyoideus muscles in neognathous birds insert onto the basihyal and affect lingual function. They appear to affect lingual flexion at the same time that they contribute to hyobranchial retraction. The basihyal serves as a fulcrum ("secondary fulcrum") for paraglossal flexion. Retractive forces ex-
erted by the stylohyoideus muscles pull the basihyal posteriorly from the protracted position as the ceratobranchials rotate posterolaterally around the primary fulcra, thus rotating the basihyal dorsally at its articulations with the ceratobranchials while it is retracted. The paraglossal flexes relative to the basihyal at the secondary fulcrum while it (and the tongue) is lifted (Fig. 11.17). The retracting tongue either lifts from the floor of the mouth or its midsection bulges upward, presumably depending on the activity of intrinsic hyolingual muscles. Hyolingual retraction in this way lifts or pushes food posteriorly toward the pharynx during lingual feeding. Although neognathous hyolingual morphology promotes rapid lingual transport, it would seem to limit effective swallowing. Food cannot be actively pushed from the oropharyngeal cavity into the esophagus unless the tongue retracts against the palate. However, when the hyobranchium is retracted during intraoral transport cycles, the basihyal and tongue are forced away from the palate due to the limited degree of movement allowed at the primary fulcra. The ceratobranchials are forced to rotate at these points so that the basihyal (and tongue) swings posteroventrally, away from the palate. Jaw-hyolingual coordination exacerbates the problem—mandibular depression
Primary fulcrum Occipital region
Fascial attachment
Comua
M. stylohyoideus (extrinsic retractor) Paraglossal (and tongue) lifted Mandible^/^
Mandible depressed
Basihyal dorsiflexed
Hyoid retracted
FIGURE 11.17. The hyobranchial mechanism during lingual feeding in neognathous wild turkey, Meleagris gallopavo. See text for discussion.
379
11. Feeding in Paleognathous Birds during hyobranchial retraction also pulls the tongue away from the palate. During swallowing, this pattern is circumvented by postural changes of the head and neck, limited depression of the mandible, and contraction of the constrictor muscles in order to overcome the effects of the primary fulcra so that the tongue is held against the palate during retraction. B. Cranioinertial Feeding in Paleognaths The same kinematic parameters shown in Figs. 11.15 and 11.16 for the turkey during lingual feeding are illustrated for the rhea (Fig. 11.18) and emu (Fig. 11.19) during cranioinertial feeding. Ostrich data were hand generated and are not illustrated. Cranioinertial feeding is virtually identical in ratites, as depicted for the rhea in Fig. 11.20. Food is released from the beak tips as the head is raised by a sinuous movement of the long, S-shaped neck. Cranial motion imparts kinetic energy to the food item so that it travels inertially into the oral cavity following release. Numerous small food items are sometimes transported at once, moving directly into the rostral esophagus solely by inertia, without hyolingual involvement. However, large food items are ingested individually. Any food items remaining in the oral cavity or oropharynx after an inertial transport cycle (small food sometimes hit the pharyngeal roof) are then pushed into the rostral esophagus with the tongue. In all cases, the protracted tongue is located anterior to the food item following cranioinertial transport and the tongue retracts following jaw closure. When the tongue retracts, constriction of the throat indicates that the constrictor muscles are active. Therefore, each retraction of the tongue appears to serve as a swallowing cycle. Intraoral transport and swallowing are thus combined within one hyolingual cycle. 1. Jaw-Hyolingual
Coordination
Coordination of the jaws and hyolingual apparatus in ratites contrasts profoundly from the pattern observed in primarily lingual-feeding (neognathous) birds and is attributable in part to morphological differences that result in a distinct hyobranchial mechanism. During intraoral transport of small or large food items, the hyolingual cycle coincides with the mandibular cycle in the following way: hyolingual protraction occurs during one gape cycle (open-close) and hyolingual retraction occurs with the jaws closed. During retraction, a second, small gape sometimes occurs in rhea and ostrich, but not in the emu (Figs. 11.18 and 11.19). Unlike the neognath pattern, the ratite pattern is in
contrast to the "basic tetrapod pattern" of jaw and hyolingual coordination. The extreme rapidity and stereotypy of synchronous movements in avian feeding suggest that a "neurological shift" in the brainstem (Zweers, 1985, 1991a,b; Zweers et al, 1994) may have occurred during the evolution of cranioinertial feeding in paleognathous birds. 2. Hyobranchial
Mechanism
The ratite pattern of jaw and hyolingual coordination requires that the tongue be short in order to avoid injury—a protracted tongue of greater length would be caught between the rapidly closing beak tips. Protraction is apparently affected by the branchiomandibularis muscles acting on the epibranchials, genioceratohyoideus muscles acting on the ceratobranchials, and genioglossus muscles acting on the paraglossal (with a distinct role in the ostrich, see later). The basihyal orbit is roughly triangular, and "early protraction" occurs with the mandible elevated, whereas "late protraction" occurs during the jaw opening (Figs. 11.14B-11.14D). During late protraction, the genioceratohyoideus muscles apparently exert a strong pull on the ceratobranchials toward the mandibular symphysis while the mandible is depressed and elevated (note that a dermal muscle, the cleidohyoideus, clearly plays a role in directing the basihyal ventrally during protraction in ratites, but it is here considered a minor participant and is not discussed further). Hyolingual retraction is by means of different sets of muscles in each ratite species. Retractors act either on the hyoid horns alone or on the hyoid horns and larynx. In the rhea (and tinamou), retractors act indirectly on the hyoid horns (serpihyoideus, constrictor colli intermandibularis, constrictor colli cervicalis). In the emu, retractors act directly on the ceratobranchials (hyomandibularis lateralis) and urohyal (hyomandibularis medialis). In the ostrich, retractors act directly on the ceratobranchials (hyomandibularis) and larynx (serpihyoideus; although insertion on the larynx is unique to the ostrich, this muscle name is tentatively retained based on presumed functional similarity in retracting the hyobranchium in all birds). Regardless of variation in retractor muscles and their insertion sites, the pattern of hyolingual retraction is similar in all three ratite species—retraction occurs with the mandible essentially elevated and the retractive path of the basihyal follows a dorsally convex arc (Figs. 11.14B-11.14D). The retractive path of the basihyal is due in part to the morphology of the primary fulcra. The short (rhea, emu) or downward-curving (ostrich) epibranchials are loosely anchored to the ear region, unlike
380
Carole A. Bonga Tomlinson
initial head rise
Food c
3 O
^_
O)
E o
Food released by beak tips
»•—
o c
Qj^iiiiiiiii|iiiiiiiiM|iiiiiiiiii|iiimiiii|iiiiiiiiii|ii
CO
^-^ (O
T5
Upper beak tip
1
To O
E
3 *C (0
0)
>
o o c
2
• • - "..•••
3
(0
4
15 T3 O
5
*-• (/>
.-.
•
Lower beak tip
• ••
1
> 6
11 • 1 1 1 1 1 1
• •
II Ml l i l l l
1111111111111 MII1111111111111111111111
Beak tips 0)
.9-
^CO (D
^
E o
^
•4—
CD O
c
0) "D "cO
*-* o
N c o X k»
J
2
'•.. Food
4 Protraction
6
T
Retraction
8 10 12 lllllllllllllllll
0
10
I i 111111111111111111111111111111111111111111111111111111111
0
10
20 30 1 frame = 5 msec
40
50
FIGURE 11.18. Kinematic profile of cranioinertial transport in a paleognathous ratite, Rhea americana. (Top) Vertical displacement of the head and food item. (Center) Vertical displacement of upper and lower beak tips relative to a fixed point on the cranium showing gape distance. (Bottom) Horizontal displacement of hyobranchium and food item relative to a fixed mandible. The initial stages of the transport cycle shown on the left were obtained from a separate film sequence.
11. Feeding in Paleognathous Birds 16-
i. i
the basihyal in a dorsally convex arc that enables the tongue to scrape the palate (see Figs. 11.20C and 11.20D) to affect swallowing during each hyolingual cycle. Tongue movement in ratites is variable. Extension of
Head y / ' ^ N .
1412-
/
10-
Food ^""^N^
X
..••'
381
"••..^^
864 -
Reha
Food released by beak tips
2-
..' 11111111II1 1 i 111111II111II If 11 Ml 11111 I I I i l l 1111 • 1111II1IIIIII11II111
20
10
I
1
30
40
50
"•••^
7-
•..
Upper beak •
m
^m'
1
•::::i
..-.•:: m
'•
8-
60
u
•
-••
•
Lower beak
9 •
10•••••
I I I l l | M i l l 1111i111111••11111111H11 M111111111111i 111 i1111111111111
-10
0
10 20 30 Frame (1 fr = 5 msec)
Beak tips <
40
"•"•©. Food
Release
I 3i •i
5-
f
6H
o
I
Retraction
Hyoid
Esophagus
iiiii|iiiiiiiiii|iiiiniiiniiiiiiiiii[iiiiiiiiii|iiini
1 frame = 5 msec
FIGURE 11.19. Kinematic profile of cranioinertial transport in the paleognathous emu, Dromaius novaehollandiae. See legend for Fig. 11.18.
the condition in neognaths. This loose suspension permits the hyolingual apparatus to be retracted while the hyoid horns remain elevated and, significantly, the mandible is essentially elevated (Fig. 11.21). Mandibular elevation and the presumed contraction of constrictor muscles both contribute to lift the tongue against the palate during each hyolingual retraction. In combination, the loose nature of the primary fulcra, elevation of the hyoid horns and mandible, and constriction of the throat region result in retraction of
350 msec
FIGURE 11.20. Overview of cranioinertial transport in Rhea based on tracings made of selected individual frames of cineradiographic film. (A) Ingestion of small food items from ground level precedes a cranioinertial toss. (B) Tongue protraction accompanies maximum gape following release of food items. (C) Tongue protraction continues as mandible is elevated and food travels inertially into the rostral esophagus. (D) Hyolingual retraction and pharyngeal emptying (swallowing). Head rise (shown on left) usually occurs only when large food items are transported. Arrows depict pathways of movement of various elements. Position of radio-opaque markers are indicated by (+) for tongue, ( • ) for mandible, and ( • ) for cranium and upper beak. Positions of the left ceratobranchial, its articulation with the basihyal, and the inferred position of the cartilaginous paraglossal within the tongue are shown in each step. Proximal (posterior) ends of cartilaginous epibranchials were not visible.
382
Carole A. Bonga Tomlinson Primary fulcra (ear region)
Mandibular depression F I G U R E 11.21. for discussion.
Early hyoid protraction
Hyobranchial mechanism during cranioinertial feeding in paleognathous ratites. See text
the tongue anterior to the basihyal enhances ventrojElexion of the tongue, but the presence of a urohyal limits it, thus the rhea tongue is most flexible and the ostrich tongue the least flexible among the ratites. In the rhea and emu, the role of the basihyal as a fulcrum for paraglossal flexion is limited, and the paraglossal and tongue can only be ventroflexed by means of the ceratoglossus and the sling-like hyoglossus muscles. In these taxa, lingual ventroflexion occurs during hyolingual retraction. In the ostrich, tongue movements are extremely limited, yet muscles acting on the paraglossalia appear to close the lingual pocket during late protraction (genioglossus) and maintain closure during subsequent retraction (ceratoglossus). It is possible that side-to-side movements of the tongue (not observed in lateral projection of the cineradiographs) are also due to the actions of these muscles acting bilaterally (Bock and Btihler, 1988). C. Comparison of Ratite Cranioinertial and Neognathous Lingual Feeding Jaw and basihyal (hyobranchial) cycles in paleognathous and neognathous taxa are compared in Fig. 11.22. Coordination of jaw and hyolingual cycles char-
acteristic of paleognathous ratites, including a oneto-one correspondence between retraction cycles and swallowing, is possible partly because of hyobranchial mechanics permitted by the distinct morphology of the caudal suspensoria (primary fulcra) (Figs. 11.13 and 11.21). In contrast, the swinging motion of the hyoid horns in neognathous species (Figs. 11.13 and 11.17) precludes swallowing during each hyolingual cycle, but enhances rapid lingual transport of numerous food items prior to a distinct swallowing stage. Although coordination of jaws and tongue in the neognath apparently follows a previously proposed basic tetrapod pattern, this pattern applies to coordination of the mandible and tongue. However, care must be taken in comparing these patterns among tetrapods because not all studies employ the same method of measurement. For example, in some reptile studies tongue and hyobranchial movements are "averaged" (e.g., Cleuren and De Vree, 1992), whereas in mammals, hyobranchial movement is usually measured relative to the palate (see Thexton et ah, 1998). Because hyolingual protractors originate on the mandible in all tetrapods, the basihyal orbit determined relative to the mandible represents a precise measure of hyolingual function that could be applied in all
11. F e e d i n g in P a l e o g n a t h o u s Birds
Ratite crainioinertial feeding-Rhea Hyoid
s I
Beaks dosed
910-1
11A
5
|niiiiiiii[iiii'fHiii|Hniiini[iniiiinnirninnniniiui
Beaks open
Ratite crainioinertial feeding-Emu Hyoid .^..... ^.?P® .Beaks ctosed
4.0-1
h-0.5 H-LO
4.5H
h-1.5 I- -2.0
I*
5.5-
h-2.5
6.OH
h--3.0 |inn III inmmni{ni mini mil 11111(1 Hill
h-3.5 Beaks qpen
Avian lingual feeding-large food item Beaks dosed
H i n m i i i j i 111111111} m i n i Hi 11111111111(1 Beaks open
Avian lingual transport-small food item
|iiiiiinMiiTiiiiiii|iiiiHiiiiiiiiiiiiii|iiiHiiiiiinimmfmmiinmnHiii{nmmimiiii 80 40 60 20 Frame (1 fr = 5 msec) F I G U R E 11.22. Coordination of jaw and hyoid cycles during feeding in birds. (Top) Cranioinertial intraoral transport in the rhea. (Second from top) Cranioinertial intraoral transport in the emu. (Third from top) Combined inertial and lingual transport of a large food item in a neognathous bird, Meleagris gallopavo, possessing a "normal-length tongue.'' (Bottom) Lingual transport of small food in Meleagris.
383
384
Carole A. Bonga Tomlinson Gape
'|iiiiiiiiniiiiiiiiii|niiiriiiiiiniininiimiiiiiiimiiiiniiiii
ii|iiiiiiiiirtiiiiiiiii[iiiiiiiiiiiiiiiiiiii)iiiiiiiiiiiiiiiiiiii|iinniiiiiiiiiiiiii|iiiiiiiiiiiiiiiiiiinnmiiimiiiiiim|iimiiiiii
0
80
20
40
60
100
120
140
160
180
200
Time (flim frame) FIGURE 11.23. Kinematic profile of hyoid movement during intraoral transport in a generalized, lepidosaurian reptile, the tuatara, Sphenodon punctatus. (Top) Coordination of jaw and hyobranchial (basihyal) cycles (compare to Fig. 11.22). (Bottom) Anteroposterior (protraction-retraction) and dorsoventral movements of the basihyal relative to the mandible. Overall the basihyal follows a figure-eight orbit (not illustrated), which seems to reflect the motions of the muscular, manipulative tongue, as compared to the simpler, ovoid orbits ofbirds (Fig. 11.14).
taxa to elucidate underlying differences in pattern generation. To help polarize functional patterns in the hyobranchial apparatus, kinematic data for birds were compared to similar data obtained from cineradiographic films of feeding in a generalized lepidosaurian reptile, the tuatara (Sphenodon). The basihyal orbit in Sphenodon while chewing is more complex than in birds. Sphenodon has a large, muscular tongue (Schwenk, 1986) and exhibits the basic tetrapod pattern of mandible and tongue coordination. The basihyal follows a modified figure-eight pattern, probably reflecting a more complex lingual manipulation of food as compared to birds. Nevertheless, basihyal protraction occurs during jaw opening and closing, as it does in ratite cranioinertial feeding (Fig. 11.23). In both neognathous and paleognathous birds, the basihyal orbit has clearly been simplified (i.e., retractive and protractive paths do not cross) relative to a reptilian pattern and is very likely due to the origin on the mandible of both the major protractors and major retractors. Nevertheless, the coordination of hyolingual retraction or protraction with jaw opening clearly has been modified in both the ratites and the wild turkey and the underlying control mechanism remains unknown. The simple basihyal orbit in birds results in
rapid intraoral transport, presumably by permitting rapid coordination of tongue and jaws, yet basihyal movement is controlled by distinct means in paleognathous and neognathous birds.
V. EVOLUTION OF THE FEEDING SYSTEM A. Avian Phylogeny and Outgroup Choice During much of the past century, paleognathous birds were excluded from discussions on the evolutionary origins of birds because the primitive-appearing attributes of the group were assumed to result from reversals brought about through the processes of paedomorphosis, pachyostosis, and/or loss of flight (Bledsoe, 1988; de Beer, 1956; Feduccia, 1996; Martin, 1983; Olson, 1985). Elzanowski (1986) refuted paedomorphosis in paleognaths and proposed separate evolutionary processes for many paleognathous attributes. Nevertheless, for many years the Paleognathae were commonly regarded as an early, monophyletic group descended from some unknown group of neognathous birds (Bock, 1963; Cracraft, 1973, 1974, 1986, 1988) rather than a truly primitive neornithine group, as had been suggested earlier (Pycraft, 1900; Lowe, 1928). The
11. Feeding in Paleognathous Birds ancestral type of the Paleognathae was believed to be a "proto-tinamou" (Parkes and Clark, 1966). Fossil discoveries of extinct, volant paleognaths in North America (Paleocene) and Europe (Eocene) (Houde, 1986,1988; Houde and Haubold, 1987; Houde and Olson, 1981) and a flightless ostrich precursor from the Eocene of Europe {Palaeotis; Peters, 1988) demonstrated that paleognathous birds were once more widespread than today and that dispersal routes included Laurasia (Houde, 1988; Olson, 1985, 1986). Although Houde (1988) suggested that tinamous are related more closely to neognathous birds than to ratites (see also Sibley and Ahlquist, 1990), the view that ratites descended from a tinamou-like precursor has been generally accepted, and many molecular studies interpret results on this basis. Results of molecular studies examining the relationships of neognathous birds to either tinamous or ratites, or relationships among ratites, are contradictory (e.g.. Cooper et ah, 1992; Harlid et al, 1997; Hedges, 1994; Hedges et al, 1996; Prager et al, 1976; Sibley and Ahlquist, 1990; van Tuinen et al, 1998). Molecular studies may lack the ability to resolve relationships above the family level and the great length of time since the divergence between paleognathous and neognath taxa, and divergences among the ratite lineages, may be responsible for the confusing results (Cracraft and Mindell, 1989; Nei, 1996; Rambaut and Bromham, 1998). Moreover, few molecular studies (Hedges et ah, 1995; Stapel et al, 1984, updated by Caspers et ah, 1994) compare all three groups (ratites, tinamous, neognathous taxa) with reptilian outgroups. Although Hedges et al. (1995) and Stapel et al (1984) did not include both paleognathous groups (ratites and tinamous) in a single analysis, each study nevertheless placed paleognathous taxa between reptiles and neognathous taxa. The revised view of paleognath relationships coincides with a growing number of analyses of fossil Mesozoic birds (many of them toothed) identifying cranial structures present in paleognaths but absent in neognaths (e.g., Elzanowski, 1991, 1995; Elzanowski and Gallon, 1991; Elzanowski and Wellnhofer, 1992, 1993, 1996; Kurochkin, 1995; Zhou, 1995). Studies on palate morphology (Peters, 1987; Weber, 1992; Witmer and Martin, 1987), egg shell histology (Houde, 1988), amino acid sequences of eye-lens protein (Stapel et al, 1984; Caspers et al, 1994), embryological development (neognathous pterygoid, Jollie, 1957; tarsus, McGowan, 1984,1985; but see Martin and Stewart, 1987), gross karyotype (Takagi et al, 1972, 1974), and immunological distance based on DNA hybridization (Prager and Wilson, 1976; reinterpreted in Lee et al, 1997) all indicate that paleognathous birds are less derived relative to either living reptiles or dinosaurs (depend-
385
ing on outgroup used) than are neognathous birds. On this basis, the most suitable outgroups for determining primitive avian characteristics of the hyolingual apparatus are toothed Mesozoic birds, theropod dinosaurs and/or earlier thecodont reptiles, living archosaurs, and living lepidosaurs (see Chiappe, 1995; Feduccia, 1994,1996; Hecht, 1985; Martin, 1983,1985,1987; Molnar, 1985; Ostrom, 1969, 1973, 1985, 1991; Welman, 1995). B. Primitive Condition of the Neornithine Hyolingual Apparatus Bock and Biihler (1988) regarded a "normal-length" tongue (i.e., filling the available space within the oral cavity) and the more extensively ossified hyobranchium of neognathous birds to represent the primitive neornithine condition. Nevertheless, direct evidence for the condition of the primitive avian or neornithine hyobranchium is lacking as hyobranchia are rarely reported, preserved, and/or detected in fossils. However, polarities of hyobranchial skeletal and muscular characters can be evaluated by comparison with extant reptiles and a few, recently reported fossil hyobranchia associated with theropod dinosaurs and toothed Mesozoic birds. 1.
Ossification
In modern reptiles, including crocodilians and lepidosaurs, the only ossified hyobranchial elements are the ceratobranchials (e.g., Oelrich, 1956; Schumacher, 1973; Smith, 1984,1986; Sondhi, 1958). Fossilized hyobranchia in the possible ancestors of modern birds (toothed Mesozoic birds, theropod dinosaurs, and thecodont reptiles) indicate that only the ceratobranchials were ossified in these animals as well (Chiappe et al, 1998; Dal Sasso and Signore, 1998; Elzanowski and Wellnhofer, 1996; Weishampel et al, 1990). There is no evidence, therefore, that the ancestral avian basihyal was ossified. The first appearance of the neomorphic avian paraglossal (Crompton, 1953) is unknown and, if cartilaginous, would not be preserved in fossils. Thus, a cartilaginous basihyal is likely to have been present and a cartilaginous paraglossal cannot be ruled out in ancestral neornithine birds. Thus, the paleognathous hyobranchium appears to retain the ancestral condition of these characters, whereas an ossified basihyal and the capacity for paraglossal ossification seem to be derived traits of Neognathae. Extensive hyobranchial ossification and diversification in the form of the hyolingual apparatus in neognathous birds may have arisen from a basic framework retained only in modern paleognathous birds.
386
Carole A. Bonga Tomlinson
2. Position and
Orientation
The hyobranchium in most reptiles is located in the neck region, but in birds it is located directly beneath the raandible. Orientation of the hyoid horns (comprising first ceratobranchials and epibranchials in reptiles, and the cerato-epibranchials in birds) is toward the sites of origin for the retractor musculature—in modern birds the ceratobranchials are aligned with the mandibular rami in the resting position, whereas the hyoid horns in all modern reptiles are directed toward the sterno-pectoral region to which the retractor muscles attach. In fossil theropods and toothed birds, only ceratobranchials are preserved, but in most specimens these are displaced and their natural position is unclear. Two exceptionally well-preserved Cretaceous fossils have been described in which some of the animals' soft tissues were revealed and the ceratobranchials positioned apparently as in life. The ceratobranchials of one of these fossils, a small bird-like theropod (Dal Sasso and Signore, 1998), are located posteriorly in the neck region and project posteroventrally toward the pectoral region. Ceratobranchials of the second fossil, a mononykine (toothed) bird (Chiappe et al, 1998), lie directly beneath the mandible and are oriented along its axis. Although identification of the mononykines as avian is disputed (Feduccia, 1994; Zhou, 1995), the ceratobranchials are clearly bird-like. These fossils indicate that transformation in the hyobranchial position did not occur in ancestral theropods, but later during the evolution of early birds. Given that the ancestral condition was orientation of the cerato-epibranchials toward the sternum (as in living reptiles and theropod dinosaurs), then their upward curvature and occipital attachment in neognaths is obviously derived. The paleognathous condition, in which the epibranchials are short or downcurved and attachment is to the ear, is then reasonably interpreted as an intermediate condition. The paleognathous condition therefore appears to be more primitive than that of neognaths. The slight curvature of the epibranchials in a tinamou may thus be interpreted as an incipient neognathous condition, which is most parsimoniously considered a synapomorphy of tinamous and neognaths. The condition of the tongue during these hyobranchial transformations remains wholly unknown. 3.
Musculature
Intrinsic lingual muscles in reptiles (Smith, 1984, 1986, 1992; Schwenk, 1986, 1988) and mammals (Crompton, 1989; Hiiemae and Crompton, 1985; Hiiemae ei al, 1978; Thexton et al, 1998; Schwenk, 2000) comprise longitudinal and transverse fiber systems that occur within the fleshy body of the tongue and
do not connect directly to the hyobranchial skeleton. Such muscles are absent in all modern birds. Although Homberger (1986) and Homberger and Meyers (1989) refer to muscles that occur within the tongue of birds as "intrinsic lingual muscles,'' these muscles nevertheless attach to hyobranchial elements and are therefore referred to here as "intrinsic hyolingual muscles" (as suggested by Baumel et al, 1993). In birds these muscles arise from hyobranchial elements and only some, such as the hyoglossus anterior in neognathous species, occur exclusively within the body of the tongue. Other intrinsic hyolingual muscles, such as the hyoglossus in paleognaths and the hyoglossus obliquus in neognaths, occur partially within the tongue and partially within the lingual base. Another, the ceratoglossus, inserts on the paraglossal but originates on the ceratobranchial lateral to the larynx. The avian tongue is thus defined less by its muscular content than by its skeletal and epithelial components. The paraglossal occupies much of the tongue body and, except in parrots (Homberger, 1989), all muscles attached to it are located ventral to it. The dorsal surface of the tongue is therefore usually rigid and covered with papillae to provide a "frictional surface" for holding and moving food items. At what point in history the unique form of the avian tongue evolved is unknown. Its extreme reduction is derived relative to other tetrapods, as the ances-' tral amniote condition was almost certainly a mobile, muscular tongue (see Chapter 8). Reduction occurred when reptile-like, intrinsic lingual muscles were replaced by the neomorphic paraglossal and avian hyolingual muscles. There is no evidence for a paraglossal in taxa earlier than the neornithine common ancestor of paleognathous and neognathous birds, thus direct evidence for the timing of this transformation is lacking. However, it is probable that reduction and modification were associated with other changes in the avian (or proto-avian) feeding apparatus, such as the loss of teeth and the repositioning of the hyobranchial apparatus. The simple (although variable) arrangement of musculature within the tongues of paleognathous birds may represent the primitive condition. This hypothesis is supported circumstantially by the fact that the extrinsic hyolingual musculature appears to be more primitive in paleognaths than in neognaths. The floor of the oral cavity in modern reptiles is filled by complex protractor muscles that connect the mandible to the ceratohyals, ceratobranchials, or base of the tongue (Oelrich, 1956; Schumacher, 1973; Schwenk, 1986, 1988; Smith, 1984, 1986, 1988; Sondhi, 1958). The protractor musculature of all ratites exhibits a consistent pattern, and the genioceratohyoideus appears to represent retention with a slight modification
11. Feeding in Paleognathous Birds of reptilian protractor musculature (mandibulohyoideus or geniohyoideus) that originates near the mandibular symphysis and inserts on the ceratobranchials. Virtually all such muscles are absent in neognathous birds. Neognathous protractors (branchiomandibularis muscles) appear to represent a more derived avian condition than present in paleognaths. Extrinsic hyolingual retractor muscles in modern birds and reptiles reflect the transition that apparently occurred in the evolution of the avian hyobranchium from a reptile-like hyobranchium. The major hyobranchial retractors in modern lizards (in addition to the constrictor colli and constrictor colli intermandibularis, which are present in all reptiles and birds) are the omohyoideus and the sternohyoideus, which originate in the sterno-pectoral region (Smith, 1986). In birds, the major hyobranchial retractor muscles originate on the laterocaudal surface of the mandible. [The avian cleidohyoideus is a dermal muscle and is unlike any reptilian retractor. However, in the paleognathous kiwi {Apteryx), a sternohyoideus is said to be present as it is in reptiles (attributed to Gadow and Selenka, 1891; in Baumel et al, 1993)]. Muscles originating on the mandible could only act as hyobranchial retractors following the evolutionary migration of the hyobranchial apparatus anteriorly to its avian position within the mandibular rami. Thus modification of reptile-like retractor musculature into its avian form must have occurred concomitant with this evolutionary transformation. The serpihyoideus is present in all birds and was probably the earliest avian hyobranchial retractor to originate on the mandible. In the paleognathous rhea and tinamou, it is the only hyobranchial retractor, apart from the constrictors. In this respect, the rhea and tinamou appear to possess the most primitive hyolingual apparatus among modern birds. Although the ostrich and emu have additional retractors (hyomandibularis muscles), their differences do not appear to affect hyobranchial function during intraoral transport and probably arose subsequent to divergence among different continental assemblages of paleognathous taxa. In contrast, the presence of the retractor stylohyoideus in neognathous birds is clearly a derived avian character, as indicated by its absence in reptiles, and as is true for the additional retractors in the ostrich and emu, it may have evolved from the primitive avian condition putatively retained in the paleognathous rhea and tinamou. 4. Tongue Size The paleognathous tongue is short not only relative to the ancestral amniote condition, but relative to neognaths as well. Rapid and efficient cranioinertial feeding in ratites would seem to require that the tongue
387
be reduced in length (see earlier discussion) and it is possible that this represents the primitive condition for neornithine birds. If early avian hyobranchia were mostly cartilaginous, as suggested by fossil evidence, it is possible that extensive lingual reduction occurred prior to, and not following, the divergence of paleognathous and neognathous lineages. Forfeiture of biting and intraoral repositioning of food with a mobile, muscular tongue clearly occurred early in avian history and conceivably was driven by strong environmental selection pressures to feed rapidly. As such, efficient cranioinertial feeding using a small tongue may have preceded the neognathous form of lingual feeding (see later). The neognathous feeding system could have been derived from the basic, putatively ancestral form of the hyolingual apparatus present in the paleognathous rhea or tinamou. This scenario is supported by the fact that lingual feeding is dependent on structures that are derived relative to the paleognathous condition (see later). C. Changes in Feeding Function During the Theropod-Bird Transition Location of the hyobranchium in the neck of a fossil theropod suggests that intraoral transport in theropods was unlike that in modern birds and more similar to living reptiles, such as lepidosaurs in which intraoral transport and swallowing take place gradually by means of repetitive hyolingual and jaw cycles (Busbey, 1989; Cleuren and De Vree, 1992; Smith, 1984, 1986; Throckmorton, 1976,1980; see Chapters 2 and 8). Presence of teeth in theropods likewise indicates retention of a reptile-like feeding pattern. Nevertheless, dinosaurs were bipedal, like birds, and presumably their cranioinertial capabilities were better than those of quadrupedal reptiles (Smith, 1986; Cans, 1969). Efficient cranioinertial feeding is likely to have originated in modern birds, however. All modern birds possess longer necks (12-24 cervical vertebrae) than theropod dinosaurs and Mesozoic toothed birds (10 cervical vertebrae), reflecting an improvement of cranioinertial feeding in neornithines. Paleognaths lie in the middle of the avian range with 13-20 cervical vertebrae (Pycraft, 1900). The loss of intrinsic lingual musculature indicates that intraoral manipulation of food became less important in ancestral birds, a trait that in modern birds is associated with the absence of teeth and chewing behavior. Changes in feeding behavior in toothed birds may have occurred prior to the appearance of the paraglossal, which provides attachment sites for hyolingual muscles that flex the tongue relative to the basihyal and ceratobranchials. Tongue flexion partially compensates for the loss of a muscular, manipulative
388
Carole A. Bonga Tomlinson
tongue. The modern avian tongue may have evolved in neornithine ancestors, which decreasingly relied on killing or grasping bites with toothed jaws. As such, changes in avian feeding mechanics and hyolingual form probably occurred as birds came to rely more on small, relatively inert food items (e.g., omnivory, granivory) and less on large prey items (e.g., carnivory). Formation of a modern avian tongue may thus have occurred in response to the same selection pressures that resulted in complete loss of teeth. Loss of avian teeth presumably occurred no earlier than the neornithine common ancestor of paleognaths and neognaths. In this putative ancestor, cranioinertial feeding was probably well developed. D . Proposed Functional Evolution of Early Avian Transport Mechanisms The caudal suspensoria of avian hyobranchia are a critical mechanical component (primary fulcra) of the feeding system that, to a large extent, dictate the nature of jaw and tongue coordination in birds. Ratite suspensoria permit hyolingual retraction when the jaws are closed and may represent an evolutionary stage that first appeared as the necks of birds elongated and cranioinertial feeding improved. Ratite cranioinertial feeding requires not only this moveable form of hyobranchial suspension, but also a very small (anteriorly truncated) tongue so that that tongue can be fully protracted as the jaws close. A small tongue may have evolved in a paleognathous lineage only, but if selection to improve cranioinertial behavior was strong in early birds, and was responsible for increased neck length, loss of teeth, and loss of intrinsic lingual musculature in the descendants of toothed Mesozoic birds, then significant lingual reduction may have occurred in the common ancestor of paleognaths and neognaths in association with the origin of the paraglossal. If so, ratite cranioinertial feeding and the paleognathous form of the hyolingual apparatus are likely to have preceded the origin of avian lingual feeding, which may require an ossified basihyal. The kinematics of hyolingual protraction and the form of the extrinsic hyobranchial musculature also suggest that ratite cranioinertial feeding preceded avian lingual feeding. In ratite cranioinertial feeding, the principal protractor (genioceratohyoideus) runs from the mandible to the anterior surface of ceratobranchials, conforming to the plesiomorphic tetrapod condition. It protracts the hyobranchium when the mandible is elevated. This action does not occur in avian lingual feeding. In neognaths, hyolingual protraction and retraction during intraoral transport result primarily from a swinging motion of the hyoid
horns, which are elevated and depressed, pivoting at their points of suspension on the occiput. Thus neognaths lack the plesiomorphic form of the protractor musculature. The connective tissue attachments of the upcurved epibranchials to the occiput in neognathous birds are also clearly derived relative to other tetrapods and are associated with the unique mechanism of hyobranchial movement. The unique form of lingual transport in neognathous birds is also associated with the presence of a novel muscle, the stylohyoideus. During lingual feeding, muscular forces applied directly to the basihyal, in conjunction with the nature of the articulation between the basihyal and the paraglossal, cause the tongue to lift and/or flex as it is retracted (and the hyoid horns depressed) to move food posteriorly. In sum, avian lingual feeding relies on structures and hyolingual movements that are unique to neognaths and, as such, the neognathous mechanism of hyolingual transport is probably unrelated (i.e., not directly homologous) to the hyolingual mechanism evident in most modern reptiles. Reptiles rely extensively on intrinsically generated, manipulative movements of a complex muscular tongue completely lacking in birds. The implication is that despite the shared presence of hyolingual food transport in living reptiles and neognathous birds, the avian mechanism is secondarily derived. I suggest that hyolingual feeding in neognathous birds evolved from a cranioinertially feeding ancestor whose feeding mechanism probably resembled that of the living ratite, Rhea. E. Evolutionary Morphology: A n Overview Bock and Biihler (1988) suggested that two paleognathous groups (ostrich vs all other paleognaths) independently specialized on large food items and thus the functional role of the tongue during intraoral transport was lost convergently. McLelland (1979), however, maintained that the paleognathous hyolingual apparatus is specialized for swallowing, not for transporting large food items. Although past definitions of swallowing in birds require clarification (Tomlinson, manuscript in preparation), functional data support McLelland's assertion and falsify Bock and Biihler's assertion that the tongue in paleognaths does not assist in the transport of large food items. Although lingual manipulation of food does not occur in ratites, the hyolingual apparatus nonetheless plays a critical role during all food transport. Ecological data and dietary data for living paleognaths (discussed earlier) also fail to support the Bock and Biihler (1988) scenario. Obligate cranioinertial feeding is one means of
11. Feeding in Paleognathous Birds rapidly transporting food of any size through the oropharyngeal cavity and into the expanded rostral esophagus. In paleognathous birds, the form of the hyolingual apparatus permits it to function in a distinctive manner in which swallowing appears to occur during each hyolingual cycle. The mechanism is similar in the South American rhea, the Australian emu, and the African ostrich, which suggests its common evolutionary origin for all paleognaths. Morphology of the hyolingual apparatus in a tinamou strongly suggests that its function is similar, but this needs to be tested experimentally (see Chapter 12). Efficient cranioinertial feeding in paleognathous birds requires not only hyolingual structures that appear to be primitive in modern birds (e.g., protractor muscle form and primary fulcra attached to the ear region), but a very short tongue that does not protrude beyond the beak tips when protracted, a paleognathous trait that may be derived. Whether or not a reduced tongue occurred primitively in neornithine birds is unresolved, but evolution of efficient cranioinertial feeding is postulated to have been the context in which the transformation of the hyolingual apparatus, from reptile-like to avian, occurred. During the evolution of Mesozoic birds, omnivorous gleaning of small food items would no longer have required biting jaws, chewing, or lingual manipulation, at which point teeth and a muscular tongue could have been reduced, as in the common ancestor of neo- and paleognaths. Many neognathous birds retain the ability to swallow relatively large prey whole (e.g., owls, roadrunners, piscivores), but the paleognaths are not especially adept at this and their long, slender necks would seem to provide an obstacle to this practice. It remains possible that a small tongue, as seen in paleognaths, was the primitive neornithine condition, providing the means for efficient cranioinertial feeding in response to selection for rapid feeding. Lingual feeding in neognathous birds is another means to transport food rapidly and is accomplished with a hyolingual apparatus that can be derived from the basic structure evident in paleognaths. Hyolingual function in lingual feeding depends on structures (e.g., primary fulcra attached to the occipital region and a novel retractor, the stylohyoideus) that are clearly more derived relative to the reptilian condition than their paleognathous counterparts. An intermediate condition, combining a ''normal-length" tongue with cranioinertial feeding, is functionally improbable. Moreover, morphological trends in Mesozoic birds, including the loss of teeth, origin of the paraglossal and loss of intrinsic tongue musculature, and elongation of the neck, are better explained by the evolution of a cranioinertial feeding system than a lin-
389
gual feeding system. Nonetheless, it is not presently possible to rule out unequivocally the presence of a lingual feeding system in the common neornithine ancestor. If this were the case, then further lingual reduction and obligate cranioinertial feeding may have occurred exclusively in a paleognathous lineage. F. Conservation of Pattern Generation Brain stem control of jaw and hyolingual cycles are thought to be evolutionarily conservative; however, significant alterations would seem to have been necessary in order for either avian feeding system to have evolved from the other. However, the basis of such a "neurological shift" is completely unknown. As argued earlier, the basihyal orbit offers a standardized way to compare kinematic patterns among tetrapods to establish if a "basic tetrapod pattern" truly exists (see Thexton et al, 1998). The absence of a chorda tympani nerve in paleognaths reported by Starck (1995) is not corroborated in another study (Miiller, 1963), but if true it indicates one type of neural modification that has occurred in ratites. The chorda tympani is related to salivary secretions in ducks (Dubbeldam et ah, 1976) and Starck (1995) suggested that its absence in ratites supported the notion that cranioinertial feeding is derived. However, the chorda tympani is a visceral nerve not directly related to movement of the hyolingual apparatus, and a relationship to the short length of the tongue (attributed to Bock and Biihler, 1988; Starck, 1995) is entirely unsubstantiated. In any case, repatterning of neural control of the feeding system is likely to be a function of central changes in brain stem connectivity not necessarily manifested in the periphery. G. Phylogenetic Relationships Ratites appear to be an ancient group in which unique modifications have occurred in each taxon subsequent to an early diversification, as first suggested by Lowe (1928). A similar phenomenon occurs in expressions of paleognathous cranial kinesis: although each mechanism is based on shared structures and clearly differs from the types of kinesis present in neognathous birds (Bock, 1963,1964; Biihler, 1981; McDowell, 1948; Peters, 1987; Pycraft, 1900; Simonetta, 1960; Weber, 1992; Zusi, 1984), kinetic mechanisms differ somewhat in the rhea, emu, and ostrich (Tomlinson, 1997b; Tomlinson, manuscript in preparation). The form of the extrinsic hyolingual retractor musculature in the ostrich and emu (hyomandibularis in the ostrich, hyomandibularis lateralis and medialis in the emu) and in neognathous taxa (stylohyoideus) can
390
Carole A. Bonga Tomlinson
be derived from the condition in the rhea, which is the simplest known in any bird, possibly representing the primitive neornithine condition. Differentiation of the hyobranchial retractors appears to have occurred in opposite evolutionary directions in neognaths and paleognaths, however, and is clearly related to function. This is possible if the common ancestor of all modern birds was paleognathous and the modern "Paleognathae" are a paraphyletic group (see Houde, 1988; Kurochkin, 1995; Lowe, 1928; McGowan, 1984,1985; Pycraft, 1900; Stapel et al, 1984; Takagi et al, 1972,1974; Witmer and Martin, 1987). If modern ratites and neognaths both diverged from a paleognathous common ancestor, subsequent morphological modification in each lineage would not necessarily follow the same pattern, particularly if cranioinertial function was retained in ratites but altered later in neognaths. Whereas functional innovation may require morphological innovation, integration of complex functional systems (e.g., intraoral transport in birds, which required modifications in the tongue, hyoid, hyolingual musculature, jaw suspensoria, palate, pharynx, rostral esophagus, and teeth), once evolved, exerts strong selection to maintain or enhance the function of the system, as a whole, potentially limiting morphological change in individual elements and stabilizing the functional complex (Csanyi, 1989; Vrba, 1989; Wake and Roth, 1989; Wagner and Schwenk, 1999; Schwenk, 2000a). Differences in hyolingual and hyolaryngeal musculature among the three ratite species result in no significant functional differences in intraoral food transport, but may provide clues to the relative timing of divergence from a common stem. Similarities in ostrich and emu retractors suggest that these lineages diverged from a common stem later than the rhea lineage and support the passage of ratites from Gondwana to Australia, or vice versa, via Antarctica. Whereas vicariance of the ratites due to continental drift in the Cretaceous or early Tertiary is supported by the apparently primitive condition of the hyolingual apparatus, the tinamous seem to have diversified more recently (based on immunological distance; Prager and Wilson, 1976). Although the common ancestor of ratites and tinamous was clearly a volant bird, evidence does not support the contention that the tinamou lineage (or "proto-tinamou"; Parkes and Clark, 1966) is ancestral to ratites. It seems likely that the volant, paleognathous Lithornithiformes, known from North America (Paleocene) and Europe (Eocene) (Houde, 1986,1988; Houde and Haubold, 1987; Houde and Olson, 1981), were among paleognathous birds dispersed around the globe in the Mesozoic. Crania and beaks of some lithornithids resemble those of ratites, and lithornithids may have been a sister group to the common ratite an-
cestor. Although Paleotis, a flightless paleognath that occurred in Europe in the early Tertiary (Eocene), is proposed to be an ostrich precursor (Peters, 1988), a linear relationship proceeding from North American lithornithids through Europe to the African ostrich appears unlikely. Similarities in the beak and cranium in tinamous and neognathous galliforms (personal observation) suggest that tinamous are closer to neognathous birds than ratites; however, characteristics of the avian hyolingual apparatus point to derivation of neognathous birds from a paleognathous ancestor. Moreover, the tinamou beak and cranium, not including retained features of the paleognathous palate, more closely resemble galliforms than Lithornithif ormes. I propose that modern birds share a common, paleognathous ancestor with a basal split between ratites and all other birds and a subsequent divergence between tinamous, which retain the ancestral paleognathous condition, and the Neognathae (Fig. 11.1). This phylogenetic hypothesis renders 'Taleognathae" an invalid, paraphyletic taxon and accommodates intermediate features of the tinamou hyolingual apparatus. It is also consistent with plausible and functionally defensible transformations in the avian feeding system. Acknowledgments Citations of manuscripts in preparation refer to papers to be submitted as separate publications. These studies are part of the doctoral research conducted by the author at Harvard University. I express my great appreciation for the guidance, perseverance, patience, and support of A. W. Crompton and K. Schwenk who shared with me their enthusiasm for this topic.
References Alvarenga, H. F. (1983) Uma ave ratitae do Paleoceno Brasileiro: Bacia Calcaria de Itaborai, Estado do Rio de Janeiro, Brasil. Boletim do Museu Nacional 41:1-11. Alvarenga, H. R, and J. R Bonaparte (1988) A new flightless landbird from the Cretaceous of Patagonia. Pp. 51-64. In: Papers in Avian Paleontology, No. 36. K. E. Campbell (ed). Baumel, J. J., A. S. King, J. E. Breazile, H. E. Evans, and J. C. Vanden Berge (eds.) (1993) Handbook of Avian Anatomy: Nomina Anatomica Avium, 2nd Ed. Publications of the Nuttal Ornithological Club, No. 23. Cambridge, MA. Bhattacharyya, B. N. (1980) The morphology of the jaw and tongue musculature of the common pigeon, Columba livia, in relation to its feeding habit. Proc. Zool. Soc. Calcutta 31:95-127 (1978). Bledsoe, A. H. (1988) A phylogenetic analysis of postcranial skeletal characters of the ratite birds. Ann. of Carnegie Mus. 57:73-90. Bock, W. J. (1963) The cranial evidence for ratite affinities. Proc. XIII Intern. Ornithol. Congr.: 39-54. Bock, W. J. (1964) Kinetics of the avian skull. J. Morph. 114:1-42. Bock, W. J., and P. Biihler (1988) The evolution and biogeographical history of the palaeognathous birds. Pp. 31-36. In: Proceedings of the International Centennial Meeting of the Deutsche OrnithologenGesellschaft, Current Topics in Avian Biology, Bonn.
11. F e e d i n g in P a l e o g n a t h o u s Birds Bubien-Waluszewska, A. (1981) The cranial nerves. Pp. 385-438. In: Form and Function in Birds, Vol. 2. A. S. King and J. McLelland (eds.). Academic Press, New York. Bramble, D. M. (1980) Feeding in tortoises and mammals: why so similar? Am. Zool. 20:931 Bramble, D., and D. B. Wake (1985) Feeding mechanisms of lower tetrapods. Pp. 230-261. In: Functional Vertebrate Morphology. Hildebrand, D. M. Bramble, K. F Liem, and D. B. Wake (eds.). Belknap Press, Cambridge. Biihler, P. (1981) Functional anatomy of the avian jaw apparatus. Pp. 439-468. In: Form and Function in Birds, Vol. 2. A. S. King and J. McLelland (eds.). Academic Press, New York. Burton, P. J. K. (1984) Anatomy and evolution of the feeding apparatus in the avian orders Coraciiformes and Piciformes. Bull. Br. Mus. Nat. Hist. (Zool.) 47(6): 331-443. Busbey, A. B. (1989) Form and function of the feeding apparatus of Alligator mississippiensis. J. Morphol. 202:99-127. Caspers, G.-J., J. Wattel, and W. W. de Jong (1994) aA-crystallin sequences group tinamou with ratites. Mol. Biol. Evol. 11:711-713. Chiappe, L. M. (1995) The first 85 million years of avian evolution. Nature 378:349-355. Chiappe, L. M., A. Norell, and J. Clark (1998) The skull of a relative of the stem-group bird Mononykus. Nature 392:275-278. Cleuren, J., and F De Vree (1992) Kinematics of the jaw and hyolingual apparatus during feeding in Caiman crocodilus. J. Morph. 212: 141-154. Cooper, A., and D. Penny (1997) Mass survival of birds across the Cretaceous-Tertiary boundary: molecular evidence. Science 275: 1109-1113. Cooper, A., C. Mourer-Chauvire, G. K. Chambers, A. von Haeseler, A. C. Wilson, and S. Paabo (1992) Independent origins of New Zealand moas and kiwis. Proc. Natl. Acad. Sci. USA 89:87418744. Cracraft, J. (1973) Continental drift, paleoclimatology, and the evolution and biogeography of birds. J. Zool. Lond. 169:455-545. Cracraft, J. (1974) Phytogeny and evolution of the ratite birds. Ibis 116:494-521. Cracraft, J. (1986) The origin and early diversification of birds. Paleobiology 12:383-399. Cracraft, J. (1988) The major clades of birds. Pp. 339-361. In: The Phytogeny and Classification of the Tetrapods, Vol. 1. M. J. Benton (ed.). Clarendon Press, Oxford. Cracraft, J., and D. P. Mindell (1989) The early history of modern birds: a comparison of molecular and morphological evidence. Pp. 389-403. In: The Hierarchy of Life. B. Fernholm, K. Bremer, and Jdrnval (eds.). Elsevier, Amsterdam/New York. Crompton, A. W. (1953) The development of the chondrocranium of Spheniscus demersus with special reference to the columella auris of birds. Acta Zool. 34:70-146. Crompton, A. W. (1989) The evolution of mammalian mastication. Pp. 23-40. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, New York. Csanyi, V. (1989) Origin of complexity and organizational levels during evolution. Pp. 349-360. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, New York. Dal Sasso, C , and Signore (1998) Exceptional soft-tiisu^ preservation in a theropod dinosaur from Italy. Nature 392:383-387. de Beer, G. (1956) The evolution of ratites. Bfull^ Br. Mus. Nat. Hist. 4:59-71. Deich, J. D., and P. D. Balsam (1994) Development of prehensile feeding in ring doves {Streptopelia risoria): learning under organismic and task constraints. Pp. 160-181. In: Perception and Motor Control in Birds. M. N. O. Davies and P. R. Green (eds.). Springer-Verlag, Berlin.
391
Delheusy, V, G. Toubeau, and V. L. Bels (1994) Tongue structure and function in Oplurus cuvieri (Reptilia: Iguanidae). Anat. Rec. 238: 263-276. del Hoyo, J., A. Elliott, and J. Sargatal (eds.) (1992) Handbook of the Birds of the World, Vol. 1. Lynx Edicions, Barcelona. Dellow, P. G., and J. P. Lund (1971) Evidence for central timing of rhythmical mastication. J. Physiol. 215:1-13. De Vree, F, and C. Cans (1989) Functional morphology of the feeding mechanisms in lower tetrapods. Pp. 115-127. In: Trends in Vertebrate Morphology. Fortschritte der Zoologie, Vol. 35. Splechtna and Hilgers (eds.). Gustav Fischer Verlag, Stuttgart. Dubbeldam, J. L. (1984) Brainstem mechanisms for feeding in birds: interaction or plasticity, a functional-anatomical consideration of the pathways. Brain Behav. Evol. 25:85-98. Dubbeldam, J. L., H. Karten, and S. B. J. Menken (1976) Central projection of the chorda tympani nerve in the mallard. Anas platyrhynchos L. J. Comp. Neurol. 170:415-420. Elzanowski, A. (1986) Ontogeny and evolution of the ratites. Pp. 2037-2046. In: Acta XIX Congressus Internationalis Ornithologici. Vol. II. Ottawa, Canada. Elzanowski, A. (1991) New observations on the skull of Hesperornis with reconstructions of the bony palate and otic region. Postilla 207:1-20. Elzanowski, A. (1995) Cretaceous birds and avian phytogeny. Courier Forschungsinstitut Senckenberg, 181:37-53. Elzanowski, A., and P. Galton (1991) Braincase of Enaliornis, an Early Cretaceous bird from England. J. Vert. Paleo. 11:90-107. Elzanowski, A., and P. Wellnhofer (1992) A new link between theropods and birds from the Cretaceous of Mongolia. Nature 359: 821-823. Elzanowski, A., and P. Wellnhofer (1993) Skull of Archaeornithoides from the Upper Cretaceous of Mongolia. Am. J. Sci. 293A-A: 35-252. Elzanowski, A., and P. Wellnhofer (1996) Cranial morphology of Archaeopteryx: evidence from the seventh skeleton. J. Vert. Paleo. 16: 81-94. Feduccia, A. (1994) The great dinosaur debate. Living Bird 13:28-33. Feduccia, A. (1996) The Origin and Evolution of Birds. Yale Univ. Press, New Haven, CT. Fiirbringer, M. (1922) Das Zungenbein der Wirbeltiere insbesondere der Reptilien und Vogel. Abhandlungen der Heidelberger Akademie der Wissenschaften, math.-naturw. Kl. Abt. B 1 1 : 1 164. Cans, C. (1969) Comments on inertial feeding. Copeia 1969:855857. Gorniak, G. C , H. I. Rosenberg, and C. Cans (1982) Mastication in the Tuatara, Sphenodon punctatus (Reptilia: Rhynchocephalia): structure and activity of the motor system. J. Morph. 171:321353. Gussekloo, S. W. S., and G. A. Zweers (1997) Feeding adaptations in the greater rhea (Rhea americana; Ratitae). J. Morph. 1?>1:161. Harlid, A., A. Janke, and U. Arnason (1997) The mtDNA sequence of the ostrich and the divergence between paleognathous and neognathous birds. Mol. Biol. Evol. 14:754-761. Hecht, K. (1985) The biological significance of Archaeopteryx. Pp. 149-160. In: The Beginnings of Birds. Proc. Int. Archaeopteryx Conf., Eichstatt, (1984). M. K. Hecht, J. H. Ostrom, G. Viohl, and P. Wellnhofer (eds.). Freunde des Jura-Museums Eichstatt, Willibaldsburg, Eichstatt. Hedges, S. B. (1994) Molecular evidence for the origin of birds. Proc. Natl. Acad. Sci. USA 91:2621-2624. Hedges, S. B., P H. Parker, C. G. Sibley, and S. Kumar (1996) Continental breakup and the ordinal diversification of birds and mammals. Nature 381:226-229. Hedges, S. B., D. Simmons, A. van Dijk, G.-J. Caspers, W. W. de Jong, and C. G. Sibley (1995) Phylogenetic relationships of the hoatzin.
392
Carole A. Bonga Tomlinson
an enigmatic South American bird. Proc. Natl. Acad. Sci. USA 92: 11662-11665. Hiiemae, K., and A. W. Crompton (1985) Mastication, food transport and swallowing. Pp. 262-290. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. F. Liem, and D. B. Wake (eds.). Harvard Univ. Press, Cambridge, MA. Hiiemae, K. M., A J. Thexton, and A. W. Crompton (1978) Intra-oral transport: a fundamental mechanism of feeding? Pp. 181-208. In: Muscle Adaptation in the Craniofacial Region. Monograph No. 8, University of Michigan. Homberger, D. G. (1986) The lingual apparatus of the African grey parrot, Psittacus erithacus Linne (Aves: Psittacidae): description and theoretical mechanical analysis. Omith. Monogr. No. 39: 1-233. Homberger, D. G. (1988) Comparative morphology of the avian tongue. Pp. 2427-2435. In: Acta XIX Congressus Internationalis Ornithologici, Vol. II. National Museum of Natural Sciences, Univ. of Ottawa Press. Homberger, D. G. (1989) Correlations between morphology of the lingual apparatus and feeding mechanics in birds. Forsch. Zool. 35:14-150. Homberger, D. G. (1999) The avian tongue and larynx: multiple functions in nutrition and vocalization. In: Proc. 22nd Int. Ornithol. Congr. N. Adams and R. Slotow (eds.). Univ. of Natal, Durban. Homberger, D. G., and R. A. Meyers (1989) Morphology of the lingual apparatus of the domestic chicken. Callus gallus, with special attention to the structure of the fasciae. Am. J. Anat. 186:217-257. Houde, P. (1986) Ostrich ancestors found in the Northern Hemisphere suggest new hypothesis of ratite origins. Nature 324: 563-565. Houde, P. W. (1988) Paleognathous Birds from the Early Tertiary of the Northern Hemisphere. Publications of the Nuttall Ornithological Club, No. 22. Cambridge, MA. Houde, P., and H. Haubold (1987) Paleotis weigelti restudied: a small Middle Eocene ostrich (Aves: Struthioniformes). Palaeovertebrata 17:27-42. Houde, P., and S. L. Olson (1981) Paleognathous carinate birds from the Early Tertiary of North America. Science 214:1236-1237. Jean, A. (1984) Brainstem organization of the swallowing network. Brain Behav. Evol. 25:109-116. Jollie, T. (1957) The head skeleton of the chicken and remarks on the anatomy of this region in other birds. J. Morph. 100:389-436. Kallius, E. (1905) Beitrage zur Entwickelung der Zunge. II. Teil. Vogel. {Anas bochas L., Passer domesticus L.). Anat. Hefte. 85/86(2/3): 21-586. Kesteven, H. L. (1944) The evolution of the skull and the cephalic muscles: a comparative study of their development and adult morphology. III. The Sauria (Reptilia). Austral. Mus. Mem. 8: 237-269. Kesteven, H. L. (1945) The evolution of the skull and the cephalic muscles: a comparative study of their development and adult morphology. III. The Sauria (Aves). Austral. Mus. Mem. 8:270293. Kontges, G., and A. Lumsden (1996) Rhombencephalic neural crest segmentation is preserved throughout craniofacial ontogeny. Development 122:3229-3242. Kurochkin, E. N. (1995) Morphological differentiation of palaeognathous and neognathous birds. Courier Forschungsinstitut Senckenberg 181:79-88. Lang, C. (1956) Das Cranium der Ratiten mit besonderer Beriicksichtigung von Struthio camelus. Zeit. Wissensch. Zool. 159:165224. Lee, K., J. Feinstein, and J. Cracraft (1997) The phylogeny of ratite birds: resolving conflicts between molecular and morphological
data sets. Pp. 173-208. In: Avian Molecular Evolution and Systematics. D. P. Mindell (ed.). Academic Press, San Diego. Lowe, P. R. (1928) Studies and observations bearing on the phylogeny of the ostrich and its allies. Proc. Zool. Soc. Lond. 1928: 185-247. Martin, L. D. (1983) The origin and early radiation of birds. Pp. 291353. In: Perspectives in Ornithology. Martin, L. D. (1985) The relationship of Archaeopteryx to other birds. Pp. 177-183. In: The Beginnings of Birds. Proc. Int. Archaeopteryx Conf., Eichstatt, (1984). M. K. Hecht, J. H. Ostrom, G. Viohl, and P. Wellnhofer (eds.). Freunde des Jura-Museums Eichstatt, Willibaldsburg, Eichstatt. Martin, L. D. (1987) The beginning of the modern avian radiation. Docum. Lab. Geol. Lyon 99:9-19. Martin, L. D., and J. D. Stewart (1985) Homologies in the avian tarsus. Nature 315:159-160. McDowell, S. (1948) The bony palate of birds. I. The Palaeognathae. Auk 65:520-549. McGowan, C (1984) Evolutionary relationships of ratites and carinates: evidence from ontogeny of the tarsus. Nature 307:733-735. McGowan, C. (1985) Tarsal development in birds: evidence for homology with the theropod condition. J. Zool. Lond. 206:53-67. McLelland, J. (1979) Digestive system. Pp. 69-181. In: Form and Function in Birds, Vol. 1. A. S. King and J. McLelland (eds.). Academic Press, New York. Molnar, R. E. (1985) Alternatives to Archaeopteryx: a survey of proposed early or ancestral birds. Pp. 209-217. In: The Beginnings of Birds. Proc. Int. Archaeopteryx Conf., Eichstatt, (1984). M. K. Hecht, J. H. Ostrom, G. Viohl, and P. Wellnhofer (eds.). Freunde des Jura-Museums Eichstatt, Willibaldsburg, Eichstatt. Miiller, H. J. (1963) Die Morphologie und Entwicklung des Craniums von Rhea americana Linne. II Viszeralskelett, Mittelohr, und Osteocranium. Zeit. Wissensch. Zool. 168:35-118. Nei, M. (1996) Phylogenetic analysis in molecular evolutionary genetics. Annu. Rev. Genet. 30:371-403. Noden, D. M. (1983a) The embryonic origins of avian cephalic and cervical muscles and associated connective tissues. Am. J. Anat. 168:257-276. Noden, D. M. (1983b) The role of the neural crest in patterning of avian cranial skeletal, connective, and muscle tissues. Dev. Biol. 96:144-165. Oelrich, T. M. (1956) The anatomy of the head of Ctenosaura pectinata (Iguanidae). Misc. Publ. Mus. Zool. Univ. Michigan, No. 9 4 : 1 122. Olson, S. L. (1985) The fossil record of birds. Pp. 79-238. In: Avian Biology, Vol. VIII. D. S. Farner, J. R. King, and K. C. Parkes (eds.). Academic Press, London. Olson, S. L. (1986) Aspects of global avifaunal dynamics during the Cenozoic. Pp. 2023-2029. In: Acta XIX Congressus Internationalis Ornithologici, Vol. II. Ottawa. Ostrom, J. H. (1969) Osteology of Deinonychus antirrhopus, an unusual theropod from the Lower Cretaceous of Montana. Peabody Mus. Nat. Hist. Bull. 30:31-165. Ostrom, J. H. (1973) The ancestry of birds. Nature 242:136. Ostrom, J. H. (1985) The meaning of Archaeopteryx. Pp. 161-176. In: The Beginnings of Birds. Proc. Int. Archaeopteryx Conf., Eichstatt, 1984. M. K. Hecht, J. H. Ostrom, G. Viohl, and P Wellnhofer (eds.). Freunde des Jura-Museums Eichstatt, Willibaldsburg, Eichstatt. Ostrom, J. H. (1991) The question of the origin of birds. Pp. 467-484. In: Origins of the Higher Groups of Tetrapods, Controversy and Consensus. H.-P Schultze and L. Trueb (eds.). Comstock, Ithaca, NY. Padian, K., and L. Chiappe (1998) The origin and early evolution of birds. Biol. Rev. 73:1-42.
11. Feeding in Paleognathous Birds Parker, T. J. (1891) Observations on the anatomy and development of Apteryx. Phil. Trans. Roy. Soc. Lond. 182(B): 25-134. Parker, W. K. (1866) On the structure and development of the skull in the Ostrich Tribe. Phil. Trans. 1866:113-183. Parkes, K. C , and G. A. Clark (1966) An additional character linking ratites and tinamous and an interpretation of their monophyly. Condor 68:459-471. Peters, D. S. (1987) Mechanische Unterschiede palaognather und neognather Vogelschadel. Natur und Museum 117:173-182. Peters, D. S. (1988) Ein vollstandiges Exemplar von Paleotis weigelti (Aves, Palaeognathae). Cour. Forsch.-Inst. Senckenberg 107: 223-233. Prager, E. M., A. C. Wilson, D. T. Osuga, and R. E. Feeney (1976) Evolution of flightless land birds on southern continents: transferrin comparison shows monophyletic origin of ratites. J. Mol. Evol. 3:283-294. Pycraft, W. P. (1900) On the morphology and phylogeny of the Palaeognathae (Ratitae and Crypturi) and Neognathae (Carinatae). Trans. Zool. Soc. Lond. 15:149-290. Rambaut, A., and L. Bromham (1998) Estimating divergence dates from molecular sequences. Mol. Biol. Evol. 15:442-448. Sampson, S. D., L. Witmer, C. A. Forster, D. W. Krause, P. O'Connor, P. Dodson, and F. Ravoavy (1998) Predatory dinosaur remains from Madagascar: implications for the Cretaceous biogeography of Gondwana. Science 280:1048-1051. Schumacher, G.-H. (1973) The head muscles and hyolaryngeal skeleton of turtles and crocodilians. Pp. 101-199. In: Biology of the Reptilia, Vol. 4. C. Gans (ed.). Academic Press, New York. Schwenk, K. (1986) Morphology of the tongue in the tuatara, Sphenodon punctatus (Reptilia: Lepidosauria), with comments on function and phylogeny. J. Morph. 188:129-156. Schwenk, K. (1988) Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny. Pp. 569-598. In: Phylogenetic Relationships of the Lizard Families. R. Estes and G. Pregill (eds.). Stanford Univ. Press, Stanford, CA. Schwenk, K. (2000a) Functional units and their evolution. In: The Character Concept in Evolutionary Biology. G. P. Wagner (ed.). Academic Press, San Diego. Schwenk, K. (2000b) Intrinsic versus extrinsic lingual muscles: a false dichotomy? Bull. Mus. Comp. Zool. Schwenk, K., and G. S. Throckmorton (1989) Functional and evolutionary morphology of lingual feeding in squamate reptiles: phylogenetics and kinematics. J. Zool. Lond. 219:153-175. Sibley, C. G., and J. E. Ahlquist (1990) Phylogeny and Classification of Birds: A Study in Molecular Evolution. Yale Univ. Press, New Haven, CT. Sibley, C. G., and B. L. Monroe, Jr. (1990) Distribution and Taxonomy of Birds of the World. Yale Univ. Press, New Haven, CT. Simonetta, A. M. (1960) On the mechanical implications of the avian skull and their bearing on the evolution and classification of birds. Quart. Rev. Biol. 35:206-220. Smith, K. K. (1984) The use of the tongue and hyoid apparatus during feeding in lizards (Ctenosaura similis and Tupinambis nigropunctatus). J. Zool. Lond. 202:115-143. Smith, K. K. (1986) Morphology and function of the tongue and hyoid apparatus in Varanus (Varanidae, Lacertilia). J. Morph. 187: 261-287. Smith, K. K. (1988) Form and function of the tongue in agamid lizards with comments on its phylogenetic significance. J. Morph. 196:157-171. Smith, K. K. (1992) The evolution of the mammalian pharynx. Zool. J. Linn. Soc. 104:313-349. Sondhi, K. C. (1958) The hyoid and associated structures in some Indian reptiles. Ann. Zool. 2:155-240.
393
Stapel, S. O., J. A. Leunissen, Versteeg, J. Wattel, and W. W. de Jong (1984) Ratites as oldest offshoot of avian stem-evidence from alp/ifl-crystallin A sequences. Nature 311:257-259. Starck, J. M. (1995) Comparative anatomy of the external and middle ear of palaeognathous birds. Adv. Anat. Embryol. Cell Biol. 131: 1-137. Suzuki, and S. Nomura (1975) Electromyographic studies on the deglutition movement in the fowl. Jap. J. Vet. Sci. 37:289-293. Takagi, N., and Sasaki (1974) A phylogenetic study of bird karyotypes. Chromosoma (Berl.) 46:91-120. Takagi, N., Itoh, and Sasaki (1972) Chromosome studies in four species of Ratitae (Aves). Chromosoma (Berl.) 36:281-291. Tambussi, C. P. (1995) The fossil Rheiformes from Argentina. Courier Forschungsinstitut Senckenberg 181:121-129. Thexton, A. J., A. W. Crompton, and R. Z. German (1998) Transition from suckling to drinking at weaning: a kinematic and electromyographic study in miniature pigs. J. Exp. Zool. 280:327343. Throckmorton, G. S. (1976) Oral food processing in two herbivorous lizards. Iguana iguana (Iguanidae) and Uromastix aegyptius (Agamidae). J. Morph. 148:363-390. Throckmorton, G. S. (1980) The chewing cycle in the herbivorous lizard Uromastix aegyptius (Agamidae). Arch. Oral Biol. 25:225-233. Tomlinson, C. A. (1997a) Anomalous pattern of jaw-hyolingual synchrony during true cranioinertial feeding by paleognathous birds. J. Morph. 232:330. Tomlinson, C. A. (1997b) Evolutionary transformation of prokinetic neognathous birds from ''akinetic" paleognathous grade based on feeding apparatus. J. Morph. 232:331. Van den Heuvel, W. R (1992) Kinetics of the skull in the chicken {Callus gallus domesticus). Neth. J. Zool. 42:561-582. van Tuinen, M., C. G. Sibley, and S. B. Hedges (1998) Phylogeny and biogeography of ratite birds inferred from DNA sequences of the mitochondrial ribosomal genes. Mol. Biol. Evol. 15:370-376. Vrba, E. S. (1989) What are the biotic hierarchies of integration and linkage? Pp. 379-401. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, Chichester. Wagner, G. P., and K. Schwenk (1999) Evolutionarily stable configurations: functional integration and the evolution of phenotypic stability. In: Evolutionary Biology, Vol. 32. M. K. Hecht, R. J. MacIntyre, and M. T. Clegg (eds.). Plenum Press, New York. Wake, D. B., and G. Roth (1989) The linkage between ontogeny and phylogeny in the evolution of complex systemanuscript in preparation Pp. 361-377. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.). Wiley, Chichester. Webb, M. (1957) The ontogeny of the cranial bones, cranial peripheral and cranial parasympathetic nerves, together with a study of the visceral muscles of Struthio. Acta Zool. 38:81-203. Weber, V. E. (1992) Zur Evolution basicranialer Gelenke bei Vogeln, insbesondere bei Hiihner- und Entenvogeln (Galloanseres). Z. Zool. Syst. Evolut.-forsch. 31:300-317. Weishampel, D. B., P. Dodson, and H. Osmolska (1990) The Dinosauria. Univ. of California Press, Berkeley, CA. Welman, J. (1995) Euparkeria and the origin of birds. S. Afr. J. Sci. 91: 533-537. Witmer, L., and L. D. Martin (1987) The primitive features of the avian palate, with special reference to Mesozoic birds. Docum. Lab. Geol. Lyon 99:21-39. Zeigler, H. P., R. Bermejo, and R. Bout (1994) Ingestive behaviour and the sensorimotor control of the jaw. Pp. 182-200. In: Perception and Motor Control in Birds. M. N. O. Davies and P. R. Green (eds.). Springer-Verlag, Berlin.
394
Carole A. Bonga Tomlinson
Zeigler, H. P., P. Welch Levitt, and R. R. Levine (1980) Eating in the pigeon {Columba livia): movement patterns, stereotypy, and stimulus control. J. Comp. Physiol. Psych. 94:783-794. Zeigler, H. P., Miller, and R. R. Levine (1975) Trigeminal nerve and eating in the pigeon {Columba livia): neurosensory control of the consummatory responses. J. Comp. Physiol. Psych. 89:845858. Zhou, Z. (1995) The discovery of Early Cretaceous birds in China. Courier Forschungsinstitut Senckenberg 181:9-22. Zusi, R. L. (1984) A functional and evolutionary analysis of rhynchokinesis in birds. Smiths. Contr. Zool. 395:1-40. Zweers, G. A. (1974) Structure, movement, and myography of the feeding apparatus of the mallard {Anas platyrhynchos L.): a study in functional anatomy. Neth. J. Zool. 24:323-467. Zweers, G. A. (1982a) Pecking in the pigeon {Columba livia). Behaviour 81:173-230.
Zweers, G. A. (1982b) The Feeding System of the Pigeon (Columba livia L). Springer-Verlag. Zweers, G. A. (1985) Generalism and specialism in avian mouth and pharynx. Pp. 189-201. In: Functional Morphology in Vertebrates (Fortschritte der Zoologie, Vol. 30). H.-R. Duncker and G. Fleischer (eds.). Gustav Fischer, Stuttgart. Zweers, G. A. (1991a) Pathways and space for evolution of feeding mechanisms in birds. Pp. 530-547. In: The Unity of Evolutionary Biology. Dioscorides Press. Zweers, G. A. (1991b) Transformation of avian feeding mechanisms: a deductive method. Acta Biotheor. 39:15-36. Zweers, G. A., H. Berkhoudt, and J. C. Vanden Berge (1994) Behavioral mechanisms of avian feeding. Pp. 241-279. In: Biomechanics of Feeding in Vertebrates. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Adv. Comp. Environ. Physiol., Vol. 18. Springer-Verlag, Berlin.
C H A P T E R
12 Feeding in Birds: Approaches and Opportunities MARGARET RUBEGA Department of Ecology and Evolutionary Biology University of Connecticut Storrs, Connecticut 06268 covered a previously unknown feeding mechanism, surface tension transport (STT) (Fig. 12.2), in phalaropes (Phalaropus), despite decades of previous study on their foraging (e.g.. Bent, 1927; Tinbergen, 1935; Mercier and Gaskin, 1985; Jehl, 1986). Similarly, Piersma et ah (1998) demonstrated the existence of a previously unknown prey detection mechanism in red knots {Calidris canutus). Quite aside from our interest in the evolution of feeding mechanisms per se, this lack of understanding has important consequences for the evolutionary and ecological study of birds. First, without an informed understanding of feeding mechanisms, we may seriously err in our ideas about the dietary and energetic strategies available to birds, and hence about one of the most fundamental aspects of the selective regimes they operate, and evolve, under. Note that I am distinguishing here (and hereafter) between our (frequently extensive) information about what birds eat, and our relatively poor understanding of how they eat it ("food acquisition'' sensu Zweers et ah 1994) , and of how bill morphology influences the latter. Second, our knowledge of avian feeding mechanics circumscribes our ability to understand how foraging relates to other behaviors. Prey intake rates are an important component of many behavioral and energetic models of the process of habitat choice (Krebs and Davies, 1991; Sutherland, 1996). Indeed, much of optimal foraging theory was built upon studies of avian subjects (Stephens and Krebs, 1986). Yet because we rarely understand the functional relationship of feeding movements in birds to actual ingestion rate, our data frequently constitute estimates of intake rates with unknown error terms. Often we are unable even to distinguish the specific prey being taken.
I. INTRODUCTION 11. PATTERNS OF ANALYSIS A. Systematics and Choice of Taxa B. Inferring Function from Structure versus Tests of Hypotheses C. Statistical Analysis, Sample Sizes, and the Importance of Variation III. CONCLUSION References
L INTRODUCTION The feeding structures of birds are probably more diverse than those in any animals except insects (Fig. 12.1). This dramatic modification of the feeding structures in birds has attracted a good deal of attention, historically, on the basis of the idea that where there is a crossbill, there must be an interesting feeding mechanism. In addition, bird beaks and their workings have long been attractive subjects because students of evolution reasonably presume that extreme modification of structures as fundamental to survival as mouthparts is likely the result of strong selection. Indeed, Darwin's (1859) ideas about evolution by natural selection were influenced by variation in beak size and shape in Galapagos finches (Geospizinae). Studies of the influence of this variation on survival via the ability to crack hard seeds in hard times remain a classic demonstration of evolution in the wild (Boag and Grant, 1981; Grant, 1985). Nonetheless, to a great degree, avian feeding mechanics and functional morphology remain poorly understood. For example, Rubega and Obst (1993) disFEEDING (K. Schwenk, ed.)
395
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
396
Margaret Rubega
no review that examines the entire published literature with a view specifically toward identifying areas where new efforts are liable to produce new insights into the evolution of avian feeding mechanisms. Rather than merely repeat an overview of the existing literature, this chapter aims to (1) identify patterns in the nature of past and present work on avian feeding mechanics and (2) suggest areas where new investigations might be particularly informative, and approaches which I believe will be especially productive.
IL PATTERNS OF ANALYSIS A. Systematics and Choice of Taxa A survey of the literature (see earlier list of reviews) reveals that feeding structure and/or mechanisms have been investigated in a wide variety of avian taxa, but opportunities for significant new contributions are abundant. The majority of the published analyses I could locate are single species ("idiographic"; see Chapter 1) studies, and I have used the term "analysis" very broadly. In a large number of cases, the analysis consisted of an examination of beak morphology and the subsequent generation of (frequently untested) hypotheses about the functional significance of features of the beak (see Section II,B), rather than direct examinations of feeding patterns and mechanics.
FIGURE 12.1. Diversity of feeding structures in birds. (A) Hyacinth macaw (Anodorhynchus hyacinthinus), (B) southern giant petrel {Macronectes giganteus), (C) parakeet auklet {Aethia psittacula), (D) wrybill (Anarhynchus frontalis), (E) Andean avocet (Recurvirostra andina), (F) whippoorwill {Caprimulgus vociferus), and (G) African spoonbill {Platalea alba). Feeding structures and feeding mechanics of these species are unstudied. Drawings by M. J. Spring.
Because of the long history of interest in feeding in birds, there is a large literature, with multiple reviews published since the mid-1980s. Gottschaldt (1985) reviewed sensory receptors in the bill; Berkhoudt (1985) summarized information on taste receptors; Zusi (1984) presented a detailed review and analysis of avian rhynchokinesis; Vanden Berge and Zweers (1993) reviewed the myology of the avian feeding apparatus; and Zweers et al. (1994) reviewed behavioral aspects of feeding mechanisms. To date, however, there has been
FIGURE 12.2. Surface-tension feeding in phalaropes. Small invertebrate prey are seized in the tips of the jaws and are transported in the water that adheres to the bill. Water is adhesive to the surface of the bill; by spreading its jaws the bird stretches the drop. The increase in potential energy resulting from the increase in the surface area of the drop drives the drop and prey along the bird's bill into the buccal cavity. Reproduced from Rubega (1997), with permission.
12. Feeding in Birds Throughout the remainder of this chapter I refer only to feeding analyses of the following types: (1) detailed anatomical descriptions of the feeding apparatus [e.g., Homberger's (1986) now-classic treatment of the tongue in the African grey parrot, Psittacus erithacus], (2) experimental (or at least controlled) examinations of motor patterns and feeding mechanisms in live animals (e.g., Tomlinson's analysis of cranioinertial feeding in paleognaths. Chapter 11), or (3) studies that combine the two [e.g., Zweers et al.'s (1977) brilliant and comprehensive analysis of feeding and the feeding apparatus in mallards. Anas platyrhynchos]. I specifically exclude uni- or bidiraensional comparisons of bill size [e.g., tables of bill lengths, commonly found in, but not restricted to, identification guides, such as Prater et al.'s (1984) guide to Holarctic waders], casual observations of free-living birds, and untested speculation about feeding mechanisms based on either. Species in approximately 19 of 25 orders have been the subject of some form of feeding analysis. Although this may seem like rather extensive coverage of the class, it should be noted that the majority of bird families remain completely unexamined. Published work to date covers only about 49 of 158 families, i.e., details of feeding structure and mechanics are unknown for almost 70% of all families of birds. Table 12.1 identifies the orders and families of birds for which no published work on the details of either bill morphology or feeding mechanism could be located. It can be assumed that I have failed to locate every published feeding study on birds. Also, my assessment of the degree to which we are uninformed about avian feeding depends on the classification of birds used. I have used a traditional classification (Morony et ah, 1975; del Hoyo et ah, 1992), rather than a newer, still controversial classification (Sibley et ah, 1988, Sibley and Ahlquist, 1990) with fewer orders and families (see Section II,C). Nonetheless, even if my estimate of the number of published studies was doubled, my overall conclusion would not change: there are about 9000 extant species of birds, and we know little or nothing about feeding structure and mechanics for the majority of them. This survey reveals that the field lacks a phylogenetic strategy with respect to the taxa investigated. In a few cases, systematic and purposeful within-clade comparative work has been done [e.g., passerines, Passeriformes (Bock, 1960); waterfowl, Anatidae (Goodman and Fisher, 1962); woodpeckers, Piciformes (Spring, 1965); kingfishers and allies, Coraciiformes (Burton, 1984)]; however, most of the literature on avian feeding seems to have been largely driven by (a) convenience [e.g., the investigator works with a common, or commonly available, species such as
397
chickens, Gallus domesticus (e.g., Calhoun, 1933; White, 1968; Lucas and Stettenheim, 1972; McClelland, 1979; Bhattacharyya, 1980; Berkhoudt, 1985; Homberger and Meyers, 1989; Van den Fleuvel, 1992), or domestic pigeons, Columba livia (e.g., Lucas and Stettenheim, 1972; Zeigler et al, 1980; Zweers, 1982a,b, 1985; Bermejo et al, 1989), thus each of these are disproportionally well known relative to their phylogenetic importance]; (b) serendipity, in which the investigator is studying something else and makes a chance observation [e.g., observations leading to the discovery of surface tension feeding in phalaropes (Rubega and Obst, 1993) were made during the making of an educational film (University of California 1985)]; or (c) the allure of the extreme [e.g., flamingo, Phoenicopteriformes, feeding mechanisms have been much more intensively studied (e.g., Jenkin, 1957; Kear and Duplaix-Hall, 1975; Zweers et al, 1995) than those of tyrant flycatchers, Tyrannidae, for example, which are far more speciose (~ 374 species: del Fioyo et ah 1992) and widely distributed)]. To be sure, these criteria have produced a wealth of information about the diverse ways in which birds capture and process their food. Yet available data are so thinly scattered across taxa that it would be impossible to confidently assert anything about the higher-level evolution of avian feeding systems (Table 12.1). In fact, to date the field not only lacks a widely accepted general theory explaining the evolution and diversity of avian feeding mechanisms (Lauder, 1989), but lacks a core group of plausible hypotheses, which are being systematically evaluated. For example, in the context of attempting to construct a general theory, Zweers (1991a,b) has stated that pecking mechanisms occur in all modern birds, and thus pecking is the ancestral condition (Zweers et ah, 1994; Zweers and Gerritsen, 1997; Zweers and Vanden Berge, 1997). This assertion is intuitively attractive, but made in the absence of information about the feeding mechanics in two-thirds of the families of birds, its accuracy remains to be demonstrated. The hypothesis is certainly true if pecking is defined sufficiently broadly. This is not mere hairsplitting; if defined sufficiently broadly, pecking is present in all reptiles as well. What, if anything, makes avian pecking characteristically avian, as opposed to reptilian in nature? Is there only one kind of avian pecking, arising from one conserved motor pattern underlying this approach to food grasping? This would be an interesting and impressive finding. If not, how many kinds of pecking are there, how are they distributed among taxa, and what is their relationship to the diversity of feeding structures in birds? An even more compelling question is what, if anything, makes avian feeding, as a whole, characteristically avian?
