Advances in the Study of Behavior, Volume 34 (Advances in the Study of Behavior)

Advances in

THE STUDY OF BEHAVIOR VOLUME 34

Advances in THE STUDY OF BEHAVIOR Edited by

Peter J. B. Slater Jay S. Rosenblatt Timothy J. Roper Charles T. Snowdon H. Jane Brockmann Marc Naguib

Advances in THE STUDY OF BEHAVIOR Edited by Peter J. B. Slater School of Biology University of St. Andrews Fife, United Kingdom

Jay S. Rosenblatt

Timothy J. Roper

Institute of Animal Behavior Rutgers University Newark, New Jersey

Department of Biology and Environmental Science University of Sussex Sussex, United Kingdom

Charles T. Snowdon

H. Jane Brockmann

Department of Psychology University of Wisconsin Madison, Wisconsin

Department of Zoology University of Florida Gainesville, Florida

Marc Naguib Department of Animal Behavior University of Bielefeld Bielefeld, Germany

VOLUME 34

Elsevier Academic Press 525 B Street, Suite 1900, San Diego, California 92101-4495, USA 84 Theobald’s Road, London WC1X 8RR, UK

This book is printed on acid-free paper. Copyright ß 2004, Elsevier Inc. All Rights 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. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (www.copyright.com), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2004 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-3454/2004 $35.00 Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, E-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting ‘‘Customer Support’’ and then ‘‘Obtaining Permissions.’’ For all information on all Elsevier Academic Press publications visit our Web site at www.books.elsevier.com ISBN: 0-12-004534-6 PRINTED IN THE UNITED STATES OF AMERICA 04 05 06 07 08 9 8 7 6 5 4 3 2 1

Contents

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix xi

Reproductive Conflict in Insect Societies ¨ RGEN HEINZE JU I. II. III. IV. V. VI.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Causes of Conflict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Conflict and Mechanisms of Control . . . . . . . . . . The Behavioral Side of Conflict Resolution . . . . . . . . . . . . Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 7 35 39 43 43

Game Structures in Mutualistic Interactions: What Can the Evidence Tell Us About the Kind of Models We Need? REDOUAN BSHARY AND JUDITH L. BRONSTEIN I. II. III. IV. V. VI. VII. VIII. IX. X. XI.

The Puzzle of Cooperative Behavior . . . . . . . . . . . . . . . . . . Game Theoretical Approaches to Mutualism . . . . . . . . . . . Goals of This Article . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Parameters Considered for the Assessment of Game Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of the Literature . . . . . . . . . . . . . . . . . . . . . . . . . How Similar/Different Are Mutualisms?. . . . . . . . . . . . . . . The Importance of Ecology . . . . . . . . . . . . . . . . . . . . . . . . . Future Avenues with Respect to Evaluation of Game Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

59 60 63 64 65 68 86 92 95 96 96 97

vi

CONTENTS

Neurobehavioral Development of Infant Learning and Memory: Implications for Infant Attachment TANIA L. ROTH, DONALD A. WILSON, AND REGINA M. SULLIVAN I. II. III. IV.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unique Characteristics of Infant Learning . . . . . . . . . . . . . . Early Experiences Affect Brain and Behavior . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

103 104 121 122 123

Evolutionary Significance of Sexual Cannibalism MARK A. ELGAR AND JUTTA M. SCHNEIDER I. II. III. IV. V. VI. VII. VIII. IX. X.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Natural History and Taxonomic Distribution . . . . . . . . . . . The Timing of Sexual Cannibalism . . . . . . . . . . . . . . . . . . . . Natural Selection of Sexual Cannibalism . . . . . . . . . . . . . . . Sexual Selection and Sexual Cannibalism . . . . . . . . . . . . . . Postinsemination Sexual Cannibalism, Self-Sacrifice, and Monogyny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sexual Conflict over Mating Rate and the Duration of Copulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sexual Cannibalism and Male Mate Choice . . . . . . . . . . . . Sexual Cannibalism and Sexual Size Dimorphism . . . . . . . Outlook and Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

135 136 139 140 146 151 153 155 156 157 159

Social Modulation of Androgens in Vertebrates: Mechanisms and Function RUI F. OLIVEIRA I. II. III. IV.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Androgens as Causal Agents of Behavior . . . . . . . . . . . . . . Behavioral Feedback on Endocrine Function . . . . . . . . . . . Proximate Mechanisms for the Social Modulation of Androgens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Ontogeny of the Social Modulation of Androgens . . . . . . . VI. Adaptive Significance of Social Modulation of Androgens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Evolution of the Social Modulation of Androgens . . . . . . .

165 166 175 188 195 197 206

CONTENTS

VIII. Social Modulation of Androgens in Men . . . . . . . . . . . . . . IX. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii 216 218 219

Odor Processing in Honeybees: Is the Whole Equal to, More Than, or Different from the Sum of Its Parts? HARALD LACHNIT, MARTIN GIURFA, AND RANDOLF MENZEL I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Pavlovian Conditioning and Models of Compound Stimulus Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Olfactory Pavlovian Conditioning and Olfactory Compound Stimulus Processing in the Honeybee . . . . . . . IV. Physiological Correlates of Odor Processing and Element/ Compound Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Functional Model of the Olfactory System as a Neural Substrate for Elemental and Compound Processing . . . . . VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

241 242 245 250 255 259 260 261

Begging, Stealing, and Offering: Food Transfer in Nonhuman Primates GILLIAN R. BROWN, ROSAMUNDE E. A. ALMOND, AND YFKE VAN BERGEN I. II. III. IV.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adult–Adult Food Transfer . . . . . . . . . . . . . . . . . . . . . . . . . Food Transfer to Infants from Parents and Helpers . . . . . Does Food Transfer Influence Infant Growth and/or Survival? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Functional Explanations of Infant Food Transfer . . . . . . . VI. Information Donation and Teaching . . . . . . . . . . . . . . . . . . VII. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

265 266 270 275 277 285 286 287

viii

CONTENTS

Song Syntax in Bengalese Finches: Proximate and Ultimate Analyses KAZUO OKANOYA I. II. III. IV. V. VI. VII. VIII.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Origin of Bengalese Finches . . . . . . . . . . . . . . . . . . . . . . . . . Analyses of Bengalese Finch Songs . . . . . . . . . . . . . . . . . . . Tinbergen’s Four Questions. . . . . . . . . . . . . . . . . . . . . . . . . . Further Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Remaining Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scenario for the Evolution of Complex Syntax . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

297 302 304 309 334 338 339 340 341

Behavioral, Ecological, and Physiological Determinants of the Activity Patterns of Bees P. G. WILLMER AND G. N. STONE I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Intrinsic Factors Affecting Bee Activity Patterns . . . . . . . . III. Sexual Differences Affecting Bee Activity Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Extrinsic Factors Structuring Bee Activity . . . . . . . . . . . . . . V. Special Effects of Sociality on Bee Activity Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Summary: Why Bee Activity Patterns Matter . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

347 352

432 445 446

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

467

Contents of Previous Volumes . . . . . . . . . . . . . . . . . . . . . . .

497

369 392

Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.

ROSAMUNDE E. A. ALMOND (265), Department of Psychology, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA JUDITH L. BRONSTEIN (59), Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721, USA REDOUAN BSHARY (59), Institut de Zoologie, Universite´ de Neuchaˆtel, Neuchaˆtel, Switzerland YFKE VAN BERGEN (265), Sub-Department of Animal Behaviour, University of Cambridge, Cambridge CB3 8AA, United Kingdom GILLIAN R. BROWN (265), School of Psychology, University of St. Andrews, Fife KY16 9JU, United Kingdom and Sub-Department of Animal Behavior, University of Cambridge, Cambridge CB3 8AA, United Kingdom MARK A. ELGAR (135), Department of Zoology, University of Melbourne, 3010 Victoria, Australia MARTIN GIURFA (241), Centre de Recherches sur la Cognition Animale, CNRS, Paul-Sabatier-University, UMR 5169, 31062 Toulouse Cedex 4, France ¨ RGEN HEINZE (1), Biologie I, Universita¨t Regensburg, 93053 JU Regensburg, Germany HARALD LACHNIT (241), Department of Psychology, PhilippsUniversity Marburg, 35032 Marburg, Germany RANDOLF MENZEL (241), Department of Neurobiology, Free University of Berlin, 14195 Berlin, Germany RUI F. OLIVEIRA (165), Instituto Superior de Psicologia Aplicada, 1149-041 Lisbon, Portugal

ix

x

CONTRIBUTORS

KAZUO OKANOYA (297), Faculty of Letters, Chiba University, Chiba 263-8522, Japan and Precursory Research for Embryonic Science and Technology (PRESTO), Japan Science and Technology Agency, Tokyo 102-8666, Japan TANIA L. ROTH (103), Department of Zoology, University of Oklahoma, Norman, Oklahoma 73019, USA JUTTA M. SCHNEIDER (135), Institute of Evolutionary Biology and Ecology, University of Bonn, D-53121 Bonn, Germany REGINA M. SULLIVAN (103), Department of Zoology University of Oklahoma Norman, Oklahoma, 73019, USA G. N. STONE (347), Institute of Evolutionary Biology, University of Edinburgh, Edinburgh EH9 3JT, United Kingdom DONALD A. WILSON (103), Department of Zoology, University of Oklahoma, Norman, Oklahoma 73019, USA P. G. WILLMER (347), School of Biology, University of St. Andrews, Fife KY16 9TS, United Kingdom

Preface

The aim of Advances remains as it has been since the series began: to serve the increasing number of scientists who are engaged in the study of animal behavior by presenting their theoretical ideas and research to their colleagues and to those in neighboring fields. We hope that the series will continue its ‘‘contribution to the development of cooperation and communication among scientists in our field’’, as its intended role was phrased in the Preface to the first volume in 1965. Since that time, traditional areas of animal behavior research have achieved new vigor by the links they have formed with related fields and by the closer relationship that now exists between those studying animal and human subjects. Scientists studying behavior today range more widely than ever before: from ecologists and evolutionary biologists, to geneticists, endocrinologists, pharmacologists, neurobiologists and developmental psychobiologists, not forgetting the ethologists and comparative psychologists whose prime domain the subject is. It is our intention not to focus narrowly on one or a few of these fields, but to publish articles covering the best behavioral work from a broad spectrum. The skills and concepts of scientists in such diverse fields necessarily differ, making the task of developing cooperation and communication among them a difficult one. But it is one that is of great importance, and one to which the Editors and publisher of Advances in the Study of Behavior are committed. We will continue to provide the means to this end by publishing critical reviews, by inviting extended presentations of significant research programs, by encouraging the writing of theoretical syntheses and reformulations of persistent problems, and by highlighting especially penetrating research that introduces important new concepts. The present volume illustrates the breadth of our subject matter very well. Several of the chapters deal with invertebrates (particularly social insects) and show just how intricate and exquisite their behavior can be. At the functional and evolutionary level, Heinze discusses sexual conflict and Elgar and Schneider sexual cannibalism, while Lachnit et al., and Willmer and Stone take a more mechanistic approach in their discussions of odor processing and of activity patterns. As we have just passed the 40th anniversary of Niko Tinbergen’s seminal ‘‘Aims and methods’’ paper (1963. Z. Tierpsychol. 20, 410–433), it is both appropriate and gratifying that two authors, Oliveira on social effects on hormone secretion and Okanoya on bird song organization, consider their subjects in the light of xi

xii

PREFACE

the famous ‘‘four questions’’. Brown et al., on food transfer in primates and Bshary and Bronstein on game theory approaches to mutualism, concern themselves with problems that have certainly been persistent as far as functional and evolutionary interpretation is concerned. And at the other extreme, the chapter by Roth et al. is concerned with neural mechanisms of learning and memory, and how these change during infant development. What we can say from this wide spectrum of subject matter is that the study of behavior continues to advance well on a broad front. We have to report with sadness that this volume will be the last with Professor Jay S. Rosenblatt as a member of the editorial team. No one has done more to shape this series than Jay. He was initially an Associate Editor for one volume before taking over as Chief Editor upon the premature death in 1972 of the legendary Daniel S. Lehrman, the founding Editor of the series. Jay then served as Chief Editor for 13 volumes, and has continued as an Associate Editor for the subsequent 16 volumes. This is a remarkable contribution and one for which his fellow editors, the publishers and, not least, many authors have reason to be grateful. If, as we believe, the series has high standing in the field, this is above all because of the reputation established for it from the very start by Danny Lehrman and then by Jay Rosenblatt. Beyond this series Jay has made an outstanding research contribution and has trained several generations of first class scholars. Thank you Jay! At the same time, we welcome Professor H. Jane Brockmann to the editorial team: her expertise will help us to fulfill our aim that the interests of the editors, the questions that concern them, the approaches they adopt and the organisms they study, have a scope that mirrors the subject matter of the series so ensuring that its breadth is maintained.

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Reproductive Conflict in Insect Societies Ju¨rgen Heinze biologie i, universita¨t regensburg 93053 regensburg, germany

I. Introduction Being social and living in groups is often advantageous over leading a solitary life. At the same time, however, cooperative systems are susceptible to exploitation and cheating by individuals, and disagreements among group members abound. Within-group conflict has been reported from most social systems, from cooperative bacteria and social amoebae to the societies of humans and other group-living primates (Aureli and De Waal, 2000; Strassmann et al., 2000a; Velicer, 2003). Eusocial insects, certainly one of the pinnacles of social evolution, are no exception. Fighting and the formation of hierarchies among group members in the annual, more ‘‘primitively eusocial’’ societies of wasps have already been intensively studied during the first half of the twentieth century (Pardi, 1940, 1946), and aggression among members of bumblebee societies has even been reported more than 200 ago (Huber, 1802). In contrast, life in the nests of the more advanced, perennial social Hymenoptera, especially ants and honeybees, is classically regarded as harmonious and essentially free of aggression. Conflicts of interest among nestmates are often thought to be mitigated at the genetic level, because of their presupposed close relatedness (e.g., Aureli and de Waal, 2000). The societies of ants and honeybees have been likened to conflict-free, multicellular organisms (Ho¨lldobler and Wilson, 1990; Moritz and Southwick, 1992; Seeley, 1989; Wheeler, 1911). Just as the individual cells in the body of a multicellular animal or plant differentiate and efficiently take over different tasks, nestmates in the ‘‘superorganism’’ of an ant or bee colony specialize for various tasks, such as reproduction, foraging, brood care, nest defense, and maintenance. Hundreds, thousands, or even millions of nonreproductive females (workers) thus cooperate in a well-organized, complex manner, apparently with the common aim of increasing the output of 1 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

2

¨ RGEN HEINZE JU

the reproductive females, the queens (or, rarely, mated workers, also known as gamergates; Peeters and Crewe, 1984). Research on social insects has consequently focused predominantly on features of the insect society that reflect the cohesiveness of this superorganism and on properties that emerge at the level of the whole society, such as the well-ordered exploitation of food sources, joint territory defense, food flow within the nest, and cooperative brood care (e.g., Wheeler, 1910; Wilson, 1971). However, insect societies differ from organisms in that they, with rare exceptions, do not consist of genetically identical components. Individuals in insect societies may therefore have diverging interests concerning the partitioning of reproduction, the allocation of resources toward one type of offspring or the other, or whether female larvae develop into sexuals or workers. Investigations over the last 25 years, stimulated by an increased general interest in conflict and conflict resolution, have indeed demonstrated that the resulting conflict may be expressed through overt aggression among nestmates, mutual policing, punishment, ritualized domination, and egg cannibalism (Beekman and Ratnieks, 2003; Beekman et al., 2003; Bourke and Franks, 1995; Crozier and Pamilo, 1996; Heinze et al., 1994a; Pamilo et al., 1997; Ratnieks and Reeve, 1992; Sundstro¨m and Boomsma, 2001). In this article, the causes of the most important types of kin conflict are briefly delineated, followed by an examination of how these conflicts are expressed. It is not the aim of this article to give a complete and balanced overview of this whole field of research. Instead, an attempt is made to point out some of the current trends in the study of kin conflict in insect societies, focusing in particular on genetic aspects of conflict in ants and honeybees and only occasionally referring to important results from other social insects.

II. Causes of Conflict Some of the types of conflict occurring in insect societies are identical to those commonly observed in other animal groups. Many insect societies contain multiple queens (or gamergates), which can often lay more eggs than the workers can rear to adulthood. Therefore, individuals may have conflicting interests about the number of reproductives and about who produces how many offspring. This conflict resembles that commonly observed in cooperatively breeding mammals or birds. In the Hymenoptera, however, its details are shaped by the special genetic relationships

REPRODUCTIVE CONFLICT IN INSECT

3

resulting from haplodiploidy. These also give rise to a number of other types of conflict, which are without parallel in diploid organisms. Haplodiploid sex determination in Hymenoptera leads to the wellknown close genetic relationships among nestmates (Fig. 1). Diploid, fertilized eggs develop either into diploid female sexuals or workers, depending on environmental or social conditions, whereas unfertilized eggs always develop into haploid males (arrhenotokous parthenogenesis; e.g., Cook, 1993). The female offspring of a singly mated mother share all the genes inherited from their common father, but only half of the genes inherited from their mother. As a result, 75% of the genes in sisters are identical by descent (relatedness r ¼ 0.75). This is considerably more than the relatedness between full sibs in diploid animals and between a hymenopteran mother and her daughter (r ¼ 0.5) and has been seen as an important catalyst in the evolution of sterile castes in bees, wasps, and ants (Hamilton, 1964; Wilson, 1971). A female hymenopteran seems to propagate copies of

Fig. 1. Genetic relationships in a hymenopteran society with a single, singly mated queen (monogyny and monandry). The numbers illustrate the ‘‘life-for-life relatedness’’ of a worker to her sister, her brother, her own son, and the son of another worker. Life-for-life relatedness takes the sex-specific reproductive values into account and therefore is the appropriate measure when considering relationships in haplodiploid organisms (Bourke and Franks, 1995; Hamilton, 1972).

4

¨ RGEN HEINZE JU

her genes more efficiently by helping her mother to rear additional female sexual sisters than by mating and producing her own offspring (for caveats going beyond this simplified presentation see, e.g., Bourke and Franks, 1995). In many bees and ants, the majority of females have lost the ability to mate and lack a spermatheca to store sperm. They can thus no longer produce their own daughters, but by functioning as workers they increase the reproductive output of their mother, the queen, and in this way indirectly augment their own inclusive fitness. Because of haplodiploid sex determination, the relatedness of a female to her brother is only 0.25, that is, considerably less than that to her sisters. Furthermore, workers are more closely related to their own sons (r ¼ 0.5) and, in societies with a single, singly mated queen (monogyny and monandry), also to the sons of other workers, their nephews (r ¼ 0.375), than they are to their brothers. Two fundamental types of conflict result from this relatedness asymmetry. First, when they retain functioning ovaries, workers could potentially produce their own sons from unfertilized eggs instead of bringing up brothers or could at least rear worker-produced males (Bourke, 1988a; Choe, 1988; Hamilton, 1972). Workers of honeybees and most ant species have retained functioning ovaries, even though these lack a spermatheca and typically consist of fewer ovarioles than those of the queen, and they are capable of producing haploid males from unfertilized eggs. As the queen is more closely related to her own sons (r ¼ 0.5) than to her grandsons (i.e., the sons of workers, r ¼ 0.25), a potential conflict results between queen and workers about who produces the males in the society. Second, whereas the queen is equally related to her female and male sexual offspring and thus is selected to produce a 50:50 sex ratio, workers benefit from their ‘‘altruistic’’ help only if they invest more in female than male sexuals (Trivers and Hare, 1976). They might attempt to increase their inclusive fitness by manipulating the sex ratio in favor of female sexuals. Thus, potential conflict exists even in hymenopteran societies with the simplest pedigree and with a high relatedness among nestmates of 0.75. One might therefore expect that conflict becomes even more acute when additional parties are introduced, that is, when the queen mates with several partners (polyandry) or when multiple female reproductives lay eggs (polygyny; in social insect research polygyny traditionally means the presence of multiple female reproductives per colony and not a single male mating with several females; see also Choe, 1995). Instead, however, polyandry and polygyny to some extent alleviate the queen–worker conflict about sex ratio and male origin, while at the same time introducing conflict about other issues, such as the partitioning of reproduction among

REPRODUCTIVE CONFLICT IN INSECT

5

queens or the allocation of resources toward different genetic lineages (Bourke and Franks, 1995; Ratnieks and Reeve, 1992). The genetic relationships between an individual worker and the males produced by different egg layers in the society change when the queen is multiply mated, such as in honeybees, Apis mellifera, and related species (Fig. 2). With increasing mating frequency of the queen, more and a more of a worker’s nestmates are her half-sibs, sharing the same mother but having a different father (r ¼ 0.25). Workers are less closely related to the males produced by their half-sibs (r ¼ 0.125) than they are to queen-produced males. If workers cannot distinguish between sons of full-sibs and sons of half-sibs, they are expected to favor queen-produced over worker-produced males at effective mating frequencies of more than 2. As each worker is still most closely related to her own sons, a conflict results about who produces the males between each individual worker and the rest of the colony. A similar logic applies when a colony contains several related queens. When workers cannot distinguish between the males produced by sisters and other workers, they are again expected to favor queen production of males (Pamilo, 1991a). Note that worker interests are not different from the monogynous situation when nestmate queens are unrelated to each

Fig. 2. Genetic relationships in a hymenopteran society with a single, double-mated queen (monogyny and polyandry). The numbers illustrate the life-for-life relatedness of a worker to her sister, her half-sister, her brother, her own son, and the son of a half-sister. As in Fig. 1, the worker is still most closely related to her own son but is more closely related to her brother than to males produced by her half-sister.

6

¨ RGEN HEINZE JU

other. This case, however, appears to be relatively rare among social insects (but see Heinze and Keller, 2000). Polyandry affects the sex ratio optimum of workers in a similar way: with increasing mating frequency, the average relatedness of a worker to female sexuals declines toward the relatedness to male sexuals. The sex ratio optima of workers and the queen may therefore converge at high mating frequencies (e.g., Bourke and Franks, 1995; Pamilo, 1991b). The situation becomes more complicated in polygynous colonies with related queens, where the sex ratio optimum of workers depends on queen number, queen relatedness, and the fitness of dispersing queens (Bourke and Franks, 1995; Pamilo, 1990). When queen mating frequencies or the number of queens per nest vary within a population, worker control of sex allocation may lead to split sex ratios. In short, colonies with lower than average queen mating frequencies or queen numbers will produce female sexuals, and colonies with higher than average queen mating frequencies and queen numbers will produce male sexuals (Boomsma and Grafen, 1990, 1991). For more accurate and more detailed reviews on sex ratio theory see, for example, Bourke and Franks (1995), Crozier and Pamilo (1996), and Mehdiabadi et al. (2003). Kin structure thus creates a multitude of potential conflicts. However, theory does not a priori predict whether and how these conflicts are realized and which party ‘‘wins,’’ that is, who has the power to enforce her reproductive optimum (e.g., Beekman and Ratnieks, 2003; Beekman et al., 2003). Instead, there are numerous constraints that limit the occurrence of conflict and affect its outcome. For example, workers of some ant species do not have any ovaries and are simply incapable of laying eggs. Although possibly adaptive in a polygynous or polyandrous ancestor, the evolutionary loss of worker ovaries might later constitute an insurmountable anatomical constraint, limiting the power of workers to enforce their interests. Workers might also lack the information necessary to enforce their interests. To replace haploid queen-laid eggs by haploid worker-laid eggs requires that workers can reliably distinguish between diploid and haploid queen-laid eggs, because sacrificing a female sexual sister in favor of a son or nephew would decrease the workers’ inclusive fitness. Similarly, to manipulate sex allocation, workers must be able to determine the queen’s mating frequency and to distinguish between male and female offspring of the queen. The larger the probability of making a mistake, the less likely manipulation is to evolve. Queens may actively camouflage the sex of their own offspring at least during the early larval stages (Nonacs, 1993; Nonacs and Carlin, 1990). Although the lack of clear differentiation between male and female pupae in the ant Camponotus floridanus was taken as evidence

REPRODUCTIVE CONFLICT IN INSECT

7

of sexual deception (Nonacs and Carlin, 1990), studies of other species suggest that workers can distinguish between the sexes. For example, in the ant Lasius niger, workers sorted the sexual larvae by sex into different parts of the nest (Jemielity and Keller, 2003). Similar arguments have been made to explain the apparent absence of nepotism in polyandrous or polygynous societies. Kin recognition might enable workers to distinguish between members belonging to different genetic lineages and to feed and groom full-sibs instead of investing in less related nestmates (Ho¨lldobler and Michener, 1980; Michener and Fletcher, 1987). Here, again, active scrambling of recognition cues or the costs of making mistakes may prevent discrimination (Ratnieks and Reeve, 1991, 1992; Reeve, 1998). Even when individuals have the information and power to enforce their own fitness interests, to do so may be associated with costs that lead to a decrease in the performance of the colony as a whole. As colonies compete with other colonies, selection at the colony level might then act against conflict. For example, it might be feasible although time-consuming to distinguish between haploid and diploid larvae or between full-sibs and half-sibs. Rather than transferring food to their nearest hungry nestmate, workers would then waste time and energy by searching for a hungry full-sib (Page et al., 1989; Ratnieks and Reeve, 1992). Similarly, fighting among workers about egg-laying rights and the division of labor might also substantially decrease colony efficiency (Cole, 1986). An increase in an individual’s direct fitness may thus be canceled out by a reduction in its indirect fitness, leading to the evolution of self-restraint (self-policing; Ratnieks, 1988). Even when worker reproduction was associated with a net increase in the reproductive worker’s fitness, other, nonreproducing workers might still suffer a decrease in their inclusive fitness. Workers may therefore actively prevent worker reproduction by attacking them or destroying their eggs (worker-policing; Monnin and Ratnieks, 2001; Ratnieks, 1988). Although such efficiency costs are generally difficult to measure, they have repeatedly been invoked to explain the absence of overt conflict or the particular outcome of conflict resolution.

III. Types of Conflict and Mechanisms of Control Research on conflict in insect societies has focused on a relatively small number of taxa, but has covered a large part of the enormous range of colony sizes and social organizations found in the social Hymenoptera. The honeybees, in particular Apis mellifera, have become the main study system for investigations on queen–worker and worker–worker conflict in

8

¨ RGEN HEINZE JU

large societies with complex genetic structure due to the queens mating with on average 5 to 30 males. At the other end of the scale of colony sizes are some paper wasps and several ants from the subfamily Ponerinae and the myrmicine tribe Formicoxenini, whose colonies often consist only of a few dozen individuals. In wasps and several Ponerinae, the caste dimorphism between nonreproductives and reproductives is not pronounced. Workers are ‘‘totipotent,’’ meaning that they have retained a spermatheca and therefore can in principle all mate and lay fertilized, diploid eggs. In several ponerines, morphologically specialized queens have been replaced completely by gamergates, that is, mated workers (Peeters, 1991, 1993). In contrast, queens in formicoxenine ants are typically considerably larger than workers and morphologically specialized for dispersal and reproduction; formicoxenine workers lay only haploid eggs. The next sections illustrate in more detail how the various types of conflict described above are expressed and with which mechanisms they are associated. The well-studied queen–worker conflict about the origin of males will serve as a starting point, because the control mechanisms mentioned in this context are of importance also in other types of conflict. A. Conflict over the Origin of Males Considering relatedness alone, queens prefer to monopolize the production of both male and female offspring in a society. Likewise, the reproductive optimum of an individual worker is to produce her own sons from unfertilized eggs. Whereas workers in monogynous, monandrous societies favor nephews over brothers, workers in a society with a more complex genetic structure (polygyny and/or polyandry) do not gain from worker reproduction and instead favor the production of brothers over that of sons of half-sibs. In addition, workers may be in conflict about who lays eggs and who engages in foraging and other costly tasks in the daily life of the colony. Workers by far outnumber the queen, and from this asymmetry in power one might expect that they are capable of enforcing their reproductive interests against those of the queen. It is therefore surprising that, with only a few exceptions, males in queenright insect societies are mostly offspring of the queen. Workers lay eggs in the presence of the queen (Bourke, 1988a; Choe, 1988; Hamilton, 1972), but these are often ‘‘trophic eggs,’’ destined to be eaten by larvae or adult nestmates (e.g., Crespi, 1991; Dietemann and Peeters, 2000; Gobin and Ito, 2000; Koedam et al., 1996). Using genetic markers, substantial worker reproduction has been documented only in a comparatively small number of species, including several stingless bees (Meliponinae; Paxton et al., 2003; To´th et al., 2002a,b) and

REPRODUCTIVE CONFLICT IN INSECT

9

a few ants [e.g., Myrmica tahoensis (Evans, 1998) and Protomognathus americanus (Foitzik and Herbers, 2001)]. For example, in accordance with predictions from kin selection theory, a large percentage of all males in the monogynous, monandrous stingless bee Paratrigona subnuda, are worker offspring. Although the queen patrols the comb and pushes workers ready to lay an egg from the brood cells, she is apparently not capable of preventing worker reproduction (To´th et al., 2002a). Another example of worker egg laying is the subterraneous ant Crematogaster smithi, in which some workers are morphologically and functionally specialized for the production of unfertilized eggs. Most of these eggs might serve as storable food in an unpredictable environment, but at least some develop into males (Heinze et al., 1999a, 2000). Despite these and other exceptions, queens appear to be mostly the winners in the queen–worker conflict about the origin of males. The next sections examine how queens keep workers from reproducing and why the latter typically do not revolt against the queen’s reproductive monopoly. 1. Queen Control Small societies with only a few dozen workers are often clearly structured by rank orders, which include the queen and at least some of the workers, whereas other workers, especially the foragers, form a class of equally subordinate individuals somewhat outside of the hierarchy. Such rank orders appear to be important in the regulation of reproduction and, as a rule, queens (or gamergates) are on top. They can punish workers who attempt to produce males by physically dominating them or by feeding on their eggs. In ants, honeybees, and vespine wasps, queens are typically larger and heavier than the workers and therefore physically dominant in a one-to-one interaction. Furthermore, in most species they are the only mated females and thus the only source of diploid offspring in the society. Whereas a queen can risk injuring or killing a worker during an escalating fight, it is normally not in the worker’s best interest to injure the queen (Trivers and Hare, 1976). Indeed, workers cooperate to attack and kill their mother queen and take over reproduction only when the queen’s reproductive output has decreased, for example, toward the end of the breeding season in the annual societies of bumblebees and wasps (e.g., Bourke, 1994a; Strassmann et al., 2003), or because of injuries or old age (Heinze, unpublished observations). In perennial species, queen killing is commonly observed in foundress associations (e.g., Choe and Perlman, 1997; Heinze, 1993; Rissing and Pollock, 1988) and in polygynous societies, when there is an overabundance of queens (reviewed by Deslippe, 2002). Queen attacks toward nestmate workers are known from colonies of bumblebees, wasps, and several ants (e.g., Fletcher and Ross, 1985; Franks

10

¨ RGEN HEINZE JU

and Scovell, 1983; Heinze et al., 1994b; Kardile and Gadagkar, 2002; Reeve, 1991; Spradbery, 1991). In addition to suppressing worker reproduction, physical aggression in small societies may also play a role in the regulation of foraging activities (e.g., Kolmer and Heinze, 2000; O’Donnell and Jeanne, 1995; Powell and Tschinkel, 1999; Premnath et al., 1995). In colonies of the boreal slave-making Formicoxenini, Protomognathus americanus and Harpagoxenus sublaevis, the queen patrols through the nest and violently antennates the head and thorax of high-ranking workers, whose ovaries have started to develop (Franks and Scovell, 1983; Heinze et al., 1994b). Workers attempt to avoid the queen, but if antennated try to escape or exhibit a submissive crouching posture, while the queen starts to bite and pull their legs and antennae. Queens may also beg food preferentially from workers with developed ovaries and in these ways perhaps deplete workers of resources needed for the completion of egg maturation. Nevertheless, genetic data reveal that queen attacks in Protomognathus americanus are not efficient in preventing worker egg laying: substantial worker reproduction was found in both queenright and queenless colonies (Foitzik and Herbers, 2001). Aggressive interactions involving the queen typically occur only when the queen’s ovaries do not yet contain many mature eggs, for example, in boreal ants during the first few weeks after hibernation. Queen–worker aggression becomes less frequent later in the season and, rather than being avoided, the queen may than be surrounded by a retinue of workers that groom and feed their mother, suggesting a switch from behavioral control to chemical signaling (e.g., Ortius and Heinze, 1999; Premnath et al., 1996; Sledge et al., 2001). Physical control is also known from several Ponerinae with totipotent workers. Here, and also in other social insects without morphologically differentiated castes, the conflict between reproductives and nonreproductives has facets that go beyond ordinary queen–worker conflict in other taxa, because reproductives not only prevent workers from producing sons but also from mating and producing daughters. Gamergates of Diacamma australe and related species exercise control in a particularly striking way. All Diacamma workers eclose with ‘‘gemmae,’’ a pair of bladder-like appendages on the thorax filled with glandular cells. The gamergate bites off the gemmae from all newly eclosed workers, which results in dramatic behavioral changes, including the loss of their capacity to attract males and to mate (Fukumoto et al., 1989; Peeters and Higashi, 1989). Young workers that eclose immediately after an old gamergate has died and therefore retain their gemmae can mate and will become the colony’s new reproductives. These future gamergates attack nestmate workers by antennation bouts and biting and thus keep them from laying haploid eggs

REPRODUCTIVE CONFLICT IN INSECT

11

Fig. 3. Dominance interactions between workers of the ponerine ant, Dinoponera quadriceps. (A) A high-ranking worker blocks a subordinately crouching worker, and (B) rubs the antenna of the subordinate worker against the tip of her forward bent gaster. From Monnin and Peeters (1999), by permission of Oxford University Press.

(e.g., Cuvillier-Hot et al., 2001). Similarly, through antennation, immobilization, and other attacks, dominant workers of Dinoponera quadriceps, Gnamptogenys menadensis, and other ponerines monopolize mating and become gamergates (e.g., Gobin et al., 2001; Monnin and Peeters, 1999) (Fig. 3). Egg cannibalism constitutes the second important tactic by which dominant reproductives can control worker reproducing. The eating of viable worker-laid eggs is known from wasps, for example, Dolichovespula arenaria (Greene et al., 1976), and several ants, including the formicoxenine Leptothorax acervorum (Bourke, 1991) and the ponerine Dinoponera quadriceps (Monnin and Peeters, 1997).

12

¨ RGEN HEINZE JU

2. Queen Pheromones: Manipulation or Honest Signaling? Queen control by physical attack or egg eating is obviously not effective in larger colonies (Wilson, 1971). Furthermore, because the chances of a worker reproducing decrease with increasing colony size, conflict over reproduction is expected to be less intense in larger societies (Bourke, 1999). In colonies with tens or hundreds of thousands of workers, such as those of weaver ants or honeybees, the mere presence of a queen appears to suffice to prevent workers from laying viable eggs (e.g., Ho¨lldobler and Wilson, 1983; Seeley, 1985). In the red imported fire ant, Solenopsis invicta, where workers do not have ovaries, the presence of a fertile queen prevents the production of new sexuals from the brood and keeps virgin queens from shedding their wings and starting to lay haploid eggs (Fletcher and Blum, 1981; Vargo and Fletcher, 1986). There is experimental evidence that queens, which are hindered from physically contacting workers by caging, and even fresh corpses or hexane rinses of queens, are sometimes sufficient to suppress worker reproduction (e.g., Ho¨lldobler and Wilson, 1983). This suggests that queens signal their presence by volatile or transmissible chemical compounds: queen pheromones. The composition and origin of queen pheromones and the physiological mechanisms that mediate the response of workers are at present mostly unknown, and the ultimate significance of queen pheromones is still a matter of discussion. In the honeybee, secretions from several of the queen’s glands elicit retinue behavior and maintain social cohesiveness. In particular, her mandibular gland pheromone appears to be involved in the regulation of reproduction (e.g., Hoover et al., 2003; Katzav-Gozansky et al., 2001a,b; Keeling et al., 2003; Wossler et al., 1999). One of its major compounds, 9-oxodecenoic acid, is also produced by queens of the hornet Vespa structor and inhibits egg laying by hornet workers (Blum et al., 1997). In Solenopsis invicta, the queen pheromone that induces workers to kill sexual larvae probably stems from the poison gland, a gland associated with the sting (Klobuchar and Deslippe, 2002), whereas substances that elicit retinue behavior and prevent the shedding of wings in virgin queens come from several other glands (Vargo and Hulsey, 2000). In other taxa, certain long-chained cuticular hydrocarbons appear to be associated with fertility. In general, the blend of nonpolar chemicals on the cuticula is known to communicate information about an individual’s colony membership, age, sex, and task (e.g., Lenoir et al., 1999; Singer, 1998), but cuticular hydrocarbon patterns also vary with reproductive status. Gamergates of Dinoponera quadriceps have significantly higher amounts of the alkene 9-C31:1 on their cuticula than do nonreproductive workers (Monnin et al., 1998), and the relative proportion of this compound

13

REPRODUCTIVE CONFLICT IN INSECT

increases when an individual is allowed to become fertile (Peeters et al., 1999). Similar qualitative or quantitative differences between the cuticular hydrocarbon mixtures distinguish reproductives—both queens and gamergates—and nonreproductives also in other ants, bumblebees, and wasps (Ayasse et al., 1995; Cuvillier-Hot et al., 2001; Dietemann et al., 2003; Hannonen and Sundstro¨m, 2002; Liebig et al., 2000; Sledge et al., 2001). Interestingly, long-chained hydrocarbons also are correlated with dominance in male cockroaches (Roux et al., 2002; Table I). Whether such cuticular substances indeed serve as pheromones is at present unknown. The evidence that they communicate reproductive status is merely correlative and a direct proof is still missing. In the ponerine Pachycondyla inversa, the alkane 3,11-dimethyl-C27 constitutes up to 50% of the cuticular hydrocarbons of fertile queens, but is found only in traces on the cuticula of nonreproductive workers (Heinze et al., 2002) (Fig. 4). Gas chromatography coupled with electroantennographic detection demonstrates that the antennae of workers indeed react to 3,11-dimethyl-C27, TABLE I Cuticular Hydrocarbons Associated with High Fertility in Reproductive Females of Social Insects and, for Comparison, in Male Cockroaches

Taxon

Cuticular hydrocarbons associated with reproductive dominance

Ref.

Polistes dominulus Bombus hypnorum Dinoponera quadriceps Diacamma ceylonense Formica fusca

9-C29:1, 9-C31:1, n-C33, n-C35:2

Sledge et al. (2001)

Branched alkanes, alkadienes, geranyl citronellol 9-C31:1

Ayasse et al. (1995)

Methylated C25 and C27

Cuvillier-Hot et al. (2001)

5-MeC25, 5,13-diMeC25

Harpegnathos saltator Myrmecia gulosa Pachycondyla inversa Platythyrea punctata

13,23-diMeC37

Hannonen and Sundstro¨m (2002) Liebig et al. (2000)

9-C25:1, 3-MeC25

Dietemann et al. (2003)

3,11-diMeC27

Heinze et al. (2002)

3-MeC25

Hartmann (personal communication)

Methylated C35, C36, C38

Roux et al. (2002)

Nauphoeta cinerea

Monnin and Peeters (1999)

14

¨ RGEN HEINZE JU

Fig. 4. Chromatogram of cuticular hydrocarbons from a nonreproductive worker, a reproductively active worker, and a reproductively active queen of Pachycondyla inversa. The arrows indicate the peak of 3,11-diMeC27, a substance that appears to be involved in signalling fertility. Modified from Heinze et al. (2002), by permission of Springer-Verlag.

but behavioral responses of workers to the synthetic substance have not yet been quantified (D’Ettorre et al., 2004a). Physiological studies in the honeybee suggest that queen pheromones have a primer effect directly on the endocrine system of workers (e.g., Kaatz et al., 1992). Queen pheromones have therefore long been regarded as a chemical analog to manipulative antennation and biting, by which queens inhibit the development of worker ovaries against the workers’ own interests (e.g., Fletcher and Ross, 1985; Wilson, 1971). More recently,

REPRODUCTIVE CONFLICT IN INSECT

15

however, queen pheromones have been interpreted as honest signals, which truthfully communicate the fertility of the queen and to which workers react in their own interest by refraining from reproducing (Ho¨lldobler and Wilson, 1990; Keller and Nonacs, 1993; Seeley, 1985; Woyciechowski and Łomnicki, 1987). Pheromonal control is argued to be unstable in evolution, because natural selection would reward nonresponsive workers, leading to a costly arms race between queens and workers. Zahavi and Zahavi (1997) even suggested that queen pheromones are toxic and therefore true handicaps, which only the most healthy and fertile reproductives can risk producing. At present, many authors agree that queen pheromones are honest rather than manipulative, but the unexpected sterility of workers in several monogynous, monandrous species has prompted a critical reexamination of the arguments against pheromonal control. For example, an arms race between queens and workers might result in an evolutionarily stable alternation between a limited set of inhibitory substances, instead of escalating to continually increasing quantities of a single manipulative pheromone or to more and more complex mixtures (Foster et al., 2000). In addition, physiological constraints might simply prevent workers from becoming insensitive to queen manipulation (Heinze et al., 1994a). The honest signaling hypothesis requires that an individual’s pheromone output be directly correlated with her present or future fecundity and also that it reflects her mating status. To avoid cheating, the production of the signal should also be proximately linked to the production of eggs (Keller and Nonacs, 1993). Fertile, mated queens of Pachycondyla inversa have typically much higher amounts of the presumed fertility signal, 3,11-dimethyl-C27, than less productive virgin queens (D’Ettorre et al., 2004a). Queens in founding associations and small colonies have low egg-laying rates and physically dominate their nestmates, and the proportion of 3,11-dimethyl-C27 on their cuticula is comparatively small. In larger colonies, queens are more fertile and less aggressive, and have higher quantities of this substance on their body surface. Similar changes from physical intimidation to signaling have been observed in future gamergates of Diacamma australe, which aggressively dominate their nestmates before they mate but later appear to signal their mating status chemically (Cuvillier-Hot et al., 2002). The honest signaling hypothesis seems to be at odds with the fitness interests of workers in monogynous and monandrous societies. Ignoring possible constraints and costs and considering only the workers’ genetic relationships to their different male relatives, workers would in this case never benefit from letting the queen monopolize reproduction. Queen signaling is more easily understood at effective queen mating frequencies

16

¨ RGEN HEINZE JU

greater than 2, as here the reproductive interests of workers and the queen converge. It is not surprising, therefore, that queen signaling was first suggested for honeybees, where queens mate with multiple males and workers appear to readily distribute queen pheromones throughout the entire hive (Seeley, 1985). Nevertheless, queen pheromones are similarly common and efficacious in monogynous and monandrous societies. This was explained by worker egg laying being associated with costs that considerably reduce the total colony productivity (Cole, 1986; Hamilton, 1972). Such costs may arise from reproductive workers refusing to forage, from fighting among workers about who reproduces and who does not (see below), or from worker-laid eggs being more costly to produce than queen-laid, haploid eggs, because of the often pronounced morphological and physiological differences between the two castes. If, for example, the same amount of investment that gives one worker-laid egg would yield more than two queen-laid eggs, workers would benefit from raising their cheaper brothers instead of the more expensive sons or nephews. In such cases, workers should completely refrain from egg laying (self-policing; Ratnieks, 1988). Self-policing might lead to the complete loss of reproductive organs in workers, when worker reproduction is too costly or when it is highly unlikely that an individual worker can lay eggs after the loss of the colony’s queens. Total worker sterility is known from a number of ant genera, such as Solenopsis, Monomorium, Cardiocondyla, and others (Bourke, 1988a; Bourke and Franks, 1995). Regular polygyny with queen replacement appears to promote ovary loss, and worker sterility, together with polygyny and nest founding by budding or colony fission, has been recognized as one of several characteristics causing the success of invasive ant species (Passera, 1994). 3. Worker Policing and Punishment Workers may oppose egg laying by other workers when it is associated with costs that decrease the total productivity of the colony, or when workers produce males of lower average reproductive value to the workers than does the queen. At the same time, however, individual workers may still benefit from attempting to produce their own sons (Fig. 2). This conflict between individual workers and the rest of the society may result in worker policing: workers actively prevent other workers from laying eggs through aggression, immobilization, or egg cannibalism (Ratnieks, 1988; Seeley, 1985; Starr, 1984). The concepts of policing, dominance, and punishment, all regulating conflict over reproduction in societies, widely overlap and are not unequivocally separated. According to Monnin and Ratnieks (2001), worker

REPRODUCTIVE CONFLICT IN INSECT

17

policing refers to situations in which subordinate workers increase their indirect fitness, for example, by eating the eggs of other workers, whereas dominance means that an individual increases its own direct fitness through aggression against subordinates. Punishment prevents individuals from repeating their actions, whereas policing usually does not—policing by itself therefore does not necessarily enhance cooperation and efficiency (Frank, 2003). Thus, policing and dominance are often difficult to distinguish. In queenless ants, low-ranking workers have been seen policing highranking workers through immobilization, but in other cases workers might selfishly exploit the benefits from apparent ‘‘policing’’ by increasing their own direct fitness instead of augmenting colony efficiency (e.g., Frank, 2003). In the honeybee, Apis mellifera, most workers refrain from reproduction in response to queen pheromones. However, about 1 in 10,000 workers has fully developed ovaries and lays eggs (e.g., Ratnieks, 1995). These fertile workers are more frequently attacked than others (Sakagami, 1954), and most of their eggs are destroyed and eaten by other workers (Ratnieks and Visscher, 1989). Workers probably recognize worker-laid eggs by the lack of certain queen-specific pheromones. Substances from the queen’s Dufour’s gland, a sting-associated gland, have been suggested to play a role in egg discrimination (Ratnieks, 1995), but they more likely constitute a component of the complex queen signal (Katzav-Gozansky et al., 2001a,b, 2003; Martin et al., 2002). Worker policing by discriminative egg eating has also been found in several wasps (Foster and Ratnieks, 2000, 2001), and egg eating in general is quite common in ants (e.g., Crespi, 1991; Ho¨lldobler and Carlin, 1989; Kikuta and Tsuji, 1999). Pachycondyla inversa workers preferentially destroy worker-laid eggs, when queen-laid and worker-laid eggs are simultaneously introduced into their colonies. Egg discrimination appears to be based on 3,11-dimethyl-C27, which is present in considerable amounts on the surface of queen-laid, but not on the surface of worker-laid, eggs (D’Ettorre et al., 2004b). In species with totipotent workers, policing serves an additional role in the regulation of the number of laying workers and/or gamergates. In Dinoponera quadriceps, workers may prevent nestmates from developing their ovaries and replacing the present gamergate by antennation bouts, biting, grasping their legs and antennae with the mandibles, and spreadeagling them. Such ‘‘immobilization’’ may continue over several days and often results in loss of rank or even death of the attacked individual (Fig. 5) (Monnin and Peeters, 1999). Immobilization may be elicited by the gamergate rubbing her sting against the body of a challenging worker and besmearing her with secretions from the Dufour’s gland (Fig. 3B). In such

18

¨ RGEN HEINZE JU

Fig. 5. Immobilization of a high-ranking worker of Dinoponera quadriceps by several lower ranking nestmates. From Monnin and Peeters (1999), by permission of Oxford University Press.

a case, the gamergate and her lower ranking nestmates cooperate in punishing a high-ranking individual (Monnin et al., 2002). Conversely, immobilization may eventually also be used to force a weakening gamergate to abscond (Monnin and Peeters, 1999). Worker policing has also been demonstrated in the ponerines Gnamptogenys menadensis and Harpegnathos saltator by experimentally splitting colonies into two parts and reuniting them after one or several workers have become fertile in the part without an old gamergate. Workers typically attack and immobilize new reproductives and thus force them to stop laying eggs (Gobin et al., 1999; Liebig et al., 1999). In some of the above-mentioned species, above all honeybees, worker policing is expected purely on grounds of relatedness. The importance of colony kin structure seems to be corroborated by facultative policing in the wasp Dolichovespula saxonica, where worker-laid eggs survive in monandrous colonies but are cannibalized in polyandrous colonies (Foster et al., 2001). In contrast, workers of the common wasp, Vespula vulgaris, mutually police each other’s egg laying although they are equally related to the queen’s and other workers’ sons (Foster and Ratnieks, 2000). Likewise, in

REPRODUCTIVE CONFLICT IN INSECT

19

several of the ant species that have been studied, worker policing cannot be explained solely by relatedness. For example, although Pachycondyla inversa is facultatively polygynous, worker policing is not expected because queens are on average unrelated and presumably mostly singly mated (Heinze et al., 2001a). Obviously, worker policing may also evolve when worker egg laying results in a decrease in other workers’ inclusive fitness through efficiency costs at the level of the colony. However, at present good evidence that worker policing indeed increases colony productivity is still lacking. An attempt to quantify costs associated with additional reproductives has been made in the ponerine Platythyrea punctata, one of the few social insects in which all workers can produce diploid offspring from unfertilized eggs by thelytokous parthenogenesis (Heinze and Ho¨lldobler, 1995; Schilder et al., 1999). In such species, workers may essentially be clone mates and, as explicitly modeled for the thelytokous Cape honeybee, Apis mellifera capensis, policing is not expected to arise from kin structure because it should not matter to a worker by whom offspring are produced (Greeff, 1996). Nevertheless, P. punctata workers attack new reproductives after the reunification of two colony fragments and force them to degenerate their ovaries. Adding brood into experimental colonies did not result in increased offspring production, suggesting that one single reproductive can produce all the brood the colony can take care of (Hartmann et al., 2003). Worker-laid diploid eggs in the thelytokous Cape honeybee are also efficiently removed despite the absence of relatedness benefits, but here the costs arising from selfish worker reproduction have not yet been quantified (Pirk et al., 2003; but see Moritz et al., 1999). 4. Escaping Control: Anarchic Honeybees Through intensive search by beekeepers and researchers, a number of honeybee colonies have been detected in which numerous workers, mostly from a single worker lineage, appear to disregard the presence of the queen and lay eggs. These eggs evade policing, possibly because they are marked with the same egg-marking pheromone as queen-laid eggs. ‘‘Anarchist workers’’ produce numerous males and therefore obviously benefit greatly from their selfish behavior at the cost of other worker lineages. Breeding experiments suggest that the anarchic syndrome has a strong heritable component. Nevertheless, anarchist workers appear to be so rare in field colonies that they typically do not reduce colony productivity. The complexity of the syndrome, which is favored by selection only when both ovary activation and escape from policing are expressed together, probably explains its rarity in nature (e.g., Barron et al., 2001). At present it is unknown whether similar countermanipulation against worker or queen policing has evolved in other social insects. Honeybees