398
Margaret Rubega TABLE 12.1
Unstudied Orders and Families in the Class Aves''
Common name
Approx. no. of species in family
Order
Family
Struthioniformes
Struthionidae Rheae Casuariidae Dromaiidae Apterygidae
Ostrich Rheas Cassowaries Emu Kiwis
Tinamiformes
Tinamidae
Tinamous
47
Sphenisciformes Gaviiformes
Spheniscidae Gaviidae
Penguins Loons (divers)
17 4
Podicipediformes
Podicipedidae
Grebes
22
Diomedeidae Procellariidae Hydrobatidae Pelecanoididae Phaethontidae Pelicanidae Sulidae Phalacrocoracidae Anhingidae Fregatidae
Albatrosses Petrels, shearwaters Storm petrels Diving petrels
14 70 20 4
Tropic birds Pelicans Gannets, boobies Cormorants Darters Frigate birds
3 7 9 39 2 5
Ardeidae Scopidae Ciconiidae Balaencipitidae Threskiornithidae
Herons Hamerkop Storks ShoebiU Ibises, spoonbills
60 1 19 1 32
Procellariformes
Pelecaniformes
Ciconiiformes
Phoenicopteriformes
1 2 3 1 3
Order
Family
Common name
Approx. no. of species in family 8 2 1 7 1 13 9
Charadriidae Scolopacidae Thinocoridae Chionididae Stercorariidae Laridae Rynchopidae Alcidae
Jacanas Painted snipe Crab plover Oystercatchers Ibisbill Avocets, stilts Thick knees Coursers, pratincoles Plovers Sandpipers, snipe Seedsnipe Sheathbills Skuas Gulls, terns Skimmers Auks
Columbiformes
Pteroclididae Columbidae
Sandgrouse Pigeons, doves
16 283
Psittaciformes
Loriidae Cacatuidae Psittacidae
Lories Cockatoos Parrots
55 18 271
Cuculiformes
Musophagidae Cuculidae
Turacos Cuckoos
19 130
Strigiformes
Tytonidae Strigidae
Barn owls "Typical" owls
12 134
Caprimulgiformes
Steatornithidae Podargidae Nyctibiidae Aegothelidae Caprimulgidae
Oilbird Frogmouths Potoos Owlet-nightj ars Nightjars
1 13 5 8 76
Apodidae Hemiprocnidae Trochilidae
Swifts Tree swifts Hummingbirds
82 4 338
Charadriiformes
5
Jacanidae Rostratulidae Dromadidae Haematopodidae Ibidorhynchidae Recurvirostridae Burhinidae Glareolidae
16 64 86 4 2 5 90 3 23
Phoenicopteridae
Flamingoes
Anseriformes
Anhimidae Anatidae
Screamers Ducks, geese, swans
Falconiformes
Cathartidae Pandionidae Accipitridae Sagittariidae
7 New World vultures, Osprey 1 217 Hawks, eagles Secretary bird 1
Apodiformes
Coliiformes
Coliidae
Mousebirds
Galliformes
Megapodiidae Cracidae
Trogoniformes
Trogonidae
Trogons
37
44 213 1
Coraciiformes
Phasianidae Opisthicomidae
Megapodes Guans, chachalacas. currasows Pheasants, grouse Hoatzin
Mesitornithidae Turnicidae Pedionomidae Gruidae Aramidae Psophiidae Rallidae Heliomithes Rhynochetidae Eurypygidae Cariamidae Otididae
Mesites Button quails Plains wanderer Cranes Limpkin Trumpeters Rails, coots Finfoots Kagu Sunbittern Seriemas Bustards
3 14 1 15 1 3 133 3 1 1 2 24
Alcedinidae Todidae Motmotidae Meropidae Coraciidae Brachypteraciidae Leptosomatidae Upupidae Phoeniculidae Bucerotidae
Kingfishers Todies Motmots Bee eaters Rollers Ground rollers Cuckoo roller Hoopoe Woodhoopoes Hornbills
90 5 9 21 11 5 1 1 8 44
Piciformes
Galbulidae Bucconidae Capitonidae Indicatoridae Ramphastidae Picidae
Jacamars Puffbirds Barbets Honeyguides Toucans Woodpeckers
17 34 81 14 33 204
Gruiformes
3 147
12
6
(continues)
399
12. Feeding in Birds TABLE 12.1 (continued)
Order
Family
Passeriformes
Eurylaimidae Dendrocolaptidae Furnariidae Formicariidae Conopophagidae Rhinocryptidae Cotingidae Pipridae Tyrannidae Oxyruncidae Phytotomidae Pittidae Xenicidae Philepittidae Menuridae Atrichornithidae Alaudidae Hirundinidae Motacillidae Campephagidae Pycnonotidae Irenidae Laniidae Vangidae Bombycillidae Dulidae Cinclidae Troglodytidae Mimidae Prunellidae Muscicapidae
Common name
Approx. no. of species in family
14 Broadbills 52 Woodcreepers Ovenbirds 218 Antbirds 228 Gnateaters 11 Tapaculos 30 Cotingas 79 57 Manakins Tyrant flycatchers 374 Sharpbill 1 Plantcutters 3 Pittas 24 New Zealand wrens 4 Asities 4 Lyrebirds 2 Scrub birds 2 Larks 77 Swallows, martins 80 Wagtails, pipits 54 Cuckooshrikes 70 Bulbuls 123 Leafbirds, ioras. fairy bluebirds 14 Shrikes 74 Vanga shrikes 13 Waxwings 8 Palmchat 1 Dippers 5 Wrens 59 Mockingbirds, thrashers 31 Accentors 12 Thrushes and allies 1423
Order Passeriformes (continued)
Family Aegithalidae Remizidae Paridae Sittidae Certhiidae Rhabdornithidae Climacteridae Dicaiedae Nectariniidae Zosteropidae Meliphagidae Emberizidae Parulidae Drepanididae Vireonidae Icteridae Fringillidae Estrilididae Ploceidae Sturnidae Oriolidae Dicruridae Callaeidae Grallinidae Artamidae Cracticidae Ptilonorhynchidae Paradisaeidae Corvidae
Common name
Approx. no. of species in family
Long-tailed tits Penduline tits Tits, chickadees Nuthatches Treecreepers Philippine creepers Australian creepers Flowerpeckers Sunbirds White eyes Honeyeaters Buntings, cardinals. tanagers New World warblers Hawaiian honeycreepers Vireos New World blackbirds Finches Waxbills Weavers, sparrows Starlings Orioles Drongos Wattlebirds Magpie-larks Woodswallows Butcherbirds Bowerbirds Birds of paradise Crows, jays
8 10 27 25 6 2 6 58 116 83 171 558 126 23 43 95 122 127 143 111 28 20 3 4 10 8 18 42 105
^Taxa for which I could identify no published studies of the functional morphology of the feeding structures or feeding mechanics are bold faced. Classification is traditional and follows del Hoyo et al (1992) and Morony et ah (1975).
Progress currently is hampered by our lack of focus, phylogenetically speaking. A long list of authors have persuasively stated the case for a phylogenetic approach to understanding the evolution of complex behaviors and their ecological relevance (see citations summarized in Brooks and McClennan, 1991; Losos and Miles, 1994). It is therefore striking to note that the modern study of feeding mechanics in all vertebrates, except birds (and possibly mammals), is proceeding within an explicitly evolutionary framework, using phylogenetic tools and approaches (see other chapters in this book). What factors have prevented investigators of avian
feeding mechanisms from following suit? As with other groups of vertebrates, the bulk of all work to date was done prior to the rise of phylogenetic methods. More recently, ornithologists have been hampered by the lack of a rigorous phylogeny for the class. The ordinal level relationships of birds are still poorly understood (Raikow, 1985; Cracraft, 1988). The development of a phylogeny for the class is impeded by an insufficient inventory of cladistic characters (Cracraft, 1988). Such a phylogeny is essential to understanding the evolution of avian feeding mechanisms, as a basis for the generation of sampling schemesn, and for mapping character states.
400
Margaret Rubega
This is not to say that no higher-level phylogenetic analyses exist; Sibley and associates have provided a phylogenetic hypothesis for the whole class (Sibley et ah, 1988; Sibley and Ahlquist, 1990). Unfortunately, their characters, methods, and assumptions have serious weaknesses and have been widely criticized (e.g., Krajewski, 1991; O'Hara, 1991; Raikow, 1991; Lanyon, 1992). It has been argued, however, that this phylogeny at least presents us with a working hypothesis (O'Hara, 1991), and thus a starting place, and an opportunity for tests of hypotheses about the evolution of feeding mechanisms. To date, there are no alternative higherlevel phylogenies for the whole class. Nonetheless, ornithologists have been among the most active vertebrate systematists, and phylogenies below the ordinal level now are available for many groups of birds. A shortage of phylogenies, of course, is not the only factor preventing a shift toward phylogenetic (cladistic) methods. Some of the most active investigators of avian feeding methods and evolution have deliberately eschewed a phylogenetic approach in order to pursue alternate strategies of investigation. Zweers and colleagues, the most prolific and productive group currently working on avian feeding mechanics, have been developing an approach that deduces likely pathways of phenotypic transformation in avian feeding systems (Zweers, 1991a,b; Zweers and Vanden Berge, 1997; Zweers and Gerrittsen, 1997). They use functional optimality as the criterion for morphological and mechanistic change in order to generate testable ideas about the domain of all possible feeding mechanisms and evolutionary pathways to those mechanisms (see Chapter 1). In other words, they are using mechanical principles to generate ideas about how feeding mechanisms and bill morphology might have evolved, rather than looking at the distribution of feeding characters on a phylogeny, and then inferring the direction and nature of evolution. This approach shows some promise as a tool for generating hypotheses, but suffers from potential circularity. (The choice of optimal morphologies and mechanisms is unavoidably drawn from extant examples, which are then mapped onto the resulting transformation scheme, and are found to match.) Such a method will ultimately require grounding in a phylogenetic framework if it is to serve as a tool for understanding the actual evolution of existing feeding mechanisms. For instance, Zweers and associates (Zweers, 1991a,b; Zweers et al, 1994; Zweers and Gerritsen, 1997; Zweers and Vanden Berge, 1997) have proposed that avian filter-feeding mechanisms arose via modification of a charadriiform-like bill structure, and of the motor patterns associated with pecking. This idea posits the following sequence of events: (1) Shorebird-like bill structures are a modification of a more generalized
(pigeon-like) ancestral beak. In this first modification, elongation and slenderization contributed to improved probing performance. (2) Surface-tension transport, a mechanism by which birds transport prey along the bill using the physical properties of water droplets (Rubega and Obst, 1993; Rubega, 1997), arose as an epiphenomenon of this change in bill morphology. (3) Further modification of the bill structure and motor patterns arose as a consequence of specialization. (4) Modifications that increased the volume of prey and water processed in one feeding cycle led to flamingo- and duck-like bill morphologies, and hence to filter-feeding mechanisms. The initial generation of these ideas took place in a strictly biomechanical rather than phylogenetic context, and thus was not accompanied by the directed sampling of feeding structure and mechanism necessary to test most of the resultant hypotheses. An additional problem is that trophic mechanisms or morphologies are not coded in a manner that would lend itself to cladistic analysis. Zweers and Vanden Berge (1997) do provide a phenogram in which key trophic mechanisms and transitions are overlaid with the names of taxa that putatively have the mechanisms (Fig. 12.3), and then compare their scheme of "trophic radiation" to available phylogenetic analyses. Unfortunately, their phenogram includes many taxa for which detailed analysis of feeding mechanisms have not been conducted (e.g., stone curlew, Burhinus oedicnemus; spoonbill sandpiper, Eurynorhynchus pygmeus; crab plover. Dramas ardeola; ruff, Philomachus pugnax; Eurasian curlew, Numenius arquata; the screamers Anhimidae), or can never be done (e.g., Preshyornis, a fossil bird with a duck-like head and a shorebird-like axial skeleton). Also, they fail to map their character states (mechanisms or morphologies) directly onto existing phylogenies. Thus, it is difficult to evaluate their conclusion that their scheme of phenotypic transformation is largely congruent with cladistically produced phylogenies. Nonetheless, this detailed set of hypotheses can provide a basis for designing a sampling scheme that would contribute to our understanding of the evolution of feeding mechanisms in the shorebirds (Charadriiformes). There are a number of explicit predictions resulting from the Zweers model that can be tested. First, the model postulates that the ability to use surface-tension prey transport is simply a consequence of the basic shorebird bill structure. If true, then the capacity to employ this feeding mechanism should be found not only in the species in which it was discovered, the highly aquatic red-necked phalarope {Phalaropus lohatus), but in every shorebird with a straight, needlelike bill. Initial steps in a survey of the whole shorebird clade indicate that surface-tension transport of prey is indeed available to other phalaropes (Wilson's phala-
401
12. Feeding in Birds
organic ooze scraping mechanism
t
catching fish I shoveling shells 1 ^ grazing m's \ 1 A dabbling m's
\lt/
scaling of filter-feeding mechanisms
scaling of filter-feeding mechanisms
\t/
\t/
M
size & hardness recording touch mechanism: along jaw rami L- , « , . back & forth pump mechanism
at jaw tips
scaling of remote touch mechanisms stretching curved beak curving beak ^ combined vertical wedge mechanism \ remote 1 touch & \ 1 penetration filter-feeding m.
stab & crunch/ split & cut mechanisms
M e r g u s (merganser)
Polysticta Somateria (eider) Aythya (scaup)
Limnodromus (dowitcher) Scolopax (woodcock) Calidris (stint, sandpiper) Gallinago (snipe) T r i n g a (shank)
A n a s (wigeon)
L i m o s a (godwit) Numenius (curlew) Philomachus (ruff)
\u
\
grubbing '^'''®^^ ^^^^^^ hunting m's horizontal wedge mechanism curving/scaling/widening' mechanism
t suction pressure pump mechanism
\\t
grazing & filter-feeding (water) hole inspection mechanism mechanisms A T / scaling curved beak / probing mechanisms i ^
substrate penetration mechanisms
deep probe-hunting mechanism
grasp-pump graze-filter
\f
fetch and carry mechanism
^
»-
inspection/ turn-over/ chase m's
combined sight-peck & touch-probe m.
sit-watch & run-peck m. & superficial-probe m. stretch-catch m. & scaling , ^ _ ® T walk & peck mechanism top-soil breaking m. , ^ ^ fastened ingestion m. I sight-peck mechanism
f
grasp mechanism shores/ wetlands
Ciconidae (storks) Ardeidae (herons)
Burhinidae (stone curlews) shorebird-like ancestor
G a l l i f o r m e s (fowl) C o l u m b i f o r m e s (doves) pecking ancestor
FIGURE 12.3. A hypothesis of phenotypic transformations of avian feeding systems resulting in probing and filter feeding. (A) The branching pattern and the mechanisms along it were deduced by modifications of a pecking mechanism, optimized for probing and filter-feeding functions. (B) A phenogram of hypothetical evolutionary change in avian feeding systems, produced by overlaying taxa with appropriate feeding systems on (A). Reproduced from Zweers and Vanden Berge (1997), with permission.
rope, P. tricolor), as well as other species of shorebirds, including western {Calidris mauri) and least (C. minutilla) sandpipers (Rubega, 1997). The latter two species generally feed by probing in sandy or muddy substrates, hence it is unlikely that STT is a specialization for an aquatic lifestyle. Second, it follows from the optimization criteria used by Zweers that character states within the shorebirds that deviate from this basic needle-like bill morphology are derived and thus will be accompanied by improved performance of some other (new) feeding mechanism. The physical model for STT requires that deviations from a straight needle-like bill will result in a reduction in performance of surface-tension feeding (Rubega and
Obst, 1993). Some evidence for intra- and interspecific STT performance variation exists (Rubega, 1996,1997), but sampling of a broader array of shorebird bill morphologies would be informative. One interesting observation points to the importance of detailed and quantitative performance testing: American avocets {Recurvirostra americana) hatch with a needle-like bill that subsequently develops into a structure that is markedly dorsoventrally flattened and recurved. Hatchlings employ surface-tension transport throughout the transition from one morphology to another (Harker, Rubega, and Oring, unpublished observation). Field observations indicate that mean feeding performance increases as chicks grow (i.e., during
402
Margaret Rubega
the transition from a needle-shaped bill to one that is dorsoventrally flattened and recurved) (Harker and Rubega, unpublished observation). This appears to contradict the prediction that STT performance should decrease with deviations from a needle-shaped bill. These field observations, however, cannot distinguish improved performance at the individual level from improvements in the mean performance of age classes, which could result from the elimination of poorly performing individuals from the population by selection. Only quantitative measures of clearly defined performance characters under controlled conditions (i.e., in the laboratory) allow us to test directly the relationships between variation in morphology and feeding mechanics (e.g., Rubega, 1996). A third prediction of Zweers' model for the evolution of filter feeding is that character transformations, including increased bill volume, a tongue-based water pump, and straining structures, lead to a filter-feeding mechanism (Zweers et al., 1994). In its general outline of progression from a simple bill with low internal volume to a higher-volume, complex bill with filtering structures, this model is plausible, even compelling. As a phylogenetic hypothesis for the evolution of filter feeding in extant birds, the model is at odds with phylogenetic information, as it appears to suggest that the Anseriformes (the avian lineage in which filter feeding is most widespread and developed) arose from a shorebird (charadriiform) ancestor (Zweers and Vanden Berge, 1997; see Fig. 12.3B). All available evidence points to a sister-group relationship between the Anseriformes and Galliformes, which together form a clade that is the sister group of all other neognaths, with no special relationship to the charadriiform clade (Ho et al, 1976; Sibley et al, 1988; Cracraft, 1988). Nonetheless, it is conceivable that, despite their galliform relationship, anseriform ancestors had a simple, plover-like bill, which might have been subsequently modified for filter feeding. There is no paleontological evidence for or against this idea. With the proper sampling opportunities, however, we could test Zweers' ideas about the evolution of filter feeding within the Anseriformes. Although there are no anseriform extant taxa with Zweers' hypothesized ancestral plover-like (simple) bill morphology, the basic ideas of the model may be testable within the charadriiform lineage instead. For example, red phalaropes {Phalaropus fulicaria), which are so closely related to red-necked phalaropes as to be virtually indistinguishable genetically (Dittman et al, 1989; Dittman and Zink, 1991), have a bill that is wider and deeper (i.e., has a larger internal volume). Red phalaropes also have small internal bill structures that may be simple filtering systems (personal observation). Evidence shows that they select prey within
a rather narrow size range (Dodson and Egger, 1980; Mercier and Gaskin, 1985), as would be expected if filter size limits prey-capture performance. Red phalaropes thus provide a putative intermediary in which to examine the mechanistic predictions of the Zweers model for the evolution of filter feeding. If these predictions appear to be supported, it would be of interest to consider what factors may have prevented further evolution of filter feeding, which is otherwise not known to be present among the shorebirds. B. Inferring Function from Structure versus Tests of Hypotheses Possibly no group of vertebrates has a richer history of anatomical description than birds. Beginning with Aristotle, a long line of investigators has been carefully dissecting and describing birds in an attempt to understand their anatomy. From the 16th century onward, the detailed description oLanatomy was particularly important, second only to plumage descriptions (which were frequently all an investigator had, prior to the discovery of methods to preserve tissues) as a means of classification (Stresemann, 1975). When binoculars became widely available early in the 20th century, the focus of mainstream ornithology shifted to avian behavior, but beautiful and useful anatomical descriptions continued to be produced, particularly among German and Dutch investigators (e.g., Fiirbringer, 1922). While the anatomical descriptions in these studies were masterful and comprehensive, they constituted only the first step in understanding the role morphological structures play in avian feeding. Attempts to understand the anatomy they were describing naturally led investigators to formulate hypotheses about the function and evolutionary significance of various structures and structural complexes. Unfortunately, there has been a tendency in the ornithological literature to elevate such hypotheses to the status of fact. Nowhere has this practice been more evident than in the description of the avian feeding apparatus. Because of the obvious and dramatic modifications of the bill, there has been much speculation on the functional relationship between bill structures and feeding mechanics. In the most common approach, feeding mechanics are inferred from morphological features revealed by anatomical dissection, rather than directly observed or experimentally verified (see discussion in Chapter 1). This approach, dubbed "adaptive storytelling" in Gould and Lewontin's (1979) now-famous "Spandrels" paper and extensively criticized since, has been slow to fade in the avian literature and is still surprisingly common. An examination of the beaks of different birds that feed on the same prey provides
403
12. F e e d i n g in Birds
ample demonstration that many tools can do one job (Fig. 12.4) and implies that they are unlikely to do the job in simple, easily predicted ways. It seems likely that adaptive storytelling about feeding in birds is driven more by what is already known about the diets of birds than by informed understanding about the relationship between feeding structures and function. Occasionally, the storytelling approach is given a veneer of experimentalism via manipulation of specimens in ways meant to reveal the functional relationship among structures (e.g., pulling on a muscle to see if the mouth opens). The weakness of this approach is nicely illustrated by Dial's (1992) study of the avian flight apparatus, which showed clearly that flapping flight in birds is powered by different muscles, firing at different times, than anatomical description and manipulation of dead specimens had led investigators to believe. The frequency of the earliest adaptive storytelling is not surprising, not least because many of the hypotheses generated by anatomists would have been difficult or impossible to test without modern technology. Nonetheless, adaptive storytelling appears to have persisted for a series of obvious and not-so-obvious reasons. First, it is, quite simply, easier to formulate a hypothesis than it is to test it. Even in the post-Spandrel era, when most investigators have learned to label clearly untested hypotheses as speculation, such hypotheses are often untested. For example, Zusi (1984) pointed out that hypotheses concerning cranial kinesis arising from his anatomical analysis of bony hinges in bird skulls should be tested against observations of bill movement in living birds. It appears they never have been. Sometimes this failure to follow-up tends to occur because our ideas about the relationship between form and function in avian feeding have not been translated into formalized, falsifiable null and alternative hypotheses, and thus are weak generators of testable predictions. For instance, ornithologists have long guessed that the unique spinning behavior of phalaropes serves to "stir u p ' ' prey from the bottoms of ponds and lakes where they were feeding, but this idea failed to explain why birds spin while at sea over water many fathoms deep. It was only when this idea was formalized into testable hypotheses about the specific patterns of water flow generated by spinning that it became possible to show that spinning by phalaropes does draw prey to the surface, but by creating an upwelling rather than by stirring {Ohstetal,1996). As in other vertebrates, it often is difficult even to guess how intricate structural complexes might function (e.g., there are many unanswered questions about the functioning of the avian tongue in feeding and drinking; Homberger, 1988), hence it is difficult to generate clear, testable predictions about the relation-
K ^ ^ . l } . . . . , "I rry
F I G U R E 12.4. An example of the diversity of feeding structures associated with feeding on a single prey type. All these species eat fish. (A) Brown pelican {Pelecanus occidentalis), (B) horned puffin {Fratercula comiculata), (C) common loon (Gavia immer), (D) shoebill {Balaeniceps rex), and (E) red-breasted merganser {Mergus senator). Drawings by M. J. Spring.
ship between parts and feeding mechanics. There appears to be no cure for this problem but creativity and empiricism.
404
Margaret Rubega
To some extent, untested hypotheses based only on descriptions of structure have also accumulated because birds are uniquely vagile organisms, thus the range of feeding circumstances is huge, and the opportunities for observation (much less manipulation) of live, feeding birds are limited. Taxa that feed in midair (e.g., flycatchers, see discussion later) or underwater (e.g., penguins, Spheniscidae, and auks, Alcidae) are particularly poorly known. Even when observation is possible and good hypotheses exist, in many cases there are significant technical barriers to testing hypotheses directly, e.g., the feeding event happens too rapidly to be discernable with the naked eye. For example, in rednecked phalaropes, surface tension transport can be completed in as little as 0.002 sec (Rubega and Obst, 1993) and would have been undetectable without the aid of high-speed videography. Occasionally, we fail even to discard hypotheses that have been falsified. For example, ornithologists have long guessed that the brush-like array of rictal bristles (feathers modified into fine, rather stiff, whiskers; Fig. 12.IF) present around the mouth margins of many species of fly-catching birds function to funnel prey into the gape. Lederer (1972) presented strong evidence that this is unlikely, and Conover and Miller's (1980) study of willow flycatchers {Empidonax traillii) clearly demonstrated that this is not the case. Birds caught prey equally well both before and after rictal bristles had been removed. Yet textbooks continue to assert that rictal bristles function as insect nets (e.g.. Gill, 1995). This may simply be a demonstration of the difficulty inherent to the dissemination of results at a time when investigators are more overwhelmed with new literature than ever before. Alternatively, this example may merely demonstrate that we are fonder of a good story than of the facts. This is unfortunate, as the facts generally are more interesting than any story we could make up. What are rictal bristles for? They are present to a greater or lesser extent in many birds (Lederer, 1972) (including birds such as kiwis, Dinornithiformes, flightless birds, which forage in leaf litter), but dense, basketlike arrays of them around the margins of the mouth have apparently arisen independently more than once in birds. A partial list of birds with prominent rictal bristles includes New World flycatchers (Tyrannidae), Old World flycatchers (Muscicapidae), shrikes (Lannidae), and frogmouths (Caprimulgidae). In all cases, insect capture on the wing is a significant part of the feeding biology. Conover and Miller (1980) presented evidence that rictal bristles may function to protect the eyes from strikes by missed prey or from parts of prey that may break up when seized. Increased emphasis on experimentalism and availability of new tools (e.g.. X-ray cineradiography and high-speed film and video) have contributed to a wel-
come and growing tendency to approach avian feeding with formalized hypothesis testing. Some recent examples include Benkman's elegant analysis of crossbill {Loxia sp.) feeding (Benkman, 1987,1988; Benkman and Lindholm, 1991); the impressive body of work amassed by Zweers and associates on greater flamingoes Phoenicopterus ruber (Zweers et al, 1995), pigeons (Zweers, 1982), ducks {Anas platyrhynchos, A. clypeata, and Ay thya fuligula) (Zweers et ah, 1977; Kooloos et ah, 1989), and sandpipers (Calidris sp.) (Gerritsen et ah, 1983; Gerritsen and van Heezik, 1985; Gerritsen and Meiboom, 1986); my own work on feeding mechanics in phalaropes (Rubega and Obst, 1993; Rubega, 1996,1997); Hulscher and Ens's (1991) analysis of the functional significance of bill shape in Eurasian oystercatchers {Haematopus ostralegus), and the clever experimental work of Piersma et al. (1998) on red knot prey detection mechanisms. In all these cases, real progress in our understanding of feeding in birds was achieved by the observation of live animals at close range under controlled (i.e., laboratory) conditions, inventive experimental approaches, the application of appropriate technology to reveal details of feeding mechanics, or all three. Most importantly, all these tools were employed in the deliberate testing of formalized, falsifiable hypotheses about the relationship of feeding structures to mechanisms of food capture and processing. C. Statistical Analysis, Sample Sizes, and the Importance of Variation Historically, investigators of feeding in birds have tended to base their studies on observations of few individuals. In some situations this is acceptable, but most of our understanding of avian feeding mechanisms is hampered by reliance on small sample sizes. Studies of avian feeding can be broken down into (1) those that primarily describe phenomena and (2) those that compare groups of organisms. Descriptions of phenomena do not require large sample sizes. A sample of one is sufficient to demonstrate that a structure or mode of feeding exists. Even in these studies, however, assessing whether the phenomenon occurs in more than one or two individuals is important to ensure that the observations are not aberrant. When one wants to compare groups (e.g., comparing structure among species or trying to relate performance to variation in morphology), a rigorous statistical analysis becomes important (Shaffer and Lauder, 1985a,b). Results of statistical analyses are only meaningful when applied to appropriate sample sizes. The vast majority of published studies of the feeding apparatus and feeding function, however, are based on fewer than five in-
12. Feeding in Birds dividuals (I am guilty of this myself: Rubega and Obst, 1993,1997); in many cases, the sample size is one. Why do avian feeding specialists persist in presenting results from such small samples? One of the most obvious reasons for this problem is the difficulty in obtaining, and keeping, sufficient numbers of live, healthy specimens. This problem is not unique to birds, but perhaps uniquely complicated by their volant nature. Birds can be much more difficult to catch than fish, lizards, or small mammals. Once caught, all but the smallest species of birds also require significantly more space and attention for captive maintenance. Experimental feeding setups for some species of birds (e.g., pursuit diving birds) can be too demanding of space and resources anywhere outside of a zoological park. Birds held in zoos are only rarely available for manipulative experiments. These problems are real, but by no means sufficient to explain the widespread lack of statistical rigor in the field. For example, warblers are completely unstudied with respect to feeding mechanics. Yet many species are widespread, abundant, easily caught in nets (as evidenced by the thousands banded yearly for studies of movement patterns), and require no more space for captive maintenance than a typical lizard or snake. The same is true for many other families of passeriform birds. Our failure to direct our attention to the opportunities present in these taxa is probably due to patterns identified earlier (see Section II,A). An important contributor to the lack of statistical rigor in the field is that journal editors and reviewers have continued to allow investigators of avian feeding mechanics to publish with small samples. This appears to be due, at least in part, to a tradition of belief that feeding patterns are "hardwired" (genetically inherited, rather than learned, and therefore largely invariant), thus a sample of one is as representative, and as informative, as a larger sample. As I have repeatedly pointed out, we actually have very little detailed information on the feeding process in birds, but we have enough to know that the notion of feeding patterns as invariant within a species must be at least partly false. First, although there is certainly a genetic component to control of the feeding process and development of the feeding apparatus in birds, it would be surprising in the extreme to find complete genetic fixation for most traits in the feeding complex. The huge range of variance in bill shape and feeding patterns across the class Aves attests, at a minimum, to the historical availability of population-level variation in feeding structures and selection for their modification. Further, evidence shows that extrinsic factors may influence adult bill morphology (and presumably feeding performance, if not pattern) via developmental plasticity James (1983) and NeSmith (1984; cited in Travis,
405
1994), for example, showed that variation in temperature and humidity induces variation in bill shape of nestling red-winged blackbirds {Agelaius phoeniceus). The existence and importance of interspecific feeding variation have mostly been assumed on the basis of observed variation in bill morphology among related species or inferred from observations of differential habitat use and diets. Few direct comparisons of multiple species employing a common feeding mechanism on a standard food type have been conducted. Exploratory analyses (of data from a small sample of individuals) indicate that there is significant interspecific variation among four species of shorebird (red-necked phalarope, Wilson's phalarope, western sandpiper, and least sandpiper) in the performance of surface-tension feeding. Motor patterns are similar, but vary quantitatively among species (Rubega, 1997). Variance in feeding structures, process, and performance clearly exists within species as well. For example, it has been demonstrated repeatedly that juvenile birds exhibit poor feeding performance (usually expressed as feeding efficiency, or the catch-to-attack ratio; poor performance is also inferred from diet restricted to prey that is presumed to be less preferred) relative to their adult conspecifics (for reviews, see Marchetti and Price, 1989; Wunderle, 1991). Increasing age is associated with improved feeding performance. Explanations offered for this pattern of an ontogenetic feeding shift include learning, physical maturation of the feeding apparatus, variance in the nutritional status of juveniles relative to adults, and competitive suppression of juvenile feeding by adults. To date, none of these have been accompanied by formalized, testable hypotheses linking them to the feeding mechanism itself. Additional explanations that are well worth pursuing include the effects of neurological maturation, the possibility that juveniles may exhibit superior performance (relative to adults) of "juvenile" feeding mechanisms, and the likelihood that the perceived improvement in mean feeding performance with age is due to the elimination of poorly performing individuals from the population due to selection. Finally, significant variation in feeding within groups (among individuals) has also been demonstrated. Red-necked phalaropes exhibit significant among-individual variation in the performance of surface-tension feeding as a function of morphological variation of the inside of the upper jaw (Rubega, 1996). It should be apparent by this point in this chapter that, aside from the importance of accounting for variability when assessing the generality of our conclusions about avian feeding mechanisms, variance in avian feeding structures and mechanics is (or should be), in itself, a statistic of interest to us. This is especially true given the extreme degree of variation in feeding in birds overall relative to other vertebrates. In any group of
406
Margaret Rubega
vertebrates, variation (especially at the individual level) is the basis for selection and thus reflects the opportunity for, and outcome of, the evolution of feeding mechanics. In particular, attention to individual-level variation may be revealing with respect to the direction of selection, which in turn may provide us with clues as to which aspect of function is selectively most important.
III. C O N C L U S I O N The exceptional degree of variation in the avianfeeding apparatus has stimulated a large and interesting literature on beak and tongue morphology, feeding mechanics, and behavioral aspects of foraging. Much, however, remains to be done. I have tried to show that opportunities for significant new contributions to our understanding of feeding in birds are abundant. There is no widely accepted general theory explaining evolution of the observed diversity in avian feeding mechanisms. This is at least partly due to the complete lack of information about many species of birds: detailed analyses of feeding structures and function are lacking for more than half of all families of birds. To date our choice of taxa has been largely opportunistic. Significant advances will be made only when a phylogenetic strategy is applied to the problem of choosing study taxa. Further advances will also require the controlled testing of formal hypotheses about the relationship of feeding structures to some aspect of function or performance, and coding of feeding mechanisms in ways that allow cross-taxa comparisons. Although the difficulties inherent in maintaining birds in captivity are not trivial, our confidence in the outcome of comparative studies will depend on statistically appropriate sample sizes, however difficult they may be to attain. Some of the tools required to achieve these goals (such as high-speed video cameras and family-level phylogenetic hypotheses) are increasingly available. Widely and properly applied, they could produce a renaissance in the study of avian feeding. Even in the absence of a renaissance, we stand to learn a great deal more about feeding in birds. References Benkman, C. W. (1987) Crossbill foraging behavior, bill structure, and patterns of food profitability. Wilson Bull. 99:351-368. Benkman, C. W. (1988) On the advantages of crossed mandibles: an experimental approach. Ibis 130:288-293. Benkman, C. W., and A. K. Lindholm (1991) The advantages and evolution of a morphological novelty. Nature 349:519-20.
Bent, A. C. (1927) Life histories of North American shore birds. Order Limicolae (Part I). U. S. National Museum Bulletin 142. Berkhoudt, H. (1985) Structure and function of avian taste receptors. In: Form and Function in Birds, Vol. 3. A. S. King and J. McClelland (eds). Academic Press. Bermejo, R., R. W. Allan, D. Houben, J. D. Deich, and H. P Zeigler. (1989) Prehension in the pigeon. I. Descriptive analysis. Exp. Brain Res. 75:569-576. Bhattacharyya, B. N. (1980) The morphology of the jaw and tongue musculature of the common pigeon, Columba livia, in relation to its feeding habit. Proc. Zool. Soc. Calcutta 31:95-127. Boag, P. T, and P. R. Grant (1981) Intense natural selection in a population of Darwin's finches (Geospizinae) in the Galapagos. Science 214:82-85. Bock, W. J. (1960) The .palatine process of the premaxilla in the Passere. Bull. Mus. Comp. Zool. 122(8):361-488. Burton, P. J. K. (1984) Anatomy and evolution of the feeding apparatus in the avian orders Coraciiformes and Piciformes. Bull. Brit. Mus. (Nat. Hist.) 47(6): 331-443. Brooks, D. R., and D. A. McLennan. (1991) Phylogeny, Ecology and Behavior; a Research Program in Comparative Biology. University of Chicago, Chicago. Calhoim, M. L. (1933) The microscopic anatomy of the digestive tract of Gallus domesticus. Iowa State Coll. J. Sci. 7:261-382. Conover, M. R., and D. E. Miller (1980) Rictal bristle function in willow flycatcher. Condor 82:469-471. Cracraft, J. (1988) The major clades of birds. In: The Phylogeny and Classification ofTetrapods, Vol. 1. M. J. Benton (ed). Systematics Association Special Volume No. 35A. Clarendon Press, Oxford. Darwin, C. 1859. On the Origin of Species hy Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray, London. Del Hoyo, J., A. Elliott, and J. Sargatal (eds.) (1992) Handbook of the Birds of the World Vol. 1. Lynx Edicions, Barcelona. Dial, K. P. (1992) Avian forelimb muscles and nonsteady flight: can birds fly without using the muscles in their wings? Auk 109: 874-885. Dittman, D. L., and R. M. Zink (1991) Mitochondrial DNA variation among phalaropes and allies. Auk 108:771-779. Dittman, D. L., R. M. Zink, and J. A. Gerwin. (1989) Evolutionary genetics of phalaropes. Auk 106:326-331. Dodson, S. I., and D. L. Egger. (1980) Selective feeding of red phalaropes on zooplankton of Arctic ponds. Ecology 61:755-763. Fiirbringer, M. (1922) Das Zungenbein der Wirbeltiere insbesondere der Reptilien und Vogel. Abhandlungen der Heidelberger Akademie, math.-nathurw. Kl. Abt.B 11:1-164. Gerritsen, A. F. C , and A. Meijboom (1986) The role of touch in prey density estimation by Calidris alba. Neth. J. Zool. 36:530-562. Gerritsen, A. F. C , and Y. M. van Heezik (1985) Substrate preference and substrate related foraging behavior in three Calidris species. Neth. J. Zool. 35:671-692. Gerritsen, A. R C , Y. M. van Heezik, and C. Sweenen (1983) Chemoreception in two further Calidris species: Calidris maritima and C. canutus; a comparison of the relative importance of chemoreception during foraging in Calidris species. Neth. J. Zool. 33: 485-496. Gill, R B. (1995) Ornithology, 2nd Ed. Rreeman, New York. Goodman, D. C , and H. I. Fisher (1962) Functional Anatomy of the Feeding Apparatus in Waterfowl (Aves: Anatidae). Southern Illinois University Press, Carbondale, IL. Gottschaldt, K. M. (1985) Structure and function of avian somatosensory receptors. In: Form and Function in Birds, Vol. 3. A. S. King and J. McClelland (eds). Academic Press, New York. Gould, S. J., and R. C. Lewontin (1979) The spandrels of San Marco
12. Feeding in Birds and the Panglossian paradigm: a critique of the adaptationist programme. Proc. R. Soc. Lond. B 205:581-598. Grant, P. R. (1985) Selection on bill characters in a population of Darwin's finches: Geospiza conirostris on Isla Genovesa, Galapagos. Evolution 39:523-532. Ho, C. Y.-K., E. M. Prager, A. C. Wilson, D. T. Osuga, and R. E. Feeney (1976) Penguin evolution: protein comparisons demonstrate phylogenetic relationships to flying aquatic birds. J. Mol. Evol. 8: 271-82. Homberger, D. G. (1986) The Lingual Apparatus of the African Grey Parrot Psittacus erithacus Linne (Aves: Psittacidae): Description and Theoretical Mechanical Analysis. Ornithological Monograph No. 39. American Ornithologists' Union, Washington. Homberger, D. G. (1988) Comparative morphology of the avian tongue. In: Acta XIX Congressus Internationalis Ornithologici, Vol. II. H. Ouellet (ed). University of Ottawa Press, Ottawa. Homberger, D. G., and R. A. Meyers (1989) Morphology of the lingual apparatus of the domestic chicken, Gallus gallus, with special attention to the structure of the fasciae. Am. J. Anat. 186:217-257. Hulscher, J. B., and B. J. Ens (1991) Somatic modifications of feeding system structures due to feeding on different foods with emphasis on changes in bill shape in Oystercatchers. Acta XX Congr. Inter. Ornith. Symposium 13:889-896. James, F. C. (1983) Environmental component of morphological differentiation in birds. Science 221:184-186. Jehl, J. R. (1986) Biology of the red-necked phalarope (Phalaropus lobatus) at the western edge of the Great Basin in fall migration. Great Basin Nat. 46:185-197. Jenkin, P. M. (1957) The filter feeding and food of flamingoes (Phoenicopteri). Phil. Trans. Roy. Soc. Lond. B 240:401-493. Kear, J. and N. Duplaix-Hall (1975) Flamingos. Poyser, Berkhamsted, UK. Kooloos, J. G. M., A. R. Kraaijeveld, G. E. J. Langenbach, and G. A. Zweers (1989) Comparative mechanics of filter feeding in Anas platyrhynchos, Anas clypeata, and Aythya fuligula (Aves, Anseriformes). Zoomorphology 108:269-290. Krajewski, C. (1989) Phylogeny and classification of birds: a study in molecular evolution. Auk 108:987-990. Krebs, J. R., and N. B. Davies (1991) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford. Lanyon, S. M. (1992) Phylogeny and classification of birds: a study in molecular evolution. Condor 94:304-307. Lauder, G. V. (1989) How are feeding systems integrated, and how have evolutionary innovations been introduced? In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds). Wiley, Chichester. Lederer, R. J. (1972) The role of avian rictal bristles. Wilson Bull. 84: 193-197. Losos, J. B., and D. B. Miles (1994) Adaptation, constraint, and the comparative method: phylogenetic issues and methods. In: Ecological Morphology, P. C. Wainwright and S. M. Reilly (eds). University of Chicago, Chicago. Lucas, A. M., and P. R. Stettenheim (1972) Avian Anatomy, Integuement. Agricultural Handbook No. 362. U. S. Government Printing Office, Washington, DC. Marchetti, K., and T. Price (1989) Difference in the foraging of juvenile and adult birds: the importance of developmental constraints. Biol. Rev. 64:51-70. McClelland, J. (1979) Digestive system. In: Form and Function in Birds, Vol. 1. A. S. King and J. McClelland (eds). Academic Press, New York. Mercier, F., and D. E. Gaskin (1985) Feeding ecology of migrating red-necked phalaropes {Phalaropus lobatus) in the Quoddy region. New Brunswick, Canada. Can. J. Zool. 63:1062-1067
407
Morony, J. J., W. J. Bock, and J. Farand. (1975) Reference List of the Birds of the World. American Museum of Natural History, New York. NeSmith, C. C. (1984) The Effect of the Physical Environment on the Development of Red-Winged Blackbird Nestlings: A Laboratory Experiment. M.S. thesis, Florida State University, Tallahassee, FL. Obst, B. S., W. M. Hamner, E. Wolanski, P. P. Hamner, M. A. Rubega, and B. Littlehales (1996) Kinematics and fluid mechanics of spinning in phalaropes. Nature 384:121. O'Hara, R. J. (1991) Phylogeny and classification of birds: a study in molecular evolution. Auk 108:990-993. Piersma, T., R. van Aelst, K. Kurk, H. Berkhoudt, and L. R. M. Maas (1998) A new pressure sensory mechanism for prey detection in birds: the use of principles of seabed dynamics? Proc. Roy. Soc. Lond. B 265:1377-1383. Prater, A. J., J. H. Marchant, and J. Vuorinen (1984) Guide to the Identification and Aging of Holarctic Waders. British Trust for Ornithology Field Guide 17. Raikow, R. (1985) Problems in Avian Classification. Current Ornith. 2:187-212. R. J. Johnson (ed). Plenum, New York. Raikow, R. (1991) Phylogeny and classification of birds: a study in molecular evolution. Auk 108:985-987. Rubega, M. A. (1996) Sexual size dimorphism in red-necked phalaropes and functional significance of the nonsexual bill structure variation for feeding performance. J. Morph. 228:45-60. Rubega, M. A. (1997) Surface tension prey transport in shorebirds: how widespread is it? Ibis 139:488-493. Rubega, M. A., and B. S. Obst (1993) Surface tension feeding in phalaropes: discovery of a novel feeding mechanism. Auk 110:169178 + frontispiece. Shaffer, H. B., and G. V. Lauder (1985a) Aquatic prey capture in ambystomatid salamanders: patterns of variation in muscle activity. J. Morphol. 183:273-326. Shaffer, H. B., and G. V Lauder (1985b) Patterns of variation in aquatic ambystomatid salamanders: kinematics of the feeding mechanism. Evolution 39:83-92. Sibley, C. G. J. E. Ahlquist, and B. L. Monroe, Jr. (1988) A classification of the living birds of the world based on DNA-DNA hybridization studies. Auk 105:409-423. Sibley, C. G., and J. E. Ahlquist (1990) Phylogeny and Classification of Birds: A Study in Molecular Evolution. Yale Univ. Press, New Haven, CT. Spring, L. W. (1965) Climbing and pecking adaptations in some North American woodpeckers. Condor 67:457-488. Stephens, D. W., and J. R. Krebs (1986) Foraging Theory. Princeton Univ. Press, Princeton, NJ. Stresemann, E. (1975) Ornithology: From Aristotle to the Present. Harvard Univ. Press, Cambridge. Sutherland, W. J. (1996) From Individual Behaviour to Population Ecology. Oxford Univ. Press, Oxford. Tinbergen, N. (1935) Field observations of east Greenland birds. I. The behavior of the red-necked phalarope (Phalaropus lobatus, L.) in spring. Ardea 26:1-42. Travis, J. (1994) Evaluating the adaptive role of morphological plasticity. In: Ecological Morphology. P. C. Wainwright and S. M. Reilly (eds). University of Chicago Press, Chicago. University of California. (1985) Phalarope Feeding Behavior (film). From the film series Aspects of Animal Behavior. Office of Instructional Development, University of California, Los Angeles. Van den Heuvel, W. R (1992) Kinetics of the skull in the chicken (Callus gallus domesticus). Neth. J. Zool. 42:561-582. Vanden Berge, J. C , and G. A. Zweers (1993) Myology. In: Handbook of Avian Anatomy. J. J. Baumel (ed). Nuttall Ornithological Club, Cambridge.
408
Margaret Rubega
White, S. S. (1968) Mechanisms involved in deglutition in Gallus domesticus. J. Anat. 104:177. Wunderle, J. M., Jr. (1991) Age-specific foraging proficiency in birds. In: Current Ornithology, Vol. 8. D. M. Power (ed). Plenum Press, New York. Ziegler, H. P., P W. Levitt, and R. Levine (1980) Eating in the pigeon {Columba livia): movement patterns, stereotypy and stimulus control. J. Comp. Physiol. Psychol. 94:783-794. Zusi, R. L. (1984) A functional and evolutionary analysis of rhynchokinesis in birds. Smith. Contrib. Zool. 395:1-40. Zweers, G. A. (1982a) Pecking of the pigeon {Columba livia L.). Behaviour 81:173-230. Zweers, G. A. (1982b) The feeding system of the pigeon (Columba livia L.) Adv. Anat. Embryol. Cell Biol. 73:VII+108. Zweers, G. A. (1985) Generalism and specialism in the avian mouth and pharynx. Fortschr. Zool. 30:189-201. Zweers, G. A. (1991a) Transformation of avian feeding mechanisms: a deductive approach. Acta Biotheor. 39:15-36.
Zweers, G. A. (1991b) Pathways and space for evolution of feeding mechanisms in birds. In: The Unity of Evolutionary Biology, E. C. Dudley (ed). Dioscorides Press, Portland. Zweers, G. A., H. Berkhoudt, and J. C. Vanden Berge (1994) Behavioral mechanisms of avian feeding. Pp. 241-279, In: Advances in Comparative and Environmental Physiology, Vol. 18. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Zweers, G. A., F. de Jong, H. Berkhoudt, and J. C. Vanden Berge (1995) Filter feeding in flamingos (Phoenicopterus ruber). Condor 97:297-324. Zweers, G. A., and A. F. C. Gerritsen (1997) Transitions from pecking to probing mechanisms in waders. Neth. J. Zool. 47:161-208. Zweers, G. A., A. F. C. Gerritsen, and P. J. van Kranenburg-Vood (1977) Mechanics of feeding of the mallard (Anas platyrhynchos L.; Aves, Anseriformes). Contrib. Vert. Evol, Vol. 3. Karger, Basel. Zweers, G. A., and J. C. Vanden Berge (1997) Evolutionary transitions in the trophic system of the wader-waterfowl complex. Neth. J. Zool. 47:255-287.
S E C T I O N
VI
This Page Intentionally Left Blank
C H A P T E R
13 Feeding in Mammals KAREN M. HIIEMAE Department of Bioengineering and Neuroscience Institute for Sensory Research Syracuse University Syracuse, New York 13244
tion can be associated with the availability of habitats previously occupied by the highly successful dinosaur radiation whose species then occupied most, if not all, the ecological niches now occupied by mammals. Survival to reproductive age, followed by successful reproduction, requires an adequate food intake: in this context "adequate food intake" is defined as that which yields more biochemical energy than required for its collection, ingestion, and digestion. While that statement holds true for all vertebrates, warm-blooded mammals and birds require a higher and a sustained level of energy to maintain homeostatic mechanisms. This positive energy balance is needed to support tissue turnover and the maintenance of all body systems, as well as intrauterine fetal development followed by lactation (a uniquely mammalian reproductive mechanism). For most mammals in nontropical climates, there is also a need for the acquisition of metabolic reserves to be drawn upon in seasons where the available food supply diminishes and drives the metabolic equation into negative (e.g., winter in high latitudes or drought in all arid regions). It follows that not only must food be available to meet physiological demand, but it must be accessible and processable. Some mammals make food caches, others brown fat, and still others migrate in search of food and water. For survival:
I. INTRODUCTION 11. MAMMALIAN FEEDING SYSTEM A. Overview B. Approaches to the Study of Feeding in Mammals III. THE "PROCESS MODEL" FOR MAMMALIAN FEEDING IV. MECHANICAL PROPERTIES AND TEXTURES OF FOODS V. THE FEEDING APPARATUS A. Jaw Complex B. Oropharyngeal Complex VI. FEEDING FUNCTION A. Tongue-Jaw Linkages B. Food Manipulation and Movement VII. CONTROL OF FEEDING BEHAVIORS References
I. INTRODUCTION About 60 million years ago, an explosive evolutionary radiation of mammals took place (Romer, 1974). Although dental and cranial evidence shows that the earliest mammals had appeared some 120 million years ago, the extraordinary proliferation of mammalian genera occurring during and after dinosaur extinc-
metabolic metabolic cost metabolic cost reserves reserves yield of = of food + of body + for a n d / o r for seasonal food intake acquisition maintenance reproduction food shortage
FEEDING (K.Schwenk,ed.)
411
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
412
K a r e n M. H i i e m a e
During the evolution of mammals, the reptilemammal transition was characterized by the unique development of several fundamental and linked structural innovations. First, the dentary expanded to become the single lower jaw bone, articulating with the squamosal to form a new jaw joint. The quadrate and articular bones of the reptilian jaw joint were coopted to form the ossicular chain of the mammalian middle ear (see Crompton, 1995). Second, the palatal processes of the premaxilla and maxilla fused in the midline, forming a true hard palate completely separating the oral and nasal cavities. A new structure, the muscular soft palate, attached to the posterior margin of the hard palate, appeared as a mobile flap separating the airway from the oral cavity (see Smith, 1992). Third, the appearance of a highly differentiated tooth row restricted to, at most, a deciduous and a permanent dentition developed for both the acquisition and the processing of food (mechanical digestion). There were also associated fundamental changes in oropharyngeal soft tissues. The fiber directions of the adductor muscles of the jaw provided for mandibular movement in the anteroposterior and mediolateral directions, as well as the vertical. A functionally integrated intrinsic and extrinsic tongue musculature allowed for complex tongue movements, including differential expansion and contraction. A completely new neuromuscular complex of longitudinal and circular muscles forming the pharynx provided a new method of swallowing (see Smith, 1992). This system allowed for the more or less continuous movement of tidal air (respiration) as well as the intermittent transmission of swallowable material into the gastrointestinal tract. Although the taxonomy and putative monophyly of most mammalian orders is well established and reasonably stable, higher-level phylogenetic relationships of mammals (i.e., among orders) remain contentious (for reviews, see Novacek, 1992; Honeycutt and Adkins, 1993; also Meng et al., 1994; Springer ei al., 1997). Difficulties in resolving higher-level relationships might relate to the explosive nature of mammalian adaptive radiation during the Cretaceous (but see Hedges ei al., 1996). Such extreme rapidity of cladogenesis and phenotypic evolution may have led to extensive homoplasy in both molecular and morphological characters, thus confounding cladistic character analysis. Figure 13.1 illustrates a generally accepted phylogeny of mammals based on Novacek (1992), but it must be acknowledged that this cladogram is neither wholly agreed upon now nor likely to persist unchanged for long. Possibly as a consequence of uncertainty in the higher-level relationships of mammals, virtually no study has attempted an overarching, phylogenetic
MONOTREMATA MARSUPIALIA PHOLIDOTA XENARTHRA CARNIVORA INSECTIVORA MACROSCELIDEA LAGOMORPHA RODENTIA PRIMATES SCANDENTIA DERMOPTERA CHIROPTERA TUBULIDENTATA ARTIODACTYLA CETACEA PERISSODACTYLA HYRACOIDEA SIRENIA PROBOSCIDEA
F I G U R E 13.1. Phylogenetic relationships among mammalian orders based on Novacek (1992). See text for discussion.
analysis of feeding system evolution in mammals. Rather, there is a general dogma that most eutherian orders arose from a generalized, insectivorous ancestor with subsequent divergence and specialization. Thus, comparative approaches to mammal feeding typically are typological in the sense that feeding systems are characterized order by order (e.g., TurnbuU, 1970) with little attention paid to evolutionary transformations among systems. With a phylogenetic approach it should now be possible to reconstruct aspects of the feeding system at ancestral nodes and to examine patterns of character evolution. Unfortunately, such an analysis is beyond the scope of this chapter. Rather, this chapter establishes the fundamentals of mammalian feeding and reviews much of the known diversity in feeding systems, with emphasis on those relatively few taxa for which significant functional data are
13. Feeding in Mammals
available. Chapters 15 and 16 provide detailed coverage of specialized myrmecophagous and marine feeding systems, respectively. It is a telling fact that some mammal orders are named for dietary habit, e.g., Carnivora and Insectivora, indicating both the importance of feeding system characters in mammal taxonomy and the presumed stability of feeding system phenotype within (in contrast to among) orders. However, it is important to note that the actual diets of species in a given order range across the available food source spectrum (Table 13.1 and Fig. 13.2). Despite various modifications to the basic ordinal Bauplan in response to the mechanical demands of food collection or processing, members of each order retain the fundamental characters diagnostic of its group. For example, Ailuropoda (the giant panda) retains features identifying its carnivoran (ursid) origins, despite its highly specialized diet of bamboo. In short, phylogeny dictates the overall musculoskeletal anatomy of the orofacial complex, but the details can be very specific to genera, even species, sub-
TABLE 13.1 An Overview of the Range of Food Sources Utilized by Members of the Major Orders of Terrestrial Mammals'' Order
Diet
Marsupialia
Insectivores, omnivores, carnivores, herbivores
Insectivora
Insects, small vertebrates, blood, pollen, nectar, fruit (extreme specialization: anteaters^)
Chiroptera
Insects, blood, fruit, honey
Primates
Insects, fruit, leaves, nuts, small mammals (termites) (extreme generalization: H. sapiens)
Carnivora
Insects, Crustacea, fish, small/large animals, including carrion, fruit, honey/nectar, leaves [extreme variant: baleen whales (Spermaceti)]
Perissodactyla and Artiodactyla
Some omnivores; insects, roots, bulbs, fruit, buds, shoots, young and old leaves, grasses, aquatic vegetation (extreme variants: elephants,^ dugong, manatee'')
Hystricomorpha and Lagomorpha
Grasses, leaves, nuts, fruits
Rodentia
Insects, small animals, fish, nuts, leaves, grasses, seeds, bark, fruit
Edentata
Sloths: leaves Anteaters: termites (see Chapter 15)
^This list is not intended to be exhaustive, but rather indicative of the dietetic opportunism within groups. ^Indicates members of separate "specialized" orders derived from the primary order cited.
413
sequently adapted for the acquisition and digestion of specific foods. The American opossum, Didelphis marsupialis, has been used as the extant exemplar of the ancestral mammalian condition (Turnbull, 1970; Hiiemae and Crompton, 1985; earlier references cited therein). Whether or not modern Didelphis has itself evolved (other than in its opportunistic ability to take advantage of all available food sources, i.e., its northward spread into suburbia raiding garbage cans), it remains true that the oropharyngeal behavioral mechanisms of Didelphis provide a valuable baseline for other mammalian forms. In his "heroic" (Herring, 1993) monograph. Mammalian Masticatory Apparatus, Turnbull (1970) attributed the anatomy of the jaws, teeth, and the major adductor jaw muscles to four dietetic groups: (1) generalized group, exemplified by Didelphis and Echinosorex, including the primates; (2) specialized group I (carnivores), typified by the domestic cat {Telis domesticus); (3) specialized group II (ungulates), typified by the horse {Equus caballus), a deer {Odocoileus virginianus), and a sheep {Ovis aries); (4) specialized group III ("rodent/gnawing mammals"), exemplified by a squirrel {Sciurus niger), a rat (Rattus norvegicus), and a porcupine (Hystrix). Turnbull was also forced to recognize a miscellaneous group of "oddball" mammals that did not fit into his major categories. His primary focus was on the jaw musculature and its bony attachments, not on the dentition, still less on the actual mechanics of the feeding process. Nor was Turnbull much concerned about the evolutionary history of the mammals he used as exemplars, as noted earlier. However, he made an invaluable contribution, in part because he included primary source references for mammals other than those he used as exemplars for his groups. Since the 1950s, there has been an important shift in the focus of studies of the mammalian feeding apparatus from morphology (shape and structure) to mechanism (behavior and biomechanics), especially of the teeth and jaws (see reviews by Hiiemae, 1978; Hiiemae and Crompton, 1985; Herring, 1993; Weijs, 1994). Nevertheless, Turnbull's classification of dietary types is used here as the scaffold on which to base this review. Mammals are so diverse and their feeding mechanisms so varied that this chapter is designed to discuss (1) the processes involved in food acquisition and mechanical digestion; (2) the biomechanics of that process in the context of an "archetypal" primitive mammal; the American opossum (Didelphis virginiana); (3) the hroad brush variations on the mammalian Bauplan associated with broad dietetic categories; and (4) what is now known about the linkages between jaw and tongue functions in feeding. Primary source references
414
K a r e n M. H i i e m a e
OMNIVORES
ANIMAL = high protein
VEGETABLE = variable protein and lignin/celiulose
i
Mature leaves,^ Grasses Buds; S h o o t s /
Marine mammals/,^ large fish V ^
^^
Seeds, Nuts
Fish ^
Aquatic herbivores
Small mammals, ,,SbiKlS.
Moliusca Krill
^^ASW N9
Insects
Nectar
Eggs
BODY SIZE
Pollen
Generalized Mammal e.g. Opossum
BODY SIZE
F I G U R E 13.2. Schematic diagram showing broad categories of food sources available to mammals. Although there is some correlation betw^een food type and body size, this is in no w^ay absolute. The largest w^hales (Spermaceti) consume the smallest of Crustacea (krill). There is also no precise correlation between phylogenetic history and dietary habit (see Table 13.1).
are cited, but emphasis has been given to recent review articles to facilitate access to the relevant literature for those wishing to pursue a given area in depth. Substantial emphasis is placed on "process/' as recent experimental work has shown that jaw and tongue movements are linked and that this linkage has a far greater effect on jaw movement patterns than generally recognized to date (Hiiemae and Palmer, 2000). Given societal imperatives. Homo sapiens has become essentially a "species apart," a focus for anthropologists (human evolution), social scientists (behavior), and clinicians (disease). Although the only true facultative biped, the human is, otherwise, only "peculiar" in having evolved adaptations in the orofacial complex to accommodate a fundamental change in dominant oropharyngeal function from feeding to speech. Those changes are worthy of some attention. IL M A M M A L I A N FEEDING SYSTEM A. Overview The mammalian feeding system, like that of other vertebrates, can best be characterized as a tube or luminal space of variable dimensions, lined by epithelia with different properties, into which the products of exocrine glands are poured in the presence of food. In mammals, this tube, whose lumen is, sensu strictu, extracorporeal, subserves several functions: its anterior or rostral end is the oral cavity, where food ingestion and, almost uniquely among tetrapods, food break-
down occurs as a function of patterned jaw movements and tooth-food-tooth interaction (physiologically, "mechanical digestion"); immediately distal is the oropharynx, a lumen used both for food transport to the gastrointestinal tract and for movement of tidal air in respiration; the esophagus is a simple transthoracic/ transdiaphragmatic tube leading to the stomach, in turn to the intestines, all of which are involved in chemical digestion, in some cases with the assistance of symbiotic bacteria, before fecal formation in the colon for emission through the rectum and anus. Importantly, this tube is interrupted by a series of sphincters, dividing it into separable sections. These sphincters control the rate of passage of material along its length. This tube is supported directly by skeletal elements (oral cavity and oropharynx) or indirectly by suspension (peritoneum) from the posterior abdominal wall. Its included volumes are altered by the actions of the jaw and tongue muscles (oral cavity); by the striated longitudinal and circular striated pharyngeal muscles (pharynx); and then, in the lower esophagus and gastrointestinal (GI) tract, by the circular and longitudinal muscles of the gastric and intestinal walls. It follows that the upper part of the system is under "voluntary control," whereas motility in the remainder is governed by the electrotonic coupling of smooth muscle cells and is controlled by their intrinsic neural networks associated with the autonomic nervous system. This feeding system subserves four vital physiological functions: (1) the ingestion of nutrients (liquids
415
13. Feeding in Mammals and solids), (2) the preparation of solids for transmission across the oropharynx to the esophagus, (3) the breakdown of materials such that their included nutrients can be absorbed across the intestinal epithelium, and (4) the elimination of waste products, including indigestible materials. While for most mammals, the system serves as a "one-way channel," this is not always the case. For example, in the ruminant herbivores, material that has been transported to the stomach is regurgitated for further mechanical digestion. Conversely, some mammals ingest foods that are not processed in the oral cavity and are made nutritionally useful only when they reach the GI tract (e.g., ants, krill, see Chapters 15 and 16). The biochemistry of digestion and the not so subtle variations in gastrointestinal anatomy and physiology associated with particular diets cannot be addressed here. Similarly, adaptations in metabolism, indicated by food intake patterns to form reserves of brown fat, e.g., dictated by overwintering strategies such as hibernation, are far beyond the scope of this chapter. Important elements in the investigation of the interactions among the ecology, anatomy, and physiology of mammals are body size, metabolic rate, and the "metabolic yield" of preferred foods. These have to be factored into the interpretation of behavior, and the underlying physiological imperatives, especially when modeling causation for evolutionary change. Conversely, sympatric species may be able to coexist simply because each has developed subtle specializations in dietary preference within the same, even very specialized, food source habitat. As Birt et ah (1997) pointed out for Australian megachiropteran bats, this degree of niche specialization may guarantee the extinction of species as habitats are isolated and lost as a result of environmental degradation coupled with the deraands of expanding human populations, especially when the animals are competitors with farmers for a fruit crop (e.g., a BBC World News Service 12/19/98 report about the cherry crop and a cull of fruit bats in northern Australia). Ironically, the success of H. sapiens, now a competitor for resources with many tetrapod species, is itself a product of an extreme specialization of the oropharyngeal complex coupled with expansion of the central nervous system (CNS) for a new function, speech. Nevertheless, when considering the history of mammalian taxa and assessing the environmental factors that may have produced gradual or stochastic change in craniofacial anatomy, interpretations of climatic change may be important. For example, evidence shows that preferred diets changed between wet and dry seasons in some fossil east African mammals. The sophisticated techniques now available for the deter-
mination of the botanical environment in fossil sites are helping in such evaluations. That said, traditional methods, such as comparative anatomy and, importantly, comparative physiology, cannot be relegated to the realm of "old hat." DNA analysis, however seductive a means of attempting to establish not only relationships among taxa, but the temporal pattern of their divergence, is still a very inexact science. We do not have a "genetic blueprint" for feeding mechanisms. We almost certainly never will. Given the complexity of "the system, reliance will have to continue to be placed on careful experimentation and observation (see Chapter 1 for further discussion). B. Approaches to the Study of Feeding in Mammals Differences in the shape of mammalian skulls, including jaws and their associated teeth, have intrigued naturalists since before written history (cf. Lascaux and comparable cave art). Given that such observations were rarely formally codified until the creation of the scientific journal and the scientific meeting (initiated by the Royal Society and the Academie Frangaise in the 17th century), it is difficult to establish the genesis of preceding observations that are now taken as "given." However (and importantly), major studies of mammalian comparative anatomy were published beginning in the late 1700s, continuing through the 1800s and to the present. We continue to rely on them, e.g., Magendie (1825) on swallowing, Dobson (1882) on the digastric muscle, and Edgeworth (1911, 1914) on cranial and hyoid muscles. Comparative studies of the tongue are more recent (e.g., Doran and Baggett, 1971; Doran, 1975). TurnbuU (1970) includes an invaluable bibliography of the source literature for jaw muscle anatomy and associated feeding mechanisms and Smith (1992) for the oropharyngeal soft tissues. If one wants to know how a system functions, then the obvious approach is to study it in action. This approach was used in the last century, but without the technology to create an analyzable permanent record of events. Mammalian cyclical feeding behaviors occur too fast for accurate visual recording, so no quantifiable measures of events could be obtained. The first calibrated records of jaw movements were made in either 1899 or 1908 with a human subject (Lord, 1913). Recently (post-1960), two parallel trains of investigation into feeding mechanisms in mammals have been ongoing (see Herring, 1993): (1) experimental and theoretical studies of selected species or groups and (2) studies seeking to explore morphological changes during evolution. Both have adopted, in the last two to three decades, very sophisticated approaches, but are
416
Karen M. Hiiemae
always dictated by technological developments applicable to the problem. It is proper to say that the "silicon revolution" has allowed major changes in the methods available for (1 and 2), but that comparative and evolutionary anatomists (2) are still constrained by the specimens available to them. Legislative efforts, such as the U.S. Endangered Species Act and the international conventions on trade in endangered species, as well as the vocal concerns of those objecting to experiments on animals, have restricted what is now possible in the laboratory. Offsetting these limitations, the extraordinary expansion of public interest in "natural history" has led organizations such as the National Geographic Society to sponsor major expeditions to a wide range of environments, which have yielded a treasure trove of film and tape showing feeding behavior in natural habitats for many mammals, which can greatly illuminate the evidence to be gleaned from bones, teeth, and soft tissues. Specific examples of the experimental methods changing functional analysis are (1) the development of the fine wire in-dwelling electrode (Basmajian and Stecko, 1957); (2) the use of strain gauges to measure forces on the lower jaw and skull (e.g., Hylander, 1977, 1984); (3) the development of the scanning electron microscope (SEM), allowing high magnification studies of tooth surfaces (e.g., Rensberger, 1978; von Koenigswald, 1982; Teaford and Runestad, 1992); (4) the advent of small powerful computers capable of manipulating complex data sets and able to support modeling software for techniques such as motion analysis and finite element analysis (e.g., De Jongh et ah, 1989; Hart et al, 1992; Spears and Macho, 1998); and (5) refinements in scintillation technology, coupled with video techniques (particularly S-VHS), which have improved not only the image intensifier (the fineresolution screen required for recording movement events using radiography) but which have also allowed a drastic reduction in radiation exposure making videofluorography (VFG) applicable to research on normal humans within the constraints dictated by federal regulations for the involvement of human subjects. It remains true that the use of X-ray techniques pioneered by Ardran and Kemp (1958) provide the primary data base for the interpretation of intraoral behaviors and mechanisms in mammals. Similar methodology has been used to establish the mechanisms of food transit through the GI tract. Mammal species studied to date using X-ray include rabbit (Ardran and Kemp, 1958; Weijs and Dantuma, 1981, Anapol, 1988; Cortopassi and Muhl, 1990), rats (Hiiemae and Ardran, 1968; Weijs and Dantuma, 1975), opossum (Hiiemae and Crompton, 1971; Hiiemae et al, 1978), cat (Hiiemae et al, 1981; Thexton et al, 1982; Thexton and Mc-
Garrick, 1988, 1989), pig (Herring and Scapino, 1973), hyrax (Janis, 1979; Franks et al, 1985; German and Franks, 1991), tenrec (Oron and Crompton, 1985), goat (de Vree and Gans, 1976), and bats (Kallen and Gans, 1972; de Guelde and de Vree, 1984). In the period between World Wars I and II, A. V. Hill, as well as Sherrington and his student Mountcastle, made seminal discoveries about the behavior of muscle and the central nervous system, respectively. Bremer (1923) was the first to demonstrate that rhythmic behaviors in feeding, such as chewing, were controlled by a "centre de correlation," now called a central pattern generator (CFG). The existence of such a CFG was confirmed in an elegant study by Dellow and Lund (1971). Since then, Jean (1984) and others have developed the concept of a "swallowing center." An as yet important unresolved question is the issue of how these two centers, both located in the pontinemedullary region of the hindbrain, are connected to produce the smooth integration of food processing and swallowing. Given the comparatively recent discovery (Hiiemae et al, 1978; Hiiemae and Falmer, 2000) that rhythmic tongue movements are linked to those of the jaw, but can occur independently of jaw movement as in suckling (Chapter 14) and at some stages in feeding (Hiiemae et al, 1996), substantial questions as to the neural control of feeding remain. Given the "process model" (see Section III), the stages in food processing, i.e.. Ingestion, Stage I Transport, Frocessing (Reduction), Stage II Transport, and Swallowing, have to be addressed sequentially. Our current knowledge base is uneven. For example, far more is known about the biomechanics of food processing than of ingestion or bolus formation and deglutition. Because the processes of ingestion, reduction, and swallowing largely involve the same musculoskeletal elements, a "generalized mammalian model" is proposed; significant variants associated with dietetic specialization (insectivory, frugivory, carnivory, herbivory) are addressed in each of the following sections.
III. THE "PROCESS MODEL" FOR M A M M A L I A N FEEDING The process model is shown in Figure 13.3. Originating in cinefluorographic studies of the rat (Hiiemae, 1967), but fully developed from data on opossum and cat (see Hiiemae and Crompton, 1985; Hiiemae et al, 1978), it has been tested in other mammals, including the macaques (Hiiemae et al, 1995) and humans (Hiiemae and Palmer, 1999), although H. sapiens displays one important difference (see later). Descriptions of
417
13. F e e d i n g in M a m m a l s
INGESTION Food moved into front of mouth
TRANSPORT ?
EJECTION
•NO-
YES
n
-i2
ii-
STAGE I TRANSPORT I to postcanine area ^^^^A^XM
PROCESSING Chew or tonguepalate compression
'by mastication'
TRANSPORT ?
¥
NO
YXAX/'X/yA/yyj
YES
1
STAGE II TRANSPORT through fauces for bolus formation
^
'
r^
I
THRESHOLD?
H. Sapiens Liquids and semi-solids only
•f
SWALLOW
Bolus formation continues
»<><x/x;ovxxx^ NO
J vxxt^
Liquids, soft semi-solids
5?5^
Solids
F I G U R E 13.3. The process modeL First developed to explain feeding in the rat (Hiiemae, 1967), this model has been revised over the years as more data have become available. It now includes what is known of dual bolus formation and the swallowing mechanism in humans (see text). The distinction between mechanical elements of the model and postulated sensorimotor ''gates'' is indicated by shaded boxes.
the feeding process in other mammals indicate that the process model applies to a wide variety of species. The model is predicated on the fact that (a) most mammals ingest both liquids and solids, but (b) solids are typically not swallowable as ingested, i.e., are reduced by some form of processing. The model embodies two major hypotheses, one mechanical, the other sensorimotor. The mechanical hypothesis argues that ingested material is transported through the oral cavity to the
pharynx and then the GI tract, using two transport mechanisms (which may not be identical): stage I transport, in which food is moved from the incisal area to the postcanine region, followed by stage II transport, in which swallowable food is passed through the fauces for bolus formation and deglutition. The sensorimotor hypothesis postulates a series of sensorimotor "gates" that regulate the process. It argues that gate 1 provides for an evaluation of the nature of the ingested
418
Karen M. Hiiemae
material (i.e., "palatability")- Noxious or otherwise unacceptable material can be ejected ("spat out" in common parlance). Gate 2 postulates a sensorimotor mechanism for the determination of the "swallowability" of food in the postcanine area. If the material is not suitable for bolus formation, then the mechanisms for processing are triggered. Gate 3 postulates a "threshold" below which a swallow does not occur. This is the most problematical feature of the model. While based on data for Didelphis and cat, there is no corroborating evidence available for other nonhuman mammals. The substantial literature on swallowing behavior in humans does not appropriately address the problem as the paradigm on which the human experiments were designed is based on liquid swallows rather than swallows of processed natural bites (for an exhaustive review of swallowing in mammals, see Thexton and Crompton, 1998). The model accommodates the natural variations in initial food consistency and feeding behavior among mammals. Clearly, liquids are swallowable "as is": giraffes at a watering hole are lapping up water and swallowing in a definite rhythm. Water and maternal milk (see Chapter 14) clearly meet the "swallowable standard" (gate 2). In fact, the processes of stage I and stage II transport become a continuum. In the more artificial laboratory environment, semisolid foods such as tinned cat food mixed to varying consistencies may be treated in much the same way (Thexton and McGarrick, 1988). In these circumstances, the ratio between "lap" (ingestion of an aliquot of fluid) and swallow may change (Hiiemae et al, 1978; see also Thexton and Crompton, 1998). However, at some consistency threshold, this transport mechanism is interrupted because the ingested material cannot be formed into a swallowable bolus. With the exception of those mammals that swallow their prey whole (e.g., insects, eggs, or whole fish by dolphins), this triggers processing, defined as the reduction of the food material to a consistency "acceptable" for swallowing. The model makes no attempt to define "acceptable consistency." Clearly, this must differ among taxa and types of food. For example, it is alleged that felid carnivores can swallow fairly large lumps of fresh prey meat, whereas it is well known that herbivorous grazers spend considerable time chewing the foliage they have ingested. However, at some point a proportion, if not all, of the ingested material can be swallowed. In rendering it swallowable, the effect of the mixing of the food particles with saliva cannot be ignored. Saliva has a "wetting function." Indeed Prinz and Lucas (1997) argue that for a bolus (in humans) to be swallowed, certain rheological criteria must be met. Those criteria involve a combination of food particle sizes and salivary wetting with
their combined surface tension and packing effects. Their model is compelling. In this context, variations in the pattern can be noted. De Gueldre and De Vree (1984) reported that the frugivorous, megachiropteran bat, Pteropus, swallows only the juice extracted from the ingested fruit, spitting out pellets of cellular residue. Many fruits have hard shells or peel. Incisors are used to break open the fruit, accessing the pulp. During this process, pieces of peel or shell may enter the anterior mouth only to be ejected (primates, rodents). When fed 1-cm^ lumps of hard liver, cats cut them into two lumps, ejected one and then processed the remainder (Thexton et al., 1980), suggesting that some acceptable food may be rejected because it is initially too large for intraoral management. Once food in the mouth is "acceptable for swallowing," stage II transport (i.e., passage through the fauces to the oropharynx) occurs. Bolus formation ensues and a swallow is initiated. Although experimental evidence shows a rhythm for ingestion to deglutition for liquids and soft semisolids (see Thexton and Crompton, 1998), the same data (Thexton et al., 1980) also show that there is essentially a one-to-one relationship between ingested "bite" and bolus formation. If true, and this is by no means clear for the broad range of mammals (the question was not asked in most studies), a given volume of solid food is ingested and processed until swallowed. While true for solid foods in opossum, cat, and possibly hyrax, it is not the case for the macaque and modern man. We have almost no data on other mammals. Does a cow swallow some portion of ingested grass and retain the inadequately triturated remainder in the oral cavity as do pigs or does it take a mouthful of silage and chew it all until swallowable? Why should this be an issue? The concern directly relates to our understanding of the neural control of the process. This must depend on sensory input to the central (hindbrain) effector motor systems. The nature of this input is currently disputed. The functional problem hinges on the issue of "segregation"—how can adequately triturated material be segregated from that which is inadequately reduced? Work in progress (human studies, see Hiiemae and Palmer, 1999, 2001) suggests a surprising answer, but one that may not be applicable to other mammals, as a single feeding sequence in humans may involve multiple swallows. The model posits that stage II transport moves food to the site of bolus formation. This aspect of the model is pivotal. Bolus formation does not occur in the oral cavity in most mammals studied. It occurs in the oropharynx around a larynx whose aditus (entrance or approach) is in the nasopharynx (see Thexton and Crompton, 1998). Triturated food is moved through
419
13. Feeding in Mammals the fauces (Fig. 13.13) and accumulates in the valleculae and piriform fossae. Swallowing (deglutition) occurs as a powerful ejection of the material from this perilaryngeal space into the esophagus. Unlike other mammals, the human larynx is positioned well below the soft palate in adults. The oropharynx is, therefore, much longer, and the larynx (and respiratory tract) is not shielded from the aspiration of food, except by the sphincter afforded by the epiglottis and vocal folds. Nevertheless, it has been determined (Palmer ei al., 1997; Hiiemae and Palmer, 1999) that bolus formation in humans, when fed solid foods, typically occurs in the oropharynx. This is the case regardless of whether subjects are in the upright (normal human position) or on "all fours" mimicking the feeding position of most mammals (Palmer, 1998). Beyond its implications for the management of patients with swallowing disorders, this finding is important because it suggests that H. sapiens retains a basic mammalian bolus formation mechanism, at least when solid foods in normal bite sizes are ingested. However, it is indisputable that liquid boli are formed in the oral cavity and swallowed by expulsion through the fauces. We therefore postulate that liquid bolus formation in the mouth is a uniquely human specialization (Palmer ei al., 1997; Fiiiemae and Palmer, 1999).
IV. MECHANICAL PROPERTIES A N D TEXTURES OF F O O D S The classification of mammals depends in large part, although not exclusively (e.g., differences be-
tween Perissodactyla and Artiodactyla) on the anatomy of the jaws and teeth. This practice follows from the fact that tooth and jaw form in mammals are highly sensitive indicators of diet, in a general sense (e.g., herbivory, carnivory, insectivory), and that diets are relatively stable at higher taxonomic levels (e.g., families and orders in some cases; see earlier discussion). Although food scientists have developed a large body of information on the rheology, mechanical properties, and texture perception of foods, this work is focused on natural and engineered components of the human diet. Lucas, with his colleagues and students, has contributed the most significant body of work on the properties of normal mammalian dietary items, with a particular emphasis on those consumed by cercopithecid primates [see Lucas and Corlett (1991) for an overview; Sibbing (1991), in the same volume, describes food processing by the pharyngeal jaws of cyprinid fish, which show analogous adaptations in tooth form related to diet]. Nonetheless, it remains true that, to date, we have a very poor understanding of the relationship between the mechanical properties of foods and tooth form. Figure 13.2 presents a broad-brush overview of food sources available to mammals. Table 13.2 summarizes the mechanical properties of common mammalian food items, coupled with the basic mechanical requirements for their reduction (processing). For food to be "reduced," its structure has to be disrupted. The most "effective" (energy conserving) method depends on the intrinsic mechanical structure of the particular food item. While the terms used in Table 13.2 have specific meanings for mechanical engineers and food
TABLE 13.2 The Mechanical Properties of Typical Mammalian Food Items/ the "Engineering" Required for Reduction (Appropriate Technique), and Its Morphological Translation into Tooth Form^ Appropriate ToughNotch sensitivity reduction ness
Tooth morphology
Found in
Type
Examples
Hard brittle
Seeds, nuts, unripe fruit, some tubers, roots, some adult insects
Stiff elastic
High
Low
High
Crushing, splitting to^ grinding
Mortar-pestle
Primitive mammals, insectivores, primates, omnivorous herbivores
Turgid
Ripe juicy fruits, some insect larvae
Plastic flow
Low
Variable
Variable
Crushing to^ grinding
Mortar-pestle
Primitive mammals, insectivores, primates, omnivorous herbivores
Soft tough
Animal soft tissues, some insects, young leaves (?), young grasses (?)