20

¨ RGEN HEINZE JU

are by far the most systematically monitored social insects, and large numbers of hives are regularly inspected by beekeepers. Nevertheless, some peculiarities in the pattern of occurrence of worker reproduction might be attributed to similar selfish unresponsiveness of workers to queen signaling or policing. Whereas in the majority of Leptothorax species, worker reproduction is absent or rare in colonies with a queen (e.g., Hammond et al., 2003; Heinze et al., 1997a), workers lay about one-fifth of all haploid eggs in Leptothorax allardycei from Florida (Cole, 1981). It was estimated that the negative effects of worker reproduction on colony productivity are too small for countermeasures to evolve (Cole, 1986), but an analogous calculation might give similar results also for those Leptothorax species in which the workers do not lay eggs. It would therefore be interesting to investigate the phylogenetic or phylogeographic pattern of worker reproduction in Leptothorax, as well as the basis of queen signaling or control. 5. Worker–Worker Conflict In colonies of species, such as Leptothorax allardycei, in which workers lay eggs and presumably produce their own sons despite the presence of the queen (Cole, 1981), as well as in orphaned colonies of other social insects, a second level of disagreement concerning the origin of males arises: which of the workers will lay eggs and which will continue to work? Small insect societies appear to be structured as linear or near-linear dominance hierarchies that are established through aggressive interactions either already in the presence of the socially and reproductively dominant queen or once she has died or been experimentally removed. Although worker hierarchies have long been known from paper wasps and bumblebees, many behavioral studies suggest that ant workers treat each other amicably without observable aggression among nestmates. The observation of ritualized dominance interactions and rank orders in Leptothorax allardycei and Protomognathus americanus therefore was quite unexpected (Cole, 1981; Franks and Scovell, 1983). Just as queens patrol their nests and pummel the head and thorax of all nestmates they encounter, some workers engage in violent antennation to which others react by avoidance, submissive crouching, or retaliation. These interactions do not occur at random but mostly in a consistent, unidirectional fashion that suggests that some workers are dominant over others. Indeed, the existence of linear or near-linear rank orders structuring the social relationships at least among some of the group members has meanwhile been deduced by intensive observations on queenright and orphaned colonies of several ant species, in particular within the Formicoxenini (Bourke, 1988b; Heinze, 1996; Heinze

REPRODUCTIVE CONFLICT IN INSECT

21

et al., 1997a) and the Ponerinae (e.g., Ito and Higashi, 1991; Liebig et al., 1997; Monnin and Peeters, 1999; Oliveira and Ho¨lldobler, 1990; Peeters and Tsuji, 1993). Outside these two taxonomic groups, worker hierarchies have as yet been studied only in Eutetramorium mocquerysi, a Madagascan myrmicine without morphologically distinct queens (Heinze et al., 1999b), and Acanthomyrmex ferox, a myrmicine with polymorphic workers, of which the large ‘‘major workers’’ engage in ‘‘sumo wrestling’’-like shaking contests (Gobin and Ito, 2003). It is likely that additional studies will reveal rank orders also in species with small colonies in other subfamilies of ants. The Australian Myrmeciinae and slave-making taxa within the Formicinae are especially promising candidates. Indeed, Hung (1973) mentioned worker aggression in the slave-making ants Formica pergandei and Polyergus breviceps, but this preliminary observation has never been followed up by a full investigation. In the Formicoxenini, top-ranking workers are typically fed and groomed more often than their lower-ranking nestmates and they remain near or on the brood pile instead of engaging in activities outside the nest (e.g., Bourke, 1988b; Cole, 1981; Franks and Scovell, 1983). Whereas the role of social dominance in evolution is often controversial in vertebrates, in particular primates (e.g., Ellis, 1995), reproductive status and hierarchy rank are typically closely correlated in insect societies. This can easily be proven in those species in which only one or two high-ranking individuals lay eggs, as when a socially dominant individual monopolizes reproduction. In other species, such as Pachycondyla inversa, many workers develop their ovaries after queen removal and egg-laying rates do not always match the social hierarchy as deduced from aggressiveness. Using artificially assembled colonies with genetically identifiable workers, Trunzer et al. (1999) showed that, although in some colonies workers of lower rank laid more eggs during a certain observation time than did higher ranking individuals, the alpha worker typically had the highest share in the maternity of the adult males eclosing in the colony. This is presumably due to differential egg eating. Just as fertile queens and gamergates are characterized by special chemical signatures (see Section III.A.2), the reproductive status of workers appears to be associated with variation in cuticular hydrocarbons. In Harpegnathos saltator, all egg layers—queens, gamergates, or unmated workers—are characterized by the presence of hydrocarbons with particularly long chain length (Liebig et al., 2000; Table I). Similarly, the cuticular bouquet of dominant, egg-laying workers of Pachycondyla inversa resembles that of fertile queens although they apparently do not produce the same quantity of 3,11-dimethyl-C27 as do the queens. This probably allows

22

¨ RGEN HEINZE JU

workers to discriminate among fertile queens and workers (Fig. 4) (Heinze et al., 2002). At present it is unknown how reproduction is regulated in larger, orphaned societies, where physical dominance is not a suitable mechanism. Typically, only young workers lay eggs, but it appears that in large colonies not all young individuals become reproductive. B. Conflict over Sex Allocation Trivers and Hare (1976) first recognized the existence of a queen–worker conflict over sex allocation. Their seminal paper inspired a still growing number of theoretical and empirical studies on sex allocation patterns in the social Hymenoptera. The refinements of sex ratio theory and the accumulated mass of data have frequently been reviewed (e.g., Bourke and Franks, 1995; Chapuisat and Keller, 1999; Crozier and Pamilo, 1996; Pamilo et al., 1997; Mehdiabadi et al., 2003; Queller and Strassmann, 1998; Reuter and Keller, 2001) and it is not an aim of this article to replicate these efforts in detail. Observations that population-wide sex allocation ratios are female biased in monogynous species, but more male biased in polygynous ones, quite convincingly suggest worker control of sex allocation (e.g., Boomsma, 1989; Nonacs, 1986; Pamilo, 1990; Trivers and Hare, 1976). Although the queen wins the conflict about the origin of males, workers appear to enforce their interests concerning sex allocation. Slave-making ants are thought to be a particularly good system in which to test predictions from sex ratio theory. Slave-making ants are social parasites, whose queens usurp the nests of other ant species and take over their brood, from which slaves are reared, which then nurse the slavemaker queen’s young. Slave-maker workers neither forage nor provide care for the brood, but instead replenish the stock of slave workers by pillaging brood from neighboring nests of the host species (Buschinger, 1986; D’Ettorre and Heinze, 2001; Ho¨lldobler and Wilson, 1990). Because slave-maker workers have little contact with the brood and thus appear to lack the opportunity to manipulate sex allocation (Trivers and Hare, 1976), sex ratios in monogynous, monandrous parasitic ants are predicted to be close to the queen’s optimum of 1:1. Sex allocation ratios are indeed unbiased in some slave makers (Bourke, 1989; Bourke and Franks, 1995; Savolainen and Deslippe, 1996; Trivers and Hare, 1976) but in others exhibit variation from colony to colony (Herbers and Stuart, 1998; Pamilo and Seppa¨, 1994). This suggests that additional features of the biology of slave-making ants affect sex allocation, including local competition of a colony’s female offspring for host nests, local mate competition

REPRODUCTIVE CONFLICT IN INSECT

23

among males, the misdirected interests of the slaves, and others. Sex allocation patterns therefore are not exclusively shaped by queen–worker conflict. Attention has switched from population-level sex allocation ratios to the often considerable variation between colonies. According to split sex ratio theory, differences in the relatedness of workers to male and female sexuals (relatedness asymmetries), resulting from varying queen number or queen mating frequency, will be reflected in differences in sex ratios for those colonies (Boomsma and Grafen, 1990, 1991; Ratnieks and Boomsma, 1997). A number of field studies, in particular on Formica ants, have corroborated this prediction. Colony-level sex ratios were consistently female biased over 4 years in monandrous colonies, but heavily male biased in polyandrous colonies of Formica truncorum (Sundstro¨m, 1994). Similarly, in Leptothorax acervorum, sex allocation ratios vary with queen number and/or worker relatedness (Chan et al., 1999; Heinze et al., 2001b). Workers therefore appear to be able to detect variation in queen number and/or queen mating frequency (Boomsma et al., 2003) and to react to this variation adaptively by manipulating sex ratio. They may do so by eating haploid eggs or cannibalizing male larvae, as evidenced by the difference between the primary sex ratio produced by the queen and the ratio of adult male and female sexuals (Aron et al., 1994, 1995; Chapuisat et al., 1997; Sundstro¨m et al., 1996) or by increasing the queen bias of female brood (Hammond et al., 2002). Although colony-level sex ratios are almost always more male biased in societies with low nestmate relatedness than in societies with high relatedness (but see Fournier et al., 2003), this does not necessarily mean that workers are rearing their favored sex ratios. More and more studies have reported bimodality of sex ratios without underlying variation in colony genetic structure. For example, in dense populations of the monogynous, monandrous ant Temnothorax nylanderi, young founding queens may attempt to invade alien, established colonies and to replace the resident queen when suitable nest sites are sparse. Workers respond to a successful usurpation by rearing a male-biased brood, although they are equally unrelated to female and male offspring of the usurper queen and therefore do not gain in inclusive fitness from manipulating sex allocation. A similar pattern is observed when unrelated colonies fuse or are artificially mixed in the laboratory (Foitzik and Heinze, 2000; Foitzik et al., 2003). Monogyny and monandry appear to be ancestral traits in palearctic Temnothorax and it is therefore unlikely that workers have retained this response from polygynous or polyandrous ancestors, in which it would have been adaptive.

24

¨ RGEN HEINZE JU

Colony-level variation of sex allocation without variation in relatedness asymmetries can be interpreted as evidence of an ongoing queen– worker conflict (e.g., Banschbach and Herbers, 1996). The impact of queen and workers on sex allocation might vary with colony size, the number and fertility of queens, and nest architecture. Studies even suggest that the queens, although often considered as ‘‘losers’’ in this conflict, may eventually have complete control over sex allocation. The exchange of Solenopsis invicta queens between colonies that had previously specialized in the production of either male or female sexuals reversed the sex ratio among sexuals later reared in the adopting colony (Passera et al., 2001). Here, split sex ratios may result from a mixed ESS (evolutionarily stable strategy), with queens either limiting the number of diploid eggs, resulting in male production, or providing an unlimited number of diploid eggs, leading to the production of virgin queens (Roisin and Aron, 2003). Finally, additional factors, such as variability in resource availability, the need to recruit new queens from the colony’s own offspring, or varying degrees of local mate competition, probably sometimes exert more influence on sex allocation patterns than colony kin structure and the resulting conflicts between queen and workers (e.g., Aron et al., 1999; Brown and Keller, 2000; Brown et al., 2002; Cremer and Heinze, 2002). C. Queen–Queen Conflict over the Partitioning of Reproduction In contrast to conflicts over the production of males and sex allocation, the conflict among queens has analogies in other group-living animals. As in troops of primates or packs of carnivores, reproductives have different interests concerning the partitioning of reproduction (‘‘reproductive skew’’), and the resulting disagreements often lead to dominance interactions and violent fighting. Multiple queening comprises two phenomena with different ecological and genetic causes and consequences: pleometrosis and secondary polygyny. In many wasps and ants, young queens may avoid the risky phase of solitary founding, during which they rear their first offspring without the help of adult workers, by cooperating with other foundresses (pleometrosis). Queens in such foundress associations are often unrelated and their cohabitation typically finds its end once the first workers have eclosed. Only in a few species do foundress associations grow to produce mature, multiple queened colonies (primary polygyny). In contrast, in secondary polygyny, young queens avoid solitary founding by returning to their maternal colony after mating and starting to reproduce there. In this case, queens typically are close relatives, although the repeated adoption of

REPRODUCTIVE CONFLICT IN INSECT

25

queens may lead to a strong dilution of relatedness at high queen numbers. Various authors have discussed the ecological causes of pleometrosis and polygyny, and this is therefore not reviewed here (e.g., Bernasconi and Strassmann, 1999; Bourke and Franks, 1995; Choe and Perlman, 1997; Ho¨lldobler and Wilson, 1977, 1990; Keller, 1993, 1995; Rissing and Pollock, 1988; Strassmann, 1989). The following sections describe what is known about queen–queen conflict in foundress associations and mature societies and add some words on conflict among virgin queens. Variation in the expression of queen–queen conflict can be explained by optimal skew theory, which is described in the last section of this article. 1. Conflict in Foundress Associations Many ant species are characterized by a pronounced queen–worker dimorphism: queens of leafcutter ants, harvester ants and other taxa are much larger than the respective workers and have strong wing muscles and enormous fat bodies. They completely rely on their own nutrient stores during colony founding, that is, they histolyze muscle and storage proteins and use this material to produce eggs and to nourish their first brood without ever leaving their newly founded nest (‘‘claustral founding’’; Ho¨lldobler and Wilson, 1990). In contrast, other ant queens and in particular wasp queens are not endowed with large body reserves and have to forage to provide food for their first brood (e.g., Brown and Bonhoeffer, 2003). Furthermore, whereas queens of many ant species mate during large nuptial swarms and disperse on the wing, wasp queens are often philopatric and stay in the vicinity of their maternal nest. Cofounding wasp queens are therefore often related (e.g., Queller et al., 1990; Strassmann, 1989; but see Queller et al., 2000), whereas ant queens are typically unrelated (Heinze et al., 2001c; Sasaki et al., 1996; Trunzer et al., 1998). These differences probably explain the conspicuous divergence between foundress associations of claustrally founding ants and other ants and wasps. Claustrally founding queens initially cooperate peacefully but later start to fight until all but one queen have been executed or expelled once the first workers have eclosed (Bourke and Franks, 1995; Choe and Perlman, 1997; Heinze, 1993; Rissing and Pollock, 1988). In contrast, cofoundresses of wasps and of the ponerine Pachycondyla inversa immediately engage in dominance interactions, which quickly result in a clear-cut division of labor (Itoˆ, 1993; Kolmer and Heinze, 2000; Pardi, 1946; Reeve, 1991; Ro¨seler, 1991a,b; Strassmann, 1989; Fig. 6). Whereas the dominant queen stays in the nest and guards the brood, subordinate queens leave to forage. In many annual wasp societies, dominant queens produce the majority of reproductive offspring and feed on eggs laid by the subordinate queens (Gervet, 1964; Reeve and Nonacs, 1992)—the latter wait for their time to

26

¨ RGEN HEINZE JU

Fig. 6. Two cofounding Pachycondyla queens fighting for dominance. Photograph by K. Kolmer.

come to replace the dominant and help rear a related dominant’s offspring (but see Queller et al., 2000). In contrast, all P. inversa queens lay eggs at a similar rate, but because the dominant preferentially eats the eggs laid by subordinate queens, the contribution of individual queens to the first workers is heavily skewed (Heinze et al., 2001c). Cooperation among unrelated queens is probably facilitated by the fact that founding associations of P. inversa may develop into mature, polygynous societies, in which all queens produce sexuals (Trunzer et al., 1999). 2. Conflict in Mature Societies with Multiple Queens Mature societies of about half of all ant species and of many species of wasps may contain several queens, but surprisingly little is known about how queens interact and how egg laying is partitioned. With the exception of so-called oligogynous species, in which queens are intolerant of each other and space out in different parts of the nest (Ho¨lldobler and Wilson, 1990), queens in mature ant societies typically interact amicably or, more often, ignore each other’s presence. Wilson (1974) remarked that queens of Temnothorax curvispinosus ‘‘respond to each other as if they were little more than inanimate objects.’’ Queens occasionally feed on each other’s eggs (e.g., Bourke, 1991; Medeiros et al., 1992; Wilson, 1974), but whereas

REPRODUCTIVE CONFLICT IN INSECT

27

egg eating may be directional in foundress associations and during queen or worker policing, there is little evidence that queens in polygynous societies discriminate between their own eggs and eggs laid by other queens once they have been put onto the egg pile (Bourke, 1994b). Nestmate queens often differ in their egg-laying rates, but these differences are presumably not caused by dominance interactions but instead by individual differences in fecundity. Both ovarian dissections and genetic estimates of parentage suggest that reproductive skew is on average low (Bourke et al., 1997; Fournier and Keller, 2001; Hannonen and Sundstro¨m, 2002). The egalitarian structure of polygynous low-skew societies stands in sharp contrast to the huge inequalities observed in a handful of species belonging to the genera Leptothorax and Formicoxenus. In nearctic Leptothorax sp. A, its palearctic congener L. gredleri, and a few others, queens engage in ritualized antennation bouts and biting (Heinze, 1993; Heinze and Smith, 1990; Heinze et al., 1992, 1993; Ortius and Heinze, 1995, 1999; Ito, personal communication). Queen aggression may eventually escalate to violent mandible fights and mutual sting smearing, during which secretions from the Dufour’s gland are applied to the opponent (Fig. 7). These secretions elicit worker aggression, which may result in injuries and the death of the besmeared queen (Heinze et al., 1998). Regardless of how

Fig. 7. Aggressive interaction between two Leptothorax queens. The individual depicted in black has seized the antennae of the white queen and points her sting toward the opponent. From Heinze et al. (1994a), by permission of Springer-Verlag.

28

¨ RGEN HEINZE JU

violent the interactions are, within a few weeks after the end of hibernation they result in the formation of stable hierarchies in which only the topranking queen reproduces (functional monogyny; Buschinger, 1968, 1974; Heinze and Buschinger, 1988). Monopolization of reproduction by a single queen has been demonstrated by ovary dissection and by genetic investigations (Ortius and Heinze, 1995). Subordinate queens, regardless of their hierarchy rank, almost never lay eggs, but their rank in the hierarchy appears to be important when the alpha queen is removed, because after a period of renewed aggression it is typically the beta queen that becomes fertile (Heinze and Smith, 1990). As with behavioral control by Protomognathus and Polistes queens or future Diacamma gamergates (see above), aggressive interactions among Leptothorax queens become less frequent once the top-ranking individual has started to reproduce, suggesting a switch from behavioral to chemical dominance (Ortius and Heinze, 1999). In the colonies of Formicoxenini, either all queens lay eggs at comparable rates or one queen more or less completely monopolizes reproduction. Societies with medium skew, for example, with one queen laying 75%, the other 25%, of the eggs, are rare. In contrast, reproductive skew varies widely in foundress associations of Polistes paper wasps, from functional monogyny to more or less equal contributions by individual foundresses (Field et al., 1998; Pardi, 1940, 1946; Reeve, 1991). Similarly, queens of the ponerine Odontomachus chelifer form dominance hierarchies by antennation bouts, seizing and pulling the subordinate’s head, and lifting up the subordinate. In contrast to functionally monogynous Formicoxenini, several high-ranking queens may lay eggs (Fig. 8) (Medeiros et al., 1992). The level of skew is also intermediate in the ponerine Pachycondyla tridentata, where 80 to 100% of all nestmates are mated. Mated workers and queens compete for reproduction by antennation, biting, and stinging and in the resulting hierarchies up to one-third of all individuals can lay eggs (Sommer and Ho¨lldobler, 1992; Sommer et al., 1994). Finally, mild aggression among fertile queens suggests the occurrence of dominance hierarchies and intermediate skew among fertile queens in the carpenter ant, Camponotus planatus (Carlin et al., 1993). Within a species, queens are typically less fecund in polygynous than monogynous societies and it was suggested that queens mutually suppress each other’s fertility by inhibitive pheromones (Mercier et al., 1985a,b; Vargo and Fletcher, 1989). However, Bourke (1993) showed that queens of L. acervorum did not lay more eggs when reared under monogynous conditions and concluded that other, as yet unidentified, effects cause the per capita decline in fecundity.

REPRODUCTIVE CONFLICT IN INSECT

29

Fig. 8. Dominance interactions among two queens of Odontomachus chelifer. The dominant queen (depicted in white) (A) antennates a crouching subordinate, (B) grasps her head with elongated mandibles, and (C) lifts her up. From Medeiros et al. (1992), by permission of Blackwell.

3. Competition Among Virgin Reproductives In most social insects, female sexuals leave the nest to mate outside and to independently found their own new society. Consequently, aggression among virgin female sexuals is expected only when they attempt to inherit the maternal colony after the old queen has died or emigrated from the nest. Honeybees are strictly monogynous and new colonies are founded by swarming: some of the workers leaving the hive together with the old queen (e.g., Seeley, 1985; Winston, 1987). Virgin honeybee queens attack and kill immature female sexuals in the brood cells or engage in deadly combat with adult rivals to inherit the orphaned colony (e.g., Gilley, 2001). During such fights, hind gut fluid may be sprayed onto the opponent, eliciting attacks and immobilization by workers (Tarpy and Fletcher,

30

¨ RGEN HEINZE JU

2003). Although nest inheritance is much less common in ants, aggressive interactions leading to the death of virgin queens have also been observed in orphaned colonies of some ant species (e.g., Heinze, 1993). 4. Optimal Skew Theory How reproduction is partitioned among queens differs between species and even within species between different populations. For example, reproductive skew is low in temperate and boreal European populations of Leptothorax acervorum (Bourke et al., 1997; Heinze et al., 1995a,b), but colonies from Hokkaido and from central Spain are functionally monogynous (Felke and Buschinger, 1999; Ito, 1990) and fighting has been observed among queens from colonies from Alaska (Heinze and Ortius, 1991). Optimal skew theory attempts to explain the causes of such social plasticity and how queen–queen conflict is resolved. ESS models that explain variation in the magnitude of reproductive skew were originally proposed by Vehrencamp (1983) and were later extended and refined by Ratnieks, Reeve, Johnstone, and several others (e.g., Johnstone et al., 1999; Keller and Reeve, 1994; Kokko and Johnstone, 1999; Reeve and Keller, 1995, 2001; Reeve and Ratnieks, 1993; for a review see Johnstone, 2000). The share of individual group members in the group’s reproduction is affected by factors such as the genetic relatedness among competitors, the effect subordinates have on the productivity of the dominant, the success rate of solitary breeding, and others. How these factors influence reproductive skew depends, for instance, on who is in control of reproduction, whether the dominant can punish subordinates that claim reproduction by evicting them from the nest, and how good the chances are for a subordinate to take over a nest from a weakening dominant. For example, when the dominant can determine the maximum share subordinates have in the group’s offspring, skew is expected to increase with increasing relatedness, but when the subordinates can claim unsanctioned reproduction, higher relatedness will lead to lower skew (Johnstone, 2000). Foundress associations of Polistes and allodapine bees and multiple queened societies of Leptothorax ants have been used to test skew model predictions. Dominant Leptothorax queens have the power to expel other queens from the nest either alone or together with workers (Heinze and Smith, 1990; Heinze et al., 1992; Ortius and Heinze, 1995). They therefore seem to be in control of the partitioning of reproduction, making a particular class of skew models, the so-called concession models, applicable. These predict that reproductive skew increases with relatedness among queens, the magnitude of ecological constraints on solitary founding, and the influence of subordinates on the reproductive output of the dominant queen (e.g., Johnstone, 2000; Reeve and Ratnieks, 1993; Vehrencamp, 1983).

REPRODUCTIVE CONFLICT IN INSECT

31

In contrast to these predictions, queen–queen relatedness does not differ between functionally monogynous and facultatively polygynous species: Leptothorax queens were on average closely related in all species (Heinze, 1995). Similarly, skew appears to be independent of the relatedness among foundresses in the paper wasp, Polistes bellicosus (Field et al., 1998), whereas it increases with relatedness in foundress associations of Polistes fuscatus (Reeve et al., 2000) and decreases with relatedness in an allodapine bee (Langer et al., 2004). In formicoxenine ants, ecological constraints on solitary nesting appear to be considerably higher in high-skew (functionally monogynous) than in low-skew (polygynous) species (Bourke and Heinze, 1994). Whereas colonies of polygynous species typically live in extended, homogeneous habitats, such as boreal coniferous forests, nests of functionally monogynous species occur in small habitat patches. For example, colonies of Leptothorax sp. A can be found in the half-shaded, lichen-covered zone between dense forest and sun-exposed rocky outcrops, L. gredleri live in rose and blackthorn thickets on the edge of pine–oak forests or in small, scattered pine stands, and functionally monogynous L. sphagnicolus inhabit dry patches in Sphagnum bogs in Que´bec (Bourke and Heinze, 1994). In such patches, young queens, which disperse from the maternal nest, risk getting lost in habitat unsuitable for colony founding, whereas young queens dispersing in homogeneous habitats have a high chance of locating suitable nesting sites. Concession models presuppose that subordinates somehow increase the reproductive output of the dominant queen. Otherwise, subordinates would not benefit from staying with a related dominant and should not tolerate high skew. Subordinates in foundress associations of wasps forage while the dominants guard the nest and the brood. This division of labor presumably increases the number and survival rate of the brood (e.g., Keeping, 1992). Although in Leptothorax foraging is completely taken over by workers, the presence of subordinate queens nevertheless boosts the number of offspring a dominant queen produces. In two-queen colonies of L. gredleri, the first eggs after hibernation were laid with significant delay compared with single-queen colonies, probably because of the energetic costs of hierarchy formation. However, once the rank order was firmly established, the number of eggs and larvae in two-queen colonies increased more sharply and reached higher maximum values (Heinze and Oberstadt, 2003). Higher productivity in two-queen colonies might result either from the more fertile queen becoming dominant in two-queen colonies, or from subordinates transforming their nutrient stores, which they do not need for colony founding and egg production, into glandular secretions that are then fed to the queen and/or the larvae.

32

¨ RGEN HEINZE JU

D. Other Types of Conflict in Insect Societies 1. Kin Discrimination: Conflict Among Different Genetic Lineages Among the various other types of potential conflict in insect societies, conflict between different genetic lineages has been most intensively studied. Although the presence of several queens or multiple mating by the queen lessens the conflict over sex allocation and the origin of males at the same time it introduces a new source of conflict between the offspring of different mothers or fathers. Each worker benefits most from rearing the most closely related female sexuals or, in the case of colony fission, from joining the most closely related queen. Nepotism requires that individuals can reliably distinguish between more or less closely related nestmates (kin discrimination; e.g., Ho¨lldobler and Michener, 1980; Michener and Fletcher, 1987), but after early promising studies were criticized because of a too simple or artificial experimental design, evidence of the existence of true kin discrimination has remained scarce (e.g., Breed et al., 1994; Carlin and Frumhoff, 1990; Grafen, 1990; Keller, 1997). In contrast, several studies have specifically documented the absence of kin discrimination in particular contexts. For example, queens in foundress associations of the paper wasp, Polistes carolina, do not preferentially feed their own larvae (Strassmann et al., 2000b) and colony budding of polygynous colonies of Leptothorax acervorum does not result in a separation of lineages (Heinze et al., 1997b). A notable exception is a study of two-queened colonies of the ant, Formica fusca. A positive association between queen–worker relatedness and the increase in the share the more closely related queen had in the brood between the egg and the larval stage was interpreted as evidence of worker nepotism during brood rearing (Hannonen and Sundstro¨m, 2003). As in the case of other types of conflict, actual conflict among genetic lineages may be selected against because of the risk of making costly mistakes, the reduction of group level efficiency, or the active scrambling of recognition signals (e.g., Page et al., 1989; Ratnieks and Reeve, 1991, 1992; Reeve, 1998). There is strong evidence that the chemical bouquet on the cuticula of social insects, especially long-chained cuticular hydrocarbons, is the cue that enables social insects to discriminate among individuals from different colonies, and also, within colonies, among nestmates belonging to different castes, age classes, or social and reproductive status (Lenoir et al., 2001; Singer, 1998). The dynamics of the formation of a common colony odor, especially the constant transfer of cues among individuals, appear to prevent easy discrimination among differently related nestmates. The chemical profiles of workers from polyandrous Formica colonies indeed differ enough among patrilines for workers to tell that their mother is

REPRODUCTIVE CONFLICT IN INSECT

33

multiply mated, but at the same time overlap sufficiently to make nepotism difficult (Boomsma et al., 2003). 2. Caste Differentiation and Selfish Larvae Under certain conditions, the interests of workers and queens concerning the allocation of resources to colony growth and maintenance, versus reproduction, may diverge (Bourke and Chan, 1999; Herbers et al., 2001; Pamilo, 1991b; Reuter and Keller, 2001). This conflict is obviously tightly linked with sex allocation considerations, and Herbers et al. (2001) predict it to be most pronounced in monogynous, monandrous societies. Studies of Temnothorax nylanderi found a negative association between male and reproductive allocation ratios, suggesting that workers are in control of allocation (Foitzik et al., 2003). Adult workers and queens may also have different interests from the brood about whether female larvae develop into sexuals or workers. In mature colonies, each additional worker makes a smaller contribution to colony survival and reproductive output and at a certain stage a larva might prefer to develop into a sexual, while the other adult nestmates would still gain if she became a worker (Bourke and Ratnieks, 1999). Likewise, in species in which male or female sexuals are polymorphic and have different dispersal and reproductive tactics, individual interests of sexual larvae probably do not coincide with the interests of the rest of the colony. For example, small queens of the fire ant, Solenopsis geminata, typically seek adoption into orphaned colonies after mating, whereas large queens found new colonies independently. Large queens were estimated to have a somewhat higher founding success than small queens—the optimum for a female sexual larva therefore should be to develop into a large queen. However, because small queens need less investment, workers receive a similar fitness-return from large and small queens and benefit from producing both morphs in equal numbers (McInnes and Tschinkel, 1995). The same reasoning applies to taxa with polymorphic males, as in the ant genus Cardiocondyla (e.g., Cremer and Heinze, 2003). Wingless Cardiocondyla males compete in deadly fights to become the only adult wingless male in the colony and to monopolize mating with a large number of virgin queens eclosing over the next weeks or months, while winged males emigrate a few days after eclosion to mate outside the nest. The small wingless males are much cheaper to produce and the corpses of killed males are cannibalized by larvae, suggesting that with the same amount of energy many more wingless than winged males can be produced. However, the majority of wingless males are killed by their adult rivals before they can mate, whereas almost all winged males survive and have the chance to copulate with virgin queens in the maternal nest before dispersing. Male

34

¨ RGEN HEINZE JU

larvae might therefore prefer to develop into winged males. It appears that workers are in control of larval development and allow larvae to grow into winged males only under certain stressful environmental or social conditions (Cremer and Heinze, 2003). At present it is unclear whether and how conflicts involving the brood are expressed in other species. When caste or morph is determined genetically or in the egg stage, or when the different types are reared during different seasons, selfish larvae may have little power to manipulate their own development. Furthermore, various authors have reported on what might be an aggressive regulation of larval development by adult nestmates. Adult social insects solicit secretions, food, or water from larvae and bite and chew on them so that their cuticula may eventually be covered with many scars (Buschinger, 1993). Some ant queens even regularly feed on larval hemolymph (Ito and Billen, 1998; Masuko, 1986). Bourke and Ratnieks (1999) point out that in the stingless bee genus Melipona, queens and workers are of the same size and the workers do not directly feed the brood. Female larvae may therefore selfishly develop into new queens, resulting in overproduction and subsequent elimination of surplus adult female sexuals. Similarly, caste conflict results in an occasional overabundance of female sexuals in epiponine wasps, among which queens and workers are morphologically indistinguishable (Strassmann et al., 2002; Wenseleers et al., 2003). 3. The Neglected Sex: Hymenopteran Males The societies of social Hymenoptera are built on the two female castes, and males do not play an important role in their everyday life. Males are short-lived and rarely engage actively in social interactions. In most species, their only task is to inseminate the female sexuals during a short nuptial season of a few hours or days (Boomsma et al., 2005). Nevertheless, males have fitness interests, for example concerning their own mating frequency, the queen’s mating frequency, and the origin of new males, and these interests may oppose those of other group members (Boomsma, 1996). Because of haplodiploid sex determination, males do not have sons, but they may have grandsons. Hence, whereas the queen’s optimum is to monopolize male production in the society, her mate prefers the new males to be worker offspring. While it appears that males typically lack the power to physically enforce their interests, they might benefit from the production of sperm giving rise to ‘‘anarchist workers.’’ Males appear to have few options to increase their reproductive success. The mating systems of social Hymenoptera do not leave males with much opportunity to obtain and defend a harem. Although the males of some wasp species are territorial, the chances for multiple mating in social

REPRODUCTIVE CONFLICT IN INSECT

35

Hymenoptera are on average low. Hymenopteran males therefore are not well equipped for fighting and their sperm content, which is not replenishable due to the degeneration of the testes, is sufficient only for one or a few matings (Boomsma et al., 2005). Notable exceptions are the wingless, fighting Cardiocondyla males described in the preceding section. They are characterized by life-long spermatogenesis, which probably explains why they engage in deadly combat over the chance to monopolize mating with all female sexuals eclosing in a colony (Heinze and Ho¨lldobler, 1993). 4. Conflict in Termite Societies Societies of termites consist of both male and female reproductives and nonreproductives and all individuals are diploid. The variants of conflict dictated by haplodiploidy are therefore not expected in termites, but disagreements about caste development and the partitioning of reproduction in polygamous societies might occur. Because of their concealed life style, comparatively little is known about conflict in termite societies. Aggressive manipulation, such as wing bud mutilation, may be involved in the regulation of sexual development (Roisin, 1994), but studies suggest that most of the observed injuries result from disturbances during colony handling (Korb, personal communication; Thorne and Traniello, 2003).

IV. The Behavioral Side of Conflict Resolution A. Dominance, Subordination, and Reconciliation Dominance behavior and subordination displays are now known from a large range of social Hymenoptera, including bees, wasps, and ants from several subfamilies. It is not surprising that overt aggression, such as immobilization, biting, or stinging, looks similar throughout all of these taxa (Figs. 3B and 5–7). However, the same applies to stereotyped antennation and subordination displays. Although the repertoire and especially the frequencies of the different types of behavior vary between and also within species, it appears that a ‘‘universal’’ set particularly of subordination displays exists that is exhibited not only within the society, but also during territorial displays and in other inter- and intraspecific situations. Violent antennation (‘‘antennal boxing’’; Peeters and Tsuji, 1993) is probably the most widespread dominance display in the social Hymenoptera (Fig. 8A). It is characterized by much more rapid movements than the gentle, inspecting antennation commonly observed among social insects, for example, between a nest guard and a forager returning to the nest. The duration and frequency of antennation (bouts per second) may differ

36

¨ RGEN HEINZE JU

consistently between pairs of opponents, but at present it is not known whether this variation has any particular meaning (Heinze et al., 1999b). Antennal boxing closely resembles food begging and individuals may respond to both dominance and food-begging antennation by opening their mandibles and regurgitating food. Such food offering is a widespread appeasement behavior, exhibited both within and between species, and regular food exchange among nestmates (‘‘trophallaxis’’) might have evolved from conciliatory feeding (Heinze, 1996; Liebig et al., 1997), retaining the appeasing function it has during dominance contests. In the slave-making ant Chalepoxenus muellerianus, subordinate workers typically respond to antennation by opening their mandibles and sticking out their maxillary and labial palps, which the attacker touches with her mouth for a second or so. Clearly, not much food can be transferred during this short ‘‘kiss,’’ although it is unclear whether secretions are exchanged from the postpharyngeal gland, a gland associated with the mouth and that plays an important role in the homogenization of colony odors (Lahav et al., 1999; Lenoir et al., 1999). Another widespread response to an attack is crouching in a pupa-like posture with withdrawn legs and antennae (Figs. 3A and 8). Submissive crouching has been observed not only in wasps, bumblebees, and ants, but in a similar form also in interactions between cockroaches (e.g., Moore et al., 2002). The dominant may climb onto the crouching subordinate and stand on her for a few seconds with extended legs while antennating her head. Subordinate Leptothorax workers have been seen to remain frozen in a crouching position for minutes after the dominant has moved on. The stilting position of the dominant, often associated with a somehow inflated gaster (in particular in Eutetramorium mocquerysi; Heinze et al., 1999b), closely resembles the ritualized display of workers of Myrmecocystus honey ants during territorial contests (Ho¨lldobler and Wilson, 1990), again suggesting that the same behavior may be exhibited in widely varying contexts. Several special behavior patterns serve to mark the opponent with chemicals from glands associated with the sting, or at least the threat to do so. In various species, the dominant may direct the tip of her gaster toward the opponent (‘‘pygidial display’’ or ‘‘gaster curling’’; e.g., Bourke, 1988b; Heinze et al., 1999b; Monnin and Peeters, 1999), touch the opponent with the gaster tip or even grab an antenna and rub it over the gaster tip. Stereotyped gaster rubbing has been observed in ponerines, such as Dinoponera quadriceps (Fig. 3B; Monnin and Peeters, 1999) and Platythyrea punctata (Hartmann et al., 2003), whereas more openly aggressive sting smearing occurs in Leptothorax gredleri, among which queens apply Dufour’s gland secretions during dominance interactions (Fig. 7; Heinze et al.,

REPRODUCTIVE CONFLICT IN INSECT

37

1998). Interestingly, sting smearing is also known from attacks of slavemaking ants on host workers during slave- raids, where substances from the Dufour’s gland are used to deter and confuse host workers (Allies et al., 1986). Neither the ‘‘propaganda substances’’ of slave-making Formicoxenini nor the Dufour’s gland secretions of L. gredleri queens contain specific toxic or repellent compounds. It would be interesting to know whether Dufour’s gland secretions originally served in an intraspecific context to regulate reproduction and were secondarily used to stigmatize host ants as disorderly reproducing workers. Aggressive interactions among workers remain mostly ritualized and intensify into prolonged biting only when individuals are of similar rank or when young workers, whose hierarchy relationships are not yet resolved, are involved. Series of mutual antennation bouts have been observed among callow workers of several Leptothorax species, and a peculiar form of prolonged mandible fencing occurs in Harpegnathos saltator (Liebig, 1998). From fitness and life history considerations it is likely that interactions among queens are more violent than those among workers: workers cannot emigrate to found a new society and the life expectancy of a queenless colony is limited. A comparison of quantity and quality of interactions in Leptothorax gredleri indeed shows that antagonism escalates to overt fighting more frequently among queens than among workers (Fig. 9).

Fig. 9. Occurrence of aggressive interactions of different intensity among Leptothorax queens (total, 1550 interactions in 120 h of observation) and Leptothorax workers (total, 2851 interactions in 68.75 h of observation). Interactions among queens are more likely to escalate to overtly aggressive biting and mandible fights (R  C test of independence, G ¼ 243.3, p < 0.0001).

38

¨ RGEN HEINZE JU

Queen interactions, but not worker interactions, may result in injuries and cause the emigration of one or several opponents from the nest. In primates and other vertebrates, aggressive interactions among group members are often followed by affiliative postconflict behavior, such as allogrooming, that helps to maintain social cohesion (called reconciliation; e.g., Preuschoft and van Schaik, 2000). According to casual observations, dominants in Leptothorax and related genera may intensively lick the head and mouthparts of the attacked individual after submissive crouching or food offering. In the slave-making ant Chalepoxenus muellerianus, workers that exhibited much aggression after experimental removal of the queen showed a suite of previously rarely observed, altruistic behaviors when the queen was returned into the nest. For example, they groomed and offered food to the slave workers (Heinze, 1996). In contrast, a detailed comparison of behavior exhibited by high-ranking Pachycondyla inversa workers before queen removal, during queen absence, and after the queen was returned did not give evidence of behavioral changes that could be interpreted as ‘‘peace-making’’ (Kellner and Heinze, unpublished results). B. The Shape of Hierarchies Insect societies, particularly those consisting of totipotent females, provide suitable systems for testing predictions from models on conflict resolution and hierarchy formation in animals in general. Rank orders in social insects show the typical properties of hierarchies observed among other group-living animals. In most cases they are linear or nearly linear, often with most aggression being directed toward individuals that are immediately following in rank. This suggests feedback with loser and winner effects (e.g., Dugatkin, 1997) and that individuals somehow recognize each other’s rank. Individual recognition, as in the societies of vertebrates, is unlikely to occur in social insects (for a possible exception in the paper wasp Polistes fuscatus, see Tibbetts, 2002). Alternatively, individuals might be characterized by rank-specific chemical ‘‘badges of status,’’ as suggested from the association of cuticular hydrocarbon patterns and social and reproductive status in several taxa (e.g., Cuvillier-Hot et al., 2001, 2002; Heinze et al., 2002; Liebig et al., 2000; Sledge et al., 2001). Because of the colony specificity of cuticular hydrocarbons it has as yet been difficult to determine whether an alpha individual recognizes the social status of an unfamiliar beta individual from an alien colony and vice versa. Research on the establishment and function of hierarchies in social Hymenoptera has remained mostly descriptive and little is known about

REPRODUCTIVE CONFLICT IN INSECT

39

the assembly rules of rank orders and which characteristics of the individuals determine the outcome of dominance interactions. As shown above, queens and gamergates can be in control without exhibiting aggression. Hence, it is not easy to determine, from observations of dominance interactions alone, which individual is on top of the hierarchy; high rankers immediately below the top are often much more aggressive. Typically only a minority of the individuals in the society is aggressive. Foragers, which often are among the oldest workers in the colony, rarely engage in fighting, and most dominance interactions are observed among young individuals, which mostly stay close to the brood. Consequently, in some species, such as Pachycondyla sublaevis, a clear association exists between age and rank (Higashi et al., 1994; Ito and Higashi, 1991). However, this rule is far from universal and, in particular in wasps, old individuals are dominant over younger nestmates and may replace the dominant (e.g., Strassmann and Meyer, 1983). In other species, however, neither age, body size, nor fluctuating asymmetry varies consistently with social status (Heinze and Oberstadt, 1999). Overall, it appears that individuals, once they have obtained their high rank and have become fully fertile, are not easily replaced. In contrast to other animals, reproductives of perennial social insects typically outlive nonreproductives. This has long been known for queens, which in some ant species may live for up to 30 years, but appears also to be true for gamergates (Tsuji et al., 1996) and unmated workers in the clonal ant Platythyrea punctata (Hartmann and Heinze, 2003). In totipotent, queenless ants, hierarchy length depends on the chance that an individual has to replace the reproductive and on the colony-level costs arising from high-ranking individuals that do not work. Hierarchy length was predicted to increase with colony size (Monnin and Ratnieks, 1999), as has indeed been corroborated by an analysis of hierarchy length in three species of Dinoponera with different average colony size (Monnin et al., 2003). The association will be qualitatively similar in hierarchies among workers that can produce only male offspring, and the number of fighting workers was indeed correlated with colony size in orphaned colonies of Leptothorax nylanderi (Pirner and Heinze, unpublished results). Obviously, more empirical and theoretical studies are needed to better understand the dynamics of hierarchy formation in social insects.

V. Conclusion and Perspectives Research has clearly documented that neither the ‘‘primitively’’ eusocial annual societies of wasps and bumblebees nor those of the ‘‘highly’’ eusocial ants and honeybees are free of potential and actual conflict.

40

¨ RGEN HEINZE JU

A multitude of potential conflicts are dictated by the kin structure of the society, with haplodiploidy giving rise to a number of disagreements among group members that are without parallel in other social animals. But how important is the genetic composition of a colony in determining whether a certain conflict is expressed or not? It is widely agreed that kin selection is the best explanation for the frequent occurrence of cooperative behavior among related animals and that the genetic relationships resulting from haplodiploidy have facilitated the evolution of eusociality and altruistic worker castes in the social Hymenoptera (Bourke and Franks, 1995; Crozier and Pamilo, 1996; Ho¨lldobler and Wilson, 1990; Queller, 2000). Kin structure has also provided robust predictions concerning sex allocation, for example, in facultatively polyandrous Formica ants (e.g., Sundstro¨m, 1994), and nicely explains the differences between stingless bees and honeybees concerning, for example, the occurrence of worker reproduction and worker policing (Peters et al., 1999). However, relatedness is only one of three factors in the central theorem of kin selection (Hamilton, 1964, 1972), and without considering the other two factors—the fitness benefits to the recipient of the help and the costs to the helper—kin theory cannot make predictions about how conflicts are resolved. Therefore, variation in nestmate relatedness appears to play only a minor role in shaping the outcome of nestmate conflicts. A growing number of studies explicitly document that the pattern of worker reproduction, the degree of reproductive skew, the occurrence of worker policing, and so on, cannot easily be explained by relatedness theory alone and that other factors must be more important. For example, workers do not reproduce in colonies of the stingless bee, Schwarziana quadripunctata, despite monogyny and monandry (To´th et al., 2003); and workers lay eggs neither in monogynous nor polygynous colonies of the ant, Leptothorax acervorum (Hammond et al., 2003), but do so in colonies of facultatively polygynous Myrmica tahoensis (Evans, 1998). In contrast to expectations from kinship, workers police worker-laid eggs in monogynous, monandrous Vespa crabro (Foster et al., 2002) and reproductive skew among queens appears to be independent of relatedness (Field et al., 1998; Heinze, 1995; Seppa¨ et al., 2002). Although many species are monogynous and presumably have low effective mate numbers, workers in the ant genera Tetramorium and Cardiocondyla have completely lost their ovaries (Bourke, 1988a; Bourke and Franks, 1995; Heinze unpublished results), and many social insects exhibit split sex ratios independent of any variation in relatedness asymmetries (e.g., Aron et al., 1999; Brown and Keller, 2000; Fjerdingstad et al., 2002; Foitzik and Heinze, 2000; Helms, 1999). Finally, most studies have as yet failed to document differences between polyandrous and monandrous societies regarding disease susceptibility, colony

REPRODUCTIVE CONFLICT IN INSECT

41

homeostasis, or brood production (e.g., Crozier and Fjerdingstad, 2001; but see Cole and Wiernasz, 1999; Tarpy, 2003). In general, the social structure of insect societies appears to be remarkably unresponsive to variation in its genetic structure, with clonal societies of thelytokously parthenogenetic ants showing a division of labor similar to that of colonies of highly polygynous or unicolonial ants, which consist of so many genetic lineages that average nestmate relatedness is close to zero (Hartmann and Heinze, 2003; Ho¨lldobler and Wilson, 1990; Korb and Heinze, 2004; Tsuji, 1988). As the social phenotype is remarkably robust in all these respects, variation of the colony genotype might not be as important in sex allocation patterns as currently thought. Instead, the commonly observed association of low relatedness and a more male-biased sex ratio might arise because nestmate workers belonging to different lineages or species cooperate less efficiently than do workers in a genetically homogeneous colony. The resulting suboptimal acquisition and distribution of resources might result in a lowered total output of sexuals, in particular of costly young queens. Male-biased sex ratios in usurped, parasitized, polygynous or polyandrous societies might therefore all be nonadaptive consequences of worker inefficiency. It might be informative to compare sex ratios in natural colonies of slave-making ants simultaneously containing one or two slave species or experimentally assembled colonies with workers from one, a few, or many lineages: if worker efficiency was important, sex ratios would be more male biased in the more heterogeneous colonies. Why variation in kin structure is relatively unimportant in shaping the expression of potential conflict is explained by the constraints mentioned earlier. Genetic relationships may be masked or actively camouflaged, and manipulation may be impossible because of the lack of power or feasible only at a high cost to the society as a whole (Beekman and Ratnieks, 2003; Beekman et al., 2003; Ratnieks and Reeve, 1992). Competition among colonies therefore appears to override conflict among individual group members and, in particular in larger societies, colony-level properties probably have a higher impact on the inclusive fitness of group members than their individual selfishness (Hammond and Keller, 2004; Korb and Heinze, 2004). The manifestation of conflict is also constrained because different types of conflict are strongly interrelated (Bourke and Franks, 1995; Crozier and Pamilo, 1996). Workers attempting to maximize their inclusive fitness cannot draw from an unlimited number of eggs but must partition their investment into male-, worker-, and queen-destined brood. Sex ratio considerations are therefore not independent from interests concerning the ratio of investment into sexuals and workers. In most species, worker-laid eggs are haploid and might be more easily recognized than queen-laid

42

¨ RGEN HEINZE JU

haploid eggs. Sex allocation and worker reproduction are therefore tightly connected. Similarly, the magnitude of reproductive skew may be constrained by queen–worker conflict over sex allocation (Bourke, 2001). The few studies that have simultaneously investigated several types of conflict have come to the conclusion that sex ratio biasing and male production may be mutually exclusive strategies (Brown et al., 2003; Walin et al., 1998). Potential conflict is therefore expressed only in certain species and under certain conditions, explaining why even a detailed observation of an ant or honeybee colony will rarely reveal aggression among nestmates. Whereas a plethora of models explain the details of conflict in social insects, in particular sex allocation and reproductive skew, proximate aspects other than the mere description of the relevant behavior have long been neglected. The physiological mechanisms involved in rank acquisition and the resolution of conflict are currently poorly understood. It is unclear which anatomical or physiological characters make an individual dominant over another and which role social experience plays. Social and reproductive dominance appear to be intimately linked: individuals that have become reproductively active are typically also socially dominant and vice versa. Nevertheless, it was possible to separate social and reproductive dominance in Polistes dominulus by ovariectomizing dominant foundresses (Ro¨seler and Ro¨seler, 1989); and previously subordinate queens of Leptothorax gredleri became the undisputed dominants of the colonies when they were allowed to prematurely develop their ovaries by keeping the dominant queens under prolonged hibernation (Heinze, unpublished). It is quite obvious that the endocrine system links dominance behavior and reproduction (Ro¨seler, 1991a,b), but how the receipt of antennation bouts affects the hormonal situation of the attacked individual, and how repeated losing leads to a lowered chance of winning future contests, remains unclear (but see Kravitz, 2000, as an example for advances made in nonsocial arthropods). Physiological studies might also give more insight into the proximate aspects of queen signaling. In a number of taxa, long-chained hydrocarbons are associated with reproductive status but, as mentioned above, a direct proof that these substances act as pheromones is as yet lacking. Furthermore, it is unclear whether and why changes in cuticular hydrocarbons are directly linked with ovary development and whether the production of honestly signaling queen pheromones is costly. A comparison of queen pheromones from different populations of the same species or in clusters of sibling species might help us to understand whether they are indeed honest or manipulative—in the latter case, an arms race between reproductives and nonreproductives would perhaps lead to different outcomes in different clades.