Viscoelastic/ pliant
Moderate
High
Low
Cutting, shearing, piercing
Blades
Carnivores (especially carnassials), premolars in insectivores, primitive mammals
Tough fibrous
Grass, fruit skin
Viscoelastic/ pliant
Variable, High depends on fiber direction
Low
Serial arrays of Cutting, lacerating, low-profile shearing blades
Deformability
"See text for explanation. ^After Lucas and Luke (1984) and Sibbing (1991).
strength
All facultative herbivores to varying degrees; extreme development in some groups
420
K a r e n M. H i i e m a e
scientists (see Jeronimidis, 1991; Purslow, 1991; Vincent, 1991; Rensberger, 1995), they can be briefly defined as follows: hardness (probably stiffness) is a measure of deformability (elastic or plastic) and is expressed as stress per unit of strain (i.e.. Young's modulus, E). These materials fail under load after a period in which they are loaded beyond recovery. Depending on the material, the required load may be relatively small or very large. A large load can be delivered most effectively by a sn\all, sharp tooth cusp moved by powerful adductor contraction, penetrating the material held between upper and lower teeth. (I once recorded a force of 17 kg in Didelphis produced at a premolar cusp tip over an area of about 1 mm^.) Brittle materials are readily cracked. For example, in eating a nut, compressive force is applied and the nut shatters. Hard and brittle materials break when a crack is created in the material by the application of external force and the crack is propagated throughout the material, resulting in breakage. A different problem is created by materials that are ductile, i.e., can flow or be deformed beyond their elastic limit and so resist crack propagation (viscoelastic materials). This makes such foods tough, i.e., more work is needed to generate fracture. Notch sensitivity refers to the likelihood of penetration leading to fracture. As Purslow (1991) shows, the fracture properties of meat are highly dependent on the orientation of the muscle fibers when subjected to occlusal forces. The external "working surfaces" of mammalian teeth, with a few exceptions (e.g., sloths, manatees), are covered (at least initially) with enamel, the hardest known biological tissue. Mammalian enamel is prismatic. As such, the enamel is organized microscopically into columns, which means that the orientation of the prisms relative to the direction of loading affects the strength of the enamel, and thus its resistance to fracture (see Rensberger, 1978,1995; von Keonigswald, 1982). Prisms are oriented perpendicular to the typical direction of loading. While enamel is highly resistant to wear, it is abraded in normal use. Thickness of crown enamel varies within groups and has been used as an indicator of likely wear stress in primates (Kay, 1975; Kay and Hylander, 1978; Spears and Macho, 1998) and ungulates (Janis and Fortelius, 1988). In some groups, the teeth reach full functionality once dentine is exposed (Fig. 13.4). It is alleged that mortality in wild mammals, such as shrews, is closely correlated with dental wear—once the teeth have lost their shearing surfaces, the animals can no longer process sufficient material to keep the "metabolic equation" in balance. It is also said that tigers in India become "man-eaters" when their teeth are so degraded that they are unable to bring down normal prey. Inspection of Anasazi skulls in southwestern U.S. mu-
1^ •m-wA
MORTAR-PESTLE
SHEARING BLADES
LTLJlA/
rmful
SERIAL ARRAYS -LOW PROFILE BLADES F I G U R E 13.4. Mechanical principles of tooth design in relation to food type. (Top) Pestle and mortar, cusp (pestle) acting against a mortar (fossa in upper tooth), and molars of a cercopithecoid frugivore. (Middle) Cutting and shearing blades, section through dental blades, and carnassial teeth found in felid carnivores. (Bottom) Grinding blades, cross section through lophodont cheek teeth (note alternating enamel and dentine), and molars of a selenodont artiodactyl. Enamel is shaded in all cross-sectional drawings. Note the direction of movement (heavy arrows) in each mechanical model and the degree to which this is mimicked in the dentitions shown.
seums provides a dramatic demonstration of the effect of a tough diet (Table 13.2) on human postcanines— and the Anasazi had fire! The bulk (crown and root) of the mammalian tooth is formed by dentine (see Chapter 2). In contrast to enamel, which is acellular, this tissue is supported by odontoblasts (analogous to osteocytes), which have their cell bodies in the pulp cavity, but long processes extending through the dentine. Dentine can be formed throughout life. It is softer than enamel and is therefore abraded more rapidly. This difference in wear resistance has been exploited by some groups of mammals, particularly herbivores and rodents. Both consume highly abrasive (siliceous) foods. The difference
13. Feeding in Mammals in hardness between enamel and dentine is used to enhance the efficiency of food reduction, as the enamel "edges" act as blades with the more rapidly abraded dentine as catchment areas for the product of toothtooth interaction (Fig. 13.4, and see Rensberger, 1995). Herbivorous mammals have evolved three approaches (often combined) to the challenge of tough, fibrous food sources. They can be characterized as (1) the fusion of discrete cusps into complex patterns of ridges or lophs (see Janis, 1995); (2) designs to preserve tooth function for as long as possible; and (3) designs to optimize the effectiveness of the teeth in accessing or processing available food. The herbivore solutions to the problem of maintaining tooth function have been (a) a greatly prolonged period of tooth development, i.e., hypsodonty ("high-crowned" teeth), which continue to form for years after initial eruption (e.g., Perissodactyla), but do have a finite life or (b) persistent growth (e.g., rodent incisors). Rodent incisors have an active tooth germ in the base of their alveolus that generates new enamel and dentine as the erupted tooth is worn away by normal gnawing activity. Some rodents (e.g., hystricomorphs) also have persistently growing cheek teeth and (c) "molarization" and fusion of teeth. If the "metabolic equation" requires the ingestion of very large volumes of vegetation (since its nutrient value is low), either (i) remodel the premolars into molariform teeth so as to increase the postcanine dental working surface for food reduction or (ii) make each tooth (premolars and molars) the equivalent of a complete cheek tooth row and erupt them seriatim over a very long lifetime. Fiorses, most ungulates, and some rodents have adopted the former approach; elephants, uniquely, the latter.
V. THE FEEDING APPARATUS The feeding apparatus can be divided, for convenience only, into two functional complexes: (1) the jaw complex, including jaws, teeth, and the muscles of mastication. The jaw complex is used primarily in food acquisition and processing. (2) The oropharyngeal complex, comprising mostly soft tissues. The oropharyngeal complex manages all intraoral food transport, bolus formation, and deglutition (and in some taxa, ingestion, as well). Clearly, more is known about the evolution of the former, given the nature of the fossil record. A. Jaw Complex 1, The Masticatory
Cycle
Using changes in the rate of jaw movement in the opossum as the criterion, the masticatory cycle was
421
divided into a number of stages (Hiiemae, 1976, 1978; see Chapter 2). Jaw closing was described as having fast close (FC) and slow close (SC), or power stroke (PS), phases; the latter being distinguished by a reduction in the rate of closure when teeth met the resistance of food. With more recent work, the splitting of jaw opening into slow open (SO) and fast open (FO) phases has proved a problem, as pointed out by Schwartz et al. (1989). In their study of feeding in the rabbit, they identified three phases in opening: O l , 0 2 , and 0 3 . 0 3 is not always present, but when it is, it is always rapid, ending at maximum gape (see Fig. 13.5). This four/ five-phase pattern in the masticatory cycle has been observed in most nonhuman mammals studied, including those that feed in unusual postures such as bats (de Greet and de Vree, 1984) and sloths (Naples, 1985). Originally, an intercuspal phase (IP), in which no discernible vertical movement occurs, was included in the power stroke. We are now treating this, at least for higher primates, as a distinct element in the feeding cycle (Fig. 13.5), not least because it occupies a significant proportion of total cycle time (Hiiemae et al, 1995, 1996; Palmer et al, 1997; Hiiemae and Palmer, 2001). Events in IP involve the transition from "jawbased" to "tongue-based" behaviors (see Section V,B). 2. Jaws and Teeth The shape and proportions of the skeletal elements of the feeding apparatus in mammals vary widely. The early evolution of the mammals dictated the fundamentals of their craniofacial anatomy. There is a large literature on mammalian cranial osteology (see Turnbull, 1970) and dental anatomy [see Peyer's (1968) classic text, which also includes a useful review of the early papers on the origin of the tribosphenic molar]. Other studies (e.g., Radinsky, 1981a, 1981b, and 1982 for carnivores) have examined the evolution of skull shape within major groups. The primitive mammalian dentition is highly differentiated, i.e., heterodont: the basic mammalian formula is three incisors, one canine, four premolars, and three molars in each jaw quadrant with each upper tooth occluding with its opposite lower jaw equivalent, but somewhat posterior to it (one-half tooth length behind in the case of the molars). There is general agreement that the molars of mammals were derived from an ancestral tribosphenic molar by selection for the features shown in Fig. 13.4 (Butler, 1952; Mills, 1955; 1966; Crompton, 1971; Hiiemae and Kay 1972; Kay and Hiiemae, 1974; Janis, 1979; Janis and Fortelius, 1988). The overall shape of the skull is affected by a wide variety of factors, including size and position of the orbits, the complexity of the nasal airway, the proportions of the muscles, and the size of the brain (e.g., van
422
K a r e n M. H i i e m a e
A. DIdelphls TIME
Tongue Reversal
Puncture Crushing: 315 ms.
Chewing: 400 ms
B. Macaca fascicularis GAPE
FIGURE 13.5. Gape-time plots for Didelphis and Macaca fascicularis. (A) Opossum. Mean values for the duration: total and included phases for the two cycle types identified in Hiiemae (1976). The first cycle shown is a puncture-crushing cycle. This occurs early in a feeding sequence and is used to crush/fracture the ingested food (N = 18). The teeth do not achieve occlusion but there is a clearly defined slow close phase in closing. Slow open (SO) and fast open (FO) phases are present in opening. In the longer chewing cycle (N = 24), the teeth reach occlusion. Arrows show the primary direction of jaw movement: vertical in puncture-crushing and lateral to medial in chewing. Tongue protraction occurs in the first part of opening, reversing at the transition between SO and FO. (B) Macaque. Analysis of jaw, tongue, and hyoid movement has shown that there are no "'stereotypicar' feeding cycles as measured by jaw movement. Cycle profiles are highly correlated with the type of food (initial consistency) and stage in the feeding sequence (Hiiemae et ah, 1995; Thexton and Hiiemae, 1997). Movements of a radio-opaque marker on the anterior tongue, and of the hyoid, measured from the anterosuperior aspect of the hyoid show that there are patterned movements of both tongue and hyoid during masticatory cycles. The one predictable relationship between jaw and tongue/hyoid movement is the reversal from tongue protraction to tongue retraction at the SO-FO transition. In macaque (Hiiemae, 1995), this linkage occurs within a 30-msec "window.'' Hyoid movement does not exactly parallel that of the tongue surface (see Section V,B).
der Klaauw, 1945; Schwenk, 2000b). Nonetheless, long tooth rows require long snouts and jaws. The relative proportions of braincase and snout (defined here as the combined anteroposterior length of the premaxilla and maxilla, inclusive of the tooth row, to the anterior margin of the orbit) are illustrated in Fig. 13.6 (Didelphis) where the ratio of braincase to snout length is approximately 1:2. Using slightly different criteria, Fearnhead et al. (1955) measured the same ratio using the glenoid (squamosal-dentary joint) as the reference point and showed that for insectivorous mammals (presumed to reflect the ancestral condition), the ratio averaged 1:37 for erinacids, 1:5 for tenrec, but only 1:1.4 overall for soricids, which have relatively expanded braincases.
As different mammal lineages evolved, the ratios changed. For example, Fearnhead et al. (1955) give the ratio as 1:1 in mustelids (Carnivora). Such changes reflect the characteristics of each order and the dietetic specializations within it (Fig. 13.6). It is important to note that the lateral surface of the braincase provides attachment for the temporalis muscle. Where the underlying brain is comparatively small, as in Didelphis, the presence of a large temporalis may lead to the development of sagittal crests. A combination of powerful temporalis and nuchal (neck) musculature can produce a T-shaped pattern of crests. This development occurs in primitive mammals, in many carnivores, and some primates. Similarly, the zygomatic arch may be
13. F e e d i n g in M a m m a l s
423
FIGURE 13.6. Simplified lateral views of the skulls of (A) Didelphis, (B) Cehus, (C) cat, (D) sheep, and (E) rat to illustrate the different proportions of the bony elements in examples of each of Turnbuirs ''dietetic groups/' All skulls are drawn to the same overall length to allow direct comparisons of the proportions of the cranium, the lower jaw, and the length of the tooth row. Areas of attachment for the major muscles groupings (see Fig. 13.7) are shaded (light stipple, temporalis complex; darker stipple, masseter complex). (A) Based on Hiiemae and Jenkins (1969), (B) on Le Gros Clark (1959), and (C-E) on TurnbuU (1970). Arrows are intended as a simplified indication of the overall line of action of each complex and are "weighted" in accordance with the proportions of the muscles shown in Fig. 13.7.
attenuated or missing in some insectivores, but very robust in herbivorous mammals with a powerfully developed masseter. Each half of the lower jaw (hemimandible) is composed of a single bone, the dentary. Embryologically, the dentary has three major elements (e.g., Atchley, 1993): (1) a basal component, the dentary proper, which encases the inferior dental nerve and extends posteriorly to form the mandibular condyle; (2) an alveolar component, i.e., the bone associated with tooth
attachment (above the nerve); and (3) posteriorly, the coronoid and angular processes of the mandibular ramus, which develop around the core dentary axis. The size and shape of these processes are dictated by the relative development and organization of the adductor muscles: the coronoid process serves primarily as a site of attachment for the temporalis muscle whereas the masseter and internal (medial) pterygoid muscles insert primarily on the angular process. It is important to note that experimentally impaired development or
424
Karen M. Hiiemae
ablation of developing temporalis or masseter/internal pterygoid muscles results in attenuated process development (Avis, 1961). The mandible develops as an intramembranous (dermal) ossification investing Meckel's cartilage. The latter is later resorbed anterior to the middle ear. Embryonically, the two halves are separated anteriorly by a fibrocartilaginous joint {mandibular symphysis). This joint may persist throughout life, allowing some independent movement of the two hemimandibles, or it may ossify. In the former case, the symphyseal joint may have a complex arrangement of the constituent fibers to regulate the range and direction of intramandibular movement [e.g., Didelphis, tenrec, some rodents, and herbivorous marsupials and placentals (Beecher, 1979)]. A moveable symphysis is often associated with the presence of a transversus mandibulae muscle. It follows that (a) the length of the upper and lower jaws is highly correlated with the length of the tooth row; (b) the vertical height of the body of the mandible is highly correlated with the length of the tooth roots; (c) the anatomy of the coronoid and angular processes reflects the relative development of the muscles attached to them (i.e., a large angular process is associated with a well-developed masseter and internal pterygoid and, conversely, a small coronoid process with a weakly developed temporalis); and (d) the morphology of the lateral margins of the premaxilla and maxilla is also affected by the alveolar bone associated with the anchorage of the upper dentition, although this bone may be "subsumed" into the lateral walls of the nasal cavity, thus masking the length of the tooth roots and creating a relatively flat hard palate from which only the crowns of the upper teeth appear to project. The mandible articulates with the skull through the squamosal-dentary joint (otherwise, the jaw-joint, cranio-mandibular joint, CMJ, temporo-mandibular joint, or TMJ—the latter term most often applied to humans). The squamosal, or the bone forming the cranial articular surface for the CMJ/TMJ, is fused with other ossifications during development of the mammalian skull to form a composite temporal bone, hence the mandible articulates with the "squamous portion" of the temporal bone. Movement of the lower jaw relative to the upper is produced by integrated actions of the jaw muscles acting to move the dentary condyle on the squamosal fossa. For a mammal to ingest and process food, it must be possible for the jaws to open as wide as needed to acquire or ingest the food item and then to close, bringing the entire tooth row into full occlusion, while at the same time facilitating the movements of the teeth into and out of occlusion (arrows in Fig. 13.4). In primitive mammals, this is achieved by a CMJ positioned just above the level of the occluded tooth row. The joint position is higher in other mam-
malian groups (see Fig. 13.6). The elevated position of the jaw joint relative to the bite point permits upper and lower teeth to engage more or less simultaneously along the entire tooth row during occlusion (as opposed to a pure "scissors" action in which the teeth would engage sequentially from posterior to anterior; see Greaves, 1995). The evolution of the squamosal-dentary joint (see Crompton, 1995) created the potential for mandibular movement in the mediolateral (coronal, M-L,) and anteroposterior (sagittal, A-P) planes in addition to the vertical movement common to all vertebrates. (In much of the older literature the following terms are used: orthal for vertical, ectental for mediolateral, and propalinal for anteroposterior movement.) Didelphis and its placental equivalents with basically tribosphenic molars utilize movement in all three axes coupled with hemimandibular rotation. To optimize the shearing capacity of their cheek teeth, the lower jaw on the working side (the side on which food is positioned; see later) is moved upward, medially, and forward. In addition, each hemimandible can rotate about its longitudinal (A-P) axis. Morphology of the putatively ancestral mammalian CMJ, which facilitated movement in all three planes, was modified in the various ordinal lineages to optimize movement in one or another of the primary directions. For example, in carnivorans, especially some felids and mustelids, movements are essentially restricted to the vertical and M-L movement is variably constrained, sometimes extremely so; in most rodents the capacity for A-P movement is greatly increased, whereas chewing in most other herbivores uses a primarily transverse (ML) jaw stroke, during which the condyles must pivot about a vertical axis on the glenoid fossa (Table 13.3, Fig. 13.8). Although there are exceptions, most mammals chew food on only one side of the mouth at a given time. This side is designated the working side, the other side, the balancing side. Experimental and theoretical (modeling) work on the biomechanics of the jaws has shown both how forces are transmitted between the two sides and how patterns of muscle activity not only produce the compressive or shearing forces needed to disrupt the integrity of food items, but also maintain the integrity of the CMJ (Greaves, 1978, 1995; Hylander, 1979). 3. Muscles of
Mastication
As shown in Fig. 13.6, the areas and shapes of the major adductor muscle attachment sites differ among the dietetic groups. The muscles of mastication can be divided into jaw closers {adductors) and jaw openers (or depressors) {abductors). The digastric muscle is generally considered the principal abductor, although this
TABLE 13.3 Primary Anatomical and Dental Features of the Craniofacial Complex in Terrestrial Mammalsa Skull Primitive mammals, e.g. Didelphis marsupialis
Lower iaw
Squamo-dentarv ioint
Dentition
Long snout, complete zygomatic arch, sagittal and nuchal crests
Well-developed coronoid Inflected angular process Shallow body Mobile symphysis
Transverse, postglenoid process Condyle convex A-P Joint just above occlusal plane
Is 4, C, PMs 4, Ms 3 Projecting canine Tribosphenic molars
Placental insectivores
Braincase expanded (Soricids) Some: incomplete zygomatic arch (Soricids, tenrec)
Big coronoid, hooked angular process Shallow body Mobile symphysis
As for Didelphis in most Soricids: double joint
Incisors may be reduced or modified depending on habit Tribosphenic or bunodont molars
Bats
Generally as for insectivores
Primates Prosimians
Anthropoids
Carnivores Ancestral, and canids, mustelids, many viverrids
Highly modified anterior dentition in many forms Molars tribosphenic with heavy emphasis on connecting ridges ('W' shape) Bunodont in fruit bats
Primitive (e.g., Tupaia)-like insectivores Progressive expansion of braincase, especially orbits Relative reduction in snout length
Primitive, as for insectivores Coronoid and angular processes well developed Shallow body Mobile symphysis
Glenoid fairly flat, condyle rounded both axes
Tends to reduce number of Is, PMs Molars tribosphenic or with shearing planes or bunodont derivative (four cusps)
Progressive migration of foramen magnum toward skull base Highly variable snout length, even within related groups Relatively flat palates
Well-developed processes Angles grossly enlarged in some cebids Simian shelf with recessed attachments for genial muscles Fused symphysis
Transversely oriented condyle Convex A-P Postglenoid process Well above occlusal plane
Spatulate incisors, diastema upper lateral I, and canine for occlusion lower C Canine sexual dimorphism Two PMs, lower first bladelike Molars generally four-cusped (uppers), sometimes five Molars tend to increase in size distally
Expanded cranial volume Expanded orbits Variable snout length
Large coronoid, smaller projecting angular process Mobile symphysis (but fused larger forms, e.a., " ursids)
Pre and postglenoid processes, glenoid sharply concave Condyle transverse and sharply convex Joint at level occlusal plane
Reduced in total number Powerful canines Last upper premolar, first lower molar modified to form carnassial shearing blades
Omnivorous, herbivorous ursids, procyonids, many viverrids Felids (pure carnivores) Ungulates Hyracoidea, Perissodactyla, Artiodactyla (Lagomorphsfi)
Rodents Sciuromorpha, Myomorpha, Hystricomorpha
Molars may be bunodont for puncture-crushing/ grinding Expanded cranium Short snout, heavy zygomatic arch palate
Short, robust Large coronoid, small angular process Deep masseteric fossa
As described above
Reduced molar series Camassials dominate, are very large
Skull dominated by large snout, making cranium appear small Reduced temporal fossa Short stout zygomatic arch
Long, body fairly deep Slender, often recurved coronoid process Masseteric area on mandibular ramus (including angle) expanded greatly Mobile symphysis in most (fused adult horses, swine)
Glenoid relatively flat Retroarticular process big artiodactyls Condyle small, flattened Condyle rotates horizontally (lateral movement jaws) Joint close to occlusal plane in primitve forms, rises significantly in advanced forms
Upper incisors replaced by keratinous pad (artiodactyls) Some large canines (pigs, peccaries, hippos) Canine dimvorohism
Small coronoid, projecting welldeveloped angular process Lower border curved anteriorly (lower incisor alveolus) Mobile symphysis
Glenoid oriented anteroposteriorly Concave in M-L section Condyle elongated A-P, convex M-L May have two articular facets (one for anterior occlusion = incisors; second for posterior occlusion (molars)
Reduced greatly One I each quadrant, persistent growth No canines Reduced premolars (usually 1) Lophodont molars, can be very complex (hystricomorphs) and of persistent growth Molar rows parallel or diverge posteriorly Isognathy Two separate occlusal planes
Rounded braincase (most) Large orbits Snout dominated by upper incisor alveolus Zygomatic arch expanded anteriorly for part masseter (sciuromorphs, myomorphs) Infra-orbital foramen for part massetel (myomorphs, hystricomorphs)
(b) hypsodonty Molarisation of premolars Absolute increase in molar size
aBased on Turnbull (1970), Weijs (1994), and basic texts such as James (1960) and Peyer (1968). For comparison with Fig. 13.7. This table is not intended to illustrate phylogeny or primitive and derived characteristics. b ~ l t h o u g lagomorphs h have rodent-like incisors, Weijs (1994) considers rabbits and hares functional ungulates.
426
Karen M. Hiiemae
may not be the case when the jaw is depressed in normal feeding (see later). The adductors can be broadly identified as temporalis, masseter, internal, and external pterygoid muscles. However, I concur with Herring (1993) when she states "unfortunately, jaw-closing muscles are not only heterogeneous internally but are also linked to each other externally, so their homologies, while undoubted, are imprecise." In fact, nomenclature for the various "parts" of the major muscle complexes is overly complex; same-named parts may have very different actions and mechanical effects (see Turnbull, 1970; Weijs, 1994, and literature therein). Herring continues by arguing that "for functional purposes, it might be better to designate muscles as vectors rather than as named parts." This approach was developed by some of the earlier comparative anatomists (Arendson, 1951; Becht, 1953, Maynard Smith and Savage, 1959), used by Hiiemae and Jenkins (1969), and was the foundation of later biomechanical modeling of muscle action and forces on the jaws (e.g., Scapino, 1972; Greaves, 1978, 1988; Smith, 1978; Bramble, 1978; see also review by Hylander et al, 1992). That said, mammalian jaw adductors have probably received more attention than perhaps any other muscles in the body. Their relative mass and fiber orientation (Turnbull, 1970) and internal architecture, as well as their motor unit organization and firing patterns, have all been examined (see reviews by Herring, 1992,1993,1994). In an exhaustive review, Weijs (1994) synthesized all the information then available to provide an invaluable discussion of the behavior of the muscles and the mandibular movements they produce based on Turnbull's classification. The following is, therefore, a brief of summary based on these reviews. Differences in the proportions of the major muscles shown in Fig. 13.7 reflect the linkage among jaw movement, tooth form, and diet. The dominance of the temporalis in Didelphis and Echinosorex is associated with powerful jaw-closing movements for crushing and piercing of prey followed by crushing and cutting. A similar pattern in the cat reflects the dominance of the canines and carnassials in felids, which use their adductors during both prey capture and ingestion. In carnivorans, generally, the line of action of the temporalis is directed so posteriorly that a large, recurved postglenoid flange has developed in the upper jaw joint to prevent posterior dislocation of the condyle. In felids and some mustelids, the development of an additional preglenoid flange has produced a deeply concave upper articular surface into which a nearly cylindrical mandibular condyle fits snugly, restricting jaw movement to the vertical with only the slightest "play" in the transverse plane permitted to adjust carnassial shearing (this jaw joint is often so tight that even when a prepared skull is picked up the mandible remains "attached"). A
sizable masseter adds additional power but with a slightly anteriorly directed line of action. Although the rodents and the large herbivores shown in Fig. 13.7 appear to have very similar percentages of niasseteric muscle mass, this masks a real difference: in the former, the superficial masseter is very large, whereas in the latter, the deep masseter and zygomaticomandibularis dominate. To engage their incisors, rodents protract the lower jaw. The power stroke during mastication also has a large anterior component and molar grinding of food is mostly A-P (Hiiemae and Ardran, 1968; Weijs, 1974; Weijs and Dantuma, 1975). These movements are powered by a large and nearly horizontal superficial masseter and a complex, anteriorly placed "deep" masseter, with some assistance from the internal pterygoid. In an earlier review (Hiiemae, 1978), I argued that the "pattern of EMG activity was broadly similar in all the mammals studied so far (to 1976) despite the differences in their profile of jaw movement," implying some generic (mammalian) pattern of central motor control. Given the differing muscle architectures in ordinal groups of mammals, I did not suggest that this "generic" pattern of activity would produce the same amplitude or direction of movement. This proposition was interpreted, not without justification, as embodying an hypothesis, i.e., the existence of a "mammalian effector Bauplan" for the activity pattern of the jaw adductor muscles emanating from a masticatory CPG central pattern generator in the hindbrain. This idea was followed by a paper by Bramble and Wake (1985) and a response from Smith (1994). Both papers are important, given the fact that mammals have evolved from vertebrates relying on a hyobranchial feeding apparatus. Weijs (1994) has systematically reviewed all available data, including his and other's work post-1976, and has developed an important synthesis, which demonstrates adjustments in the "basic mammalian pattern" associated with dietetic adaptations. Weijs (1994) examined electromyographic (EMG) data for the numerous species now studied, arguing that it warrants a "functional activity" approach to the interpretation of jaw adductor motor patterns during the closing (power) stroke of the jaw. His conclusions are based on motor activity patterns in several generalist mammal species presumed to represent the primitive condition. He proposes that there are muscle groupings that function as "triplets," in effect a " p u l l pull" system, to generate the laterally directed mandibular movement, which positions the teeth for the medially directed power stroke on the active side. The jaws are closed from initial maximum gape by the symmetrical action of vertically oriented muscle fibers (e.g., zygomaticomandibularis and allied fibers such as those in the most anterior and deep parts of tempo-
427
13. F e e d i n g in M a m m a l s PRIMITIVE MAMMAL
Didelphis
INSECTIVORA
Echlnosorex
CARNIVORA
Felis
PERISSODACTYLA Equus i
ARTIODACTVLA
Odocolleus
RODENTIA
Sculrus
10
20
30
40
50
60
70
80
90
100
Percentage Total Jaw Muscle Mass (Adductors and Digastric) Masseter Complex
Temporalis Complex
Internal Pterygoid
I '
I External ' Pterygoid
Digastric
FIGURE 13.7. Relative proportions of the adductor musculature, with digastric, expressed as a percentage of the total muscle mass (data from Turnbull, 1970). Relative proportions of the masseteric complex (superficial, deep, and zygomatico- mandibularis) in herbivores and rodents as compared with the temporalis (anterior, posterior, deep) in carnivores and insectivores is clearly shown. It should be noted that the proportions of the muscle complexes identified vary within the dietetic groups such that, for example, the actual percentage of temporalis/masseter muscle mass may differ substantially but the dominance of one or other complexes remains. Interestingly, in marine carnivores the proportion of digastric may rise to as much as 25% of the total.
ralis). As closing proceeds, the jaw is moved laterally by the combined activity of the working side, posteriorly directed fibers of the temporalis coupled with anteriorly directed fibers of the masseter and medial pterygoid on the balancing side. This grouping is Weijs' trip-
let 1. As the teeth approach tooth-food-tooth contact and occlusion, a second reciprocal grouping of the balancing side, posteriorly directed fibers and the active side, anteriorly directed fibers (triplet II) come into play carrying the teeth through to centric occlusion.
428
K a r e n M. H i i e m a e
The asynchrony in the onset of activity in triplet II varies within the generalized mammal group depending on the degree of transverse movement during the power stroke. This is small in opossum but large in tenrec and the little brown bat. Contraction of muscle fibers involved in both triplets contributes both to movement direction (until restricted by occlusal contacts) and to masticatory force.
Four variants of this basic pattern are identified (Weijs, 1994). In mammals where the transverse (L-M) movement is severely constrained either by carnassials (felid and mustelid carnivorans) or by interlocking canines (e.g., fruit-eating bats and most carnivores), triplets I and II fire synchronously. This is Weijs' carnivore symmetric pattern (Fig. 13.8.) A similar, but not identical, symmetrical pattern is found in at least three
I Fast Close i Slow Close Early Open PRIMITIVE
L
• ws+bs vertical fibers
M
)W////////////////A
WAm^Mmm ;sssssssssssssssssssss
V///////////////////A
L
CARNIVORE SYMMETRIC
• ws posterior fibers (TEM)" bs anterior fibers (MAS) • bs anterior fibers (MPT) _J ' bs posterior fibers (TEM) - ws anterior fibers (MAS) II • ws anterior fibers (MPT) _J
M
%mmw/////M y///////////////////A -1-
RODENT SYMMETRIC P^
p
^A
A )%^
Chewing
Gnawing
j^^^;^^^^?;?^^^ W///////////////A
low level or inconsistent activity "^-ws lateral pterygoid (LPT)
7/////////////////A
I |
TRANSVERSE L
I
RODENT ALTERNATE
r
^cis r^
Chewing
^"Vl
& 1
E55555555555^^^^
7///////////////////////^^^^
^m '
FIGURE 13.8. Patterns of adductor activity during the closing stroke of masticatory cycles in mammals (modified from Fig. 1 in Weijs, 1994). To illustrate the path of jaw movement during the cycle, typical, if somewhat diagrammatic, jaw movement "profiles'' are shown. The ''primitive,'' "carnivore symmetric," and "transverse" groups are based on data from Hiiemae (1978); for the "rodent symmetric" group from Hiiemae and Ardran (1968) and Weijs (1975); and for the "rodent alternate" group from Byrd (1981). In all cases (except rodent gnawing where the occlusal profile of an upper incisor is shown), the horizontal line represents the upper occlusal plane. The vertical dashed line for nonrodent groups represents the position of centric occlusion (maximal intercuspation) on the working side (L, lateral; M, medial with respect to centric occlusion). The direction of chewing and gnawing strokes in the "rodent symmetric" group is shown in the sagittal projection (P, posterior; A, anterior). The alternating profile for Cavia, "rodent alternate" group, is drawn in the coronal plane, ws, working side;bs, balancing side; TEM, temporalis (posteriorly directed fibers); MAS, masseter (anteriorly directed fibers, usually referred to as "superficial masseter"); MPT, medial (internal) pterygoid. Shading is intended to indicate the line of fiber pull, with anterior to the right of the figure. Triplets are indicated by brackets.
429
13. Feeding in Mammals rodent lineages (murids, pedetids, and some hystricomorphs), all of which have isognathous dentitions (Table 13.3) and chew simultaneously on both sides of the jaws. The rodent symmetric pattern involves early firing of the posteriorly directed fibers to retract the jaw, followed by anteriorly directed fibers, which protract it and power the anteriorly directed chewing stroke. In this case (Fig. 13.8), activity within both triplets is temporally spaced. Conversely, in those mammals with flattened (lophodont, bunodont) occlusal planes and anisognathous tooth rows (e.g., ungulates, lagomorphs, and higher primates), the power stroke is heavily transverse. In the transverse pattern, triplet I acts both to close the jaws and to swing the mandible out toward the working side. In humans, this lateral swing often occurs at maximum gape (Hiiemae and Palmer, 2001). Weijs suggests that activity in triplet II may be triggered by tooth-food-tooth contact at the beginning of the power stroke, as these muscles often show greater EMG activity than those of triplet I. Pterygoids on the working side are the last to stop firing; their activity may extend into the intercuspal phase or early opening (Ol, Weijs, 1994), facilitating the disengagement of the teeth on the working side. A few mammals (two rodent lineages, e.g., Cavia, Myocaster, Hydrochoerus, and the mole rat, Tachyoryctes), as well as the camel (Hiiemae, 1978), the warthog (Ewer, 1958), and other suids (Fierring and Scapino, 1973; FFerring, personal communication), chew by alternating working and balancing sides. The activity pattern of the muscles in the guinea pig, Cavia, is known (Byrd, 1981). Symmetrical closer and triplet I activity is reduced to very low levels, with dominant triplet II activity (Fig. 13.8). In his review, Weijs (1994) focuses on muscle activity during jaw closure and the power stroke, commenting that there may be some adductor activity during opening, as well as some digastric activity during closing. Hiiemae et al (1995) and Sher et al (1997) suggest that the explanation for these patterns is associated with tongue function during processing and intraoral transport. 4. Regulation of Adductor Muscle
Activity
Food processing in mammals, i.e., chewing sensu strictu, is rhythmic. The lower jaw opens and closes, changing gape. Gape change may also be augmented by movement of the cranium on the vertebral column (Hiiemae, 1978), as often occurs in nonmammalian tetrapods (Chapter 2). Considerable attention has been given to the neural control of this process [see Taylor (1990) and other papers in that symposium volume]. Given the general agreement that the process of
chewing is maintained, once initiated, by a CPG in the hindbrain (Bremer, 1923; Dellow and Lund, 1971), there is a serious question as to the sources of the sensory input that regulate motor outflow to the adductor and depressor musculature (Hiiemae, 1993). Two such sources have been identified: (1) spindles in the adductor muscles and (2) receptors in the periodontal ligaments of the teeth. Both clearly play a significant role, the former monitoring stretch in the muscles and the latter occlusal load on the teeth. It is easy to understand how these receptors can provide sensory input regulating jaw closing, especially a response to the resistance afforded by food between the teeth. A problem arises, however, when the regulation of jaw opening is considered (see Section VII; Thexton and Crompton, 1989), given the presumption that the periodontal receptors are "off loaded" as the tongue cycles relative to the palate as food is transported or repositioned. It has to be said that the complexity of the behaviors intrinsic to normal feeding in mammals has only become apparent over the last two decades and well documented in recent years. It is no longer sufficient to address the regulation of orofacial mechanisms in feeding solely in terms of jaw movement cycles. Activities of the hyoid, the tongue, and the soft palate, some of these also highly rhythmic and linked to jaw movements, must also be considered.
B. Oropharyngeal Complex 1.
Overview
In 1978,1 and my colleagues published a paper entitled, 'Tntra-oral food transport—a fundamental mechanism of feeding?" We had previously observed cycling of food within the mouth in the opossum (Hiiemae, 1976), but only then began to focus on the behavior of the tongue, hyolaryngeal complex, and pharynx. The oropharyngeal complex can be defined as two spaces, separated by a sphincter-like mechanism (see also Chapter 2): (1) the oral cavity extending from the lips and incisors anteriorly to the fauces posteriorly; roofed by the hard palate, walled laterally by the teeth and cheeks, and with a muscular mobile floor; largely, but not completely, occupied by a highly mobile tongue; and (2) the oropharynx, extending from the fauces back to the esophagus; with the soft palate and nasopharynx above, v/alls of pharyngeal musculature above or behind the hyoid, with the epiglottis, larynx and esophageal aditus, behind or below. It must be emphasized that the larynx and the esophageal aditus are intimately associated with this complex. Indeed, its physiology is designed to assure the passage of (a) air into the larynx and (b) swallowable food into the esophagus. The sphincter mechanism consists of the
430
Karen M. Hiiemae
palatoglossal arches, with their included muscle (otherwise, faucial arches or pillars), the soft palate (also muscular), and the posterior surface of the tongue. Not only is the tongue an essential component of both parts, so is the hyoid complex. However, as far as is known, there appears to be much less gross variation in the function of this complex (although substantial anatomical variation within each component) than in the jaws, teeth, and muscles of mastication (with the possible exception of intrinsic tongue anatomy). This is hardly surprising as the core functions of the oropharyngeal complex—food transport, bolus formation, and swallowing—are common to all mammals, in contrast to the variety of mechanisms used to process different types of foods by the teeth and jaws (see earlier discussion). However, other oral and perioral soft tissues play an important role in the management of food within the oral cavity and the processes illustrated in Fig. 13.3. These are the cheeks, lips, and the epithelium of the hard palate (Chapter 2). Epithelial surfaces of the oropharyngeal complex are lubricated by widely distributed mucus-producing glands. Serous, salivary glands generate copious volumes of saliva in response to the presence of food in the mouth (Hector and Linden, 1987) or in response to odors or other stimuli (e.g., Ludwig, 1850; Bernard, 1852; Pavlov, 1910). These glands are anatomically discrete, extraoral, and, in the case of the parotid and submandibular glands, have well-defined ducts. Secretions often include some enzymes, such as salivary amylase, which begin chemical breakdown of starches to simpler sugars. A discussion of salivary gland anatomy, variations in the chemistry of saliva, and the mechanisms by which salivation is controlled is beyond the scope of this chapter. The last topic has been reviewed by Garrett and Proctor (1998). 2. Oral Cavity The cheeks form the outer walls of the oral cavity, lateral to the tooth rows and attached to the alveolar bone above and below the teeth, creating vestibules within which food can be held or stored during food processing. In many mammals, especially rodents and some primates, the vestibules are highly extensible (e.g., the chipmunk, whose expanded cheeks are wider than its thorax when full of nuts) or have diverticuli to form extraoral pouches (e.g., Brylski and Patton, 1988). These are "food-gathering/hoarding" extensions of the vestibular spaces. The muscle of the cheek is the buccinator, whose fibers decussate around a modiolus ("hub") to pass into either the upper or the lower lip (Fig. 13.9). The anteroposterior position of the modiolus dictates the extent to which the cheeks can
control food position (Heath, 1991). In many mammals, e.g., Didelphis (Hiiemae and Jenkins, 1969), the modiolus is lateral to the first molar, limiting the area of the tooth row in which tongue and cheeks can work to control food position. The inability of the cat to "manage" 1-cm^ cubes of hard liver (see earlier discussion; Thexton et ah, 1980) is almost certainly a function of limited vestibular space due to a posteriorly placed modiolus. Fibers of the buccinator muscle merge anteriorly with those of the lip muscles, specifically the orbicularis oris. Mammalian lips can be highly mobile, as in anthropoid primates and humans, but more typically the upper lip is "tied down" anteriorly by the rhinarium. A complete anterior oral seal can, nevertheless, be formed, as its creation depends on the approximation of upper and lower lip margins, even if the modiolus is positioned posteriorly as in Didelphis. Intriguingly, rats have been observed to "close off" the mouth behind the incisors; indigestible, gnawed material then spills laterally behind the incisors, never entering the oral cavity (personal observation). There are little understood functional variations in the organization of the cheeks and lips in mammals. Many mammals, for example, are unable to appose mobile upper and lower lips in front of the incisors, as do higher primates and ungulates. This special facility allows for a novel range of facial expression and communication (see any film of a distainful camel or baboons or chimpanzees, where the eversion of the upper lip is a definite, and now well-documented, social message). It is worth noting, especially regarding the evolution of primates, that as the modiolus has moved anteriorly, the palatal rugae have become attenuated. Rugae are ridges on the palatal surface of the oral cavity, formed by dense accretions of connective tissue below the epithelium, which are often keratinized (Chapter 2). In most mammals the entire surface of the hard palate is rugose. The ridges are positioned relative to the postcanines so as to optimize their ability to collect triturated food at the end of each power stroke (see Crompton and Hiiemae, 1970; Fig. 13.13). The asymmetric architecture of rugae creates asymmetric, concave catchment areas on the hard palate. The anterior slope of each ruga is shallower than the posterior to prevent spillage of liquid, or semisolid foods, as the jaws open (see later). In higher primates, e.g., the macaque, rugae are reduced, although they still contribute to the management of food during stage I transport (German ei ah, 1989). In humans, they are represented by only a few small mediolateral elevations of the palatal epithelium behind the upper incisors. This pattern suggests an, as yet, untested evolutionary hypothesis for anthropoid primates: the
431
13. F e e d i n g in M a m m a l s
development of a peripheral oral muscular wall (cheeks and apposable lips) with an anteriorly positioned modiolus has provided an expanded containment mechanism for food in the oral cavity, which has replaced that offered by the rugose hard palate.
D.Mass
3. Hyoid Complex and Tongue Base
S. Mass
Dig (pb)
The hyoid is a branchial arch derivative: the medioventral portion of the hyoid apparatus (basihyoid) and first branchial (thyrohyoid) arches form a C-shaped structure that cradles the larynx and is connected to it by muscles and ligaments. In all mammals, it forms a rigid arch buttressing the anteroinferior pharynx. In opossum, tenrec, lagomorphs, and primates, it has no connection to other skeletal elements. In ungulates, carnivores, and some rodents, it may be connected to the tympanic area of the cranial base by the remainder of the hyoid arch, which forms a jointed series of bony or cartilaginous rods. Since Edgeworth's (1914) classic paper, the comparative anatomy of the mammalian hyoid has been discussed by a number of authors; Hilloowala (1975) for primates; Woods (1975), who investigated the hyoid, laryngeal, and pharyngeal region in selected rodents; and Anapol (1988), who describes its anatomy in the rabbit and examined its movements. Hyoid movement patterns, based on Xray studies [cinefluorographic (CFG) or videofluorographic (VFG)], have been studied in opossum (Crompton et al, 1977), cat (Thexton et al, 1980; Thexton and McGarrick, 1988,1989), tenrec (Oron and Crompton, 1985), rabbit (Anapol, 1988), hyrax (German and Franks, 1991), macaque (Hiiemae et al, 1995; Hiiemae and Crompton, 1985; Thexton and Crompton, 1998), and human (Palmer et ah, 1997; Hiiemae and Palmer, 1999, 2000). In addition to contributing to our understanding of hyoid function during feeding in mammals, this work is proving to be relevant to questions regarding the capacity for speech in Neanderthal man. Comparative data provide a context in which attempts to extrapolate from the form of the only known fossilized human hyoid (the Kebara hyoid) to its functions have been made (e.g., Reidenberg and Laitman, 1992; Arensburg et al, 1989; Arensburg and Tillier, 1991; Laitman et al, 1990; Lieberman et al, 1992). As shown in Fig. 13.9, the hyoid is linked to the symphyseal region of the mandible by the geniohyoid muscle. In some mammals, the ventral fibers of genioglossus may also reach the hyoid (e.g., rodents; Woods, 1975). The anterior belly of the digastric is usually connected to the basihyoid bone through its tendon and the fascial plane overlying the hyoid periosteum. Posterodorsally, the stylohyoid connects the basicranium to the ventral surface of the hyoid, whereas
modiolus
Dig (ab)
SupC StyPii IVlidC
InfC GHy
HyG
Trachea
RQ
StHy FIGURE 13.9. (Top) Lateral view of the head of Didelphis to show the external anatomy of adductors (temporalis and masseter) as well as buccinator (with the modiolus, see text), and the arrangement of the superficial muscles attaching to the hyoid. (Bottom) A schematic sagittal hemisection to showing the deep hyoid muscles and the muscles of the pharynx. The symphysis and hyoid are shaded. Genioglossus radiates into the body of the tongue (lightly shaded) from the genial area on the symphysis and deep to the fibers of Hyoglossus, which radiate anteriorly from their origin on the hyoid (see also Fig. 13.12). Biomechanics of the hyoid musculature in the opossum are described in Crompton et al. (1977). [Redrawn from Fig. 14-4 in Hiiemae and Crompton (1985).] Tem, temporalis; D. Mass, deep masseter; S. Mass, superficial masseter; Bucc, buccinator; Dig (ab) and Dig (pb), anterior and posterior bellies of digastric; OmHy, Omohyoid; StThy, sternothyroid; StHy, sternohyoid; StyPh, stylopharyngeus; PC, palatoglossus; HyG, hyoglossus; GHy, geniohyoid; Sup C, Mid C, and Inf C, superior, middle, and inferior constrictors of the pharynx, respectively.
posteroventrally, the sternohyoid attaches to the sternum and the omohyoid to the scapula. These muscles form an adjustable suspensory mechanism (Crompton et al, 1977). Anapol (1988) modeled the possible range of theoretical hyoid movement using values of 30% shortening over resting length for muscle contraction and 130% length for maximum stretching. He calculated theoretical "domains" within which hyoid movement could occur and then recorded actual hyoid movement in feeding, mapped it onto his predicted domains, and found that observed movement domains correspond closely to the central and posterior sagittal
432
K a r e n M. H i i e m a e
et al. (1977) for Didelphis]. During normal feeding, the hyoid is in continuous motion, although its amplitude of movement may vary among feeding stages and among taxa (Figs. 13.5 and 13.10). Importantly, the hyoid moves upward and forward during the early part of opening (Ol and 02) and then downward and backward during 0 3 and closing in opossum, hyrax, cat, and human. It is also noteworthy that, with some slight delay in the macaque, the reversal of hyoid movement from forward to backward occurs at, or just after, the start of 0 3 (see Fig. 13.5B). Two issues immediately arise: (1) why this pattern and (2) what is the muscle activity responsible, or rather, what are the functions of the jaw abductors?
areas of the predicted horizontal domain. This approach is potentially valuable for studies in other mammals. However, as Anapol points out, hyoid movement during pellet feeding in rabbits is of very low amplitude. It would be worth applying such methods to mammals feeding on more natural foods. Classically, human anatomists taught that the jaw was opened by contraction of the digastrics, with geniohyoids acting against a "fixed'' or stabilized hyoid, i.e., activity in the posterior suprahyoids and the infrahyoid muscles prevented geniohyoid (and anterior digastric) contraction from pulling the hyoid forward as the jaw opened. EMG and movement studies have shown that this view is wrong [see Crompton
-10
t%»
I
«
*
' *'
*i
.A*
•* V
GO
70
80
90
HYOYLT -H*H
1
1
H
1
H
-f—I
1
H
h-
sec FIGURE 13.10. Gape-time plot for jaw and hyoid movement in a complete feeding sequence in human (chicken spread). Hyoid movement is shown in the horizontal axis (X) and the vertical (Y) relative to the upper (HYOU prefix) or the lower (HYOL prefix) occlusal planes. The fine dotted line marks the end of stage I and the thick dotted lines the start of stage II, i.e., the period in which food is (intermittently) moved through the fauces for bolus formation (oropharyngeal bolus accumulation time or OPAT). Fine arrows indicate cycles in which stage II transport occurred: the thick arrow represents an attenuated cycle in which a major food collection and transport movement took place. Swallows are shown as shaded bars. The continuous movement of the hyoid is clearly shown. The short period of clearance (with irregular jaw movement) is indicated by the horizontal bar between the two swallows. The horizontal axis shows time in seconds. The vertical axis gives a scale for the amplitude of each movement in millimeters but does not represent the actual range of movement overall as traces have been separated for clarity. Reproduced with permission from Hiiemae and Palmer (1999).
1
\r
13. Feeding in Mammals The answers hinge on the dual role of the hyoid: it is the posterior anchor of the floor of the mouth and the mobile but rigid arch to which the hyoid muscles are attached. In addition to anteroposterior^ oriented fibers of the geniohyoid, the floor of the mouth is generally considered to be formed by obliquely posteriorly directed fibers of mylohyoid, which link the hemimandibles to each other through a median raphe and to the hyoid. Contraction of the mylohyoid can pull the hyoid forward, as does the geniohyoid, and also elevate it. This action then changes the relationship of the body of the tongue to its base, shortening it, thus affecting the orientation of the lingual muscles, genioglossus (fanning into the tongue from the symphyseal region), and hyoglossus (fanning into the tongue from the hyoid). We have argued (Hiiemae and Palmer, 2000) that this shortening of the tongue base facilitates movements of the tongue associated with food transport mechanisms (see later). It is also important to note that the observed pattern of hyoid movement explains the recordings of adductor EMG activity in jaw opening and depressor activity in closing (Weijs, 1994; Hiiemae et al, 1995; Hiiemae and Palmer, 2000). Anterior digastric, posterior digastric, and geniohyoid have multiple bursts of activity in the macaque (Sher and Hiiemae, 1998) and human (Palmer et al, 1992). Activity during SC and IP is associated with tongue movement and then the forward movement of the hyoid in IP, O l , and 0 2 . The second major burst in 0 3 depresses the jaw to maximum gape for that cycle. In the macaque, where low-level activity in the superficial masseter or the internal pterygoid has been recorded in opening, it occurs during late O l and throughout a prolonged 0 2 . This strongly suggests that these muscles are acting to restrain any tendency for the jaw to open beyond the narrow range of gape (approximately equivalent to the maximum gape during lapping) associated with the 0 2 - 0 3 transition, given powerful activity in geniohyoid. It also suggests that large protrusive transport movements of the tongue can only occur within a narrow range of gape. Although anthropoid primates have shown a general tendency for the foramen magnum to migrate from the back of the skull to its base, such that in the macaque, the oropharynx lies somewhat below the oral cavity as opposed to behind it as in opossum (Thexton and Crompton, 1998); humans are the only mammals in which the long axes of the oral cavity and oropharynx are essentially at right angles (Thexton and Crompton, 1998). This shift, coupled with facial shortening and an increase in height of the oral cavity, has resulted in (a) a hyoid complex below, rather than behind, the tongue; (b) a lengthening of the oropharynx; and (c) an increase in tongue height and reduction in
433
length (Fig. 13.13). It is generally believed that these changes are associated with changes in primate vocalizations such that humans could, biomechanically, develop these sounds into the complex phonemes of speech (Lieberman et at, 1992, and earlier papers). It has long been thought that the human uniquely long separation of the soft palate-tongue sphincter (commonly called the posterior oral seal in humans) from the laryngeal aditus required that human bolus formation occurs in the oral cavity and, that once formed, it must be explosively propelled across the oropharynx while the vocal folds were closed and the laryngeal aditus protected by a lowered epiglottis. However, as described later, bolus formation for solid foods occurs in the oropharynx (not the oral cavity) in humans. Even more intriguing are observations (manuscript in preparation) on the domains of hyoid movement during feeding and speech. Clearly, given the data in Fig. 13.11, we can draw the following inferences: (1) the hyoid moves in both feeding and speech with a smaller amplitude in the latter and (2) to produce the range of movement seen (and with minimal overlap in the two functional domains), there must be different patterns of activity in the hyoid musculature, such that geniohyoid and mylohyoid are changing length around a shorter modal length in speech, while the posterior suprahyoids must be exhibiting the reverse pattern. This raises interesting questions about the motor control system for these muscles in an activity for which no one, as yet, has suggested a CPG. It is interesting to note that in our Xray studies of human speech, in the very few cases in which a subject swallowed while reading, hyoid movements were those of a conventional swallow, suggesting a capacity for instantaneous transition between functional domains (speech to feeding). 4. Tongue The mammalian tongue is an extraordinary organ. It has at least three basic functions: food collection, intraoral food management, and as an organ subserving the "special sense" of taste. In an important paper, Kier and Smith (1985) described the tongue as a "muscular hydrostat," characterized by constant volume such that a change in one dimension produces compensatory change in another (further developed in Smith and Kier, 1989) (see Chapter 2). While some have argued that the details of their model are unproven, the principle appears to hold for all mammals with muscular tongues (Doran, 1975). A simple, if deficient, analogy for the muscular hydrostat model is to see the tongue as a water balloon: decrease its diameter by squeezing and it elongates; squash it and it
434
K a r e n M. H i i e m a e -20
2b
25
3P
35
45
m
sp
56
60
65
-25 5<
^••
-30
* Banana 11 Chicken Spread A Cookie X Speech
Hyoid X Domain in mm
FIGURE 13.11. Spatial domains of the anterosuperior angle of the hyoid during complete feeding sequences (lateral projection VFG) on soft (banana, chicken spread) and hard food (shortbread cookie fingers) and, for the same subject, when reading the "Grandfather Passage." This takes about 52 sec and includes all the essential vowel-consonant combinations in English. The plot is made by taking the Cartesian coordinates (X and Y) for the hyoid calculated with reference to the upper occlusal plane for each video field in each record and plotting all as a single figure. The total number of points in each data set is between 2500 and 3000. Not only is the domain for hyoid movement in speech smaller than that for feeding, it is also anterior. Statistical tests for significance for two-dimensional areas show that the difference is so highly significant that p > 0.0000001. Data for eight subjects (both sexes) show the same pattern and no sexual dimorphism. From Hiiemae et al. (1999, 2001).
spreads, changing shape in all three dimensions. Compensatory shape changes are required by the fact that the tongue maintains a constant volume at all times due to the incompressibility of its constituents (primarily muscle tissue). In practice, the behavior of the tongue is far more complex than this simple analogy. Most studies of tongue movement reported in Section VI,A have used implanted, radio-opaque markers to provide consistent (within subject) reference points: just under the dorsal (gustatory) epithelium with surgically inserted tongue markers (opossum, tenrec, cat, macaque, the bat Pteropus) or by gluing markers to the tongue surface (human). These studies have shown that the tongue is capable of intrinsic expansion and contraction {Pteropus, de Greet and de Vree, 1984; cat, Thexton and McGarrick, 1989; rabbit, Cortopassi and Muhl, 1990; macaque, Hiiemae et al, 1995). The tongue
can be protruded extensively and rapidly retracted (as in lapping). Shortening of the tongue base (cf. contraction of geniohyoid) facilitates extensive protrusion of the intraoral part of the tongue. Hyoid retraction and elongation of the tongue base augment its retraction. A highly elastic connective tissue tunic surrounding the tongue permits such extensive shape changes and possibly helps restore the tongue to its resting shape (Schwenkeffl/., 1989). 5. Tongue as a Sensory Organ Doran (1975) described two types of mammalian tongues: type I, whose function is primarily intraoral, found in most mammals, and type II, found in mammals such as anteaters (Chapter 15), where tongue function is primarily food gathering. All mammalian
13. Feeding in Mammals tongues have four types of papillae on their oral surface: filiform, foliate, fungiform, and circumvallate. Doran argues that the filiform and foliate papillae have primarily mechanical functions, whereas the fungiform and circumvallate papillae are primarily gustatory. Taste and smell are the primary determinants of "palatability" (gate I, process model. Fig. 13.3). Taste receptors are organized in various areas of the mouth, but famously in the gustatory surface of the tongue. Gilbertson (1998) argues that the ''taste system'' in mammals can be considered as having two vital roles: (1) serving to identify the presence of essential nutrients such as minerals (e.g., salts), carbohydrates, proteins (amino acids), and fats and (2) identifying harmful or potentially toxic substances before they are ingested. Taste transduction mechanisms have been reported for all the essential dietary constituents in mammals. Gilbertson (1998) reports that many toxic substances, such as plant alkaloids and animal venoms, have an intensely bitter taste, and without the ability to detect such substances, the consuming animal might suffer fatal consequences. He argues that, despite an approach that has historically devalued the importance of taste perception as a regulator of nutritional state, new research is giving support to the hypothesis that taste receptor cells are not merely transducers of transient chemical signals, but may play a role in the longer term management of dietary intake via nutritional cues. Similarly, smell plays a vital role in feeding. "The olfactory system is a critical component in the location, identification, and hedonic perception of food" (Wilson and Sullivan, 1998). In mammals, odor molecules must reach the olfactory receptors in the cribriform plate of the ethmoid (the interface between the cranium and nasal cavity) and pass through a thin mucus layer before reaching them. Clearly, there are two possible routes for such molecules: either directly through the nares during inspiration or via the nasopharynx by means of retronasal airflow. Wilson and Sullivan (1998) suggest a third possible route through the circulation of blood-borne odorants (e.g., garlic?). While there is general agreement that respiration is suspended during swallowing (swallowing apnea), there is little useful information in the literature on the relationship between respiration and mastication during intraoral food processing. Triggered by our observation (Palmer and Hiiemae, 1997; Fiiiemae and Palmer, 1999) that bolus formation can occur in the oropharynx in humans, we have embarked on a series of experimental studies to investigate the relationship between respiration and mastication using a nasal cannula and pressure transducer (Arvas et ah, 1999). While very preliminary, and restricted to human subjects, our results
435
indicate an oscillation in basal respiratory rate directly correlated with adductor activity in humans, i.e., jaw closing during the feeding cycle is correlated with a pressure wave passing subposteroanteriorly through the nasal cavity. One can hypothesize that food processing may release important odorants not detectable before food breakdown. If food processing releases such odor molecules, then a retronasal route for their detection may have an important biological role. Although further experimentation is needed, this preliminary finding not only suggests how odor molecules may reach the olfactory epithelium during feeding through retronasal airflow, but may also explain our surprising finding that there is no predictable complete posterior oral seal in human (see later). Given the "intranarial" position of the larynx in all but hominids, and the possibility of reverse airflow round the larynx into the nasopharynx, it is possible that the valleculae in nonhuman mammals provide an "odor transmission route" to the olfactory epithelium when the airspace in the oral cavity is reduced by jaw closure and tongue elevation, creating a pressure wave such as recorded in humans. Chemoreception issues are not usually considered by those interested in explaining the functional morphology and evolution of feeding behavior. Nevertheless, they should not be neglected. It remains the case, however, that this is an area easiest to study in humans—human subjects can be asked to eat and report on their perceptions while eating (Mioche and Martin, 1998). 6. Tongue Musculature and Tongue Function Almost all mammals have type I tongues (Doran, 1975). Such tongues are spatulate and can protrude to a maximum of 50% resting length. The tongue is anchored to the mandibular symphysis, the hyoid, and, more remotely, the skull base via so-called "extrinsic muscles" (genioglossus, hyoglossus, and styloglossus, respectively). The palatoglossus is also included in this group, although it has functions associated with the posterior oral seal and the oropharyngeal sphincter. Within the body of the tongue, there are so-called "intrinsic muscles," longitudinal, transverse, and vertical, which have their attachments within the tongue and its included aponeuroses and tendinous septa. Livingstone (1956) argues that movement of the tongue depends largely on movement of the hyoid and that change of overall tongue position results from extrinsic muscle action, whereas intrinsic muscles provide for changes in shape and tongue mobility. Studies using radio-opaque markers to map change in tongue shape show, unequivocally, that the tongue not only changes its overall position relative to the hard palate during
436
Karen M. Hiiemae
feeding as a function of change in tongue base length (see earlier discussion), but that it normatively changes its shape (expands/contracts) within its intraoral body as processing proceeds (Thexton and McGarrick, 1989, Hiiemae et al., 1995). These results strongly argue for "revisiting" the traditional view that "intrinsic tongue muscles" (longitudinal, transverse, and vertical) function "independently" of the "extrinsic" muscles. All the experimental evidence to date indicates that the tongue functions in three dimensions as an integrated whole and that its high-speed changes in shape, both intrinsic and measured relative to the hard palate or the lower jaw, are the result of complex neuromuscular integration in the hindbrain of which we have essentially no understanding. It follows that the traditional distinction between "intrinsic" and "extrinsic" muscles is an "anatomical convenience" and has little to do with physiological reality. Parallel work (Schwenk, 2000a) confirms this view. There have been few studies of the tongue muscles addressing organization, fiber type, and contraction times. Hellstrand (1980, 1981) carried out an exhaustive study of both extrinsic and intrinsic muscles in the cat using anatomical and histochemical methods and used a novel technique (light reflection) to measure contraction times. Contraction times of intrinsic and extrinsic muscles averaged 20 and 33 msec, respectively. These values fall within the range of some fast muscles of the limbs and some laryngeal and facial muscles, but are longer than that reported for digastric, although shorter than for parts of the adductors (Thexton and Hiiemae, 1975). Tongue movements, in vivo, are incredibly complex. To attempt their description, four levels of analysis are required: (1) position of the lower jaw relative to the upper jaw and palate (change in gape), (2) position of the hyoid relative to the mandibular symphysis (length and orientation of the tongue base), (3) the gross relationship of the body of the tongue to its base (A-P and vertical relationships relative to hard and soft palates), (4) the actual shape of the tongue surface (intrinsic expansion, contraction, and rotation). In normal feeding, all four levels of activity occur concurrently, although experimental evidence to date suggests that some tongue movements are not only "patterned" but occur in relation to specific phases in the jaw movement cycle. It has, so far, proved impossible to "dissect" the totality of tongue muscle activity using EMG and to precisely delineate/ascribe specific roles for any muscle in vivo, despite reliable electrode placement in the extrinsic muscles. We cannot yet model the contributions to activity producing the changes in tongue shape fundamental to its intraoral food management and transport functions. However, the insertion/attachment of radio-opaque markers below or on the gustatory epi-
thelium has allowed the measurement of movements of its surface during feeding (albeit largely in the lateral projection). These findings further suggest that a functional separation of intrinsic vs extrinsic muscles is arbitrary and may be unhelpful (see also Schwenk, 2000a). Although the intrinsic biomechanics of the tongue remain poorly understood in all mammals, its general role in feeding is now well documented. The tongue muscles are innervated by cranial nerve XII, the hypoglossal. Sensory innervation of the tongue reflects its complex embryological development [see Grays Anatomy (1995) for a detailed description]. What is becoming clear, however, is the inadequacy of our knowledge of the relationship between oral sensation and tongue movement. Although the neurophysiological literature includes some very simplistic explanations for the mechanisms of, for example, tongue tip deviation to one or the other side, we have essentially no information as to the neural circuits that govern such functionally important behaviors as intrinsic contraction and expansion. The fact that these shape changes occur primarily during the IP and opening phases of the chewing cycle, when one might expect the periodontal receptors to be progressively "off loaded," raises intriguing questions. Furthermore, such intrinsic behaviors vary with feeding stage and food condition (cat, Thexton and McGarrick, 1989; rabbit, Cortopassi and Muhl, 1990; macaque, Hiiemae et ah, 1995). The tongue is not simply an intraoral organ—it has two parts, as conventionally described: (1) within the oral cavity and (2) in the pharynx (Fig. 13.12). Physiologically, these form a functional continuum; the oral part has a fundamental role in the management of food within the oral cavity, but when the food has reached a swallowable consistency, it is moved posteriorly onto the pharyngeal surface. The spatial relationships and functional implications thereof are very different between most mammals and humans. These differences hinge on the anatomy of the oropharynx and the sphincter separating the oropharynx from the oral cavity. 7.
Oropharynx
In anatomical terms, the oropharynx is the functional space between the soft palate and tongue anteriorly, in which bolus formation occurs, and the esophageal aditus (upper esophageal sphincter) posteriorly, through which the bolus enters the digestive tract (sensu strictu). In most mammals, but not humans, the oropharynx is separated from the airway such that the tidal airflow passes directly from the trachea, through the larynx to the nasopharynx. Oropharyngeal bolus accumulation involves (1) the
13. F e e d i n g in M a m m a l s
^^
GHy
Trans M.
HyG
StyG
GHy Longitudinal muscle fibers Densely interwoven intrinsic fibers Indication of transverse and vertical intrinsic fibers i^^j,^^^^ Tongue and oropharyngeal epithelium (in contact with food).
FIGURE 13.12. Diagrammatic transverse sections of the tongue and pharyngeal regions in a 1-day-old rat (drawn from photomicrographs in Fig. 11, Smith, 1992). (Top) Anterior section through the tongue at the level of the premolar tooth germs. While the bulk of the fasciculi of hyoglossus and genioglossus are clearly identifiable, their fibers are merging with those of the "intrinsic" muscles. The central region below the gustatory epithelium consists of a dense mesh of transverse and vertical fibers originating from fibrous septa and intermingling with "intrinsic" longitudinal fibers peripherally and with those of hyoglossus and genioglossus (see Fig. 13.9 for the orientation of those fibers in the sagittal plane). (Bottom) Posterior section from the back of the tongue, just posterior to the hard palate. Hyoglossus is fanning forward into the oral part of the tongue and geniohyoid is close to its hyoid insertion. Styloglossus (not shown in Fig. 13.9) is passing into the tongue, whereas Palatoglossus is still above the oropharynx, although it will pass into the tongue to form the "pillars of the fauces" (see text and Fig. 13.13). NC, nasal cavity; OC, oral cavity; OP, oropharynx; PT, pterygoid bone; UPM, upper premolar tooth germ; LPM, lower premolar tooth germ; LI, lower incisor tooth germ; MyH, mylohyoid; Trans. M, transversus mandibulae; TVP, tensor veli palatini; StyG, styloglossus. Otherwise as listed for Fig. 13.9.
437
movement of food through the posterior oral seal/pillars of the fauces (stage II transport, see later) and (2) to be collected in the valleculae and piriform fossae, lateral to the larynx. The bolus is then propelled into the esophagus in a single "swallowing action/' The anatomy and evolution of the oropharynx have been described by Smith (1992), and its anatomy and role in swallowing by Thexton and Crompton (1998). Figure 13.13 shows its basic organization in the opossum and human. The structures involved in each species—posterior (pharyngeal) surface of the tongue, the soft palate, the pharyngeal constrictors posteriorly and laterally, and the hyolaryngeal complex anteroinferiorly—are the same, but they differ in proportion and details of arrangement. It is important to emphasize that the oropharynx is a dynamic space: the tongue, soft palate, and pharyngeal constrictors are all muscular and all move, especially during feeding. Spatial relationships can change dramatically with wide jaw opening: for example, the hyoid can be pulled down and back and, with it, the larynx and the epiglottis; when pigs vocalize, the epiglottis can have an "intraoral'' position (Herring, personal communication). Postmortem specimens, the basis for most anatomical drawings, can give a distorted impression of the functional relationships of these structures. While difficult to visualize, even with drawings such as Fig. 13.13, the pharynx in Didelphis can best be described as a tube within a tube. The inner tube consists of the larynx with the trachea below. The entry to this tube lies in the posterior nasopharynx, a condition termed the "intranarial larynx." The outer tube is formed by pharyngeal constrictors, which originate above from the cranial base, and anteriorly from the pterygomandibular raphe, as well as the hyoid and its associated cartilages (thyroid, cricoid), descending to merge with the muscular esophagus. This tube is, therefore, technically open to the oral cavity anteriorly. In practice, that opening is gated by the soft palate, above, and the oropharyngeal surface of the tongue and the epiglottis, ventrally and laterally. Dimensions of the space between then\ are regulated by the position of the soft palate, but also activity in the palatoglossus muscles, forming the palatoglossal arch and defining the "pillars of the fauces." Spaces anterior to the fauces are considered to be within the oral cavity, posterior to the fauces, within the oropharynx. In Didelphis, there is a distinct lateral separation between the outer wall of the inner tube and the inner wall of the outer. The most anterior part (walled by the tongue, soft palate, and epiglottis) forms the valleculae, whereas the more posterior (lateral to the larynx) are the piriform fossae. It follows that food moves through the fauces (stage II transport), enters the valleculae, and, as volume builds, fills the expandable
skuii base esophagus B
nasopharynx
soft palate
j^y^j^ I larynx epiglottis palatal rugae
valleculae palatoglossus
nasopharynx palatoglossus soft palate genloglossus
pharyngeal constrictors valleculae epiglottis
geniohyoid mylohyoid upper esophageal sphincter
trachea F I G U R E 13.13. Midsagittal sections showing the anatomy of the oropharynx in the opossum and modern man. (A) Diagrammatic sagittal section through the head of the opossum to show the position of the oropharyngeal complex relative to the oral cavity in a normal feeding position. The area included within the rectangle is enlarged in B. (B) The airway is shown by the thicker arrow: Air enters the nasopharynx and passes directly into the larynx, given that the epiglottis projects above the soft palate. In contrast, food accumulating to form a bolus in the valleculae passes to the esophagus by going ''round'' the larynx through the piriform fossae (dotted section fine arrow). See text and Thexton and Crompton (1998). (C) While the same anatomical elements are present (e.g., palatoglossus, which forms the "pillars of the fauces"), their spatial relationships are very different. The tongue has an expanded oropharyngeal surface (shaded), and the soft palate and epiglottis are well separated. In normal respiration, air passes through the nasopharynx, across the oropharynx, and into the larynx (heavy arrow). Three basic bolus formation and swallowing mechanisms have been described: (a) small volumes of liquids are accumulated in the oral cavity anterior to palatoglossus, which contracts to form a posterior "oral seal" between the tongue and the soft palate; the bolus is then propelled through the oropharynx and into the esophagus with the hyoid elevated and the epiglottis depressed over the laryngeal aditus; (b) semisolid and sometimes processed solid food may pass into the oropharynx, reaching the valleculae in one chewing cycle; in the immediately following cycle, a bolus is expelled from the oral cavity, "collects" the material in the valleculae, and all is swallowed; and (c) bolus accumulation occurs on the oropharyngeal surface of the tongue, with the bolus moving into the valleculae before propulsion into the esophagus (see text and Hiiemae and Palmer, 1999).
439
13. Feeding in Mammals piriform fossae from which the bolus is propelled into the esophagus during swallowing. Similar arrangements are found in cats (Thexton and McGarrick, 1988, 1989) and pigs (Herring and Scapino, 1973), as well as goats and the tenrec (see Thexton and Crompton, 1998). It n\ust be emphasized that the spatial relationships of the larynx, epiglottis, and soft palate are dynamic in all mammals. In the macaque, the larynx is still intranarial, but its position is controlled more loosely and is somewhat lower relative to the soft palate, as compared to Didelphis (Thexton and Crompton, 1998). The important difference from the opossum is the reduction in the size of the vallecular space and the narrowing of the piriform fossae, which form channels rather than spaces. These fossae do not serve as bolus accumulation sites as in Didelphis, although some material can accumulate in the valleculae during mastication. Consequently, while a small volume of triturated food can move through the fauces directly to the valleculae, the bulk of the bolus is propelled through the fauces into the oropharynx, which then "collects'' the material in the valleculae from where it is moved into the esophagus during a swallow. Ontogenetically, H. sapiens matures from a neonatal oropharyngeal anatomy comparable to that seen in the macaque to the unique morphology shown in Fig. 13.13. Much has been made of the vertical separation between the posterior oral seal and the esophageal aditus. There is no question that, uniquely among mammals, the airway and foodway share a common passage—the oropharynx—in humans. Until recently, it was assumed that this arrangement required the bolus to be accumulated in the posterior oral cavity and then propelled directly into the esophagus by a coordinated activity of tongue, soft palate, and pharyngeal musculature, as well as by hyoid and epiglottal movement. As described earlier, evidence now suggests that bolus formation in humans can occur (1) in the oropharynx, analogous to the mechanism in Didelphis; (2) in both the oropharynx and the oral cavity with fusion of the two masses to form a single bolus for swallowing, as in the macaque; and (c) in the oral cavity as traditionally described. The movement of triturated solid food into the oropharynx for bolus accumulation or deglutition is generally regarded as controlled by the posterior oral sphincter, i.e., the "gate" afforded by activity in the muscles of the soft palate and the palatoglossus muscles. The latter muscles lift the back of the tongue, ensuring a tight contact between the tongue and the soft palate to form the seal. Food meeting the "swallowable" criterion (Fig. 13.3) is allowed passage to the oropharynx. Unfortunately, there is a problem with this interpretation. There is no study of the relation-
ship between the soft palate and the tongue during feeding reported in any mammal other than humans. Furthermore, data supporting the mechanical definition of a posterior oral seal are wholly dependent on human studies (e.g., Dantas et ah, 1990). Even more provocative, work on feeding in normal human subjects (Palmer and Hiiemae, 1997; Hiiemae and Palmer, 1999) has shown that the "seal" in humans is largely inoperative during mastication of solid food; rather, there is a wide-open, two-way channel. This is not to say that H. sapiens has no seal. It is vital to the management of ingested liquids. It may well prove that these observations on humans are irrelevant to the functions of the oropharynx in nonhuman mammals. Although the larynx is allegedly intranarial in most mammals, Larson and Herring (1996) report that the epiglottis descends to cover the laryngeal aditus during swallowing in pigs and ferrets. Mention of this issue is intended to draw attention to the hazards of extrapolating data from a peculiar, but much studied, species (H. sapiens) to other mammals.
VL FEEDING FUNCTION A. T o n g u e - J a w Linkages During the feeding sequence, and in each constituent cycle, three patterns of movement are occurring, linked, but not necessarily in complete synchrony: (1) the jaws are moving through the gape cycle (FC, SC, IP, and opening), (2) the hyoid is traveling upward and forward, and downward and backward, and (3) the tongue surface is both expanding and contracting, as well as cycling in an orbit that carries its surface downward and backward, then upward and forward (Fig. 13.14). Experimental evidence suggests that the jaw movement cycle can be considered as involving two discrete processes: (1) food processing, which occurs during the SC or PS phases of closing (Fig. 13.5); and (2) intraoral food management, which occurs primarily during IP and early opening (Ol, 02). It can be argued that the FO (03) and FC phases of the jaw movement cycle are for "repositioning," i.e., to return either the jaws and teeth to a position in readiness for the next chewing stroke or the tongue to a position from which it can collect and organize the food particles in the mouth. It is important to note that tongue surface movement, especially in its anterior and middle parts (see Hiiemae et ah, 1995), often precedes movement of its posterior (oral) surface and the hyoid may "lag" behind that. The tongue is not behaving as a monolithic unit, but rather one capable of subtle, instantaneous, and localized intrinsic responses to changes in the
440
Karen M. Hiiemae
Hard food - shortbread cookie
Soft food - chicken spread
Upper Occlusal Plane ANTEfitOR
1 cm
JAW OPEN
1 cm
FIGURE 13.14. Orbits of tongue surface markers during the feeding sequence when feeding on foods of different consistency (redrawn from Fig. 4 in Palmer et. ah, 1997). Positions of the anterior tongue marker relative to the upper occlusal plane at its maximum down (TD), back (TB), upward (TU), and forward (TF) positions are shown for early (bottom orbits), middle, and late sequence cycles and for swallows (uppermost orbits). The direction of tongue marker movement is shown by arrows. As is the case for hyoid, the tongue rises as the sequence progresses. This distinct vertical progression has not been observed in other mammals (e.g., opossum, cat, macaque) where the tongue has a much longer anteroposterior dimension than vertical (see Fig. 13.13). However, the direction of movement and its temporal relationship to jaw movement are homologous with that in other mammals studied (Hiiemae and Palmer, 2000).
intraoral environment. This pattern is interrupted only when a swallow occurs. Swallows are discrete but intercalated events: the IP and early SO phases of the masticatory cycle are prolonged as the swallowing CPG is activated (see Thexton and Crompton, 1998). Experimental data for Didelphis and hyrax (see Hiiemae and Crompton, 1985) suggest that the switch from processing (jaw-based behavior) to food management (tongue-based behavior) occurs at minimum gape. Analysis of all available published data (Hiiemae and Palmer, 2000) suggests that it would be more accurate to state that the changeover occurs after the teeth have reached maximal approximation in each cycle, and always within the IP period. Although not readily evident from published, time-compressed jaw movement plots, there is no precise moment within the jaw movement cycle, recorded with CFG or VFG, that can be unequivocally identified as minimum gape, or the transition point between the two behaviors. That said, the application of computer analysis to fully digitized data for which the Cartesian coordinates of pivotal reference points have been established and manipulated does allow such a point to be calculated. Such electronic precision may create an artificially precise time point, but render the result biologically questionable. When data for macaques (Hiiemae et ah, 1995) and humans (Palmer et ah, 1997) are examined, the j a w tongue linkage is present, but with slippage. None-
theless, "turn points" in the tongue movement cycle occur predictably in relation to defined jaw movement events. Accumulating evidence shows that tongue movements may, in certain circumstances, regulate jaw movements. Lapping, for example, occurs within a very narrow range of gapes (Thexton and McGarrick, 1988, for cats). Extensive tongue protrusion appears not to occur outside that gape range. In the macaque (Hiiemae et al, 1995; unpublished data), 0 2 always occurred within a narrow gape range (macaques do not, typically, lap), but within 0 2 extensive tongue protrusion could occur. Such protrusion never occurred at wider gapes. This suggests that the biomechanics of the hyoid musculature (with contributions from the adductors, see earlier discussion) are such that significant tongue base shortening, carrying the tongue base forward so that the body of the tongue has a positional advantage, may depend on the control of gape. This hypothesis needs testing. However, there is evidence (Palmer et al, 1992, 1997; Hiiemae et al, 1996) that in the apparently unique human behavior we have termed "clearance" (Fig. 13.10), movements of the tongue in collecting and aggregating food for bolus formation occur independently of jaw movement. Therefore, it appears that the jaw-tongue linkage found in regular processing cycles can be decoupled.
13. Feeding in Mammals If jaw and tongue movements are linked in normal masticatory cycles, but can be decoupled, then, ineluctably, an important question follows. It is generally agreed that jaw movement cycles are regulated by a CPG in the hindbrain. The concomitant occurrence of rhythmic tongue movements, despite a decade in which their occurrence has been reported, has not yet been considered in the context of central nervous system control mechanisms. The currently accepted models for the control of jaw movement (see Taylor, 1990) do not seem applicable to the tongue. Such models rely on sensory feedback from the periodontal ligament receptors and the spindles in the adductor muscles. Given what we now know about masticatory behavior, to attribute complex tongue behaviors to those sensory inputs is naive. Nevertheless, as should be clear from the foregoing, the behavior of the tongue is not readily analyzed using conventional neurophysiological techniques. The contraction pattern of the tongue muscles is clearly governed by hypoglossal nerve activity. How this activity pattern is generated remains to be determined. B. Food Manipulation and Movement Five discrete processes are involved in the movement of food into the mouth, within the oral cavity, through the fauces, and then into the esophagus (see Chapter 2): (1) ingestion, i.e., placement of food into the anterior oral cavity; (2) stage I transport, i.e., movement of the ingested material from the anterior oral cavity to the cheek tooth region; (3) manipulation, i.e., the cycling of food within the postcanine region as it is reduced; (4) stage II transport, i.e., movement through the fauces to the oropharynx; and (5) deglutition {swallowing). All involve the tongue and palate. As shown in Fig. 13.3 and thoroughly described in previous papers (e.g., Hiiemae and Crompton, 1985; Thexton and McGarrick, 1988; Thexton and Crompton, 1998) and in Chapter 14, liquids are ingested, transported, and swallowed in a continuous process in which the tongue acts as a conveyor belt, moving successive small volumes from the external source to the oropharynx in a definite rhythm. In contrast, processes used for solid foods are n\ore complex and may be temporally segregated. 1. Food Acquisition and Ingestion Mammals use a wide variety of techniques to acquire their food. Many, including primitive mammals, use a forefoot or hand to pick up food items and bring them to the mouth. The squirrel gnawing at a nut held in the forepaws or the anthropoid primate tearing at a hand-held, hard-skinned fruit are using their ante-
441
rior teeth to access the digestible material. Artiodactyls have no such option. Some use the tongue to wrap round leaves and then strip them; others crop vegetation using the lower incisors against a keratinized pad, which has replaced the upper incisors. The most dramatic use of teeth and jaws in food acquisition is seen in felid carnivores, where powerful canines and adductor muscles are used to bring down prey. Any avid watcher of natural history documentaries knows that the process of food acquisition in mammals can involve a complex pattern of limb, head, lip, tongue, and jaw miovements, and even those of an elongated airway, the highly specialized elephant trunk! Moreover, a single species may use more than one method of ingestion depending on the nature of the food. For example, grizzly bears catch spawning salmon in their forepaws or in their mouth and use yet another technique when it comes to obtaining honey. Regardless of how food is primarily acquired (a topic worthy of a separate review), it is first placed in the anterior oral cavity. The only exception occurs when the cheek teeth are actually used to separate pieces (bites) of food from their matrix using the postcanines. Many carnivores do this, as does Didelphis, a behavior described as ''ingestion by mastication" (Hiiemae, 1976). Although not strictly within the scope of this chapter, attention should be drawn to the important ''ancillary" uses to which the jaw apparatus is put in many mammals. We are all familiar with the beaver dam, made from logs trimmed by powerful, gnawing incisors. We know that the canine sexual dimorphism found in many mammals, as well as in primates, has a social function and that elephant trunks are used for digging and allied activities. Such behaviors can be assumed to have influenced the evolution of craniofacial and dental anatomy. 2. Stage I Transport In most mammals for which we have data, food deposited in the front of the mouth is moved to the postcanine region using patterned movements of the tongue and jaws. The alternative, seen in carnivores, primitive mammals such as Didelphis, tenrec, and the prosimian Tupaia, is inertial transport—the jaws are opened and the head is moved rapidly forward to surround the food item (see Fiiiemae and Crompton, 1985). Stage I transport and inertial transport are not mutually exclusive, but rather are two methods of dealing with solid material. The former is always used for liquids, however. The basic mechanism of stage I transport depends on the protraction of the tongue surface below the solid bite as the jaw opens (Ol and 02), its cradling on the tongue surface, and then posterior movement as the
442
Karen M. Hiiemae
tongue retracts in jaw closing. (It is important to note that CFG/VFG evidence for this process as described here for solid foods is based on data for only three species: cat, hyrax, and macaque.) As the jaw opens again, the bite is held against the palate in the concavity formed between rugae as the tongue again moves forward beneath it. This ratchet-like, pull-back mechanism may continue for several low-amplitude cycles at the beginning of the feeding sequence (see Hiiemae and Crompton, 1985; Franks et ah, 1985; German et al, 1989). In hyrax (Franks et al, 1985), the food may not be in contact with the rugae as the tongue protracts and retracts. The same principle is used in lapping (cats, opossum, and other species), although in this case the tongue is protruded into the liquid and rapidly retracted, carrying an aliquot of fluid on its surface. The posterior three-quarters of the tongue continue to retract, carrying it distally, while the anterior quarter remains a constant length, cradling the aliquot. Figure 14-11 in Hiiemae and Crompton (1985) details this process. The number of cycles needed to move food from the incisors to the postcanines appears correlated with snout (and, therefore, tongue) length. Even in the comparatively short-snouted macaque, two or three low-amplitude cycles are used (German et al, 1989). Little attention has been paid to the possibility of stage I transport in humans, probably because the shape of the dental arcade suggested that minimal distal transport would be needed. In fact (Hiiemae et al, 1996; manuscript in preparation), a distinct transport mechanism has been observed. After ingestion, the jaw is opened wide and the tongue surface is depressed below the occlusal plane of the lower molars. The food rests on the tongue surface even though it is posteriorly heaped. The body of the tongue is then sharply retracted with the hyoid moving both downward and backward. This retraction, or pull back, occurs while the jaws are held open. As they close, the tongue begins to rise and the food is brought into contact with the molars by elevation and axial rotation of the tongue to "tip" the bite onto the working side teeth. The whole process takes about 280 msec. 3.