REPRODUCTIVE CONFLICT IN INSECT

43

To conclude, the increasing interest in kin conflict has definitively led to a deeper and more complete understanding of how insect societies function and how a balance between the selfish interests of the individual and the interests of the whole society is achieved. Despite idiosyncrasies, such as eusociality and haplodiploidy, data obtained from particular social Hymenoptera are not only of interest for research on similar taxa, but certainly also have relevance to group-living animals in general.

VI. Summary Insect societies are often considered to be smoothly functioning superorganisms, in which all individuals cooperate to increase the reproductive output of the society as a whole. However, group members are genetically not identical and therefore have conflicting interests concerning, for example, the partitioning of reproduction, the origin of males, and the allocation of resources toward male or female sexuals. Empirical studies have clearly shown that potential conflict may be expressed in ritualized domination, overt fighting, policing, and punishment. The aim of this review is to describe current findings on genetic conflicts among nestmates in insect societies, in particular those of ants, and to outline the theoretical background of conflict. Potential conflict dictated by kin structure seems to lead to overt conflict only under certain conditions, suggesting that the degree of relatedness, although of major importance in the evolution of eusociality, is less influential on colony organization and the regulation of reproduction. The expression of conflict appears to be constrained by the lack of power or information to allow adaptive manipulations or by costs resulting from lowered performance at the level of the colony. Acknowledgments Supported by the Deutsche Forschungsgemeinschaft (DFG) and the INSECTS research network of the Universities of Copenhagen, Florence, Keele, Lausanne, Oulu, Regensburg, Sheffield, and the ETH Zurich; financed by the European Commission via the Research Training Network established under the Improving Human Potential Programme. J. Brockmann, T. Roper, and P. J. B. Slater made helpful comments on the manuscript.

References Allies, A. B., Bourke, A. F. G., and Franks, N. R. (1986). Propaganda substances in the cuckoo ant Leptothorax kutteri and the slave-maker Harpagoxenus sublaevis. J. Chem. Ecol. 12, 1285–1293. Aron, S., Passera, L., and Keller, L. (1994). Queen–worker conflict over sex ratio: A comparison of primary and secondary sex ratios in the Argentine ant, Iridomyrmex humilis. J. Evol. Biol. 7, 403–418.

44

¨ RGEN HEINZE JU

Aron, S., Vargo, E. L., and Passera, L. (1995). Primary and secondary sex ratios in monogyne colonies of the fire ant. Anim. Behav. 49, 749–757. Aron, S., Campan, E., Boomsma, J. J., and Passera, L. (1999). Social structure and split sex ratios in the ant Pheidole pallidula. Ethol. Ecol. Evol. 11, 209–227. Aureli, F., and de Waal, F. B. M. (2000). Why natural conflict resolution? In ‘‘Natural Conflict Resolution’’ (F. Aureli and F. B. M. de Waal, Eds.), pp. 3–10. University of California Press, Berkeley, CA. Ayasse, M., Maelovits, T., Tengo¨, J., Taghizadeh, T., and Francke, W. (1995). Are there pheromonal dominance signals in the bumblebee Bombus hypnorum? Apidologie 26, 163–180. Banschbach, V. A., and Herbers, J. M. (1996). Complex colony structure in social insects. II. Reproduction, queen–worker conflict, and levels of selection. Evolution 50, 298–307. Barron, A. B., Oldroyd, B. P., and Ratnieks, F. L. W. (2001). Worker reproduction in honeybees (Apis) and the anarchic syndrome: A review. Behav. Ecol. Sociobiol. 50, 199–208. Beekman, M., and Ratnieks, F. L. W. (2003). Power over reproduction in social Hymenoptera. Philos. Trans. R. Soc. Lond. B 358, 1741–1753. Beekman, M., Komdeur, J., and Ratnieks, F. L. W. (2003). Reproductive conflict in social animals: Who has power? Trends Ecol. Evol. 18, 277–282. Bernasconi, G., and Strassmann, J. E. (1999). Cooperation among unrelated individuals: The ant foundress case. Trends Ecol. Evol. 14, 477–482. Blum, M. S., Raghuvarman, A., Dkhar, B., Kharbuli, B., and Marak, K. R. (1997). Wasp ‘‘queen substance’’ 9-oxodec-trans-2-enoic acid. J. Endocrinol. Reprod. 1, 20–22. Boomsma, J. J. (1989). Sex-investment ratios in ants: Has female bias been systematically overestimated? Am. Nat. 133, 517–532. Boomsma, J. J. (1996). Split sex ratios and queen–male conflict over sperm allocation. Proc. R. Soc. Lond. B 263, 697–704. Boomsma, J. J., and Grafen, A. (1990). Intraspecific variation in ant sex ratios and the Trivers– Hare-hypothesis. Evolution 44, 1026–1034. Boomsma, J. J., and Grafen, A. (1991). Colony-level sex ratio selection in the eusocial Hymenoptera. J. Evol. Biol. 3, 383–407. Boomsma, J. J., Nielsen, J., Sundstro¨m, L., Oldham, N. J., Tentschert, J., Petersen, H. C., and Morgan, E. D. (2003). Informational constraints on optimal sex allocation in ants. Proc. Natl. Acad. Sci. USA 100, 8799–8804. Boomsma, J. J., Baer, B., and Heinze, J. (2005). The evolution of male traits in social insects. Annu. Rev. Entomol. 50, in press. Bourke, A. F. G. (1988a). Worker reproduction in the higher eusocial Hymenoptera. Q. Rev. Biol. 63, 291–311. Bourke, A. F. G. (1988b). Dominance orders, worker reproduction, and queen–worker conflict in the slave-making ant Harpagoxenus sublaevis. Behav. Ecol. Sociobiol. 23, 323–333. Bourke, A. F. G. (1989). Comparative analysis of sex-investment ratios in slave-making ants. Evolution 43, 913–918. Bourke, A. F. G. (1991). Queen behaviour, reproduction and egg cannibalism in multiplequeen colonies of the ant Leptothorax acervorum. Anim. Behav. 42, 295–310. Bourke, A. F. G. (1993). Lack of experimental evidence for pheromonal inhibition of reproduction among queens in the ant Leptothorax acervorum. Anim. Behav. 45, 501–509. Bourke, A. F. G. (1994a). Worker matricide in social bees and wasps. J. Theor. Biol. 167, 283–292. Bourke, A. F. G. (1994b). Indiscriminate egg cannibalism and reproductive skew in a multiplequeen ant. Proc. R. Soc. Lond. B 255, 55–59.

REPRODUCTIVE CONFLICT IN INSECT

45

Bourke, A. F. G. (1999). Colony size, social complexity and reproductive conflict in social insects. J. Evol. Biol. 12, 245–257. Bourke, A. F. G. (2001). Reproductive skew and split sex ratios in social Hymenoptera. Evolution 55, 2131–2136. Bourke, A. F. G., and Chan, G. L. (1999). Queen–worker conflict over sexual production and colony maintenance. Am. Nat. 154, 417–426. Bourke, A. F. G., and Franks, N. R. (1995). ‘‘Social Evolution in Ants.’’ Princeton University Press, Princeton, NJ. Bourke, A. F. G., and Heinze, J. (1994). The ecology of communal breeding: The case of multiple-queen leptothoracine ants. Philos. Trans. R. Soc. Lond. B 345, 359–372. Bourke, A. F. G., and Ratnieks, F. L. W. (1999). Kin conflict over caste determination in social Hymenoptera. Behav. Ecol. Sociobiol. 46, 287–299. Bourke, A. F. G., Green, H. A. A., and Bruford, M. W. (1997). Parentage, reproductive skew and queen turnover in a multiple-queen ant analysed with microsatellites. Proc. R. Soc. Lond. B 264, 277–283. Breed, M. D., Welch, C. K., and Cruz, R. (1994). Kin discrimination within honey bee (Apis mellifera) colonies: An analysis of the evidence. Behav. Proc. 33, 25–40. Brown, M. J. F., and Bonhoeffer, S. (2003). On the evolution of claustral colony founding in ants. Evol. Ecol. Res. 5, 305–313. Brown, M. J. F., Schmid-Hempel, R., and Schmid-Hempel, P. (2003). Queen-controlled sex ratios and worker reproduction in the bumble bee Bombus hypnorum, as revealed by microsatellites. Heredity 12, 1599–1605. Brown, W. D., and Keller, L. (2000). Colony sex ratios vary with queen number but not relatedness asymmetry in the ant Formica exsecta. Proc. R. Soc. Lond. B 267, 1751–1757. Brown, W. D., Keller, L., and Sundstro¨m, L. (2002). Sex allocation in mound-building ants: The roles of resources and queen replenishment. Ecology 83, 1945–1952. Buschinger, A. (1968). Mono- und Polygynie bei Arten der Gattung Leptothorax Mayr (Hymenoptera Formicidae). Insectes Soc. 15, 217–226. Buschinger, A. (1974). Monogynie und Polygynie in Insektensozieta¨ten. In ‘‘Sozialpolymorphismus bei Insekten’’ (G. H. Schmidt, Ed.), pp. 862–896. Wissenschaftliche Verlagsgesellschaft, Stuttgart, Germany. Buschinger, A. (1986). Evolution of social parasitism in ants. Trends Ecol. Evol. 1, 155–160. Buschinger, A. (1993). Ra¨tselhafte Narben auf Ameisenlarven. Ameisenschutz Aktuell 7, 54–56. Carlin, N. F., and Frumhoff, P. C. (1990). Nepotism in the honey bee. Nature 346, 706–707. Carlin, N. F., Reeve, H. K., and Cover, S. P. (1993). Kin discrimination and division of labor among matrilines in the polygynous carpenter ant, Camponotus planatus. In ‘‘Queen Number and Sociality in Insects’’ (L. Keller, Ed.), pp. 362–401. Oxford University Press, Oxford. Chan, G. L., Hingle, A., and Bourke, A. F. G. (1999). Sex allocation in a facultatively polygynous ant: Between-population and between-colony variation. Behav. Ecol. 10, 409–421. Chapuisat, M., and Keller, L. (1999). Testing kin selection with sex allocation data in social Hymenoptera. Heredity 82, 473–478. Chapuisat, M., Sundstro¨m, L., and Keller, L. (1997). Sex-ratio regulation: The economics of fratricide in ants. Proc. R. Soc. Lond. B 264, 1255–1260. Choe, J. C. (1988). Worker reproduction and social evolution in ants (Hymenoptera: Formicidae). In ‘‘Advances in Myrmecology’’ (J. C. Trager, Ed.), pp. 163–187. Brill, Leiden, The Netherlands.

46

¨ RGEN HEINZE JU

Choe, J. C. (1995). Plurimatry: New terminology for multiple reproductives. J. Insect Behav. 8, 133–137. Choe, J. C., and Perlman, D. L. (1997). Social conflict and cooperation among founding quenes in ants (Hymenoptera: Formicidae). In ‘‘The Evolution of Social Behavior in Insects and Arachnids’’ (J. C. Choe and B. J. Crespi, Eds.), pp. 392–406. Cambridge University Press, Cambridge. Cole, B. J. (1981). Dominance hierarchies in Leptothorax ants. Science 212, 83–84. Cole, B. J. (1986). The social behavior of Leptothorax allardycei (Hymenoptera, Formicidae): Time budgets and the evolution of worker reproduction. Behav. Ecol. Sociobiol. 18, 165–173. Cole, B. J., and Wiernasz, D. C. (1999). The selective advantage of low relatedness. Science 285, 891–893. Cremer, S., and Heinze, J. (2002). Adaptive production of fighter males: queens of the ant Cardiocondyla adjust the sex ratio under local mate competition. Proc. R. Soc. Lond. B 269, 417–422. Cremer, S., and Heinze, J. (2003). Stress grows wings: Environmental induction of winged dispersal males in Cardiocondyla ants. Curr. Biol. 13, 219–223. Crespi, B. J. (1991). Cannibalism and trophic eggs in subsocial and eusocial insects. In ‘‘Cannibalism: Ecology and Evolution among Diverse Taxa’’ (M. A. Elgar and B. J. Crespi, Eds.), pp. 176–213. Oxford University Press, Oxford. Cook, J. M. (1993). Sex determination in the Hymenoptera: A review of models and evidence. Heredity 71, 421–435. Crozier, R. H., and Fjerdingstad, E. J. (2001). Polyandry in social Hymenoptera—disunity in diversity? Ann. Zool. Fennici 38, 267–285. Crozier, R. H., and Pamilo, P. (1996). ‘‘Evolution of Social Insect Colonies: Sex Allocation and Kin Selection.’’ Oxford University Press, Oxford. Cuvillier-Hot, V., Cobb, M., Malosse, C., and Peeters, C. (2001). Sex, age and ovarian activity affect cuticular hydrocarbons in Diacamma ceylonense, a queenless ant. J. Insect Physiol. 47, 485–493. Cuvillier-Hot, V., Gadagkar, R., Peeters, C., and Cobb, M. (2002). Regulation of reproduction in a queenless ant: Aggression, pheromones and reduction in conflict. Proc. R. Soc. Lond. B 269, 1295–1300. Deslippe, R. J. (2002). The killing of nestmate queens in ant colonies. Southw. Entomol. 25, 111–121. D’Ettorre, P., and Heinze, J. (2001). Sociobiology of slave-making ants. Acta Ethol. 3, 67–82. D’Ettorre, P., Heinze, J., Schulz, C., Francke, W., and Ayasse, M. (2004a). Does she smell like a queen? Chemoreception of a cuticular hydrocarbon signal in the ant Pachycondyla inversa. J. Exp. Biol. 207, 1085–1091. D’Ettorre, P., Heinze, J., and Ratnieks, F. L. W. (2004b). Worker policing by egg-eating in the ponerine ant Pachycondyla inversa. Proc. R. Soc. Lond. B 271, 1427–1434. Dietemann, V., Peeters, C., Liebig, J., Thivet, V., and Ho¨lldobler, B. (2003). Cuticular hydrocarbons mediate discrimination of reproductives and nonreproductives in the ant Myrmecia gulosa. Proc. Natl. Acad. Sci. USA 100, 10341–10346. Dietemann, V., and Peeters, C. (2000). Queen influence on the shift from trophic to reproductive eggs laid by workers of the ponerine ant Pachycondyla apicalis. Insectes Soc. 47, 223–228. Dugatkin, L. A. (1997). Winner and loser effects and the structure of dominance hierarchies. Behav. Ecol. 8, 583–587. Ellis, L. (1995). Dominance and reproductive success among nonhuman animals: A crossspecies comparison. Ethol. Sociobiol. 16, 257–333.

REPRODUCTIVE CONFLICT IN INSECT

47

Evans, J. D. (1998). Parentage and sex allocation in the facultatively polygynous ant Myrmica tahoensis. Behav. Ecol. Sociobiol. 44, 35–42. Felke, M., and Buschinger, A. (1999). Social organization, reproductive behavior and ecology of Leptothorax acervorum (Hymenoptera, Formicidae) from the Sierra de Albarracin in central Spain. Insectes Soc. 46, 84–91. Field, J., Soli´s, C. R., Queller, D. C., and Strassmann, J. E. (1998). Social and genetic structure of paper wasp cofoundress associations: Tests of reproductive skew models. Am. Nat. 151, 545–563. Fjerdingstad, E. J., Gertsch, P. J., and Keller, L. (2002). Why do some social insect queens mate with several males? Testing the sex-ratio manipulation hypothesis in Lasius niger Evolution 56, 553–562. Fletcher, D. J. C., and Blum, M. S. (1981). Pheromonal control of dealation and oogenesis in virgin queen fire ants. Science 212, 73–75. Fletcher, D. J. C., and Ross, K. G. (1985). Regulation of reproduction in eusocial Hymenoptera. Annu. Rev. Entomol. 30, 319–343. Foitzik, S., and Heinze, J. (2000). Intraspecific parasitism and split sex ratios in a monogynous, monandrous ant (Leptothorax nylanderi). Behav. Ecol. Sociobiol. 47, 424–431. Foitzik, S., and Herbers, J. M. (2001). Colony structure of a slavemaking ant. I. Intracolony relatedness, worker reproduction, and polydomy. Evolution 55, 307–315. Foitzik, S., Stra¨tz, M., and Heinze, J. (2003). Ecology, life history, and resource allocation in the ant, Leptothorax nylanderi. J. Evol. Biol. 16, 670–680. Foster, K. R., and Ratnieks, F. L. W. (2000). Facultative worker policing in a wasp. Nature 407, 692–693. Foster, K. R., and Ratnieks, F. L. W. (2001). Convergent evolution of worker policing by egg eating in the honeybee and common wasp. Proc. R. Soc. Lond. B 268, 169–174. Foster, K. R., Ratnieks, F. L. W., and Raybould, A. F. (2000). Do hornets have zombie workers? Mol. Ecol. 9, 735–742. Foster, K. R., Gulliver, J., and Ratnieks, F. L. W. (2002). Worker policing in the European hornet Vespa crabro. Insectes Soc. 49, 41–44. Fournier, D., and Keller, L. (2001). Partitioning of reproduction among queens in the Argentine ant, Linepithema humile. Anim. Behav. 62, 1039–1045. Fournier, D., Keller, L., Passera, L., and Aron, S. (2003). Colony sex ratios vary with breeding system but not relatedness asymmetry in the facultatively polygynous ant Pheidole pallidula. Evolution 57, 1336–1342. Frank, S. A. (2003). Repression of competition and the evolution of cooperation. Evolution 57, 693–705. Franks, N. R., and Scovell, B. (1983). Dominance and reproductive success among slavemaking worker ants. Science 304, 724–725. Fukumoto, Y., Abe, T., and Taki, A. (1989). A novel form of colony organization in the ‘‘queenless’’ ant Diacamma rugosum Physiol. Ecol. Jpn. 26, 55–61. Gervet, J. (1964). Le comportement d’oophagie diffe´rentielle chez Polistes gallicus L. (Hymen. Vesp.). Insectes Soc. 11, 343–382. Gilley, D. C. (2001). The behavior of honey bees (Apis mellifera ligustica) during queen duels. Ethology 107, 601–622. Gobin, B., and Ito, F. (2000). Queens and major workers of Acanthomyrmex ferox redistribute nutrients with trophic eggs. Naturwissenschaften 87, 323–326. Gobin, B., and Ito, F. (2003). Sumo wrestling in ants: Major workers fight over male production in Acanthomyrmex ferox. Naturwissenschaften 90, 318–321. Gobin, B., Billen, J., and Peeters, C. (1999). Policing behaviour towards virgin egg layers in a polygynous ponerine ant. Anim. Behav. 58, 1117–1122.

48

¨ RGEN HEINZE JU

Gobin, B., Billen, J., and Peeters, C. (2001). Dominance interactions regulate worker mating in the polygynous ant Gnamptogenys menadensis. Ethology 107, 495–508. Grafen, A. (1990). Do animals really recognize kin? Anim. Behav. 39, 42–54. Greeff, J. (1996). Effects of thelytokous worker reproduction in kin-selection and conflict in the Cape honeybee, Apis mellifera capensis. Philos. Trans. R. Soc. Lond. B 351, 617–625. Greene, A., Akre, R. D., and Landolt, P. (1976). The aerial yellowjacket Doichovespula arenaria (Fab.): Nesting biology, reproductive production, and behavior (Hymenoptera: Vespidae). Melanderia 26, 1–34. Hamilton, W. D. (1964). The genetical evolution of social behaviour. I and II. J. Theor. Biol. 7, 1–52. Hamilton, W. D. (1972). Altruism and related phenomena, mainly in social insects. Annu. Rev. Ecol. Syst. 3, 193–232. Hammond, R. L., Bruford, M. W., and Bourke, A. F. G. (2002). Ant workers selfishly bias sex ratios by manipulating female development. Proc. R. Soc. Lond. B 269, 173–178. Hammond, R. L., Bruford, M. W., and Bourke, A. F. G. (2003). Male parentage does not vary with colony kin structure in a multiple-queen ant. J. Evol. Biol. 16, 446–455. Hammond, R. L., and Keller, L. (2004). Conflict over male parentage in social insects. Publ. Libr. Science Biol. 2, e248. Hannonen, M., and Sundstro¨m, L. (2002). Proximate determinants of reproductive skew in polygyne colonies of the ant Formica fusca. Ethology 108, 961–973. Hannonen, M., and Sundstro¨m, L. (2003). Worker nepotism among polygynous ants. Nature 421, 910. Hartmann, A., and Heinze, J. (2003). Lay eggs, live longer: Division of labor and life span in a clonal ant species. Evolution 57, 2424–2429. Hartmann, A., Wantia, J., Torres, J. A., and Heinze, J. (2003). Worker policing without genetic conflicts in a clonal ant. Proc. Natl. Acad. Sci. USA 100, 12836–12840. Heinze, J. (1989). A biochemical approach toward the systematics of the Leptothorax ‘‘muscorum’’ group in North America (Hymenoptera: Formicidae). Biochem. Syst. Ecol. 17, 595–601. Heinze, J. (1993). Queen–queen interactions in polygynous ants. In ‘‘Queen Number and Sociality in Insects’’ (L. Keller, Ed.), pp. 334–361. Oxford University Press, Oxford. Heinze, J. (1995). Reproductive skew and relatedness in Leptothorax ants. Proc. R. Soc. Lond. B 261, 375–379. Heinze, J. (1996). Reproductive hierarchies among workers of the slave-making ant, Chalepoxenus muellerianus. Ethology 102, 117–127. Heinze, J., and Buschinger, A. (1988). Polygyny and functional monogyny in Leptothorax ants (Hymenoptera: Formicidae). Psyche 95, 309–325. Heinze, J., and Ho¨lldobler, B. (1993). Fighting for a harem of queens: Physiology of reproduction in Cardiocondyla male ants. Proc. Natl. Acad. Sci. USA 90, 8412–8414. Heinze, J., and Ho¨lldobler, B. (1995). Thelytokous parthenogenesis and dominance hierarchies in the ponerine ant, Platythyrea punctata (F. Smith). Naturwissenschaften 82, 40–41. Heinze, J., and Keller, L. (2000). Alternative reproductive strategies: A queen perspective in ants. Trends Ecol. Evol. 15, 508–512. Heinze, J., and Oberstadt, B. (1999). Worker age, size and social status in queenless colonies of the ant Leptothorax gredleri. Anim. Behav. 58, 751–759. Heinze, J., and Oberstadt, B. (2003). Costs and benefits of subordinate queens in colonies of the ant, Leptothorax gredleri. Naturwissenschaften 90, 513–516. Heinze, J., and Ortius, D. (1991). Social organization of Leptothorax acervorum from Alaska (Hymenoptera: Formicidae). Psyche 98, 227–240.