Manipulation
The cycling movements of the tongue described earlier (Fig. 13.14) reposition food within the oral cavity by carrying material backward and downward (FO and FC phases), forward and upward, and forward and downward (IP and SO phases). This appears to be a common pattern in all mammals studied (see Fig. 1 4 13 in Hiiemae and Crompton, 1985; Hiiemae et al, 1995; Palmer et al, 1997). In many lateral projection CFG or VFG records of
feeding sequences, occasional attenuated cycles with long opening phases are seen in which the teeth do not reach occlusion. Such a cycle is shown (heavy arrow) in Figure 13.10. These "manipulation cycles" have been interpreted as associated with intraoral food positioning or side changing. In the case of the cycle shown in Fig. 13.10, the manipulation involved a complex tongue "sweeping" movement, which collected triturated material from the anterior hard palate where it had accumulated as a function of the forward thrusts of the tongue during preceding cycles (Fig. 13.14). The intriguing questions are (1) how does the tongue reposition inadequately triturated food on the postcanine occlusal table and (2) segregate triturated material from that requiring further processing? Dorsoventral CFG or VFG records are more difficult to interpret than those taken in the lateral projection. Cortopassi and Muhl (1990) carefully examined tongue behavior in the rabbit in the D-V projection. Rabbits have an intermolar eminence on the tongue, which Ardran and Kemp (1958) had argued prevented the movement of food across the midline. Cortopassi and Muhl (1990) found that the tongue, particularly the intermolar eminence, twisted toward the working side and suggest that this movement functions to keep food between the cheek teeth. Posteroanterior projection VFG in humans also shows rotation of the tongue, as measured by the change in shape and marker position, about its anteroposterior long axis. Originally described by Malik (1955), this movement certainly repositions food onto the working side occlusal table. However, work in progress (Palmer, Hiiemae, Mioche) suggests that, at least in humans, natural sized bites of solid food are managed in an unanticipated way. After the initial one or two puncture-crushing cycles, the bite is separated into two or sometimes three "lumps." All but one, which is retained on the working side, are transferred to the balancing side and held primarily in the vestibule on that side. The tongue affects this transfer. When the lump on the working side is reduced, the triturated product is moved posteriorly for stage II transport and bolus formation. The remaining lumps are then chewed seriatim. This new finding goes a long way in explaining the pattern of intermittent swallows during feeding sequences in humans (Hiiemae et al, 1996) and provides a simple explanation for intraoral segregation of triturated and barely triturated material. There is some consensus that mammals with isognathous postcanine occlusions, e.g., rodents, could either chew on both sides of the mouth simultaneously or at least use both occlusal planes alternately (Fig. 13.8; Byrd, 1981). It may be that many anisognathous mammals eating softer foods use a combination of tongue-palate compression coupled with a chewing
443
13. Feeding in Mammals stroke on one side. The net effect, at least for food such as bananas in humans, is to spread the material to both sides of the mouth, with incidental balancing side compression as the working side teeth leave occlusion and the jaw moves medially. 4. Stage 11 Transport Liquids and triturated (swallowable) food are moved posteriorly through the fauces, either for oropharyngeal bolus formation or as the first stage of the actual swallow. Two distinct stage II transport mechanisms have been identified in mammals for which CFG or VFG data are available. In primitive mammals, e.g., Didelphis, triturated food is accumulated in a hollow on the posterior surface of the tongue, anterior to the soft palate-tongue seal. In SO (Ol, 02) the tongue expands around this aggregate, compressing it against the hard palate. As the jaws open in FO (03), the middle part of the tongue rises in front of the material, cradling it within a surface depression while carrying it downward and backward away from the hard palate. At the same time, the soft palate flattens out, opening up the fauces. As the jaws begin to close in FC, the tongue rises and continues to move backward, pushing the food through the fauces into the oropharynx. We described this process as a ''squeeze wedge" (for a detailed description, see Hiiemae and Crompton, 1985). In the opossum, several stage II transport cycles are needed to form a bolus in the valleculae and piriform fossae, with the number of cycles depending on the food. The mechanism of stage II transport in macaque and human is rather different in that it uses a different part of the tongue movement cycle to propel material through the fauces. In both anthropoid species, the forward movement of the tongue during IP and SO (Ol) creates a contact zone between the tongue surface and the hard palate, which travels backward along the tongue as the tongue surface rises and moves anteriorly (see Fig. 13.14). Because the contact zone moves progressively backward, food between the tongue and the hard palate is squeezed posteriorly along the tongue. The triturated food is then pushed through the (opened) fauces into the oropharynx. It is important to note that the description of this process in human (Hiiemae and Crompton, 1985) is only partially correct. Liquid boli are assembled within the oral cavity, propelled through the fauces, and, by a sequence of highly coordinated tongue, epiglottal, and pharyngeal mechanisms, cross the oropharynx and enter the opened esophageal aditus (Fig. 13.3). The same can occur for soft or semiliquid foods. However, when natural-sized bites of soft or hard foods are consumed, several stage II transport cycles (small ar-
rows. Fig. 13.10) transfer material to the oropharynx, where the bolus is assembled (Hiiemae and Palmer, 1999; see earlier discussion). 5. Deglutition
(Swallowing)
The major issues involved in mammalian swallowing have already been addressed (e.g., in discussion of the oropharynx), albeit parenthetically. The mechanisms involved, and their control, have been exhaustively reviewed by Thexton and Crompton (1998). In part because of the clinical significance of swallowing in human, it has received more attention from the full panoply of basic and clinical scientists than any other component of the feeding process. They have used animal and human subjects in hundreds of studies. Given this background, and the availability of a very current review, the following treatment only briefly summarizes the process. Importantly, swallow cycles in all mammals studied are "intercalated'' among processing and transport cycles (in contrast to nonmammalian tetrapods; see Chapter 2). The IP and SO (Ol) phases of the chewing cycle are prolonged, but once the bolus has started to enter the esophagus, cyclical jaw movement resumes. The only exception to this, in feeding, is the terminal swallow, when the last "bits" of food are gathered and swallowed, clearing the mouth. The jaws then return to the resting position. The actual swallow, i.e., bolus propulsion from the oropharynx (hypopharynx in human), involves (1) upward and forward movement of the hyoid, which assists in opening the upper esophageal sphincter; (2) depression of the epiglottis; (3) a propulsive, posterior movement of the oropharyngeal surface of the tongue; and (4) a peristaltic wave of contraction in the pharyngeal constrictors beginning above, and traveling toward, the esophageal aditus. In mammals for which we have data, the bolus is propelled from the valleculae and piriform fossae. In the macaque, the process may start in the oral cavity with a stage II transport movement, which continues seamlessly into the pharyngeal process. The same may be the case in human, where the movement of the soft palate to close off the nasopharynx and the depression of the epiglottis are clear indications of a swallow. When the bolus has formed in the oropharynx, the same pattern of tongue, palate, and epiglottal movement is seen. There has been discussion as to whether, in human, swallowing is a truly "active" process or is at least assisted by gravity (see Palmer, 1998). This again illustrates the dichotomy between comparative and clinically focused approaches! In almost all mammals, the orientation of the oral cavity and the oropharynx is much as shown for Didelphis in Fig. 13.13. The bolus
444
K a r e n M. H i i e m a e
has to be propelled posteriorly, and somewhat ventrally. In megachiropteran bats, the bolus has to be propelled upward, given its "upside-down" feeding posture (de Greet and de Vree, 1984). In none of these mammals could gravity assist deglutition! However, to test this assertion in humans. Palmer (1998) had human subjects feed and swallow in the normal, upright position and then on "all fours," simulating the general mammalian posture. The process was the same regardless of posture. It is not unreasonable to conclude that the highly "programmed" or CPG-controUed process of deglutition is much the same in all mammals.
VII. CONTROL OF FEEDING BEHAVIORS The feeding process involves a variety of behaviors: some rhythmic (chewing, probably gnawing), some apparently "packaged" (swallowing), and others dictated by instantaneous circumstance. Clearly, the CNS control mechanisms involved in behaviors such as the capture of live prey by felids are very different from those required for the management of gum chewing in humans—an apparently absurd comparison, but one that highlights the issues: the former behavior requires the integration of the whole body's sensorimotor systems, whereas the latter is a mechanistic, virtually subconscious, use of the oropharyngeal complex's capacity for subcortically maintained rhythmic behavior. While the neurophysiology of mastication (chewing sensu strictu) is reasonably well understood, the linkage between rhythmic tongue movement and jaw movement is not. Similarly, Jean (1984, 1990) and others have argued for the existence of a swallowing CPG, but how this would be triggered in the IP phase of a chewing cycle is not clear. The source(s) of the sensory inputs regulating the complex tongue movements involved in feeding, beyond the circuits for the special senses, is also a matter of speculation. How is swallowing triggered? Clearly, the argument, for humans, that the trigger site is pillars of the fauces cannot apply to boli, which have accumulated in the oropharynx. As for other species, we know even less. What we do know beyond any doubt is that brain stem lesions can disrupt these behaviors, especially the ability to swallow. Far more research is needed before we will have a full understanding of the central control of this vitally important process. Acknowledgments I especially thank Drs. Susan Herring and Kurt Schwenk for reading this manuscript and for their helpful advice and suggestions. Drs. Allan Thexton
and Jeffrey Palmer also reviewed earlier drafts. Much of the research reported here was funded by a series of USPHS NIH Awards. Continuing support from the National Institute for Dental Research and the National Institute for Communications Disorders and Stroke has made possible many of the experimental studies of the last 20 years. References Anapol, R (1988) Morphological and videographic study of the hyoid apparatus and its function in the rabbit (Oryctolagus cuniculus) J. Morphol. 195:141-157. Ardran, G., F. Kemp et al. (1958) A radiographic analysis of mastication and swallowing in the domestic rabbit {Oryctolagus cuniculus L.) Proc. Zool. Soc. Lond. 130:257-274. Arendsen, d. W-E. (1951) On differences in the lower jaw of animalivorous and herbivorous mammals. Proc. Koninklijke Nederlandse Akad. van Wetenschappen Ser. C 54:237-246, 405-410. Arensburg, B. (1989) The hyoid bone from the Kebara 2 hominid. Investigation in South Levantine Prehistory. O. B-Y. a. B. Vandermeersch. Oxford, BAR International Series. No. 497:337-342. Arensburg, B., and A. M. Tillier (1991) Speech and the Neanderthals. Endeavour 15:26-28. Arvas, T., K. M. Hiiemae, J. B. Palmer, and L. Mioche (1999) How mastication affects respiration. J. Dent. Res. Special Issue lADR/ AADR/CADR Meeting, March 1999, Abstract. Atchley, W. R. (1993) Genetic and developmental aspects of variability in the mammalian mandible. In: The Skull, Vol. 1. J. Hanken and B. K. Hall (eds.). University of Chicago Press, Chicago. Avis, A. (1961) The significance of the angle of the mandible: an experimental and comparative study. Am. J. Phys. Anthropol. 19: 55-61. Basmajian, J., and G. Stecko (1957) A new bipolar electrode for electromyography. J. Appl. Physiol. 17:849. Becht, G. (1953) Comparative biological-anatomical researches on mastication in some mammals. I and II. Proc. Netherlands Acad. Sci. Ser. C 56:508-527. Beecher, R. (1979) Functional significance of the mandibular symphysis. J. Morphol. 159:117-130. Bernard, C. (1858) Legons de Physiologie Experimentale. Appliquee a la Medecine. Paris, J. B. Bailliere. Birt, P, L. S. Hall, and G. C. Smith (1997) Ecomorphology of the tongues of Australian Megachiroptera (Chiroptera: Pteropodidae). Aust. J. Zool. 45:369-384. Bramble, D. (1980) Feeding in tortoises and mammals: why so similar? Am. Zool. 20:931. Bramble, D., and D. Wake (1985) Feeding mechanisms of lower vertebrates. Pp. 230-261. In: Functional Vertebrate Morphology. M. Hildebrand, D. Bramble, D. Wake, and K. Liem (eds.). Belknapp Press of Harvard Univ. Press, Cambridge, MA. Bramble, D. M. (1978) Origin of the mammalian feeding complex: models and mechanisms. Paleobiology 4:271-301. Bremer, F. (1923) Physiologie nerveuse de la mastication chez la chat et lapin. Arch. Intl. Physiol. 21:309-352. Brylski, P., and J. Patton (1988) Ontogeny of a macroevolutionary phenotype: the external cheek pouches of geomyoid rodents. Evolution 42:391-394 Butler, P. (1952) Some functional aspects of molar evolution. Evolution 26:474-483. Byrd, K. (1981) Mandibular movement and muscle activity during
13. F e e d i n g in M a m m a l s mastication in the guinea pig (Cavia porcellus). J. Morphol. 170: 147-169. Clark, W. Le. Gros. (1959) The Antecedents of Man. The University Press, Edinburgh. Cortopassi, D., and Z. Muhl (1990) Videofluorographic analysis of tongue movement in the rabbit (Oryctolagus cuniculus). J. Morphol. 204:139-146. Crompton, A. (1971) The origin of the tribosphenic molar. In Early Mammals. D. Kermack and K. Kermack (eds.). Academic Press, London. Crompton, A. (1995) Masticatory function in nonmammalian cynodonts and early mammals. Pp 55-75. Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge, Cambridge Univ. Press, Cambridge. Crompton, A., and K. Hiiemae (1969) Functional occlusion in tribosphenic molars. Nature 222:678-679. Crompton, A., and K. Hiiemae (1970) Molar occlusion and mandibular movements during occlusion, Didelphis marsupialis L. Zool. J. Linn. Soc. 49:21-47. Crompton, A., A. Thexton et al. (1977) The activity of the hyoid and jaw muscles during chewing of soft food in the opossum. Pp. 287-305. The Biology of Marsupials, Vol. 2. B. Stonehouse and D. Gillmore, MacMillan, London. Dantas, R., W. Dodds et al. (1990) Manometric characteristics of the glossopalatal sphincter. Digest. Dis. Sci. 35:161-166. Davis, D. D. (1964) The giant panda: a morphological study of evolutionary mechanisms. Fieldiana: Zool. Mem. 3:1-339. Dellow, P., and J. Lund (1971) Evidence for the central timing of rhythmical mastication. J. Physiol. (Lond.) 215:1-13. De Vree, P., and C. Cans (1976) Mastication in pygmy goats {Capra hircus). Ann. Roy. Belg. Zool. Soc. 105:255-306. Dobson, G. (1882) On the digastric muscle, its modifications and functions. Trans. Linn. Soc. Lond. Zool. Ser. 2:259-264. Doran, G. (1975) Review of the evolution and phylogeny of the mammalian tongue. Acta Anat. 91:118-129. Doran, G., and H. Baggett (1971) A structural and functional classification of mammalian tongues. J. Mammal. 52:427-429. Edgeworth, R (1911) On the morphology of the cranial muscles in some vertebrates. Quart. J. Microsc. Sci. 56:167-316 (N). Edgeworth, F. (1914) On the development and morphology of the mandibular and hyoid muscles of mammals. Quart. J. Microsc. Sci. 59:573-645 (N). Ewer, R. (1958) Adaptive features of the skulls of African Suidae. Proc. Zool. Soc. Lond. 131:135-155. Fearnhead, R. W., C. C. D. Shute et al. (1955) The temporo-mandibular joint of shrews. Proc. Zool. Soc. Lond. 125:795-806. Fortelius, M. (1985) Ungulate cheek teeth: developmental, functional and evolutionary interrelations. Acta Zool. Perm. 180:1-76. Franks, H. A., A. W. Crompton, and R. Z. German (1984) Mechanism of intraoral food transport in macaques. Am. J. Phys. Anthro. 65: 275-282. Franks, H. A., R. Z. German, A. W. Crompton, and K. M. Hiiemae (1985) Mechanism of intra-oral transport in a herbivore, the hyrax {Procavia syriacus). Arch. Oral Biol. 30:539-544. Garrett, J., and G. Proctor (1998) Control of salivation. Pp. 135-155. In: The Scientific Basis of Eating. R. W. A. Linden (ed.). Karger, Basel. German, R., and H. Franks (1991) Timing in the movements of the jaws, tongue and hyoid during feeding in the hyrax {Procavia syriacus). J. Exp. Zool. 257:34-42. German, R., S. Saxe, A. W. Crompton, and K. M. Hiiemae (1989) Mechanism of food movement through the anterior oral cavity in anthropoid primates. Am. J. Phys. Anthropol. 80:765-775. Gilbertson, T. (1998) Peripheral mechanisms of taste. Pp. 1-28. In: The Scientific Basis of Eating. R. W. A. Linden (ed.). Karger, Basel.
445
Gorniak, G. (1977) Feeding in golden hamsters, Mesocritus auratus. J. Morphol. 154:427-458. Grays Anatomy (1995) Churchill Livingstone, London. Greaves, W. (1974) Functional implications of mammalian jaw joint position. Form. Funct. 7:363-376. Greaves, W. (1978) The jaw lever system in ungulates: a new model. J. Zool. 184:271-278. Greaves, W. (1988) The maximum average bite force for a given jaw length. J. Zool. 214:295-306. Greaves, W. (1995) Functional predictions from theoretical models of the skull and jaws in reptiles and mammals. Pp. 99-115. In: Functional Morphology in Vertebrate Paleontology. P. Thomason (ed.). Cambridge Univ. Press, Cambridge. Gueldre, G. de., and F. de. Vree (1984) Movements of the mandible and tongue during mastication and swallowing in Pteropus giganteus (Megachiroptera). J. Morphol. 179:95-114. Hamlett, S. (1989) Dynamic aspects of lingual propulsive activity in swallowing. Dysphagia 4:136-145. Hart, R. T. V, V. V. Hennebel, N. Throngpreda, W. C. van Buskirk, and R. C. Anderson (1992) Modelling the biomechanics of the mandible: a three dimensional finite element study. J. Biomech. 25: 262-268. Heath, M. (1991) The basic mechanics of mastication: man's adaptive success. Pp. 144-166. In: Feeding and the Texture of Food. J. Vincent and P. Lillford (eds.) Cambridge Univ. Press, Cambridge. Hector, M., and R. Linden (1987) The possible roles of periodontal mechanoreceptors in the control of parotid secretion in man. Quart. J. Exp. Physiol. 72:285-301. Hedges, S. B., P H. Parker, C. G. Sibley, and S. Kumar (1996) Continental breakup and the ordinal diversification of birds and mammals. Nature 381:226-229 Hellstrand, E. (1980) Morphological and histochemical properties of tongue muscles in cat. Acta Physiol. Scand. 110:187-198. Hellstrand, E. (1981) Contraction tinies of the cat's tongue muscles measured by light reflection: innervation of individual tongue muscles. Acta Physiol. Scand. Ill:417-423. Herring, S. (1992) Muscles of mastication: architecture and functional organisation. Pp. 541-548. In: The Biological Mechanisms of Tooth Movement and Craniofacial Adaptation. Z. Davidovitch (ed.). Ohio State University, Columbus, OH. Herring, S. W. (1993) Functional morphology of mammalian mastication. Am. Zool. 33:289-299. Herring, S. (1994) Functional properties of the feeding musculature. Pp. 5-30. In: Biomechanics of Feeding in Vertebrates, Vol. 18. V. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Herring, S. W., and R. P. Scapino (1973) Physiology of feeding in miniature pigs. J. Morphol. 141:427-460. Hiiemae, K. (1967) Masticatory function in mammals. J. Dent. Res. 46:883-893. Hiiemae, K. (1976) Masticatory movements in primitive mammals. Pp. 105-117. In: Mastication. D. Anderson and B. Matthew (eds.). John Wright, Bristol. Hiiemae, K. (1978) Mammalian mastication: a review of the activity of the jaw muscles and the movements they produce in chewing. Pp. 359-398. In: Development, Function and Evolution of Teeth. P. Butler and K. Joysey (eds.). Academic Press, London. Hiiemae, K. (1993) Process and mechanism: mechanoreceptors in the mouth as primary modulators of rhythmic behaviors in feeding? Pp. 263-284. In: Sensory Research: Multimodal Perspectives. R. Verillo (ed.). Lawrence Erlbaum, Hillsdale, NJ. Hiiemae, K., and G. Ardran (1968) Mandibular movement in the rat: a cineradiographic study of mastication in Rattus norvegicus (L). J. Zool. 154:139-154. Hiiemae, K., and A. Crompton (1971) A cinefluorographic study of feeding in in the American opossum (Didelphiis marsupialis L.)
446
K a r e n M. H i i e m a e
Pp. 299-334. In: Dental Morphol and Evolution. A. Dahlberg (ed.). University of Chicago Press, Chicago. Hiiemae, K., and A. Crompton (1985) Mastication, food transport and swallowing. Pp. 262-290. In: Functional Vertebrate Morphology. M. Hildebrand, D. Bramble, K. Liem, and D. Wake (eds.). Belknapp Press, Harvard Univ. Press, Cambridge MA. Hiiemae, K., M. R. Heath, G. Heath, E. Kazazoglu, J. Murrray, D. Sapper, and K. Hamblett (1996) Natural bites, food consistency and feeding behavior in man. Arch. Oral Biol. 41:175-189. Hiiemae, K., and F. Jenkins (1969) The anatomy and internal architecture of the muscles of mastication in Didelphis marsupialis L. Postilla 140:1-49. Hiiemae, K., and R. Kay (1972) Trends in the evolution of primate mastication. Nature 240:486-487. Hiiemae, K., A. Reese, and S. Hayenga (1995) Patterns of tongue and jaw movement: a cinefluorographic study of feeding in the macaque. Arch. Oral Biol. 40:229-246. Hiiemae, K. M., and J. B. Palmer (1999) Food transport and bolus formation during complete feeding sequences on foods of different initial consistency. Dysphagia 14:31-42. Hiiemae, K. M., and J. B. Palmer (2000) Tongue-jaw linkages in feeding. Bull. Mus. Comp. Zool. Harvard University. Hiiemae, K. M., A. J. Thexton, and A. W. Crompton (1978) Intra-oral food transport: a fundamental mechanism of feeding? Pp. 181208. In: Muscle Adaptation in the Craniofacial Region, Vol. 8. D. Carlson and J. McNamara (eds.). University of Michigan, Ann Arbor. Hiiemae, K. M., A. J. Thexton et al. (1981) The movement of the cat hyoid during feeding. Arch. Oral Biol. 26:65-81. Hiiemae, K. M., S. W. Medicis, J. B. Palmer, B. S. Jackson, and J. Hegener (1999) How we can eat and speak: hyoid movements in feeding and speech are different. Am. J. Phys. Anthropol. Suppl. 28:151. (Abstract) Hilloowala, R. (1975) Comparative anatomical study of the hyoid apparatus in selected primates. Am. J. Anat. 142:367-384. Honeycutt, R. L., and R. M. Adkins (1993) Higher level systematics of eutherian mammals: an assessment of molecular characters and phylogenetic hypotheses. Annu. Rev. Ecol. System. 24:279305. Hylander, W. (1977) In vivo bone strain in the mandible of Galago crassicaudatus. Am. J. Phys. Anthropol. 46:309-326. Hylander, W. (1979) Mandibular function in Galago crassicaudatus and Macaca fascicularis: an in vivo approach to stress analysis of the mandible. J. Morphol. 159:253-296. Hylander, W. (1984) Stress and strain in the mandibular symphysis of primates: a test of competing hypotheses. Am. J. Phys. Anthropol. 64:1-46. Hylander, W. L., K. R. Johnson, and A. W. Crompton (1992) Muscle force recruitment and biomechanical modelling: an analysis of masseter muscle function during mastication in Macaca fascicularis. Am. J. Phys. Anthropol. 72:287-314 James, W. (1960) The Jaws and Teeth of Primates. Pitman Medical Publishing, London. Janis, C. (1979) Mastication in the hyrax and its relevance to ungulate dental evolution. Paleobiology 5:50-59. Janis, C. (1995) Correlations between craniodental morphology and feeding behavior in ungulates: reciprocal illumination between living and fossil taxa. Pp. 76-98. In: Functional Morphology in Vertebrate Paleontology. T. Thomason (ed.). Cambridge Univ. Press,
Jean, A. (1990) Brainstem control of swallowing. Pp. 294-321. In: Neurophysiology of the Jaws and Teeth. A. Taylor (ed.). MacMillan, London. Jeronimidis, G. (1991) Mechanical and fracture properties of cellular and fibrous materials. Pp. 1-17. In: Feeding and the Texture of Food. J. Vincent and P. Lillford (eds.). Cambridge University Press, Cambridge. Jongh, H. de., R. Dantuma, and H. J. M. Sluijsmans (1989) The shape of the mandible in the domestic sheep: a biomechanical analysis using EMG as an estimator of muscle force. Acta Morphol. NeerlScand. 27:63-73. Kallen, F , and C. Cans (1972) Mastication in the little brown bat, Myotis lucifugus. J. Morphol. 136:385-420. Kay, R. (1975) The functional adaptations of primate molar teeth. Am. J. Phys. Anthropol. 43:195-216. Kay, R., and K. Hiiemae (1974) Jaw movement and tooth use in recent and fossil primates. Am. J. Phys. Anthropol. 40:227-256. Kay, R., and W. Hylander (1978) The dental structure of mammalian folivores with special reference to Primates and Phalangeroidea. Pp. 173-191. In: The Ecology of Arboreal Folivores. G. Montgomery (ed.). Smithsonian Institution Press, Washington DC. Kier, W., and K. Smith (1985) Tongues, tentacles and trunks: the biomechanics of movement in muscular hydrostats. Zool. J. Linn. Soc. 83:307-324. Laitman, J., J. Reidenberg et al. (1990) The Kebara hyoid: what can it tell us about the evolution of the hominid vocal tract? Am. J. Phys. Anthropol. 81:254. Larson, J., and S. Herring (1996) Movement of the epiglottis in mammals. Am. J. Phys. Anthropol. 100:71-82. Lavigne, G., J. Kim et al. (1987) Evidence that periodontal pressoreceptors provide positive feedback to jaw closing muscles during changes in the pattern brought about by objects between the teeth. J. Neurophysiol. 58:342-358. Lieberman, P. (1992) On Neanderthal speech and Neanderthal extinction. Curr. Anthropol. 33:409-410. Lieberman, P., J. Laitman et al. (1992) The anatomy, physiology, acoustics and perception of speech: essential elements in analysis of the evolution of human speech. J. Hum. Evol. 23:447-467. Livingstone, R. (1956) Some observations on the natural history of the tongue. Ann. Roy. College Surg. Engl. 19:185-200. Lord, F. (1913) Observations on the temporo-mandibular articulation. Anat. Rec. 7:355-367. Lucas, P. (1982) Basic principles of tooth design. Pp. 154-162. In: Teeth: Form, Function and Evolution. B. Kurten (ed.). Columbia Univ. Press, New York. Lucas, P., and R. Corlett (1991) Quantitative aspects of the relationship between dentitions and diets. Pp. 93-121. In: Feeding and the Texture of Food, Vol. 44. J. Vincent and P. Lillford (eds.). Cambridge Univ. Press, Cambridge. Lucas, P., and D. Luke (1984) Optimum mouthful for food comminution in human mastication. Arch. Oral Biol. 29:205-210. Ludwig, C. (1850) Neue Versuche iiber die Beihiilfe der Nerven zu der Speichelsekreiten. Naturforsch Ges Zurich 53/54:210-239. Magendie, F. (1825) Precis Elementaire de Physiologic. MequignonMarvis, Paris. Malik, A. E. (1955) The part played by the tongue in mastication and deglutition. J. Anat. 89:250-254. Maynard Smith, J., and R. Savage (1959) The mechanics of mamma-
13. F e e d i n g in M a m m a l s Mills, J. (1966) The functional occlusion of the teeth of Insectivora. J. Linn. Soc. 46:1-25. Mioche, L., and J. Martin (1998) Training and sensory judgement effects on mastication as studied by electromyography. J. Food Sci. 63:1-5. Naples, V. (1985) Form and function of the masticatory musculatures in the tree sloths, Bradypus and Choloepus. J. Morphol. 183:25-50. Naples, V. (1986) The morphology and function of the hyoid region in the tree sloths, Bradypus and Choloepus. J. Mammal. 67:712724. Novacek, M. (1992) Mammalian phylogeny: shaking the tree. Nature 356:121-125 Orchardson, R., and S. Cadden (1998) Mastication. Pp. 76-121. In: The Scientific Basis of Eating. R. W. A. Linden (ed.). Karger, Basel. Oron, U., and A. Crompton (1985) A cineradiographic and electromyographic study of mastication in Tenrec ecaudatus. J. Morphol. 185:155-182. Palmer, J. (1998) Bolus aggregation in the oropharynx does not depend on gravity. Arch. Phvs. Med. Rehabilitation 79:691-696. Palmer, J., K. M. Hiiemae, and J. Lui (1997) Tongue-jaw linkages in feeding: a preliminary videofluorographic study. Arch. Oral Biol. 42: 429- 441. Palmer, J., N. Rudin, G. Lara, and A. W. Crompton (1992) Coordination of mastication and swallowing. Dysphagia 7:187-200. Pavlov, I. (1910) Lectures on the Work qf the Digestive Glands. Charles Griffen, London. Peyer, B. (1968) Comparative Odontology. University of Chicago Press, Chicago. Prinz, J., and P. Lucas (1997) An optimization model for mastication and swallowing in mammals. Proc. Roy. Soc. Lond. Ser. B 264: 1715-1721. Purslow, P. (1991) Measuring meat texture and understanding its structural basis. Pp. 35-56. In: Feeding and the Texture qf Food. J. Vincent and P. Lillford (eds.). Cambridge Univ. Press, Cambridge. Radinsky, L. (1981a) Evolution of skull shape in carnivores. 1. Representative modern carnivores. Biol. J. Linn. Soc. 15:369-388. Radinsky, L. (1981b) Evolution of skull shape in carnivores. 2. Additional modern carnivores. Biol. J. Linn. Soc. 16:337-355. Radinsky, L. (1982) Evolution of skull shape in carnivores. 3. The origin and early radiation of modern carnivores. Paleobiology 8: 177-195. Reidenberg, J., and J. Laitman (1994) Anatomy of the hyoid apparatus in Odentoceti (toothed whales): specializations of their skeleton and musculature compared with those of terrestrial mammals. Anat. Rec. 240: 598- 624. Rensberger, J. (1973) An occlusal model for mastication and dental wear in herbivorous mammals. J. Paleontol. 47:515-528. Rensberger, J. (1978) Scanning electron microscopy of wear and occlusal events in some small herbivores. Pp. 415-438. In: Development, Function and Evolution of Teeth. P. Butler and K. Joysey (eds.). Academic Press, London. Rensberger, J. (1995) Determination of stresses in mammalian dental enamel and their relevance to the interpretation of feeding behaviors in extinct taxa. Pp. 151-172. In: FunctionaI Morphol. in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Romer, A. (1974) Vertebrate Paleontology. University of Chicago Press, Chicago. Scapino, R. (1.972) Adaptive radiation of mammalian jaws. Pp. 3339. In: Morphology of the Maxillo-Mandibu]ar Apparatus. G. Schumacher (ed.). Leipzig, G. Thieme. IX International Congress of Anatomists. Schwenk, K. (2000a) Intrinsic versus extrinsic lingual muscles: a false dichotomy? Bull. Mus. Comp. Zool. Harvard Univ.
447
Schwenk, K. (2000b) Functional units and their evolution. In: The Character Concept in Evolutionary Biology. G. P. Wagner (ed.). Academic Press, San Diego. Schwenk, K., E. Brainerd, and A. W. Crompton (1989) Biomechanics of the mammal tongue: role of the lingual tunic during hydrostatic deformation. Am. Zoot. 29:35A. Sher L., K. M. Hiiemae, and J. B. Palmer (1997) Activity in the suprahyoid muscles of Macaca fascicularis during normal feeding. J. Dent. Res. Special Issue 77. (Abstract # 1512) Sibbing, F. (1991) Food processing by mastication in cyprinid fish. Pp. 57-92. In: Feeding and the Texture of Food. J. Vincent and P. Lillford (eds.). Cambridge Univ. Press, Cambridge. Smith, K. (1978) Electromyographic and cineradiographic studies of feeding in lizards. Am. Zool. 18 : 623. Smith, K. (1992) The evolution of the mammalian pharynx. Zool. J. Linn. Soc. 104: 313- 349. Smith, K. (1994) Are neuromotor systems conserved in evolution? Brain Behav. Evol. 43:293-305. Smith, K. K., and W. M. Kier (1989) Trunks, tongues and tentacles: moving with skeletons of muscle. Am. Sci. 77:29-35. Spears, I., and G. Macho (1998) Functional adaptations of modern human molars: implications for the fossil record. Am. J. Phys. Anthropol. 106: 467-482. Springer, M. S., G. C. Cleven, O. Madsen, W. W. de Jong, V. G. Waddell, H. M. Amrine, and M. J. Stanhope (1997) Endemic African mammals shake the phylogenetic tree. Nature 388:61-64. Taylor, A. (1990) Proprioceptive control of jaw movement. Pp. 237267. In: Neurophysiology of the Jaws and Teeth. A. Taylor (ed.). Macmillan, London. Teaford, M., and J. Runestad (1992) Dental microwear and diet in Venezuelan primates. Am. J. Phys. Anthropol. 88:347-364. Teaford, M., and A. Walker (1984) Quantitative differences in dental microwear between primate species with different diets and a comment on the presumed diet of Sivapithecus. Am. J. Phys. Anthropol. 64:191-200. Thexton, A. J. (1982) Mastication and swallowing: an overview. Br. Dent. J. 173:197-206. Thexton, A., and A. Crompton (1989) Effect of sensory input from the tongue on jaw movement in normal feeding in the opossum. J. Exp. Zool. 250:233-243. Thexton, A., and A. Crompton (1998) The control of swallowing. Pp. 168-222. In: The Scientific Basis of Eating. R. W. A. Linden (ed.). Karger, Basel. Thexton, A., and K. Hiiemae (1975) The twitch-tension characteristics of opossum jaw musculature. Arch. Oral Biol. 20:743-748. Thexton, A., and K. Hiiemae (1997) The effect of food consistency on jaw movement in the macaque: a cineradiographic study. J. Dent. Res. 76: 552-560. Thexton, A., K. Hiiemae et aI. (1980) Food consistency and bite size as regulators of jaw movement during feeding in the cat. J. Neurophysiol. 44: 456-474. Thexton, A., and J. McGarrick (1988) Tongue movement of the cat during lapping. Arch. Oral Biol. 33:331-339. Thexton, A., and J. D. McGarrick (1989) Tongue movement in the cat during the intake of solid food. Arch. Oral Biol. 34:239-248. Thexton, A., J. McGarrick, K. M. Hiiemae, and A. W. Crompton (1982) Hyo-mandibular relationships during feeding in the cat. Arch. Oral Biol. 27:793-801. Turnbull, W. (1970) Mammalian masticatory apparatus. Fieldiana: Geology 18(2): 153-356. van der Klaauw, C. J. (1945) Cerebral skull and facial skull: a contribution to the knowledge of skull-structure. Arch. Neerl. Zool. 7: 16-37. Vincent, J. (1991) Textures of plants and fruits. Pp. 19-33. In: Feeding
448
Karen M. Hiiemae
and the Texture of Foods. J. Vincent and P. Lillford (eds.). Cambridge Univ. Press, Cambridge. von Koenigswald, W. (1982) Zum verstandnis der Morphologie der Wiihlmausmolaren (Arvicolae, Rodentia, Mammalia). Zool. Geol. Wittenschaft Berlin 10:951-852. Weijs, W. (1994) Evolutionary approach of masticatory motor patterns in mammals. In: Biomechanics of Feeding in Vertebrates. V. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Weijs, W., and R. Dantuma (1975) Electromyography and mechanics of mastication in the albino rat. J. Morphol. 146:1-34.
Weijs, W., and R. Dantuma (1981) Functional anatomy of the masticatory apparatus in the rabbit {Oryctolagus cuniculus L.). Netherlands J. Zool. 31:99-147. Wilson, D., and R. Sullivan (1998) Peripheral mechanisms of smell. Pp. 29-39. In: The Scientific Basis of Eating, Vol. 9. R. W. A. Linden (ed.). Karger, Basal. Woods, C. (1975) The hyoid, laryngeal and pharyngeal regions of bathygerid and other selected rodents. J. Morphol. 147:229250.
C H A P T E R
14 The Ontogeny of Feeding in Mammals A. W. CROMPTON Museum of Comparative Zoology Harvard University Cambridge, Massachusetts 02138
R. Z. GERMAN Department of Biological Sciences University of Cincinnati Cincinnati, Ohio 45221
the laryngeal opening, both a bony hard and a muscular soft palate, and a large muscular tongue capable of complex motions. However, the significant differences in anatomy and the constraints suggest that the mechanisms used for feeding in infants will differ from those of adults (Ardran et al, 1958). Despite the differences between infant and adult feeding, and the significant variation in adult feeding mechanisms, there are many similarities in suckling across species of mammals. The transition to drinking from sources other than a teat and eating solid food, termed weaning, occurs at different points of development for different species, but presents similar functional challenges to the young animal. While there is a great deal known about suckling in infant humans from a clinical perspective (Logemann, 1983; Koenig et al, 1990; Jones and Conner, 1991), much less is known about variation among species of mammals, and even less about the evolution of suckling. This chapter begins with a review of the relevant morphology and then discusses what is known about the mechanics and function in a comparative framework. The next sections focuses on three related issues: the rythmicity of suckling; the coordination of respiration and swallowing in infants; and the maturation of feeding and the transition to eating solid food. The chapter concludes with some brief speculation on the evolution of suckling in mammals.
I. INTRODUCTION 11. MORPHOLOGY III. FUNCTION AND MECHANICS OF SUCKLING A. General Features B. Suckling in Infant Pigs C. Suckling in Infant Macaques D. Suckling in Infant Opossums IV. RHYTHMICITY AND CONTROL OF SUCKLING V. COORDINATION OF SWALLOWING AND RESPIRATION VI. TRANSITION FROM SUCKLING TO DRINKING AT WEANING VII. EVOLUTIONARY CONSIDERATIONS References
L INTRODUCTION One of the defining characteristics of mammals is the presence of specialized mammary glands in females that produce milk for feeding infants (Eisenberg, 1981). Infant mammals, including all marsupials and placentals, suckle on the maternal teat for some period of time after birth. Several constraints on suckling make it different from adult feeding: delivery from the teat and a purely liquid food source. These functional differences are associated with differences in infant anatomy, including lack of teeth, relative size of the tongue, and pharyngeal morphology (Bosma, 1985; Crelin, 1987; German et al, 1992). Both functional and morphological differences place constraints on feeding. Some aspects of basic mammalian oropharyngeal anatomy are constant through development, including relative posterior position of the esophageal opening to
FEEDING (K. Schwenk, ed.)
IL MORPHOLOGY Infant mammals are born at varying degrees of maturity, but all have the ability to suckle within hours, if
449
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
450
R. Z. German and A. W. Crompton
not minutes, of birth. Altricial animals, including marsupials, carnivores, and many rodents, are incapable of even immature locomotion and their eyes are not open. More precocious mammals, including ungulates, are alert and capable of locomotion immediately. However, most infant mammals are edentulous. Some taxa are born with teeth that are not functional for feeding. Infant pigs, for example, have "needle teeth" that are used for fighting with littermates (Fraser and Thompson, 1991). All infant mammals are characterized by heads and tongues that are large relative to adult proportions. The structure of the oral cavity and oropharynx differs from adults mainly in relative proportion, with a few additional differences. The viscerocranium, relative to the neurocranium, is smaller in infants than in adults. The face tends to be shorter in the anterior/ posterior dimension and the tongue larger (DuBrul, 1988). Rugae found on the roof of the mouth, critical for adult feeding, are either reduced or nonexistent in infants. Finally, the relative position of the airway to the esophagus is different in young mammals. In all infants the epiglottis is relatively high and is located internarially, i.e., reaching into the nasopharynx (Crelin, 1987; Crompton et ah, 1997). This provides a continuous, patent airway from the nares through to the larynx. This arrangement has significance for the relationship between respiration and swallowing, and whether the two processes can occur simultaneously in infants, but not adults.
III. FUNCTION A N D MECHANICS OF SUCKLING A. General Features The basic mechanism of suckling, in all infant mammals studied to date, is a pumping mechanism using dorsoventral movements of the tongue body. There is some debate concerning this mechanism. Early cineradiographic studies by Ardran and colleagues (1958) on infant lambs suggest that milk is obtained by expression, mechanical pressure on the teat by the tongue, and is accompanied by strong jaw movements. Ardran and Kemp (1959) also measured pressure changes in the nipple and oral cavity during suckling in infant lambs and humans to support their hypothesis that suckling was accomplished by positive pressure. They considered that suction, or relative negative pressure, was important only for holding the nipple in the oral cavity and refilling the artificial teat. One difficulty with these early results may have been a reliance on the outline of the milk containing barium to characterize
the tongue position; because of intense longitudinal furrowing, the outline corresponding to the middle of the tongue can be confused with the outline of the raised edges in the absence of markers (Thexton and McGarrick, 1988; German et al, 1992). More recent studies describe a different mechanism of suckling, but one that is consistent across a number of different species. For the initial acquisition of milk, similar mechanisms were found in marsupials (German and Crompton, 1996), ungulates and primates (German et al, 1992), and carnivores (unpublished data). In each of these species, suckling was characterized by a number of distinctive kinematic patterns of tongue movement. In all cases, the anterior portions of the tongue moved very little. Usually, the tip of the tongue was wrapped around the nipple. The middle of the tongue, lying beneath the hard palate and hard palate/soft palate junction, moved in a dorsal/ventral pathway, with little or no anterior/posterior movement. Posterior portions of the tongue tended to move in larger circular patterns. The orientation of this portion of the tongue is not anterior/posterior, but slopes downward. Thus, tongue movements are still orthogonal to the direction of milk flow. B. Suckling i n Infant Pigs The species with the most complete description of suckling is the domestic pig {Sus scrofa). Infant pigs weighed approximately 1.0-2.0 kg during the time these data were collected. This species uses a two-step mechanism, with each step involving pumping (German et al, 1992). In the first step, milk is pumped out of the nipple and into the oral cavity. The second step transports milk from the oral cavity and the oral pharynx and into the esophagus. Initially, the infant would wrap its tongue around the tip of the nipple, and the tip of the tongue curled around the nipple was visible outside the oral cavity. The tongue sealed against the hard palate and the anterior end of the soft palate (Fig. 14.1, frame 78). A small space existed in the anterior oral cavity, immediately posterior to the nipple. Between frames 80 and 87 the jaw opened slightly (Fig. 14.2), and the space between the tongue and the hard palate expanded as follows: (a) the middle of the tongue, near marker three, moved strongly downward and (b) the posterior contact between tongue and soft palate, as well as the positions of markers four and five, shifted in a posterior direction. These movements leave two seals, or potential seals, between tongue and palate: (i) an anterior seal between the hard palate and the anterior tongue and (ii) a posterior seal between the soft palate and the posterior tongue. Initially there was an increase in the space between the tongue and that
14. T h e O n t o g e n y of F e e d i n g in M a m m a l s
I^PPIQ
451
appearance on X-ray, suggesting air mixed in with the milk. By the time the jaw began to close and marker one (lying underneath the nipple) began to rise, most of the space between palate and tongue was filled with milk, and little, if any, more was seen to enter the cavity (Fig. 14.1, frame 92). The tongue reformed the anterior seal with the hard palate, just posterior to the nipple. The second stage began as the jaw opened between frames 92 and 98, and the middle of the tongue, near marker three, rose to contact the hard palate (Fig. 14.1, frame 98). The posterior tongue, near markers four and five, moved anteriorly and ventrally or downward (Fig. 14.2). This broke the posterior seal and formed a large space beneath the soft palate. At this time, the aliquot of liquid, which had been lying in the space between the anterior and the posterior seals underneath the hard palate, was moved through the pillars of the fauces and into the vallecular space above markers four and five (Fig. 14.1, frame 98). The jaw continued to open between frames 98 and 107 (Fig. 14.2). The posterior tongue markers moved upward and backward and the liquid moved out of the valleculae. By frame 113, the liquid had passed through the piriform recesses and into the esophagus. These swallows were regular, occurring in every second cycle. Rarely, a swallow occurred in the third cycle, but this only happened when little or no milk was acquired in one of the preceding cycles. C. Suckling in Infant Macaques
FIGURE 14.1. Lateral view of infant pig suckling by pumping mechanism. Each of five frames are taken from cine X-ray film of an infant pig sucking. Filming was at 100 fps, so each frame was 10 msec apart. Frame numbers are indicated at upper left. Anterior is to the right, posterior to the left. Five radio-opaque markers are in the body of the tongue and are numbered on the first frame. The pumping movements of the tongue are indicated with arrows and, given the seal of the tongue to the palate, produce a reduced pressure inside the oral cavity.
palate, which would imply the existence of a reduced pressure if the seals were complete. As marker three moved downward, the anterior seal broke in front of the nipple, and milk moved into the space between the middle tongue and the hard palate. Milk that entered the oral cavity as the space was created had a frothy
Suckling in macaques is also based on a dorsal/ventral pumping mechanism, although there were some differences from the infant pigs (German et ah, 1992). These animals were smaller than the infant pigs, weighing approximately 500 g. The position of the infant macaque tongue during suckling was similar to that of the infant pig, despite the macaque face and tongue being shorter in an anterior/posterior direction. The tongue was elevated and sealed against a large area of the hard and soft palates, leaving a small space in front of the nipple at the anterior end of the oral cavity (Fig. 14.3, frame 254). Over the next 500 msec the jaw opened and the middle of the tongue moved strongly downward, breaking its seal with the hard palate (Fig. 14.4). A small portion of the anterior tongue remained elevated without contacting the palate, leaving a smiall gape through which liquid flowed (Fig. 14.3, frame 261). The posterior of the tongue remained firmly sealed to the posterior region of the hard palate and most of the soft palate. As milk passed through the now incomplete anterior seal into the oral cavity, it did not have a uniform radiodensity but appeared to contain bubbles. At frame 288 (Fig. 14.4), the jaw began to close, and milk
452
R. Z. G e r m a n a n d A. W. C r o m p t o n
FIGURE 14.2. Movement of jaw and tongue markers from Fig. 14.1 vs time, in frame number. On the left is anterior/posterior movement, on the right is dorsal/ventral movement. Jaw movement, or gape, and movement of the anterior tongue are minimal. Movement of the posterior markers is large.
no longer left the nipple. The anterior tongue moved upward, and the posterior tongue downward and forward (Fig. 14.3, frame 282). Liquid transport continued with these movements for several frames as the jaw continued closing (Fig. 14.3, frame 296). At frame 300 (Fig. 14.3), the anterior of the tongue was still rising and resealing with the hard palate. The posterior tongue changed direction and began to move upward and somewhat backward, moving the contact between it and the palate (Fig. 14.3, frames 300 and 303). Through these frames, and to the end of this sequence, the posterior tongue established a seal with the hard palate initially, and subsequently with the soft palate, suggestive of a squeezing action. At the same time the liquid moved into the valleculae and then through the piriform recesses, culminating in a swallow. When tongue movement is measured relative to mandible movement, anterior tongue markers, lying underneath the nipple, moved very little relative to the lower jaw. However, marker 3 (Fig. 14.3), nearer the middle of the tongue, moved downward (dorsally) relative to the lower jaw, at the same time the jaw moved downward. The extent of upward movement of the anterior tongue was the extent of jaw movement, although again, marker three had a large excursion of movement in addition to jaw movement. Pressure change in the nipple during a single cycle
was a cyclic rise and fall that was correlated with tongue movement and milk flow. Change in pressure closely followed change in gape and middle tongue movement, with pressure relatively low at maximum gape, after liquid has been acquired from the previous cycle, and rising as the jaw moves upward and the milk is transported through the oral cavity. D . Suckling in Infant O p o s s u m s Suckling in infant opossums is also characterized by a dorsal/ventral pumping mechanism (German and Crompton, 1996). A single cycle of tongue movement indicated several similarities to the kinematics of suckling in other mammals. These infants were the smallest studied, less than 50 g, and thus had only two tongue markers. Initially the tongue formed a seal with the hard palate, just anterior to the hard palate/soft palate junction (Fig. 14.5, frames 00-10). In front of this seal, the tongue moved downward, as milk flowed from the nipple into the oral cavity (Fig. 14.5, frames 00-20). Between frames 00 and 10, the anterior tongue moved upward, maintaining the tongue to palate seal. The seal moved posteriorly between frames 10 and 20, as a slightly more posterior portion of the tongue contacted the soft palate. Next, the tongue near the first marker moved downward, increasing the space inside the oral
14. T h e O n t o g e n y of F e e d i n g in M a m m a l s 254
296
261
300
side the posterior region of the oral cavity. Finally, the tongue posterior to the seal, near marker two, moved upward to reach first the posterior-most edge of the hard palate and a large portion of the soft palate, forcing the milk in the anterior oral cavity into the oropharynx (Fig. 14.5, frames 49-63). The main movement of the tongue during suckling was in a dorsal/ventral direction (Fig. 14.6A).
^O
269
303
282
309
453
FIGURE 14.3. Lateral view of infant macaque suckling by pumping mechanism. Each of eight frames are taken from cine X-ray film of an infant macaque sucking. Filming was at 100 fps, so each frame was 10 msec apart. Frame numbers are indicated at upper left. Anterior is to the right, posterior to the left. Six radio-opaque markers are in the body of the tongue and are numbered on the first frame. The pumping movements of the tongue are indicated with arrows and, given the seal of the tongue to the palate, produce a reduced pressure inside the oral cavity. The final swallow is accomplished through a squeezing mechanism, rather than a pumping one. Stippled area is milk.
cavity. Between frames 20 and 34 (Fig. 14.5) the tongue near the first marker continued to move downward and slightly backward. There was considerable movement in the tongue that was not reflected in marker movement. The seal between palate and tongue shifted further in a posterior direction along the soft palate (Fig. 14.5, frames 24-34), and a portion of the tongue anterior to the first marker moved dorsally, upward toward the hard palate. These movements created a second space within the oral cavity, and milk flowed backward into this space. The rising portion of the tongue continued to rise, and by frame 49 again contacted the hard palate, sealing this bolus of milk in-
IV. RHYTHMICITY A N D CONTROL OF SUCKLING Suckling, as is true of drinking and mastication in adult mammals, is a rhythmic activity. In older animals, movement of jaws, tongue, and hyoid is regular and cyclic (German and Franks, 1991). In infants the behavior of tongue and jaws is also cyclic. The rate of this cyclicity varies with species, and there is significant variation in the rate among individuals of a single species (German et al, 1992,1997). The nature of neural control of rhythmic behavior has been studied extensively in adult mammals (Lund and Dellow, 1971; Lund and Olsson, 1983; Lambert et al, 1986; Goldberg and Chandler, 1990), as well as the extent to which peripheral sensory information can modify the ongoing rhythms of feeding. Little is known about the ontogeny of neural control of feeding. However, some information exists as to the rhythmic nature of infant feeding, in particular, what peripheral stimulation is necessary to elicit rhythmic activity and to what extent that rhythm is flexible and malleable. Mechanical stimulus by a nipple is not sufficient to elicit rhythmic jaw and tongue activity by infant pigs (German et ah, 1997). However, delivery of milk, often a single drop, or delivery at a low rate (an order of magnitude less than normal feeding frequency) was sufficient to elicit rhythmic movement. The frequency of the rhythmic jaw/tongue movement in this situation was significantly higher than when milk was delivered near the preferred rate of approximately 4 Hz (German et al, 1997). When the rate of delivery is varied from 50 to 150% of preferred suckling rate, there was no effect on the animal's suckling frequency. Individuals had a preferred suckling frequency independent of the rate of milk delivery. Given that milk delivery in these experiments was fixed over a long period of time, milk was inevitably delivered at all phases of the jaw cycle. This strongly suggests that subsequent transport and swallowing of the aliquot of milk are decoupled from the movements, and rhythm, of jaw and tongue movements involved in the acquisition of milk.
454
R. Z. G e r m a n a n d A. W. C r o m p t o n
t6
t6 t4 t3
t2 t1
gape 260
260
1
270
310
280
posterior
dorsal { anterior FIGURE 14.4. Movement of jaw and tongue markers from Fig. 14.3 vs time, in frame number. On the left is anterior/posterior movement, on the right is dorsal/ventral movement. Movement of the anterior tongue is minimal. However, jaw movement, measured through gape, and movement of the posterior markers are large.
o cycle start # cycle end 0.5 mm
FIGURE 14.5. Lateral view of suckling in an infant opossum. Ten frames are taken from cine X-ray film, at 100 fps, so each frame is 10 msec apart. Anterior is to the left, posterior to the right. There were two radio-opaque markers in the tongue indicated by filled-in circles. Open circles indicate the position of the markers in the above frame. The stippled area is milk. In infant opossums, the tongue forms a seal with the hard palate. The tongue pumps primarily in a dorsal/ventral direction. This creates a reduction in intraoral pressure, which in turn moves the milk through the oral cavity.
(
dorsal anterior
FIGURE 14.6. (A) Two-dimensional movement of radio-opaque markers in an infant opossum tongue while suckling. (B) Twodimensional movement of radio-opaque markers in an infant opossum who has been removed from the mother and consequently laps instead of sucks. Dorsal is up, and anterior is the right. Each curve traces the movement in a lateral plane. The unfilled circle is the beginning of the cycle, the filled circle indicates the end of the cycle. (A) Cycles are dominated by dorsal/ventral movements, associated with the pumping mechanism seen in other species. (B) Movements are entirely anterior/posterior characteristic of lapping in other adult mammals.
14. The Ontogeny of Feeding in Mammals V. C O O R D I N A T I O N OF SWALLOWING A N D RESPIRATION Feeding, particularly swallowing, is linked to breathing because both functions utilize common anatomical spaces. In mammals of all ages, including humans, inspired air must pass through the pharynx into the glottis. Swallowed food also passes through the pharynx, crosses the laryngeal opening, and must be excluded from the glottis. Mammals have devised various anatomical structures and neuromuscular mechanisms to minimize the disruption of air flow during mastication. All mammals, excluding adult humans, can lock the larynx (including the epiglottis) into the nasopharynx so that an airway extends from the external nares to the trachea, bypassing the oral cavity (Wood Jones, 1940; Negus, 1949). This arrangement permits suckling, and mastication and food manipulation (in nonhuman adults), to take place within the oral cavity, without interfering with rhythmic breathing. A great deal is known about the mechanics of swallowing in all mammals, especially humans (see Logemann, 1983; Hiiemae and Crompton, 1985; Jones and Dormer, 1991), but the relationship between swallowing and respiration is more contentious. Negus (1949) suggested that breathing need not necessarily be interrupted by swallowing because of the continuous airway described earlier. Negus's thesis that breathing and swallowing could and did occur simultaneously generated controversy and extensive contradictory data (Ardran et aZ., 1958; Lieberman, 1984; Crelin, 1987; Laitmann and Reidenberg, 1988,1993; Larson and Herring, 1996). Adult opossums are in the curious position of swallowing both ways; the airway is clearly interrupted when they swallow solid food, but it remains open when they swallow liquids. The relationship between respiration and swallowing is best known in adult humans, where the situation is different from all other mammals, given the unique, low position of the adult human larynx. These data suggest a complex, linked relationship between the two behaviors (Selley et ah, 1989,1990; Logemann et ah, 1992; Maddock and Gilbert, 1993; Martin et al, 1994; Paydarfar et al, 1995). A number of factors complicate the relationship between swallowing and respiration in infant mammals. The position of the epiglottis during swallowing depends on the liquid the animal is drinking (Laitman et ah, 1977). Several aspects of respiration increase with ontogeny, including respiratory drive (Farber, 1988; Barrington and Finer, 1991) and regularity of respiration during feeding (German et ah, 1992). Coordination between feeding and respiration also changes during ontogeny, as well as the pathway of the swallow through the oropharynx and the timing of swallows
455
relative to the respiratory cycles (German et ah, 1996; Crompton et ah, 1997). One interpretation of these ontogenetic changes is that, initially, breathing and swallowing are not correlated but that a linkage develops early in infant maturation. If this is the case, the question arises of why breathing and swallowing should need to be linked, given that infants maintain an intranarial larynx during a swallow. Some of the changes documented in respiration and swallowing may explain the discrepancies reported in the literature in relation to breathing and swallowing patterns in human infants. Laitman et ah (1977) claimed that, in human infants, a patent airway could be retained during a saliva swallow and that respiration was only briefly interrupted when milk or a barium milk mixture was swallowed. However, several authors (Ardran et ah, 1958; Wilson et ah, 1981; Koenig et ah, 1990; Selley et ah, 1990; Medoff-Cooper et ah, 1993) all agree that swallowing disrupts breathing patterns in human infants. Swallows can, nevertheless, occur at various times during the breathing cycle of human infants and we have found the same to be true in infant macaques. Bamford et ah (1992) found that swallows could occur at any time in the respiratory cycle of the human infant, whereas Selley et ah (1990) found that it could only occur (a) between inspiration and expiration or (b) during the middle of expiration. Most of these human studies, however, are limited to a small sample of infants of a particular age. Given the great variations in the methodologies employed, it is difficult to integrate the results obtained by one research group working on one age group with any other set of data.
VI. T R A N S I T I O N FROM SUCKLING TO D R I N K I N G AT W E A N I N G With maturation, infant mammals stop suckling and begin ingestion of solid food and drinking from sources other than their mothers. This change occurs as teeth erupt and is associated with morphological changes in the face, particularly an increase in the anterior/posterior direction (Helm and German, 1996). Little is known about initial consumption of solid food, but some data exist on the transition from suckling to juvenile drinking (Thexton, et ah, 1998). Longitudinal data, showing the change in tongue movements in the same individual, indicate significant differences between suckling and drinking. In suckling, movements of any magnitude are restricted to the posterior twothirds of the tongue. This region is involved in the aquisition of milk from the nipple and the transport into the vallecular recess. In drinking, the whole tongue is involved in the ingestive movements, which
456
R. Z. G e r m a n a n d A. W. C r o m p t o n
reflect a combination of lapping and sucking movements. Jaw and hyoid movements are minimal in both activities. Consequently, the different movements of the tongue during the two activities are generated by changes in the contraction patterns of muscle within the tongue, i.e., they are produced by the contraction of different groups of intrinsic and extrinsic tongue muscles and not by changes in the movement of the tongue base (hyoid). In suckling, the marked cyclical depression in the midregion of the tongue anterior to the fauces lowers intraoral pressure at that point and assists in drawing milk from the teat. This movement is initially coupled with high levels of electromyographic (EMG) activity in those muscles forming a base to the tongue (digastric, geniohyoid); subsequently, the genioglossus and hyoglossus increase in activity and then the styloglossus, omohyoid, and sternohyoid show activity before the cycle repeats. Swallows occur during the jaw-opening phase of every other cycle so that a suckling sequence can be divided into suckling and suckle/swallow cycles. The swallow is characterized by increased levels of activity in the hyoglossus, styloglossus, and omohyoid muscles, which pull the tongue back to force milk out of the valleculae and into the esophagus. In contrast, when a pig drinks from a bowl there is a reduced sucking action combined with lapping movements and the swallows occur every third or fourth cycle. Activity levels of the hyoglossus and omohyoid, during both drinking cycles and drink/swallow cycles, are increased relative to suckling while the activity of the other monitored muscles decreased significantly. It would appear that the generation of negative pressure is less important for ingestion by drinking whereas the anterior/posterior movement of the tongue surface is more important. Differences in the activity levels of the monitored muscles during suckling and drinking may consequently be explained on the basis of the work necessary to generate the negative pressures while keeping the tongue in a stable overall position. Recordings from a larger suite of muscles are required to confirm or reject this suggestion. The timing of the transition from suckling to drinking may not be a strict function of age (German and Crompton, 1996). Evidence from infant opposums suggests that suckling and drinking can exist contemporaneously in littermates in different environments. Infants that are not yet mature enough to exist independent of n\aternal care, i.e., incomplete fur, inability to thermoregulate, and endentulous, will no longer suckle if they are removed from maternal care. Instead they exhibit only adult drinking mechanisms, characterized by the anterior/posterior movements of the tongue (Fig. 14.6B). Littermates remaining in maternal
care continue to suckle using a dorsal/ventral pumping mechanism that characterizes other infant mammals.
VII. EVOLUTIONARY CONSIDERATIONS Suckling is clearly a mammalian synapomorphy. Only mammals suckle, and all extant mammals suckle as infants. Even monotremes, which lack true mammary glands, suckle as infants (Griffths, 1978). Furthermore, suckling mechanisms appear to be homologous in widely separated taxa. In all species studied to date, suckling is characterized by dorsal/ventral movement of the tongue, and at least one seal between tongue and palate to isolate the oral cavity from the the oropharynx. This common suckling mechanism is correlated with a number of anatomical traits common to infant mammals: a relatively short face or viscerocranium, especially relative to the neurocranium, and relatively large tongues (Clark and Smith, 1993; Maunz and German, 1996; Helm and German, 1996). Despite these similarities in infant form and function, considerable variation exists in adult feeding mechanisms. Differences in feeding structures, particularly teeth, associated with different feeding mechanics are remarkable and are one of the hallmarks of mammalian evolution (Crompton and Jenkins, 1979; Eisenberg, 1981; Hiiemae and Crompton, 1985). Suction feeding occurs in nonmammalian groups, particularly snakes (Kardong and Haverly, 1993; Cundall, 1995). In these animals, the floor of the oropharyngeal cavity rises and falls, functioning as a pump that aspirates water into the cavity and forces it posteriorly into the esophagus. The expansion and compression of the buccal cavity are achieved through n\ovement of the oral floor, and a "bellows-like" displacement of the mandible (Kardong and Haverly, 1993). The tongue functions as a guide for the water, but is not an active part of a pumping mechanism. Cundall (1995) cautions that a more complex pumping mechanism is likely, given that water appears to be moving not only during depression of the buccal floor, but also during elevation, when compression within the cavity must be occurring. It is clear, however, that these mechanisms are used only by infant mammals in suckling. Whereas both snakes and infant mammals produce a pressure gradient to move liquid, the function of the mammalian tongue is unique to mammals. References Ardran, G. M., and R H. Kemp (1959) A correlation between sucking pressures and movement of the tongue. Acta Paediatr. 48:261272.
14. T h e O n t o g e n y of F e e d i n g in M a m m a l s Ardran, G. M., R H. Kemp, and J. Lind (1958) A cineradiographic study of bottle feeding. Br. J. Radiol. 31:11-22. Bamford, O., V. Taciak, and I. H. Gewolb (1992) The relationship between rhythmic swallowing and breathing during suckle feeding in term neonates. Pediatr Res. 31:619-624. Barrington, K., and N. Finer (1991) The natural history of the appearance of apnea of prematurity. Pediatr. Res. 29(4 Ft 1): 372-375. Bosma, J. (1985) Postnatal ontogeny of performances of the pharynx, larynx and mouth. Ann. Rev. Respir. Dis. 131: Supp. S10-S15. Clark, C. T, and K. K. Smith (1993) Cranial osteogenesis in Mondelphis domestica (Didelphidae) and Macropus eugenii (Macropodidae). J. M o r p h 215:103-114. Crelin, E. S. (1987) The Human Vocal Tract: Anatomy, Function, Development and Evolution. Vantage Press, New York. Crompton, A. W., and F. A. Jenkins (1979) Origin of mammals. In: Mesozoic Mammals: The First Two-Thirds of Mammalian History. J. Lillegraven, Z. Kielan-Jaworoska, and W. A. Clemens (eds). University of California Press, Berkeley, CA. Crompton, A. W., R. Z. German, and A. J. Thexton (1997) Protection of the airway during swallowing in infant mammals. J. Zool. Lond. 241:89-102. Cundell, D. (1995) Drinking in snakes. Am. Zool. 35:106A. DuBrul, E. L. (1988) Oral Anatomy. Ishiyaku EuroAmerica, St. Louis. Eisenberg, J. F. (1981) The Mammalian Radiations: An Analysis of Trends in Evolution, Adaptation, and Behavior. University of Chicago Press, Chicago. Farber, J. P. (1988) Medullary inspiratory activity during opossum development. Am. J. Physiol. 254(4 Ft 2) :R578-584. Eraser, D., and B. K. Thompson (1991) Armed sibling rivalry among sickling piglets. Behav. Ecol. Sociobiol. 29:9-15. German, R. Z., and H. A. Franks (1991) Timing in the movement of jaws, tongue and hyoid during feeding in the Hyrax, Procavia syriacus. J. Exp. Zool. 257:34-42. German, R. Z., A. W. Crompton, L. C. Levitch, and A. Thexton (1992) The mechanism of suckling in two species of infant mammal: miniature pigs and long-tailed macaques. J. Exp. Zool. 261:322330. German, R. Z., and A. W. Crompton (1996) Ontogeny of suckling mechanisms in opossums. Brain Behav. Evol. 48:157-164. German, R. Z., A. W. Crompton, C. McCluskey, and A. J. Thexton (1996) Coordination between respiration and deglutition in a preterm infant mammal. Arch. Oral. Biol. 41(6): 619-622. German, R. Z., A. W. Crompton, and A. J. Thexton (1997) Determinants of rhythm and rate in suckling. J. Exp. Zool. 278:1-8. Goldberg, L. J., and Chandler, S. H. (1990) Central mechanisms of rhythmical trigeminal activity. Pp. 268-289. In: Neurophysiology of the Jaws and Teeth. A. Taylor (ed.). Macmillan Press, London. Griffths, M. (1978) The Biology of the Monotremes. Academic Press, New York. Helm, J. W., and R. Z. German (1996) The epigenetic impact of weaning on craniofacial morphology during growth. J. Exp. Zool. 276: 243-253. Hiiemae, K. M., and A. W. Crompton (1985) Mastication, food transport and swallowing. Pp. 262-290. Functional Vertebrate Morphology. M. E. Hildebrand et al. (eds.). Harvard Univ. Press, Cambridge. Jones, B., and M. W. Dormer (1991) Normal and Abnormal Swallowing. Springer-Verlag, New York. Kardong, K. V., and J. E. Haverly (1993) Drinking by the common boa. Boa constrictor. Copeia. 1993(3):808-818. Koenig, J. S., A. M. Davies, and B. T Thach (1990) Coordination of breathing, sucking, and swallowing during bottle feedings in human infants. J. Appl. Physiol. 69:1623-1629. Laitman, J. T, E. S. Crelin, and G. J. Conlogue (1977) The function of the epiglottis in monkey and man. Yale J. Biol. Med. 50:43-48.
457
Laitman, J. T, and J. S. Reidenberg (1988) Advances in understanding the relationship between the skull base and larynx with comments on the origins of speech. Hum. Evol. 3:99-109. Laitman, J. T, and J. S. Reydenberg (1993) Specializations of the human upper respiratory and upper digestive systems as seen through comparative and developmental anatomy. Dysphagia 8: 318-325. Lambert, R. W., L. J. Goldberg, and S. H. Chandler (1986) Comparison of mandibular movement trajectories and associated patterns of oral muscle electomyographic activity during spontaneous and apomorphine-tnduced rhythmic jaw movements in the guinea pig. J. Neurophysiol. 55:301-315. Larson, J. E., and S. W. Herring (1996) Movement of the epiglottis in mammals. Am. J. Phys. Anthrop. 100:71-82. Lieberman, P. (1984) The Biology and Evolution of Language. Harvard Univ. Press, Cambridge. Logemann, J. A. (1983) Evaluation and Treatment of Swallowing Disorders. College-Hill Press, San Diego. Logemann, J. A., P. J. Kahrilas, J. Cheng, B. R. Pauloski, P. J. Gibbons, A. W. Rademaker, and S. Lin (1992) Closure mechanisms of the laryngeal vestibule during swallow. Am. J. Physiol. 262 (Gastrointest. Liver Physiol. 25), G338-G344. Lund, J. P., and P. G. Dellow (1971) The influence of interactive stimuli on rhythmical masticatory movements in rabbits. Arch. Oral. Biol. 16:215-223. Lund, J. P., and K. A. Olsson (1983) The importance of reflexes and their control during jaw movement. Trends Neurosci. 6:458-463. Maddock, D. J., and R. J. Gilbert (1993) Quantitative relationship between liquid bolus flow and laryngeal closure during deglutition. Am. J. Physiol. 265 (Gastrointest. Liver Phsyiol. 28), G704-G711. Martin, B. J. W., J. A. Logemann, R. Shaker, and W. J. Dodds (1994) Coordination between respiration and swallowing: respiratory phase relationships and temporal integration. J. Appl. Physiol. 76:714-723. Maunz, M., and R. Z. German (1996) Craniofacial heterochrony and sexual dimorphism in the short-tailed opossum {Monodelphis domestica). J. Mammol. 77:992-1005. McFarland, D. H., J. P. Lund, and M. Gagner (1994) Effects of posture on the coordination of respiration and swallowing. J. Neurophysiol. 75(5): 2431-2437. Medoff-Cooper, B., T. Verklan, and S. Carlson (1993) The development of sucking patterns and physiologic correlates in very-lowbirth-weight infants. Nurs. Res. 42(2): 100-105. Negus, V. E. (1949) The Comparative Anatomy and Physiology of the Larynx. Heinmann, London. Paydarfar, D, R. J. Gilbert, C. S. Poppel, and R F Nassab (1995) Respiratory phase resetting and airflow changes induce by swallowing in humans. J. Physiol. 483:273-288. Selley, W. G., F C. Flack, R. E. Ellis, and W. A. Brooks (1989) Respiratory patterns associated with swallowing. 1. The normal adult pattern and changes with age. Age Aging. 18:168-172. Selley W. G., R. E. EUis, F C. Flack, and W. A. Brooks (1990) Coordination of sucking, swallowing and breathing in the newborn, its relationship to infant feeding and normal development. Bri. J. Disorders Commun. 25:311-327. Thexton, A. J., A. W. Crompton, and R. Z. German (1998) Transition from suckling to drinking at weaning: a kinematic and electromyographic study in miniature pigs. J. Exp. Zool. 280:327-343. Thexton, A. J., and J. D. McGarrick (1988) Tongue movement of the cat during lapping. Arch. Oral. Biol. 39:599-612. Wilson, S. L., B. T Thach, R. T Brouillette, and Y. K. Abu-Osba (1981) Coordination of breathing and swallowing in human infants. J. Appl. Physiol. 50:851-858. Wood Jones, F (1940) The nature of the soft palate. J. Anat. 74: 147-170.
This Page Intentionally Left Blank
C H A P T E R
15 Feeding in Myrmecophagous Mammals KAREN ZICHREISS^ Section of Ecology and Systematics Cornell University Ithaca, New York 14853
I. INTRODUCTION A. Defining the Problem B. Taxonomy and Phylogeny of Mammalian Myrmecophages II. FORAGING ECOLOGY A. Prey Characteristics B. Myrmecophage Foraging Ecology and Behavior III. MORPHOLOGY OF THE FEEDING APPARATUS A. The Myrmecophagous Morphotype B. Exceptions to the Morphotype IV. FUNCTIONAL MORPHOLOGY A. Jaw Movements B. Tongue Movements C. Pharynx and Soft Palate Movements D. Feeding Stages in Myrmecophages V. EVOLUTION OF MYRMECOPHAGOUS SPECIALIZATIONS A. Phylogenetic Pathways to Myrmecophagy B. Structural Pathways to Myrmecophagy C. Primitive and Derived Features in the Myrmecophagous Feeding Apparatus VI. DIRECTIONS FOR FUTURE RESEARCH A. Form B. Function C. Evolution References L INTRODUCTION A. Defining tlie Problem Many ant- and termite-eating mammals share a set of anatomical features, including a long, thin, and 1. Present address: Department of Biological Sciences, Humboldt State University, Areata, CA 95521, e-mail:
[email protected].
FEEDING (K Schwenk, ed.)
459
highly extensible tongue, viscous saliva produced by hypertrophied salivary glands, reduction or loss of the teeth, v^ell-developed olfactory structures, a muscular and gizzard-like stomach, and forelimbs with robust flexor musculature and enlarged claws (Griffiths, 1968). Similarities in feeding structures prompted Cuvier (1817) to classify echidnas, anteaters, armadillos, pangolins, and aardvarks, as well as platypuses and sloths, as the Edentata. From an evolutionary perspective, however, this is not a natural group. Even though the phylogenetic interrelationships of some of its members are uncertain, Cuvier's Edentata contains four modern mammalian orders and at least five separate radiations of ant and/or termite specialists. Because the similarities seen in these ant and termite eaters are thought to be independently derived, they are frequently used as textbook examples of convergent evolution (Simpson and Beck, 1965; Eisenberg, 1981; Savage and Long, 1986; Pough et al, 1989). Knowing when convergence is an appropriate label for patterns in morphological evolution requires both robust phylogenetic hypotheses and close study of morphology. The supposed convergence in the feeding apparatus of ant-eating mammals has never been examined explicitly and in detail. Relatively little is known about the natural history of most ant-eating species, few comparative studies of the relevant anatomy exist, functional studies are altogether lacking, and the phylogeny of the taxa under consideration is debated. These are serious deficits, but ant-eating mammals present an interesting set of problems in both evolutionary and functional morphology, and a critical review of the available literature is overdue.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
460
Karen Zich Reiss B. Taxonomy and Phylogeny of Mammalian Myrmecophages
Many mammals eat ants and/or termites (myrmecophagy) but far fewer subsist exclusively on this restricted diet. Redford (1987) reviewed the literature on mammalian stomach and fecal contents extensively and found a wide-ranging proportion of ants and termites in mammal diets. Committed myrmecophages are those mammals in which at least 90% of the diet is ants and/or termites (Table 15.1) and their phylogenetic distribution among mammals is summarized in Fig. 15.1. Myrmecophages are found in each of the three major groups of mammals: the egg-laying Monotremata, the pouched Marsupialia, and the placental Eutheria. This tripartite division on the basis of reproductive characters (Blainville, 1816) is supported by cladistic studies, although molecular studies emphasize the closeness of the three groups (Westerman and Edwards, 1992; Gemmel and Westerman, 1994; Retief et ah, 1994) whereas morphological studies suggest that the Marsupialia and Eutheria are sister taxa (Marshall, 1979; Rowe, 1988; 1991; 1993; Wible and Hopson, 1993; Zeller, 1993) that together comprise the Theria. TABLE 15.1 Myrmecophagous Mammals'' Taxon
Diet"
Monotremata
Tachyglossidae
Tachyglossus aculeatus
4A,T
Marsupialia
Myrmecobiidae Didelphidae Thylacomidae
Myrmecobius fasciatus Metachirus nudicaudatis Macrotis lagotis
2A,3T 3T 3A,T
Manidae Myrmecophagidae
Manis sp. Cyclopes didactylus Tamandua sp. Myrmecophaga jubata Dasypus sp. Cabassous sp. Tolypeutes tricinctus Priodontes maximus Parascalops breweri Emballonura nigrescens Perodictus potto Cercopithecus sp. Otocyon megalotis Bdeogale cmssicauda Cynictis pencillata Mungos mungo Rhynchogale melleri Proteles cristatus Orycteropus afer Funisciurus sp. Prionomys batesi Oxymycterus rufus Praomys erythroleucus
4A,T 4A,2T 4A,T 4A,T 2-3A, T 4A,T 4A,T 4A,T 3A 3A, IT 3A 3 - 4 A , IT 2A,3T 3A,T 2A,3T 3A,2T 4T 2A,4T 4A,T 3-4A,T 4A 4A 1-4A,T
Eutheria Pholidota Xenarthra
Dasypodidae
Insectivora Chiroptera Primates Camivora
Tubulidenatata Rodentia
Talpidae Vespertilionidae Lorisidae Cercopithecidae Canidae Herpestidae
Protelidae Orycteropidae Sciuridae Cricetidae Muridae
^After Redford (1987). ^A, ants; T, termites; numbers indicate proportion ants or termites in diet: 1, <10%; 2,11-50%; 3,51-90%; 4, >90%.
There is one monotreme myrmecophage, the echidna (spiny anteater) Tachyglossus aculeatus, found in Australia and Tasmania. The only other extant echidna is Zaglossus bruijni, from New Guinea, but it appears to be an earthworm specialist (Griffiths, 1968; Nowak, 1991). Monotremes are considered highly autapomorphic descendents of the common ancestor of extant mammals. Echidnas are thought to be derived with respect to the only other extant monotremes, the platypuses, mainly because the oldest known fossil monotremes are more platypus-like than echidna-like (Woodburne and Tedford, 1975; Archer et al, 1985; 1992; Pascual et al, 1992). Among marsupial mammals, the numbat Myrmecobius fasciatus, is the most myrmecophagous, but it is considered to be an "amateur" ant eater (Griffiths, 1968) along with aardvarks and aardwolves (discussed later). Numbats are terrestrial squirrel-sized inhabitants of wandoo eucalyptus forests of southwestern Australia, where they are sympatric with echidnas (Fleay, 1942; Calaby, 1960). They are monotypic members of the Myrmecobiidae, which are most closely related to either the carnivorous Dasyuridae or Thylacinidae (the marsupial wolves) (Luckett, 1994). Among the placental mammals, or Eutheria, most myrmecophages are in the orders Xenarthra (anteaters, sloths, and armadillos; synonomous with the Edentata of some recent authors) and Pholidota (pangolins, or scaly anteaters). Extant xenarthrans are found only in the New World. There are four species of anteaters in the Myrmecophagidae. Cyclopes didactylus, the silky anteater, is a small (about 250 g) tropical forest canopy dweller. The two species of Tamandua, the vested or collared anteaters, are about the size of a fox terrier and semiarboreal. Myrmecophaga jubata, the giant anteater, is large (about 30 kg) and predominantly terrestrial. Myrmecophaga and Tamandua are sister taxa that are derived with respect to Cyclopes (Engelmann, 1978; 1985; Patterson et al, 1992; Gaudin and Branham, 1998). Sloths (Tardigrada) are the likely sister group of anteaters, although the extant sloths (Bradypodidae) are arboreal folivores only distantly related through numerous fossil sloth taxa (Patterson and Pascual, 1972; Webb, 1985; Gaudin, 1990; 1992; 1995). Sloths and anteaters together form the Pilosa, which is the sister group to the other major xenarthran radiation, the Dasypodidae (or Cingulata, armadillos) (Engelmann, 1978, 1985; Gaudin, 1990). Some armadillos are myrmecophagous: Priodontes maximus is the rare giant armadillo; Cabassous contains four species of naked-tailed armadillos Tolypeutes contains two species of three-banded armadillos, and Dasypus contains five species of long-nosed armadillos. These myrmecophagous armadillos com-
461
15. Feeding in Myrmecophagous Mammals
4?
^y'
.^^ J^^
^^ :^ ^
^
,W'
>^
^^ J>'
.cS' A>
<^ .tf
^°'
^' V* .
<> *
O'#
^
c^^ .,^^ J^^
•'i
-^^#^s:>°*
Xenarthra
pitheria Eutheria Mammalia FIGURE 15.1. Phylogenetic distribution of mammalian myrmecophages. Interrelationships of the extant eutherian orders are based on a 50% majority rule consensus of published analyses of both morphological and molecular data sets (from Honeycutt and Adkins, 1993); interrelationships of xenarthran families follow Engelmann (1978) and Gaudin (1995). Bolded clades contain committed myrmecophages with feeding apparatus specializations, and number of myrmecophagous species and common names are in parentheses. Asterisked clades also contain myrmecophages, but no anatomical specializations have been reported in these taxa. There appear to have been two independent origins of myrmecophagy in the Xenarthra, and despite controversy over the phylogenetic position of Pholidota, there have been at least six independent origins of anatomically specialized myrmecophages among extant mammals.
prise the tribe Dasypodini, within which Dasypus is probably the most derived taxon (Engelmann, 1978; 1985). The sister group of the Dasypodini is the Euphractini, which contains all the nonmyrmecophagous armadillos (Redford, 1985, 1987; Smith and Redford, 1990). All armadillos are terrestrial and fossorial. In contrast to the Xenarthra, the extant Pholidota are found only in the Old World. The Pholidota consists of the single family Manidae, which contains seven species of pangolin, all in the genus Manis. The Asian species {M.crassicaudata, M.pentadactyla, M.javanica) can be distinguished from the African species (M. gigantea, M. temmincki, M. tricuspis, M. longicaudata) on the basis of morphology (Pocock, 1924; Emry, 1970; Patterson, 1978). The monophyly of order and the Asian subgroup have recently been confirmed, but the African species appear to be polyphyletic (Gaudin and Wible,
1999). Within each geographic grouping there is a predominantly arboreal species, a semiarboreal species and a fully terrestrial species (Pages, 1970; Kingdon, 1974; Patterson, 1978) paralleling the situation in the anteaters. The phylogenetic positions of Xenarthra and Pholidota are unclear. An analysis based on morphological data that has dominated higher-level mammalian systematics for a decade (Novacek and Wyss, 1986) places Xenarthra and Pholidota as sister taxa, which together form the basal clade of extant eutherian mammals. The idea of a particularly close relationship between these orders is an old one (Storr, 1780; Blainville, 1816; Cuvier, 1817; Gill, 1870; Gregory, 1910) and there has been some molecular support for this alliance (McKenna, 1992; Norman and Ashley, 1994). Moreover, cladistic studies that investigate relationships among
462
Karen Zich Reiss
xenarthran and pholidotan taxa without constraining the monophyly of the orders have resulted in the inclusion of pangolins within Xenarthra (Engelmann, 1978; MacPhee, 1994; Norman and Ashley, 1994; Reiss, 1997a). However, there are vociferous opponents to these hypotheses. Simpson (1945, 1978) believed throughout his lifetime that Xenarthra was an endemic South American group and that common ancestry with Pholidota was so remote as to be irrelevant. Rose and Emry (1993) reviewed morphological evidence for the alliance and found it unconvincing due to polarity uncertainties, uneven within-taxon character distributions, and the likelihood of adaptive convergence. Some molecular studies place pholidotes with carnivores or a carnivore-ungulate assemblage (dejong, 1982; Miyamoto and Goodman, 1986; Shoshani, 1986; McKenna, 1987; Ohnishi, 1991) and there is some morphological support for this idea (Novacek and Wyss, 1986; Szalay and Schrenk, 1998). Aside from some xenarthrans and all pholidotans the only other eutherians that are committed myrmecophages are the sole extant member of the Tubulidentata (the aardvark Orycteropus afer), a single hyaenid in the Carnivora (the aardwolf Proteles cristatus), and scattered rodents and primates. These myrmecophagous taxa are found only in Africa. Affinities of aardvarks are controversial and range from fairly unresolved (Novacek and Wyss, 1986; 1988) to ungulate (Shoshani, 1993; Shoshani and McKenna, 1995) or paenungulate (dejong, 1982; Miyamoto and Goodman, 1986) alliances, and some cranial similarities with lipotyphlan insectivores have been noted (Thewissen, 1985; Novacek, 1986). The aardwolf is clearly a feloid carnivore in the family Hyaenidae, although feloid interrelationships are unresolved (Flynn et al, 1988; Wayne et ah, 1989; Wozencraft, 1989). Despite these controversies regarding eutherian interrelationships, outgroup criteria make it clear that myrmecophagy has evolved several times (see Fig. 15.1). The only taxa that might have inherited myrmecophagy from a common ancestor are anteaters and pangolins (Reiss, 1997a,b), and this only if the Xenarthra are not monophyletic. Myrmecophagy arose at least once in Monotremata, once in Marsupialia, twice in Xenarthra, once in Tubulidentata, once in Carnivora, and perhaps one more time in the Pholidota. All these myrmecophages, defined on the basis of trophic ecology, can be further subdivided into those that are anatomically specialized (i.e., have many of the features that were cited earlier) and those that are not. As we will see, the most anatomically specialized myrmecophages are echidnas, pangolins, anteaters, and, to a lesser degree, the armadillos. These taxa are the core of this review. Aardvarks, aardwolves, and numbats
are less anatomically distinct, but each has a few interesting anatomical features that will be discussed.
IL FORAGING ECOLOGY A. Prey Characteristics Ants and termites are ubiquitous, abundant, and live at high densities. They are most concentrated pantropically, but also extend into temperate regions, ants more so than termites. Ants and termites are found throughout all vertical levels of tropical forests, although ants are more common high in the canopy and termites are more common on the ground (Redford, 1987). Also, most species are colonial and thus, conveniently for their predators, exist at high local densities. Finally, ants and termites are geologically old. Ants are first known from the late Cretaceous, and termites are thought to be older. Their ubiquity and abundance appear to have been long-standing. Thus, ants and termites are likely to have been an abundant food source throughout the entire history of mammalian myrmecophagy. Unfortunately, ants and termites are also aggressive, noxious, nutritionally poor, and small. Aggression may involve stinging, biting, and spraying with endogenous irritants. Kinds and levels of aggression depend both on the species and on the status of an individual in a species' social system. However, because ant and termite sociality frequently involves a caste system with specialized soldiers, even milder individuals can have formidable armies protecting them. Soldiers are generally mobilized in response to perceived threats, either at the nest site or along frequently used paths to and from the nest. The time it takes for soldiers to be mobilized varies, but it is finite and characteristic for each species. Ant and termite aggressive/defensive behaviors are either based on or complemented by allelochemicals, whether poison in an ant's sting or the spray dispensed by some termites (genus Nasutitermes). The high concentrations of these various secondary compounds must be ingested and metabolized by a predator, along with the structural compounds that make up ant and termite bodies. Most nitrogen is in the chitinous exoskeleton and the presence of chitinases in mammals is only beginning to be explored. Also, with the exception of certain reproductive and larval stages, ants and termites have virtually no fat (Redford and Dorea, 1984). In addition to being nutritionally poor, ants and termites are quite small compared with the relatively large mammals that feed on them, thus predator/prey size ratios are large.
15. Feeding in Myrmecophagous Mammals B. Myrmecophage Foraging Ecology and Behavior Prey distribution and abundance, prey aggressive/ defensive strategies, the prey's low nutritional value, and the large predator/prey size ratio are all factors one expects to affect myrmecophage foraging. Not surprisingly, the global distribution of mammalian myrmecophages correlates with prey distribution and abundance. Moreover, sympatry of myrmecophagous species occurs in both New and Old World tropics and in Australia (Pages, 1965,; 1970; Lubin et al, 1977; Montgomery, 1985). Stomach content analyses show that diets of sympatric species may overlap partially but never entirely. This prey resource partitioning is, in part, passively dependent on predator microhabitat associations. In general, arboreal myrmecophages eat more ants than terrestrial myrmecophages, and even among Tamandua, individuals that are niore arboreal ingest more ants than terrestrial individuals in the same habitat (Pages, 1970; Montgomery, 1985). Passive partitioning also results from diurnal and seasonal variation in prey activity patterns (Montgomery, 1985; Redford, 1987). Additionally, there is some evidence that myrmecophages actively select preferred prey species and avoid others. Certain ant taxa, ponerines (Ponerinae), army ants (Dorylinae), and leaf cutters (Myrmicinae), are conspicuously absent from the diets of myrmecophagid anteaters, despite their abundance within feeding territories (Montgomery and Lubin, 1977). Where echidnas and numbats are sympatric, echidnas eat a higher percentage of ants, whereas numbats prefer termites, although away from numbats echidna preference for termites correlates with termite biomass (Abensperg-Traun and De Boer, 1992). This prey species selectivity appears to be olfaction mediated. Nasal turbinates are particularly well developed in myrmecophages, and there are numerous field observations that foraging myrmecophages will periodically stop, sniff in different directions, and then walk directly toward a particular prey species' nest where feeding is resumed (e.g., Sweeney, 1956; Montgomery and Lubin, 1977). Maze experiments have demonstrated that captive giant anteaters will preferentially walk toward food and away from other volatile compounds (McAdam and Way, 1967). Electroreception may also play a role in some myrmecophages. It has been suggested that the fleshy tentacles on the aardvark snout are electrosensory (Kingdon, 1974), and the echidna also has electroreceptors in its snout (Andres et a/., 1991; Augee and Gooden, 1992; Proske et al, 1998), which may assist in the detection of subterranean prey. With respect to prey aggressive/defensive strate-
463
gies, we would expect that myrmecophages are either resistant or use behavioral means to minimize the impact. Resistance could be conferred by integumental specializations, such as the scales of pangolins and the carapace of armadillos. Some anecdotal evidence suggests that pangolins are unperturbed by prey attacks. They have been seen rolling in prey nests with their scales elevated, after which they clamp down the scales, move to a body of water, elevate their scales again, and finally dart their tongue over the water's surface, collecting the now floating prey (Hatt, 1934; Nowak, 1991). This is an amusing feeding strategy that could serve other purposes as well, e.g., prey allelochemicals may offer integumental parasite protection. There have also been observations that pangolins shake their scales when swarmed over, sending prey flying (Hatt, 1934; Kingdon, 1974). I know of no reports on armadillo behavior in the face of angry prey, but one would expect their carapace to be particularly effective protection. The other myrmecophages, all of whom are described as thick skinned, show obvious displeasure during prey attacks. I have seen Tamandua feeding furiously as their forelimbs expose more and more of a nest, while a hind limb scratches at the nape and shoulders just as furiously. On the assumption that myrmecophages are bothered by prey defenses, patterns of foraging behavior in myrmecophagid anteaters have been interpreted as adaptations for avoiding prey defenses (Montgomery and Lubin, 1977; Montgomery, 1985). Foraging patterns in anteaters are usually described as "farming." A nest is broken into and fed upon for a short period of time, and then the anteater moves on, taking isolated prey on the way to another nest. Foraging grounds are generally large, many nests are visited in a single activity cycle, and damaged nests are not returned to for days. A similar pattern is seen in three species of African pangolins and the aardvark (all sympatric in West Africa) and in numbats (Calaby, 1960; Pages, 1970). The presumed advantages of this kind of foraging pattern are twofold: exposure to prey soldier defenses are minimized by short feeding bouts and prey colonies are maintained as long-term food resources. There is both interspecific and individual variation in the duration of feeding bouts at nest sites and this variation has been ascribed both to the predator's sensitivity to prey attacks and to seasonal variation in prey nutritional content (Kruuk and Sands, 1972; Lubin and Montgomery, 1981; Redford and Dorea, 1984). Pangolins feed from termite nests for much longer than Tamandua, perhaps because the pangolin's scaly covering affords it some protection. Tamandua does, however, spend an unusually long time at termite nests when fat-rich alates are present (Lubin and Montgomery,
464
Karen Zich Reiss
1981), perhaps providing incentive for Tamandua to suffer more annoyance than usual. The echidna Tachyglossus also shows a preference for alates. Indeed, the only time they feed on Iridomyrmex ants is when alates are present, and this is correlated with the echidna's emergence from hibernation and presumably higher energy needs (Griffiths and Simpson, 1966; Griffiths, 1968). Among myrmecophages, aardwolves and numbats are the main exception to nest-based feeding. Aardwolves specialize on a single genus of a surfaceforaging termite {Trinivertermes) and usually feed from aggregations of termites on the veld floor and only sometimes lick termites off the surface of a nest. They will not (and perhaps cannot) break into nests (Richardson, 1987,1990). Numbats feed mostly from the termite-infested bases of wandoo trees, as well as from the forest floor generally (Calaby, 1960). Prey size, indigestibility and noxiousness all suggest that myrmecophages should have unusual metabolic physiology and biochemistry. However, aside from comparative studies of mammalian basal metabolic rates, virtually nothing is known. Eutherian myrmecophages do have unusually low metabolic rates. A low metabolic rate may be a retained primitive character, but it has also been suggested that a low metabolic rate is an adaptation for myrmecophagy or, alternatively, is the result of the ingestion of concentrated allelochemicals (McNab, 1984). Regardless of controversies over its causal basis, a low metabolic rate probably facilitates subsistence on a small and nutritionally poor food source. Other physiological adaptations for myrmecophagy are likely, e.g., specialized metabolic pathways may also be present. However, in a strict sense, myrmecophages are committed but not obligate trophic specialists. Several species are successfully raised in captivity on a mash diet that does not include ants and termites, although this diet can lead to obesity (Heath and Vanderlip, 1988).
III. MORPHOLOGY OF THE FEEDING APPARATUS The preceding section discussed the exigencies imposed by prey characteristics on mammalian myrmecophages. There is some indication that myrmecophage foraging behavior is structured to cope with prey aggressive/defensive strategies. Also, mammalian myrmecophages may have physiological mechanisms that allow them to cope with their prey's small size, poor nutritional value, and perhaps even toxicity. These ecological, behavioral, and physiological issues should all be explored further. However, the most
obvious features that set myrmecophages apart from other mammals are in the feeding apparatus. A. The Myrmecophagous Morphotype The frequently cited features in which myrmecophages resemble one another are often superficially defined. It remains to be seen how closely they resemble one another if the anatomy is scrutinized. The following sections review osteology and the muscles of the jaws, tongue, pharynx, and soft palate. Each section begins with an overview of the anatomy of the region and is followed by a comparative review of the morphology in echidnas, pangolins, anteaters, and dasypodine armadillos. The goal is to discuss the comparative anatomy with particular emphasis on issues of homology and to provide a basis for a subsequent discussion of function. 1.
Osteology
Studies of the echidna skull (Griffiths et al, 1991), pholidotan skull (Emry, 1970; JoUie, 1968), edentate jaw joint (Lubosch, 1908), xenarthran auditory region (Gaudin, 1990, 1992; Patterson et al, 1989), xenarthran temporal region (Guth, 1961), hyoid apparatus of echidnas, anteaters, and pangolins (Gasc, 1967; Schneider, 1964), and the mammalian skull generally (Novacek, 1986, 1993) are available. This section is based on these accounts and my own observations and summarizes features that are directly relevant to feeding apparatus anatomy and function: the shape of the skull and dentary, their muscular processes, the jaw joint, dentition, the hyoid apparatus, and, for some species, the xiphoid processes of the sternum to which tongue muscles are attached. The myrmecophage skull is typically long, with a small braincase and relatively elongated rostrum (Fig. 15.2). The zygomatic arch is incomplete in anteaters and in most of the pangolins {Manis javanica and M. pentadactyla are variable. Rose and Emry, 1993), whereas echidnas and armadillos have a complete arch. The dorsal surface of the skull is smooth and without pronounced sagittal crests, and the temporal fossa is always reduced (Rose and Emry, 1993). The paraoccipital and styloid processes are either reduced or absent, and nuchal crests and mastoid processes are well developed only in armadillos (Gaudin, 1990). A long hard palate extends beneath the elongate rostrum. Palatine shelves extend nearly to the jaw joint, and the pterygoids extend to or beyond the otic region. In the anteaters Tamandua and Myrmecophaga, the pterygoids have medially expanded laminae that meet in the midline, prolonging the hard palate to the level of
465
15. Feeding in Myrmecophagous Mammals
iof
FIGURE 15.2. Comparison of skulls and dentaries of myrmecophages. Lateral views are to the left and palatal views are to the right. (A) echidna, Tachyglossus aculeatus; (B) pangolin, Manis pentadactyla; (C) anteater, Myrmecophaga tridactyla; (D) anteater, Cyclopes didactylus; and (E) armadillo, Dasypus novemcinctus. Skull sizes have been normalized to equivalent cranial lengths despite a 20-fold difference between C, the largest, and D, the smallest. In A through D the skulls and dentaries are streamlined, muscular processes are weak, and teeth are absent. (E) A myrmecophagous armadillo stands out in being relatively robust, but its skull and dentary are lightly built compared with nonmyrmecophagous armadillos.
the caudal bulla and eliminating the interpterygoid vacuity across which the soft palate usually lies (see Fig. 15.3 for a more detailed illustration of the skull of Tamandua). In the anteater Cyclopes, pangolins, and armadillos the pterygoids are similarly long but do not meet in the midline; only pangolins and Cyclopes have a defined pterygoid hamulus. Teeth are either reduced or absent altogether. Reduction, seen in the armadillos, typically consists of the loss of incisors and canines and the conversion of cheek teeth into homodont pegs that lack enamel. The dentary of the myrmecophagous armadillos is weak compared with nonmyrmecophagous relatives but remains L-shaped with prominent muscular processes
fo
at
pif
he
FIGURE 15.3. Skull and dentary of Tamandua mexicana (traced from photographs of CU 17743, Cornell Vertebrate Collections, Ithaca, NY). Dorsal view is at the top; palatal view in the middle; and lateral view of skull and dentary at the bottom. Arrowhead indicates the coronoid process of the dentary. In particular, note the streamlined appearance of the skull and dentary, indicative of the absence of robust processes for the muscle attachment. On the lateral view, also note the incomplete zygomatic arch and the unusual caudoventral position of the auditory foramen, from which the soft tissue auditory tube runs caudally. On the ventral view, note the absence of teeth and the hard palate that extends nearly the entire length of the skull. Scale: 1 cm. Abbreviations used in this figure and in Figs. 15.4 and 15.9. a, arytenoid; ac, anterior cornu; as, alisphenoid; at, auditory tube; bo, basioccipital; c, cricoid, ch, ceratohyal; co, corpus hyoideum; cs, constrictor salivaris; d, dentary; eam, external auditory meatus; eh, epihyal; es, esophagus; f, frontal; flp, foramen lacerum posterior; fo, foramen rotundum; gg, genioglossus; gh, geniohyoideus; he, hypoglossal canal; hg, hyoglossus; ih, interhyoideus; iof, infraorbital foramen; j , jugal; 1, lacrimal; Ivp, levator veli palatini; mh, mylohyoideus; mvp, medialis veli palatini; mx, maxilla; n, nasal; of, optic foramen; om, oral mucosa; p, parietal; pc, posterior cornu; pg, palatoglossus; pi, palatine; plf, posterior lacerate foramen; pm, premaxilla; pp, palatopharyngeus; pt, pterygoid; sd, salivary duct; sg, sternoglossus; sh, stylohyal; smf, stylomastoid foramen; so, supraoccipital; sof, sphenorbital fissure; sp, stylopharnygeus; sq, squamosal; sta, sternal aponeurosis; t, thyroid bone; to, tongue; and ty, tympanic bulla.
(Smith and Redford, 1990). Echidnas, anteaters, and pangolins are all toothless. In all but the anteater Cyclopes, the dentary is an elongate and bowed splint that
466
K a r e n Zich Reiss
has no prominent coronoid, angular, or condylar processes. The dentary of Cyclopes is similarly slender but has a peg-like coronoid and a defined angular process. The jaw joint in most myrmecophages is relatively low on the skull, at the same level as the hard palate (Rose and Emry, 1993). Among armadillos, this characteristic is most pronounced in Dasypus, and the jaw joint of other armadillos is higher (Patterson et al, 1989; Smith and Redford, 1990). The articular surfaces of the jaw joint are relatively flat in all taxa. The anteaters Tamandua and Myrmecophaga have flat, narrow, and rostrocaudally elongate facets, whereas other myrmecophages have oval to triangular facets that are slightly dished. The hyoid apparatus in most myrmecophages is well ossified and placed close to the sternum. The echidna rostral cornu (cornu hyale) is composed of the usual three elements, hypohyal, ceratohyal (= epihyal), and stylohyal, and is distally in contact with the auditory region (Gasc, 1967; Schneider, 1964). The pangolin hyoid apparatus has short rostral comua containing only the hypohyal, and a stylohyal ligament continues to the skull (Gasc, 1967). The anteaters are similar to one another (Fig. 15.4 illustrates the hyoid apparatus of T. mexicana): the elongate stylohyal lies superficially along the side of the neck, nearly reaching the skull in Tamandua and Cyclopes but not Myrmecophaga (Reiss, 1997b). The usual proximal articulation between the rostral cornua and the basihyal is absent;
instead, contact is at the epihyal-ceratohyal junction with the thyrohyal (caudal cornu or cornu branchiale) (Schneider, 1964). This condition is also seen in the armadillo Cabassous (Starck, 1967). The xiphoid processes of the sternum give rise to some extrinsic tongue muscles in echidnas, anteaters, and pangolins. In African pangolins these processes are elongated magnificently (Fig. 15.5). They extend from the sternum down the ventral abdominal wall, follow the curve of the pelvic basin to turn cranially, and end freely at the level of the diaphragm (Ehlers, 1894). In Asian pangolins the xiphoids are shorter but
B
B FIGURE 15.4. Hyoid apparatus of Tamandua mexicana (traced from photographs of CU 17743, Cornell Vertebrate Collections, Ithaca, NY): (A) dorsal view and (B) ventral view. Anterior cornua are especially elongate and, in situ, give skeletal support to the long postcranial pharynx. The proximal ends of the epihyals sit freely on the surface of the corpus hyoideum and the anterior cornu is attached to the hyoid apparatus mainly via the thyrohyal. Scale: 1 cm. See Fig. 15.3 legend for abbreviations.
FIGURE 15.5. Xiphoid process of the sternum in pangolins. (Top) Ventral views of an isolated xiphoid process of (A) an Asian pangolin and (B) an African pangolin. Notice the dramatic difference in the length of the process. (C) The bottom is a right lateral view of the abdomen of an African pangolin lying on its back; the abdominal wall has been cut away to reveal the position of the xiphoid process in situ (drawn from Manis tricuspis, AMNH 86944, American Museum of Natural History, New York, NY). Notice that the xiphoid process runs down the abdominal wall, curves around the pelvic basin, and terminates dorsal to the abdominal viscera; i, intestines; k, kidney; 1, liver; arrow denotes gall bladder.
15. Feeding in Myrmecophagous Mammals variable (Chan, 1995). A synovial joint capsule is present between the caudal sternebra and the xiphoid process (Doran and AUbrook, 1973; Sikes, 1966). 2. Jaw Musculature Mammalian jaw-closing muscles primitively include a masseter, a temporalis, and a pterygoid complex (Crompton, 1989; Davis, 1961; Hiiemae, 1976). The masseter typically has superficial and deep portions, and the deep portion may be further subdivided. The temporalis is usually a large and architecturally complex muscle. The pterygoideus complex is composed of a pterygoideus externus and a pterygoideus internus. All these muscles are derived from the embryonic mandibular muscle plate and are innervated by the mandibular branch of the trigeminal nerve (V3). In monotremes, jaw opening is accomplished by another trigeminal-innervated muscle, the detrahens mandibulae, which has no clear homologue in therians (Edgeworth, 1931; Schulman, 1906). Jaw opening in therians is accomplished by the digastricus, which, as its name suggests, has two bellies. The anterior belly is derived from the embryonic mandibular muscle plate, is innervated by the trigeminal, and is homologous to the monotreme intermandibularis (along with the therian mylohyoideus). The posterior belly is derived from the hyoid muscle plate, is innervated by the facial nerve (VII), and is homologous to the monotreme interhyoideus (Edgeworth, 1931). The adductors of myrmecophages are small and architecturally simple compared with nonmyrmecophagous close relatives. Echidnas and pangolins have superficial and deep components to their small masseter whereas anteaters have only a superficial layer (Edgeworth, 1923; Naples, 1999; Reiss, 1997a,b; Schulman, 1906). The echidna temporalis has a three-part structure that is characteristic of monotremes, but its main portion is small and architecturally simple, as is the temporalis of pangolins and anteaters (Reiss, 1997a; Schulman, 1906). Dasypodine armadillos have robust and architecturally complex adductors for myrmecophages, although they are reduced and simple compared with nonmyrmecophagous armadillos (Edgeworth, 1923; Lubosch, 1908; Reiss, 1997a; Smith and Redford, 1990). Pterygoideus muscles vary among myrmecophages. In the echidna, the pterygoideus internus and externus fuse during development (Edgeworth, 1931). In pangolins, the pterygoideus externus is present, although Edgeworth (1923) claims that the pterygoideus internus atrophies in the embryo. I interpret this muscle to be present and partially merged with the mylohyoideus (Reiss, 1997a). In anteaters and dasypodine arma-
467
dillos, both pterygoideus muscles are well developed and the pterygoideus externus has two distinct heads (Edgeworth, 1923; Naples, 1999; Reiss, 1997a,b; Kiihlhorn, 1939). Jaw-opening muscles are also variable. As mentioned earlier, the echidna has a detrahens mandibulae that runs from in front of the external auditory meatus to the angle of the mandible (Saban, 1971). In pangolins, the anterior portion of the digastric originates from the mandible and splits distally to insert on a salivary gland, the deep surface of the rostral cornu of the hyoid, and the fascia overlying the scalenus anterior (Edgeworth, 1923; Chan, 1995; Reiss, 1997a). The homologue of the posterior digastric forms the outermost layer of the pangolin's glossal tube (Chan, 1995; see Section III,A,3,b). Anteaters also lack the typical compound digastric. Instead, a sternomandibularis muscle runs from the sternum to the dentaries. The compound innervation of this muscle suggests that its cranial portion is homologous to the anterior digastric, and its caudal portion is homologous to the sternohyoideus (Reiss, 1997b). The posterior digastric of anteaters forms an interhyoideus (see Section III,A,3,b). Armadillos have variable digastric anatomy: Dasypus has a sternomandibularis and an interhyoideus like anteaters (Reiss, 1997a,b), whereas Tolypeutes has an anterior digastric that attaches to the transverse tendon of the interhyoideus (Edgeworth, 1923,1935). This apparent interspecific variation could represent individual variation (Reiss, 1997a,b). Finally, all myrmecophages have an unusually well-developed mandibulo-auricularis, a facial muscle that extends from the angle of the mandible (or the squamosal in pangolins) to the tympanic bulla, rostral to the external auditory meatus (Edgeworth, 1923; Reiss, 1997a,b). 3. Tongue a. External Features Tetrapod tongues are epithelium-covered muscular bags, usually anchored proximally to the hyoid apparatus. Most mammals have chunky, dexterous, and moderately extensible tongues (Doran and Baggett, 1971). The dorsal surface is typically covered with numerous specialized epithelial papillae. Conical papillae (filiform and foliate) are the simplest and most numerous and probably increase surface area and act as mechanoreceptors. Fungiform papillae bear specialized taste buds and are interspersed among the conical papillae. Vallate papillae, unique to mammals, are found at the junction of the oral and pharyngeal parts of the tongue and are the most specialized morphologically. They sit in taste bud-littered epithelial depressions into which both mucous and serous glands empty. Much effort has been made to characterize and
468
K a r e n Zich Reiss
classify the distribution of papillae types in mammalian tongues (see references in Doran, 1975). The tongues of echidnas and armadillos, both of which maintain hyoid attachments (Fig. 15.6), are long and thin rostrally, but widen substantially toward the root. The tongues of pangolins and anteaters, free of hyoid attachments, are long and thin throughout, and the diameter increases only slightly as one proceeds from rostral to caudal. The stratum corneum of the dorsal tongue epithelium is unusually thick in myrmecophages (Sonntag, 1925), and echidnas have toothlike keratinized papillae on the base of the tongue (Griffiths, 1968; Sonntag, 1923). Myrmecophages also have what appear to be reservoirs of extra tongue mucosa surface area. The mucosa is folded into numerous
transverse sulci at the base of the tongue of echidnas, pangolins, and anteaters (personal observations; Oppel, 1899). Anteaters and pangolins also have a mucosal hood that overlies the root of the tongue, and in pangolins, tongue mucosa is reflected onto the lining of their unique glossal tube. All types of tongue papillae are sparse in myrmecophages. Echidnas have two vallate papillae and small foliate papillae on the base of the tongue, whereas the free portion of the tongue is bare (Doran and Baggett, 1972; Griffiths, 1968; Oppel, 1899; Sonntag, 1925). Pangolins have three vallate papillae, and sparse fungiform and filiform papillae rostral to these (Kubota et al, 1962b; Oppel, 1899; Sonntag, 1923). All xenarthrans have only two vallate papillae (Sonntag, 1923). In anteaters, the tongue is otherwise bare. Taste buds are found only on the walls of the vallate papillae, and the bases of the vallate papillae bear the openings of serous gland ducts (Kubota et ah, 1962a; Sonntag, 1923). Armadillos have filiform and fungiform papillae scattered all over the tongue and particularly concentrated at its tip. Reports on the distribution of other papillae types and taste buds in armadillos are conflicting (Sonntag, 1923). In some myrmecophages there are specialized structures on the tip of the tongue. The tip of the pangolin tongue bears two well-innervated tactile organs that resemble Pacinian corpuscles (Kubota et ah, 1962b; Doran and AUbrook, 1973; Cheng, 1986). In xenarthrans, the tip of the tongue is richly innervated but no differentiated tactile organs are seen (Sonntag, 1923). In Dasypus, two small muscularized projections extend beneath the tip of the tongue but their histological structure is unknown (Sonntag, 1923). b. Extrinsic Musculature
F I G U R E 15.6. The tongue, soft palate, and oropharyngeal specializations of myrmecophages. Schematic illustration comparing (A) the echidna Tachyglossus, (B) the pangolin Manis, (C) the armadillo Dasypus, (D) the anteater Cyclopes, and (E) the anteater Myrmecophaga. The skull, hyoid, and laryngeal elements are stippled, the soft palate is hatched, and the tongue is bold. The esophagus (e) and the trachea (t) are illustrated in A through E but labeled only in A. A sternoglossus muscle, formed from fused hyoglossus and sternohyoideus muscles, makes up the tongue in all but the armadillo, but note that the echidna retains remnants of the primitive muscle attachments to the hyoid. The sternoglossus is entirely free from the hyoid in the other taxa. Also notice that the ventral muscular sheet, indicated by the large arrowhead, terminates on the hyoid in the echidna and armadillo, but continues caudally in the other taxa. This sheet contains the mylohyoideus and interhyoideus, which contribute to a muscular sheath that surrounds the tongue and pharynx of anteaters and pangolins. Finally, anteaters and pangolins possess a specialization of the oral muscosa called the prehyoid pouch, indicated by the asterisk, that overlies the root of the tongue, but that only pangolins have a glossal tube (shaded) that completely surrounds the caudal portions of the tongue.
The extrinsic musculature of the mammalian tongue includes muscles that originate from outside the tongue and terminate in the tongue (the -glossus muscles) and muscles that insert on the hyoid and indirectly affect movements of the tongue (the -hyoid muscles). The first group includes the genioglossus, hyoglossus, and styloglossus, all derived from embryonic hypobranchial musculature and innervated by the hypoglossal nerve (XII). Also in this group is the palatoglossus, derived from caudal arch branchiomeric musculature and innervated by the vagus nerve (X). Muscles of the second group have diverse embryonic origins and innervations. The mylohyoideus is derived from the embryonic mandibular muscle plate (which also gives rise to the anterior digastric) and is innervated by the trigeminal (V3). The stylohyoideus is derived from the hyoid muscle plate (which also gives rise to the posterior digastric) and is innervated by the facial nerve
469
15. Feeding in Myrmecophagous Mammals
1973), but in pangolins and anteaters the sternoglossus bypasses the hyoid apparatus entirely (Chan, 1995; Doran and AUbrook, 1973; Owen, 1862; Pouchet, 1867; Naples, 1999; Reiss, 1997a,b; Sikes, 1966). The sternoglossus is probably derived from a fusion of hyoglossus and sternohyoideus: it is innervated by the descending branch of the hypoglossal nerve, the presumed ancestral muscles appear to be missing (except see the anteater hyoglossus discussed later), and the morphology in echidnas suggests an intermediate condition. Myrmecophages with a sternoglossus muscle have modified the rest of their feeding apparatus so that most of the primitive extrinsic tongue muscles have little to do with the tongue. There are two common patterns to the modifications seen. First, a muscular sheath formed in part by homologues of the primitive intermandibularis and interhyoideus surrounds the tongue and may contribute circumferential fibers to its base (see Section III,A,3,c). Second, although some of the extrinsic tongue and hyoid muscles are unusually modified, others appear to be altogether absent. In echidnas (Fig. 15.7) the intermandibularis (myloglossus of Doran, 1973; mylohyoid of Duvernoy, 1830) arises from the dentaries, the hard palate, and distal tip of the rostral cornua of the hyoid apparatus (Duvernoy, 1830; Edgeworth, 1935). It is bilaminar (the deep layer is probably the annulus inferior of Doran), reminiscent of the differentiation of its therian homologue into the anterior digastric and mylohyoideus. It is continuous caudally with the interhyoideus (styloglossus of Doran), which arises from the skull and the rostral cornua (Doran, 1973; Edgeworth, 1931). The superficial
(VII). The geniohyoideus, sternohyoideus, sternothyroideus, and thyrohyoideus are hypobranchial muscles that are innervated by the descending branch of the hypoglossal nerve (XII). There is much conflict an\ong reports on the extrinsic musculature of echidna tongues (Doran, 1973; Duvernoy, 1830; Edgeworth, 1931,1935; Schulman, 1906), pangolin tongues (Chan, 1995; Cheng, 1986; Doran and AUbrook, 1973; Edgeworth, 1923; Ehlers, 1894; Nene, 1978; Reiss, 1997a; Sikes, 1966; Sonntag, 1923), and anteater tongues (Duvernoy, 1830; Owen, 1862; Pouchet, 1867; Naples, 1999; Reiss, 1997b). My own dissections of anteaters and pangolins, along with careful scrutiny of the literature, have suggested that at least some of the conflict is attributable to differing assessments of muscle homologies and varying anatomical nomenclature. To avoid compounding the confusion I make homology arguments explicit wherever there are points of contention and give muscle synonymies throughout the following text. Table 15.2 contains a summary of the conditions of the extrinsic tongue musculature in echidnas, anteaters, pangolins, and armadillos. The bulk of the tongue of echidnas, pangolins, and anteaters is an unusual muscle called the sternoglossus (see Fig. 15.6), also present in some nectarivous bats (Griffiths, 1982). The sternoglossus arises from the xiphoid process of the sternum, passes through the thorax on the deep surface of the sternum, passes through the neck ventral to the larynx and hyoid apparatus, and ultimately rises up into the oral cavity to form the free part of the tongue. In echidnas, some sternoglossus muscle fibers attach to the basihyal (Doran,
TABLE 15.2 Homologies of Tongue and Hyoid Muscles of Echidnas, Pangolins, Anteaters, and Armadillos^ Anteaters
Armadillos
modified +, in soft palate
sternomandibularis +, in soft palate
+ /sternomandibularis
interhyoideus
In glossal tube
interhyoideus
+ /interhyoideus
Genioglossus Hyoglossus Styloglossus Palatoglossus
+ +, sternoglossus
+, in glossal tube iternoglossus
+ +, sternoglossus
+, not in tongue
+, not in tongue
Geniohyoideus Sternohyoideus Sternothyroideus Thyrohyoideus
+ +, as sternoglossus +
+, glossal tube sternoglossus
+ sernoglossus + +
Muscle
Echidnas
Anterior digastric Mylohyoideus
intermandibularis intermandibularis
Posterior digastric Stylohyoideus
Sternoglossus
+
Pangolins
+ + +
+
+
+ + + + + + + ~
^A plus sign indicates the muscle is present and unmodified from the primitive condition; modifications from the primitive condition are described; a slash indicates anatomical variation within the taxon; and a minus sign indicates that no homolog is present.
470
K a r e n Zich Reiss
FIGURE 15.7. Ventral view of echidna cranial musculature (A and B from Schulman, 1906; C from Edgeworth, 1945). Progressively deeper dissections of echidna throat muscualture are shown. Abbreviations follow the nomenclature of the original authors: a, annulis intimus; b, buccinator; ch, ceratohyal; dm, detrahens mandibulae; gge, genioglossus extemus; ggl, genioglossus lateralis; gh, geniohyoideus; ih, interhyoideus; ims, intermandibularis; m, masseter; md, mandible; oh, omohyoideus; pi, pterygoideus lateralis; sg, stemoglossus; sm, stemomastoideus; st, stemothyroideus; t, temporalis; zm, zygomaticomandibularis. Notice, in particular, the contributions of the intermandibularis and interhyoideus to a sheath that surrounds the entire tongue as well as to the circumferential muscualture that is applied to the surface of the stemoglossus muscles. These are structural characteristics common to echidnas, anteaters, and pangolins.
portions of both these muscles insert onto a ventral midline raphe and the basihyal, whereas deeper fibers wrap around the lateral edges of the sternoglossi and insert on the ventral surface of the base of the tongue (Doran, 1973; Duvemoy, 1830). The genioglossus arises from the mandible; its median portion attaches directly to the ventral tongue base, but lateral fasiculi wrap around the lateral edge of each stemoglossus before entering the base of the tongue (Duvemoy, 1830; Saban, 1971). The geniohyoideus is not unusual and both styloglossus and palatoglossus are absent (Edgeworth, 1931,1935; Saban, 1971). The sheath surrounding the tongue of pangolins (Ehlers, 1894; Sikes, 1966) has been dubbed the glossal tube (Chan, 1995) (Fig. 15.8; also see Fig. 15.6). It is formed, in part, by contributions from the posterior digastric, geniohyoideus, and genioglossus (Chan, 1995; Edgeworth, 1923, 1935; Reiss, 1997a; Sikes, 1966). The glossal tube begins in the oropharynx and extends into the neck ventral to the hyoid apparatus. It terminates by attaching to the surface of the paired stemoglossus muscles at a level reported to be species specific (Chan, 1995) but that I have found to vary among individuals of the same species (personal observations). The lumen of the oral cavity extends into the glossal tube for its
full extent, so the tongue lies first within the oral cavity and then within the glossal tube. The glossal tube is made up of four concentric layers that are muscle ventrally and fascia dorsally and an inner layer of connective tissue and epithelium that is continuous with the oral mucosa (Chan, 1995). The outermost muscular layer (hyoglossus of Cheng, 1986; glossovaginalis superficialis of Ehlers, 1894) arises from the rostral cornu of the hyoid and is cranially continuous with an extensive mylohyoideus that arises from the dentaries, palatines, pterygoids, and, remarkably, the soft palate (Reiss, 1997b). This outer layer has transverse fibers, is innervated by the facial nerve (VII) and is the likely homologue of the posterior digastric. The next two layers of the glossal tube (glossovaginalis externum stratum externum and medium of Ehlers) are formed by the geniohyoideus, which arises from the mandible and splits distally into layers of longitudinal muscle fibers. The innermost muscular layer (glossovaginalis externum stratum internum of Ehlers) is formed by the genioglossus, which arises from the mandible and diverges so its muscle fibers lie transversely. There is some question as to the ultimate insertion of the geniohyoideus- and genioglossus-derived layers of the glossal tube. Doran describes a disk of dense
471
15. Feeding in Myrmecophagous Mammals
ghl(II) ghm(II) ghm(II)
gg(IV) fp
ghl(IIl) V FIGURE 15.8. Ventral view of pangolin cranial musculature (from Chan, 1995). The illustration on the left depicts the caudal extent of the mylohyoideus and its continuity with the outer layer of the glossal tube, derived from the posterior digastric. These muscles are reflected in the illustration on the right revealing the deeper layers of the glossal tube that surround the sternoglossus muscles. Abbreviations follow the nomenclature of the original author: fp, free portion of tongue; gg, genioglossus; gh, geniohyoideus; gh, geniohyoideus lateralis; ghm, geniohyoideus medialis; gpn, glossopharyngeal nerve; gt, glossal tube layers I-IV; he, hyoid cartilage; hn, hypoglossal nerve; hnd, descending branch of hypoglossal nerve; la, lingual artery; In, lingual nerve; mh, mylohyoideus; mha, mylohyoideus aponeurosis; mhp, mylohyoideus, pterygoid portion; mht, mylohyoideus, tympanic portion; nmh, nerve to mylohyoid and anterior digastric; sm, sternomastoideus; smd, submandibular duct; smg, submandibular gland; slg, sublingual gland; st, sternothyroideus; th, thyrohyoideus; tmh, tympanohyoid; vh, ventral hiatus.
connective tissue interrupting the fiber continuity of the sternoglossus and to which the caudal ends of the geniohyoideus and genioglossus are attached (Doran and AUbrook, 1973). Chan (1995) did not see this structure and claims that the geniohyoideus dissipates in loose connective tissue on the surface of the sternoglossus whereas the genioglossus inserts on the ventral midline. My dissections of three species of pangolin are in accordance with Chan (Reiss, 1997a). Edgeworth (1923, 1935) claims that the palatoglossus is absent in pangolins but that they have a unique pterygohyoideus muscle. I disagree with his interpretation and view his pterygohyoideus as a palatoglossus with an unusual origin from the hyoid apparatus resembling the conditon seen in anteaters (Reiss, 1997a,b). The styloglossus and stylohyoideus are absent in pangolins (Edgeworth, 1923; Reiss, 1997a). The muscular sheath formed by the mylohyoideus and interhyoideus that surrounds the anteater tongue also lines the oropharynx [Fig. 15.9; also see Reiss (1997b) for additional figures of Tamandua cranial my-
ology]. As in pangolins, the mylohyoideus arises from the dentaries, the hard palate, and the soft palate and is continuous with the interhyoideus (interstylohyoideus of Naples, 1999; ceratohyoideus of Owen, 1862). The interhyoideus arises from the rostral cornua of the hyoid apparatus and runs ventrally to meet its antimer via a stout tendon that crosses the ventral midline. It is innervated by the facial nerve and is the homologue of the posterior digastric (and the outer layer of the pangolin glossal tube). The geniohyoideus, which has the usual attachments, and sternomandibularis (discussed earlier) are closely applied to the sheet of muscle formed by the mylohyoideus and interhyoideus (Reiss, 1997b). Thus, in the neck, the sternoglossus lies in a space that is defined dorsally by pharyngeal mucosa and the hyoid and laryngeal elements, and laterally and ventrally by the multilayered sheet that consists of mylohyoideus, geniohyoideus, and interhyoideus. The genioglossus lies on the dorsal surface of this sheet and rises up between the paired sternoglossus muscles to form the frenulum (Reiss, 1997b, musculaire du frein
472
K a r e n Zich Reiss
F I G U R E 15.9. Schematic parasagittal view of the interior of the oropharynx of Tamandua mexicana. Rostral is to the left. The vertebral column and associated musculature, the dorsal wall of the pharynx, and the ventral stemomandibularis have been removed. The skull is tilted away from the page and the soft palate is pulled dorsally to afford view of the ventral surfaces of the hard and soft palate; the asterisk indicates where the left stemoglossus was cut and removed. Note the extensive contributions of mylohyoideus to the wall of the oropharynx and the mylohyoideus and stylopharyngeus to the soft palate. Also note that the hyoglossus and the palatoglossus terminate in the oropharyngeal walls.
of Pouchet, 1867), more lateral fibers bind the sternoglossi together where they rise up into the oropharynx (musculaire annulaire of Pouchet, 1867) and envelop each stemoglossus caudal to the root of the tongue (ringmuskel of Dingier, 1964a; musculaire spiraux of Pouchet, 1867). The hyoglossus is present as a distinct muscle (epihyoglossus of Owen, 1862; part of the hyobuccalis of Pouchet, 1867), even though its anlage also contributes to the stemoglossus. It originates from the free proximal end of the epihyal, runs forward in the floor of the pharynx dorsal to the still submerged stemoglossus, and dissipates in the mucosal hood overlying the root of the tongue. This hood forms the roof of a space that Owen calls the prehyoid pouch (Owen, 1862), present in both anteaters and pangolins (see Fig. 15.6). The styloglossus and stylohyoideus are absent in anteaters, and the palatoglossus is modified like that of pangolins (although it does not reach the palate in Tamandua). Armadillos do not have a stemoglossus and their extrinsic tongue muscles are fairly typical for eutherians (Doran, 1975; Edgeworth, 1923; Saban, 1968; Smith and Redford, 1990; Sonntag, 1923; Windle and Parsons, 1899). However, they do lack a palatoglossus and a stylohyoideus (Edgeworth, 1923; 1935; Reiss, 1997a). c. Internal Structure of the Tongue In addition to the extrinsic muscles that originate outside the tongue, the mammalian tongue contains in-
trinsic muscle fibers that both begin and end in the tongue. These intrinsic fibers, the lingualis propria, usually consist of transverse and vertical bundles of muscle. These, along with the longitudinal fibers provided by extrinsic muscles, are complexly organized by tongue connective tissues and give the mammalian tongue its notable dexterity. Primitively, a median connective tissue septum separates the right and left extrinsic muscles, and within this connective tissue usually lies the lingual artery and vein. The tongue is permeated by nerve fibers serving motor and general and special (taste) sensory functions. This general description is fairly representative of the armadillo tongue because armadillos do not have a stemoglossus (Edgeworth, 1923). In contrast, the internal structure of the tongue of echidnas, pangolins, and anteaters is comparatively simpler (Fig. 15.10). In general, the paired sternoglossi are dominant and the lingualis propria is sparse at the base of the free portion of the tongue. As one proceeds toward the tip of the tongue, the sternoglossi subdivide into a variable number of longitudinal muscle bundles and the intrinsic fibers become progressively more abundant. Myrmecophages differ in the source and geometry of the circumferential fibers, the geometry of the subdividing sternoglossi, and in the presence of specialized connective tissue structures (lyttae) in the tongue's tip. In the echidna, a layer of spiraling muscle fibers, the annulis intimis (annular muscle of Edgeworth, 1931; not the annulus inferior of Doran, 1973), is closely apposed to each stemoglossus (Duvernoy, 1830; Griffiths, 1968; Saban, 1971). These fibers appear to originate and insert on tongue connective tissues, although this is not explicitly stated by any author. In the free portion of the tongue, a tortuous artery lies between the sternoglossi and communicates with extensive vascular cavernous sinuses that lie in both the dorsal and the ventral midline of the free portion of the tongue. Some circumferential muscle fibers attach to the walls of the sinuses, and transverse and vertical fibers are sparse (Doran and AUbrook, 1973; Doran and Baggett, 1970). The origin of the circumferential fibers at the base of the pangolin tongue is siniilarly unclear, although at least some appear to have extrinsic origin (personal observations). The circumferential fibers attenuate rostrally and the sternoglossi subdivide into 20 peripherally placed bundles between which lie interwoven vertical and transverse fibers of the lingualis propria (Dingier, 1964a; Doran and AUbrook, 1973). Further rostrad, the tongue flattens and 10 bundles of longitudinal fibers lie on each side of a central connective tissue lytta. Vertical and transverse fibers are sparse and lie deep to the longitudinal muscles. The lytta surrounds some loosely packed longitudinal muscle fibers, a central artery (Cheng, 1986; Kubota et ah,
473
15. Feeding in Myrmecophagous Mammals
B
echidna
pangolin
anteater
FIGURE 15.10. Schematic transverse sections through the tongue of echidnas, anteaters, and pangolins. All illustrations are schematics drawn from published micrographs: echidna after Griffiths (1968) and Doran (1970); pangolin after Doran (1973) and Dingier (1964b); and anteater after Dingier (1964a). (A) A section near the root of the tongue, (B) near the middle of the free portion of the tongue, and (C) a section from the tip of the tongue. Longitudinal muscle bundles are shaded darkly; lingual nerves are solid circles; wavy circles represent vascular sinuses; and all other internal lines represent muscle fiber directions. Myrmecophages have a tongue muscle architecture that is simpler than most mammals. Longitudinal muscle fibers derived from the paired sternoglossus muscles are dominant, circumferential muscle fibers derived in part from extrinsic muscles surround these, and the transverse and vertical intrinsic fibers are sparse throughout but become more abundant toward the tip of the tongue. Section A of the echidna tongue shows the dual contribution of sternoglossus and hyoglossus (see Fig. 15.6 for a lateral view); other notable differences between taxa include the geometry of the subdividing bundles of longitudinal fibers and the presence of possibly erectile vascular sinuses in the tip of the tongue of echidnas and pangolins but not anteaters.
1962b), and small vascular sinuses (Doran and Allbrook, 1973). The fibers that initially envelop the anteater sternoglossi are clearly derived from the genioglossus (Dingier, 1964; Pouchet, 1867; Reiss, 1997b). Rostrally, the sternoglossi subdivide repeatedly into peripheral longitudinal muscle bundles, and transverse and vertical fibers are interwoven throughout the central portion of the tongue (Dingier, 1964a). These bundles extend the length of the tongue, although their number progressively diminishes. Toward the tip of the tongue, radial fibers dominate the tongue's core and there is no centrally placed connective tissue lytta or vascular sinus (Kubota et al, 1962a). 4. Pharynx and Soft Palate The pharynx is a tube of variable length that is continuous with the oral and nasal cavities. Its lateral and dorsal walls are muscularized and its floor is formed by the root of the tongue and the hyolaryngeal struc-
tures. Pharyngeal muscles consist of a set of transversely oriented constrictors and a set of retractors or dilatators that tend to be more longitudinal in orientation. Eutherians typically have a superior constrictor that originates from the skull and various cranial fascia and ligaments, a middle constrictor that originates from hyoid elements, and an inferior constrictor that originates from laryngeal elements. Each of these is often subdivided further into several muscles named for the precise site of origin. The retractor/dilatators originate from various places on the skull and run more or less longitudinally to insert on the dorsolateral pharynx. A stylopharyngeus and a palatopharyngeus are fairly consistently present. Marsupials differ from eutherians in that they lack a superior pharyngeal constrictor and their stylopharyngeus has both transverse and longitudinally oriented fibers (Edgeworth, 1916,1935; Smith, 1992,1994). The stylopharyngeus is innervated by the glossopharyngeal nerve (IX). All other pharyngeal muscles are innervated by the vagus nerve (X). The soft palate is a muscular structure that divides
474
Karen Zich Reiss
the rostral pharynx into ventral oral and dorsal nasal portions. It originates from the caudal hard palate and is usually suspended between the caudally extending pterygoid processes. Eutherian palatal muscles include the tensor veli palatini (innervated by the mandibular branch of the trigeminal nerve, V3) and levator veli palatini (whose innervation is controversial, but is probably from both facial and vagus nerves, VII and X) (Dellon, 1989; Furusawa et ah, 1991a,b; Hecht et al, 1993; Ibuki et al, 1978; Keller et al, 1984; Nishio et ah, 1976b; Van Loveren et al, 1983). Both of these muscles arise from the pterygoids and the auditory tube and insert on the connective tissue of the soft palate. Marsupials do not have a levator veli palatini (Edgeworth, 1935; Saban, 1968). The medialis veli palatini (= uvularis, palatinus) arises from the midline of the hard palate and runs caudally in the midline of the soft palate. The palatopharyngeus (mentioned earlier) spans the wall between the palate and the pharynx, and finally, a pterygopalatinus sometimes joins the other palatal muscles arising from the pterygoids. These muscles are all innervated by the vagus nerve (X). The pharynx and soft palate of echidnas are notably simpler than the preceding description based on therians. There is no superior constrictor and the stylopharyngeus is wholly transverse in orientation. An undifferentiated constrictor pharyngeus lies caudally and the palatopharyngeus is the sole longitudinal muscle (Edgeworth, 1935). The soft palate is composed mainly of connective tissue (Edgeworth, 1916). Edgeworth describes the pangolin pharynx to be typically eutherian with a full set of pharyngeal constrictors and dilatators that include a wholly longitudinal stylopharyngeus and a palatopharyngeus (Edgeworth, 1923) but my results differ. I could not identify a palatopharyngeus and found the pangolin stylopharyngeus to have both transverse and longitudinal bundles characteristic of the marsupial condition (Reiss, 1997a). Also, Edgeworth noted only the presence of a tensor veli palatini in pangolins (Edgeworth, 1923), whereas I found that the mylohyoideus and the transverse component of stylopharyngeus also enter the soft palate (Reiss, 1997a). Perhaps the messy interdigitation of palatal and pharyngeal muscle fibers and the deep position of the palatal portions of mylohyoideus and stylopharyngeus account for our differing opinions. Anteaters also have a two-part stylopharyngeus, and its transverse portion joins the mylohyoideus in the soft palate (Reiss, 1997b), as in pangolins. A superior pharyngeal constrictor is present (although weak in some species) palatopharyngeus is present, and the tensor and levator veli palatini vary among species (Reiss, 1997b).
Armadillos, like anteaters and pangolins, have a two-part stylopharyngeus (Edgeworth, 1923; Reiss, 1997a). It does not enter the soft palate, however, and in all other ways, armadillos have characteristic eutherian pharyngeal and palatal anatomy (Reiss, 1997a). B. Exceptions to the Morphotype The feeding apparatus of the more derived eutherian myrmecophages is generally indicative of ancestry, e.g., the aardwolf looks more or less like a hyena. Myrmecophagous squirrels and primates have no unusual anatomical features, whereas numbats, aardvarks, and aardwolves each have a few tell-tale features that are uncharacteristic of their lineage and probably represent myrmecophagous specializations. Numbats have in common with the specialized myrmecophages a long snout and hard palate, as well as simple cheek teeth that are variable in number. The mandible is reduced only slightly and the forelimbs are not particularly robust (Calaby, 1960; Fleay, 1942). Their tongue is long and flattened and the epithelium is highly keratinized. They have no sternoglossus, but cross sections of the tongue show numerous peripheral bundles of longitudinal muscle between which lie a dense array of vertical and transverse fibers (Griffiths, 1968). Aardwolf skulls are generally hyena-like, but both the skull and the hard palate are significantly wider than similar-sized relatives (Anderson et al, 1992). The cheek teeth are peg-like and do not occlude when the jaws are closed, although the jaw musculature is nevertheless robust. The aardwolf tongue is somewhat longer than that of its close relatives, but is mostly distinguished by having numerous highly cornified papillae (Anderson et al, 1992). Aardvark skulls have an elongate rostrum, but the mandible is L-shaped and has typical muscular processes. The dentition is reduced and similar to myrmecophagous armadillos in that cheek teeth are homodont, peg-like, and lack enamel. However, aardvark teeth are unique among Mammalia in having parallel pulp-filled tubules of dentine that give the group its ordinal name, Tubulidentata. Jaw musculature is robust and well differentiated, and a digastric is present (Frick, 1951). The aardvark tongue is thick at the root and pointed at the apex, and it bears filiform, fungiform, and three vallate papillae, but no terminal papillary concentrations or specializations (Sonntag, 1923). The hyoid musculature is typical for Epitheria, e.g., hypobranchial muscles are all present and attach to the hyoid, and a styloglossus and stylohyoideus are both present (Edgeworth, 1924; Frick, 1951; Shoshani, 1993).
15. Feeding in Myrmecophagous Mammals
475
IV. FUNCTIONAL MORPHOLOGY
B. Tongue Movements
Patterson (1975) suggested that the main functional requirement of the myrmecophage feeding apparatus is speed of ingestion. Myrmecophages must ingest a large number of individual prey daily, and if myrmecophages are sensitive to prey defenses then consuming prey quickly before colony defenses ensue will also be advantageous. Testing the hypothesis that the myrmecophage feeding apparatus is designed for speed (or any other adaptationist hypothesis) requires a basic understanding of the functional correlates of structure. Unfortunately, dynamic studies of myrmecophage feeding are lacking except for a single study that compares feeding in a dasypodine and a euphractine armadillo (Smith and Redford, 1990). This study's singularity gives it importance, but because armadillos lack a sternoglossus muscle it tells us little about echidnas, anteaters, or pangolins. The absence of functional morphological analyses of feeding in these latter taxa requires that the following section be based largely on inference from structure. It begins with a consideration of the possible functional consequences of the unusual structural features of the jaws, tongue, pharynx, and soft palate. This is followed by a brief consideration of how the basic stages of tetrapod feeding—lingual ingestion, prey processing, intraoral food transport, and swallowing—might be modified in myrmecophages.
Myrmecophages are noted for their long tongues capable of extreme protrusion. Two issues arise: (1) How is it protruded so far? and (2) how is it retracted? Tongue protrusion in mammals results, in part, from shape changes affected by the intrinsic musculature. Tongues can be viewed as constant volume sacs of muscle that function as a muscular hydrostat (Kier and Smith, 1985). Intrinsic muscle fiber orientation determines the kinds of shape changes that can occur, and the resting size and shape of the tongue determine the magnitude of shape changes that can occur. In the case of myrmecophage tongues, contraction of circumferential muscle, both intrinsic and extrinsic, will decrease the cross-sectional area of the tongue, which will be compensated for by an increase in length such that the total volume of the tongue remains constant. Because of geometric considerations, a tongue that is initially long and thin will get much longer in response to a relatively little shortening of circumferential fibers than would a shorter tongue under the same circumstances (Kier and Smith, 1985,1992). In myrmecophages, the level at which extrinsic muscles apply circumferential fibers to the sternoglossi, which is always caudal to the first appearance of intrinsic circumferential fibers, should determine the length of the tongue acting as a muscular hydrostat. Tongue protrusion in mammals is also partially dependent on the extrinsic muscles genioglossus and geniohyoideus pulling the base of the tongue forward (Crompton et al, 1977). A tongue free of hyoid attachments should have a more mobile tongue base than one that retains hyoid attachments. The long skull base, the caudal position of the root of the tongue, and reports that genioglossus fibers run caudally after entering the tongue in anteaters and pangolins (Dingier, 1964b) should all contribute to the absolute value of this displacement. The effect of extrinsic muscle contraction in echidnas and armadillos will depend largely on hyoid mobility, and because the echidna hyoid is anchored to the skull, their extrinsic muscles may not have much of an effect on tongue protrusion. The pangolin glossal tube is unique and the role of extrinsic musculature in pangolin tongue protrusion may be more complex. By manipulating fixed specimens, Chan (1995) observed that the glossal tube is everted and obliterated when the tongue protrudes, and he proposed that its function was to provide extra mucosa to the lengthening tongue. This may indeed be a role of the structure but it does not explain why the tube is so well muscularized. The orthogonal muscle fiber orientations in different layers of the glossal tube and the attachment of the longitudinal layer to caudal
A. Jaw Movements Myrmecophages have a small gape, relatively flat jaw joint facets, tooth reduction, and an overall reduction in jaw muscle mass compared with close relatives. These features, along with evidence from anteater stomach contents (Lubin et al, 1977; Montgomery, 1985), suggest that mastication is absent. At the same time, the degree of jaw muscle differentiation suggests that some mandibular movements do occur. The shape of the jaw joint, the fiber direction of the masticatory muscles, and the robust pterygoid muscles suggest that mandibular movements are primarily parallel to the skull base and may include protraction/retraction, medial/lateral movement, and/or rotation. In echidnas, manipulation of fresh specimens has led to the hypothesis that mouth opening is the result of the rotation of the dentary along its long axis (Murray, 1981). Due to the curvature of the dentaries, rotation causes the tips of the dentaries to move away from one another, thus opening the mouth. This hypothesis has recently been advanced for the anteater Myrmecophaga as well (Naples, 1999).
476
Karen Zich Reiss
portions of the sternoglossi suggest that contraction of the glossal tube may actively extrude the sternoglossi. Assuming that the different muscles of the tube can contract independently, contraction of layers with circumferential fibers and the attending sternoglossus shape change would affect partial tongue protrusion, whereas full tongue extension would require the added contraction of the longitudinal layer, which would evert the tube. The portion of the sternoglossus caudal to the attachment of extrinsic fibers presumably functions as a tongue retractor, as does its primitive homologue, the sternohyoideus (Crompton et al, 1977). Because a contracting muscle can shorten to only a finite fraction of its resting length (Gans and Bock, 1965), greater absolute shortening requires a longer muscle. In addition, the tongue is at a mechanical disadvantage when fully protruded due to the relationship between muscle fiber length and half-sarcomere overlap (Gans, 1982; Gordon et al, 1966). Any added tongue length not involved in the muscular hydrostat (i.e., not surrounded by circumferential muscle) reduces the average extension experienced over the entire muscle, which allows more forceful contraction upon retraction. These arguments parallel Smith's discussions of the biomechanical implications of tongue muscle length in lizards with protrusible tongues (e.g.. Smith, 1986) and are based on a simple model in which the sternoglossus is made up of long and parallel muscle fibers. Muscle fiber lengths, fiber type, filament and fiber packing, connective tissue architecture, and motor unit structure have complex and not fully understood effects on whole muscle behavior (Gans and Gaunt, 1991, 1992, and references contained therein) and not one of these factors has been investigated for the sternoglossus of any myrmecophage. Most myrmecophagous species have extra lingual mucosa in the form of mucosal pleats or the glossal tube so it is unlikely that the lingual tunic mechanically limits absolute displacement. Also, both pangolins and anteaters exhibit a behavior that appears somewhat like a yawn and involves a very slow protrusion of the tongue to a distance that is roughly twice that observed when wild Tamandua are feeding on surface-foraging termites, captive Manis and Myrmecophaga are feeding from a bowl of mush, or captive Myrmecophaga are feeding through clear, flexible, plastic tubing (personal observations). This suggests that mechanical limits may be irrelevant to normal feeding behavior. Unfortunately, when wild myrmecophages are feeding from nests, tongue displacement is obscured by the nest itself. The myrmecophage tongue appears designed for large excursions but excursion has its costs. For one.
excursion takes time [see Kier and Smith (1985) for a theoretical discussion of this issue in muscular hydrostats] and if speed is an important factor in myrmecophage feeding then the tongue should protrude as little as is sufficient in any given situation. That captive animals can protrude their tongue farther than they typically do while feeding is consistent with this expectation. However, there has been no empirical description of the kinematic relationship between displacement and speed in any myrmecophage with a sternoglossus. Discussions of muscular hydrostats emphasize the role of a structure's length/width ratio (e.g., Kier and Smith, 1985, 1992) in achieving displacement, but the thinness of myrmecophage tongues may also serve to minimize inertia, which would have effects on speed. Also, muscle fiber physiology and muscle architecture can affect muscle contraction velocity, but as mentioned earlier, these factors remain unexplored. Thin tongues are, in part, the result of fewer muscles in the tongue and this presumably has effects on tongue dexterity. In most mammals, the spatially diverse origins of the extrinsic muscles and the diversity of fiber orientations in the lingualis propria give remarkable dexterity to the tongue. In myrmecophages with a sternoglossus, the absence of any other -glossus muscle, the predominance of circumferential fibers, and sparse vertical and transverse intrinsic fibers all suggest that tongue dexterity is minimal. Popular accounts depict the myrmecophage tongue as prehensile and dexterous, snaking its way through labyrinthine tunnels in ant and termite colonies. However, such meandering could also result from the tongue passively following the contours of prey nests. My own preliminary experiments filming Myrmecophaga tongues in curved clear tubing suggest that this is the case. Moreover, it appears that one thing anteaters cannot do is lick their own snout deliberately (personal observations). C. Pharynx and Soft Palate Movements Echidnas, anteaters, and pangolins all have a muscle sheet derived from the intermandibularis (or mylohyoideus) that at least partially surrounds the oropharynx. This sheet of transverse muscle extends into the soft palate in anteaters and pangolins, forming a nearly complete ring of muscle. Its lateral portions are topographically reminiscent of the buccinator muscle and could function similarly, squeezing prey from the buccal chamber back into the lingual chamber of the caudal oral cavity. It is also reminiscent of reptilian pharyngeal constrictors, particularly in anteaters which have a long oropharynx, and could conceivably aid in prey transport. Finally, it may serve to elevate the tongue's base, a function usually carried out by
15. Feeding in Myrmecophagous Mammals muscles that are absent in anteaters and pangolins (styloglossus and palatoglossus). This has been proposed as the mechanism by which echidnas chew, apposing their robust tongue pad against keratinous palatal teeth (Griffiths, 1968). In pangolins, anteaters, and armadillos (as well as in echidnas) the stylopharyngeus has transverse fibers, which suggest function as a pharyngeal constrictor. In pangolins and anteaters, these transverse fibers also enter the soft palate and, along with palatopharyngeus, form a ring of musculature around the soft palate and nasopharynx. This well-developed palatopharyngeal sphincter would, at the very least, prevent live unmasticated prey from crawling up into the nasopharynx. Another unusual feature of anteaters and pangolins is the prehyoid pouch. In reference to Myrmecophaga, Owen (1862) suggested that it was a distensible holding chamber for the accumulation of prey items prior to swallowing. If this is true, the muscularization of the roof of this chamber in both Tamandua and Myrmecophaga could facilitate its eversion and emptying, but there is no muscle in the prehyoid pouch of Cyclopes or pangolins. D . Feeding Stages in Myrmecophages Mammalian feeding is generally broken down into four stages: lingual ingestion, prey processing, intraoral food transport, and swallowing. Myrmecophages differ from most other mammals in the relative importance of lingual ingestion and prey processing. 1. Lingual
Ingestion
Lingual ingestion involves tongue protrusion and retraction, which have been discussed earlier. Other important considerations are the roles of sensory cures in prey detection and the mechanics of prey acquisition. As discussed previously, initial orientation to prey appears to be olfactory mediated and, in some species, electroreception may also play a role. At closer range, the absence of taste buds or other chemoreceptors on the tongue's tip suggests that chemical clues are not a factor. However, the role of tactile organs in some species needs explanation. With respect to mechanisms of acquisition, prey are generally observed all over the length of the anteater tongue (personal observations), suggesting that local deformation of the tongue's surface, as is seen in fluid lapping in other mammals, is not occurring. Myrmecophage saliva seems the most likely n\eans of prey acquisition. It is extremely viscous and prey may simply stick to the tongue. Additionally, myrmecophage saliva could be chemoattractive to prey.
477
2. Prey Processing In most mammals, prey processing involves mastication with occluding dentition. Echidnas lack true teeth and so do not masticate in the usual sense, but can be observed chewing using cornified structures on their tongue base and hard palate (Crompton, personal communication). The absence of teeth in pangolins and anteaters, along with the presence of whole prey in the stomach, argues strongly for an absence of mastication in these species. Even in myrmecophages with teeth (armadillos, numbats, aardvarks, aardwolves, and Funisciurus squirrels), observations of feeding and stomach content analyses suggest that chewing occurs rarely. Northern populations of Dasypus novemcinctus are generalized insectivores that can be seen laboriously chewing insect prey both in the wild and in captivity (personal observations; Smith and Redford, 1990), whereas southern myrmecophagous conspecifics do not chew (Redford, 1987). Aardvarks eat and chew an African wild cucuniber {Cucumis humifructus) in what may be a symbiotic relationship (Meeuse, 1963), but they do not chew their usual majority diet of termites and some ants (Kingdon, 1974; Patterson, 1975). Aardwolves have not been observed chewing, but do use their teeth in intraspecific social displays and defense (Ewer, 1973; Richardson, 1985, 1991). Finally, numbats and Funisciurus squirrels do not chew when feeding on social insects (Emmons, 1975). These observations suggest that chewing is either disadvantageous or not necessary for myrmecophages. Chewing would slow down the rapid ingestion of prey (Emmons, 1975; Redford, 1987), prey chemical defenses might be released and irritate sensitive oral tissues (Anderson et ah, 1992), and these chemicals, as well as the large amounts of soil frequently ingested with ants and termites, could exacerbate tooth wear. Of course, mechanical breakdown is not the only way food can be processed in mammals. Chemical digestion is also a form of processing and occurs, in part, in the mouth. Despite a great deal of speculation, virtually nothing is known about myrmecophage saliva or other glandular secretions. Oral secretions may be important in neutralizing the effects of noxious prey secretions (Anderson et ah, 1992; Kratzing and Woodall, 1988), could assist in the digestion of chitin, and, as mentioned earlier, could be chemoattractive to prey species. 3. Intraoral Food Transport Before food can be swallowed it must be transported from the oral cavity to the oropharynx. Intraoral food transport mechanisms seem to correlate with the length of the oral cavity, surface texture of the hard
478
Karen Zich Reiss
palate, and size of the vallecular pouches (Franks et ah, 1984). While no myrmecophage has the pronounced rigid palatal rugae seen on the hard palate of many mammals, anteaters have epithelial papillae on the tongue and lining of the oral cavity that point toward one another (Dingier, 1964b; Kuhlhorn, 1939) and a ratcheting mechanism could work between these surfaces. This is supported by films of anteater feeding that show several tongue protrusion/retraction cycles after food is taken into the mouth but before it is swallowed (personal observations). At least one species of pangolin has caudally pointing fringed epithelial folds on the hard palate (personal observation), but whether these are coupled with surface irregularities on the tongue is unknown. It has been suggested that the basihyal might scrape prey off the protruding pangolin tongue (Doran and AUbrook, 1973) so that prey accumulate in the caudal oropharynx. Whether the tongue's excursion upon retraction is great enough to allow this to be an important mechanism of intraoral transport in any myrmecophage is uncertain. 4.
Swallowing
Mammalian swallowing is a reflex that involves the coordinated actions of the tongue, palatal, pharyngeal, and laryngeal striated muscle, and esophageal smooth muscle. The complexity of the behavior is related, in part, to the necessary coordination between feeding and respiration that results from the crossing of air and food passageways in the pharynx . The number and extent of oropharyngeal modifications in anteaters and pangolins suggest that mechanisms of swallowing may be substantially different from those reported for other mammals (e.g., De Gueldre and De Vree, 1984; German and Franks, 1991; Herring, 1993). However, the unknown functional consequences of each of these modifications make speculation futile. Mechanisms of bolus formation depend in part on mechanisms of intraoral transport, the prehyoid pouch and glossal tube could play a role both in bolus formation and in swallowing, and there may be coordination between the timing of tongue movements and the initiation of a swallow. 5. Neonate
Suckling
Few myrmecophages have bred successfully in captivity and nursing is rarely witnessed in the wild. It has been reported that echidna newborns do not suckle, but instead lap milk, which has been cited as evidence that lapping is a primitive mammalian feeding behavior (Crompton, 1989). In those mammals that do suckle, apposition of the tongue to the soft palate is an important component of forming a seal around the nipple. As mentioned in Section IV,C, the muscles that
usually cause this movement are absent or modified in anteaters and pangolins, but the unusual mylohyoideus may be a functional replacement. However, reports on captive-raised anteaters claim that the tongue hangs freely out of the mouth during nursing, and infants appear to be pumping rather than sucking fluids (M. Flint, personal communication). The generality of these phenomena for anteaters and other myrmecophages remains to be confirmed. V. EVOLUTION OF MYRMECOPHAGOUS SPECIALIZATIONS It is certain that myrmecophagy among extant mammals is the product of numerous independent evolutionary events. However, for none of these events do we have a good series of intermediates between a nonmyrmecophagous and a myrmecophagous form. This section reviews what we know about the history of myrmecophagy in diverse lineages, which suggests the phylogenetic pathways by which myrmecophagy can evolve. Following this, I construct a morphoseries of extant forms that suggest the pathways of structural change that lead to anatomically specialized myrmecophages. Finally, I discuss the implications of feeding apparatus character polarity and character state distributions in anteaters and pangolins. A. Phylogenetic Pathways to Myrmecophagy Patterson suggested that there are two evolutionary pathways to myrmecophagy: (1) from generalized fossorial insectivores and (2) from carnivores (Patterson, 1975). Consideration of the known fossils referred to each of the extant myrmecophagous taxa along with the phylogenetic structure of each taxon suggests that myrmecophagy is more easily achieved. Echidnas are not known as fossils. Moreover, the oldest known monotreme, the early Cretaceous Steropodon from Australia (Archer et ah, 1985), and the Miocene platypus-like Obdurodon (Archer et ah, 1992) are well toothed and are both more like platypuses than echidnas. This suggests that the immediate ancestors of echidnas may have been herbivorous. Myrmecobius, the sole marsupial myrmecophage, is itself known only from the Pleistocene. However, its phylogenetic relationships suggest a carnivorous ancestry. The other marsupials that have a high percentage of ants or termites in their diets (Table 15.1) are also marsupicarnivores. There are no fossil marsupials thought to be myrmecophagous. The history of myrmecophagy in the Pholidota is confused by uncertainties regarding the sister group to
15. Feeding in Myrmecophagous Mammals Pholidota (see Section I,B) and their uncertain relationship to the fossil taxon Palaeanodonta. Palaeanodonts are small fossorial insectivores known from the Late Paleocene and Early Oligocene. Some palaeanodonts have craniodental and limb modifications similar to those seen in extant myrmecophages and may themselves have been myrmecophagous (Rose and Emry, 1983; 1991). The oldest presumed pholidotan myrmecophage is the Eocene Eomanis from the Messel formation of Germany (Storch, 1978). Its morphology is similar to extant pangolins, including the presence of scales (Koenigswald et al, 1981), although some evidence points to affinities with the Palaeanodonta (Rose and Emry, 1993). The Oligocene Patriomanis from the Big Horn Basin of Wyoming (Emry, 1970) is also clearly a myrmecophage, although Simpson doubted its identification as a pangolin and suggested that it too was a palaeanodont (Simpson, 1978). The controversies surrounding pholidotan fossils, if anything, suggest a close relationship between palaeanodonts and pholidotes (Emry, 1970; Rose and Emry, 1993). This would imply insectivorous or perhaps myrmecophagous ancestry of Pholidota. The trophic diversity within extant Xenarthra and the affinities between armadillos and fossil glyptodonts on the one hand and anteaters and sloths on the other hand make it clear that myrmecophagy is not primitive for xenarthrans. The common ancestor of Xenarthra is debated and palaeanodonts have been suggested here as well (Simpson, 1931). Even so, myrmecophagy certainly evolved independently in armadillos and anteaters. At what point anatomical specializations arose in each of these two lineages is less certain. The fossil armadillo Stegotherium, from the Miocene of Argentina, is an undisputed myrmecophage on the basis of cranial anatomy, but this genus is generally placed as the sister group to Dasypus, the most derived of the extant dasypodines (Engelmann, 1978, 1985). The morphology of other fossil armadillos is not indicative of incipient myrmecophagy. Outgroup criteria applied to dasypodine armadillos suggest that ancestral armadillos were herbivorous (as are sloths, glyptodonts, and some other armadillos). The origin of anteater myrmecophagy is similarly unclear. Eurotamandua, also from the Messel formation (Storch, 1981), was originally described as a myrmecophagid. Radiographic analyses of the hard palate and middle ear show striking similarity to Tamandua (Storch and Habersetzer, 1991), but this fossil has been referred, alternatively, to the Pilosa (Gaudin and Branham, 1998) and the Pholidota (Rose, 1988, 1993; Shoshani et al, 1997). The Early Miocene Protamandua, from Argentina, is an undisputed myrmecophagid (Ameghino, 1904; Hirschfeld, 1976) and, like the extant Cyclo-
479
pes, its palatal shelves do not meet in the midline (Patterson et ah, 1992). Both these fossil taxa were clearly myrmecophagous, but a great deal of morphological change separates sloths from any of these anteaters (Reiss, 1997a) and the fossil taxa are of little value in reconstructing these events. The earliest aardvark, the Miocene Myorycteropus, as well as the Pliocene Leptorycteropus, both have more generalized anatomy than Orycteropus, suggesting that myrmecophagy was not primitive for that group. The Paenungulata, often suggested as close relatives of aardvarks, are all herbivorous. The Recent Plesiorycteropus is more specialized for myrmecophagy than Orycteropus (Patterson, 1975) and, as its name suggests, had been thought to be an aardvark. However, this genus is distinct from aardvarks in many ways and is not clearly allied with any other myrmecophagous eutherian. It has been diagnosed as a new eutherian order that evolved in isolation on Madagascar (MacPhee, 1994), suggesting yet one more independent origin of myrmecophagy in the Mammalia. Finally, while aardwolves are not known as fossils, other extant hyaenids are carnivorous. Clearly, we do not have all the desired evidence to confirm or refute Patterson's hypothesis, but it is likely that several of the extant myrmecophagous lineages arose from herbivorous ancestors. This suggests that the evolution of myrmecophagy is not constrained by ancestral trophic specializations. Moreover, the fact that the earliest pangolin and anteater fossils are so similar to extant forms suggests that whatever one's ancestry and whatever phylogenetic pathway is taken to niyrmecophagy, the resulting ecological niche and morphology are extremely stable. B. Structural Pathways to Myrmecophagy Unusual anatomical specializations are seen in the feeding apparatus of many myrmecophages. At the same time, there are several mammals that are ecologically specialized ant and termite feeders that have few anatomical specializations. The evidence that ancestral trophic specializations do not constrain the evolution of myrmecophagy suggests that a general morphoseries can be constructed from the least to the most specialized feeding apparatus. This morphoseries can give us insight as to how myrmecophagy arises in such diverse lineages. The most common feature shared by myrmecophages is the loss, reduction, or disuse of the teeth. Even those species that have well-developed teeth do not use them when eating ants or termites. That this feature is so commonly shared suggests that teeth are dispensable to a myrmecophage and that teeth are
480
Karen Zich Reiss
particularly evolutionarily malleable in mammals. This is supported by observations that tooth count is variable in toothed myrmecophages (Anderson et al., 1992; Calaby, 1960). The next most common feature shared by myrmecophages is an unusually long and thin tongue. A tongue's resting length confers a displacement advantage and a tongue's slenderness may confer a speed advantage. Even aardvarks have a tongue that is relatively long and slender despite the overall resemblance of their feeding apparatus to that of other ungulates. The repeated occurrence of this feature suggests that a quick and extensible tongue is advantageous if a mammal eats ants and termites or at least that a squat tongue is disadvantageous. The most anatomically specialized myrmecophages, echidnas, pangolins, and anteaters, have a long slender tongue that is composed primarily of the sternoglossus muscles. Other modifications of the feeding apparatus are also seen in taxa with a sternoglossus. Hyoid arch musculature is unusual and the palate and pharynx contain an unusual complement of muscles. Speculation regarding the functional significance of these anomalies was discussed in Section IV, but these shared features of anatomically specialized myrmecophages might also have phylogenetic implications.
blance has no basis in homology. Lyttae in mammalian tongues are remnants of the primitive reptilian entoglossum, but this potential homology has never been thoroughly explored. Most of the anatomiical features that characterize the feeding apparatus of anteaters and pangolins are derived features. Moreover, in most cases these derived features arise from an ancestral condition that is characteristically eutherian (the one exception is the stylopharyngeus, see Reiss, 1997a). While this is true regardless of where pangolins are placed with respect to other mammals, the homologous and derived similarities between anteaters and pangolins are so extensive that the most parsimonious interpretation of the anatomy is derivation from a common myrmecophagous ancestor within the Xenarthra (Reiss, 1997a). That this result is in keeping with some analyses based on broader data sets is encouraging (see Section I,B), but until there is more widespread consensus on pholidotan relationships, the evolutionary interpretation of shared derived features in anteater and pangolin anatomy will remain open.
C. Primitive and Derived Features in the Myrmecophagous Feeding Apparatus
A. Form
One result of the discussion just given is that the most specialized anatomy occurs in taxa that are relatively basal, particularly if one accepts the hypothesis of close relationship between Xenarthra and Pholidota. This raises the question of whether some of these anatomical features are, in fact, primitive characteristics retained from the ancestral mammalian feeding apparatus. This would lead to the hypothesis that the evolution of anatomically specialized myrmecophages in more derived mammals was, perhaps, constrained by the acquisition of a more derived feeding apparatus. Because it is very difficult to choose an appropriate outgroup with which to compare monotreme anatomy, the following discussion is restricted to the eutherian myrmecophages, anteaters and pangolins. Cladistic analysis of xenarthran and pangolin cranial muscles, using marsupials as an outgroup, demonstrates that only the absence of the stylohyoideus and the presence of a complex stylopharyngeus can be interpreted as retained primitive features (Reiss, 1997a). Other characteristics that are primitive in appearance are secondarily derived. Cross sections through a sternoglossus-based tongue shows marked resemblance to some reptilian tongues, but this resem-
VI. DIRECTIONS FOR FUTURE RESEARCH
A great deal is known about the structure of the feeding apparatus in myrmecophagous mammals, but there are still holes in data that derive both from insufficient taxon sampling and from relatively superficial analysis of a very complex set of structures. Some of the remaining conflict in published anatomical accounts may be due to misunderstanding the taxonomic level of patterns of variation, and there are always controversies in the determination of homology. Many of these animals are exotic, studies are often based on dissection of single specimen of a species, and the more speciose groups (armadillos and pangolins) have not been fully sampled. Knowledge of the full diversity of adults is important, but closely staged developmental series are also invaluable in homology discussions. Closely staged developmental series are likely to remain rare for many of the myrmecophages discussed here, but a series of Dasypus novemcinctus should be readily obtainable because Dasypus are easily bred and colonies are used extensively in leprosy research. Also, much can be learned from studies of mammals that have the primitive and generalized form of the mammalian feeding apparatus. The study on the development of craniofacial musculature in the marsupial Monodelphis (Smith, 1994) is incredibly valuable in
481
15. Feeding in Myrmecophagous Mammals
this regard, but many questions still remain. One important task that needs to be undertaken is an attempt to correlate this immunohistochemical analysis with Edgeworth's classic observations on mammalian cranial development that have formed the basis for many of the homology assessments in this review. B. Function With respect to function, organismal natural history is always a good place to start. Observations of what animals actually do with their anatomy helps define functionally and evolutionarily important features (Bock, 1979), which can guide functional investigations. In the laboratory, the potential value of X-ray cineradiographic studies of normal feeding in anatomically specialized myrmecophages cannot be overemphasized. Many of the competing hypotheses discussed earlier could be evaluated with simple, noninvasive X-ray cineradiographic observations employing radio-opaque food items. Basic patterns of food transport and swallowing could be easily observed, and some insight could be gained into the coordination of tongue movements and swallowing and respiration even without the surgical implant of markers into the tongue. More invasive techniques, such as electromyography, are difficult in exotic and endangered species, but the captive breeding of many of these species is a high priority at many zoos and surgical techniques may become feasible. Zoo animals have a great deal to offer even without resorting to invasive techniques. They provide ample opportunity to observe basic feeding behavior and also offer a means to test biomechanical predictions based on structure. Finally, there are some functional questions that require fixed specimens, specifically muscle fiber type histochemistry and muscle architecture. Museum specimens are sometimes too precious or inappropriately prepared for certain techniques, but there are other means of obtaining specimens. I have had good fortune collecting freshly road-killed Tamandua, and several zoos have given me former captives that died or required euthanasia. The functional morphology of myrmecophage feeding will never be as well understood as it is in Didelphis or macaques, but the problem is too interesting to be left untouched and persistence and ingenuity pay off. C. Evolution Our knowledge of mammalian structure and function is always enriched by an evolutionary perspective. Fiowever, in the case of myrmecophagous mammals, we are limited by unresolved questions regarding
mammalian phylogenetic history. This review has emphasized the importance of determining the phylogenetic positions and within-group relationships of both Xenarthra and Pholidota. This is a general problem. Numerous researchers have pointed out that modern mammalian groups were formed in an early and explosive radiation, the cladistic pattern of which may be irretrievable. Moreover, a great deal of subsequent anagenetic change may have obscured what little information existed. I believe that this is an excessively pessimistic viewpoint. Cladistic methodologies are still young, and new techniques (e.g., those devised to handle total evidence analyses) are continually being developed and explored. Moreover, slowly but surely, new comparative studies (both morphological and molecular) and new fossil finds are ever expanding the data set. Nevertheless, it cannot be denied that specialized basal groups such as monotremes and xenarthrans and (perhaps) pholidotans are particularly problematic. Abundant plesiomorphy, adaptation, and the consequent abundance of autapomorphies and the difficulty of choosing a suitable outgroup are all difficulties that must be reckoned with if we are to ever fully comprehend the evolution of mammalian myrmecophagy.