REPRODUCTIVE CONFLICT IN INSECT

49

Heinze, J., and Smith, T. A. (1990). Dominance and fertility in a functionally monogynous ant. Behav. Ecol. Sociobiol. 27, 1–10. Heinze, J., Lipski, N., and Ho¨lldobler, B. (1992). Reproductive competition in colonies of the ant Leptothorax gredleri. Ethology 90, 265–278. Heinze, J., Gu¨bitz, T., Errard, C., Lenoir, A., and Ho¨lldobler, B. (1993). Reproductive competition and colony fragmentation in the guest ant Formicoxenus provancheri. Experientia 49, 816–816. Heinze, J., Ho¨lldobler, B., and Peeters, C. (1994a). Conflict and cooperation in ant societies. Naturwissenschaften 81, 489–497. Heinze, J., Ortius, D., Ho¨lldobler, B., and Kaib, M. (1994b). Intracolonial discrimination among heterospecific slaves of the ant, Harpagoxenus sublaevis. Behav. Ecol. Sociobiol. 35, 75–83. Heinze, J., Lipski, N., Bourke, A. F. G., and Ho¨lldobler, B. (1995a). Geographical variation in the social and genetic structure of the ant Leptothorax acervorum. Zoology 98, 127–135. Heinze, J., Lipski, N., Schlehmeyer, K., and Ho¨lldobler, B. (1995b). Colony structure and reproduction in the ant, Leptothorax acervorum. Behav. Ecol. 6, 359–367. Heinze, J., Puchinger, W., and Ho¨lldobler, B. (1997a). Worker reproduction and social hierarchies in Leptothorax ants. Anim. Behav. 54, 849–864. Heinze, J., Elsishans, C., and Ho¨lldobler, B. (1997b). Sociogenetic consequences of budding in the ant, Leptothorax acervorum. Naturwissenschaften 84, 249–250. Heinze, J., Oberstadt, B., Tentschert, J., Ho¨lldobler, B., and Bestmann, H. J. (1998). Colony specificity of Dufour gland secretions in a functionally monogynous ant. Chemoecology 8, 169–174. Heinze, J., Foitzik, S., Oberstadt, B., Ru¨ppell, O., and Ho¨lldobler, B. (1999a). A female caste specialized for the production of unfertilized eggs in the ant Crematogaster smithi. Naturwissenschaften 86, 93–95. Heinze, J., Ho¨lldobler, B., and Alpert, G. (1999b). Reproductive conflict and division of labor in Eutetramorium mocquerysi, a myrmicine ant without morphologically distinct female reproductives. Ethology 105, 701–717. Heinze, J., Stra¨tz, M., Pedersen, J. S., and Haberl, M. (2000). Microsatellite analysis suggests occasional worker reproduction in the monogynous ant Crematogaster smithi. Insectes Soc. 47, 299–301. Heinze, J., Trunzer, B., Ho¨lldobler, B., and Delabie, J. H. C. (2001a). Reproductive skew and queen relatedness in an ant with primary polygyny. Insects Soc. 48, 149–153. Heinze, J., Hartmann, A., and Ru¨ppell, O. (2001b). Sex allocation ratios in the facultatively polygynous ant, Leptothorax acervorum. Behav. Ecol. Sociobiol. 50, 270–274. Heinze, J., Trunzer, B., Ho¨lldobler, B., and Delabie, J. H. C. (2001c). Reproductive skew and queen relatedness in an ant with primary polygyny. Insectes Soc. 48, 149–153. Heinze, J., Stengl, B., and Sledge, M. F. (2002). Worker rank, reproductive status and cuticular hydrocarbon signature in the ant, Pachycondyla cf. inversa. Behav. Ecol. Sociobiol. 52, 59–65. Helms, K. R. (1999). Colony sex ratios, conflict between queens and workers, and apparent queen control in the ant Pheidole desertorum. Evolution 53, 1470–1478. Herbers, J. M., and Stuart, R. J. (1998). Pattern of reproduction in slave-making ants. Proc. R. Soc. Lond. B 265, 875–887. Herbers, J. M., deHeer, C. J., and Foitzik, S. (2001). Conflict over sex allocation drives conflict over reproductive allocation in perennial social insect colonies. Am. Nat. 158, 178–192.

50

¨ RGEN HEINZE JU

Higashi, S., Ito, F., Sugiura, N., and Ohkawara, K. (1994). Worker’s age regulates the linear dominance hierarchy in the queenless ponerine ant, Pachycondyla sublaevis (Hymenoptera: Formicidae). Anim. Behav. 47, 179–184. Ho¨lldobler, B., and Carlin, N. F. (1989). Colony founding, queen control, and worker reproduction in the ant Aphaenogaster (¼Novomessor) cockerelli (Hymenoptera: Formicidae). Psyche 96, 131–151. Ho¨lldobler, B., and Michener, C. D. (1980). Mechanisms of identification and discrimination in social Hymenoptera. In ‘‘Evolution of Social Behavior: Hypotheses and Empirical Tests’’ (H. Markl, Ed.), pp. 35–58. Verlag Chemie, Weinheim, Germany. Ho¨lldobler, B., and Wilson, E. O. (1977). The number of queens: An important trait in ant evolution. Naturwissenschaften 64, 8–15. Ho¨lldobler, B., and Wilson, E. O. (1983). Queen control in colonies of weaver ants (Hymenoptera: Formicidae). Ann. Entomol. Soc. Am. 76, 235–238. Ho¨lldobler, B., and Wilson, E. O. (1990). ‘‘The Ants.’’ Harvard University Press, Cambridge, MA. Hoover, S. E. R., Keeling, C. I., Winston, M. L., and Slessor, K. N. (2003). The effect of queen pheromones on worker honey bee ovary development. Naturwissenschaften 90, 477–480. Huber, P. (1802). Observation on several species of the genus Apis, known by the name of humble-bees, and called Bombinatrices by Linnaeus. Trans. Linn. Soc. Lond. 6, 214–298. Hung, A. F. C. (1973). Reproductive biology in dulotic ants: Preliminary report (Hymenoptera: Formicidae). Entomol. News 84, 253–259. Ito, F. (1990). Functional monogyny of Leptothorax acervorum in northern Japan. Psyche 97, 203–211. Ito, F., and Billen, J. (1998). Larval hemolymph feeding and oophagy: Behavior of queens and workers in the primitively ponerine ant Prionopelta kraepelini (Hymenopera: Formicidae). Belg. J. Zool. 128, 201–209. Ito, F., and Higashi, S. (1991). A linear dominance hierarchy regulating reproduction and polyethism of the queenless ant Pachycondyla sublaevis. Naturwissenschaften 78, 80–82. Itoˆ, Y. (1993). ‘‘Behaviour and Social Evolution of Wasps: The Communal Aggregation Hypothesis.’’ Oxford University Press, Oxford. Jemielity, S., and Keller, L. (2003). Queen control over reproductive decisions—no sexual deception in the ant Lasius niger. Mol. Ecol. 12, 1589–1597. Johnstone, R. A. (2000). Models of reproductive skew: A review and synthesis. Ethology 106, 5–26. Johnstone, R. A., Woodroffe, R., Cant, M. A., and Wright, J. (1999). Reproductive skew in multimember groups. Am. Nat. 153, 315–331. Kaatz, H.-H., Hildebrandt, H., and Engels, W. (1992). Primer effect of queen pheromone on juvenile hormone biosynthesis in adult worker honey bees. J. Comp. Physiol. B 162, 588–592. Kardile, S. P., and Gadagkar, R. (2002). Docile sitters and active fighters in paper wasps: A tale of two queens. Naturwissenschaften 89, 176–179. Katzav-Gozansky, T., Soroker, V., and Hefetz, A. (2001a). Honeybees Dufour’s gland— idiosyncrasy of a new queen signal. Apidologie 33, 525–537. Katzav-Gozansky, T., Soroker, V., Ibarra, F., Francke, W., and Hefetz, A. (2001b). Dufour’s gland secretion of the queen honeybee (Apis mellifera): An egg discriminator pheromone or a queen signal. Behav. Ecol. Sociobiol. 51, 76–86. Katzav-Gozansky, T., Soroker, V., Francke, W., and Hefetz, A. (2003). Honeybee egg-laying workers mimic a queen signal. Insectes Soc. 50, 20–23.

REPRODUCTIVE CONFLICT IN INSECT

51

Keeling, C. I., Slessor, K. N., Higo, H. A., and Winston, M. L. (2003). New components of the honey bee (Apis mellifera L.) queen retinue pheromone. Proc. Natl. Acad. Sci. USA 100, 4486–4491. Keeping, M. G. (1992). Social organization and division of labour in colonies of the polistine wasps, Belonogaster petiolata. Behav. Ecol. Sociobiol. 31, 211–224. Keller, L. (Ed.) (1993). ‘‘Queen Number and Sociality in Insects.’’ Oxford University Press, Oxford. Keller, L. (1995). Social life: The paradox of multiple-queen colonies. Trends Ecol. Evol. 10, 355–360. Keller, L. (1997). Indiscriminate altruism: Unduly nice parents and siblings. Trends Ecol. Evol. 12, 99–103. Keller, L., and Nonacs, P. (1993). The role of queen pheromones in social insects: Queen control or queen signal? Anim. Behav. 45, 787–794. Keller, L., and Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends Ecol. Evol. 9, 98–102. Kikuta, N., and Tsuji, K. (1999). Queen and worker policing in the monogynous and monandrous ant, Diacamma sp. Behav. Ecol. Sociobiol. 46, 180–189. Klobuchar, E. A., and Deslippe, R. J. (2002). A queen pheromone induces workers to kill sexual larvae in colonies of the red imported fire ant (Solenopsis invicta). Naturwissenschaften 89, 302–304. Koedam, D., Velthausz, P. H., van der Krift, T., Dohmen, M. R., and Sommeijer, M. J. (1996). Morphology of reproductive and trophic eggs and their controlled release by workers in Trigona (Tetragonisca) angustula Illiger (Apidae, Meliponinae). Physiol. Entomol. 21, 289–296. Kokko, H., and Johnstone, R. A. (1999). Social queueing in animal societies: A dynamic model of reproductive skew. Proc. R. Soc. Lond. B 266, 571–578. Kolmer, K., and Heinze, J. (2000). Rank orders and division of labour among unrelated cofounding ant queens. Proc. R. Soc. Lond. B 267, 1729–1734. Korb, J., and Heinze, J. (2004). Multilevel selection and social evolution of insect societies. Naturwissenschaften 91, 291–304. Kravitz, E. A. (2000). Serotonin and aggression: Insights gained from a lobster model system and speculations on the role of amine neurons in a complex behavior. J. Comp. Physiol. A 186, 221–238. Lahav, S., Soroker, V., Hefetz, A., and Vander Meer, R. K. (1999). Direct behavioral evidence for hydrocarbons as ant recognition discriminators. Naturwissenschaften 86, 246–249. Langer, P., Hogendoorn, K., and Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature 428, 844–847. Lenoir, A., Fresneau, D., Errard, C., and Hefetz, A. (1999). Individuality and colonial identity in ants: The emergence of the social representation concept. In ‘‘Information Processing in Social Insects’’ (C. Detrain, J. L. Deneubourg, and J. M. Pasteels, Eds.), pp. 219–237. Birkha¨user, Basel, Switzerland. Lenoir, A., D’Ettorre, P., Errard, C., and Hefetz, A. (2001). Chemical ecology and social parasitism in ants. Annu. Rev. Ecol. Syst. 46, 573–599. Liebig, J. (1998). Eusociality, female caste specialization, and regulation of reproduction in the ponerine ant Harpegnathos saltator Jerdon. Ph. D. Thesis, Universita¨t Wu¨rzburg, Wu¨rzburg, Germany. Liebig, J., Heinze, J., and Ho¨lldobler, B. (1997). Trophallaxis and aggression in the ponerine ant, Ponera coarctata: Implications for the evolution of liquid food exchange in the Hymenoptera. Ethology 103, 707–722.

52

¨ RGEN HEINZE JU

Liebig, J., Peeters, C., and Ho¨lldobler, B. (1999). Worker policing limits the number of reproductives in a ponerine ant. Proc. R. Soc. Lond. B. 266, 1865–1870. Liebig, J., Peeters, C., Oldham, N. J., Marksta¨dter, C., and Ho¨lldobler, B. (2000). Are variations in cuticular hydrocarbons of queens and workers a reliable signal of fertility in the ant Harpegnathos saltator? Proc. Natl. Acad. Sci. USA 97, 4124–4131. Martin, S. J., Jones, G. R., Chaline, N., Middleton, H., and Ratnieks, F. L. W. (2002). Reassessing the role of the honeybee (Apis mellifera) Dufour’s gland in egg marking. Naturwissenschaften 89, 528–532. Masuko, K. (1986). Larval hemolymph feeding: A nondestructive parental cannibalism in the primitive ant Amblyopone silvestrii Wheeler (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 19, 249–255. McInnes, D. A., and Tschinkel, W. R. (1995). Queen dimorphism and reproductive strategies in the fire ant Solenopsis geminata (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 36, 367–375. Medeiros, F. N. S., Lopes, L. E., Moutinho, P. R. S., Oliveira, P. S., and Ho¨lldobler, B. (1992). Functional polygyny, agonistic interactions and reproductive dominance in the neotropical ant Odontomachus chelifer (Hymenoptera, Formicidae, Ponerinae). Ethology 91, 134–146. Mehdiabadi, N., Reeve, H. K., and Mueller, U. G. (2003). Queens versus workers: Sex-ratio conflict in eusocial Hymenoptera. Trends Ecol. Evol. 18, 88–93. Mercier, B., Passera, L., and Suzzoni, J. -P. (1985a). E´tude de la polygynie chez la fourmi Plagiolepis pygmaea Latr. (Hym. Formicidae). I. La fe´condite´ des reines en condition expe´rimentale monogyne. Insectes Soc. 32, 335–348. Mercier, B., Passera, L., and Suzzoni, J.-P. (1985b). E´tude de la polygynie chez la fourmi Plagiolepis pygmaea Latr. (Hym. Formicidae). II. La fe´condite´ des reines en condition expe´rimentale polygyne. Insectes Soc. 32, 349–362. Michener, C. D., and Fletcher, D. J. C. (1987). ‘‘Kin Recognition in Animals.’’ John Wiley & Sons, New York. Monnin, T., and Peeters, C. (1997). Cannibalism of subordinates’ eggs in the monogynous queenless ant Dinoponera quadriceps. Naturwissenschaften 84, 499–502. Monnin, T., and Peeters, C. (1999). Dominance hierarchy and reproductive conflicts among subordinates in a monogynous queenless ant. Behav. Ecol. 10, 323–333. Monnin, T., and Ratnieks, F. L. W. (1999). Reproduction versus work in queenless ants: When to join a hierarchy of hopeful reproductives? Behav. Ecol. Sociobiol. 46, 413–422. Monnin, T., and Ratnieks, F. L. W. (2001). Policing in queenless ponerine ants. Behav. Ecol. Sociobiol. 50, 97–108. Monnin, T., Malosse, C., and Peeters, C. (1998). Solid-phase microextraction and cuticular hydrocarbon differences related to reproductive activity in queenless ant Dinoponera quadriceps. J. Chem. Ecol. 24, 473–490. Monnin, T., Ratnieks, F. L. W., Jones, G. R., and Beard, R. (2002). Pretender punishment induced by chemical signalling in a queenless ant. Nature 419, 61–64. Monnin, T., Ratnieks, F. L. W., and Branda˜o, C. R. F. (2003). Reproductive conflict in animal societies: Hierarchy length increases with colony size in queenless ponerine ants. Behav. Ecol. Sociobiol. 54, 71–79. Moore, A. J., Haynes, K. F., Preziosi, R. F., and Moore, P. J. (2002). The evolution of interacting phenotypes: Genetics and evolution of social dominance. Am. Nat. 160, S186–S197. Moritz, R. F. A., and Southwick, E. E. (1992). ‘‘Bees as Superorganisms: An Evolutionary Reality.’’ Springer-Verlag, Heidelberg, Germany.

REPRODUCTIVE CONFLICT IN INSECT

53

Moritz, R. F. A., Kryger, P., and Allsopp, M. H. (1999). Lack of policing in the Cape honeybee, Apis mellifera capensis. Behaviour 136, 1079–1092. Nonacs, P. (1986). Ant reproductive strategies and sex allocation theory. Q. Rev. Biol. 61, 1–21. Nonacs, P. (1993). Male parentage and sexual deception in the social Hymenoptera. In ‘‘Evolution and Diversity of Sex Ratio in Insects and Mites’’ (D. L. Wrensch and M. A. Ebbert, Eds.), pp. 384–401. Chapman & Hall, New York. Nonacs, P., and Carlin, N. F. (1990). When can ants discriminate the sex of brood? A new aspect of queen–worker conflict. Proc. Natl. Acad. Sci. USA 87, 9670–9673. O’Donnell, S., and Jeanne, R. L. (1995). The roles of body size and dominance in division of labor among workers of the eusocial wasp Polybia occidentalis (Olivier) (Hymenoptera: Vespidae). J. Kansas Entomol. Soc. 68, 43–50. Oliveira, P. S., and Ho¨lldobler, B. (1990). Dominance order in the ponerine ant Pachycondyla apicalis (Hymenoptera, Formicidae). Behav. Ecol. Sociobiol. 27, 385–393. Ortius, D., and Heinze, J. (1995). Dynamics and consequences of hierarchy formation in the ant, Leptothorax sp. A. Ethology 99, 223–233. Ortius, D., and Heinze, J. (1999). Fertility signaling in queens of a North American ant. Behav. Ecol. Sociobiol. 45, 151–159. Page, R. E., Robinson, G. E., and Fondrk, M. K. (1989). Genetic specialists, kin recognition and nepotism in honey-bee colonies. Nature 338, 576–579. Pamilo, P. (1990). Sex allocation and queen–worker conflict in polygynous ants. Behav. Ecol. Sociobiol. 27, 31–36. Pamilo, P. (1991a). Evolution of colony characteristics in social insects. II. Number of reproductive individuals. Am. Nat. 138, 412–433. Pamilo, P. (1991b). Evolution of colony characteristics in social insects. I. Sex allocation. Am. Nat. 137, 83–107. Pamilo, P., and Seppa¨, P. (1994). Reproductive competition and conflicts in colonies of the ant Formica sanguinea. Anim. Behav. 48, 1201–1206. Pamilo, P., Gertsch, P., Thore´n, P., and Seppa¨, P. (1997). Molecular population genetics of social insects. Annu. Rev. Ecol. Syst. 28, 1–25. Pardi, L. (1940). Poliginia vera ed apparente in Polistes gallicus L. Atti Soc. Tosc. Sci. Nat. 49, 64–70. Pardi, L. (1946). Ricerche sui Polistini. VII. La ‘‘dominazione’’ e il ciclo ovarico annuale in Polistes gallicus L. Boll. Ist. Ent. Univ. Bologna 15, 25–84. Passera, L. (1994). Characteristics of tramps species. In ‘‘Exotic Ants: Biology, Impact, and Control of Introduced Species’’ (D. F. Williams, Ed.), pp. 23–43. Westview Press, Boulder, CO. Passera, L., Aron, S., Vargo, E. L., and Keller, L. (2001). Queen control of sex ratio in fire ants. Science 293, 1308–1310. Paxton, R. J., Bego, L. R., Shah, M. M., and Mateus, S. (2003). Low mating frequency of queens in the stingless bee Scaptotrigona postica and worker maternity of males. Behav. Ecol. Sociobiol. 53, 174–181. Peeters, C. (1991). The occurrence of sexual production among ant workers. Biol. J. Linn. Soc. 44, 141–152. Peeters, C. (1993). Monogyny and polygyny in ponerine ants with or without queens. In ‘‘Queen Number and Sociality in Insects’’ (L. Keller, Ed.), pp. 234–261. Oxford University Press, Oxford. Peeters, C., and Crewe, R. (1984). Insemination controls the reproductive division of labour in a ponerine ant. Naturwissenschaften 71, 50–51.

54

¨ RGEN HEINZE JU

Peeters, C., and Higashi, S. (1989). Reproductive dominance controlled by multilation in the queenless ant Diacamma australe. Naturwissenschaften 76, 177–180. Peeters, C., and Tsuji, K. (1993). Reproductive conflict among ant workers in Diacamma sp. from Japan: Dominance and oviposition in the absence of the gamergate. Insectes Soc. 40, 119–136. Peeters, C., Monnin, T., and Malosse, C. (1999). Cuticular hydrocarbons correlated with reproductive status in a queenless ant. Proc. R. Soc. Lond. B 266, 1323–1327. Peters, J. M., Queller, D. C., Imperatriz-Fonseca, V. L., Roubik, D. W., and Strassmann, J. E. (1999). Mate number, kin selection and social conflicts in stingless bees and honeybees. Proc. R. Soc. Lond. B 266, 379–384. Pirk, C. W. W., Neumann, P., and Ratnieks, F. L. W. (2003). Cape honeybees, Apis mellifera capensis, police worker-laid eggs despite the absence of relatedness benefits. Behav. Ecol. 14, 347–352. Powell, S., and Tschinkel, W. R. (1999). Ritualized conflict in Odontomachus brunneus and the generation of interaction-based task allocation: A new organizational mechanism in ants. Anim. Behav. 58, 965–972. Premnath, S., Sinha, A., and Gadagkar, R. (1995). Regulation of worker activity in a primitively eusocial wasp, Ropalidia marginata. Behav. Ecol. 6, 117–123. Premnath, S., Sinha, A., and Gadagkar, R. (1996). Dominance relationship in the establishment of reproductive division of labour in a primitively eusocial wasp. Behav. Ecol. Sociobiol. 39, 125–132. Preuschoft, S., and van Schaik, C. (2000). Dominance and communication: Conflict management in various social settings. In ‘‘Natural Conflict Resolution’’ (F. Aureli and F. B. M. de Waal, Eds.), pp. 77–105. University of California Press, Berkeley, CA. Queller, D. C. (2000). Relatedness and the fraternal major transitions. Proc. R. Soc. Lond. B 355, 1647–1655. Queller, D. C., and Strassmann, J. E. (1998). Kin selection and social insects. Bioscience 48, 165–175. Queller, D. C., Hughes, C. R., and Strassmann, J. E. (1990). Wasps fail to make distinctions. Nature 344, 388. Queller, D. C., Zacchi, F., Cervo, R., Turillazzi, S., Henshaw, M. T., Santorelli, L. A., and Strassmann, J. E. (2000). Unrelated helpers in a social insect. Nature 405, 784–785. Ratnieks, F. L. W. (1988). Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. Am. Nat. 132, 217–236. Ratnieks, F. L. W. (1995). Evidence for a queen-produced egg-marking pheromone and its use in worker policing in the honey bee. J. Apic. Res. 34, 31–37. Ratnieks, F. L. W., and Boomsma, J. J. (1997). On the robustness of split sex ratio predictions in social Hymenoptera. J. Theor. Biol. 185, 423–439. Ratnieks, F. L. W., and Reeve, H. K. (1991). The evolution of queen-rearing nepotism in social Hymenoptera: Effects of discrimination costs in swarming species. J. Evol. Biol. 4, 93–115. Ratnieks, F. L. W., and Reeve, H. K. (1992). Conflict in single-queen Hymenopteran societies: The structure of conflict and processes that reduce conflict in advanced eusocial species. J. Theor. Biol. 158, 33–65. Ratnieks, F. L. W., and Visscher, P. K. (1989). Worker policing in the honeybee. Nature 342, 796–797. Reeve, H. K. (1991). Polistes. In ‘‘The Social Biology of Wasps’’ (K. G. Ross and R. W. Matthews, Eds.), pp. 99–148. Cornell University Press, Ithaca, NY.

REPRODUCTIVE CONFLICT IN INSECT

55

Reeve, H. K. (1998). Game theory, reproductive skew, and nepotism. In ‘‘Theory and Animal Behaviour’’ (L. A. Dugatkin and H. K. Reeve, Eds.), pp. 118–145. Oxford University Press, Oxford. Reeve, H. K., and Keller, L. (1995). Partitioning of reproduction in mother–daughter versus sibling associations: A test of optimal skew-theory. Am. Nat. 145, 119–132. Reeve, H. K., and Keller, L. (2001). Tests of reproductive-skew models in social insects. Annu. Rev. Entomol. 46, 347–358. Reeve, H. K., and Nonacs, P. (1992). Social contracts in wasp societies. Nature 359, 823–825. Reeve, H. K., and Ratnieks, F. L. W. (1993). Queen–queen conflicts in polygynous societies: Mutual tolerance and reproductive skew. In ‘‘Queen Number and Sociality in Insects’’ (L. Keller, Ed.), pp. 45–85. Oxford University Press, Oxford. Reeve, H. K., Starks, P. T., Peters, J. M., and Nonacs, P. (2000). Genetic support for the evolutionary theory of reproductive transactions in social wasps. Proc. R. Soc. Lond. B 267, 75–79. Reuter, M., and Keller, L. (2001). Sex ratio conflict and worker production in eusocial Hymenoptera. Am. Nat. 158, 166–177. Rissing, S. W., and Pollock, G. B. (1988). Pleometrosis and polygyny in ants. In ‘‘Interindividual Behavioral Variation in Social Insects’’ (R. L. Jeanne, Ed.), pp. 179–222. Westview Press, Boulder, CO. Roisin, Y. (1994). Intragroup conflicts and the evolution of sterile castes in termites. Am. Nat. 143, 751–765. Roisin, Y., and Aron, S. (2003). Split sex ratios in perennial social Hymenoptera: A mixed evolutionary stable strategy from the queen’s perspective? Am. Nat. 162, 624–637. Ro¨seler, P.-F. (1991a). Soziale und reproduktive Dominanz bei Insekten. Naturwissenschaften 78, 114–120. Ro¨seler, P.-F. (1991b). Reproductive competition during colony establishment. In ‘‘The Social Biology of Wasps’’ (K. G. Ross and R. W. Matthews, Eds.), pp. 309–335. Cornell University Press, Ithaca, NY. Ro¨seler, P.-F., and Ro¨seler, I. (1989). Dominance of ovariectomized foundresses of the paper wasp. Polistes gallicus. Insectes Soc. 36, 219–234. Roux, E., Sreng, L., Provost, E., Roux, M., and Cle´ment, J.-L. (2002). Cuticular hydrocarbon profiles of dominant versus subordinate male Nauphoeta cinerea cockroaches. J. Chem. Ecol. 28, 1221–1235. Sakagami, S. F. (1954). Occurrence of an aggressive behavior in queenless hives, with considerations on the social organization of the honeybee. Insectes Soc. 1, 331–343. Savolainen, R., and Deslippe, R. J. (1996). Slave addition increases sexual production of the facultative slave-making ant Formica subnuda. Behav. Ecol. Sociobiol. 38, 145–148. Sasaki, K., Satoh, T., and Obara, Y. (1996). Cooperative foundation of colonies by unrelated foundresses in the ant Polyrhachis moesta. Insectes Soc. 43, 217–226. Schilder, K., Heinze, J., Gross, R., and Ho¨lldobler, B. (1999). Microsatellites reveal clonal structure of populations of the thelytokous ant Platythyrea punctata (F. Smith) (Hymenoptera; Formicidae). Mol. Ecol. 8, 1497–1507. Seeley, T. D. (1985). ‘‘Honeybee Ecology: A Study of Adaptation in Social Life.’’ Princeton University Press, Princeton, NJ. Seeley, T. D. (1989). The honey bee colony as a superorganism. Am. Sci. 77, 546–553. Seppa¨, P., Queller, D. C., and Strassmann, J. E. (2002). Reproduction in foundress associations of the social wasp, Polistes carolina: Conventions, competition, and skew. Behav. Ecol. 13, 531–542.