References Abensperg-Traun, M., and E. S. De Boer (1992) The foraging ecology of a termite- and ant-eating specialist, the echidna Tachyglossus aculeatus (Monotremata: Tachyglossidae). J. Zool. Lond. 226:243257. Ameghino, R (1904) Nuevas especies de mamiferos Cretaceos y Terciarios de la Republica Argentina. Ann. Soc. Cient. Argen. 58: 225-291. Anderson, M. D., P. R. K. Richardson, and R R Woodall (1992) Punctional analysis of the feeding apparatus and digestive tract anatomy of the aardwolf Proteles cristatus. J. Zool. Lond. 228:423-434. Andres, K. H., M. Von During, A. Iggo, and U. Proske (1991) The anatomy and fine structure of the echidna Tachyglossus aculeatus snout with respect to its different trigeminal sensory receptors including the electroreceptors. Anat. Embryol. 184(4): 371-393. Archer, M., T. R Flannery, A. Ritchie, and R. E. Molnar (1985) Pirst Mesozoic mammal from Australia: an early Cretaceous monotreme. Nature (Lond.) 318:363-366. Archer, M., P. A. Jenkins, S. J. Hand, P. Murray, and H. Godthelp (1992) Description of the skull and non-vestigial dentition of a Miocene platypus {Obdurodon dicksoni n.sp.) from Riversleigh, Queensland, Australia, and the problem of monotreme origins. Pp. 15-27. In: Platypus and Echidnas. M. L. Augee (ed.). Royal Zoological Society of New South Wales, Sydney. Augee, M. L., and B. A. Gooden (1992) Evidence for electroreception from field studies of the echidna, Tachyglossus aculeatus. In: Platypus and Echidnas. M. L. Augee (ed.). Royal Zoological Society of New South Wales, Mosman. Blainville, H. M. D. (1816) Prodrome d'une nouvelle distribution systematique de regne animal. Bull. Soc. Philomatique 1816: 105-124. Bock, W. (1979) The synthetic explanation of macroevolutionary
482
K a r e n Zich Reiss
change: a reductionistic approach. Bull. Carnegie Mus. Nat. Hist. 13:20-69. Calaby, J. H. (1960) Observations on the banded anteater Myrmecobius f. fasciatus Waterhouse (Marsupialia) with particular reference to its food habits. Proc. Zool. Soc. Lond. 135:183-207. Chan, L.-K. (1995) Extrinsic lingual musculature of two pangolins (Pholidota: Manidae). J. Mammal. 76(2): 472-480. Cheng, H. (1986) Anatomical studies of Manis pentadactyla's tongue and structures relative to the tongue. Mem. Beijing Nat. Hist. Mus. 36:1-9. Crompton, A. W. (1989) The evolution of mammalian mastication. Pp. 23-40. In: Complex Organismal Functions: Integration and Evolution in Vertebrates. D. B. Wake and G. Roth (eds.) Wiley, Dahlem. Crompton, A. W., A. Thexton, P. Parker, and K. Hiiemae (1977) The activity of the hyoid and jaw muscles during chewing of soft food in the opposum. Pp. 287-305. In: The Biology of Marsupials. B. Stonehouse and D. Gilmore (eds.). MacMillan, London. Cuvier, G. (1817). La Regne Animal. Deterville, Paris. Davis, D. D. (1961) Origin of the mammalian feeding mechanism. Am. Zool. 1:229-234. De Gueldre, G., and F. De Vree (1984) Movements of the mandibles and tongue during mastication and swallowing in Pteropus giganteus (Megachiroptera): a cineradiographical study. J. Morphol. 179:95-114. dejong, W. (1982) Eye lens protein and vertebrate phylogeny. Pp. 7 5 114. In: Macromolecular Sequences in Systematic and Evolutionary Biology. M. Goodman (ed.). Plenum Press, New York. Dellon, A. L. (1989) The levator veli palatini muscle arises from the second branchial arch: a hypothesis. Arm. Plast. Surg. 23:317319. Dingier, E. C. (1964a) Die konstructive Form der Zunge des Ameisenbaren. Verhandlungen Anat. Gesellschaft 58:405-416. Dingier, E. C. (1964b) Vergleichende funktionell-anatomischeStudien an Edentatenzungen {Dasypus sexcinctus, Smutsia temmincki, Tamandua tetradactyla und Myrmecophaga jubata). Gegenbaurs Morphologisches Jahrbuch 105(3): 375-395. Doran, G. A. (1973) The lingual musculature of the echidna, Tachyglossus aculeatus. Anat. Anzeiger 133:468-476. Doran, G. A. (1975) Review of the evolution and phylogeny of the mammalian tongue. Acta Anat. 91:118-129. Doran, G. A., and D. B. Allbrook (1973) The tongue and associated structures in two species of African pangolins, Manis gigantea and Manis tricuspis. J. Mammal. 54(4): 887-899. Doran, G. A., and H. Baggett (1970) The vascular stiffening mechanism in the tongue of the echidna. Anat. Rec. 167:197-204. Doran, G. A., and H. Baggett (1971) A structural and functional classification of mammalian tongues. J. Mammal. 52(2): 427-429. Doran, G. A., and H. Baggett (1972) The specialized lingual papillae of Tachyglossus aculeatus. I. Gross and light microscopic features. Anat. Rec. 172:157-166. Duvemoy, G. L. (1830) De la langue, consideree comme organe de prehension des alimens ou recherches anatomiques sur les mouvemens de la langue dans quelques animaux, particulierement de la classe des mammiferes et de celle des reptiles. Mem. Soc. Hist. Nat. Strasbourg 1:1-16. Edgeworth, F. H. (1916) On the development and morphology of the pharyngeal, laryngeal and hypobranchial muscles of mammals. Quart. J. Microscop. Sci. 61:383-432. Edgeworth, F. H. (1923) On the development of the cranial muscles of Tatusia and Manis. J. Anat. Physiol. Lond. 57:313-335. Edgeworth, F. H. (1924) On the masticatory, intermandibular and hyoid muscles of Orycteropus capensis. J. Anat. Physiol. Lond. 58: 134-139. Edgeworth, F. H. (1931) On the development of the external ocular.
masticatory, and hyoid muscles of Monotremata. Proc. Zool. Soc. Lond. 1931:809-815. Edgeworth, F. H. (1935) The Cranial Muscles of Vertebrates. Cambridge Univ. Press, Cambridge. Ehlers, E. (1894). Der Processus Xiphoideus und seine Muskulatur von Manis macrura Erxl. und Manis tricuspis Surdev. Zool. Miszellen I. Abhandlungen der Koniglighe Gesellschaft der Wissenschaften39:l-34. Eisenberg, J. F. (1981) The Mammalian Radiations: An Analysis of Trends in Evolution, Adaptation, and Behavior. University of Chicago Press, Chicago. Emmons, L. H. (1975) Ecology and Behavior of African Rainforest Squirrels. Doctoral Dissertation, Cornell University. Emry, R. J. (1970) A North American Oligocene pangolin and other additions to the Pholidota. Bull. Am. Mus. Nat. Hist. 142(6): 459-510. Engelmann, G. F. (1978) The Logic of Phylogenetic Analysis and the Phylogeny of the Xenarthra. Doctoral Dissertation, Columbia University. Engelmann, G. F. (1985) The phylogeny of the Xenarthra. Pp. 51-64. In: The Evolution and Ecology ofArmadillos, Sloths, and Vermilinguas. G. G. Montgomery (ed.). Smithsonian Institution Press, Washington, DC. Ewer, R. F. (1973) The Carnivores. Cornell Univ. Press, Ithaca, NY. Fleay, D. (1942) The numbat in Victoria. Vict. Nat. 59:3-7. Flynn, J. J., N. A. Neff, and R. H. Tedford (1988) Phylogeny of the Camivora. In: The Phylogeny and Classification of the Tetrapods, Vol. 2. M. J. Benton (ed.). Oxford Univ. Press, New York. Franks, H. A., A. W. Crompton, and R. Z. German (1984) Mechanisms of intraoral transport in macaques. Am. J. Phys. Anthropol. 65:275-282. Frick, H. (1951) Uber die Trigeminusmuskulatur und die tiefe Facialismuskulatur von Orycteropus aethiopicus. Zeithschrift Anat. Entwicklungsgeschichte 116:202-217. Furusawa, K., M. Yamaoka, M. Kogo, and T. Matsuya (1991a) The innervation of the levator veli palatini muslce by the glossopharyngeal nerve. Brain Res. Bull. 26:599-604. Furusawa, K., M. Yamaoka, and T. Kumai (1991b) Properties of the lingual and Ivp branches of the glossopharyngeal nerve. Brain Res. Bull. 28:1-7. Cans, C. (1982) Fiber architecture and muscle function. Exercise Sport Sci. Rev. 10:160-207. Gans, C , and W. J. Bock (1965) The functional significance of muscle architecture: a theoretical analysis. Adv. Anat. Embryol. Cell Biol. 38:115-142. Gans, C , and A. S. Gaunt (1991) Muscle architecture in relation to function. J. Biomech. 24, Supple. 1:53-65. Gans, C , and A. S. Gaunt (1992) Muscle architecture and control demands. Brain Behav. Evol. 40:70-81. Gasc, J.-P (1967) Squellete hyobranchial. Pp. 550-583. In: Traite de Zoologie, Vol. 16. P.-P Grasse (ed.). Masson et Cie, Paris. Gaudin, T. J. (1990) The ear region of the Pilosa (Mammalia, Xenarthra) and the phylogeny of the Tardigrada. J. Vertebr. Paleontol. 10(3, Supple.) :24A. Gaudin, T. J. (1995) The ear region of Edentates and the phylogeny of the Tardigrada (Mammalia, Xenarthra). J. Vertebr. Paleontol. 15(3): 672-705. Gaudin, T. J., and D. G. Branham (1998) The phylogeny of Myrmecophagidae (Mammalia, Xenarthra, Vermilingua) and the relationship of Eurotamandua to the Vermilingua. J. Mammal. Evol. 5(3): 237-265. Gaudin, T. J., and J. Wible (1999) The entotympanic of pangolins and the phylogeny of the Pholidota (Mammalia). J. Mammal. Evol. 6(1): 39-65. Gemmel, N. J., and M. Westerman (1994) Phylogenetic relationships
15. F e e d i n g in M y r m e c o p h a g o u s M a m m a l s within the class Mammalia: a study using 12s RNA sequences. J. Mammal. Evol. 2(1): 3-23. German, R. Z., and H. A. Franks (1991) Timing in the movement of jaws, tongue and hyoid during feeding in the hyrax, Procavia syriacus. J. Exp. Zool. 257:34-42. Gill, T. (1870) Arrangements of the families of mammals and synoptical table of characters of the subdivisions of mammals. Smith. Miscell. Collect. 11:1-98. Gordon, A. M., A. R Huxley, and F. Julian (1966) The variation in isometric tension with sarcomere length in vertebrate muscle fibers. J. Physiol. 184:170-190. Gregory, W. K. (1910) The orders of mammals. Bull. Am. Mus. Nat. Hist. 27:1-524. Griffiths, M. (1968) Echidnas. Pergamon Press, Oxford. Griffiths, M., and K. G. Simpson (1966) A seasonal feeding habit of spiny anteaters. CSIRO Wildlife Res. 11(2): 137-143. Griffiths, M., R. T. Wells, and D. J. Barrie (1991) Observations on the skulls of fossil and extant echidnas (Monotremata: Tachyglossidae). Aust. Mammal. 14:87-101. Griffiths, T. A. (1982) Systematics of the New World nectar-feeding bats (Mammalia: Phyllostomidae), based on the morphology of the hyoid and lingual regions. Am. Mus. Novit. 2742:1-45. Guth, C. (1961) La Region Temporale des tdentates. Le Puy, Haute Loire. Hatt, R. T. (1934) Pangolins. Nat. Hist. 34(8): 725-732. Heath, M. E., and S. L. Vanderlip (1988) Biology, husbandry and veterinary care of captive Chinese pangolins {Manis pentadactyla). Zoo Biol. 7:293-312. Hecht, C. S., P. J. Gannon, and A. R. Eden (1993) Motor innervation of the Eustachian tube muscles in the guinea pig. Laryngoscope 103:1218-1226. Herring, S. W. (1993) Functional properties of the feeding musculature. In: Biomechanics of Feeding in Vertebrates. V. L. Bels, M. Chardon, and P. Vandewalle (eds.). Springer-Verlag, Berlin. Hiiemae, K. (1976) Masticatory movements in primitive mammals. Pp. 105-118. In: Mastication. D. J. Anderson and B. Matthews (eds.). Wright and Sons, Bristol. Hirschfeld, S. E. (1976) A new fossil anteater (Edentata, Mammalia) from Colombia, S.A. and evolution of the Vermilingua. J. Vertebr. Paleontol. 50:419-432. Ibuki, K., T. Matsuya, J. Nishio, Y. Hamamura, and T. Miyazaki (1978) The course of the facial nerve innervation for the levator veli palatini. Cleft Palate J. 15:209-214. JoUie, M. (1968) The head skeleton of a new-born Manis javanica with comments on the ontogeny and phylogeny of the mammal head skeleton. Acta Zool. 49:227-305. Keller, J. T., M. C. Saunders, H. van Loveren, and M. T. Shipley (1984) Neuroanatomical considerations of palatal muscles: tensor and levator veli palatini. Cleft Palate J. 21:70-75. Kier, W. K. (1992) Hydrostatic skeletons and muscular hydrostats. Pp. 205-231. In: Biomechanics—Structures and Systems, a Practical Approach. A. A. Biewener (ed.). Oxford Univ. Press, New York. Kier, W. M., and K. K. Smith (1985) Tongues, tentacles and trunks: the biomechanics of movement in muscular-hydrostats. Zool. J. Linnaean Soc. 83:307-324. Kingdon, J. (1974) East African Mammals: An Atlas of Evolution in Africa. University of Chicago Press, Chicago. Koenigswald, W., G. Richter, and G. Storch (1981) Nachweis von Homschuppen bei Eomanis waldi aus der Grube Messel bei Darmstadt (Mammalia, Pholidota). Senckenbergiana lethaea 61(3): 291-298. Kratzing, J. E., and P. F Woodall (1988) The rostral anatomy of two elephant shrews. J. Anat. 157:135-143. Kruuk, H., and W. A. Sands (1972) The aardwolf (Proteles cristatus
483
Sparmann 1783) as a predator of termites. East Af. Wildlife J. 10: 211-227. Kubota, K., J. Kubota, N. Fukuda, and S. Asakura (1962a) Comparative anatomical and neurohistological observations on the tongue of the great anteater {Myrmecophaga jubata Linne). Anat. Rec. 143: 15-25. Kubota, K., J. Kubota, T. Nakamura, N. Fukuda, S. Asakura, S. Nakagawa, and M. Masui (1962b) Comparative anatomical and neurohistological observations on the tongue of the pangolin {Manis pentadactyla, Linneus). Anat. Rec. 144:43-55. Kiihlhorn, F. (1939) Beziehungen zwischen Ernahrungsweise und Bau des Kauapparates bei einigen Giirteltier und Ameisenbarenarten. Gegenbaurs Morphologisches Jahrbuch 84:55-85. Lubin, Y D., and G. G. Montgomery (1981) Defenses of Nasutitermes termites (Isoptera, Termitidae) against Tamandua anteaters (Edentata, Myrmecophagidae). Biotropica 13(1): 66-76. Lubin, Y D., G. G. Montgomery, and O. P Young (1977) Food resources of anteaters (Edentata: Myrmecophagidae) I. A year's census of arboreal nests of ants and termites on Barro Colorado Island, Panama Canal Zone. Biotropica 9(1):26-34. Lubosch, W. (1908) Das Kiefergelenk der Edentaten und Marsupialier. Nebst Mittheilungen iiber die Kaumuskulatur dieser Thiere. Semons Zoologische Forschungsreisen in Australien und dem Malayischen Archipel. 4(6): 520-555. Luckett, W. P. (1994) Suprafamilial relationships within Marsupialia: resolution and discordance from multidisciplinary data. J. Mammal. Evol. 2(4):255-283. MacPhee, R. D. E. (1994) Morphology, adaptations and relationships of Plesiorycteropus, and a diagnosis of a new order of eutherian mammals. Bull. Am. Mus. Nat. Hist. 220:1-214. Marshall, L. G. (1979) Evolution of metatherian and eutherian (mammlian) characters: a review based on cladistic methodology Zool. J. Linnaean Soc. 66:369-410. McAdam, D. W., and J. S. Way (1967) Olfactory discrimination in the giant anteater. Nature (Lond.) 214:316-317. McKenna, M. C. (1987) Molecular and morphological analysis of higher-level mammalian interrelationships. In: Molecules and Morphology in Evolution: Conflict or Compromise. C. Patterson (ed.). Cambridge Univ. Press, Cambridge. McKenna, M. C. (1992) The alpha crystallin A chain of the eye lens and mammalian phylogeny. Ann. Zool. Fennici 28:349-360. McNab, B. K. (1984) Physiological convergence amongst ant-eating and termite-eating mammals. J. Zool. Lond. 202:485-510. Meeuse, A. D. J. (1963) A possible case of interdependence between a mammal and a higher plant. Arch. Neerland. Zool. 13, Supple.(l): 314-318. Miyamoto, M. M., and M. Goodman (1986) Biomoloecular systematics of eutherian mammals: phylogenetic patterns and classification. Syst. Zool. 35:230-240. Montgomery, G. G. (1985) Movements, foraging and food habits of the four extant species of neotropical vermilinguas (Mammalia: Myrmecophagidae). Pp. 365-377. In: The Evolution and Ecology of Armadillos, Sloths, and Vermilinguas. G. G. Montgomery (ed.). Smithsonian Institution Press, Washington, DC. Montgomery, G. G., and Y D. Lubin (1977) Prey influences on movements of neotropical anteaters. Pp. 103-131. In: Proceedings of the 1975 Predator Symposium. R. L. Phillips and C. Jonkel (eds.), Montana Forest and Conservation Experiment Station, University of Montana, Missoula. Murray, P. F (1981) A unique jaw mechanism in the echidna, Tachyglossus aculeatus (Monotremata). Aust. J. Zool. 29:1-5. Naples, V. L. (1999) Morphology, evolution and function of feeding in the giant anteater (Myrmecophaga tridactula) J. Zool. Lond. 249: 19-41.
484
K a r e n Zich Reiss
Nene, R. V. (1978) A note on the tongue muscle of the scaly anteater, Manis gigantea. J. Biol. Sci. 21(1):48-49. Nishio, J., T. Matsuya, K. Ibuki, and T. Miyazaki (1976a) Roles of the facial, glossopharyngeal and vagus nerves in velopharyngeal movements. Cleft Palate J. 13:201-214. Nishio, J., T. Matsuya, J. Machida, and T. Miyazaki (1976b) The motor nerve supply of the velopharyngeal muscles. Cleft Palate J. 13: 20-30. Norman, J. E., and M. V. Ashley (1994) Molecular systematics of the xenarthrans. In: American Society of Mammalogists, Vol. 271 A, Washington, DC. Novacek, M. J. (1986) The skull of leptictid insectivores and the higher-level classification of eutherian mammals. Bull. Am. Mus. Nat. Hist. 183:1-111. Novacek, M. J. (1993) Patterns of diversity in the mammalian skull. Pp. 438-531. In: The Skull, Vol. 2. J. Hanken and B. K. Hall (eds.). University of Chicago Press, Chicago. Novacek, M. J., and A. R. Wyss (1986) Higher-level relationships of the Recent eutherian orders: morphological evidence. Cladistics 2:257-287. Novacek, M. J., A. R. Wyss, and M. C. McKenna (1988) The major groups of eutherian mammals. Pp. 31-71. In: The Phylogeny and Classification of the Tetrapods, Vol. 2. M. J. Benton (ed.). Clarendon Press, Oxford. Nov^ak, R. M. (1991) Walker's Mammals of the World. Johns Hopkins Univ. Press, Baltimore, MD. Ohnishi, K. (1991) A tentative evolutionary tree of mammalian orders constructed by Hennigian comparison of the amino acid sequences of alpha-crystallin A chain, myoglobin, and hemoglobin alpha chain. Sci. Rep. Niigata Univ. Ser. D (Biol.)(28): 19-31. Oppel, A. (1899) Ueber die Zunge der Monotremen, einiger Marsupialier und von Manis javanica. Zoologische Forschungsreisen in Australien und dem Malayischen Archipel. 4(2): 107-170. Owen, R. A., Sir. (1862) On the anatomy of the great anteater {Myrmecophaga jubata, Linn.). Trans. Zool. Soc. Lond. 4:117-140. Pages, E. (1965) Notes sur les pangolins du Gabon. Biol. Gabonica 1(3): 209-237. Pages, E. (1970) Sur Tecologie et les adaptations de Torycterope et des pangolins sympatriques du Gabon. Biol. Gabonica 6(1): 27-92. Pascual, R., M. Archer, E. O. Jaureguizar, J. L. Prado, H. Godthelp, and S. J. Hand (1992) The first non-Australian monotreme: an early Paleocene South American platypus (Monotremata, Ornithorhynchidae). Pp. 2-15. In: Platypus and Echidnas. M. L. Augee (ed.). Royal Zoological Society of New South Wales, Sydney. Patterson, B. (1978) Pholidota and Tubulidentata. In: Evolution of African Mammals. Maglio and Cooke (eds.). Patterson, B., and R. Pascual (1972) The fossil mammal fauna of South America. Pp. 247-309. In: Evolution, Mammals, and Southern Continents. A. Keast, F. C. Erk, and B. Glass (eds.). State University of New York Press, Albany, NY. Patterson, B., W. Segall, and W. D. Tumbull (1989) The ear region in Xenarthrans (= Edentata: Mammalia). I. Cingulates. Field. Geol. (18): 1-46. Patterson, B., W. Segall, W. D. Turnbull, and T. J. Gaudin (1992) The ear region in Xenarthrans (= Edentata: Mammalia). II. Pilosa (sloths, anteaters), palaenodonts, and a miscellany. Field. Geol. (24): 1-79. Patterson, B. K. (1975) The fossil aardvarks (Mammalia: TubuHdentata). Bull. Mus. Comp. Zool. 147:185-237. Pocock, R. I. (1924) The external characters of the pangolins (Manidae). Proc. Zool. Soc. Lond. 46:707-719. Pouchet, H. (1867) Memoires sur le grand Fourmilier (Myrmecophaga jubata). Paris.
Pough, H. P , J. B. Heiser, and W. D. McFarland (1989) Vertebrate Life. Macmillan, New York. Proske, U., J. E. Gregory, and A. Iggo (1998) Sensory receptors in monotremes. Phil. Trans. Roy. Soc. Lond. B 353:1187-1198. Redford, K. H. (1985) Food habits of armadillos. Pp. 429-437. In: The Evolution and Ecology of Armadillos, Sloths, and Vermilinguas. G. G. Montgomery (ed.). Smithsonian Institution Press, Washington, DC. Redford, K. H. (1987) Ants and termites as food: patterns of mammalian myrmecophagy. Pp. 349-399. In: Current Mammalogy, Vol. 1. H. H. Genoways (ed.). Plenum Press, New York. Redford, K. H., and J. G. Dorea (1984) The nutritional value of invertebrates with emphasis on ants and termites as food for mammals. J. Zool. Lond. 203: 385-395. Reiss, K. Z. (1997a) How the Anteater Got Its Tongue: Structure and Evolution of the Feeding Apparatus in Xenarthra and Pholidota (Mammalia: Eutheria). Ph.D. Dissertation, Cornell University. Reiss, K. Z. (1997b) Myology of the feeding apparatus of myrmecophagid anteaters (Xenarthra: Myrmecophagidae). J. Mammal. Evol. 4(2): 87-117. Retief, J. D., R. J. Winkfein, and G. H. Dixon (1994) Evolution of the monotremes: the sequences of protamine PI genes of platypus and echidna. Euro. J. Biochem. Richardson, P R. K. (1990) The lick of the aardwolf. Nat. Hist. (April): 78-84. Richardson, P. R. K. (1985) The Social Behavior and Ecology of the Aardwolf Proteles cristatus in Relation to Its Food Resources. Doctoral Dissertation, University of Oxford. Richardson, P. R. K. (1987) Aardwolf: the most specialized myrmecophagous mammal? South Af. J. Sci. 83:643-646. Richardson, P. R. K. (1991) Territorial significance of scent marking during the non-mating season in the aardwolf Proteles cristatus (Camivora: Protelidae). Ethology 87:9-27. Rose, K. D. (1988) Early Eocene mammal skeletons from the Bighorn Basin, Wyoming: significance to the Messel fauna. Cour. Forsch.Inst. Senkenberg 107:435-450. Rose, K. D., and R. J. Emry (1983) Extraordinary fossorial adaptations in the Oligocene palaeanodonts Epoicotherium and Xenocranium (Mammalia). J. Morphol. 175:33-56. Rose, K. D., and R. J. Emry (1993) Relationships of Xenarthra, Pholidota, and fossil ''edentates": the morphological evidence. Pp. 8 1 102. In: Mammal Phylogeny, Vol. 2. F. S. Szalay, M. J. Novacek, and M. C. McKenna (eds.). Springer-Verlag, New York. Rose, K. D., L. Krishtalka, and R. K. Stucky (1991) Revision of the Wind River faunas. Early Eocene of central Wyoming. II. Palaeanodonta (Mammalia). Ann. Carnegie Mus. 60(1): 63-82. Rowe, T. (1988) Definition, diagnosis, and origin of Mammalia. J. Vertebr. Paleontol. 8(3): 241-264. Rowe, T. (1993) Phylogenetic systematics and the early history of mammals. Pp. 129-145. In: Mammal Phylogeny, Vol. 1. F. S. Szalay, M. J. Novacek, and M. C. McKenna (eds.). Springer-Verlag, New York. Saban, R. (1968) Musculature de la Tete. Pp. 229-471. In: Traite de Zoologie, Vol. 16. P.-P Grasse (ed.). Masson et Cie, Paris. Saban, R. (1971) Particularites Musculaires des Monotremes: Musculature de la Tete. Pp. 681-732. In: Traite de Zoologie, Vol. 16. P.-P. Grasse (ed.). Masson et Cie, Paris. Savage, R. J. G., and M. R. Long (1986) Mammal Evolution: An Illustrated Guide. Facts On File Publications, New York/Oxford. Schneider, R. (1964) Der Larynx der Saugetiere. Pp. 1-128. In: Handbuch der Zoologie, Vol. 8. Helmcke, Lengerken, Starck, and Wermuth (eds.). deGruyter, Berlin. Schulman, H. (1906) Vergleichende Untersuchungen iiber die Tri-
15. Feeding in Myrmecophagous Mammals geminus-Musculatur der Monotremen, sowie dabei in Betracht kommenden Nerven und Knochen. Zool. Forschungsreisen Australien und dem Malayischen Archipel. 3(2): 303-400. Shoshani, J. (1986) Mammalian phylogeny: comparison of morphological and molecular results. Mol. Biol. Evol. 3:222-242. Shoshani, J. (1993) Hyracoidea-Tethytheria affinity based on myological data. Pp. 235-256. In: Mammal Phylogeny, Vol. 1. R S. Szalay, M. J. Novacek, and M. C. McKerma (eds.). Springer-Verlag, New York. Shoshani, J., and M. C. McKenna (1995) Mammalian evolution vis-avis Asian biogeography. J. Vertebr. Paleontol. 15(3): 53A. Shoshani, J., M. C. McKenna, K. D. Rose, and R. J. Emry (1997) Eurotamandua is a pholidotan not a xenarthran J. Vertebr. Paleontol. 17:76A. Sikes, S. K. (1966) The tricuspid tree pangolin, Manis tricuspis (Rafinesque): its remarkable tongue complex. The Nigerian Field 31(2): 99-110. Simpson, G. G. (1931) Metacheiromys and the Edentata. Bull. Am. Mus. Nat. Hist. 59:295-381. Simpson, G. G. (1945) A classification of the mammals. Bull. Am. Mus. Nat. Hist. 85:1-350. Simpson, G. G. (1978) Early mammals in South America: fact, controversy, and mystery. Proc. Am. Philosoph. Soc. 122(5):318-328. Simpson, G. G., and W. S. Beck (1965) Life: An Introduction to Biology. Harcourt, Brace and World, New York. Smith, K. K. (1986) Morphology and function of the tongue and hyoid apparatus in Varanus (Varanidae, Lacertilia). J. Morphol. 187: 261-287. Smith, K. K. (1992) The evolution of the mammalian pharynx. Zool. J. LinnaeanSoc. 104:313-349. Smith, K. K. (1994) Development of craniofacial musculature in Monodelphis domestica (Marsupialia, Didelphidae). J. Morphol. 222: 149-173. Smith, K. K., and K. H. Redford (1990) The anatomy and function of the feeding apparatus in two armadillos (Dasypoda): anatomy is not destiny. J. Zool. Lond. 222:27-47. Sonntag, C. F. (1923) The comparative anatomy of the tongues of Mammalia. IX. Edentata, Dermoptera, and Insectivora. Proc. Zool. Soc. Lond. XXXIV: 515-529. Sonntag, C. F. (1925) The comparative anatomy of the tongues of Mammalia. XII. Summary, classification and phylogeny. Proc. Zool. Soc. Lond. : 701-764. Starck, D. (1967) Le crane des mammiferes. Pp. 405-549. In: Traite de Zoologie, Vol. 16. P.-P. Grasse (ed.). Masson et Cie, Paris. Storch, G. (1978) Eomanis waldi, ein Schuppentier aus dem MittelEozan der Grube Messel bei Darmstadt (Mammalia: Pholidota). Senckenbergiana lethaea 59(4): 503-529. Storch, G. (1981) Eurotamandua joresi, ein Myrmecophagide aus dem
485
Eozan der Grube Messel bei Darmstadt (Mammalia, Xenarthra). Senckenbergiana lethaea 61(3): 247-289. Storch, G., and J. Habersetzer (1991) Riickverlagerte Choanen und akzessorische Bulla tympanica bei rezenten Vermilingua und Eurotamandua aus dem Eozan von Messel (Mammalia: Xenarthra). Zeitschrifte Saugetierkunde 56:257-271. Storr, G. C. C. (1780) Prodromus methodi Mammalium . . . inaugural disputationem propositus. Fridriech Wolffer, Tubingen. Szalay, R S., and R Schrenk (1998) The Middle Eocene Eurotamandua and a Darwinian phylogenetic analysis of "Edentates''. Kaupia 7: 97-185. Sweeney, R. C. H. (1956) Notes on Manis temmincki. Ann. Magazine Nat. Hist. Lond. Thewissen, J. G. M. (1985) Cephalic evidence for the affinties of Tubulidentata. Mammalia 49:257-284. Van Loveren, H., M. C. Saunders, and J. T. Keller (1983) Localization of motoneurons innervating the levator veli palatini muscle. Brain Res. Bull. 11:303-307. Wayne, R. K., R. E. Benveniste, D. N. Janczewski, and S. J. O'Brien (1989) Molecular and biochemical evolution of the carnivora. Pp. 465-494. In: Carnivore Behavior, Ecology, and Evolution. J. L. Gittleman (ed.). Cornell Univ. Press, Ithaca, NY. Webb, S. D. (1985) The interrelationships of tree sloths and ground sloths. Pp. 105-112. In: The Evolution and Ecology of Armadillos, Sloths, and Vermilinguas. G. G. Montgomery (ed.). Smithsonian Institution Press, Washington, DC. Westerman, M., and D. Edwards (1992) DNA hybridisation and the phylogeny of monotremes. Pp. 28-34. In: Platypus and Echidnas. M. L. Augee (ed.). Royal Zoological Society of New South Wales, Sydney. Wible, J. R. (1991) Origin of Mammalia: the craniodental evidence re-examined. J. Vertebr. Paleontol. 11:1-28. Wible, J. R., and J. A. Hopson (1993) Basicranial evidence for early mammal phylogeny. Pp. 45-61. Mammal Phylogeny, Vol. 1. R S. Szalay, M. J. Novacek, and M. C. McKenna (eds.). SpringerVerlag, New York. Windle, B. C. A., and R G. Parsons (1899) On the myology of the Edentata. I. Muscles of the head, neck and forelimb. Proc. Zool. Soc. Lond. 1899:314-339. Woodburne, M. O., and R. H. Tedford (1975) The first Tertiary monotreme from Australia. Am. Mus. Novitates 2588:1-11. Wozencraft, W. C. (1989) The phylogeny of the recent Carnivora. Pp. 495-535. In: Carnivore Behavior, Ecology, and Evolution. J. L. Gittleman (ed.). Cornell Univ. Press, Ithaca, NY. Zeller, U. (1993) Ontogenetic evidence for cranial homologies in monotremes and therians, with special reference to Ornithorhynchus. Pp. 95-107. In: Mammal Phylogeny, Vol. 1. R S. Szalay, M. J. Novacek, and M. C. McKenna (eds.). Springer-Verlag, New York.
This Page Intentionally Left Blank
C H A P T E R
16 Feeding in Marine Mammals ALEXANDER WERTH Department of Biology Hampden-Sydney College Hampden-Sydney, Virginia 23943 I. INTRODUCTION IL FEEDING STRATEGIES A. Filter Feeding B. Suction Feeding C. Raptorial Feeding D. Grazing III CONCLUSIONS References
or dolphin, is a full-fledged marine mammal by virtue of its subsistence entirely within the marine trophic web (nonetheless, it spends most of its life on and around sea ice and is a very impressive swimmer). Indeed, among the frequent suggestions to explain the reversion of mammals to the oceans, the promise of abundant food is regarded as a strong possibility, especially when considering critical near-shore eutrophication and other productivity changes in many Tertiary habitats (Lipps and Mitchell, 1976). Although different marine mammals variously evolved in the warm, shallow waters of the Tethys Sea during the Paleocene and Eocene (presumed to have been the case for cetaceans, sirenians, and perhaps phocid pinnipeds) or on the rocky, steeply sloping shores of the cool North Pacific beginning in the Miocene (in the case of otariid and odobenid pinnipeds, desmostylians, the polar bear, and marine mustelids), all would have avoided increasing competition from the intense radiations of contemporaneous terrestrial mammals. At the same time, these seagoing pioneers could fill niches vacated by the recent extinction of most marine reptiles. Just as feeding might have provided the original impetus for this major environmental transition, it was also a likely cause of further marine mammal diversification. For example, the cetacean suborder Mysticeti (baleen whales) is believed to have originated as a direct result of the Oligocene development of the Circum-Antarctic current, which created nutrient-rich upwelling and in turn led to huge shoals of zooplankton in the South Pacific (Fordyce, 1977,1980). Patterns of sirenian evolution show close correlation with the spread of temperate seagrasses in the Pacific (Domning, 1978b, 1982). Other examples of speciation and dispersal concurrent with changes in food supply
L INTRODUCTION The eutherians commonly known as "marine mammals" obviously do not constitute a valid systematic taxon, as they comprise diverse mammals of carnivoran, condylarth, or subungulate ancestry (approximately 120 extant species in three orders; Table 16.1, Fig. 16.1). However, like so many other living and extinct vertebrates that have, for various reasons, reverted to an aquatic habitat, all have independently evolved a suite of morphological and ecological characters that reflect their shared environment. Just as their locomotor and reproductive systems adapted to an aquatic lifestyle, so too has the feeding system been affected considerably by the new medium (see Chapter 1). This habitat shift necessitated many radical and often dramatic alterations in the feeding methods and mechanisms of their terrestrial ancestors. Marine mammals exhibit widely varying degrees of morphological and physiological adaptation for aquatic locomotion and reproduction, both because of their diverse ancestry and, perhaps more importantly, the great disparity in their times of origin. Cetaceans, for example, appeared long before marine carnivorans. However, the traditional criterion for inclusion in this polyphyletic assemblage concerns diet and foraging. Thus the polar bear, while a less adept diver than a seal FEEDING (K. Schwenk,ed.)
487
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
488
Alexander Werth
TABLE 16.1 Simplified Taxonomy of Marine M a m m a l s w i t h Estimated N u m b e r of Living G e n e r a / S p e c i e s Order Cetacea Suborder Archaeoceti'' Suborder Mysticeti Balaenidae: right whales (3/3) Balaenopteridae: fin or rorqual whales (2/6) Eschrichtiidae: gray whale (1/1) Suborder Odontoceti Physeteridae: sperm whales (2/3) Ziphiidae: beaked whales (5/19) Monodontidae: narwhal and beluga (2/2) Delphinidae: dolphins (17/34) Phocoenidae: porpoises (3/6) Platanistidae: river dolphins (4/5) Order Sirenia Trichechidae: manatees (1/3) Dugongidae: dugong (1/1) Order Desmostylia" Order Carnivora Suborder Pinnipedia Phocidae: true (earless) seals (10/19) Otariidae: sea lions and fur seals (7/14) Odobenidae: walrus (1/1) Suborder Fissipedia Mustelidae: sea otter (1/1) Ursidae: polar bear (1/1) ''Extinct.
(and, indirectly, to the development of ocean currents, thermal gradients, and upwelling processes) involve the Miocene radiation of pelagic odontocetes (toothed whales), origin of pinnipeds, and extinction of archeocetes (Lipps and Mitchell, 1976). Oceanographic conditions such as thermal stratification, salinity, depth, and currents play critical roles in marine mammal feeding (Murison and Gaskin, 1989; Simenstad et al, 1990; Smith and Whitehead, 1993). Marine trophic webs are deceptively complex. The critical ecological interaction between sea otters and algivorous sea urchins has been well studied (Estes and Steinberg, 1988; Estes et al, 1998), but it is possible that this relation in turn affected the historic distribution of sirenians (Dayton, 1975). Given the tacit importance of feeding and the urgent need for all marine mammals to adopt new feeding strategies rapidly, one might expect to find a small number of recurring solutions. The biomechanical problems associated with prey capture in an extremely dense and viscous medium should lead to rampant convergence and parallel evolution, both within this group as well as relative to other aquatic vertebrates. Thus, despite the broad taxonomic representation within marine mammals, a few basic feeding types predominate.
It is much more effective to clean small floating or suspended particles from a swimming pool with a dip net or vacuum tube than by attempting to grasp them individually. When attempting to grasp an object in the water, movement toward the object tends to push it away due to a preceding pressure wave of water. This problem is confronted by all aquatic predators who approach a food item jaws first. While many marine mammals rely on this raptorial seizing method and some have evolved elongate pincer-like jaws replete with elaborate dentition, many species, particularly within Cetacea (the oldest and perhaps most highly derived marine mammals), rely on the tried-and-true methods of suspension and suction feeding (dip net and vacuum tube approaches, respectively). These are the predominant modes of feeding among fishes and other aquatic vertebrates as well. It must be stressed, however, that although the simple solutions to these hydrodynamic difficulties are fairly universal, marine mammals (and indeed all marine tetrapods, particularly birds and reptiles) were obliged to remodel functional complexes evolved for feeding in air, and thus had a critical starting point far from that of distant piscine ancestors. In particular, the mammalian hyolingual apparatus is notably distinct from that of fishes in construction and proportions, as are the dentition, oral cavity, and pharynx (Chapter 2). Likewise it is important to remember that the feeding mechanics of fish and aquatic mammals are fundamentally different, even if basic principles remain the same. Thus, from a functional standpoint there are numerous underlying similarities but there are as many key differences. Filter feeding may be the primitive mode of prey capture for many lower vertebrates, but it is a secondary derivation for mysticete whales, as it was for certain marine reptiles of the Mesozoic Era. Because of marine mammals' relatively recent adaptation to a habitat immensely different from that of their immediate progenitors, it is often possible to separate the dual factors of internal and external evolutionary pressures—the competing influences of heredity and environment—and therefore to determine which of their features are functionally adaptive and which are merely vestiges of their terrestrial ancestry. Few organisms offer the opportunity to distinguish so clearly the plesiomorphic and apomorphic characters that create an evolutionary mosaic, and thus marine mammals constitute an ideal group with which to study the pattern and process of evolution. An example related to feeding is the compartmentalized cetacean stomach, cited by some as a practical (i.e., adaptive) derivation (Hosokawa and Kamiya, 1971) yet by others as a historical holdover from ruminant ancestry (Zhou, 1982). This chapter explores marine
MONOTREMATA MARSUPIALIA PHOLIDOTA XENARTHRA - CARNIVORA INSECTIVORA MACROSCELIDEA LAGOMORPHA RODENTIA PRIMATES SCANDENTIA DERMOPTERA CHIROPTERA TUBULIDENTATA ARTIODACTYLA CETACEA PERISSODACTYLA HYRACOIDEA SIRENIA PROBOSCIDEA tDESMOSTYLIA F I G U R E 16.1. A cladistic phylogeny of mammal orders based on Novacek (1992). The position of Desmostylia is based on McKenna and Bell (1997) (some consider them to be more closely related to sirenians). Orders containing marine mammals are in bold. Cetacea, Sirenia, and Desmostylia are entirely aquatic, whereas only some families within Carnivora contain marine taxa (Phocidae, Otariidae, Odobenidae, Mustelidae, and Ursidae). Among these families, the first three are obligate marine taxa, but ursids and mustelids have only one or a few (respectively) marine species. The Carnivoran famihes Phocidae, Otariidae, and Odobenidae have been regarded traditionally as a separate order, Pinnipedia, but current consensus places them within the Carnivora. To include them in a separate order would render the Carnivora a paraphyletic taxon. In any case, monophyly of 'Tinnipedia'' is in dispute; the families may have arisen from different carnivoran ancestors. Desmostylia is the only extinct taxon included in the cladogram.
490
Alexander Werth
mainraal feeding from a historical perspective, including consideration of the morphology and systematics of extinct taxa. However, marine mammal phylogenetics remains a highly controversial subject; even such simple presumptions as the monophyly of Cetacea and the pinniped carnivorans are hotly contested. Thus, relationships presented in Fig. 16.1 should be considered tentative. Despite the apparent advantages for preservation in aquatic environments, the fossil record of marine mammals is spotty and incomplete, particularly in early periods of evolution. Fortunately, ongoing molecular analyses may shed light on marine mammal phylogenies. Despite recent technological advances that have greatly improved our understanding of vertebrate feeding generally, much remains to be learned regarding marine mammals. This is not surprising for they are notoriously difficult animals to study in captivity, let alone in the wild. Research is hindered by obvious logistical and fiscal limitations (mostly stemming from their size and remote location) as well as by equally formidable legal restrictions. Many marine mammals remain endangered or severely threatened due to hunting and habitat destruction. Thus there is an inability to apply standard techniques of vertebrate morphology (e.g., electromyography via surface or surgically implanted electrodes, high speed video, cineradiography) because it is neither feasible to maintain marine mammals as laboratory subjects nor possible (if permissable) to use even mildly invasive or stressful procedures in the field. Nonetheless, marine mammals remain a perenially popular choice of research subject, and despite the challenges, biologists have made great advances in our knowledge of their feeding. Perhaps due to the paucity of direct experimental data, considerable attention has been paid to feeding behavior, such as the bubble netting of humpback whales {Megaptera novaeangliae) to corral prey, the remarkable diving capabilities of Weddell seals {Leptonychotes weddelli), and the ubiquitous tool use of sea otters {Enhydra lutris) to crack open hard-shelled molluscs and echinoderms. Research on captive animals is often limited to husbandry or behavioral studies. Indeed, vastly more is known about marine mammal diet and foraging ecology than of the actual mechanics of prey capture, ingestion, transport, and swallowing. Many details of marine mammal food and feeding have been gleaned from centuries-old lore of European and American whalers or from the Inuit, South Pacific islanders, and other native cultures—all venerable and generally reliable sources. Nevertheless, this information is a poor substitute for controlled experiment or observation as in most cases it has not been, nor can it be, independently substantiated.
Much of our understanding of the mechanical aspects of marine mammal feeding comes from speculative extension of anatomical knowledge, i.e., the inference of function from form (see Chapter 1). While in many cases this n\ay constitute a reasonable approximation of function, this information must be considered conjectural, for in the absence of experimental evidence (or even underwater observation of animals in natural or captive conditions) there is no way to verify it directly. However, despite the many, admittedly critical, differences between marine and terrestrial mammals, there is little reason to suspect that they differ in either the basic principles or the actual details of oral processing and swallowing. There do, however, appear to be major exceptions, such as the masticatory apparatus of sirenian jaws, which deviates markedly from that of other mammalian grazers. In cetaceans, the permanent intranarial larynx and attendant changes in pharyngeal constrictor musculature with a patent airway presumably lead to divergent mechanisms of deglutition, although this is not certain. The aim of this chapter is to review what is currently known about function and relevant structures of marine mammal feeding, but the reader should be forewarned that this information often consists purely of basic anatomical and ecological data (e.g., dental formulae, gross myology; stomach contents, captive foraging observations) or other nuggets of natural history. Ample published literature exists concerning marine mammal digestive systems, although morphological description here will largely be restricted to the oral region. Unless otherwise noted, all details of marine mammal ingestion, transport, processing, and deglutition are presumed to be similar to the basic mammalian conditions, as described elsewhere in this book (Chapters 2 and 13). Marine mammals differ from other aquatic animals in obvious yet critical ways that influence their feeding. As endotherms their feeding must be sufficient to support a high metabolic rate, yet although their elevated body temperature permits a high degree of activity and enables them to exploit many cold environments, it also constrains their morphology so as to prevent undue heat loss. As obligate air breathers, marine mammals must dwell near to the surface, or at least return there periodically. They cannot benefit from pharyngeal slits that ventilate the gills of fishes and larval amphibians and permit a unidirectional flow of prey-laden water into the oral cavity; in marine mammals, flow is necessarily bidirectional, i.e., all water that enters the mouth must be expelled from whence it came or be swallowed (excepting balaenid mysticetes, which circumvent this problem and achieve unidirectional flow because water enters the mouth anteriorly yet exits it posteriorly). Strong pharyngeal and laryngeal musculature prevents
16. Feeding in Marine Mammals
water from entering the trachea and filling the lungs when the mouth is open underwater. Several marine mammals undertake long migrations that carry them far from productive feeding grounds, necessitating long bouts of fasting. This means that certain species must not only build up plentiful nutritional stores for extended periods of time, but also maximize their intake during periodic bouts of feeding. Finally, because marine mammals must suckle (Chapter 14), their feeding apparatus is obliged to accommodate both neonate/juvenile suckling and adult feeding or else undergo some transformation between juvenile and adult phases of their life history. Not only does this affect the feeding strategies of marine mammals but also, of course, their suckling behavior and mammary morphology. In cetaceans, for example, milk is not so much pulled from the mammary gland via sucking action of the tongue as ejected by the contraction of smooth muscles surrounding the nipple (Arvy, 1974). Lactation itself also places many constraints on marine mammals: small litter size, a single pair of internal mammary glands, whose nipples are everted through slit-like openings (to maintain streamlined profile), and exceptionally fatty milk (to prevent water loss). Costa (1991) and Olesiuk (1993) discuss foraging behavior and energy budgets in relation to pinniped reproduction and life history patterns. Marine mammals occupy a unique and fascinating position, for while they are vastly unlike other aquatic vertebrates, they are at the same time highly divergent in comparison to other mammals, particularly in terms of that most celebrated mammalian feature, dentition. In numbers of teeth, marine mammals variously exceed or fall short of the typical eutherian pattern. Not only are teeth typically reduced in quantity, but they are also reduced in differentiation along the tooth row (reduc-
491
tion in heterodonty), reduced in generations, and reduced in number and complexity of cusps. Dental formulae of marine mammals cannot be considered diagnostic aids as individual variation is common. Mysticetes are edentulous as adults, and odontocetes may possess teeth far in excess of the usual eutherian maximum of 42. In the sperm whale {Physeter catodon) only mandibular teeth erupt, whereas in Risso's dolphin {Grampus griseus) there are no maxillary teeth. In many ziphiid whales teeth erupt only in males. Odontocete teeth are typically conical in shape, without cusps or shelves, and occur in just one generation. All of these circumstances—polyodonty, homodonty, and monophyodonty—diverge from standard eutherian conditions. Incisor and canine tusks are found in diverse marine mammals such as the narwhal, walrus, and dugong. Manatee teeth are replaced horizontally throughout life, as in proboscideans. In fact, aside from herbivorous sirenians, there is typically little if any reduction or other processing of food in the oral cavity, although teeth may be used to slash or tear pieces from large prey in some species. Mastication is the hallmark of mammals, yet few marine mammals chew their food. The mouth is specialized for food acquisition alone. The cetacean skull (Fig. 16.2) nicely illustrates some of the profound ways in which marine mammal cranial morphology departs from that of terrestrial mammals. In the absence of mastication the primary mammalian jaw adductors (masseter and temporalis) are reduced; pterygoids are now the dominant jaw-closing muscles. Consequently the zygoma are reduced to slender, thread-like or styliform rods in odontocetes, whereas jugal bones are likewise limited to small stubs in mysticetes; in neither case is this arch firmly fused to the squamosal. Jaws and braincase are greatly modified, as the posterior narial migrations form a long
a.
FIGURE 16.2. Skulls of mysticete and odontocete cetaceans: (a) right whale, Eubalaena glacialis, and (b) long-finned pilot whale, Globicephala melaena.
492
Alexander Werth
"telescoped" rostrum with concomitant elongation of numerous contributions of the dermatocranium andneurocranium (and asymmetrical elements in some species). The rami of the dentaries (straight and compressed in odontocetes, arched and rounded in n\ysticetes) bear rudimentary coronoid and condyloid processes, except in balaenopterid mysticetes, where the coronoid is associated with a ligamentous band that aids in mandibular kinesis. Mammalian carnivores typically exhibit a scissor-like jaw action with a jaw joint that permits no room for lateral motion, yet cetaceans have a rounded mandibular condyle instead of a hingelike one. Bite force is often weak in cetaceans, and except for mysticetes (particularly rorquals and balaenids), there is little "play" or mobility in the mandibular symphysis. The enlarged, flattened hyoid apparatus of cetaceans consists of several ossified and cartilaginous elements articulated by synovial joints. In the absence of a mobile soft palate and epiglottis, the palatal and pharyngeal musculature is significantly altered. Whereas other marine mammals do not demonstrate such extreme cranial modification, their musculoskeletal systems likewise reflect adaptations for feeding in water. The head is obviously an important functional complex. Numerous manifest activities and developmental processes affect marine mammal oral morphology, including modification of nares and nasal passages for air breathing at the surface and streamlining for improved locomotion. Hearing and other special senses are modified significantly in marine mammals, not only to suit their new environment, but as direct consequences of alterations in craniofacial anatomy for feeding. The marine mammal cochlea is attuned to infra- and ultrasonic frequencies that travel well in water. The odontocete "melon" used to focus sonic pulses greatly affects their cranial profile. While cetaceans are well known for their exceptional echolocation and other auditory abilities, they possess no olfactory sense and the presence of a gustatory sense is questionable. Vision is important for foraging, particularly in pinnipeds. Although marine mammals are an admittedly diverse group, several basic generalizations relevant to their feeding can be made. As noted, many fast for extended periods of time, up to half a year (e.g., summer in mysticetes, winter in polar bears). Sexual dimorphism is not uncommon; however, aside from dentition, this does not usually affect diet or feeding method. The stomach is often complex and compartmentalized. The alimentary canal is typically long, without clear transition between large and small intestines, and digestive organs such as the gall bladder may be absent. Several marine mammals, particularly pinnipeds and odontocetes, are known to ingest stones and pebbles inciden-
tally as well as intentionally. These gastroliths may serve as a grinding mill for mechanical digestion in the stomach, especially since prey are often swallowed whole or "bolted" (Owen, 1980), or they may aid in buoyancy regulation. It is not unusual for marine mammals to ingest a variety of other inorganic objects, especially man-made pollution. Concretions or pathological obstructions of the intestine (e.g., ambergris) are found occasionally. Cooperative foraging (with conspecifics, or other marine mammals or birds) is common, and a few odontocetes harken to their terrestrial ancestry by chasing fish onto river banks or surfing onto beaches to snatch sea lions. Pinnipeds and sea otters are so well adapted for aquatic locomotion that although they come ashore to mate and give birth, they are ungainly on land and ill-suited to catching prey there. Despite dwelling in saline habitats, some marine mammals (manatees and sea lions) demonstrate an occasional need for fresh drinking water; others gain sufficient water fron\ their animal or plant diet, from metabolic water production and blubber degradation, and from physiological or behavioral adaptations of digestion, excretion, reproduction, and thermoregulation.
IL FEEDING STRATEGIES Marine mammals employ a variety of strategies to capture and ingest prey. They subsist on virtually every multicellular form of life (including other marine mammals) found in every marine environment, from intertidal or littoral benthic zones to neritic, pelagic, and abyssal habitats of all ocean basins. Their food ranges from tiny zooplankters to large nekton (e.g., squid, sharks, bony fishes) and is obtained at all levels of the water column, including the surface and bottom. Their prey range from nonmotile to highly elusive and can be solitary or gregarious, soft, or covered with a hard shell or test (the latter often cracked or broken for feeding, but sometimes swallowed intact). As a group, marine mammals are predominantly piscivores, although most filter feeders prey on small arthropods such as euphausiids, amphipods, copepods, and mysids. Many odontocetes, especially suctionfeeding species, are almost exclusively teuthophagous (cephalopod-eating). With few exceptions (notably the killer whale, Orcinus orca), all carnivorous marine mammals eat prey whole. Baleen whales are often described in popular literature as grazers; however, this term is properly reserved for animals that consume herbage, algae, or phytoplankton rather than zooplankton, so that carnivory is an appropriate if unlikely term for mysticete feeding. However, the
493
16. Feeding in Marine Mammals presence of volatile fatty acids and symbiotic bacteria in mysticete forestomachs suggests microbial fermentation (Herwig et ah, 1984). Sirenians are the only herbivorous marine mammals; they may browse (i.e., eat leaves, stems, or other portions of plants), but they normally crop all surface growth or consume the entire plant. Catalogues of diet are commonplace and will not be reviewed here. However, there have been no previous attempts to compile a comprehensive, comparative survey of the methods and mechanisms of marine mammal feeding. Because of the tremendous diversity (taxonomic and ecological) within this artificial grouping, this chapter is outlined in a slightly different format, presenting information on structure, function, and evolution by feeding type rather than for marine mammals as a unified whole. Marine mammals can be categorized generally as filter feeders, suction feeders, raptorial ("predatory") feeders, and grazers. These groupings cross taxonomic boundaries. Aside from grazing (Sirenia alone), none of these four categories contains a single taxon exclusively. Cetaceans and pinnipeds have filter, suction, and raptorial representatives. The diversity of these taxa cannot be overemphasized. Still, the main phylogenetic division of Cetacea relates to feeding (hence the old subordinal names Filtrales and Raptoriales), and several researchers have attempted an ecological or systematic classification of cetaceans according to diet and feeding type (Tomilin, 1954). For example, Gaskin (1976) divides odontocetes into ichthyo-, teutho-, and sarcophagi, yet admits that such rigid classifications are impractical, as cetaceans, like all marine mammals, are opportunistic feeders whose preferences are dictated by seasonal migrations and circumstances of prey distribution, depth, and abundance (Frost and Lowry, 1981; Kim and Oliver, 1989; Goebel et a/., 1991; Boyd et ah, 1994). The four categories can be further subdivided into ecological guilds on the bases of habitat, prey size, and locomotion, but there is likely to be considerable overlap in these and other traits. It goes without saying that numerous mammals of various orders inhabit or visit freshwater habitats regularly. Fiowever, none demonstrate the level of separation from land seen in marine mammals, nor do any exhibit the extreme range of anatomical and physiological specialization resulting from tens of millions of years of adaptation to life in water. With few exceptions (e.g., the duck-billed platypus, Ornithorhynchus anatinus), and aside from rare differences in diet and foraging behavior, feeding in aquatic monotremes, marsupials, insectivorans, rodents, and ungulates does not differ notably from that of their terrestrial relatives. The same can be said for marine fissiped carnivorans, for even the
truly marine sea otter and polar bear do not typically ingest prey underwater. A. Filter Feeding Filter or suspension feeding (straining small prey items or particulate organic matter suspended in water) can claim a proud history among vertebrates, as it is not only presumed to be the primitive mode of chordate feeding but has typically been used by the largest aquatic animals, extant and extinct. Among fishes, the largest chondrichthyans (the manta ray and whale, basking, and megamouth sharks) are filter feeders, as are paddlefish, mackerel, and other large osteichthyans. Filter feeding has also been adopted by several tetrapod lineages that reverted independently to an aquatic heritage. Various reptiles from plesiosaurs to pterosaurs were filter feeders, as are some larval amphibians, several aquatic birds (e.g., ducks, flamingos), and some mammals. A filter feeding strategy allows animals to feed near the bottom of a trophic pyramid and thus to reap the rewards of greater biomass (and therefore more enegy) available for consumption, permitting filter feeders to attain giant body size (or, conversely, to support huge populations of smaller animals). Some of the earliest mysticete whales measured only 4 or 5 m (still rather large), yet Recent baleen whales are without question the largest animals that have ever lived. The blue whale (Balaenoptera musculus) is the terminal predator in an abbreviated food chain that often includes just diatoms (Fragilariopsis) and krill (Euphausia). Short trophic pyramids also permitted mysticetes' formerly extensive distributions prior to their decimation by hunting. The mammalian suborder Mysticeti ("mustached whales") comprises three families of baleen whales, each of which employs a distinct type of suspension feeding (Fig. 16.3). Like other tetrapods, most mammals are intermittent rather than continuous filter feeders, i.e., they engulf a single mouthful of water at a time and separate food from this water before expelling it, unlike lower vertebrates that generally pump or push water constantly and unidirectionally through the mouth (Sanderson and Wassersug, 1993). Preyladen water may be ingested either by ram filter feeding, as in balaenopterid mysticetes, or by suction filter feeding, as in the gray whale {Eschrichtius robustus). The distinction depends on how water is drawn into the mouth: by forward motion in the former case and rapid expansion of the buccal cavity to create negative intraoral pressure in the latter. In both cases the ultimate cause of water influx is muscular contraction, with energy supplied by the locomotor or feeding apparatus
494
Alexander Werth
FIGURE 16.3. Mysticete filtering methods: (a) balaerdd (right whale) surface skim feeding, (b) rorqual (humpback, Megaptera novaeangliae) lunge feeding (compare to Fig. 16.7), and (c) gray whale (Eschrichtius robustus) benthic suction feeding.
(tail or tongue, respectively), and both methods are effective at delivering water to the filtering device, the plates and filters of baleen. Members of Balaenopteridae, commonly called rorquals (from the Norwegian "furrowed whale," referring to their pleated throats), rely on ram "gulping" in which a single mouthful of prey-laden water is engulfed by rapid lunges of forward locomotion. As the mouth is closed to expel water, schooling fish, krill, or other micronekton are trapped by the baleen sieve. The gray whale (Eschrichtiidae) is also an intermittent filter feeder, yet it can feed while stationary; its mouth is filled by rapid tongue depression, enlarging the oral cavity to draw in benthic invertebrates, which again are filtered from water and sediment as water is forced from the mouth. A third family, Balaenidae, utilizes continuous ram filter feeding. Right whales slowly skim or "graze" copepods and other minute zooplankton that deposit on finely fringed baleen by a steady stream of water flowing through the mouth. With such substantial variation in mysticete diet and foraging behavior, it is not surprising that each family also possesses a distinct suite of morphological characters associated with feeding, including differences in the filter itself. Baleen is commonly known as whalebone, although it is not an ossified or mineralized tissue, but rather a keratinaceous integumentary product that develops
along the margins of the palate and is suspended from the maxillae in large, flexible laminae or plates (Fig. 16.4). Baleen ranges in color from creamy white or yellow to olive, gray, and black. Each side or "rack" may contain (depending on the species) from 100 to 400 triangular plates of baleen, ranging from 50 cm to nearly 5 m in length and only a few millimeters in thickness. The plates are arranged in transverse series, like the teeth of a comb, at intervals of roughly 1 cm. The base of the scalene triangle is rooted to a foundation layer in the gum, whereas the long outer edge is smooth and the medial side (the hypotenuse) is frayed with bristles or "hairs" that intertwine to form a fibrous mat. The continuously skim-feeding balaenids possess 250-350 plates of narrow, flexible, finely fringed baleen plates, up to 5 m long in bowheads and with 35-70 bristles per cm^ (Leatherwood et ah, 1983). The rorquals and gray whale have much wider, shorter baleen, 0.5-1 m long, with much coarser fringes. As might be expected, the porosity of the filter (number of plates and fringes per plate) correlates strongly with prey size. The finer toothed comb of balaenids captures tiny zooplankters that would pass easily through the filters of other whales, primarily piscivores or krill eaters (although the Sei whale, Balaenoptera horealis, also feeds on small zooplankton and possesses the finest fringes of any rorqual). However, the simple relation between strainer
495
16. Feeding in Marine Mammals
F I G U R E 16.4. Mouth of bowhead whale, Balaena mysticetus, showing arrangement of baleen plates.
and prey is in fact more complex, as the straight inner surface of balaenid baleen means that few fringes can be exposed here. Pivorunas (1976) suggested that smaller fringes alleviate the shortcomings of fewer fringes and
presented a mathematical model correlating plate features with fringe characteristics. Where the plates angle outward, as in rorquals, many fringes are exposed on the medial surface by friction (from water, tongue, and prey) so that the coarse fibers in this dense, tangled, interwoven mat need not have such a small diameter. Baleen is neither homologous nor histologically related to teeth. In fact, tiny tooth buds develop in most mysticetes, but these anlagen disappear before birth. Baleen appears to be homologous to the horny palatal ridges of artiodactyls, which is not surprising given their close phylogenetic affinity (Graur and Higgins, 1994). Nonetheless, teeth and baleen share some developmental similarities. Like teeth, baleen growth begins as a dermal/epidermal interaction (Fig. 16.5). Long, slender, conical papillae extend from an underlying
inner fringes keratin sheath outer fringes
papilla
baleen plate
horn tube (papilla and keratin) epithelium (gingiva) base of papillary process upper jaw bone intermediate layer (compacting horn) basal (foundation) plate F I G U R E 16.5. (1979).
Diagrammatic section of baleen developing from palatal dermal process. After Pivorunas
496
Alexander Werth
basal plate of connective tissue, the dermal process, which overlays the bones of the upper jaw and is in turn covered by a layer of epidermis in which the papillae are enveloped with a horny layer of keratin to form long tubular bristles or horn tubes. These tough, fibrous strands are likewise surrounded and cemented together by a layer of compacting horn, although the innermost tubes are loosely packed without connective tissue so that they can slide relative to one another. A softer, cushioning layer of intermediate horn substance provides a dense cortex covering the anterior and the posterior faces of the plate. The more lateral horn tubes grow faster than the medial ones, so that the long (smooth) side of the triangle faces the outside of the mouth, whereas friction abrades the matrix medially, wearing away compacting horn to reveal the hollow horn tubes, which remain as fine or coarse fringes to serve as the sieving apparatus. Cells in the rubbery, pliant epithelium of the gums anchor baleen to the palate and divide to replenish abraded gingival tissue, just as the base of each papilla grows to replace the worn-out horn tubes. Although the visible portions of baleen consist exclusively of dead, cornified cells, the dermal process remains a living tissue so that baleen is analogous to the part-living tissue in a horse's hoof (Slijper, 1962). The anisotropic nature of this growing tissue's construction, with a homogeneous cortical layer covering free, hollow, cylindrical tubes, affords maximal strength with minimal mass. Baleen is strong yet pliant and elastic, a suitable material to meet the demands of constant exposure to friction and abrasion. Unfortunately, the physical characteristics of this material also made it an extremely precious commodity that helped to fuel the whaling industry. Baleen whales eat whole organisms, whereas most tiny filter feeders, with correspondingly diminutive filters, feed on detritus (fragmented organic debris). However, like other strainers, mysticetes are not selective; they locate patchy sources of food and swallow whatever they trap. In fact, all marine mammals are opportunistic feeders and ingest their share of rubbish, but grazing, raptorial, and suction feeders exercise more discrimination in the items they ingest. Although baleen fringes probably capture prey simply by passive sieving, other mechanisms best known from studies of suspension feeding fishes—direct interception, inertial impaction, and gravitational deposition—maybe possible (Rubenstein and Koehl, 1977). Mysticetes do not rely on mucus or other sticky substances to trap prey. As noted, the porosity of the filter determines prey size, but fine filters catch large as well as small prey and hence may be more versatile. However, dietary studies indicate that coarse-fringed mysticetes are less specialized (Nemoto, 1959, 1970), ingesting large items ranging
from fish to hapless seabirds. Because baleen is not a rigid material, filter porosity varies according to hydrodynamic factors such as velocity of ram locomotion, size and density of prey, and direction and pressure of water flow (Sanderson and Wassersug, 1990). While very tiny filter feeders (such as many marine invertebrates) must contend with difficulties of moving in a dense, viscous fluid, mysticete kinematics depend more on inertial than on viscous forces. Mysticetes can exploit certain aspects of water flow and incompressibility, since by virtue of their size and speed they operate at the opposite end of the spectrum, with Reynolds numbers many orders of magnitude distant from invertebrate filter feeders. Balaenid whales (of which there at least three species: bowhead, northern/southern right, and pygmy right) are slow, rotund cruisers that filter a constant current of water through their mat of finely fringed baleen to skim dense swarms of zooplankton collectively designated by whalers as "brit" (Watkins and Schevill, 1979; Lowry and Frost, 1984, Mayo and Marx, 1990; Carroll et al, 1987). As in whale and basking sharks, the head of balaenids is enormous, measuring up to onethird of their 10- to 20-m length. The upper jaw is strongly curved (Fig. 16.2) to fit the exceptionally long, narrow baleen plates, which fold back when the scoopshaped lower jaw is closed. The giant head functions as a plankton tow net, although it is of course not pulled, but pushed while feeding at about 5 k m / h r so that these whales also share the gradual pace and gentle placidity of the largest sharks. Balaenids are the only mysticetes with a large anterior cleft between the left and the right baleen racks (Fig. 16.4). Water enters this subrostral gap, flows through the oral cavity and between baleen plates, and passes out a gutter-like orolabial sulcus at the rear of the lower lip (Fig. 16.3). It has been supposed that the great pressure drag during continuous filtration slows balaenids and prevents capture of large or evasive prey. Watkins and Schevill (1976) noted that water backs up into the mouth during surface feeding, with a level higher than that of the surrounding sea (although gravity would then force water through the baleen). Lambertsen et al. (1989) conducted a thorough photogrammetric study of bowhead baleen to document its curvature and speculate as to its role in establishing currents of water flow through the oral cavity. As in all mysticetes, plates curve somewhat from medial to lateral aspect so that they appear slightly C-shaped when viewed in transverse section, and the lateral edge of each rack bows outward as well. This creates a narrowing space between the baleen and lips that may be responsible for an inertial rather than passive flow of prey-laden water that enters the mouth, as the decreased space increases pressure and hence flow
16. Feeding in Marine Mammals velocity, setting up a Bernoulli effect (Lambertsen et a/., 1989). As this water flows posteriorly, water still in the center of the buccal cavity is induced to flow medially (perpendicularly) through the plates by a Venturi effect. Lambertsen et al (1989) also studied the morphology of the lower jaw that allows the lip to rotate outward, varying the width of this channel. Werth (1995) devised mathematical and physical models of the bowhead oral cavity to test these biomechanical predictions. Use of such hydrodynamic, rather than simple hydraulic forces, allows balaenids to reduce turbulent flow, improve filtering efficiency, and avoid creation of an anterior pressure wave so that they may in fact be able to capture elusive prey even at slow swimming speeds (Lambertsen et a/., 1989; Werth, 1995). Although these hydrodynamic effects actually draw water into the mouth, the negative intraoral pressures, if any, would be negligible and should not be considered akin to suction feeding by any means. Wiirsig (1988) suggested that foraging in tight formations also allows balaenids to consume evasive prey. Although bowhead and right whales are often observed feeding on the surface or lower in the water column, there is little doubt that they skim along the bottom as well, as evidenced by scratched, muddy heads and by the presence of benthic crustaceans in stomach contents. Right whales also feed while "tail sailing" (Hamner et al, 1988). Balaenids have an uncanny knack for locating dense slicks of zooplankton in which they turn back and forth like mowers cutting fields of grass (as beautifully described in Moby Dick; Melville, 1851). Presumably they find and judge the density of such swarms visually. Regrettably, extremely limited visibility precludes underwater filming of balaenid feeding, although films from clearer summer breeding grounds (Werth, 1990) suggest that the massive, firm, elevated tongue (the largest muscular organ in the world at lengths of nearly 5 m in large animals) sweeps laterally to deflect or channel water to the baleen racks, as suggested by Nemoto (1959) and Gaskin (1982). Rorquals (Balaenopteridae) display the most extreme morphological and mechanical specialization of any mysticete. This family includes the familiar humpback whale {Megaptera novaeangliae) and five species of the genus Balaenoptera (including blue and fin whales), all longer, slimmer, and faster than balaenids, enabling them to draw in mouthfuls of prey-laden water on an enormous scale through rapid gulps and lunges. This is accomplished by inward folding of the distinct, yet loosely organized and weakly muscular tongue into a ventral space with subsequent massive gular expansion. This pouch, which receives engulfed water and the displaced, bag-like tongue, also extends over the thorax to the umbilicus, momentarily giving the body a spec-
497
tacular bloated tadpole shape (Figs. 16.3 and 16.7). Storro-Patterson (1981) calculated that the total volume of a large, hypothetical blue whale (B. musculus) increases over 600% during feeding, from roughly 5550 to 35,700 cubic feet! He further speculated that as much as 1000 tons of water are engulfed in a single gulp. Pivorunas (1979) provided a more conservative estimate of at least 60 m^ (approximately 70 tons) of water ingested in one gulp, still an incredible amount of water equal to roughly 50% of a blue whale's total body volume. While gulping is not so pronounced in other rorquals, this impressive capacity for gular distention is seen in all balaenopterids. Despite trivial physical variation (apart from length) among species within the genus Balaenoptera, there appears to be considerable behavioral divergence and ecological partitioning. The sei whale (B. borealis) has finely fringed baleen and prefers to skim copepods (Kawamura, 1974), whereas Bryde's whale (B. edeni), another moderately sized rorqual, mainly eats fish (Nemoto, 1959). The versatile foraging oiMegaptera includes bottom feeding (Hain et al, 1995). Once depicted as stout dirigibles (based on animals stranded or hauled out on flensing platforms), underwater photographs now confirm balaenopterids' exceptionally streamlined form, particularly in Balaenoptera. Rorquals are built for speed, and rapid locomotion powers their feeding. Rather than expanding a space to create negative pressure and draw water in, rorquals unlock their jaws and relax adductor musculature to open the mouth (at least 30° and up to 90°), which suddenly fills with water in much the same way as a bag is opened by rapidly pulling it through air. Positive inertial pressure forces open the space into which the water and prey flow; seawater is passively enveloped rather than displaced forward or sucked internally (Orton and Brodie, 1987). However, smaller rorquals may also retract the tongue somewhat to pull in a stream of water that fills the oral sac toward the rear rather than the sides (Pivorunas, 1979). In any case, the goal is to avoid a wave of resistant pressure that would disperse prey at the entrance to the mouth and thus interfere with their capture. Because the expansive pouch allows huge quantities of water to enter, pressures build only when the incurrent stream slows (Pivorunas, 1979). Orton and Brodie's mathematical analysis (1987) confirmed that rorqual feeding can be powered solely by locomotion. Because gular distention depends on swimming speed, precise timing with respect to prey location is critical for successful engulfment. If jaws are opened a moment too soon the filled mouth pushes the intended prey ahead with a compressive bow wave (Brodie, 1977). Key anatomical innovations for feeding in this family include the loose tongue (distinct yet flaccid and deformable, with a central furrow) and floor of the mouth
498
Alexander Werth
(with broad, flattened mylohyoid and geniohyoid); accordion-like longitudinal elastic throat pleats; wideopening jaws with a locking joint that creates a hydrostatic seal to prevent unintentional jaw opening during rapid locomotion, and a frontomandibular stay that stores kinetic energy for jaw closure; unfused mandibular symphysis with Y-shaped fibrocartilage arms extending along the mandibular rami (Pivorunas, 1977); and flat, streamlined rostrum. The cavum ventrale, named and described by von Schulte (1916), is an intermuscular fascial cleft extending under the mouth, throat, and chest. The slick walls of this space slide freely over adjacent surfaces and the cavum enlarges to form a giant, bulging vestibule as it receives engulfed water and the displaced tongue and oral floor (Pivorunas, 1979). Dissection of fetal and adult rorquals (Pivorunas, 1979) reveals major ontogenetic changes in the tongue as it is transformed from a solid, muscular structure used in suckling to the deformable, flaccid sheet seen in adults. From around birth until weaning, the paired lingual muscles separate from the midline septum and muscle tissue is replaced with adipose and elastic connective tissues as the increasingly saccular
cavum ventrale
organ flattens and spreads laterally. Intrinsic lingual muscle fibers are scattered and poorly developed (Fig. 16.6). As in balaenids, the rorqual tongue may also serve as a seasonal store of adipose tissue (Howell, 1930; Tarpley, 1985). Lambertsen (1983) elucidated the anatomy of the balaenopterid tongue and its dynamic inversion into the intracaval position (Fig. 16.7). Experiments with the head of a minke whale (B. acutorostraia, a species with a more bulky tongue) demonstrated that when the mouth is filled with water to simulate engulfment, the loose tongue folds into the cranial portion of the cavum (between its inner and outer walls) to initiate distention of the capacious oral pouch. The inverted tongue acts as an elastic sac whose lumen is continuous with the buccal cavity and whose walls are formed by the nowinvaginated tongue and nonlingual intermandibular lining. The mouth balloons out in bullfrog or pelicanlike fashion as the distensible pleats expand. Lambertsen (1983) speculated that closure and expulsion occur by the active contraction of jaw adductors (especially the temporalis) and elastic recoil of the pouch. In their biomechanical study of the throat wall, Orton and
pouch musculature
gular pleats
F I G U R E 16.6. Diagrammatic transverse section through rorqual mouth showing relation of baleen to lips. Note the cavum ventrale (black space) separating the sheet-like lingual musculature from the longitudinal pouch muscle underlying ventral blubber.
16. Feeding in Marine Mammals
FIGURE 16.7. Rorqual engulfment mechanism, indicating normal streamlined body profile and extensive inversion of tongue and distension of cavum ventrale (dashed lines) during maximal pouch expansion. After Lambertsen (1983).