56

¨ RGEN HEINZE JU

Singer, T. L. (1998). Roles of hydrocarbons in the recognition systems of insects. Am. Zool. 38, 394–405. Sledge, M. F., Boscaro, F., and Turillazzi, S. (2001). Cuticular hydrocarbons and reproductive status in the social wasp Polistes dominulus. Behav. Ecol. Sociobiol. 49, 401–409. Sommer, K., and Ho¨lldobler, B. (1992). Coexistence and dominance among queens and mated workers in the ant Pachycondyla tridentata. Naturwissenschaften 79, 470–472. Sommer, K., Ho¨lldobler, B., and Jessen, K. (1994). The unusual social organization of the ant Pachycondyla tridentata. J. Ethol. 12, 175–185. Spradbery, J. P. (1991). Evolution of queen number and queen control. In ‘‘The Social Biology of Wasps’’ (K. G. Ross and R. W. Matthews, Eds.), pp. 336–388. Cornell University Press, Ithaca, NY. Starr, C. K. (1984). Sperm competition, kinship, and sociality in aculeate Hymenoptera. In ‘‘Sperm Competition and the Evolution of Animal Mating Systems’’ (R. L. Smith, Ed.), pp. 427–464. Academic Press, Orlando, FL. Strassmann, J. E. (1989). Altruism and relatedness at colony foundation in social insects. Trends Ecol. Evol. 38, 256–266. Strassmann, J. E., and Meyer, D. C. (1983). Gerontocracy in the social wasp, Polistes exclamans. Anim. Behav. 31, 431–438. Strassmann, J. E., Zhu, Y., and Queller, D. C. (2000a). Altruism and social cheating in the social amoeba Dictyostelium discoideum. Nature 408, 965–967. Strassmann, J. E., Seppa¨, P., and Queller, D. C. (2000b). Absence of within-colony kin discrimination: Foundresses of the social wasp, Polistes carolina, do not prefer their own larvae. Naturwissenschaften 87, 266–269. Strassmann, J. E., Sullender, B. W., and Queller, D. C. (2002). Caste totipotency and conflict in a large-colony social insects. Proc. R. Soc. Lond. B 269, 263–270. Strassmann, J. E., Nguyen, J. S., Are´valo, E., Cervo, R., Zacchi, F., Turillazzi, S., and Queller, D. C. (2003). Worker interests and male production in Polistes gallicus, a Mediterranean social wasp. J. Evol. Biol. 16, 254–259. Sundstro¨m, L. (1994). Sex ratio bias, relatedness asymmetry and queen mating frequency in ants. Nature 367, 266–268. Sundstro¨m, L., and Boomsma, J. J. (2001). Conflicts and alliances in insect families. Heredity 86, 515–521. Sundstro¨m, L., Chapuisat, M., and Keller, L. (1996). Conditional manipulation of sex ratios by ant workers: A test of kin selection theory. Science 274, 993–995. Tarpy, D. R. (2003). Genetic diversity within honeybee colonies prevents severe infections and promotes colony growth. Proc. R. Soc. Lond. B 270, 99–103. Tarpy, D. R., and Fletcher, D. J. C. (2003). ‘‘Spraying’’ behavior during queen competition in honey bees. J. Insect Behav. 16, 425–437. Thorne, B. L., and Traniello, J. F. A. (2003). Comparative social biology of basal taxa of ants and termites. Annu. Rev. Entomol. 48, 283–306. Tibbetts, E. A. (2002). Visual signals of individual identity in the wasp Polistes fuscatus. Proc. R. Soc. Lond. B 269, 1423–1428. To´th, E., Queller, D. C., Imperatriz-Fonseca, V. L., and Strassmann, J. E. (2002a). Genetic and behavioral conflict over male production between workers and queens in the stingless bee Paratrigona subnuda. Behav. Ecol. Sociobiol. 53, 1–8. To´th, E., Strassmann, J. E., Nogueira-Neto, P., Imperatriz-Fonseca, V. L., and Queller, D. C. (2002b). Male production in stingless bees: Variable outcomes of queen–worker conflict. Mol. Ecol. 11, 2261–2267.

REPRODUCTIVE CONFLICT IN INSECT

57

To´th, E., Strassmann, J. E., Imperatriz-Fonseca, V. L., and Queller, D. C. (2003). Queens, not workers, produce the males in the stingless bee Schwarziana quadripunctata quadripunctata. Anim. Behav. 66, 359–368. Trivers, R. L., and Hare, H. (1976). Haplodiploidy and the evolution of the social insects. Science 191, 249–263. Trunzer, B., Heinze, J., and Ho¨lldobler, B. (1998). Cooperative colony founding and experimental primary polygyny in the ponerine ant Pachycondyla villosa. Insectes Soc. 45, 267–276. Trunzer, B., Heinze, J., and Ho¨lldobler, B. (1999). Social status and reproductive success in queenless ant colonies. Behaviour 136, 1093–1105. Tsuji, K. (1988). Obligate parthenogenesis and reproductive division of labor in the Japanese queenless ant Pristomyrmex pungens. Behav. Ecol. Sociobiol. 23, 247–255. Tsuji, K., Nakata, K., and Heinze, J. (1996). Ants, age and reproduction. Naturwissenschaften 83, 577–578. Vargo, E. L., and Fletcher, D. J. C. (1986). Evidence of pheromonal queen control over the production of male and female sexuals in the fire ant, Solenopsis invicta. J. Comp. Physiol. A 159, 741–749. Vargo, E. L., and Fletcher, D. J. C. (1989). On the relationship between queen number and fecundity in polygyne colonies of the fire ant Solenopsis invicta. Physiol. Entomol. 14, 223–232. Vargo, E. L., and Hulsey, C. D. (2000). Multiple glandular origins of queen pheromones in the fire ant Solenopsis invicta. J. Insect Physiol. 46, 1151–1159. Vehrencamp, S. (1983). A model for the evolution of despotic versus egalitarian societies. Anim. Behav. 31, 667–682. Velicer, G. J. (2003). Social strife in the microbial world. Trends Microbiol. 11, 330–337. Walin, L., Sundstro¨m, L., Seppa¨, P., and Rosengren, R. (1998). Worker reproduction in ants: A genetic analysis. Heredity 81, 604–612. Wenseleers, T., Ratnieks, F. L. W., and Billen, J. (2003). Caste fate conflict in swarm-founding social Hymenoptera: An inclusive fitness analysis. J. Evol. Biol. 16, 647–658. Wheeler, W. M. (1910). ‘‘Ants: Their Structure, Development and Behavior.’’ Columbia University Press, New York. Wheeler, W. M. (1911). The ant-colony as an organism. J. Morphol. 22, 307–325. Wilson, E. O. (1971). ‘‘The Insect Societies.’’ Harvard University Press, Cambridge, MA. Wilson, E. O. (1974). Aversive behavior and competition within colonies of the ant Leptothorax curvispinosus. Ann. Entomol. Soc. Am. 67, 777–780. Winston, M. L. (1987). ‘‘The Biology of the Honey Bee.’’ Harvard University Press, Cambridge, MA. Wossler, T. C., and Crewe, R. M. (1999). Honeybee tergal gland secretion affects ovarian development in caged workers. Apidologie 30, 311–320. Woyciechowski, M., and Łomnicki, A. (1987). Multiple mating of queens and the sterility of workers among eusocial Hymenoptera. J. Theor. Biol. 128, 317–327. Zahavi, A., and Zahavi, A. (1997). ‘‘The Handicap Principle.’’ Oxford University Press, Oxford.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Game Structures in Mutualistic Interactions: What Can the Evidence Tell Us About the Kind of Models We Need? Redouan Bshary* and Judith L. Bronstein{ *institut de zoologie universite´ de neuchaˆtel neuchaˆtel, switzerland { department of ecology and evolutionary biology university of arizona tucson, arizona 85721, usa

I. The Puzzle of Cooperative Behavior Nature is full of examples in which individuals of different species cooperate with each other. Some of these interactions (mutualisms) are crucial to the persistence of the world that we know: most plants need mycorrhizal fungi and/or rhizobial bacteria for successful growth, as well as pollinators for reproduction; coral reefs are the result of a mutualistic symbiosis between polyps and algae; and virtually all animals appear to have endosymbionts that help with digestion of food. Other mutualisms attract human attention because of their oddity: birds and fish that enter the mouths of predators in search of food, birds that lead other animals to a mutually appreciated food source, and anemones that defend the crabs on whose backs they ride. Darwin (1859) was well aware that interspecific mutualism, like intraspecific cooperation, provided a challenge to his theory of evolution. Selection favors individuals that behave selfishly and maximize their own benefit. Cooperative behavior, however, often involves costly investment by one individual for the benefit of its partner. This puzzle of cooperative behavior is best illustrated with the so-called prisoner’s dilemma game. In this game, each of two genetically unrelated players can either cooperate or defect. Both players receive a payoff from the interaction (assumed to be of some relevance to the players’ fitness) that depends on the combination of 59 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

60

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

behaviors the two players have performed. The payoff matrix is constructed such that (1) both players receive a higher payoff from mutual cooperation than from mutual defection, (2) each player receives a higher payoff from defecting than from cooperating, irrespective of the partner’s action, and (3) the player invariably receives a higher payoff if the partner cooperates than when the partner defects. Thus, cooperative behavior is an altruistic act in this game, an investment in the partner’s fitness. Not surprisingly, mutual defection is the only evolutionarily stable outcome under these conditions. Thus, even if the partner cooperates and invests, selection should favor individuals that do not invest in return (both reduced investment and active exploitation are what we refer to as cheating). Why, then, does cooperative behavior seem to be ubiquitous in nature? Furthermore, why do individuals of one species invest in individuals of another species?

II. Game Theoretical Approaches to Mutualism Initially, most theoretical work on the evolutionary stability of cooperative behavior/altruism focused on cooperation within species. Here, the explanation for many cases seems to be linked to the existence of kin selection (Hamilton, 1964). However, examples of within-species cooperation exist in which the partners are unrelated. Furthermore, kin selection cannot explain any example of interspecific mutualism. Trivers (1971) proposed that both intraspecific cooperation and interspecific mutualism can evolve and be maintained when the same individuals interact repeatedly with each other. His idea of reciprocal altruism was formalized by Axelrod and Hamilton (1981) via an iterated version of the prisoner’s dilemma game. In brief, Axelrod and Hamilton found, in a computer simulation tournament with a variety of strategies, that a simple strategy called ‘‘Tit-for-Tat’’ could emerge as a cooperative solution to the game. Tit-for-Tat players begin the game by cooperating and then, in subsequent rounds, do what their partners did in the previous round. For some time, the iterated prisoner’s dilemma was the paradigm for theoretical studies on the evolution of cooperation (reviewed by Dugatkin, 1997). However, empiricists seem to have found it difficult to relate the cooperative strategies proposed by theoreticians, including Tit-for-Tat and its successors (Dugatkin, 1997), to real-world cooperative interactions. There are a few examples in which an observed case of intraspecific cooperation seems consistent with game theory models (Dugatkin, 1997). The situation is different with regard to interspecific mutualism, however. Researchers at the 90th Dahlem workshop, on the evolution of cooperation, argued that

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

61

there is not a single suspected example of mutualism for which (1) the payoff matrix can be described adequately by the prisoner’s dilemma game, and (2) the partners behave as predicted from cooperative solutions to the iterated prisoner’s dilemma game (Bergstrom et al., 2003). Three other game theoretic concepts that until now have attracted less attention have been seen to be more promising frameworks to explain the evolutionary stability of cooperation and mutualism: by-product mutualism, pseudoreciprocity, and biological market theory. By-product mutualism (Brown, 1983) is a confusing term, because it was developed to explain intraspecific cooperation rather than interspecific mutualism. However, the concept applies to both sets of interactions, and yields a simple, straightforward explanation for the occurrence of cooperative behavior: each individual acts selfishly, and the benefits to other individuals accrue as a simple by-product of their behavior. As there is no costly investment on the part of the partners, no altruistic behavior must be explained. This may be the reason why many researchers, in particular theoreticians, tend to ignore this explanation (Dugatkin, 1997): it takes away the most interesting aspect of cooperation. But at the same time, that is the very argument for why this form of cooperation should be found frequently in nature: because it does not pose any problems to either player. Pseudoreciprocity (Connor, 1986) differs from by-product mutualism in two ways: (1) one or both partners invest in each other, and (2) investment makes cooperative behavior the best option for the investing partner. The second point is also the crucial difference between pseudoreciprocity and the prisoner’s dilemma game, in which cheating invariably yields a higher payoff in each round. In pseudoreciprocity, cooperative behavior yields the highest possible payoff in each single round as long as the partner cooperates. The selfishness of cooperative behavior can be understood best with optimality theory. If and only if investment by the partner is above a critical threshold, it pays to perform an act that returns the investment. Thus, there is still no risk that the altruist will be exploited. Biological market theory proposes that cooperative or mutualistic interactions can be viewed as an exchange of goods or commodities between individuals that differ in the degree of control over these goods/commodities (Noe¨, 2001). Control is used in a loose sense here; it may simply imply that a commodity/good is easier to produce for one partner than for the other. Trading partners can be chosen from a number of potential partners. The focus of biological market theory has been to understand how supply and demand ratios of the goods/commodities traded in combination with partner choice opportunities determine the exchange rate (Hoeksema and Schwartz, 2001, 2003; Noe¨, 2001; Noe¨ and Hammerstein, 1994; Noe¨ et al.,

62

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

1991; Schwartz and Hoeksema, 1998). For example, market theory predicts that the nectar-provisioning rate by lycaenid butterfly larvae to tending ants (that would defend the larvae against predators) depends on variables like predation risk, and the number of ants or other larvae present: decreasing predation risk and increasing number of ants reduces the larvae’s demand for tending, hence larvae should reduce nectar production. The ability to choose between partners may also be a mechanism that promotes costly cooperative behavior in the form of investment (Bshary and Noe¨, 2003; Bshary and Scha¨ffer, 2002; Bull and Rice, 1991; Ferrie`re et al., 2002). In particular, if a cheater that refrains from investing belongs to the abundant trading class, it risks being abandoned by the choosing partner and remaining partnerless for a long time, while the choosing partner will easily find a new and hopefully more cooperative partner. The concepts of by-product mutualism, pseudoreciprocity, and partner choice may explain why cooperative behavior persists in mutualistic interactions: either cooperative behavior may not be costly, or investment may yield predictable benefits, or investment may secure repeated interactions. However, these game theoretic concepts are not the focus of most scientists working on mutualism; most work on these interactions is ecological in nature (Bronstein, 1994). One reason might be that game theoretical analyses of mutualism have tended to treat the core interaction quite abstractly, focusing on few aspects rather than explicitly considering a broad range of its ecological or behavioral features. Admittedly, these features vary enormously across particular forms of mutualism (pollination, dispersal, etc.), often obscuring fundamental similarities across interactions differing greatly in natural history (Bronstein, 1994, 2001a). On the other hand, since mutualisms are real interactions that function in real ecological settings, excessive abstraction risks meaningless results. Here, we evaluate the game theoretic approach, with its focus on behavioral strategies of individuals, to identify its potential strength but also its shortcomings in light of empirical knowledge about mutualistic systems. We identify 12 parameters that, in combination, describe features of a mutualism that are relevant for generating a meaningful so-called game structure (Table I). By meaningful, we mean that the game structure allows an exploration of the fitness consequences of individual behavioral strategies that may resemble those that partners actually use in the real world. The parameters can be seen as important modules for the construction of a game, for which evolutionarily stable behavioral strategies can be explored. For example, individuals may interact once or repeatedly, both partners (or one partner, or neither) may be mobile, and so on. We have chosen to examine eight relatively wellunderstood kinds of mutualism. We have selected these mutualisms on the basis of the following criteria: (1) the reciprocal benefits of the mutualism

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

63

TABLE I Parameters Evaluated in Mutualistic Systemsa Parameter

Possible combinations (species 1/species 2)

Dependency Specificity N interactions Offer produced Moves Mobility Active choice Partner recognition Behavioral options Investment Payoff symmetry Control over interaction

high/high, high/low, low/high, low/low high/high, high/low, low/high, low/low repeated, one-off prior/during, during/during, during/prior, during/after simultaneous, alternating, sequential mobile/mobile, mobile/sessile, sessile/mobile, sessile/sessile yes/yes, yes/no, no/yes, no/no yes/yes, yes/no, no/yes, no/no allC/allC, C or D/allC, allC/C or D, C or D/C or D yes-no/yes-no, variable/yes-no, yes-no/variable, variable/variable symmetrical, asymmetrical full/full, full/limited, limited/full, limited/limited

a

Behavioral options: individuals can either only cooperate (‘‘allC’’) or they can cooperate or defect (‘‘C or D’’).

appear clear, and (2) enough is known about the natural history of both species and their interactions to allow us to assess the various parameter states. We recognize, however, that as more is learned about these interactions, we are likely to find that we have misidentified one or more of these parameter states. Indeed, one of our aims is to point to problematic gaps in our empirical knowledge of these mutualisms.

III. Goals of This Article 1. We aim to evaluate the extent to which there are general features that underlie many different mutualisms with respect to game structure. The evolutionary and ecological backgrounds of different mutualisms are highly diverse, prompting the question of whether it is appropriate to categorize them at all. With respect to game theory and questions about the evolutionary stability of mutualisms, one must ask how many different game structures will be found. Does every mutualistic interaction exhibit a unique game structure, or are there important common features? How uniform are game structures within broad classes of mutualism, such as ‘‘cleaning mutualisms’’ or ‘‘pollination mutualisms,’’ which are defined by the actions of partners but may include a wide range of taxa? How uniform are game structures within taxonspecific mutualisms, such as ant–lycaenid interactions, which appear to be relatively uniform with respect to what is traded between partners

64

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

and with respect to phylogeny, but which involve several thousand different species with potentially very different ecological demands? 2. We aim to identify gaps in our empirical knowledge that limit our ability to assess the game structure for certain mutualistic systems. 3. We aim to identify important ecological parameters. Ecological parameters are usually not captured by game theoretic approaches, but still may be of fundamental importance for explaining individual behavior. More generally, we hope to describe natural systems in a way that will facilitate the development of theoretical concepts for mutualism. This in turn will allow us to understand the basic rules for how cooperative behavior between unrelated individuals may persist in nature. At this stage, a descriptive, rather than abstract, theory-driven approach seems necessary to facilitate communication between empiricists and theoreticians. More specifically, we hope to provide empiricists with a framework for the kinds of data theoreticians may find useful, and we hope to provide theoreticians with a wealth of empirical information as a basis for future models. This task is big enough to force us to refrain from using the data to develop theoretical predictions ourselves, or to evaluate the various existing concepts of partner control. We also restrict ourselves to interspecific mutualism and do not further discuss intraspecific cooperation. Intraspecific cooperation often involves cases in which some partners are related to varying degrees, while other partners are unrelated. In practice, it may therefore often be difficult to distinguish between reciprocity arguments and kin selection arguments for the evolution of cooperation in intraspecific interactions. Nevertheless, we would predict that the principles that lead to stable mutualism may also be relevant for intraspecific cooperation between unrelated individuals.

IV. Terminology There has been considerable confusion about the use of terms like cooperation, mutualism, and symbiosis. In brief, the verb ‘‘to cooperate’’ refers to a positive consequence (increased fitness) of the behavior of individuals on the partners with which they interact (which can either be conspecifics or allospecifics), without implying any cognitive abilities such as intentionality. Cooperative behavior may be selfish (it directly increases the fitness of the actor, irrespective of the partner’s action), cost-free, or a costly investment. In the latter case, we refer to cooperative behavior as being altruistic. We use this term with the knowledge that ultimately, we will try to give a functional explanation for such altruistic behavior by

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

65

explaining how the investment behavior promotes benefits that outweigh the costs. Note that kin selection theory too starts with the observation of one animal investing into another, and then explains this oddity away with the concept of relatedness. The same is true for reciprocity based on the iterated prisoner’s dilemma. ‘‘Cooperation’’ and ‘‘mutualism’’ refer to the outcome of interactions: the terms are used if both partners receive a net benefit from the interaction. We use the term ‘‘cooperation’’ as the outcome of intraspecific cooperative behavior, and ‘‘mutualism’’ as the outcome of interspecific cooperative behavior. ‘‘Symbiosis’’ is an intimate spatial relationship between individuals of different species; the outcome can be mutualistic, commensal, or parasitic. Another important terminological issue is what we mean by ‘‘cheating.’’ There has been reasonable confusion in the literature about what cheating refers to and who cheaters are. For a definition of the behavior, we refer to Bull and Rice (1991): an individual cheats if it provides less to its partner than the average individual of its species would provide. This definition acknowledges that cheating may be both a discrete behavior (e.g., an ant eating an aphid) but more often a continuous one (a lycaenid butterfly larva providing less nectar to tending ants than other larvae do). It also acknowledges the existence of what Sherratt and Roberts (2001) called phenotypic defectors. These are individuals who cannot invest (as much as others) because they are in poor condition. From the partner’s point of view, it should always respond to low payoffs, no matter whether its own investment is not reciprocated because the other individual defects or because it cannot reciprocate. We use the term ‘‘cheater’’ only for individuals that belong to the mutualist species under investigation. They may either be individuals who always cheat (designated by Bronstein, 2001b, as ‘‘pure exploiters’’), or individuals that cooperate under some defined range of conditions (‘‘conditional exploiters’’). Bronstein (2001b) also identified ‘‘exploiter species,’’ allospecific exploiters of mutualisms that cannot provide benefits but that take advantage of rewards and services designated for the mutualistic partner. While they are an important ecological and evolutionary problem for mutualisms (Bronstein, 2001b), we set exploiter species aside for the purposes of this article.

V. Parameters Considered for the Assessment of Game Structures The prisoner’s dilemma game is a good starting point to illustrate the kinds of parameters that might be important in defining the game structure of mutualisms, and hence in exploring how evolutionary stability of cooperative

66

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

behavior might be achieved. The version of the iterated prisoner’s dilemma used by Axelrod and Hamilton (1981) for exploring mutualism made the following assumptions: (1) players interact repeatedly; (2) players are assigned randomly to each other as partners; (3) players do not know a priori how many rounds they will play; (4) both players have the same behavioral options (i.e., to cooperate or to cheat); (5) investment in the partner is an all-or-nothing behavior; (6) the payoffs are symmetrical for both players; (7) the payoffs are constructed in such a way that cheating invariably yields a higher payoff than cooperating; (8) the players’ moves are simultaneous; and (9) the offers are produced during the interaction. No assumptions are made regarding mobility of partners, nor whether partners recognize each other on an individual basis (although it may be implied by the repeated game structure). Nor are there any assumptions regarding whether individuals are limited in any way in their choice of partner species, or whether they are strongly or weakly dependent on each other. This particular combination of assumptions allows cooperative solutions for mutualism, including Tit-for-Tat and its successors (Dugatkin, 1997). Theorists have more recently recognized the artificiality of some of these assumptions, however, and have altered them to generate alternative conditions for the evolutionary stability of mutualism. Our goal here is not to evaluate the newer modeling approaches (e.g., Doebeli and Knowlton, 1998; Ferrie`re et al., 2002; Roberts and Sherratt, 1998), but to describe empirical examples of mutualism according to the list of parameters identified below. Note that each of the following 12 parameters may have several potential states. We list possible states for all parameters based on the empirical examples that we discuss in this article in Table I, and briefly describe them below. We recognize that this list may be incomplete, since in many cases other parameter states seem possible, at least in principle. Dependency: Mutualisms vary greatly in how crucial they are to survival and reproduction of each partner. We distinguish between ‘‘high’’ dependency for obligate mutualisms and ‘‘low’’ dependency for facultative mutualisms (those in which individuals can survive and reproduce at some level without mutualist partners). Dependency can be mutually high, mutually low, or asymmetric. Specificity: In some mutualisms, single partner species are matched (‘‘high specificity’’), while in other systems, several partner species can function as mutualists (‘‘low specificity’’). Again, specificity can be mutually high, mutually low, or asymmetric. Note that specificity may be low even if one partner is very dependent on mutualistic interactions (i.e., in cleaning mutualisms).