Brodie (1987) outlined three means of pouch deflation and water expulsion: dynamic pressure of water coming to a stop at the front of the mouth, stored elastic energy in the blubber and other tissues surrounding the cavum, and possibly active contraction of muscle underlying the blubber (although they noted that this has yet to be shown). They found large amounts of elastin in ventral pleat blubber and gular musculature and quantified limits of reversible deformation in these tissues (circumferential and longitudinal expansion of groove blubber up to 4 and 1.5 times resting length, respectively). The loose articulation of the mandibular rami by fibrous ligaments rather than a bony symphysis permits a wide gape and cushions against the stresses of water entry (Gaskin, 1976). Lambertsen et al (1995) synthesized anatomical study, behavioral observation, and biomechanical analysis to describe the structure and function of a fibrous frontomandibular "stay" apparatus extending (as an appendage of the temporalis) from the supraorbital process of the skull to the coronoid process of the jaw. This taut connection optimizes gape angle and mandibular rotation during engulfment and acts as a spring to store elastic energy to facilitate gape closure by kinematic reversal (transferring the enormous momentum of the engulfed water to the jaws) and perhaps to focus the flow of expulsion through baleen. Water may also cascade out of the mouth by gravity during lunges above the surface. The gray whale {Eschrichtius robustus) employs benthic suction feeding, turning on its side and skimming the bottom while rapidly depressing and retracting the tongue to stir up sediments and suck in prey (Fig. 16.3), primarily molluscs and gammaridean amphipods that are winnowed from a single mouthful of muddy water with stiff, short, coarse baleen (Murison et al., 1984; Nerini, 1984). Scores of suction-generated pits or depressions (1-5 m deep) scar the ocean bottom along gray whale feeding grounds (Oliver and Slattery, 1985; Nel-
499
son and Johnson, 1987; Weitkamp et al, 1992), and muddy snouts, scarred and abraded jaws, and unilaterally worn baleen are common in this monotypic family. Gray whales seem predominantly "right handed," as baleen on this side is often shorter, whereas barnacles collect on the left side of the head (Kasuya and Rice, 1970). Scarring is also asymmetrical. The tongue is firm and muscular and there is a small subrostral gap as in balaenids. Although Eschrichtius occasionally scrapes prey off strands of kelp or skims fish and squid suspended in open water (even gulping like rorquals; Sund, 1975), the "mussel grubber" (as Yankee whalers called it) is the only mysticete to ingest food with strong, internally generated suction pressures. This behavior was first documented in a young captive animal given blocks of frozen squid (not a natural diet), which, when thawed and separated, sank to the floor of the pool (Ray and Schevill, 1974). The yearling whale turned with her lip just above the bottom and opened her jaws slightly, sucking in water and expanding her throat so that the gular grooves (two to five deep creases) bulged out. Lips could be moved independently and curled away from the baleen. She ate 900 kg of squid daily but spat out fish that had been engulfed in this manner. Turbid water was also seen squirting from the mouth. Although this animal always turned to the left while feeding, this is probably because she was given food on this side when trained to accept squid. Ray and Schevill (1974) cautioned that suction feeding may not be normal behavior of a wild adult, but the abundance of pits and mud plumes trailing behind gray whales argues otherwise. It has been suggested that this gouging and plowing of the sea floor actually increases productivity (Nelson and Johnson, 1987; Klaus 6f al, 1990). By the Middle Oligocene all three mysticete lineages were present, and it is interesting to consider which of the feeding mechanisms described earlier, if any, was ancestral for Mysticeti, especially since Eschrichtius is often regarded as an archaic species. Were mysticetes originally suction feeders? Did gulpers and skimmers develop from whales that captured prey with suction? Whether the balaenid method of continuous ram filter feeding or the intermittent mechanisms of other mysticetes constitutes the plesiomorphic condition is a mystery that the fossil record cannot yet resolve. Although continuous filter feeding is generally presumed to be a primitive character for vertebrates, many believe that Recent gray and rorqual whales come closer in form and function to the earliest baleen whales, suggesting that balaenids diverged from the mysticete lineage later on and that their feeding is a derived condition. Although the rostral shape of the earliest known mysticetes (Aetiocetidae) is controversial and their soft tissue anatomy obviously unknown, it is possible that some
500
Alexander Werth
were rather blunt-headed suction feeders. However, Sanderson and Wassersug (1990, 1993) proposed that intermittent suction feeding occupies an intermediate functional stage between continuous ram filter feeding and capture of individual prey items via suction, which would suggest that balaenids employ the original mysticete feeding method. As Pivorunas (1979) noted, balaenids are the only whales in which the sides of the mouth are entirely blocked by baleen when the mouth is opened to full gape, for their "whalebone'' both catches and gathers prey, unlike rorqual and gray whales who employ a two-step process in which the oral cavity catches the prey and the baleen retains it as water is rapidly expelled. The latter whales do not capture their prey directly from water per se, but instead utilize lunge feeding or suction to engulf a mouthful of preyladen water, then take advantage of the baleen filter to separate their (typically larger) food from the exiting water. The straining can occur at leisure, without fear of prey escaping, so that the baleen can be said to be of secondary importance to prey capture, as it technically plays no role in the initial stages of ingestion (Pivorunas, 1979). Another interesting point concerns intermediate species during the evolution of baleen. Early mysticetes had small, multicusped teeth that may have acted as a filter (as in filter feeding seals), whereas later forms possessed tiny teeth and jaw adaptations similar to modern mysticetes, raising intriguing possibilities of transitional forms and the gradual evolution of this filtration mechanism from simple palatal papillae. Miller (1929) presented the elevated, rugose gums that largely obscure the teeth of Dall's porpoise (Phocoenoides) as an analogue for the origin of baleen. More recently, Mitchell (1989) hypothesized that filter feeding developed in mysticetes with widely spaced, notched or serrated teeth, creating a dental straining apparatus analogous to that of the crabeater seal, Lohodon carcinophagus (Fig. 16.15). Location of new fossil material should help resolve these issues. Mysticetes are surface, midwater, and bottom feeders, yet they depend heavily on producers and thus do not stray far from the photic zone (upper 200 m). Gray whales are not found beyond the continental shelf, and balaenids likewise have a coastal distribution, feeding in productive waters near upwelling zones or at high latitudes (where long periods of daylight spawn abundant phyto- and zooplankton). Rorquals are more pelagic and cross large stretches of open ocean, but like right whales, they feed in polar regions or near seamounts, escarpments, or other major oceanographic features that augment prey density. Mysticete zoogeography is well defined. Migration patterns are firmly established and like winter breeding grounds, summer feeding grounds vary little, if any, from year to
year. However, whales are known to follow shortterm fluctuations in prey distribution and concentration caused by temperature and currents. Pivorunas (1979) speculated that while pack ice in the Arctic Ocean and other northern seas limits phytoplankton growth, it also breaks up large stretches of water where storms would tend to develop, forming heavy seas and other disturbances, as in southern oceans. He proposed that the calmer, Arctic waters might allow indigenous cetaceans (bowhead, beluga, and narwhal) to remain yearround instead of remaining only for brief summer periods, as occurs around Antarctica. The lack of strong, disruptive waves or upwelling currents might have made continuous filter feeding (particularly at the surface) a more attractive prospect for bowhead and northern right whales. Curiously, while the southern right whale {Eubalaena glacialis australis) inhabits a notoriously stormy region, it prefers to spend much of its time in sheltered bays such as Golfo San Jose near Peninsula Valdez off Patagonia, at least during winter breeding and calving seasons, although in general right whales are less migratory than rorqual and gray whales. Specialized behaviors associated with lunge feeding are as remarkable as the mechanics. Rorquals employ numerous agile feeding behaviors such as stereotyped circular "pinwheel" or figure-eight swimming patterns, breaching and surface lunging (vertical, lateral, or upside down), and flipper and fluke splashing to smack and stun prey. Humpbacks use remarkable entrapment devices such as bubble nets, clouds, columns, rings, and curtains emitted by the blowholes (Jurasz and Jurasz, 1979; Watkins and Schevill, 1979; Hain et al, 1982; Gormley, 1983; Wiirsig, 1988). Side swimming is very common and seems to be an effective means of engulfing two-dimensional swarms of prey schooling at the surface. Side swimming is seen in raptorial odontocetes and other narrow-snouted, but wide-gaped predators and has been postulated for archaeocetes as well (Barnes and Mitchell, 1978). Other 'lateralized" foraging tendencies (e.g., breaching, spinning, flipper slapping) are well documented (Clapham et ah, 1995). Humpback "flick feeding" involves wave production by slapping the tail when diving; the whale then surfaces and swims through the wave (Evans, 1987). Hays et ah (1985) described a resourceful humpback that by slowly sinking created a low pressure zone and thus concentrated its prey. "Lobtailing," in which a whale slaps its tail on the surface to stun or congregate prey momentarily, is also common. Foraging costs in mysticetes (both solitary and group feeders) correlate with prey patch density and depth (Dolphin, 1988). As with other social behaviors, mysticete foraging yields clues to their ungulate ancestry (Wiirsig, 1988). Balaenid and balaenopterid foraging is often cooperative (although
16. Feeding in Marine Mammals this may in fact simply stem from locally abundant food supplies) and may involve interspecific groups (Whitehead and Carlson, 1988) or special swimming formations such as V-shaped echelons that apparently limit prey escape (Wiirsig, 1989). Wlirsig (1989) suggested that the nongregarious gray whale may establish and defend individual feeding ranges. Photoidentification confirms that individual humpback and minke whales develop distinctive, characteristic foraging behaviors (Bonner, 1989; Hoelzel et ah, 1989). Coloration apparently plays an important role in mysticete feeding: for concealment and camouflage (countershading to disrupt body outline or mottling to blend in with shoals of prey), for attraction, or for herding prey into denser concentrations (Mitchell, 1970). Two commonly cited examples are the long, white, knobby flippers of humpbacks and the asymmetrical pigmentation of fin whale jaws (white on right side, dark on left), both of which may startle and scare dense clouds of prey toward the mouth (Brodie, 1977). Megaptera relies on complex foraging maneuvers whereas Balaenoptera uses plain, rectilinear locomotion (and typically rolls to the right) while feeding. Brodie (1993) also proposed that noise generated by synovial joint cracking during mandibular realignment in fin whales might confuse and corral shoals of prey. Unlike odontocetes, mysticetes do not use sound production to locate prey. Regarding the timing of engulfment noted earlier, visual prey detection is possible but unlikely given poor visibility underwater. Tactile mechanoreceptionis more feasible considering the vibrissae scattered over the tip of the rorqual rostrum and mandible. The large "stovebolts" or mental/genial and maxillary tubercles of Megaptera each contain several sensory hairs; Brodie (1977) speculated that since this species is less streamlined than other rorquals, the nodes allow vibrissae to project beyond the thick boundary layer created by more turbulent flow. Lingual mechanoreceptors may also function in prey detection, although these would, of course, be useless when the mouth is closed. Mysticetes consume huge quantities of prey, with estimates ranging from 200 to 1000 kg per meal and 200 to 600,000 kg annually [Gaskin (1982) summarized information on mysticete food intake, metabolism, and energy budgets]. The retention of prey by baleen is straightforward, but the mechanism of food removal from this filter is not yet understood. Several competing theories have been proposed to explain this mystery, all loosely supported by anecdotal evidence (Werth, 1990). The most common idea is that the tongue is simply elevated and retracted to scrape entrapped items from the mat. Baleen is estimated to grow roughly 30 cm per year in balaenids, and its occasional presence in whale feces is offered as evidence that it regularly wears away. Lin-
501
gual depression and retraction would allow a bolus of prey collected on the central furrow to be swallowed. Potential disadvantages of this mechanism include rapid removal of baleen and inefficient clearing of prey from fringes that do not directly contact the tongue. Another hypothesis, also mechanical in nature, depends on observations of rapid head-shaking behavior in southern right whales. Whales have been sighted shaking their heads rapidly from side to side just above the surface, with a sound that can be heard from a great distance (not unlike the "baleen rattle" during right whale skim feeding, produced by lapping of water over partially submerged plates; Watkins and Schevill, 1976). Again, this method might not release prey adequately, although because balaenids sometimes skim for hours between apparent swallows, it is likely that they would release enough food to swallow. However, right whales often make short, lunging rushes during bouts of skim feeding, which Gaskin (1982) hypothesized could agitate and remove clinging food particles. A final plausible proposal involves "backwashing" by brief entry of water into the mouth to remove effectively items from the baleen, depositing them on the tongue for swallowing. Opening (and perhaps outward rotation) of the jaws coupled with rapid tongue depression might cause sufficient negative pressure to draw water in from all directions. The backwash current need not be especially strong merely to transport prey into the middle of the oral cavity. Although this method would be very effective, it requires lingual and labial mobility. Tongue movements in all three hypotheses involve changes in position rather than shape, which is consistent with myological findings in right and bowhead whales (Werth, 1993) for which extrinsic lingual muscles (originating outside the tongue and serving to move it) are far more prominent than intrinsic muscles (which exist solely within the tongue and presumed to be the main effectors of lingual shape change). Another filter-feeding marine mammal is the crabeating seal, Lobodon carcinophagus (Phocidae), which, in fact, preys not on crabs as its name implies, but on krill. Krill are sieved from the water with uniquely lobed, ornate cheek teeth, which, when interlocked, permit water to drain from the mouth. The incisors and canines are shaped as in other phocids (although the canines are notably small), yet the five upper and lower postcanines possess a complex array of well-defined, comb-like cusps (Fig. 16.15) that interdigitate when occluded to retain prey while water is expelled from the oral cavity (King, 1961). Prey are prevented from escaping posterior to the tooth rows by bony protuberances (covered with gingiva) of the dentary and maxilla. Lobodon has been observed swimming open mouthed through dense shoals of krill; however, observations of captive
502
Alexander Werth
animals suggest that its intermittent suspension feeding is more likely due to suction generation than locomotion-based ram intake. Although this species is rarely held in captivity, two juveniles showed an ability to suck small fish into the mouth rapidly and forcefully from distances as great as 50 cm. Prey were held in the mouth by the dentition and immobilized against the palate by the tongue as the jaws closed and the lips raised (along with lingual elevation) to expel water with a "lip-smacking" action prior to deglutition, although it is not known whether some water was swallowed as well (Ross et al, 1976). The elongated mandibular symphysis affords large surfaces for genioglossal and geniohyoid muscle origins, presumably the basis of powerful tongue movements (King, 1961). This reliance on suction (for prey ingestion), combined with a straining apparatus (to filter prey from the water), demonstrates once again the versatility of marine mammals and the incomplete division of their feeding strategies. Noting the much smaller size of the head of Lobodon relative to mysticetes, Bonner (1990) proposed that the lobulated teeth are an adaptation for expelling large amounts of water engulfed with individual prey items; because less water is ingested (and swallowed) with prey, a smaller filtration mechanism suffices. Thus, rather than engulfing a mouthful of prey (as in baleen whales), crabeater seals apparently suck prey items selectively, and once sufficient krill accumulate a bolus is swallowed. According to 0ritsland (1977), the average crabeater seal meal is roughly 8 kg of krill. As each krill weighs only 1 g, Lobodon spends a large proportion of its time filter feeding, made practicable only with the immense swarms of this crustacean in Antarctic seas (Bonner, 1990). Given such abundance, it seems likely that Lobodon ingests several krill at a time (King, 1961). Although Lobodon feeds exclusively on krill throughout most of its circumpolar distribution, it may also take small fish (as in captivity). Stomach contents and captive behavior also hint at benthic foraging (Ross et al, 1976). King (1983) speculates that Lobodon feeds mostly at night when krill and other crustaceans approach the surface; feeding dives are shallow (to 25 m) and short in duration, but may occupy the entire night (Bengtson and Stewart, 1992). The crabeater seal's suction/filter feeding is doubtless highly effective, for this Antarctic phocid is variously described as the world's most abundant seal (Reeves et al., 1992) or even the world's most abundant large mammal (Laws, 1984). With an estimated annual consumption of 63 million tons of krill (Reeves et ah, 1992), crabeater seals are strong competitors of other krill-eating species, especially mysticete whales, whose severe reduction after several decades of intense hunting around Antarctica has probably contributed greatly to the success of crabeater seal populations.
Some phocids, notably the leopard seal, Hydrurga leptonyx, have complex cusped postcanines reminiscent of (but not as pronounced as in) Lobodon (Fig. 16.15). These sharp points probably serve to immobilize struggling prey, although they may permit water to flow from the mouth, as zooplankton constitutes a small yet seasonally important component of the diet of ringed seals (Phoca hispida), Antarctic fur seals {Arctocephalus gazella), and other northern and southern phocids and otariids. Noting the presence of slight yet similar bony protuberances on dentaries of several phocids. King (1961) speculated that filter feeding of the crabeater seal may represent retention of a primitive condition, from which other seals later adapted to piscivory; alternatively, this habit may have arisen twice, but evolved to a far more specialized level in Lobodon. B. Suction Feeding Suction is used not only by select suspension feeders, of course, but also by many marine mammals that ingest and transport prey with negative pressures generated via oral and pharyngeal expansion. The difference is that while gray whales and crabeater seals consume vast quantities of aggregating organisms (hence the need for filtration mechanisms), other cetaceans and pinnipeds vacuum larger and ordinarily more evasive solitary prey; again, items are retained while engulfed water is expelled back through the mouth prior to swallowing. Although broad variation exists in prey size, type, and activity, mesopelagic cephalopods are by far the most common choice of marine mammal suction feeders, with bivalve molluscs and other benthic invertebrates second. This extremely effective and versatile solution to the problems of aquatic prey capture has been independently adopted by nearly all marine and freshwater vertebrates and has been extensively studied in elasmobranchs, actinopterygian fishes, lungfishes, and the coelacanth. Among tetrapods, suction feeding has been described in larval and adult salamanders, pipid frogs, caecilians, and turtles (Lauder, 1985). Although simple in principle, suction feeding often requires substantial modification of the skull, jaws, and hyolingual apparatus. Thus it is not surprising that this mode of feeding is practiced by some of the most derived and specialized marine mammals, including several odontocetes and pinnipeds. Among the former at least five families (delphinids, monodontids, phocoenids, physeterids, ziphiids) have definite or putative suctionfeeding members, whereas there are undisputed suction feeding representatives of the carnivoran families Odobenidae and Phocidae (walruses and seals, respectively). It must be emphasized, however, that no family of marine mammals is composed entirely of suction
16. Feeding in Marine Mammals feeders. Clearly this trait has evolved multiple times. Likewise, there appear to be no species that rely exclusively on suction feeding. While the walrus and narwhal are perhaps the most specialized suction feeders (in terms of morphology and foraging ecology), they are nonetheless extremely opportunistic mammals that may utilize other feeding methods when necessary. Unfortunately, while the kinematics, muscle mechanics, and pressure attributes of the discrete phases of suction feeding have been worked out in great detail for myriad suction feeding fishes and amphibians, this is not the case for marine mammals, as research is severely hampered by the constraints noted earlier. In fact, morphologists were reluctant to accept the feasibility of suction feeding in marine mammals given the anatomy of "typical" dolphins and seals (e.g., long jaws and open "notched" gape with numerous sharp teeth), particularly their lack of features seen in suction feeding fishes, which mediate precise, undirectional water flow (e.g., pipette-like mouth, opercular ostium). Despite the traditional view of long-snouted snappers with elongate jaws and prodigious teeth, many toothed whales and pinnipeds exhibit an entirely different profile and dentition (Fig. 16.14), far from that of the wellknown bottlenose dolphin and harbor seal. Their feeding behavior is likewise divergent from the customary picture. A cursory examination reveals major variation in characters linked to suction feeding, from oral openings to gular myology. The anatomical mechanism of suction feeding involves rapid, piston-like retraction of a flat, hemicylindrical tongue, creating a negative (less than ambient) pressure in the buccal cavity into which water and prey are drawn. Ingested water in this bidirectional flow system is momentarily accommodated by the expandable, elastic pharynx and (in many cases) by external throat grooves. It is possible that the stomach participates in some species as well, although influx this far posteriorly (and internally) is doubtful, as it would result in the consumption of undesirable quantities of seawater. However, Fiarrison et at. (1967) showed that the odontocete forestomach could forcibly eject ingested water, and the excretory capabilities of marine mammals are sufficient to counteract this osmotic load. Fay (1960) described the walrus's extensive and highly elastic pharyngeal pouches, which may be used to store water ingested via suction; he suggested buoyancy regulation, storage for food or air reserves, and sound production as alternative functional possibilities. In all cases, ingested water is forced back out through the mouth when the oropharyngeal cavity is compressed by jaw closure and tongue elevation. Generation of suction by pulmonary expansion and subsequent inhalation is untenable in odontocetes, as (like all cetaceans) they maintain a permanent patent
503
airway, effectively separating the oropharynx and trachea, nor is it feasible in pinnipeds, especially when considering the great depths at which many feed, where intense pressures inhibit thoracic expansion. Nonetheless, inspiration of water remains a hazard for walruses and other suction feeding pinnipeds (Gordon, 1984). Exceptionally deep feeders such as Physeter would not need to generate any more or less suction pressure; all that matters is the relative change from ambient pressure. Difficulties of oral/gular expansion against extreme hydrostatic pressure may nonetheless explain the massive nature of the sperm whale hyoid and sternum. Floating or otherwise stationary marine mammals (often at the bottom of the water column) suction feed as well as swimming ones, demonstrating that this is not merely "ram" feeding, in which a predator relies on rapid locomotion to overtake and engulf prey. Even raptorial odontocetes and pinnipeds that seize prey rely on suction to transport it to the rear of the oral cavity for swallowing. In suction feeding whales, dolphins, and porpoises, the hyoid skeleton consists of a large, flattened basihyal and thyrohyal body with robust stylohyal arches anchored to the skull just posterior to the otic complex via tympanohyals (Fig. 16.8). Bony and cartilaginous elements of the hyoid are joined by highly mobile, synovial joint capsules (Reidenberg and Laitman, 1994); the hyoid is also connected to the immobile larynx, which is bound in an intranarial position by a powerful sphincter of pharyngeal muscles originating from the pterygoids and palate. A massive sternohyoideus muscle is largely responsible for hyoid depression, whereas the large, paired mm. hyoglossus and styloglossus project anteriorly and insert into the tongue body. Along with the m. genioglossus (which has extensive origins along the rami of the mandibles), these extrinsic lingual muscles rapidly retract the tongue to expand the oral cavity and create a space into which water and prey are sucked. Comprehensive analysis of attachments, positions, and actions of lingual and hyoid musculature disclosed significant differences between suction feeding versus raptorial dolphins in the size, shape, thickness, and curvature of the hyoid, as well as in the proportional weight and cross-sectional area and relations of the sternohyoideus and hyoglossus (Werth, 1992). The tongue itself is a large, firm, muscular hydrostat (see Chapter 2) with only a tiny free anterior tip, but an extensive system of longitudinal folds or plicae indicating marked mobility, nonetheless. The tongue has a smooth, flat dorsum and does not taper but maintains uniform height and width, making it a perfect hemicylindrical piston. A fringe of marginal papillae ( 2 5 mm) is frequently seen along the anterior and lateral edges of the tongue body (Yamasaki et al, 1976; Kastelein and Dubbeldam, 1990). Found predominantly in
504
Alexander Werth
F I G U R E 16.8. Odontocete hyolingual specializations for suction feeding: (a) fringe of marginal papillae on tongue of adult Atlantic white-sided dolphin, Lagenorhynchus acutus, and (b) robust hyoid apparatus of pygmy sperm whale, Kogia breviceps, showing muscle origin. Note enlarged insertion of sternohyoid d, digastric; gh, geniohyoid; hg, hyoglossus; ih, interhyoid; mh, mylohyoid; oh, occipitohyoid; sg, styloglossus; sh, sternohyoid; th, thyrohyoid.
suction feeding genera, these structures (Fig. 16.8) appear late in development yet persist throughout life, unlike in other mammals where they create a tight seal between the tongue and the palate for improved neonate suckling. Retention of this juvenile feature is strongly suggestive of sucking activity by the tongue in certain adult odontocetes. Sokolov and Volkova (1973) proposed a receptor function for these "fimbriae," but histological analysis does not support their claim. Examination of intrinsic lingual musculature reveals a limited capacity for shape changes and protrusion yet great retractile capabilities (Werth, 1992). A ventral sublingual space of loose areolar tissues provides little resistance and allows for rapid, piston-like withdrawal
of the tongue. The palate is generally smooth, but in beaked whales bears many papillose rugosities (unlike the ridges of many terrestrial mammals), which may function to hold cephalopods or other slippery prey (Heyning and Mead, 1991,1996). Phocoenoides dalli, the virtually toothless Dall's porpoise, also has a ribbed palate. Many large odontocetes, especially sperm and beaked whales, possess several parallel or wishboneshaped external throat grooves (Hubbs, 1946; Clarke et al, 1968). Ross (1987) proposed that these creases are found in large suction feeders where gular expansion from the engulfment of large quantities of water would otherwise be prohibited by the thickness and rigidity of the overlying blubber. Heyning and Mead (1991) confirmed that these folds are distensible and controlled by the contraction of superficial ventral musculature in ziphiids, suggesting that they may serve as active rather than merely passive expanders of the pharynx. They also found that ziphiid intrinsic lingual musculature is quite complex and that certain gular muscles (such as the interhyoid and sternohyoid) are relatively larger in beaked whales than in other putative suctionfeeding odontocetes, correlating with their larger hyoid skeletons. Several lines of circumstantial and anecdotal evidence supported the idea of marine mammal suction feeding before its actual documentation. This accumulated evidence includes anatomical and ecological data as well as direct observations by marine mammal handlers. One of the first and most compelling pieces of information to support marine mammal suction feeding indirectly is the surprising reduction or total loss of teeth, especially in "toothed" whales, where the presence of teeth is a diagnostic character if a poor descriptive one. Some species are truly edentulous, such as the narwhal {Monodon monoceros; Fig. 16.14), whose sole apparent tooth is a long, spiraled tusk that normally erupts only in males and which is clearly not used in feeding (it would, in fact, hinder raptorial feeding). Many teeth remain obscured by gingiva, however, in all odontocetes. Where present, teeth are often partially or completely worn to the root (due to the thin cap of soft, aprismatic enamel) or covered with barnacles or other epizoic organisms (Clarke, 1966; Morris and Mowbray, 1966; Rice, 1989) that prevent occlusion and indicate the nonfunctional nature of the dentition, as these would be crushed if jaws were used in typical mammalian fashion. The strap-toothed whale, Mesoplodon layardi, represents an extreme case of nonfunctional dentition (Fig. 16.9): the pair of flat lower teeth completely encircle the rostrum, restricting gape to a small, round anterior opening analogous to a vacuum cleaner attachment or "slurp gun" used by divers to collect fish. As in other beaked whales, teeth are greatly reduced in
16. F e e d i n g in M a r i n e M a m m a l s
F I G U R E 16.9. Nonfunctional jaws and dentition in odontocetes: (a) head of male strap-toothed whale, Mesoplodon layardi, showing teeth protruding outside and encircling mouth; and (b) curved jaw of sperm whale, Physeter macrocephalus (ventral view).
number and project outside the oral cavity, where they would be useless for prey capture and processing. Oddly, these edentulous species feed on squid, the most slippery of odontocete prey. M. layardi also illustrates the standard ziphiid condition of tooth eruption only in males. Far from vestigial, however, these teeth and tusks undoubtedly perform a function, suggested by their unisexual distribution, as a secondary sexual character used like horns and antlers of terrestrial mammals in dominance displays, fighting, and courtship rituals (Caldwell and Caldwell, 1966; Kleinenberg et al, 1969). Scarring along the flanks and head is common in many odontocetes (McCann, 1974; Heyning, 1984), including sperm whales (Berzin, 1972; Best, 1979; Kato, 1984), whose plentiful scratches and marks are often attributed to squid, although their patterns perfectly match the spacing of sperm whale teeth (Boschma, 1938) and are rarely found on females, whose feeding (and diet) is identical (Kawakami, 1980). Scarring is particularly common in solitary bachelor bulls that engage in the most fighting. Even in species bearing respectable dentition, teeth may not erupt until sexual maturity (long after wean-
505
ing), suggesting that they are not necessary for feeding. The reduced complexity apparent in the gross and ultrastructural morphology of odontocete teeth also reflects their apparent nonutility. Enamel is often weak, poorly developed, and limited to a small area on the tip of the crown, representing another notable degeneration from the standard eutherian dental condition (Ishiyama, 1987; Werth and Stern, 1992). Pinniped teeth are less variable in size, shape, and num^ber, but the suction feeding walrus and elephant seal show marked reduction in the standard dental battery. Flattened or tiny peg-like teeth may be used to crush or grasp prey but certainly not to obtain it; canines or incisors are enlarged (e.g., walrus tusks) only for social purposes, as in odontocetes. Not only are teeth greatly reduced or totally lost, but evidence suggests that the jaws, themselves, are nonfunctional in some odontocetes. A common congenital defect in sperm whales causes the mandible to develop in a loose spiral (Murie, 1865; Nasu, 1958; Spaul, 1964; Nakamura, 1968), and broken jaws incurred by fighting, entanglement in submarine cables (Heezen, 1957), or collision with other undersea obstacles often heal at an improper angle that precludes jaw closure (Fig. 16.9). However, sperm whales with deformed or broken jaws (or without teeth) grow as large, live as long, and, most importantly, feed on the same size and type prey as animals with normal jaws (Clarke et ah, 1988). Equally convincing evidence for marine mammal suction feeding abounds in ecological data, especially information on prey condition showing that shrimp and other delicate prey items are found intact and unharmed, without bite marks or tooth holes (Gunter, 1951; Jones, 1981). Spanish whalers have even reported live squid flopping out of the stomachs of freshly caught sperm whales (Clarke, 1955; Norris and Mohl, 1983). Although stomach content data are understandably hard to obtain for marine mammals, the limited information available suggests that food is often swallowed whole and unmarked. Odontocetes and pinnipeds ingest a large variety and quantity of inorganic debris or exotic, nonfood items ranging from rocks and sticks to fishing gear and plastic objects (Nemoto and Nasu, 1963; Ridgway, 1965; Berzin, 1972). It is not unusual to find sand dollars, brittle stars, and other poorly digested benthic invertebrates in stomach contents of belugas and pilot whales, suggesting that these species often scour and vacuum clean the bottom for food. The underslung, shark-like jaw of Kogia is an apparent adaptation for benthic feeding. Many suction feeders demonstrate impressive diving capacity (depth and duration) for protracted feeding at the thermocline or sea floor (Gaskin, 1982). Norris and Mohl (1983) presented much of the
506
Alexander Werth
circumstantial evidence for suction feeding in their review of the prey-stunning hypothesis first proposed by Bel'kovich and Yablokov (1963). The hypothesis holds that certain odontocetes are able to debilitate prey acoustically with sonic pulses generated by nasal valves and sacs and focused by the melon, perhaps developed as an extension of echolocation for prey detection (Hult, 1982). Despite intense interest and scrutiny, the acoustic stunning theory has never been substantiated (Zagaeski, 1987; MacKay and Pegg, 1988) [some cetaceans, notably killer whales, stun prey via fluke slaps (Simila and Ugarte, 1993)]. Nonetheless, Norris and Mohl outlined two major trends in odontocete evolution: the reduction or loss of teeth and the progressive widening and shortening of the skull and jaws to create a more blunt head. While these obvious and indisputable trends demonstrate the nonutility of long jaws and dentition for feeding in many odontocetes, they do not appear to relate to the sound generation needed for prey stunning. It is clear that head shape is, however, a critical factor in suction generation. Manipulative experiments on severed heads of stranded odontocetes displaying a diversity of head forms demonstrated quantitatively the effect of head shape on water flow patterns (Werth, 1989,1992). Cannulae were threaded through successive holes drilled into rostra of heads ranging from a blunt-headed harbor porpoise {Phocoena phocoena) to a long-snouted common dolphin {Delphinus delphis). Transducers introduced into these cannulae measured pressures in the anterior, middle, and posterior thirds of the oral cavity as infra- and suprahyoid muscles of submerged heads were manually retracted and depressed with string handles; site, speed, and force of pull were varied. Dowels inserted between the rostrum and the mandibular symphysis controlled gape; the esophagus was sutured closed. Pressure drops, calibrated and measured in mm Hg, were converted into digital waveforms for analysis; the time course for each pulse of negative pressure was also measured to ensure that all pulls were consistent and thus comparable. Pressures of roughly - 4 5 mm Hg were measured, suitable for capture of small prey; certainly greater pressures could be attained by live animals. More importantly, these experiments showed that shortening and widening of the rostrum greatly improve suction generation. Species with a smaller, rounder mouth opening are better able to draw in prey anteriorly, as this reduces the "notched" gape of longsnouted forms that tends to draw water in laterally. However, experiments documenting suction feeding in captive juvenile long-finned pilot whales, Globicephala melaena, showed that the side of the mouth may also be used to ingest prey via suction, at least in this remarkably adaptable species (Werth, 1992, 2000). Three pilot whales held in a large pool during rehabilitation
following stranding (in preparation for their eventual return to the wild) were filmed feeding on frozen herring, mackerel, and squid at the surface, midwater, and especially off the bottom, from which they would rapidly "pull" their admittedly nonelusive food. This footage provided confirmation of suction feeding behaviors, such as expulsion of water from the mouth (from a whale that ascended nearly vertically and "missed" a prey item at the surface) and indicated the extent to which this species rolls or rotates while ingesting prey, presumably to engulf food in an orientation so that it may be swallowed without manipulation in the mouth. It is important that fish be swallowed head first and squid mantle first to prevent spines, fins, or other protruding structures from catching in the throat. [In this light, however, it is interesting that captive juvenile crabeater seals (Ross ei a\., 1976) had difficulty orienting food in the mouth but gradually learned to suck in or manipulate prey so as to facilitate deglutition.] While feeding on the bottom, however, pilot whales showed that prey could be as easily drawn in from the side of the mouth by using the buccinatorius and orbicularis oris muscles to close off the other side and restrict the opening through which water and prey are sucked. Like belugas (Brodie, 1989), pilot whales may purse their lips to reduce the oral opening or make it rounder to improve suction ingestion of prey, although this was not seen, nor was feeding observed with live prey. Osteological data on rostral and mandibular dimensions (Werth, 1992) reveal an impressive range of odontocete head profiles (Fig. 16.14) rivaling that of other aquatic tetrapods (e.g., gracile and robust crocodilians ranging from gavials to caimans), with extremely slender-snouted platanistids and oceanic delphinids (Stenella) as well as short, bulldog-like heads in dwarf and pygmy sperm whales (Kogia) and many dolphins and porpoises. Not unexpectedly, blunt heads correlate with other anatomical, ecological, and behavioral traits associated with suction feeding. A surprise, however, is that the species for which the most compelling circumstantial evidence for suction feeding has been adduced, the sperm whale, has long, narrow jaws. However, unlike pilot whales, sperm whales have no proper oral cavity according to the standard definition, and with its pendulous jaw hanging fully open (as is often seen) the oropharyngeal isthmus of the sperm whale presents a perfect circle (Fig. 16.10), which in the absence of an oral orifice is the opening through which prey are sucked directly into the oropharynx, thus eliminating the need for intraoral prey transport and combining ingestion and swallowing into a single feeding stage! The glistening white mouth of the sperm whale, standing in stark contrast to the remainder of its jet black or charcoalcolored body, has prompted researchers to propose a passive luring stratagem for attracting squid (Beale,
16. Feeding in Marine Mammals
F I G U R E 16.10. Lateral profile of sperm whale '"oral cavity" showing round oropharyngeal opening (dashed), throat grooves, and hyoid apparatus in extended position.
1839). Gaskin (1967) confirmed that a luminous substance rubs off the bodies of bioluminescent squid easily and documented instances of glowing sperm whale mouths. Heyning and Mead (1996) suggested that white-beaked whatle mouths may also serve as squid attractants. Clarke's (1979) observation that Physeter often sounds and surfaces in the same spot casts doubt on the belief that it actively chases prey, and Fristrup and Harbison (1993) thought that vision plays a critical role in passive foraging. Cooperative foraging may also allow sperm whales to catch elusive prey (Whitehead, 1989). The huge, barrel-shaped head of Physeter has been explained as a ballast tank (Clarke, 1970,1976) or as a sonic lens for echolocation (also stunning; Norris and Harvey, 1972; see earlier discussion). The potential oropharyngeal nature of odontocete suction feeding is unique relative to other vertebrates, but it also hints at a scenario for the evolution of suction feeding in marine mammals. Although suction feeders possess clear conformities in morphology and ecology, it is important to remember that this feeding method was independently adopted by several distinct lineages; clearly the plesiomorphic condition is raptorial feeding (seizing and snapping). The use of suction to transport grasped prey in gars and other long-snouted fishes suggests an identical function in marine mammals, especially odontocetes, which have no alternatives, as swallowing occurs entirely underwater (Pilleri et ah, 1970). Odontocetes do not shake or rotate the head, use inertial or gravity transport, or any cranial kinesis to transport prey with a racheting action. Lingual transport is also precluded by the extremely long mandibular symphysis of many long-snouted odontocetes, in which the tongue does not reach the anterior extent of the oral cavity. Gavials (long-snouted piscivorous croc-
507
odylids) are known to take prey above the surface, where they transport it inertially by lifting the head and throwing it back into the pharynx, but such behaviors have never been described in cetaceans despite years of intensive observation of morphologically similar river dolphins in captivity. Suction ingestion in marine mammals appears to have evolved from the initial use of suction for intraoral transport of prey grasped in the jaws (Fig. 16.11). The grasp and separate transport steps were later lost (along with elongate jaws and elaborate dentition), simplifying and expediting the feeding process and freeing teeth for adaptation to social functions. Evolution of suction feeding from neonate suckling is unlikely given the anomalous nature of odontocete lactation, in which milk is actively pumped into the calf's mouth by the contraction of smooth muscles surrounding the mammary glands (Slijper, 1962; Arvy, 1974). Werth (2000) described the kinematic sequence of suction feeding events in Globicephala (Fig. 16.12), including a preparatory phase with partial gape followed by jaw opening and rapid hyoid depression to suck prey in at a mean distance of 14 cm, although some prey were taken from much greater distances. The actual suction or gular expansion phase averages 90 msec in duration; the entire ingestion cycle, not including the initial approach, elapses in one-third of a second. Although prey items were no longer visible at the conclusion of most ingestion events, it was still seen at the rear of the buccal cavity in a few feeding sequences. Unfortunately, jaw closure prevents clear observation of prey transport. However, slight movements of the mouth floor and throat are evident, suggesting that lingual, hyoid, or mandibular motions (or perhaps some combination of the three) are responsible for prey transport to the faucial pillars prior to deglutition. Heyning and Mead (1996) reported that fish placed in the mouths of captive Cuvier's beaked whales {Ziphius cavirostris) disappeared in less than 0.08 sec. They also noted that when live-stranded Hubb's beaked whale {Mesoplodon carlhubbsi) calves suck on fingers, hyoid motion and suprahyoid muscle contraction can be palpated externally. Despite the lack of functional analysis for other suction feeding cetaceans, several observers offer unequivocal reports of this ability in captive odontocetes. While perhaps less instructive than experimental studies of structure and function, these accounts provide incontrovertible evidence of suction feeding in a wide range of marine mammals and offer clues as to their feeding behavior. Ray (1966) described belugas sucking coins from the bottom of their pool and spitting them out. Brown (1962) and Donaldson (1977) recount similar sucking and expulsion behaviors in short-firmed pilot whales {Globicephala macrorhynchus) and killer whales (Orcinus orca), respectively. Numerous unpublished anecdotes of captive marine mammals (especially
508
Alexander Werth
PREY CAPTURE IN INIA VS. GLOBICEPHALA
1. Ingestion
1. Ingestion (SUCTION)
2. Transport (SUCTION)
3. Deglutition
2. Deglutition
FIGURE 16.11. Evolution of odontocete suction ingestion (as in pilot whale, Globicephala spp.) from transport of grasped prey in long-snouted odontocetes (e.g., the boutu or Amazon river dolphin, Inia geoffrensis) by loss of transport step and attendant elongated jaws and elaborate dentition.
belugas and beaked whales) describe repeated sucking and blowing out of small fish, leaves, or other debris. Among pinnipeds there is compelling anatomical and ecological evidence for suction feeding in several large, predominantly teuthophagous phocids such as the Ross seal, Ommatophoca rossi, bearded seal, Erignathus harhatus, and elephant seals, Mirounga sp.; its use is also suspected in an otariid, the cape fur seal, Arctocephalus pusillus (King, 1983). The Ross seal possesses large, well-developed lingual, suprahyoid, and pharyngeal musculature to aid in holding and swallowing elusive, slippery cephalopods (King, 1964; Bryden and Felts, 1974). Jaw adductors are strong despite the weak dentition (Fig. 16.15), and the tongue and epiglottis are situated far posteriorly. Muscles involved in tongue elevation and deglutition (especially hyoglossus and styloglossus and pharyngeal constrictors) appear to be adapted to gripping and swallowing large squid (King, 1964), although they may also relate to the loud bellowing as the throat expands during sound production (Bryden and Felts, 1974). King (1964) speculated that the exceptionally long soft palate (over 60% of total palate
length) enlarges the oral capacity longitudinally and vertically, whereas the ventral position of tracheal cartilages aids in esophageal expansion; she suggested that similar palatal and tracheal modifications permit the leopard seal to swallow bulky prey. However, the champion pinniped suction feeder is unquestionably the walrus (Odobenus rosmarus), which produces powerful suction currents to dislodge softbodied prey (e.g., annelid, sipunculid, or priapulid worms) along with clams, cockles, mussels, and other benthic molluscs from the gravelly or muddy substrate of the sea floor. Walruses use suction not only for ingestion, but for processing. They are able to remove bivalve siphons and feet from their shells while rarely swallowing shell fragments, a feat (like that of bearded seals sucking live whelks from their shells; Bonner, 1990) requiring remarkable suction strength and control. The shells are presumed to be held in the lips and rejected once the meat is removed by suction. Natives often rinse and eat undigested clams taken from stomachs of hunted walruses (Kenyon, 1986b). Not only can walruses obtain six or more clams per minute (for a meal
509
16. F e e d i n g in M a r i n e M a m m a l s
TIME (msec)
£ o UJ
o z <
GAPE HYOID
0)
PREY
o
GULAR DEPRESSION
JAW CLOSING
F I G U R E 16.12. Composite kinematic profile of suction feeding cycle in captive juvenile Globicephala melaena, plotting distance as measured by depression of mandible and hyoid against elapsed time. The whale's distance from the prey object is also indicated; note that jaw opening commences when prey is roughly 25 cm away. Hyoid depression occurs at prey distances averaging 14 cm; externally observed hyoid depression ranged from 2 to 6 cm. The actual suction or gular expansion phase averages 90 msec in duration. The entire ingestion cycle, not including the initial phase, lasts roughly 300 msec.
of up to 6000!) by sucking in water, but they may also squirt out powerful jets to stir the bottom and excavate prey (Riedman, 1990). Captive walruses have been known to strip flesh from fish skeletons via suction, or to suck heavy items (e.g., filter grates) from the floors of their pools (King, 1983). Again, unpublished anecdotes offer tales of captives demonstrating suction capabilities while playing with food, sometimes sucking it in from as far as 1 m away. Experiments on captive walruses provide impressive suction measurements of up to one atmosphere negative pressure (Fay, 1981; Kastelein^fa/.,1994). Gordon (1980, 1984) documented walrus suction feeding and conducted a detailed study of its anatomical mechanism using high-speed cinematography, radiography, and vector analysis of tongue movement based on dissection of retractor and protractor musculature. He compared models of lingual motion ranging from depression to retraction and concluded that the applied contractile muscle forces and resistance of the
tongue support piston-like withdrawal of the tongue directly to the rear of the oral cavity. The rather posterior origin of the genioglossus on the mandible results in a posteroventral line of action, and the extremely large paired styloglossi are powerful tongue retractors. Independent examination of tongue-induced abrasive wear patterns on the 16 nontusk teeth (Gordon, 1984) also support a fully posterior movement of the tongue. Gordon (1980) described how the short, wide shape of the walrus head and muzzle facilitate suction feeding, as does its rigid, domed tongue, which perfectly fits the concave, V-shaped palate. This buccal cavity can be greatly enlarged by lingual retraction, while the sides of the oral cavity are sealed when the jaw is adducted with only a narrow groove at the front of the palate, not blocked by the tongue tip, providing a round, anterior oral orifice. Gordon (1984) described key physical and behavioral differences between tongue movements of walrus suction feeding and of suckling in infant humans, pigs.
510
Alexander Werth
and dogs. The rear of the tongue is depressed to enlarge the oral cavity and create negative pressures during suckling in the latter species, yet Odobenus generates intraoral suction solely by posterior retraction; the tongue moves ventrally only to allow food to be swallowed. Gordon's model of lingual movement fits gross and microscopic anatomical evidence and affords the most efficient muscular force and least resistance for suction feeding (especially at foraging depths of 80 m), allowing the walrus to exploit more habitats for foraging. While profound interspecific variation exists in tongue movements during suckling, swallowing, and drinking via suction in mammals, the many contrasts in lingual mechanics and timing and degree of motion during the sucking cycle reinforce the contention that marine mammal suction feeding evolved not from suckling, but from prey transport. King (1972) observed that another pinniped skull feature may also correlate with suction feeding, as the long symphysis and outwardly bowed dorsal edges of the jaws of Odobenus and other species (notably bearded, crabeater, ringed, and harp seals) form a scoop-like mandible that may facilitate sucking or be associated with an arrangement of appropriate muscles. Although the roof of the mouth is not especially arched or vaulted in the latter three seals, all have been seen sucking in small crustaceans, and like the walrus, the bearded seal sucks in molluscs. The unique elastic pharyngeal diverticula of walruses (Fay, 1960), described earlier, may relate to suction generation, water accommodation, or holding of molluscs during processing. Despite this wealth of morphological and functional specializations, Odobenus is not a choosy eater; reported stomach contents include a variety of fish and remains of birds and marine mammals, including ringed and bearded seals, narwhals, and even yoimg walruses (King, 1983). This food may be taken only when molluscs are scarce, and may come from carrion. It is not known how walruses could capture or dispatch such large prey, although Inuit hunters claim walruses attack seals and whales (Reeves et al, 1992). Conceivably, the tusks are used for stabbing (Riedman, 1990) as in Smilodon and other saber-toothed Pleistocene carnivores. Although Odobenus does not chew its food, the low cheek teeth are worn smooth due to the abrasive diet and from sand and gravel inadvertently ingested with prey. Wear patterns on the anterior surfaces of tusks and mystacial vibrissae suggest that walruses stand on their heads during foraging dives (5-10 min), using foreflippers for stability and hindflippers for propulsion while locating prey with heavy whisker pads. Experiments on captives demonstrate the heightened tactile sensitivity of the 450 coarse vibrissae, which grow long in captivity (where food is provided) instead of wearing, suggest-
ing that they are actively used in feeding, perhaps to examine and manipulate food as well as to find it. A large portion of the cranial musculature is devoted to the control and movement of the vibrissae (Kastelein et al, 1991). Normally thick and cornified, walrus skin is thin on the front of the snout, and it is generally acknowledged that Odobenus engages in pig-like rooting of bottom sediments. Long furrows or sinuous feeding tracks scar the continental shelf where they forage (Nelson and Johnson, 1987). Controlled observations of captives (Kastelein and Mosterd, 1989) indicate that these disturbances arise not so much from rooting as from hydraulic jetting to expose molluscs and vacuuming to excavate and process them; captive walruses also ingested quantities of sand with live prey. Based on anatomical and histological data, Kastelein et al. (1997) suggested that the sensitive, mobile walrus tongue is well suited for manipulating bivalve molluscs. Although there are many purported uses of the evergrowing canine tusks in walruses, which emerge in both sexes and average 35 cm but can extend to 1 m, they do not appear to function in feeding except to act as a sled or stabilizer, allowing walruses to glide along the bottom and maintain upright orientation (Fig. 16.13). The once common explanation of their use as hoes, picks, rakes, or other implements to dislodge molluscs or excavate sediments has been discounted, for these are mainly social structures designed for intra- and interspecific display and defense, although they may also be used to chop ice, enlarge breathing holes, or hold and pull a walrus up and over the edge of an ice floe (Odobenus = "tooth walker"). Now a monotypic family, odobenids were widespread and diverse in the Late Miocene and Pliocene, some showing the heavy jaws and tusks of Odobenus while others resembled large otariids. However, the systematic relationship between walruses and the other two pinniped families remains extremely contentious. The presence of strikingly similar tusks and a walruslike skull in Odobenocetops peruvianus, a recently described Pliocene odontocete from Peru (De Muizon, 1993), suggests a remarkable convergence of cetacean and pinniped suction feeding. Known from cranial material alone, Odobenocetops shares several important morphological features with Odobenus, such as the deep, vaulted palate bearing long, ventrally projecting tusks and reduced maxillary dentition (absent altogether in Odobenocetops, as are postcanines in some fossil walruses) and wide, blunt, highly vascularized premaxillae with deep muscle scars, indicating a powerful, mobile, tactile upper lip, probably sporting vibrissae [e.g., see Kastelein et al. (1991) for Odobenus]. The exceptional autopomorphies of this delphinoid (closely related to monodontids), including the absence of a
16. Feeding in Marine Mammals
511
FIGURE 16.13. Convergent suction feeders: (a) walrus feeding (after Kastelein and Mosterd, 1989), and (b) sketch of fragmentary skull of Odobenocetops peruvianus, a fossil cetacean, in oblique anterolateral view.
rostrum (and melon, presumably), anteriorly placed nares, and modified skull telescoping with dorsally directed orbits and temporal fossa, all warrant placement in a novel family, Odobenocetopsidae, and are strongly suggestive of walrus-like feeding on benthic molluscs (Fig. 16.13). Although the asymmetrical incisor tusks of Odobenocetops erupt from protruding, sheath-like alveolar processes of the premaxillae (unlike the canine tusks of the walrus), they probably served the same social and sled functions. The eyes offered good binocular vision facing upward (or forward when feeding vertically), perhaps compensating for the lack of a melon for echolocation. As mentioned, walruses occupied the same ecological niche in the Pliocene. Not only does this present another extraordinary example of convergent evolution between contemporary mammals of northern and southern hemispheres, it demonstrates yet again the limited functional and morphological solutions to tetrapod feeding in an aquatic environment. C. Raptorial Feeding An obvious if somewhat less effective method of catching small items underwater involves seizing or grasping prey in the jaws with explosive forward or lateral movements of the head, neck, or entire body. This foraging method, herein referred to as raptorial feeding, is clearly the simplest and thus probably the original mode of prey capture in marine mammals, necessitating few changes from the terrestrial bauplan.
Hence, the most recently evolved marine mammals, including most within the order Carnivora, rely on raptorial feeding. However, some highly derived odontocetes have retained or secondarily adopted this feeding method, making marine mammal raptors a large and diverse group that combines generalists with extreme specialists. Raptorial feeding encompasses ram feeding described earlier as well as prey capture by snapping or biting. This is sometimes described as predatory feeding (Owen, 1980), but other feeding methods (e.g., suction) also involve active search and strike predation. A distinct feature of raptors is prey seizure with jaws. Specialized raptorial predators typically possess elongate "pincer" jaws and rostra bearing sharp teeth, a high degree of cervical flexibility, and a laterally directed visual field. Snappers often have a dorsoventrally compressed snout, in contrast to the high, laterally flattened skull of many terrestrial carnivores (Taylor, 1987). A flattened head, body, or tail serves as the fulcrum to brace rapid lateral or forward lunges. Smaller snappers often propel themselves forward with a burst of speed, whereas larger predators (e.g., sea lions, diving birds, and plesiosaurs) strike by swinging the head on an extensible, often S-shaped neck. The remarkably bent or "coiled" neck of the leopard seal can be extended in an instant to snatch prey. Numerous slender, pointed teeth retain and pierce grasped prey. In aquatic birds, modifications of the beak and tongue may perform the same grasping function. Seizers often
512
Alexander Werth
have enlarged pterygoid musculature (Taylor, 1987). Seagars (1982) showed that these jaw adductors predominate in several odontocete genera. An extreme group of "snapping" marine mammals are the river dolphins (Platanistidae, although these diverse and isolated species are often placed in up to four families), whose feeding is similar to that of gavials (Crocodylidae), diving birds, certain large raptorial fish such as gars (Lepisosteidae) and needlefish (Belonidae), and is probably much like that of numerous extinct reptiles (e.g., phytosaurs, ichthyosaurs, plesiosaurs, nothosaurs, mosasaurs). These long-beaked odontocetes (Fig. 16.14) have narrow jaws and rostra bearing up to 300 needle-like teeth, with which they stir the bottom or dart between roots or rocks, rapidly snapping their flexible necks to seize and impale single prey items (generally small fish). With its strong, cusped, vaguely molar-
F I G U R E 16.14. Variation in head profiles of odontocetes (not to scale): (a) pygmy sperm whale {Kogia breviceps), (b) northern bottlenose whale (Hyperoodon ampullatus), (c) Blainville's beaked whale (Mesoplodon densirostris), (d) killer whale {Orcinus orca), (e) narwhal {Monodon monoceros), (f) harbor porpoise {Phocoena phocoena), (g) longsnouted spinner dolphin (Stenella longirostris), and (h) Ganges susu {Platanista gangetica).
iform rear teeth, Inia geoffrensis, the bouto or Amazon river dolphin, can crush turtles, crustaceans, bivalve moUusks, and spiny and armored catfish. Although this falls well short of mastication or other precise prey processing, it is by far the most mechanical reduction seen in any cetacean. Some river and estuarine dolphins (e.g., Sotalia; Slijper, 1962) were once thought to eat fruit, leaves, and other vegetation, but the sparse evidence is highly doubtful. All raptors depend on highly refined senses to detect large, evasive prey. Platanista is functionally blind but possesses heightened tactile and auditory senses; the so-called "side-swimming" dolphins of this genus, equipped with exquisite echolocatory abilities, deftly snap tiny fish out of murky waters with quick lateral movements (actually dorsoventral flexion). Inia, which may detect prey via tiny tactile bristles on the rostrum, is also capable of similarly swift and extensive motions (Layne, 1959). Some fossil odontocetes display adaptations for raptorial feeding as extreme as those of platanistids, notably Miocene eurhinodelphids (with an exceptionally long, needle-like, overhanging upper jaw). Several abundant genera of coastal and oceanic delphinids (e.g., Stenella, Delphinus, Tursiops) clearly obtain prey with the aid of long jaws and numerous teeth (Fig. 16.14). These long-snouted forms diverge considerably from the suite of suction feeding characters described previously. At first glance, in fact, nearly all odontocetes seem to fit the raptorial category nicely, with their long, fusiform bodies, elongate jaws and projecting rostra, and array of small, conical teeth. However, as has been seen, these features do not apply to all toothed whales and dolphins, as most appear to possess bifocal visual fields and many lack mobile necks (the relatively flexible necks of monodontids and the Irrawaddy dolphin, Orcaella hrevirostris, are exceptions, presumably aiding in location and capture of benthic prey in shallow waters). Dentition, when present, is often greatly reduced in size and number, or is largely unerupted, so as to rule out any role in grasping prey. There is no question that odontocetes of several lineages are suction feeders, although virtually all dolphins (Delphinidae) and porpoises (Phocoenidae) are capable of raptorial feeding and probably rely on a combination of grasping and suction for prey capture, perhaps depending on prey type and habitat. The chisel-like, flattened, spatulate teeth of porpoises may be useful for slicing or shearing large prey. Many wild and captive delphinids have been observed shaking large prey at the surface (Norris and Prescott, 1961; Marsh et ah, 1989). Between this sharp dichotomy of suckers and snappers lie the killer whale, Orcinus orca, and its close relatives Pseudorca and Feresa. Although these blunt-headed
16. Feeding in Marine Mammals delphinids may be capable of suction generation (films of captive orca feeding support this presumption), the role of their formidable interlocking teeth is apparent, and indeed these species reduce large items using shark-like biting and tearing (Gladstone, 1988) or simply bolt whole prey. Gaskin (1976) thus described these species as the sole "sarcophagous" or meat-eating cetaceans. Pods of transient orcas display sophisticated pack-hunting patterns and are known to attack large mysticetes, tearing off chunks of the fat-rich tongue or blubber from live whales while surrounding their prey to prevent its escape and fatiguing or even asphyxiating it (e.g., Hancock, 1965; Baldridge, 1972; Whitehead and Glass, 1985; Silber et ah, 1990). Transient and resident orca pods may partition prey by size to avoid competition (Baird, 1994). Orcas often toy with prey and share it among pod members (Connor and Peterson, 1994). Johnson (1982) reported food sharing in captive Inia, in which a dominant male held prey in its jaws while smaller dolphins tore pieces off of it. Hult et al (1980) confirmed that raptorial odontocetes sometimes die from ingestion of prey too large to be swallowed. Raptorial odontocetes often feed on schooling fishes and invertebrates that are corralled into tight balls and trapped against the surface or bottom (Fertl and Wiirsig, 1995). Dolphins forage in large pods—thousands of individuals, when food is plentiful—to better locate and contain panicked prey (which may also suffer from oxygen deprivation) and then take turns passing through the school to seize individual items at random. Such foraging appears to be very effective; whether it is, indeed, cooperatively organized is uncertain, although intense vocal activity concurrent with herding suggests a concerted effort (Evans, 1987). Many odontocetes associate with feeding seabirds (Ridoux, 1987). Overall, foraging behaviors are reminiscent of the terrestrial grazers from which these social animals evolved (Wtirsig, 1989). Several instances of cooperative fishing between dolphins and wading fishermen in Africa and Central America have also been reported (Busnel, 1973; Pryor et al, 1990). Bottlenose dolphins {Tursiops truncatus) drive small fish onto the beach in shallows and tidal flats (Hoese, 1971) or sharp slopes (Norris and Dohl, 1980), repeatedly trapping schools of sciaenids (spot and silver perch) on the sand, where they are deftly snapped up; this has also been documented in humpback dolphins, Sousa plumbea (Peddemors and Thompson, 1994). Tursiops regularly follows trawlers and scavenges discarded fish and offal, and like some other odontocetes, procures prey from nets, long lines, and other fishing gear. Killer whales momentarily strand by riding tall waves onto steeply sloping beaches to pick off sea lions and elephant seals and then turn and wait for another wave to carry them off the shore. This species
513
may find breaks in offshore sandbars and reefs where such feeding is possible, and perhaps teach this behavior to juveniles (Lopez and Lopez, 1985). Orcas jump onto land or ice to grab prey (Norris and Prescott, 1961); other dolphins leap into the air to chase prey (Wells et al, 1980). Many odontocetes are nocturnal or crepuscular feeders that prey on the deep scattering layer of light-sensitive organisms that display diel vertical migrations (rising in the water column at night to feed on zooplankton). As in mysticetes, coloration appears important in odontocete predation. Species that feed below the photic zone are uniformly pigmented, whereas those that feed in upper layers are counter shaded. Lip and eye patches and strips alter or mask the head outline, whereas patterns on the back and flanks (striping, spotting, criss-crossing, saddling) may provide visual cues for schooling fish (Mitchell, 1970; Gaskin, 1982). Wiirsig (1986), Pryor and Norris (1991), and Connor and Peterson (1994) offer comprehensive reviews of delphinid foraging ecology and behavior. The standard pinniped dentition (Fig. 16.15), lacking carnassials or other sharp shearing or grinding surfaces, necessitates that prey too large to be swallowed whole be reduced by crushing, gnawing, or violent shaking or thrashing of the head, usually at the surface. Clawed phocids occasionally use foreflippers to grip and tear prey items. Because teeth posterior to the canines are essentially undifferentiated in all pinnipeds, cheek teeth are collectively designated as postcanines rather than as premolars and molars. Dental formulae are not diagnostic as individual variation is common. Postcanines of predominantly teuthophagous pinnipeds (e.g., elephant seals, Mirounga spp.) are remarkably reduced, particularly when compared to the enormous, recurved canines, although the latter are greatly enlarged for male-male fighting, dominance displays, and other social purposes in this and other polygynous species. As squid and other soft prey are often swallowed whole, postcanines are largely useless. The wide, flattened postcanines of monk seals (Monachus spp.), however, are presumably adapted for crushing fish. Riedman (1990) provided an extensive review of pinniped feeding ecology, including diet, dive patterns, cooperative foraging strategies and behaviors (often with other marine mammals or birds), and interference with fishing gear. A few marine mammals that normally grasp prey with limbs instead of teeth also fit the raptorial category. Marine fissiped carnivorans (polar bear, sea and marine otters, and extinct sea mink) do not differ appreciably from their terrestrial relatives in feeding ecology and morphology, which is to be expected given their relatively late reversion to the sea. The polar bear, Ursus maritimus, is probably the most recently evolved marine
514
Alexander Werth
F I G U R E 16.15. Upper dentition of pinnipeds (all to same scale, except h; bar: 5 cm): (a) gray seal {Halichoerus grypus), (b) bearded seal {Erignathus barbatus), (c) crabeater seal {Lobodon carcinophagus), (d) leopard seal (Hydrurga leptonyx), (e) Ross seal {Ommatophoca rossi), (f) southern elephant seal (Mirounga leonina), (g) Steller sea lion {Eumetopias jubatus), and (h) walrus {Odobenus rosmarus), with postcanines concealed medial to tusk alveolus.
mammal, thought to have diverged in the Pleistocene from subarctic brown bears, with which they successfully interbreed. Otters diverged from other mustelids in the Miocene, roughly 20 mya, but it is not known how recently the sea otter, Enhydra, took to the sea. The smaller eyes and poorer vision of these fissiped carnivorans relative to pinnipeds are taken as indications of their recent move to the ocean. The most significant changes in these marine mammals are dental adaptations to cope with new diets. The dentition of the sea otter, Enhydra lutris, differs markedly from that of other mustelids, as it is well adapted to crushing the shells, tests, and carapaces of sea urchins, crustaceans, and gastropod and bivalve molluscs, and scooping food from within (Fig. 16.16). The molars are unusually round, broad, and flat, with no trenchant cusps, and the blunt "carnassials" retain no shearing function. In mature adults, cheek teeth be-
come worn, broken, or pitted with caries due to the rough diet. Even the canines are rounded, and the shovel-like incisors are inclined anteriorly to extract the soft internal parts of broken prey items {Enhydra is the only carnivoran with four lower incisors and with no sharp teeth whatsoever). Nonelusive fishes are sometimes taken (particularly in Alaskan waters), although the diet consists mainly of invertebrate prey that are usually captured with the forepaws rather than the jaws (Chanin, 1985). During foraging dives of a minute or more (day and night, in depths ranging from 20 to 50 m) they gather food items, sensed mainly by long tactile whiskers, and store them in a large axillary or chest pouch formed by a loose flap of skin under the forelimb. Although the hind feet are flattened and modified as flippers, the forelimbs, with retractile claws, are highly mobile and adapted to locating, grasping, and manipulating prey. Curiously, sea otters invariably gather food
16. Feeding in Marine Mammals
515
FIGURE 16.16. Skulls and dentition of raptorial marine raptorial carnivorans (not to scale): (a) sea otter, Enhydra lutris, and (b) polar bear, Ursus maritimus.
with the right paw and store it under the left (Kenyon, 1986a). Food is always eaten at the surface. Sea otters are the smallest marine mammals (although the largest mustelids) and the only ones without a layer of insulating blubber. Consequently they require a great deal of food to survive in such cold waters and have ravenous appetites, consuming 20-25% of their body weight daily (Kenyon, 1986a). Food passes through the digestive system in approximately 3 hr (Kenyon, 1969). Individual sea otters often develop preferences for one or a few types of prey, although as a species, Enhydra is a generalist that forages on perhaps 50 invertebrate species at sea, on shore, and in tide pools (e.g., Ostfeld, 1982). Kelp is ingested only incidentally with food and is not digested; according to Kenyon, Enhydra cannot digest seal or bird meat. Gravid sea urchins are nutritious energy sources, and purple dyes found in urchin tests can stain sea otter skeletons (Loughlin, 1984). Sea otters have gained fame for their critical impact as "keystone" predators of kelp forest communities in the temperate North Pacific (Kvitek et al, 1992; Estes et al, 1998) and are also well known for their stereotypical feeding behavior, including unequivocal tool use. Feeding sea otters lie on their backs, knocking shells together or using a flattened stone as an anvil (and rarely as a hammer) on which to crack shells. They display many inventive feeding behaviors, such as puncturing discarded soda cans to get at octopi inside (Mason and Macdonald, 1986) or using bottles and other man-made objects as tools to crack shells, dig for bivalves, or pry abalones from rocks; their innovation and skillful dexterity often result in the destruction of pools and other captive accommodations. Whether tool use is learned or innate is unclear, despite extensive observation and experiment. Kenyon (1969) suggested that this behavior may derive from expressions of frustration. Lutra felina, the marine otter of coastal and riverine regions of tropical western South America, is often clas-
sified as a river otter, but is ecologically intermediate between the truly marine Enhydra and the seven other amphibious species of Lutra. Morphologically the marine otter is very similar to its freshwater congeners, showing few adaptations for life at sea. It spends much more time on land than the sea otter and does not venture far from shore, yet feeds on a variety of marine fish, crustaceans, molluscs, echinoderms, and worms (Reeves et al, 1992). The extinct sea mink, Mustela macrodon, lived on rocky shores and islands of the western North Atlantic, priniarily the Gulf of Maine and Bay of Fundy. Preying on marine fish and molluscs, it possessed adaptations for life in cold coastal waters, such as a larger body size, coarser pelt, and, as the name suggests, larger dentition relative to other minks and weasels, but was exterminated by intense hunting at the end of the 19th century (Nowak, 1991). The polar bear's scientific name, U. maritimus, reflects the seagoing nature of this largest bear species. Feeding in polar bears is generally similar to that of other ursids, aside from differences in foraging behavior related to their arctic habitat and prey. Polar bear dentition (Fig. 16.16) reflects a strongly carnivorous diet relative to the somewhat omnivorous habits of its terrestrial counterparts. The massive skull holds 38 to 42 sharp cutting teeth, including three simple incisors and one large canine in each tooth row. Molars are much smaller and narrower than in other modern ursids, and in general the head is narower. The staple of the polar bear diet is the ringed seal, Phoca hispida, although other species (including bearded, hooded, and harp seals) of all ages are consumed, especially pups and juveniles (often found by digging out subnivean dens). Seal blubber constitutes an important part of the polar bear's nutrition, as fat is assimilated and metabolized extremely well and stored for lean periods (Stirling, 1990). Polar bear distribution and migration are thus closely linked to seal availability.
516
Alexander Werth
Polar bears are very stealthy, solitary hunters, locating food by scent and often stalking it slowly while swimming or crawling around ice. Their acute olfactory sense enables them to identify and surreptitiously approach prey from great distances. Because their habitat is visually uniform, they keep low and hide behind ridges or other relief, blending into the icy terrain. However, the familiar tale of a polar bear sneakily covering its conspicuous black nose with a white paw is probably apocryphal, as are stories of bears hurling stones or blocks of ice to kill prey (Stirling, 1990). Like other ursids, the polar bear can run very rapidly; it often charges over the last 10-30 m to pounce upon prey. Polar bears also hunt patiently, resting for long periods near ice holes through which seals breathe and haul out, lying in wait to club startled seals with a forepaw and then grasping and dragging them out of the water with their teeth. Polar bears also prey on narwhals and belugas trapped in shallow water or cracks in ice, biting the blowhole of the whale to disable and asphyxiate it (Norris, 1994). There are numerous accounts of polar bears taking shore and water birds and eggs, small mammals, and possibly fish, although they often ignore abundant fish supplies (Reeves et al, 1992). They do not cache food, may leave large portions of carcasses uneaten, and are known to plunder carrion of whales, walruses, and caribou (and perhaps other bears). Scavenging on objects at dumps or other sites of human habitation is also well documented. Although frequently described as purely carnivorous, polar bears occasionally eat seaweed, grass, and berries, especially during summer and fall. Like other ursids, polar bears are capable of fasting for days between meals and for prolonged periods when denning in winter (having built up blubber stores). According to Stirling (1990), bears have been found with 70 kg of food in their stomachs, and Best (1985) estimated that the stomach can hold up to 20% of body weight. Polar bears not only destroy but also consume nonfood items ranging from foam and plastic to oil and other fluids (USFW, 1995). In a lengthy account of feeding methods, Stirling (1990) described how polar bears dispatch their prey with bites to the head and neck, hold down the carcass with a forepaw, and then tear off pieces of flesh and swallow them, head up, with little or no chewing. Food transport undoubtedly involves lingual and inertial forces. Initial feeding is voracious, as nearby bears may catch the scent and come to share the meal. Stirling (1990) also details the shear-like use of incisors to delicately snip away preferred seal blubber from underlying musculature; the skin is also peeled from the fat and seldom consumed. Fat is sometimes neatly "shaved" from the skin in this manner, leaving flayed hides on the ice.
D . Grazing The order Sirenia contains the sole herbivorous marine mammals. While mammalian success stems largely from the intense radiation of numerous lines of terrestrial herbivores, mammals have not profited similarly by exploiting primary producers in marine habitats (Anderson, 1979). Indeed, among extant sirenians or "sea cows," only the dugong {Dugong dugon) of the South Pacific and Indian Oceans can be considered exclusively marine, for the three manatee species are largely and, in some cases, purely riverine in distribution, rarely if ever venturing into salt water, so that they may be more accurately called "river cows." Nonetheless, all sirenians possess anatomical and physiological (including osmoregulatory) adaptations to the marine environment, and they undoubtedly constitute a monophyletic group, moreover one that clearly originated in the ocean. The preferences of West Indian {Trichechus manatus), west African (T. senegalensis), and Amazon (T. inunguis) manatees for freshwater springs, streams, and lakes or brackish estuaries niay depend on water temperature and food supply, although Hartman (1979) claimed that access to fresh water is an important factor in West Indian manatees. Manatees feed on the blades, leaves, and stems of a variety of aquatic terrestrial plants, including grasses, sedges, forbs, herbs, and mangroves; water hyacinth (Eichhornia) and Hydrilla are favorite foods, eaten in tremendous quantities. In fact, manatees were once prized for their ability to keep canals and rivers open and free of rapidly growing nuisance plants (AUsopp, 1960). Although their metabolic rate is low (among the lowest of all mammals), captive manatees can consume up to one quarter of their body weight daily (Nowak, 1991). Manatees will take floating, emergent, or overhanging vegetation, although all species seem to prefer submerged plants and often carry floating material underwater for feeding (Best, 1981; Domning, 1980) or fully uproot plants for consumption. They may graze on flood-inundated meadows and can partially beach themselves on banks to reach certain items (Reeves etal, 1992). Various floating fruits and seeds are also eaten. West Indian manatees forage extensively on acorns, locating oak trees where bird activity increases acorn availability as well as circular fish nests that collect fallen acorns (O'Shea, 1986). Manatees also graze or uproot marine grasses and epiphytic algae. Invertebrates ingested incidentally during feeding may provide critical dietary supplements. Captive manatees are given vitamins in fish or with horse chow, and wild manatees of the Caribbean are known to pull fish from nets, although fishermen may exaggerate the frequency of this behavior. West African manatees are often hunted as
16. Feeding in Marine Mammals
pests for purported damage to rice paddies and fishing gear. The dugong grazes on offshore and intertidal seagrass beds (and less frequently on green algae), plowing tracks through mudflats to dislodge anchored vegetation or stripping leaves from taller stems (Reeves et ah, 1992). Feeding dugongs seem to prefer the carbohydrate-rich rhizomes or storage roots of shorter seagrasses (e.g., Zostera), where foraging creates plumes of sediment and meandering tracks on the muddy bottom that can be seen from great distances. However, dugongs also crop taller, more delicate seagrasses (e.g., Thalassia, Amphibolis), presumably expending considerably less energy and subjecting the snout to less abrasion (Anderson, 1979). Sirenians are hindgut fermenters, like horses and rabbits, with extremely long intestines and a large, branched cecum harboring a rich bacterial microflora to facilitate cellulose digestion. The simple stomach is subdivided by a low gastric ridge but lacks a rumen or other chambers (Kenchington, 1972). Because of the high volume of low-quality food ingested, sirenian defecation and flatulation are nearly continuous, and regurgitation is frequent (Anderson, 1979; Hartman, 1979). Although no sirenians venture far from the surface, feeding dives generally last 5 min or less. Sirenians have few if any natural predators and are not especially agile or rapid swimmers (as evidenced by their frequent covering with algae and collisions with boats), but these placid and sluggish grazers possess sufficient sensory abilities to locate food at night or in turbid waters. Bristles on the lips and oral disc function in tactile sensation as well as manipulation (Marshall et ah, 1998a,b; Reep et ah, 1998). Other receptors include snout bristles, sinus hairs, cutaneous pressure receptors, proprioceptors, and rheoceptors. The sirenian postcranial skeleton resembles that of a cetacean, but Trichechus possesses only six cervical vertebrae (rather than the usual mammalian complement of seven) and all sirenians have exceptionally dense and pachyostotic bones, apparently an adaptation to achieve neutral buoyancy. Although the head often seems small in relation to the rotund, fusiform body, the skull is large and dense but delicately constructed, with the posteriorly directed external nares far back near the orbits. Reflecting differences in snouts, Trichechus has a long nasal cavity but small premaxillaries, whereas Dugong has a short nasal cavity (and lacks nasal bones) and long premaxillaries (Fig. 16.17). The deep, heavy mandible has a solid symphysis and teeth only on the posterior portion of each ramus. Sirenian dentition is highly modified, with incisors greatly reduced or absent and canines present only in fossil taxa (Nowak, 1991).
517
FIGURE 16.17. Skulls and dentition of (a) manatee, Trichechus manatus, and (b) dugong, Dugong dugon (not to scale).
Manatee dentition is notable for the continuous horizontal replacement of supernumerary cheek teeth, somewhat similar to that of the closely related proboscideans, although elephants have a limited number of teeth (Domning and Fiayek, 1984). Fiusar (1978), however, suggested that tooth replacement may cease in manatees of advanced age. Newborn calves have premolars as well as molars, but after weaning each entire tooth row begins to move anteriorly, perhaps due to mechanical stimulation from chewing (Best, 1984). An indefinite number of replacements migrate forward at a rate of roughly 1 mm per month (Domning and Hayek, 1984) to replace the worn teeth that fall out in front, an obvious adaptation to the abrasive diet of silicaceous plants. At any time from five to eight functional peglike molars are exposed in each tooth row. These lowcrowned, bunodont, double- and closed-rooted teeth have a thick layer of transverse-ridged cuspidate enamel but lack cementum. As teeth migrate forward, their roots are resorbed. Florny pads used to crop and grind plant matter rest on the front of the upper and lower jaws, covering the tiny 2/2 incisors that are lost before maturity. Dugongs, in contrast, possess only a few peg-like molars located far back on the jaws. Like young manatees, juvenile dugongs have premolars (three in each quadrant), but these are shed early in life (Reeves et ah, 1992), again by forward migration. The two or three columnar, single- and open-rooted adult cheek teeth have wrinkled occlusal surfaces that lack enamel yet are completely covered with cementum (Vaughan, 1986), conditions totally opposite those of Trichechus. Older males also sport a prominent pair of short, stubby incisor tusks, partially covered with enamel, that project from the jowly upper lip behind the facial disk but anterior to the mouth; these erupt at 12-15 years of age (Reeves et ah, 1992). The fact that these incisors rarely penetrate the gums of females, along with the presence of scars in mature males, again suggests a secondary sexual role (in competition or maneuvering during courtship)
518
Alexander Werth
rather than a food-gathering function (Nowak, 1991), although they may be used by males to scour or plow the bottom, as was suggested formerly (and erroneously) for walrus tusks. Chewing occurs also by rough, horny grinding plates (Fig. 16.18) that cover the distal surfaces of the palate and mandible (extending onto the lips). At the end of the dugong's blunt, squarish snout is a highly flexible, horseshoe-shaped, ventrally directed oral disk with a slit-like mouth (Fig. 16.19). This extremely supple, sensitive, bristle-covered disk (actually a greatly expanded and flattened upper lip) is pressed into the bottom to extract food and to direct it, with fanning waves of muscular contraction as in a precisely controlled conveyor system, to the n\outh (Anderson, 1984). As in manatees, the lips can be extended to grasp (with the help of the oral pads), uproot, and mash plants; however, because of the steep downward deflection of the rostrum and oral disk, Dugong is an obligate feeder on benthic macrophytes, whereas the more generalized Trichechus feeds throughout the water column, often taking floating or even overhanging vegetation (Anderson, 1979; Domning, 1980), although as noted earlier it typically pulls food underwater for processing. The thick, fleshy cleft or bilobed upper lips of the manatee's prominent, bristly muzzle, which hang down to cover the lower lips, are mobile and can grasp and pull plant matter toward the mouth (Marshall et ah, 1998a,b). This is accomplished by means of a muscular
a.
F I G U R E 16.18. Occlusal surfaces (anterior at top) of oral grinding plates on palates of (a) dugong (drawn from life, showing portions of lips) and (b) Steller's sea cow, Hydrodamalis gigas (from osteological specimen).
F I G U R E 16.19. Anterior and lateral views of oral discs of (a) manatee and (b) dugong showing a strongly bilobed lip (and more anterior nares) of former and ventral reflection of latter.
hydrostatic system in the snout and lips, which modulates movement and shape changes in the oral tissues, in particular controlling the position and movement of the perioral bristles used to sweep food into the mouth (Marshall et al, 1998a,b; Keep et al, 1998). The long, flexible forelimbs also manipulate food; Hartman (1979) described how pectoral flippers are used to scratch and clean the mouth. The rough tongue is small, with a tiny free anterior tip incapable of protrusion or other marked mobility. The tongue is covered with a thick, keratinized corium, yet possesses taste buds used in the selection of food and recognition of scent marks left by other manatees (Best, 1984). According to Hartman (1979), manatees chew incessantly, with a masticatory cycle of roughly two chews per second. Murie's (1872) thorough anatomical treatise and Domning's (1977, 1978a) detailed descriptions of dugong and manatee myology provide compelling insights into the functional morphology of sirenian feeding, including descriptive morphology of tongue, hyoid, jaws, palate, and pharynx. Yamasaki et al. (1980) noted the presence of mucous glands, taste buds, and tactile papillae in their comparison of the gross and microscopic anatomy of manatee and dugong tongues. The role of the lips in grasping, pulling, and conveying food is noteworthy, as is the extent to which the lips can be elevated, depressed, and flattened or pulled laterally. The extremely complex rostral musculature that permits precise movenients of the lateral and posterior margins of the oral disc and coordinates food grasping
16. Feeding in Marine Mammals and conveying functions involves a system of muscles that connect to the dense fibrous tissue of the snout (Domning, 1977,1978a; Marshall et al, 1998a,b). A deep, anterodorsally directed m. lateralis nasi also inserts in and may move the snout, although this muscle is obviously not typically associated with the lips. Other oral and facial muscles, such as the orbicularis oris and sphincter colli profundus pars oris, appear to assist food transfer from the lips to the mouth; Marshall et al. (1998a,b) suggested that this is a chief function of the buccinator and mentalis muscles. The mm. mandibulari and depressor anguli oris are small and lack the normal attachment to the corner of the mouth. The large, stiff bristles of the upper lip can be deliberately extended and retracted to assist further in manipulating food. Domning (1978a) proposed that the contraction of small muscle fibers surrounding the roots of these bristles and running to the skin is responsible for bristle extrusion, whereas on relaxation the bristles retract elastically into their sheaths; however, Marshall et al. (1998a) proposed a somewhat more complex model of bristle movement that combines direct muscle action and hydrostatic deformation of the snout and lips. Behavioral observation and morphological examination (both gross myology as well as study of tooth wear patterns) confirm that the sea cow masticatory apparatus relies primarily on horizontal motion of the mandible, rather than orthal movements found in some "ungulates." Sirenians share several features with pigs (Herring, 1972), such as a thick capsule surrounding the loose, unrestricted temporomandibular joint with rounded glenoid condyle; large pterygoid and coronoid processes; and highly pinnate jaw adductor muscles, including a reduced (single belly) digastric originating on the hyoid. However, a detailed quantitative vector analysis of Trichechus jaw mechanics (Domning, 1978a) based on Turnbull's (1970) formula of useful muscle power (and relying on fiber weights and orientations) revealed that sirenian jaw adductors, in fact, are quite different from those of ungulates, bearing more in common with primates (in terms of relative muscle masses) and carnivores (in terms of useful power). Domning proposed that jaw mechanics of Sirenia are so different from those of any of TurnbuU's categories (most notably in the large temporalis and small pterygoids) as to merit their placement in a unique group. The extremely large, pillar-like pterygoid processes, with apparent articular surfaces on the lateral and palatine sides, parallel to the coronoid arch, also suggest a singular arrangement. Mastication appears to be quite asymmetrical, with the center of mandibular rotation lying posteromedial to the massive pterygoid process on the active side and the contralateral temporalis dominant in transverse chewing (Domning, 1978a). This pterygoid-mandibular ar-
519
ticulation may absorb forces from transverse chewing, which occurs mainly by action of the temporalis, masseter, and internal pterygoid (the minor external pterygoid probably aiding in recovery rather than adduction, especially on the contralateral side), whereas the masseter and zygomaticomandibularis (the latter almost a superficial temporalis) are important in orthal chewing when the condylar joint is used as a fulcrum (Domning, 1978a). Mastication occurs mostly by lateral grinding, on one side of the mouth at a time, as the jaw is retracted by the paired digastrics and orthal force is provided by the adductors, chiefly temporalis (Domning, 1977). Elevation of the temporomandibular joint above the occlusal plane provides mechanical advantage for mastication, as the lowered digastric provides an advantage for jaw opening and improves jaw retraction (see Greaves, 1995). The more extensive, anteriorly projecting coronoid process of Trichechus (relative to Dugong) may provide a better lever arm for the temporalis appropriate to its longer tooth row and hence greater surface area for chewing (Domning, 1977). The extinct Steller's sea cow {Hydrodamalis gigas), a huge sirenian discovered in 1741 and exterminated due to hunting by 1768, had a head that was very small relative to its long (8 m), rotund body, and spent most of its time grazing on blades and stems of kelp and other marine algae of shallow waters around temperate islands of the North Pacific (Nowak, 1991). Lacking functional dentition, Hydrodamalis possessed a pair of white, horny (keratinous), tubercle-filled plates or rostral pads, long and flat yet ribbed with rugose V-shaped ridges or crests, one attached to the palate and the other to the mandible (Fig. 16.18), which fitted together to effectively crop and grind seaweed grasped or picked by the forelimbs (Forsten and Youngman, 1982). The thick, boat-shaped masticatory plates, similar to but proportionally much larger than those of Dugong, were composed of tightly packed, keratinized, tube-like cylinders. Steller described this species' short tongue and divided "internal and external lips," the protruding exterior lip thick and covered with stiff bristles (Haley, 1986). Whereas the mouth of Dugong projects ventrally for grazing on submerged plants, in Hydrodamalis the mouth was directed anteriorly for browsing on surface vegetation (Anderson, 1984). Steller's sea cow was formerly placed in its own family, Hydrodamalidae (DeBlase and Martin, 1981), although it is now recognized as a giant dugong, albeit the only strictly algivorous and cold-water one. The order Sirenia is traditionally classified with Proboscidea and Desmostylia in the superorder Tethytheria and with Hyracoidea in the more inclusive group Paenungulata (Reinhart, 1959). A recent, cladistic taxonomy includes all these taxa within an inclusive order
520
Alexander Werth
Uranotheria, with suborders Hyracoidea and Tethytheria (McKenna and Bell, 1997). With perhaps only 60,000 sirenians alive today, Sirenia is considered by far the rarest of all mammalian orders (Nowak, 1991). Sirenians originated in the Eocene Epoch from as yet unknown ancestors, presumably in the Tethys Sea, although the oldest fossils are from the New World (North, South, and Central America; Reinhart, 1959). Both dugongs and manatees were abundant, diverse, and widespread from the Oligocene through the Pliocene, especially in the Miocene when the global climate was relatively tropical. Although paleontological evidence suggests that manatees and dugongs diverged in the Middle Eocene (45 mya), Rainey et ah (1984), using sera and bone-extracted proteins for an immunological study of sirenian relationships, proposed a rather recent divergence in the Miocene (20 mya). Domning (1978b, 1981,1982) has written extensively on the evolutionary ecology of sirenians, discussing the role of numerous food and feeding-related factors, including the Tertiary expansion of seagrass meadows and freshwater macrophytes, especially true grasses (Gramineae), and the historical significance of such morphological characters as wear-resistant and replaceable dentition, tusks, jaw musculature, and rostral deflection, all of which doubtless had important effects on niche selection and specialization. Sirenians in turn may have had critical influences on the evolution of marine plant communities, among the most productive ecosystems (Barnes et al, 1985). Climatic cooling, marine regression, and a resulting transition from angiosperm to alga-dominated flora led to the extinction of certain North Pacific dugongids, yet allowed perpetuation of the line, including Hydrodamalis. In yet another example of the ecological and evolutionary interrelationships of marine mammals, Anderson (1995) suggested that sea otter carnivory on urchins permitted kelp growth to support hydrodamaline herbivory. Near extinction of sea otters by hunting may have contributed to the decline of Steller's sea cow prior to its demise at the hands of humans. Surely many marine mammals went extinct even in the absence of human interference. Desmostylians or "sea horses" constitute an extinct order of marine mammals of mysterious origin, affinity, and habits. Although several complete skeletons of four genera have been excavated from Late Oligocene and Miocene coastal deposits of the North Pacific, this group remains poorly understood. Desmostylians may have spent much of their time lounging on shore like walruses or wading through shallow coastal plains like the amphibious hippopotami, but in any case they must have been ungainly both on land and in water, and it is not known in which habitat they foraged. First described from isolated teeth and long considered a group
of sirenians, desmostylians are now presumed to share tethytherian ancestry with Sirenia (Novacek and Wyss, 1987) or are the sister group of Proboscidea (Domning et al, 1986; McKenna and Bell, 1997). The lack of living analogues makes desmostylian biology problematic; their dentition, cranial and postcranial morphology offer few clues. Paleoparadoxia ("fossil puzzle"), a Miocene desmostylian common on both sides of the Pacific basin (Fig. 16.20), is surely one of the most aptly named marine mammals. Desmostylians were pony sized and had a skull and deep mandible reminiscent of a horse, yet with decidedly nonequine teeth, including a battery of molars uniquely composed of clusters of dentine tubes surrounded by thick enamel (hence the name desmostylos = "chain pillar"), separated by a diastema from incisors and long, procumbent, tusk-like canines. The cheek teeth, which may have been continually replaced from behind as in manatees, are hypsodont in Desmostylus, yet brachyodont in Paleoparadoxia (MacLeod and Barnes, 1984). The skeleton lacked the pachyostosis of sirenians, and the stout, vaguely paddle-shaped foreand hindlimbs were curiously turned (rotated medially almost 90°), with wide feet bearing four hooved digits. The short legs and shovel-like mouth could have been used to dredge molluscs or pull vegetation from the sea floor (Mitchell, 1986; Savage and Long, 1986). The fact that walrus and desmostylian fossils are found together in the same deposits, yet walruses diversified just as desmostylians became extinct, has been proposed as evidence that "sea horses" could not compete with contemporaneous pinnipeds (although the walrus radiation could just as easily be a consequence and not a cause of desmostylian extinction). However, Domning et al. (1986), who described the apparent proboscidean roots of the Late Oligocene Behemotops, suggested that desmostylians were littoral herbivores that browsed on intertidal algae. Herbivory seems logical based on
FIGURE 16.20. Reconstruction of a desmostylian, Paleoparadoxia, after Savage and Long (1986).
521
16. F e e d i n g in M a r i n e M a m m a l s
Desmostylia's presumed tethytherian ancestry; however, even though desmostylian teeth are massive and covered with a thick layer of enamel, they lack suitable grinding ridges (each tooth bearing a single ring of thick enamel around a soft center of dentine), leading MacLeod and Barnes (1984) to speculate that desmostylians, like Triassic placodonts, used their unique cheek teeth to crush molluscs and other benthic invertebrates rooted from the bottom by their incisors. In any case, heavy wear and polishing of teeth indicate an abrasive diet.
obvious moUuscivore specialist, consumes prey ranging from minute amphipods to adult seals. As much as our knowledge of marine mammal feeding has slowly yet substantially improved in past decades, the few and limited experimental studies to date indicate how much remains to be learned. Although the obstacles to future work are great, the rewards in elucidating the ecology and evolution of these fascinating animals further should more than compensate the effort. Acknowledgments
III. C O N C L U S I O N S Reversion to marine habitats is a persistent theme of vertebrate evolution. While aquatic adaptation may be relatively easy for basal tetrapods, such as amphibians (given their low metabolism, tolerance of anoxia, capacity for anaerobic fermentation, and undulatory locomotion), it represents a dramatic shift for mammals and could not have been undertaken without significant benefits, such as the availability of plentiful food. Feeding is a crucial activity for marine mammals and, despite rampant parallel evolution, a source of critical characters in their classification. Rather than adopting a common feeding method, marine mammals radiated to exploit the varied resources and habitats of diverse marine environments. However, at the same time they were constrained by the physical limitations of life in water and thus developed a few basic strategies for prey capture and manipulation. Alhough millions of years of evolution separate the dugong, dolphin, whale, and walrus, all share a terrestrial heritage and exhibit striking convergence in their adaptations to herbivory, raptorial, filter, and suction feeding. Clearly there is considerable overlap in these divisions. The gray whale and crabeater seal (sucking filter feeders or filtering suction feeders) are prime examples of the limited solutions available. Above all, marine mammals are extremely adaptable feeders, displaying a great diversity of form and function. They exhibit complex foraging behaviors from luring, breaching, and lobtailing to cooperative herding and bubble netting. Some venture out of water briefly in their search for food, whereas others use sound to locate and possibly to debilitate prey. While versatility may be the rule, and perhaps a deciding factor in their relatively rapid evolution, there is no doubt that marine mammals demonstrate extraordinary alterations of the primitive eutherian condition. Despite such morphological specializations, they remain, for the most part, opportunistic generalists, able to exploit a variety of foods and circumstances. The walrus, an
I thank Kurt Schwenk for offering helpful comments on earlier versions of this manuscript. Portions of the research described in this chapter were conducted with the assistance of the New England Aquarium's Edgerton Research Laboratory. Illustrations were prepared by the author.
References Allsopp, W. H. L. (1960) The manatee: ecology and use for weed control. Nature 188:762. Anderson, P. K. (1979) Dugong behavior: on being a marine mammalian grazer. Biologist, 61(4): 113-144. Anderson, P K. (1984) Dugong. Pp. 298-299. In: The Encyclopedia of Mammals. D. Macdonald (ed.). Facts on File, New York. Anderson, P. K. (1995) Competition, predation, and the evolution and extinction of Steller's sea cow, Hydrodamalis gigas. Mar. Mamm. Sci. 11(3): 391-394. Arvy, L. (1974) Mammary glands, milk and lactation in cetaceans. Investi. Cetacea 5:157-202. Baird, R. W. (1994) Foraging Behaviour and Ecology of Transient Killer Whales (Orcinus orca). Ph.D. dissertation, Simon Eraser University, Burnaby, British Columbia. Baldridge, A. (1972) Killer whales attack and eat a gray whale. J. Mammal. 53(4): 898-900. Barnes, L. G., D. P. Domning, and C. E. Ray (1985) Status of studies on fossil marine mammals. Mar. Mamm. Sci. 1(1): 15-53. Barnes, L. C , and E. D. Mitchell (1978) Cetacea. Pp. 588-602. In: Evolution of African Mammals. V. J. Maglio and H. B. S. Cooke (eds.). Harvard Univ. Press, Cambridge, MA. Beale, T. 1839. The Natural History of the Sperm Whale: To Which Is Added, a Sketch of a South-Sea Whaling Voyage. John Van Voorst, London. Berkovich, V. M., and A. V. Yablokov (1963) The whale: an ultrasonic projector. Yuchni Teknik 3:76-77. Bengtson, J. L., and B. S. Stewart (1992) Diving and haulout behavior of crabeater seals in the Weddell Sea, Antarctica, during March (1986) Polar Biol. 12(6-7): 635-644. Berzin, A. A. (1972) The Sperm Whale. E. Hoz and Z. Blake, trans; A. V. Yablokov, ed. No. 600707, Israel Program for Science Translation, Jerusalem. Best, P. B. (1979) Social organization in sperm whales, Physeter macrocephalus. Pp. 227-289. In: Behavior of Marine Animals, Vol. 3. H. E. Winn and B. L. Olla (eds.). Plenum Press, New York. Best, R. C. (1981) Foods and feeding habits of wild and captive Sirenia. Mammal. Rev. 11:3-29. Best, R. C. (1984) Manatees. Pp. 295-298. In: The Encyclopedia of Mammals. D. Macdonald (ed.). Facts on File, New York.
522
Alexander Werth
Best, R. C. (1985) Digestibility of ringed seals by the polar bear. Can. J. Zool. 63:1033-1036. Bonner, W. N. (1989) Whales of the World. Facts on File, New York. Bonner, W. N. (1990) The Natural History of Seals. Facts on File, New York. Boschma, H. (1938) On the teeth and some other particulars of the sperm whale (Physeter macrocephalus L.). Temminckia III: 151-278. Boyd, I. L., J. P. Y. Arnould, T. Barton, and J. P. Croxall (1994) Foraging behaviour of antarctic fur seals during periods of contrasting prey abundance. J. Anim. Ecol. 63(3): 703-713. Brodie, P. F. (1977) Form, function, and energetic in Cetacea: a discussion. Pp. 45-48. In: Functional Anatomy of Marine Mammals. R. J. Harrison (ed.). Academic Press, New York. Brodie, P. F. (1989) The white whale Delphinapterus leucas (Pallas, 1776). Pp. 119-144. In: Handbook of Marine Mammals, Vol. 4. S. H. Ridgway and R. Harrison (eds.). Academic Press, San Diego. Brodie, P. F. (1993) Noise generated by the jaw actions of feeding fin whales. Can. J. Zool. 71:2546-2550. Brown, D. H. (1962) Further observations on the pilot whale in captivity Zoologica (New York) 47(1):59-64. Bryden, M. M., and W. J. L. Felts (1974) Quantitative anatomical observations on the skeletal and muscular systems of four species of Antarctic seals. J. Anat. 118(3):589-600. Busnel, R.-G. (1973) Symbiotic relationship between man and dolphins. Ann. N.Y. Acad. Sci. 35:112-131. Caldwell, D. K., M. C. Caldwell, and D. W. Rice (1966) Behavior of the sperm whale, Physeter catodon. Pp. 678-718. In: Whales, Dolphins, and Porpoises. K. S. Norris (ed.). University of California Press, Berkeley, CA. Carroll, G. M., J. C. George, L. L. Lowry, and K. O. Coyle (1987) Bowhead whale {Balaena mysticetus) feeding near Point Barrow, Alaska, during the 1985 spring migration. Arctic 40(2): 105-110. Chanin, P. (1985) The Natural History of Otters. Facts on File, New York. Clapham, P J., E. Leimkuhler, B. K. Gray, and D. K. Mattila (1995) Do humpback whales exhibit lateralized behaviour? Anim. Behav. 50(1): 73-82. Clarke, M. R. (1970) Function of the spermaceti organ of the sperm whale. Nature 228(5274): 873-874. Clarke, M. R. (1976) Buoyancy control as a function of the spermaceti organ in the sperm whale. J. Mar. Biol. Assoc. U.K. 58(1): 27-71. Clarke, M. R. (1979) The head of the sperm whale. Sci. Am. 240(1): 128-141. Clarke, R. (1955) A giant squid swallowed by a sperm whale. Norsk Hvalfangst-Tidende (Norwegian Whaling Gazette) 44(10): 589593. Clarke, R. (1966) The stalked barnacle Conchoderma, ectoparasitic on whales. Norsk Hvalfangst-Tidende (Norwegian Whaling Gazette) 55(8): 153-168. Clarke, R., A. L. Aguayo, and O. Paliza (1968) Sperm whales of the southeast Pacific. I and II. Introduction and size range, external characters, and teeth. Hvalradets Skrifter 51:1-80. Clarke, R., O. Paliza, and A. L. Aguayo (1988) Sperm whales of the southeast Pacific. IV. Fatness, food, and feeding. Invest. Cetacea 21:53-195. Connor, R. C , and D. M. Peterson (1994) The Lives of Whales and Dolphins. Henry Holt, New York. Costa, D. P. (1991) Reproductive and foraging energetics of high latitude penguins, albatrosses, and pinnipeds: implications for life history patterns. Am. Zool. 31:111-130. Dayton, P. K. (1975) Experimental studies of algal canopy interactions in a sea otter-dominated kelp community at Amchitka Island, Alaska. Fish. Bull. 73:230-237. Deblase, A. F., and R. E. Martin (1981) A Manual of Mammalogy with Keys to Families of the World. 2nd Ed. William C. Brown, Dubuque, lA.
De Muizon, C. (1993) Walrus-like feeding adaptation in a new cetacean from the Pliocene of Peru. Nature 365:745-748. Dolphin, W. F. (1988) Foraging dive patterns of humpback whales, Megaptera novaeangliae, in southeast Alaska, USA: a cost-benefit analysis. Can. J. Zool. 66(11):2432-2441. Donaldson, B. J. (1977) The tongue of the bottlenosed dolphin {Tursiops truncatus). Pp. 175-197. In: Functional Anatomy of Marine Mammals, Vol. 3. R. J. Harrison (ed.). Academic Press, New York. Domning, D. P. (1977) Observations on the myology of Dugong dugon (Muller). Smith. Contrib. Zool. 226:1-57. Domning, D. P. (1978a) The myology of the Amazonian manatee Trichechus inunguis (Natterer) (Mammalia: Sirenia). Acta Amaz. 8(2), Suppl. 1:1-81. Domning, D. P. (1978b) Sirenian evolution in the North Pacific Ocean. Univ. California Publ. Geol. Sci. 118:1-147. Domning, D. P. (1980) Feeding position preference in manatees {Trichechus). J. Mammal. 61(3): 544-547. Domning, D. P. (1981) Sea cows and sea grasses. Paleobiology 7(4): 417-420. Domning, D. P. (1982) Evolution of manatees: a speculative history. J. Paleontol. 56(3): 599-619. Domning, D. P., and L.-A. C. Hayek (1984) Horizontal tooth replacement in the Amazonian manatee (Trichechus inunguis). Mammalia 48(1): 105-127. Domning, D. P , C. E. Ray, and M. C. McKenna (1986) Two new Oligocene Desmostylians and a discussion of Tethytherian systematics. Smith. Contrib. Paleobiol. 59:1-56. Estes, J. A., and P. D. Steinberg (1988) Predation, herbivory, and kelp evolution. Paleobiology 14(1): 19-36. Estes, J. A., M. T. Tinker, T. M. WiUiams, and D. F. Doak (1998) Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282:473-476. Evans, P. G. H. (1987) The Natural History of Whales and Dolphins. Facts on File, New York. Fay, F. H. (1960) Structure and function of the pharyngeal pouches of the walrus (Odobenuc rosmarus L.). Mammalia 24:361-371. Fay, F. H. (1981) Walrus Odobenus rosmarus. Pp. 1-23. In: Handbook of Marine Mammals, Vol. 1. S. H. Ridgway and R. J. Harrison (eds.). Academic Press, San Diego. Fertl, D., and B. Wursig (1995) Coordinated feeding by Atlantic spotted dolphins {Stenella frontalis) in the Gulf of Mexico. Aquat. Mamm. 21(1): 3-5. Fordyce, R. E. (1977) The development of the Circum-Antarctic Current and the evolution of the Mysticeti (Mammalia: Cetacea). Palaeogeogr. Palaeoclimatol. Palaeoecol. 21:265-271. Fordyce, R. E. (1980) Whale evolution and Oligocene southern ocean environments. Palaeogeogr. Palaeoclimatol. Palaeoecol. 31:319336. Forsten, A., and P. M. Youngman (1982) Hydrodamalis gigas. Am. Soc. Mammal. Mammal. Spec. 165:1-3. Fristrup, K. M., and G. R. Harbison (1993) Vision and sperm whale foraging [Abstract]. Proceedings of the Tenth Biennial Conference on the Biology of Marine Mammals, Galveston, Texas: 49. Frost, K. J., and L. F. Lowry (1981) Foods and trophic relationships of cetaceans in the Bering Sea. Pp. 825-836. In: The Eastern Bering Sea Shelf: Oceanography and Resources. D. W. Hood and }. A. Calder (eds.). Office of Marine Pollution Assessment, National Oceanic and Atmospheric Administration. Gaskin, D. E. (1967) Luminescence in a squid Moroteuthis sp. (probably ingens Smith), and a possible feeding mechanism in the sperm whale Physeter catodon L. Tuatara 15:86-88. Gaskin, D. E. (1976) The evolution, zoogeography, and ecology of Cetacea. Oceanogr. Mar. Biol. Annu. Rev. 14:247-346. Gaskin, D. E. (1982) The Ecology of Whales and Dolphins. Heinemann Educational Books, Portsmouth, NH.
16. F e e d i n g in M a r i n e M a m m a l s Gladstone, W. (1988) Killer whale feeding observed underwater. J. Mammal. 69(3): 629-630. Goebel, M. E., J. L. Bengtson, R. L. DeLong, R. L. Gentry, and T. R. Loughlin (1991) Diving patterns and foraging locations of female northern fur seals. U.S. Nat. Mar. Fish. Serv. Fish. Bull. 89(2): 171-180. Gordon, K. R. (1980) Suction feeding in the walrus. Am. Zool. 20(4): 932. (Abstract) Gordon, K. R. (1984) Models of tongue movement in the walrus (Odobenus rosmarus). J. Morphol. 182:179-196. Gormley, G. (1983) Hungry humpbacks forever blowing bubbles. Sea Front. 29(5):258-265. Graur, D., and D. G. Higgins (1994) Molecular evidence for the inclusion of cetaceans within the order Artiodactyla. Mol. Biol. Evol. 11:357-364. Greaves, W. S. (1995) Functional predictions from theoretical models of the skull and jaws in reptiles and mammals. Pp. 99-115. In: Functional Morphology in Vertebrate Paleontology. J. Thomason (ed.). Cambridge Univ. Press, Cambridge. Gunter, G. (1951) Consumption of shrimp by the bottlenosed dolphin. J. Mammal. 32(4): 465-466. Hain, J. H. W., G. R. Carter, S. D. Kraus, C. A. Mayo, and H. E. Winn (1982) Feeding behavior of the humpback whale, Megaptera novaeangliae, in the western North Atlantic. Fish. Bull. 80(2): 259268. Hain, J. W., S. L. Ellis, R. D. Kenney, R J. Clapham, B. K. Gray, M. T. Weinrich, and I. T. Babb (1995) Apparent bottom feeding by humpback whales on Stellwagen Bank. Mar. Mamm. Sci. 11(4): 464-479. Haley, D. (1986) Steller sea cow. Pp. 264-269. In: Marine Mammals, 2nd Ed. D. Haley (ed.). Pacific Search Press, Seattle. Hamner, W. M., G. S. Stone, and B. S. Obst (1988) Behavior of southern right whales, Eubalaena australis, feeding on the Antarctic krill Euphausia superba. U.S. Nat. Mar. Fish. Ser. Fish. Bull. 86(1): 143150. Hancock, D. (1965) Killer whales kill and eat a Minke whale. J. Mammal. 46(2): 341-342. Harrison, R. J., F R. Johnson, and R. S. Tedder (1967) Underwater feeding, the stomach and intestine of some delphinids. J. Anat. 101(1): 186-187. Hartman, D. S. (1979) Ecology and Behavior of the Manatee (Trichechus manatus) in Florida. American Society of Mammalogists, Special Publication No. 5. Hays, H. E., H. E. Winn, and R. Petricig (1985) Anomalous feeding behavior of a humpback whale. J. Mammal. 66(4): 819-821. Heezen, B. C. (1957a) Whales entangled in deep sea cables. Deep-Sea Res. 4:105-115. Herring, S. W. (1972) The role of canine morphology in the evolutionary divergence of pigs and peccaries. J. Mammal. 53(3): 500-512. Herwig, R. P , J. S. Staley, M. K. Nerin, and H. W. Braham (1984) Baleen whales: evidence for forestomach microbial fermentation. Appl. Environ. Microbiol. 47(2): 421-423. Heyning, J. E. (1984) Functional morphology involved in intraspecific fighting of the beaked whale, Mesoplodon carlhubbsi. Can. J. Zool. 62:1645-1654. Heyning, J. E., and J. G. Mead (1991) Feeding mechanisms of Ziphiids: do beaked whales suck? Proceedings of the Ninth Biennial Conference on the Biology of Marine Mammals, Chicago: 33. (Abstract). Heyning, J. E., and J. G. Mead (1996) Suction feeding in beaked whales: morphological and observational evidence. Natural History Museum of Los Angeles Co., Contributions in Science 464: 31-12. Hoelzel, A. R., E. M. Dorsey, and S. J. Stern (1989) The foraging specializations of individual minke whales. Anim. Behav. 38(5): 786-794.
523
Hoese, H. D. (1971) Dolphin feeding out of water in a salt marsh. J. Mammal. 52(1):222-223. Hosokawa, H., and T. Kamiya (1971) Some observations on the cetacean stomachs, with special considerations on the feeding habits of whales. Sci. Rep. Whales Res. Insti. 23:91-101. Howell, A. B. (1930) A further note on aquatic mammals. J. Mammal. 11:432. Hubbs, C. L. (1946) First records of two beaked whales, Mesoplodon bowdoini and Ziphius cavirostris, from the Pacific coast of the United States. J. Mammal. 27(3): 242-255. Hult, R. (1982) Another function of echolocation for bottlenosed dolphins {Tursiops truncatus). Cetology 47:1-7. Hult, R. W., S. E. Dupey, and R. W. Bradley (1980) Mortalities associated with prey ingestion by small cetaceans. Cetology 38:1-2. Husar, S. L. (1978) Trichechus manatus. Am. Soc. Mammalogists Mammalian Spe. 93:1-5. Ishiyama, M. (1987) Enamel structure in odontocete whales. Scann. Microsc. 1(3): 1071-1079. Johnson, R. H. (1982) Food-sharing behavior in captive Amazon river dolphins {Inia geoffrensis). Cetology 43:1-2. Jones, R. E. (1981) Food habits of smaller marine mammals from northern California. Proc. California Acad. Sci. 42(16):409433. Jurasz, C. M., and V. P. Jurasz (1979) Feeding modes of the humpback whale, Megaptera novaeangliae, in southeast Alaska. Sci. Rep. Whales Res. Instit. 31:69-83. Kastelein, R. A., and J. L. Dubbeldam (1990) Marginal papillae on the tongue of the harbour porpoise {Phocoena phocoena), bottlenose dolphin {Tursiops truncatus), and Commerson's dolphin {Cephalorhynchus commersonii). Aquat. Mamm. 15(4): 158-170. Kastelein, R. A., and P. Mosterd (1989) The excavation technique for molluscs of Pacific walrusses (Odobenus rosmarus divergens) under controlled conditions. Aquat. Mamm. 15(1): 3-5. Kastelein, R. A., J. L. Dubbeldam, and M. A. G. De Bakker (1997) The anatomy of the walrus head {Odobenus rosmarus). 5. The tongue and its function in walrus ecology. Aquat. Mamm. 23:29-47. Kastelein, R. A., N. M. Gerrits, and J. L. Dubbeldam (1991) The anatomy of the walrus head {Odobenus rosmarus). 2. Description of the muscles and of their role in feeding and haul-out behaviour. Aquat. Mamm. 17:156-180. Kastelein, R. A., M. MuUer, and A. Terlouw (1994) Oral suction of a Pacific walrus (Odobenus rosmarus divergens) in air and under water. Zeitschrift Saugetierkunde 59:105-115. Kasuya, T., and D. W. Rice (1970) Note on baleen plates and on arrangement of parasitic barnacles of gray whale. Sci. Rep. Whales Res. Instit. 22:39-43. Kato, H. (1984) Observation of tooth scars on the head of male sperm whale, as an indication of intra-sexual fightings. Sci. Rep. Whales Res. Instit. 35:39-46. Kawakami, T. (1980) A review of sperm whale food. Sci. Rep. Whales Res. Instit. 32:199-218. Kawamura, A. (1974) Food and feeding ecology in the southern sei whale. Sci. Rep. WhaleslRes. Instit. 26:25-144. Kenchington, R. A. (1972) Observations on the digestive system of the dugong, Dugong dugon (Erxleben). J. Mammal. 53:884-887. Kenyon, K. W. (1969) The sea otter in the eastern Pacific Ocean. U.S. Fish Wildlife Serv. North Am. Fauna Ser. 68:1-352. Kenyon, K. W. (1986a) Sea otter. Pp. 254-263. In: Marine Mammals, 2nd Ed. D. Haley (ed.). Pacific Search Press, Seattle. Kenyon, K. W. (1986b) Walrus. Pp. 202-207. In: Marine Mammals, 2nd Ed. D. Haley (ed.). Pacific Search Press, Seattle. Kim, S. L., and J. S. Oliver (1989) Swarming benthic crustaceans in the Bering and Chukchi Seas and their relation to geographic patterns in gray whale feeding. Can. J. Zool. 67(6): 1531-1542. King, J. E. (1961) The feeding mechanisms and jaws of the crabeater seal {Lobodon carcinophagus). Mammalia 25:462-466.
524
Alexander Werth
King, J. E. (1964) Swallowing modifications in the Ross seal. J. Anat. 99:206-207. King, J. E. (1972) Observations on phocid skulls. Pp. 81-115. In: Functional Anatomy of Marine Mammals. R. J. Harrison (ed.). Academic Press, New York. King, J. E. (1983) Seals of the World, 2nd Ed. Cornell University Press, Ithaca, NY. Klaus, A. D., J. S. Oliver, and R. G. Kvitek (1990) The effects of gray whale, walrus, and ice gouging disturbance on benthic communities in the Bering Sea and Chukchi Sea, Alaska. Nat. Geograph. Res. 6(4): 470-484. Kleinenberg, S. E., A. V. Yablokov, B. M. Bel'kovich, and M. N. Tarasevich (1969) Beluga (Delphinapterus leucas)—Investigation of the species. O. Theodor, ed. No. 6751345, Israel Program for Science Translation, Jerusalem. Kvitek, R. G., J. S. Oliver, A. R. DeGange, and B. S. Anderson (1992) Changes in Alaskan soft-bottom prey communities along a gradient in sea otter predation. Ecology 73(2): 413-428. Lambertsen, R. H. (1983) Internal mechanism of rorqual feeding. J. Mammal. 64(1): 76-88. Lambertsen, R. H., R. J. Hintz, W. C. Lancaster, A. Hirons, K. J. Kreiton, and C. Moor (1989) Characterization of the functional morphology of the mouth of the bowhead whale, Balaena mysticetus, with special emphasis on feeding and filtration mechanisms. Report to the Department of Wildlife Management, North Slope Borough, Barrow, Alaska. Lambertsen, R., N. Ulrich, and J. Straley (1995) Frontomandibular stay of Balaenopteridae: a mechanism for momentum recapture during feeding. J. Mammal. 76(3): 877-899. Lauder, G. V. (1985) Aquatic feeding in lower vertebrates. Pp. 210229. In: Functional Vertebrate Morphology. M. Hildebrand, D. M. Bramble, K. R Liem, and D. B. Wake (eds.). Belknap/Harvard Univ. Press, Cambridge. Laws, R. M. (1984) Crabeater seal. Pp. 280-281. In: The Encyclopedia of Mammals. D. Macdonald (ed.). Facts on File, New York. Layne, J. N. (1959) Feeding adaptations and behavior of a freshwater dolphin, Inia geojfrensis. Anat. Rec. 134(3): 598. Leatherwood, S., R. R. Reeves, and L. Foster (1983) The Sierra Club Handbook of Whales and Dolphins. Sierra Club Books, San Francisco. Lipps, J. H., and E. Mitchell (1976) Trophic model for the adaptive radiations and extinctions of pelagic marine mammals. Paleobiology 2:147-155. Lopez, J. C , and D. Lopez (1985) Killer whales {Orcinus orca) of Patagonia, and their behavior of intentional stranding while hunting nearshore. J. Mammal. 66(1): 181-183. Loughlin, T. R. (1984) Sea otter. P 127. In: The Encyclopedia of Mammals. D. Macdonald (ed.). Facts on File, New York. Lowry, L. R, and K. J. Frost (1984) Foods and feeding of bowhead whales in Western and Northern Alaska. Sci. Rep. Whales Res. Instit.35:l-16. MacKay, R. S., and J. Pegg (1988) Debilitation of prey by intense sounds. Mar. Mamm. Sci. 4(4): 356-359. Marsh, H., R. Lloze, G. E. Heinsohn, and T. Kasuya (1989) Irrawaddy dolphin Orcaella brevirostris (Gray, 1966). Pp. 101-118. In: Handbook of Marine Mammals, Vol. 4. S. H. Ridgway and R. Harrison (eds.). Academic Press, San Diego. Marshall, C. D., L. A. Clark, and R. L. Reep (1998) The muscular hydrostat of the Florida manatee {Trichechus manatus latirostris): a functional morphological model of perioral bristle use. Mar. Mamm. Sci. 14:290-303. Marshall, C. D., G. D. Huth, V. M. Edmonds, D. L. Halin, and R. L. Reep (1998) Prehensile use of perioral bristles during feeding and associated behaviors of the Florida manatee {Trichechus manatus latirostris). Mar. Mamm. Sci. 14:274-289. Mason, C. R, and S. M. Macdonald (1986) Otters: Ecology and Conservation. Cambridge Univ. Press, New York.
Mayo, C. A., and M. K. Marx (1990) Surface foraging behaviour of the North Atlantic right whale, Eubalaena glacialis, and associated zooplankton characteristics. Can. J. Zool. 68:2214-2220. McCann, C. (1974) Body scarring on cetacea-odontocetes. Sci. Rep. Whales Res. Instit. 26:145-155. McKenna, M. C , and S. K. Bell (1997) Classification of Mammals Above the Species Level. Columbia Univ. Press, New York. McLeod, S. A., and L. G. Barnes (1984) Fossil desmostylians. Mem. Nat. Hist. Found. Orange County 1:39-44. Melville, H. 1851. Moby-Dick, or The Whale. Harper's Press, New York. Miller, G. S. (1929) The gums of the porpoise Phocoenoides dalli (True). Proc. U.S. Nat. Mus. 74:1-4. Mitchell, E. D. (1970) Pigmentation pattern and evolution in delphinid cetaceans: an essay in adaptive coloration. Can. J. Zool. 48:717-740. Mitchell, E. D. (1986) Origins of Eastern North Pacific sea mammal fauna. Pp. 21-28. In: Marine Mammals. 2nd Ed. D. Haley (ed.). Pacific Search Press, Seattle. Mitchell, E. D. (1989) A new cetacean from the Late Eocene La Meseta Formation, Seymour Island, Antarctic Peninsula. Can. J. Fish. Aquat. Sci. 46(12):2219-2236. Morris, R. A., and L. S. Mowbray (1966) An unusual barnacle attachment on the teeth of the Hawaiian spinning dolphin. Norsk Hvalfangst-Tidende (Norwegain Whaling Gazette) 55(1): 15-16. Murie, J. (1865) On deformity of the lower jaw in the cachalot {Physeter macrocephalus, Linn.). Proc. Zool. Soc. Lond. 1865:390396. Murie, J. (1872) On the form and structure of the manatee (Manatus americanus). Trans. Zool. Soc. Lond. 8(3): 127-202. Murison, L. D., and D. E. Gaskin (1989) The distribution of right whales and zooplankton in the Bay of Fundy, Canada. Can. J. Zool. 67(6): 1411-1420. Murison, L. D., D. J. Murie, K. R. Morin, and J. Da Silva Curiel (1984) Foraging of the gray whale along the west coast of Vancouver Island, British Columbia. Pp. 451-463. In: The Gray Whale Eschrichtius robustus. M. L. Jones, S. L. Swartz, and S. Leatherwood (eds.). Academic Press, Orlando. Nakamura, K. (1968) Studies on the sperm whale with deformed lower jaw with special reference to its feeding. Bull. Kanagawa Prefecture Mus. Nat. Hist. 1:13-27. Nasu, K. (1958) Deformed lower jaw of sperm whale. Sci. Rep. Whales Res. Instit. 13:211-212. Nelson, C. H., and K. R. Johnson (1987) Whales and walruses as tillers of the sea floor. Sci. Am. 256(2): 112-117. Nemoto, T. (1959) Food of baleen whales with reference to whale movements. Sci. Rep. Whales Res. Instit. 14:149-241. Nemoto, T. (1970) Feeding pattern of baleen whales in the ocean. Pp. 241-252. In: Marine Food Chains. J. H. Steele (ed.). Oliver and Boyd, Edinburgh. Nemoto, T., and K. Nasu (1963) Stones and other aliens in the sperm whales in the Bering Sea. Sci. Rep. Whales Res. Instit. 17:83-91. Nerini, M. (1984) A review of gray whale feeding ecology. Pp. 4 2 3 450. In: The Gray Whale Eschrichtius robustus. M. L. Jones, S. L. Swartz, and S. Leatherwood (eds.). Academic Press, Orlando. Norris, K. S. (1994) Beluga: white whale of the north. Nat. Geograph. 185(6): 2-31. Norris, K. S., and T. P Dohl (1980) The structure and functions of cetacean schools. Pp. 211-261. In: Cetacean Behavior: Mechanisms and Functions. L. M. Herman (ed.). Wiley, New York. Norris, K. S., and G. W. Harvey (1972) A theory for the function of the spermaceti organ of the sperm whale {Physeter catodon L.). Pp. 397-417. In: Animal Orientation and Navigation. National Air and Space Administration, Washington, D.C. Norris, K. S., and B. Mohl (1983) Can odontocetes debilitate prey with sound? Am. Nat. 122(1): 85-104. Norris, K. S., and J. H. Prescott (1961) Observations on Pacific ceta-
16. F e e d i n g in M a r i n e M a m m a l s ceans of Californian and Mexican waters. Univ. California Publ. Zool. 63(4):291-402. Novacek, M. J. (1992) Mammalian phylogeny: shaking the tree. Nature 356:121-125. Novacek, M. J., and A. R. Wyss (1987) Selected features of the desmostylian skeleton and their phylogenetic implications. Am. Mus. Novitates 2870:1-8. Nowak, R. M. (1991) Walkers Mammals of the World, 5th Ed. Johns Hopkins Univ. Press, Baltimore, MD. O'shea, T. J. (1986) Mast foraging by West Indian manatees (Trichechus manatus). J. Mammal. 67(1): 183-185. Olesiuk, P. F. (1993) Annual prey consumption by harbor seals (Phoca vitulina) in the Strait of Georgia, British Columbia. U.S. Nat. Mar. Fish. Serv. Fish. Bull. 91(3): 491-515. Oliver, J. S., and P. N. Slattery (1985) Destruction and opportunity on the sea floor: effects of gray whale feeding. Ecology 66(6): 19651975. 0ritsland, T. (1977) Food consumption of seals in the Antarctic pack ice. Pp. 749-768. In: Adaptations within Antarctic Ecosystems. G. A. Llano (ed.). Proceedings of the Third SCAR Symposium on Antarctic Biology. Smithsonian Institution, Washington, DC. Orton, L. S., and P. F. Brodie (1987) Engulfing mechanics of fin whales. Can. J. Zool. 65:2898-2907. Ostfeld, R. S. (1982) Foraging strategies and prey switching in the California sea otter. Oecologia 53:170-178. Owen, J. (1980) Feeding Strategy. University of Chicago Press, Chicago. Peddemors, V. M., and G. Thompson (1994) Beaching behaviour during shallow water feeding by humpback dolphins, Sousa plumbea. Aquat. Mamm. 20(2): 65-67. Pilleri, G., M. Gihr, and C. Kraus (1970) Feeding behavior of the Gangetic dolphin, Platanista gangetica, in captivity. Invest. Cetacea 2: 69-73. Pivorunas, A. (1976) A mathematical consideration on the function of baleen plates and their fringes. Sci. Rep. Whales Res. Instit. 28: 37-55. Pivorunas, A. (1977) The fibrocartilage skeleton and related structures of the ventral pouch of balaenopterid whales. J. Morphol. 151:299-314. Pivorunas, A. (1979) The feeding mechanisms of baleen whales: Am. Sci. 67:432-440. Pryor, K., J. Lindbergh, S. Lindbergh, and R. Milano (1990) A dolphin-human fishing cooperative in Brazil. Mar. Mamm. Sci. 6(1): 77-82. Pryor, K., and K. S. Norris, eds. (1991) Dolphin Societies: Discoveries and Puzzles. University of California Press, Berkeley, CA. Rainey, W. E., J. M. Lowenstein, V. M. Sarich, and D. M. Magor (1984) Sirenian molecular systematics, including the extinct Steller's sea cow {Hydrodamalis gigas). Naturwissenschaften 71(586): 1-3. Ray, C. (1966) Round table: practical problemanuscript in preparation P. 671. In: Whales, Dolphins, and Porpoises. K. S. Norris (ed.). University of California Press, Berkeley, CA. Ray, G. C , and W. E. Schevill (1974) Feeding of a captive gray whale, Eschrichtius robustus. Mar. Fish. Rev. 36(4): 31-38. Reep, R. L., C. D. Marshall, M. L. StoU, and D. M. Whitaker (1998) Distribution and innervation of facial bristles and hairs in the Florida manatee {Trichechus manatus latirostris). Mar. Mamm. Sci. 14:257-273. Reeves, R. R., B. S. Stewart, and S. Leatherwood (1992) The Sierra Club Handbook of Seals and Sirenians. Sierra Club Books, San Francisco. Reidenberg, J. S., and J. T. Laitman (1994) Anatomy of the hyoid apparatus in Odontoceti (toothed whales): specializations of their skeleton and musculature compared with those of terrestrial mammals. Anat. Rec. 240:598-624. Reinhart, R. H. (1959) A review of the Sirenia and Desmostylia. Univ. California Publ. Geol. Sci. 36(1): 1-146.
525
Rice, D. W. (1989) Sperm whale Physeter macrocephalus Linnaeus, 1758. Pp. 177-233. In: Handbook of Marine Mammals. Vol. 4. S. H. Ridgway and R. Harrison (eds.). Academic Press, San Diego. Ridgway, S. H. (1965) Medical care of marine mammals. J. Am. Vet. Medi. Assoc. 147(10): 1077-1085. Ridoux, V. (1987) Feeding association between seabirds and killer whales, Orcinus orca, around subantarctic Crozet Islands, France. Can. J. Zool. 65(8): 2113-2115. Riedman, M. (1990) The Pinnipeds: Seals, Sea Lions, and Walruses. University of California Press, Berkeley, CA. Ross, G. J. B., F Ryan, G. S. Saayman, and J. Skinner (1976) Observations on two captive crabeater seals at the Port Elizabeth Oceanarium. Int. Zoo Yearbook 16:160-164. Ross, J. P. (1987) Are odontocetes suction feeders? Proceedings of the Biennial Conference of the North Atlantic Marine Mammal Association, Boston. (Abstract) Rubenstein, D. I., and M. A. R. Koehl (1977) The mechanisms of filter feeding: some theoretical considerations. Am. Nat. 111:981-994. Sanderson, S. L., and R. Wassersug (1990) Suspension-feeding vertebrates. Sci. Am. 262(3):96-101. Sanderson, S. L., and R. Wassersug (1993) Convergent and alternative designs for vertebrate suspension feeding. Pp. 37-112. In: The Skull, Vol. 3. J. Hanken and B. K. Hall (eds.). University of Chicago Press, Chicago. Savage, R. J. G., and M. R. Long (1986) Mammal Evolution: An Illustrated Guide. Facts on File, New York. Seagars, D. J. (1982) Jaw Structure and Functional Mechanics of Six Delphinids {Cetacea: Odontoceti). M.S. thesis, San Diego State University. Silber, G. K., M. W. Newcomer, and M. H. Perez-Cortes (1990) Killer whales, Orcinus orca, attack and kill a Bryde's whale, Balaenoptera edeni. Can. J. Zool. 68(7): 1603-1606. Simenstad, C. C , L. F Small, and C. D. Mclntire (1990) Consumption processes and food web structure in the Columbia river estuary. Prog. Oceanogr. 25(1-4): 271-298. Simila, T., and F Ugarte (1993) Surface and underwater observations of cooperatively feeding killer whales in northern Norway. Can. J. Zool. 71(8): 1494-1499. Slijper, E. J. (1962) Whales. Basic Books, New York. Smith, S. C , and H. Whitehead (1993) Variations in the feeding success and behaviour of Galapagos sperm whales {Physeter macrocephalus) as they relate to ocaenographic conditions. Can. J. Zool. 71(10): 1991-1996. Sokolov, V. E., and O. V. Volkova (1973) Structure of the dolphin's tongue. Pp. 119-127. In: Morphology and Ecology of Marine Mammals: Seals, Dolphins, and Porpoises. K. K. Chapskii and V. E. Sokolov (eds.). Wiley, New York. Spaul, E. A. (1964) Deformity in the lower jaw of the sperm whale {Physeter catodon). Proc. Zool. Soc. Lond. 142(3): 391-395. Stirling, I. (1990) Polar Bears. University of Michigan Press, Ann Arbor. Storro-Patterson, R. (1981) Great gulping blue whales. Oceans 14:16. Sund, P. N. (1975) Evidence for feeding during migration and of an early birth of the Californian gray whale {Eschrichtius robustus). J. Mammal. 56:265. Tarpley, R. J. (1985) Gross and Microscopic Anatomy of the Tongue and Gastrointestinal Tract of the Bozvhead Whale (Balaena mysticetus). Ph.D. dissertation, Texas A & M University, College Station, Texas. Taylor, M. A. (1987) How tetrapods feed in water: a functional analysis by paradigm. Zool. J. Linn. Soc. 91:171-195. Tomilin, A. G. (1954) Adaptive types of the order Cetacea. Zool. Zhur. Moscow 33(3): 677-691. Bur. Trans. Lang. Div. Canada. Turnbull, W. D. (1970) Mammalian masticatory apparatus. Fieldiana (Geology) 18(2): 147-356. United States Fish and Wildlife Service (1995) Habitat Conservation
526
Alexander Werth
Strategy for Polar Bears in Alaska. Draft Report, USFWS Alaska Region. Vaughan, T. A. (1986) Mammalogy, 3rd Ed. Saunders, Philadelphia. Von Schulte, H. W. (1916) The sei whale {Balaenoptera borealis Lesson): anatomy of a foetus of Balaenoptera borealis. Monographs of the Pacific Cetacea. Mem. Am. Mus. Nat. Hist. (New Series) 1:389499. Watkins, W. A., and W. E. Schevill (1976) Right whale feeding and baleen rattle. J. Mammal. 57(1):58-66. Watkins, W. A., and W. E. Schevill (1979) Aerial observation of feeding behavior in four baleen whales: Eubalaena glacialis, Balaenoptera borealis, Megaptera novaeangliae, and Balaenoptera physalus. J. Mammal. 60(1): 155-163. Weitkamp, L. A., R. A. Wissmar, C. A. Simenstad, K. L. Fresh, and J. G. Odell (1992) Gray whale foraging on ghost shrimp, Callianassa californiensis, in littoral sand flats of Puget Sound, USA. Can. J. Zool. 70(12): 2275-2280. Wells, R. S., A. B. Irvine, and M. D. Scott (1980) The social ecology of inshore odontocetes. Pp. 263-317. In: Cetacean Behavior: Mechanisms and Functions, L. M. Herman (ed.). Wiley, New York. Werth, A. J. (1989) Suction feeding in odontocetes: water flow and head shape. Am. Zool. 29(4) :92A. (Abstract) Werth, A. J. (1990) Functional anatomy of the right whale tongue. Am. Zool. 30(4): 21 A. (Abstract) Werth, A. J. (1992) Anatomy and Evolution of Odontocete Suction Feeding. Ph.D. dissertation. Harvard University, Cambridge, MA. Werth, A. J. (1993) Functional morphology of balaenid whale tongues. Proceedings of the Tenth Biennial Conference on the Biology of Marine Mammals, Galveston, Texas: 112. (Abstract) Werth, A. J. (1995) Models of hydrodynamic flow in the bowhead filter feeding apparatus. Am. Zool. 35(5) :59A. (Abstract)
Werth, A. J., and D. N. Stern (1992) Functional influences in the evolution and devolution of odontocete enamel. J. Verteb. Paleontol. 12(3, Suppl.):59A. (Abstract) Werth, A. J. (2000) A kinematic study of suction feeding and associated behavior in the long-finned pilot whale, Globicephala melas (Traill). Mar. Mamm. Sci. 16(2):299-314. Whitehead, H. (1989) Formations of foraging sperm whales (Physeter macrocephalus) off the Galapagos Islands, Ecuador. Can. J. Zool. 67(9): 2131-2139. Whitehead, H., and C. Carlson (1988) Social behavior of feeding finback whales off Newfoundland, Canda: comparison with the sympatric humpback whale. Can. J. Zool. 66(1): 217-221. Whitehead, H., and C. Glass (1985) Orcas (killer whales) attack humpback whales. J. Mammal. 66(1): 183-185. Wiirsig, B. (1986) Delphinid foraging strategies. Pp. 347-359. In: Dolphin Cognition. J. Thomas, R. Schusterman, and F. Wood (eds.). Erlbaum, Hillsdale, NJ. Wiirsig, B. (1988) The behavior of baleen whales. Sci. Am. 258(4): 102-107. Wursig, B. (1989) Cetaceans. Science 244:1550-1557. Yamasaki, R, S. Komatsu, and T. Kamiya (1980) A comparative morphological study on the tongues of manatee and dugong (Sirenia). Sci. Rep. Whales Res. Instit. 32:127-144. Yamasaki, R, H. Satomi, and T. Kamiya (1976) An observation on the papillary projections at the lingual margin in the striped dolphin. Sci. Rep. Whales Res. Instit. 28:137-140. Zagaeski, M. (1987) Some observations on the prey stunning hypothesis. Mar. Mamm. Sci. 3(3): 275-279. Zhou, K. (1982) Classification and phylogeny of the superfamily Platanistoidea, with notes on evidence for the monophyly of Cetacea. Sci. Rep. Whales Res. Instit. 34:93-108.
Index
Alligator, see Crocodilians Amphisbaenians, see also Lepidosauria diet, 180-181, 262-263 evolution of feeding, 262-263 feeding function, 261-262 Anatomy morphology comparison, 6-7 terminology, 23-25 Anguimorpha, see Lepidosauria Ant-eating animals, see Lepidosauria; Myrmecophagous mammals Anurans, see Frog Auditory system caecilians, 151-152,154-155 lepidosaur ians, 188-189 6 Baleen comparison with teeth, 495-496 composition, 494 food removal mechanism, 501 overview, 37 structure, 37,494-495 Beak, structure and function, 36-37, 375, 395,396,400-403,405 Bill, see Beak
Biomechanics aquatic salamanders, 85-87 overview, 11 Birds, see also Neognathous birds; Paleognathous birds foraging theory, 395 function inference from structure versus tests of hypotheses, 402404 prospects for feeding studies, 396, 406 surface tension transport in phalotropes, 395,400-402 systematics and choice of taxa for feeding studies delineation of studies, 397 excluded orders and families, 397399 performance testing, 401-402 phenotypic transformations of feeding systems resulting in probing and filter feeding, 400-402 phylogenetic methods, applications, 399-400 variability of feeding structures individual variability within species, 405-406
527
interspecific variation, 395, 406 sample sizes required for studies, 404-405 statistical analysis, 404-405 Braincase, anatomy, 29 Bramble-Wake model, gape cycle, 48-49,268-269 Branchiometric musculature innervation, 40-41 muscle types, 41-42 origin, 39-41
Caecilians diet, 150 evolution feeding apparatus, 162-163 fossil record, 161 oviparous ancestor, 162 fetal feeding, 155 geographic distribution, 149 intraoral transport, 160-161 morphology auditory system, 151-152 eye, 150-151 lateral line, 151
528 Caecilians {continued) olfactory system, 151 tentacle organ, 151 musculoskeletal anatomy adult group 1,153 group II, 153 group III, 153-154 group IV, 154 group V, 154 overview, 153 fetus, 153 larva, 152-153 natural history, 149-150 phylogenetic relationships among families, 149-150 prey capture and ingestion gape cycle, 155,159-160 jaw prehension, 155-156,159-160 suction feeding, 155 prey finding auditory signals, 154-155 behavioral studies, 155 lateral line role, 155 olfaction, 155 vision, 154 prospects for feeding studies, 163-164 Carnivore abundance in ecosystem, 22 mammals, 413, 414, 425,426, 427, 428, 441 Cartilago transiliens, crocodilians, 341 Central pattern generator (CPG), feeding movement control frogs, prey capture role, 135-136 mammals, 416, 429, 444 overview, 47,49 Chameleon, see also Lepidosauria diet, 258 evolution, 261 hyolingual apparatus specializations, 258-259 ingestion kinematics, 259 lingual prehension mechanism, 260261 natural history, 257-258 skull and mandible morphology, 193-195 tongue projection mechanism, 259260 Cheeks feeding functions, 46-47 mammalian features, 430 Chewing, see Processing Chondrocranium, origin and development, 27 Cinematography, morphology studies, 15 Constructional morphology, overview, 12-13
Index Cranial kinesis birds, 403 caecilians, 162 Lepidosauria morphological basis, 199-201 functional significance mesokinesis, 256-257 streptostyly, 255-256 Komodo monitor, 264 overview, 33-34 Cranium, see also Braincase, frog morphology, 119-120 Crocodilians diet and development, 347-348 evolution diet in relation to skull morphology, 354-355 skull morphology in relation to bauplan of jaw adductors and cervical musculature, 355-357 feeding stages, 348 inertial feeding hyolingual apparatus role, 351 kinematics, 348-350 overview, 340 kinematics of feeding gape profile, 349 inertial bites duration, 349 phases, 349-350 repositioning and killing/crushing bites, 350 transport bites, 350 overview, 348-349 swallowing, 350 morphology cartilago transiliens, 341 cervical region musculature, 345-347 osteology, 345 dentition, 341 hyobranchial apparatus and musculature, 342-345 jaw joint, 341 jaw muscles, 341-343 skull, 340 tongue, 44, 343-344 motor patterns in feeding cervical musculature, 353-354 jaw musculature, 351-353 phylogenetic relationships Crocodylia to other vertebrates, 337-339 eusuchian crocodilians, 339-340
Descriptive morphology, overview, 9-10 Developmental morphology, overview, 11-12 Digital Particle Image Velocimetry, morphology studies, 16 Dissection, morphology studies, 14 Drinking, transition from suckling at weaning, 455-456 Dugong, see Marine mammals
D Dentition, see Teeth Dermatocranium, origin and development, 29
Facial skeleton, anatomy, 30-31 Fang, see Snake Feeding stages, see also individual taxa and stages
Ecological morphology, overview, 12 Ectotherm digestion rates, 23 evolution of feeding systems, 22-23 thermal regulation, 22 Electromyography crocodilians, motor patterns in feeding cervical musculature, 353-354 jaw musculature, 351-353 frog feeding studies, 127-129 Lepidosauria feeding, 228-229, 234235, 238,242,245, 247 lingual feeding in salamanders, 108 mammalian feeding, 426-429,432433,440 overview of studies, 15 snake feeding envenomation studies, 309-310 intraoral transport studies, 313, 317-318 Electron microscopy, morphology studies, 14-15,416 Elephant trunk, structure and function, 47 Emu, see Paleognathous birds Endotherm digestion rates, 23 evolution of feeding systems, 22-23 thermal regulation, 22 Envenomation, snakes fang structure, 309-310 metering, 309 musculature, 309-310 rear-f anged systems, 309 - 310 venom gland, 309 Evolutionarily stable configuration, Lepidosauria, 274-276 Evolutionary morphology, overview, 10 Eye, see Vision
Index birds ingestion, 50-51 intraoral transport, 375-384, 388 processing, 53 swallowing, 54 crocodilians, 348 caecilians ingestion, 155-160 intraoral transport, 160-161 frogs ingestion, 127-136 swallowing, 54 ingestion, 49-52 intraoral transport, 52-53 Lepidosauria Iguania ingestion, 230-235 intraoral transport, 236-238 processing, 235-236 swallowing, 238-240 ingestion, 50-51, 222-224 intraoral transport, 52, 225-226 overview, 220-222 prey capture, 222-224 processing, 224-225 Scleroglossa ingestion, 240-243 intraoral transport, 246-248 processing, 243-246 swallowing, 248-249 snakes ingestion, 299-300 intraoral transport, 52, 300-301 prey capture, 299-300 prey restraint, 301 swallowing, 54, 300-301 Sphenodon ingestion, 227-229 intraoral transport, 229-230 processing, 229 swallowing, 230 swallowing pharyngeal compression, 226227, 230, 240, 249 pharyngeal packing, 226-227, 230,238-240,248-249 mammals acquisition of food, 441 ingestion, 50-51, 441 manipulation of food, 442-443 mechanical hypothesis, 417 myrmecophagous mammals intraoral transport, 477-478 lingual ingestion, 477 prey processing, 477 swallowing, 478 overview, 416-418,441 sensorimotor hypothesis, 417-418 stage I transport, 52,441-442 stage II transport, 52,417-418, 443
swallowing, 54-55,418-419,443444 overview, 49-50 processing, 53 salamanders ingestion, 82-84,102-109 intraoral transport, 84,109-110 prey processing, 84,109-110 swallowing, 53-55 Filter feeding, marine mammals Balaenopteridae morphology, 497-499 baleen comparison with teeth, 495-496 composition, 494 food removal mechanism, 501 structure, 37,494-495 coloration in whales, 501 mysticete feeding methods, 494-501 seals, 501-502 water flow dynamics, 496-497 whales, 493-501 Fish feeding, rationale for study exclusion, 4 - 5 Form-function relationship evolution, rationale for study, 3 - 4 historical perspective, 7-8 morphological data, 8 phenotype, 8 Frog kinematic studies of feeding apparatus electromyography, 127-129 kinematic profiles, 126-127 muscle stimulation experiments, 127,129 overview, 124,126 trajectory profiles, 126 morphology of feeding apparatus buccal floor muscles, 122 cranium, 119-120 hyobranchial apparatus and musculature, 122-123 jaw muscles, 120-121 mandible, 121-122 palate, 120 teeth, 120 tongue comparison with other vertebrates, 119 correlates of protraction mechanisms, 140-141 musculature, 123-124 natural history, 118-119 neural control of prey capture central pattern generators, 135136 evolutionary transitions in mechanisms, 143 overview, 135-136 tongue afferents
529 evolution, 141-143 role in prey capture, 137-138 visual input interactions, 138 visual analysis of prey, 136-137 phylogeny of families, 118 prospects for feeding studies, 144 tongue protraction mechanisms evolutionary transitions, 139, 143-144 hydrostatic elongation, 134135 inertial elongation, 131-134 mechanical pulling, 130-131 morphological correlates, 140141 models Emerson hyoid model, 127-129 Gans and Gorniak ballista model, 127-129 Functional morphology, overview, 10-11
Gape cycle Bramble-Wake model, 4 8 - 4 9 , 2 6 8 269 caecilians, 155,159-160 crocodilian profiles, 349 definition, 47 ingestion, 51-52 Lepidosauria, 222,228,234,236,242, 247,268-269 mammals, 439-440 phases, 48 profiles, 47 Gastrolith, marine mammal food grinding, 492 Gekko, see Lepidosauria Gill slits, salamander morphology adult, 75 functional morphology, 85 larva, 70 Goniometry, morphology studies, 15-16,278 Grazing, marine mammals desmostylians, 520-521 dugong, 517-518 lips, 518-519 manatee, 516-518 overview, 516 sea cow, 519 taxonomy and evolution, 519-520 teeth, 517 Gustation Lepidosauria, 186-187 mammals, 433,435 overview, 45 salamander, aquatic feeding, 81
530 H Head, see also skull salamander morphology adult, 75 larva, 69 snake head size correlation with prey size, 322 whale profiles, 506-507 Herbivore, abundance in ecosystem, 22 Hox genes expression in snakes, 321 skull patterning, 26 Hyobranchial apparatus aquatic tetrapods, 38-39 caecilians, 153,159 crocodilian morphology and musculature, 342-345 frog morphology and musculature, 122-123 functions, 37-39 Lepidosauria, morphology musculature, 205 skeleton, 203-205 tongue relationship, 219-220 mammals, 39 morphology, 39 musculature, 42-43 myrmecophagous mammal morphology, 466 salamander morphology adult, 76-78 larva, 70-72 terrestrial feeding, 97, 99-101 terminology, 39 Hyoid definition, 38 mammalian structure and function, 431-433 Hyolingual apparatus chameleon specializations, 258-259 neognathous birds evolution, 360 functional overview, 373-375 hyobranchial elements, 362 musculature, 361,363-365 paraglossal development, 362-363 summary of morphology, 365-366 tongue morphology, 362, 365 paleognathous birds basihyal, 366-367 emu features, 370-371 evolution, 360 functional overview, 373-375 greater rhea features, 366-369 musculature, 361-363, 368-373 ostrich features, 371 summary of morphology, 372-373 timanou features, 372 tongue morphology, 366, 372-373
Index primitive condition of neornithine hyolingual apparatus musculature, 386-387 ossification, 385 position and orientation, 386 tongue size, 387
I Iguania, see Lepidosauria Image analysis, morphology studies, 15 Inertial feeding caecilians, 161 crocodilians hyolingual apparatus role, 351 inertial bites duration, 349 phases, 349-350 repositioning and killing/crushing bites, 350 transport bites, 350 kinematics, 348-350 overview, 340 lepidosaurs, inertial transport, 225, 229,238,247-248,264 overview, 52 paleognathous birds, cranioinertial feeding hyobranchial mechanism, 379, 381-382 jaw-hyolingual coordination, 379 kinematics of food transport, 379 lingual feeding comparison, 382, 384 Ingestion, see also Prey capture caecilians, 155-160 chameleon kinematics, 259 frogs, 127-136 gape cycle, 51-52 jaw prehension, 50 Lepidosauria evolution, 265-267 Iguania, 230-235 overview, 222-224 Scleroglossa, 240-243 Sphenodon, 117-119 limb role, 51 lingual prehension, 50-51 myrmecophagous mammals, 477 overview, 49-52 prey prehension, 49-50 salamander, aquatic feeding jaw prehension, 83-84 stages of feeding behavior, 82 suction feeding, 82-84 tongue prehension, 83 snakes fast systems end of strikes, 306-307 jaw behavior, 305-306
preparation for strike, 307 stabbing versus biting strikes, 305 strike features, 304-305 foraging behavior, 301-302 overview, 299-300 slow systems form and function relationships, 302-304 jaw movements, 302,304 prey characteristics, 302 suction, 51 Intraoral transport caecilians, 160-161 hydrodynamic transport, 53 hyolingual transport, 52 inertial transport, 52 Lepidosauria Iguania, 236-238 overview, 225-226 Scleroglossa, 246-248 snakes electromyography, 313, 317318 lateral jaw transport, 310-311 lateromedial jaw transport, 311, 313-315 mandibular transport, 319-320 medial jaw transport, 315-319 overview, 300-301 stretching of soft tissues, 313 suction mechanisms, 320 Sphenodon, 229-230 myrmecophagous mammals, 477478 overview, 52-53 salamander, terrestrial feeding, 110 stages, 52
J Jaw anatomy, 30 crocodilian morphology joint, 341 muscles, 341-343,351-353 frog morphology, 120-122 jaw-tongue movement linkage in mammals, 439-441 Lepidosauria musculature, 205-211 groups, 205-207 histochemistry, 210-211 mammalian morphology bone, 421,423-425 cranio-mandibular joint, 424 - 425 musculature, 424, 426 musculature, overview, 39-42 myrmecophagous mammals functional movements in feeding, 475
Index morphology joint, 466 musculature, 467 prehension, see Jaw prehension salamander morphology adult, 78 larva, 70 terrestrial feeding, 97 Jaw prehension caecilians, 155-156,159-160 Lepidosauria, 241-243, 276-277 overview, 50 salamander aquatic feeding, 68, 83-84 terrestrial feeding, 109
K Komodo monitor, see also Lepidosauria cranial kinesis, 264 ingestion, 263-264 oral transport, 264 prey capture and subjugation, 263 swallowing, 264
Labial lobes, salamander morphology adult, 75 larva, 70 Lateral line caecilians, 151,155 salamander, aquatic feeding, 81 Lepidosauria, see also Amphisbaenians; Chameleon; Komodo monitor; Snake advantages for feeding studies, 175176 cranial kinesis, functional significance mesokinesis, 256-257 streptostyly, 255-256 dentition acrodont dentition, 202-203 pleurodont dentition, 201-202 replacement, 201-202 diet amphisbaenians, 180-181, 262263 evolution, 178-180 lizard size relationship, 179 snakes, 181 specialization, 181-182 Sphenodon, 180 electromyography studies of feeding, 228-229,234-235,238,242, 245,247 evolution of feeding dietary specialization ant and termite feeders, 272-273
natural history in functional analysis, 272, 274 phenotype relationship, 271272, 274 specialized diet characterization, 272,274 distribution of feeding traits across squamate phylogeny, 265 evolutionarily stable configuration, 274-276 functional units, 274-276 gape cycle, 268-269 ingestion mode, 265-267 swallowing, 267-268 tongue, 269-271 feeding stages Iguania ingestion, 230-235 intraoral transport, 236-238 processing, 235-236 swallowing, 238-240 ingestion, 222-224 intraoral transport, 225-226 overview, 220-222 prey capture, 222-224 processing, 224-225 Scleroglossa ingestion, 240-243 intraoral transport, 246-248 processing, 243-246 swallowing, 248-249 Sphenodon ingestion, 227-229 intraoral transport, 229-230 processing, 229 swallowing, 230 swallowing pharyngeal compression, 226227,230,240,249 pharyngeal packing, 226-227, 230,238-240,248-249 food location and identification, sensory basis chemoreception gustation, 186 interrelationships, 186-187 olfaction, 186 vomeronasal organs, 186-187 integration, 187-188 Sphenodon, 188 vibratory cues, 188 -189 vision, 184-185 foraging modes, 182-183 gape cycles, 222,228,234,236,242, 247 hyobranchial apparatus morphology musculature, 205 skeleton, 203-205 tongue relationship, 219-220
531 jaw musculature adductor musculature, structural and functional aspects, 209210 groups, 205-209 histochemistry, 210-211 jaw prehension, 241-243, 276-277 lingual prehension evolution, 275-277 mechanisms adhesion, 250-251 tongue protrusion, 232-233, 249-250,270 model comparative implications, 2 5 3 254 description, 251-253 scleroglossans, 254-255 overview in species, 230-231, 241, 260-261 phylogenetic relationships, 176-178 prospects for feeding studies, 277-278 skull and mandible morphology chamaeleonids, 193-195 cranial kinesis basis, 199-201 Iguania, 193-195,197 Scleroglossa, 197-199 Sphenodon, 189,192-193,195,197 squamates, 193-195,197-199 tongue morphology connective tissue organization, 216-217 musculature classification, 217 extrinsic muscles, 217-218 foretongue-hindtongue coupling, 269-270 hyobranchial apparatus relationship, 219-220 intrinsic muscles, 218-219 overview, 211, 269 superficial form, 211-213 surface morphology histochemistry, 216 papillae, 213-216 tongue types bipartite type, 270-271 chemosensory type, 271 compromise type, 270 feeding type, 270 Lingual prehension, see also Lepidosauria chameleon mechanism, 260-261 frog, 130-135 myrmecophagous mammals, 477 overview, 50-51 salamander aquatic feeding, 67-68, 83 terrestrial feeding distance, 107-108
532 Lingual prehension {continued) kinematics, 102-103 physiology, 108-109 prehension and tongue pad, 105-107 protraction, 103,105 retraction, 107 speed, 107-108 success rate, 108 Lips feeding functions, 46-47 grazing marine mammals, 518-519 Lizard, see Lepidosauria
M Macaque, suckling function and mechanics, 451-452 Mammalian feeding, see also Marine mammals; Myrmecophagous mammals; individual species adductor muscle regulation, 429 central pattern generator, 416,429, 444 cheeks and musculature, 46, 430 dietary range, 413, 419 electromyography, 426-429, 432433,440 food intake and survival, 412 gape cycle, 439-440 hyoid structure and function, 38, 39, 431-433 jaw-tongue movement linkage, 439441 mastication, functional analysis, 426-429 masticatory cycle stages, 421 mechanical properties and textures of foods, 419-421 morphology cranio-mandibular joint, 424 425 jaw bone, 421,423-425 musculature, 41, 42, 424,426 skull, 29, 32, 421-423,425 olfaction, 435 oropharyngeal complex, overview compartments, 429 functions, 429-430 overview, 46 oropharynx function, 437 species differences, 437,439 structure, 436-437 overview of feeding system, 414415 phylogenetic relationships among orders, 412
Index physiological functions of feeding system, 414-415 process model acquisition of food, 441 ingestion, 441 manipulation of food, 442-443 mechanical hypothesis, 417 overview, 416-418,441 sensorimotor hypothesis, 417-418 stage I transport, 441-442 stage II transport, 417-418,443 swallowing, 418-419,443-444 reptile-mammalian transition, 411412 rugae, 34, 37,430-431 salivary glands, 430 speech adaptation in humans, 433 study approaches, 415-416 suckling, see Suckling teeth diet and structure, 420-421 interspecies variation, 425 overview, 36 tongue base, 433-434 functions, 433 movement, 433-434,439-440 musculature and function, 44-45, 435-436 sensory organ, 434-435 Manatee, see Marine mammals Mandible, see Jaw; Skull Marine mammals, see also Mammalian feeding comparison with feeding in other mammals, 490-491 data sources for feeding research, 490 diet variability and adaptation, 521 evolution, 487-488 feeding strategies, overview and classification, 492-493 filter feeding Balaenopteridae morphology, 497-499 baleen comparison with teeth, 495496 composition, 494 food removal mechanism, 501 structure, 37, 494-495 coloration in whales, 501 mysticete feeding methods, 494501 seals, 501-502 water flow dynamics, 496-497 whales, 493-501 gastroliths and food grinding, 492
grazmg desmostylians, 520-521 dugong, 517-518 lips, 518-519 manatee, 516-518 overview, 516 sea cow, 519 taxonomy and evolution, 519-520 teeth, 517 morphology skull, 491-492 teeth, 491 phylogeny among orders, 487, 489490 raptorial feeding foraging behavior, 513 morphology, 511-512 overview, 511 polar bear, 513-516 sea otter, 514-515 snappers, 511-512 teeth, 513-514 whales, 512-513 suction feeding distribution among taxa, 502-503 evidence anatomical evidence, 504-505 ecological evidence, 505-506 head profiles in whales, 506-507 mechanisms in whales, 503,506508 morphology, 503-504,506 seals, 508 tongue role, 503-504 walruses, 508-511 taxonomy, 487-488 water consumption, 492 Mastication, see Processing Meleagris gallopavo, lingual feeding hyobranchial mechanism, 378-379 jaw-hyolingual coordination, 377 kinematics of food transport, 375, 377 Microdissection, morphology studies, 14 Microscopy, morphology studies, 14-15 Monitor, see Komodo monitor Morphology, see also Form-function relationship; individual species and structures anatomy comparison, 6-7 biomechanics, 11 constructional morphology, 12-13 descriptive morphology, 9-10 developmental morphology, 11-12 ecological morphology, 12 evolutionary morphology, 10 evolutionary theory, 6-7
Index functional morphology, 10-11 historical perspective, 5 - 6 idiographic studies, 8-9 importance of study, 3-4,16 laboratory versus field studies, 9 morphometries, 13 nomothetic studies, 9 techniques for study anatomical techniques, 14-15 functional techniques, 15-16 theory and modeling, 13-14 transformation morphology, 12 Morphometries, overview, 13 Myrmecophagous mammals, see also Mammalian feeding convergent evolution, 459 evolution of feeding specializations derived features, 480 phylogenetic pathways, 478-479 primitive features, 480 structural pathways, 479-480 feeding stages intraoral transport, 477-478 lingual ingestion, 477 prey processing, 477 swallowing, 478 foraging ecology and behavior, 4 6 3 464 functional movements in feeding jaw, 475 pharynx, 476-477 soft palate, 476-477 tongue, 475-476 geographic distribution, 460-461 morphology of feeding apparatus exceptions to myrmecophagous morphotype, 474 hyoid apparatus, 466 jaw joint, 466 jaw musculature, 467 palate, 464-465 pharynx, 473-474 skull, 464 soft palate, 473-474 teeth, 465-466 tongue external features, 467-468 extrinsic musculature, 468-472 glossal tube, 470-471 homologies of musculature, 469 hyoglossus, 472 innervation, 468-469 intermandibularis muscle, 469470 internal structure, 472-473 sternoglossus muscle, 469,471472 xiphoid process, 466-467
neonatal suckling, 478 phylogeny, 461-462,478-479 prey characteristics, 462 prospects for feeding research function, 481 morphology, 480-481 taxonomy, 460-461
N Neognathous birds hyolingual apparatus evolution, 360 functional overview, 373-375 hyobranchial elements, 362 musculature, 361,363-365 paraglossal development, 362-363 summary of morphology, 365-366 tongue morphology, 362, 365 lingual feeding comparison with ratite cranioinertial feeding, 382, 384 Meleagris gallopavo, lingual feeding hyobranchial mechanism, 378379 jaw-hyolingual coordination, 377 kinematics of food transport, 375, 377
O Olfaction caecilians, 151,155 Lepidosauria, 186,187 mammals, 47, 435 salamander aquatic feeding, 79, 81 terrestrial feeding, 101 Opossum example of ancient feeding apparatus, 413 pharynx, 437,439 suckling function and mechanics, 452-453 Ostrich, see Paleognathous birds
Palate anatomy, 31 frog morphology, 120 lepidosaurs, 189,195,196,199 myrmecophagous mammal morphology, 464-465 Paleognathous birds cranioinertial feeding hyobranchial mechanism, 379, 381-382 jaw-hyolingual coordination, 379
533 kinematics of food transport, 379 lingual feeding comparison, 382, 384 evolutionary morphology, overview, 388-389 feeding modes, 359-360 food transport mechanisms, functional evolution, 388 hyolingual apparatus basihyal, 366-367 emu features, 370-371 evolution, 360 functional overview, 373-375 greater rhea features, 366-369 musculature, 361-363, 368-373 ostrich features, 371 summary of morphology, 372-373 timanou features, 372 tongue morphology, 366,372-373 methodology for morphological and functional comparisons, 360361, 375 pattern generation, conservation in evolution, 389 phylogeny Neognathae relationship, 359, 384-385 ratite species relationships, 389390 primitive condition of neornithine hyolingual apparatus musculature, 386-387 ossification, 385 position and orientation, 386 tongue size, 387 theropod-bird transition, changes in feeding function, 387-388 Pharyngeal emptying, see Swallowing Pharynx mammals, 429-430,438-439 morphology and function, 46 myrmecophagous mammals functional movements in feeding, 476-477 morphology, 473-474 Photography, morphology studies, 15 Phylogeny, see also Form-function relationship bird feeding applications, 399-400 caecilian family relationships, 149150 chordate relationships, 21-22 crocodilians Crocodylia to other vertebrates, 337-339 eusuchian crocodilians, 339-340 frog family relationships, 118 Lepidosauria families, 176-178
534 Phylogeny (continued) feeding traits across squamate phylogeny, 265 snake family relationships, 294296,298-299 marine mammals, 487, 489-490 myrmecophagous mammals, 461462,478-479 paleognathous birds Negnathae relationship, 359, 384385 ratite species relationships, 389390 salamander aquatic feeding form and function, 91 families, 65-66,95-96, 111 terminology, 25-26 Pig, suckling function and mechanics, 540-541 Polar bear, see Marine mammals Postcranial contribution, feeding function, 5 Prey capture, see also Jaw prehension; Lingual prehension frogs, neural control central pattern generators, 135-136 evolutionary transitions in mechanisms, 143 overview, 135-136 tongue afferents evolution, 141-143 role in prey capture, 137-138 visual input interactions, 138 visual analysis of prey, 136-137 Lepidosauria, 222-224 salamander, terrestrial feeding, 102 snakes fast systems end of strikes, 306-307 jaw behavior, 305-306 preparation for strike, 307 stabbing versus biting strikes, 305 strike features, 304-305 foraging behavior, 301-302 overview, 299-300 slow systems form and function relationships, 302-304 jaw movements, 302, 304 prey characteristics, 302 Processing Lepidosauria Iguania, 235-236 overview, 224-225 Scleroglossa, 243-246 Sphenodon, 229
Index mammalian mastication cycle stages, 421 functional analysis, 426-429 mastication, definition, 53 myrmecophagous mammals, 477 overview, 53 salamander aquatic feeding, 84 terrestrial feeding, 109-110
R Radiography, morphology studies, 14, 416 Ram feeding, salamander aquatic feeding, 68 Raptorial feeding, marine mammals foraging behavior, 513 morphology, 511-512 overview, 511 polar bear, 513-516 sea otter, 514-515 snappers, 511-512 teeth, 513-514 whales, 512-513 Ratites, see Paleognathous birds Rhea, see Paleognathous birds Rugae, mammals, 34, 37,430-431,442, 450
Salamander aquatic feeding, see Salamander, aquatic feeding diet, 67 foraging, 67 natural history, 66-67, 96 origins and outgroups. 111 phylogenetic relationships, 65-66, 95-96, 111 terrestrial feeding, see Salamander, terrestrial feeding Salamander, aquatic feeding adult morphology developmental patterns, 75 gill slits, 75 head, 75 hyobranchial apparatus, 76-78 jaws, 76 labial lobes, 75 musculature of head branchiometric series, 79 hypobranchial series, 79 metamorphisis, 78-79 skull, 75-76 teeth, 34, 35, 75 tongue, 44, 75
biomechanics of jaws, 85-87 family features Amhystomatidae, 91 Amphiumidae, 90 Cryptobranchidae, 89-90 Dicamptodontidae, 90 Hynohiidae, 89 Flethodontidae, 91 Proteidae, 90 Rhyacotritonidae, 90 Salamandridae, 90 Sirenidae, 89 feeding modes jaw prehension, 68 lingual prehension, 67-68 metamorphisis changes, 87-88 ram feeding, 68 suction feeding, 67-68 functional morphology gill slit action, 85 mouth opening and closing, 85 suction feeding, 84-85 ingestion behavior and kinematics jaw prehension, 83-84 stages of feeding behavior, 82 suction feeding, 82-84 tongue prehension, 83 larval morphology gill slits, 70 head, 69 hyobranchial apparatus, 70-72 jaws, 70 labial lobes, 70 musculature functional groups, 72 head movement, 74 hyobranchial muscles, 74 mouth opening and closing, 72-73 suction, 73-74 skull, 70 teeth, 69 tongue, 69 lateral line system, 81 motor nuclei, 81-82 olfactory systems, 79, 81 phylogenetic patterns of feeding form and function, 91 prey processing, 84 prospects for research, 92 taste buds, 81 variation interindividual variation, 88 timing of events during prey capture, 88 vision, 79 Salamander, terrestrial feeding cranial nerves, 101
Index evolution ecology and selective regime, 113114 homoplasy, 114 family features Ambystomatidae, 111 Dicamptodontidae, 111 Hynohiidae, 111-112 Plethodontidae, 113 Rhyacotritonidae, 111 Salamandridae, 112-113 feeding modes, 96 foraging, 101-102 jaw prehension, 109 lingual feeding distance, 107-108 kinematics, 102-103 physiology, 108-109 prehension and tongue pad, 105107 protraction, 103,105 retraction, 107 speed, 107-108 success rate, 108 modulation of behavior, 110-111 morphology hyobranchial apparatus, 97, 9 9 101 jaws, 97 musculature, 101 overview, 97 skull, 97 teeth, 34,35,97 tongue, 44,97 olfactory system, 101 prey processing immobilization, 109 -110 intraoral transport, 110 prey capture, 102 detection and localization, 102 prospects for research, 114 vision, 101 Salivary glands, mammals, 430 Scincomorpha, see Lepidosauria Sclerogossa, see Lepidosauria Sea cow, see Marine mammals Seal, see Marine mammals Sea otter, see Marine mammals Skull, see also Cranial kinesis anatomical units, see also individual units braincase, 29 facial skeleton, 30-31 jaws, 30 palate, 31 caecilian morphology, 152-154 chameleon morphology, 193-195
crocodilian morphology, 340,354357 Hox genes in patterning, 26 Lepidosauria morphology chamaeleonids, 193-195 cranial kinesis basis, 199-201 Iguania, 193-195,197 Scleroglossa, 197-199 snakes, 294-299,314 Sphenodon, 189,192-193,195,197 squamates, 193-195,197-199 mammalian morphology, 421-423, 425 marine mammal morphology, 491492 myrmecophagous mammal morphology, 464 origin and development chondrocranium, 27 dermatocranium, 29 splanchocranium, 27-29 salamander morphology adult, 75-76 larva, 70 terrestrial feeding, 97 temporal fenestrae function, 32-33 patterns, 31-32 Snake, see also Lepidosauria defecation, 321-322 diet, 181,293,302 digestion, 321 evolution ecological patterns, 322-325 feeding system, 264,267,294-296, 298-299,325-326 feeding stages, overview ingestion, 299-300 intraoral transport, 300-301 prey capture, 299-300 prey restraint, 301 swallowing, 300-301 head size correlation with prey size, 322 intraoral transport electromyography, 313, 317-318 lateral jaw transport, 310-311 lateromedial jaw transport, 311, 313-315 mandibular transport, 319-320 medial jaw transport, 315-319 stretching of soft tissues, 313 suction mechanisms, 320 morphological features Acrochordidae, 298 Alethinophidia, 295-296 Atractaspidae, 299 Boidae, 298
535 Colubridae, 298-299 Loxocemidae, 295-296 Macrostomata, 296,298 Pythonidae, 298 Scolecophidia, 295 Tropidophiidae, 298 Viperidae, 299 Xenopeltidae, 295 phylogenetic relationships among families, 294-296,298-299 prey capture and ingestion fast systems end of strikes, 306-307 jaw behavior, 305-306 preparation for strike, 307 stabbing versus biting strikes, 305 strike features, 304-305 foraging behavior, 301-302 slow systems form and function relationships, 302-304 jaw movements, 302,304 prey characteristics, 302 prey classification, 323-324 prey manipulation movements, 310 prey restraint blunt trauma and mechanical reduction, 307-308 constriction, 308-309 envenomation fang structure, 309-310 metering, 309 musculature, 309-310 rear-fanged systems, 309-310 venom gland, 309 prospects for feeding studies, 326-327 strike versus lunge, 301 swallowing form and function relationships, 320-321 tongue role, 320 Snout, feeding functions, 47 Soft palate, mammals functional movements in feeding, 443,450-453,454,476-477 morphology, 429,437,438,439,473474 Speech, adaptation in humans, 433 Sphenodon, see Lepidosauria Splanchocranium, origin and development, 27-29 Suckling comparison with adult feeding, 449 coordination of swallowing and respiration, 455 evolutionary considerations, 456 function and mechanics
536 Suckling {continued) macaque, 451-452 opossum, 452-453 overview, 450 pig, 540-541 morphology, 449-450 myrmecophagous mammals, 478 rhythmicity and control, 453 transition to drinking at weaning, 455-456 Suction feeding caecilians, 155 marine mammals distribution among taxa, 502-503 evidence anatomical evidence, 504-505 ecological evidence, 505-506 head profiles in whales, 506-507 mechanisms in whales, 503, 506508 morphology, 503-504, 506 seals, 508 tongue role, 503-504 walruses, 508-511 overview, 4 - 5 , 51, 53 salamander aquatic feeding functional morphology, 84-85 kinematics, 82-84 overview, 67-68 Swallowing amphibians, 54 coordination with respiration in suckling, 455 crocodilians, 350 Lepidosauria evolution, 267-268 Iguania, 238-240 pharyngeal compression, 226-227, 230, 240, 249 pharyngeal packing, 226-227, 230, 238-240, 248-249 Scleroglossa, 248-249 snakes form and function relationships, 320-321 overview, 300-301 tongue role, 320 Sphenodon, 230 mammals, 437-439,443-444 myrmecophagous mammals, 478 overview, 53-55 pharyngeal emptying, overview, 54-55 reptiles, 54-55 terminology, 53-54
Teeth crocodilian morphology, 341
Index frog morphology, 120 function and structure, 36 grazing marine mammals, 513-514 keratinized structures as functional replacements, 36-37 Lepidosauria acrodont dentition, 202-203 pleurodont dentition, 201-202 replacement, 201-202 snake envenomation fang structure, 309-310 metering, 309 musculature, 309-310 rear-fanged systems, 309-310 venom gland, 309 mammals, overview diet and structure, 420-421 interspecies variation, 425 marine mammal morphology, 491 myrmecophagous mammal morphology, 465-466 origin, 34 palatal structures, 34 prehension, 50-51 raptorial feeding marine mammals, 513-514 replacement, 25 salamander morphology adult, 75 larva, 69 terrestrial feeding, 97 tissues, 34 types, 34 Temporal fenestrae function, 32-33 patterns, 31-32 Tentacle organ, caecilians, 151 Terminology anatomy, 23-25 phylogeny, 25-26 Termite-eating animals, see Lepidosauria; Myrmecophagous mammals Thermal physiology, see Ectotherm; Endotherm Timanou, see Paleognathous birds Tongue crocodilian morphology, 343-344 frogs afferents evolution, 141-143 role in prey capture, 137-138 visual input interactions, 138 morphology comparison with other vertebrates, 119 correlates of protraction mechanisms, 140-141 musculature, 123-124 functions, 43, 45 keratinous structures, 36-37
Lepidosauria morphology connective tissue organization, 216-217 evolution, 269-271 musculature classification, 217 extrinsic muscles, 217-218 foretongue-hindtongue coupling, 269-270 hyobranchial apparatus relationship, 219-220 intrinsic muscles, 218-219 overview, 211,269 superficial form, 211-213 surface morphology histochemistry, 216 papillae, 213-216 types bipartite type, 270-271 chemosensory type, 271 compromise type, 270 feeding type, 270 mammals base, 433-434 functions, 433 jaw-tongue movement linkage, 439-441 movement, 433-434,439-440 musculature and function, 435436 sensory organ, 434-435 suckling, 450-454,454-455 musculature, 44-45 myrmecophagous mammals functional movements in feeding, 475-476 morphology external features, 467-468 extrinsic musculature, 468-472 glossal tube, 470-471 homologies of musculature, 469 hyoglossus, 472 innervation, 468-469 intermandibularis muscle, 469470 internal structure, 472-473 sternoglossus muscle, 469,471472 prehension, see Lingual prehension protraction, see Tongue protraction salamander morphology adult, 75 larva, 69 terrestrial feeding, 97 structural diversity, 43-44 suction feeding in marine mammals, 503-504 Tongue protraction chameleon mechanism, 259-260
Index frogs mechanisms evolutionary transitions, 139, 143-144 hydrostatic elongation, 134-135 inertial elongation, 131-134 mechanical pulling, 130-131 morphological correlates, 140141 models Emerson hyoid model, 127129 Gans and Gorniak ballista model, 127-129 Lepidosauria mechanisms, 232-233, 249-250,270 salamander, terrestrial feeding, 103, 105 Tooth, see Teeth
537
Transformation morphology, overview, 12 Turkey, see Neognathous birds Turtle, bibliography of feeding references, 169-171 beak, 36
Lepidosauria, 184-185 salamanders aquatic feeding, 79 terrestrial feeding, 101 Vomeronasal organs, Lepidosauria, 186-187
V Vertebrate morphology, see Morphology Vibratory cues, Lepidosauria prey capture, 188-189 Videography, morphology studies, 15, 416 Vision caecilian eye morphology, 150-151, 154 frogs, analysis of prey, 136-137
W Walrus, see Marine mammals Water consumption, marine mammals, 492 Whale, see Marine mammals
Xiphoid process, myrmecophagous mammals, 466-467
7
This Page Intentionally Left Blank
,
Tooth, see Teeth
frogs, analysis of prey, 136-137