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

67

N interactions: The number of interactions between two individual partners in their lifetimes can be mutually repeated, mutually single, or asymmetric. Offer produced: Mutualisms are based on the provision of goods and/or services to partners; we refer to these as ‘‘offers.’’ Note that offers are not necessarily investments, as some offers in mutualisms are cost-free. Of major interest to us is that offers can be produced before the interaction, during the interaction, or even after the interaction. In addition, individuals may produce an initial offer prior to the interaction, but produce more during the interaction. Consequently, there are many parameter state combinations possible for the two partners (before–during, during–during, etc.). Moves: The two players can make decisions about their behavior simultaneously or not. We refer to these decisions as ‘‘moves.’’ When moves are not simultaneous, we use the term ‘‘alternating moves’’ if there are several rounds of interactions between two individuals, and the term ‘‘sequential moves’’ if there is only one round of interactions with a fixed sequence of decisions. Mobility: Individuals may be capable of moving around freely (although they may be constrained considerably through ecological limitations), or may be sessile. Mobility can be mutual, one-sided, or nonexistent. Active choice: This refers to the ability of individuals to seek out mutualists by choosing to visit a subset of all possible partners. It is usually but not always closely linked to mobility. Active choice can be mutual, one-sided, or nonexistent. Partner recognition: Recognition may be possible either because partners have cognitive abilities that allow them to recognize partners individually, or because one or both partners show site fidelity. In the absence of empirical evidence for the presence or absence of partner recognition (as is the case for most systems), it may be more appropriate to ask, when it seems feasible, whether or not partner recognition would yield advantages. Advantages of partner recognition may be mutual, one-sided, or nonexistent. Behavioral options: Whenever individuals invest in their partner, they are able to cooperate or to cheat and hence have the option to choose between these two behaviors. If a player lacks the option to cheat in a sense that would be meaningful (receiving benefits out of cheating the partner), it is unconditionally cooperative. Options may be symmetrical in that both partners can either cooperate or cheat [i.e., both are ‘‘conditional exploiters’’ in Bronstein’s (2001b) terminology], or asymmetrical in that only one partner has this strategic option. Finally, both players may lack the option to cheat. Investment: Players may or may not invest in their partner. As any investment is by definition costly, we call the act of investing altruistic. If

68

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

a player invests, this investment can be all or nothing (a player either cooperates or cheats) or variable, in that individuals must decide precisely how much they give to their partner. Payoff symmetry: This parameter is quite specific in that it needs evaluation only when both partners can cheat. In the absence of any quantitative knowledge on exact payoff values for any mutualism, we distinguish between ‘‘asymmetric payoff values’’ (e.g., costs of being cheated are small for one partner, but fatal for the other) and ‘‘symmetric payoff values,’’ in which both partners appear to experience more or less similar costs of being cheated. When only one partner can cheat, there is an obvious asymmetry. In cases where neither partner can cheat, this parameter is not relevant. Control over interaction: Partners may have ‘‘full control’’ over the interaction, in that they can withhold the reward/service that they are offering without any costs, can steal the reward/service, or can force it to be handed over. Both, one, or neither partner may have control over the interaction. In the following section, we review the mutualism literature and report on states of these 12 parameters for each of 8 well-studied forms of mutualism. However, one must keep in mind that these 12 parameters still do not represent the full complexity of mutualistic interactions. Each mutualistic system may have its specific additional features that are important for a thorough understanding of individual behavior (and hence the outcome of the interaction). We list such additional features for each system as well. VI. Evaluation of the Literature We have organized our presentation of individual mutualisms according to the three widely recognized classes of benefits mutualists receive: transportation, protection, and nutrition (Boucher et al., 1982; Bronstein, 2001a). We briefly describe the important features of each system. To facilitate comparison of the sections, we present the parameters in the order in which they are listed in Table I. A. Transportation Mutualisms In transportation mutualisms, one partner offers the other a commodity, usually food, in exchange for transport of itself or its gametes. 1. Pollination In pollination mutualisms, plants offer a resource (usually nectar or pollen as food) in exchange for transport of pollen and hence pollination of the flowers. Most pollinator species are insects, but birds and mammals,

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

69

especially bats and primates, may also function as pollinators. The crucial issues about pollination mutualisms that game theory may be able to address are why plant investment in nectar apparently does not usually drop toward zero (it is well known that it has in some systems), and under what conditions individuals of some species shift between pollination and cheating (nectar-robbing) behaviors. 1. Dependency: Dependency can be mutually high, mutually low, or asymmetric. Certain plants (not all: some are wind-pollinated) absolutely require pollinators if they are to outbreed (although self-fertilization may often be possible). Conversely, pollinators often rely heavily or exclusively on floral resources as food. When specificity is low, however, there may be very little dependency on a particular partner species. 2. Specificity: Like dependency, specificity varies from mutually low to mutually high to asymmetric. Certain plants can be pollinated only by a single pollinator species that can obtain resources only from that plant; some of these are discussed in the following section. Most plant–pollinator interactions are considerably less specific than this, however. An increasing number of pollination mutualisms are being found to be distincly asymmetric in specificity (Va´zquez and Simberloff, 2002). 3. N interactions: Interactions may be repeated or one-off. The situation is sometimes more complex than this, however, as an individual pollinator may interact only once with a particular flower but several times with the same plant. Conversely, a plant may interact only once with an individual pollinator, but with many individuals from a single colony of Hymenoptera. Hence, the question arises whether one should look at individual flowers or plants and at individual insects or colonies. 4. Offer produced: Plants produce their offer prior to the interaction. In contrast, pollinators make their offer during the interaction (they deposit pollen, and/or collect pollen, in varying amounts) and after the interaction (they may deposit the pollen they have collected on conspecific flowers or heterospecific ones; in the latter case, the pollen is wasted). 5. Moves: In principle, moves are sequential in that the plant puts out an offer and then the pollinator decides whether or not to visit. This simple view becomes complicated through the possibility that several pollinators may visit the same flower, in which case the amount of

70

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

nectar in a flower may partially reflect how much prior visitors left behind rather than exactly what the plant offered. 6. Mobility: The pollinators are mobile, whereas plants are not. 7. Active choice: The pollinators can actively choose which plant/flower to visit, whereas the plant has no control (in a behavioral sense) over who visits. Flowers may have features that exclude certain pollinator species or individuals of a certain size, shape, or behavior, but this does not translate into active choice on the part of plants during the course of an individual flower visit. 8. Partner recognition: If interactions are repeated, it may pay the pollinator to recognize partners (‘‘site recognition’’). Indeed, certain bees pheromone-mark flowers they have visited, and avoid revisiting those flowers (e.g., Giurfa and Nunez, 1992). 9. Behavioral options: Pollinator species usually cooperate by default, as they bring in pollen from plants visited previously for food, and pollen is usually collected passively. However, there are certain pollinator species in which individuals can choose to cheat by robbing nectar without collecting or depositing pollen (Irwin et al., 2001), a strategy that may save time. Conversely, plants may invest more or less energy into the production of food for their pollinators, and low (or zero) investment can be called cheating. 10. Investment: Whenever pollinators passively transfer and collect pollen, questions about investment into the partner usually do not apply. In species in which individuals can alternatively pollinate or rob, investment in the mutualism is an all (visiting the regular way and collecting/depositing pollen) or nothing (bypassing the floral sexual organs to get directly to the nectar) phenomenon. As mentioned above, investment of plants into their partners is variable. 11. Payoff symmetry: In pollination interactions with passive pollen transfer, cheating opportunities are asymmetric (only the plant can reduce investment). When nectar robbing is possible, both partners may cooperate or cheat. For a pollinator, being cheated by a single flower probably inflicts little cost, although marking of empty flowers and the ability to learn to avoid entirely nectarless plants suggests that significant costs of fruitless visits must exist. Nectar robbing can inflict costs to plants that range from high to low to nonexistent (Maloof and Inouye, 2000).

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

71

12. Control over interaction: Individual plants may control their investment into nectar at ecological time scales (Castellanos et al., 2002). More generally, however, nectar production most likely reflects evolutionary history and is thus quite inflexible. Some pollinators may access the food source without pollinating. Some pollinators may be able to preassess the food content of flowers and avoid empty flowers. Except for the very few active pollination systems (see the next section), they do not withhold pollen or avoid collecting pollen if the flower is empty, although less pollen may be transferred if the visit is shorter. Important features of pollinator–plant mutualisms not covered by our game structure assessment: (1) When pollinators have access to alternative food sources, whether interacting with the flowering plants is beneficial depends on the quality of these alternatives; (2) a pollinator may encounter empty flowers not only because some plants invest little into nectar production, but also because the flower may have been visited recently. The effect on the pollinator is the same, however; and (3) whether interactions are repeated or not may depend on how an interaction is defined. The two crucial problems are whether an insect visiting multiple flowers of a single plant during one inspection interacts repeatedly with the plant, and whether plants pollinated by social insects interact with individual insects or with the colony, in which more foragers may be recruited through communication. 2. Pollinating Seed Parasite Mutualisms In this small subset of pollination mutualisms (reviewed by Dufay¨ and Anstett, 2003), insects pollinate plants and simultaneously lay their eggs in or near the flowers; the larvae eat some of the developing seeds. The fig–fig wasp and yucca–yucca moth interactions are the best-known pollinating seed parasite mutualisms, although several similar but independently evolved interactions have been discovered. These include the senita cactus–senita moth and globeflower–Chiastocheta fly interactions. The key questions for these mutualisms include what limits the number of eggs laid per female pollinator, how variation in population density of the pollinators affects the mutualistic outcome, how plants can cope with destructively high pollinator densities, and why a few of these plant species have evolved mechanisms to (completely) prevent oviposition by the insects, while most have not. 1. Dependency: These mutualisms are obligate for both partners. The plants can be pollinated only by these insects, and the insects can lay their eggs nowhere else.

72

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

2. Specificity: The yucca, senita, and fig systems show very high specificity, in that plant species and insect species are almost matched one to one (but see Molbo et al., 2003). For Trollius, Chiastocheta flies appear specific to a particular plant species, while the plant may be pollinated by several species of Chiastocheta. 3. N interactions: Interactions are usually one-off between fig wasps and figs, since an inflorescence is usually visited only once by an individual insect. However, in other pollinating seed parasite mutualisms, individual insects will visit more than one inflorescence on the same or on different plants. 4. Offer produced: As in other pollination mutualisms, plants produce their offer (an oviposition site, in this case) prior to the interaction, while the insects provide their offer during the interaction. The insects’ offer has a benefit component and a cost component: (a) how many flowers they pollinate, and (b) how many eggs they lay. 5. Moves: Moves are sequential: first the flower makes an offer, then the insect makes hers. Plants may make a further move later, through random (senita) or selective (yucca) fruit abortion (Holland et al., 2004). 6. Mobility: Plants are immobile, and the insects visit them. 7. Active choice: This varies across different pollinating seed parasite mutualisms. Fig wasps apparently do not compare among inflorescences; they may enter the first one that they encounter. In the process, they lose their wings and rarely leave. Thus, the wasps are unlikely to exert active choice at the level of the inflorescence (it is possible that they do make choices among flowers within it, however). In contrast, yucca moths, senita moths, and Chiastocheta flies visit several flowers, and at least have the potential to make choices among them. 8. Partner recognition: Partner recognition is unlikely to play a role in these mutualisms, although in some cases (e.g., Huth and Pellmyr, 1999), the insect has mechanisms to avoid visiting the same flower twice. 9. Behavioral options: The plants have limited options to cheat their pollinators, although some yuccas may kill the offspring of pollinators before they begin to feed (Bao and Addicott, 1998) and female fig trees (in the dioecious fig species) prevent fig wasp oviposition and thus obtain the benefits of pollination without paying a cost for it. The

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

73

insects can cheat in two possible ways: (a) they may not pollinate, and (b) they may lay more eggs than average. 10. Investment: In the dioecious fig species, cheating is all or nothing: female trees prevent fig wasp oviposition while males allow it (Kjellberg et al., 1987). Insects show variable investment, which could be expressed as the ratio of pollinating acts to eggs laid. 11. Payoff symmetry: In systems in which the plant cannot cheat, the strategy set is asymmetric in that only the insect can cheat (hence the payoff matrix is asymmetric). In the dioecious fig species, where both individual plants and insects may cheat, the payoffs are asymmetric, in that the trees (which reproduce many times in their lives) would lose little from a wasp that cheated, while the wasp dies without reproducing if it enters an inflorescence on a female fig tree. 12. Control over interaction: The pollinators have full control over the interaction, in that they come to the plant and set the conditions of the game by deciding how many flowers to pollinate and how many eggs to lay. Plants may have morphological adaptations that make ovipositing in some flowers more difficult. Important features of pollinating seed parasite mutualisms not covered by our game structure assessment: (1) Fig trees and yuccas make many flowers in each reproductive episode, whereas Trollius usually bears only a single flower at a time; (2) all plants have multiple reproductive episodes, whereas the insects have only one; (3) Trollius also offers nectar, not only oviposition space, for its mutualists. Hence, male as well as female flies visit the flowers, and may in fact be responsible for most of the pollination (Despre´s, 2003); (4) fig wasps collect and move the pollen in the inflorescence in which they develop. Any emerging female wasp thus contributes to the fitness of the very same tree whose female function was reduced by the larval seed predators. In contrast, other pollinating seed parasites do not collect pollen from the flower in which they matured (Addicott et al., 1990); (5) there is a temporal component to how cooperative the pollinators are that has nothing to do with the pollinators’ behavioral strategies (Law et al., 2001). The first individual to arrive at a flower must cooperate by pollinating, or else her offspring will starve (unless a second female visits and pollinates). Later female visitors may not contribute much more to plant fitness, as pollination has already occurred and may be sufficient to permit initiation of every seed (Bronstein, 2001c); however, they may lay some additional eggs, which decreases plant fitness; and (6) plants may abort fruits, either selectively (ones with many eggs; yucca) or unselectively (senita). This means that although plants cannot control directly what

74

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

insects are doing (see point 12; control over interaction), they may respond to insect behavior at later stages. 3. Seed Dispersal In these mutualisms, plants provide food in exchange for transport of seeds (reviewed by Herrera, 2002). Typical seed dispersers are birds, mammals, and ants. The plant benefits from the great mobility of the fruit/seed consumer that results in seeds being removed from the maternal plant. Seeds are sometimes moved long distances and/or to superior germination spots. The animal gets a meal in return. With respect to this article, the major goal is to understand what factors determine plant investment in fruit flesh. 1. Dependency: Partners are moderately dependent on each other. Plants may require seed dispersal for their seeds to have any chance to germinate, although undispersed seeds may have a low likelihood of success. Most disperser species have a wide range of food sources, although at certain times of year they may be heavily dependent on fruit. 2. Specificity: Specificity is generally low. Certain plants may rely on a given class of animals as seed dispersers (e.g., small birds, or seedharvesting ants), but in only very rare cases is specificity higher than this. Similarly, almost no animals rely on fruits of a single species for food. 3. N interactions: Interactions may often be repeated but could also be one-off: an individual fruit usually has only a single interaction—the interaction in which it gets eaten. But almost all plants bear multiple ripe fruits at once, and most animals visit the same plant repeatedly for food. 4. Offer produced: The plants produce their offer prior to the interaction, whereas the seed dispersers produce theirs during the interaction (when they select a fruit) and afterward (when they drop or defecate the seeds). 5. Moves: Moves may be termed sequential (first the plant, then the seed disperser makes a move). 6. Mobility: The plants are sessile, whereas the dispersers are mobile. 7. Active choice: The plant cannot choose (in a behavioral sense) who eats the fruits, although, as in plant–pollinator interactions, selection may have shaped the subset of animals that are attracted to and rewarded by the fruit. Seed dispersers have the potential to compare

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

75

and actively choose among several plants, although the extent to which they actually do so is little known. 8. Partner recognition: Plants lack this ability, whereas the seed dispersers have the potential to remember plants with large fruit crops and with fruits in the correct stage of ripeness. 9. Behavioral options: Plants could potentially cheat by reducing their effort in the production of fruit flesh. Seed dispersers would appear to lack the option to cheat. 10. Investment: Investment of the plants is variable; among other things, it depends on the ratio of fruit flesh to seeds. Seed disperser species do not invest in the interaction; presumably, they defecate whenever they must. 11. Payoff symmetry: As only the plant invests, the question of how mutual cheating affects each partner’s fitness does not apply. For the dispersers, picking fruits with little flesh probably bears little cost. 12. Control over interaction: The seed dispersers have full control, in that they choose the fruits they prefer. Plants would appear to have little control over where the seed disperser will defecate, although the recent discovery of laxatives in certain fruits (Murray et al., 1994) suggests some control over when they will do so (and, hence, where they may be when they defecate).

B. Protection Mutualisms In this category of mutualisms, one species offers its partner some form of protection from the abiotic environment or natural enemies. In return, the other partner receives either a food reward (e.g., in cleaning and ant protection mutualisms) or reciprocal protection (in group foraging and Mu¨llerian mimicry associations). 1. Cleaning Mutualism In cleaning mutualisms, a ‘‘cleaner’’ species benefits by obtaining access to a food source, while a ‘‘client’’ species has its ectoparasites and possibly dead or infected tissue removed (see reviews by Coˆte´, 2000; Losey et al., 1999). Cleaning mutualisms involve a very diverse group of animals. On the cleaner side, there are shrimps, fish, and birds. Clients can be fish, turtles, crocodiles, and various mammals. The key questions in understanding cleaning mutualisms are how the conflict between cleaner and client over

76

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

what the cleaner should eat is resolved, and how cleaners avoid being eaten by predatory clients. With respect to our parameters, cleaning mutualisms have the following features. 1. Dependency: Clients generally show low dependency, in that they survive long periods without being cleaned. Most cleaner species described so far are also relatively independent of their clients in that they eat plenty of alternative food items under natural conditions. However, a few cleaner fish of the genus Labroides and Elacatinus, and possibly some shrimp species, accrue 80% or more of their diet from cleaning interaction (Coˆte´, 2000). 2. Specificity: Partner species usually show very low specificity: each cleaner species interacts with several client species and vice versa (although clients are less specific to cleaners than cleaners are to clients). 3. N interactions: Depending on the home range sizes of the pair of species involved, interactions between individuals can be repeated or one-off. 4. Offer produced: Offers are produced during the interaction. Neither cleaner nor client can preassess how its partner will behave. The cleaner ‘‘produces’’ a certain level of service (i.e., parasite removal) during the interaction; similarly, during the interaction, the client makes a decision on how long to interact with the cleaner, and in the case of predatory clients whether it will try to eat it. 5. Moves: Moves are made simultaneously. 6. Mobility: Cleaner shrimps and some cleaner fish show very limited mobility, while other cleaner fish and birds are quite mobile. Similarly, there is great variation among client species in their mobility. 7. Active choice: The relatively immobile cleaner shrimps and fish mentioned above lack the option to actively visit clients and start interactions. Nevertheless, they are sometimes able to choose, when two or more clients seek their service simultaneously. Similarly, client species with small home ranges or territories usually have at best access to one cleaning station and hence cannot choose between cleaners unless they have cleaner fish and cleaner shrimp present, which happens occasionally (R. Bshary, unpublished observation). Cleaner species and client species with large home ranges, however, are in a position to actively seek the partners they prefer.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

77

8. Partner recognition: Individual recognition of clients has been shown experimentally for the cleaner wrasse Labroides dimidiatus (Tebbich et al., 2002). In this system and in some other cases, individual cleaners show strong site fidelity and hence may be recognized through location. Generally, individual recognition would be advantageous if the partners have the option to cheat and if interactions are repeated. 9. Behavioral options: It is known for some cleaner fish and bird species that they can cheat the clients by feeding on healthy client tissue (Randall, 1958; Weeks, 2000). Cleaner shrimps may lack this option, but this remains to be confirmed. Two categories of clients need to be distinguished (Bshary, 2001). Clients that are potential predators of their cleaners could cheat by eating their cleaner, whereas nonpredatory clients have no means to exploit a cleaner. 10. Investment: The investment of cleaners with cheating abilities into their clients is variable (in the sense that the rate of cheating bites to cooperative bites is variable). The investment of predators is all or nothing: they either cooperate or try to eat the cleaner. 11. Payoff symmetry: In most cases, only the cleaner is able to cheat, hence the payoff matrix is asymmetric. An exception are interactions between cleaners with cheating abilities and predatory clients. Here, the payoffs are asymmetric, as a cheated predator would lose a bit of healthy tissue while the cleaner would lose its life. 12. Control over interaction: Cleaners have full control over the service quality they provide. Clients can only respond to cleaner fish behavior. Nevertheless, most client species have high control, in that they can terminate interactions immediately whenever they decide to. However, some species may move slowly relative to cleaners and therefore have difficulties avoiding a cleaner that wants to interact with them. Important features of cleaning mutualisms not covered by our game structure assessment: (1) The benefits that cleaners can offer to the clients depend critically on the population dynamics of other species, namely the parasites. If parasite abundance is low, cleaners cannot provide great benefits to their clients (Grutter, 1997); and (2) cleaner-to-client ratios may determine whether or not cleaners get sufficient food from their interactions with clients. If cleaners are very abundant, facultative cleaners might switch to other food sources, while full-time cleaners may cheat more frequently by feeding on healthy client tissue. Finally, territorial clients might face significant costs when visiting cleaning stations means

78

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

leaving their territory, due to risk of territory loss and aggression by conspecifics (Cheney and Coˆte´, 2001). 2. Ant Protection Mutualisms In these mutualisms, ants provide protection against natural enemies, while the partners provide a food source, and in certain cases, shelter. Wellstudied examples include mutualisms between ants and lycaenid caterpillars (Pierce et al., 2002), ants and a variety of Homoptera, including aphids and treehoppers (Buckley, 1987), and ants and plants (Heil and McKey, 2003). Protection results either when ants chase away predators/herbivores, or when they actively consume them. Food sources include excretions (in the case of aphids), glandular secretions (lycaenids), and extrafloral nectar and lipid-rich food bodies (plants). The key questions for understanding ant protection mutualisms are why ants protect their partners at all, why insect mutualists are not eaten by the ants, and what keeps the production of food and shelter by the tended partners as high as it is. 1. Dependency: There is a continuum of dependence, from species that cannot survive unless ant tended, to species whose success is only marginally increased by tending. Ants usually have food sources in addition to what they receive from their partners. Only in certain highly specialized ant–‘‘myrmecophytic’’ plant mutualisms (e.g., the association between neotropical Acacia species and Pseudomyrmex ants) are the ants highly dependent on their partners. 2. Specificity: Ants and myrmecophytic plants, as well as ants and some lycaenid species, form relatively species-specific relationships; specificity in most other ant protection systems is low. 3. N interactions: Individuals interact repeatedly with each other. 4. Offer produced: Offers are generally produced continuously during the interaction, although partners in need of protection may produce an initial offer before ants are present, in order to attract them. 5. Moves: It is difficult to apply the terms ‘‘simultaneous’’ or ‘‘sequential’’ to ant protection mutualisms. The partner species continually invest in the ants by providing food, although they may modulate the amount and quality in relation to their need for protection. Ants, on the other hand, do not do anything beneficial for their partners most of the time. Benefits accrue only when the partner is at risk of attack. 6. Mobility: Partner species move little or not at all, while the ants are generally highly mobile. Exceptions are ant species associated with myrmecophytic plants that occupy nest space on or in the plant.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

79

Hence, they have strongly reduced mobility compared with most ant protectors. 7. Active choice: The tended species appear to have no ability for active choice, while ants usually have access to several partners. Again, the exceptions are ant species associated with myrmecophytic plants. Here, the ant queen could initially choose a plant on which to settle, but once the colony grows, movement to other plants and hence the ability to actively choose among partners may be constrained. 8. Partner recognition: Ants may be able to recognize individual partners through site recognition (mediated via trail-marking pheromones), since most ant-protected species are immobile or nearly so. Partners seem highly unlikely to be able to recognize individual ants. 9. Behavioral options: Ants could in principle cooperate or cheat, in one of two ways: they could avoid risks associated with defending their partners, or they could actively consume them (relevant for Homoptera and some lycaenid species). Tended partners may cheat by reducing the amount or quality of food or shelter that they provide to ants. This option is limited in aphids, as aphids honeydew is mainly an excretion that cannot be stopped completely. 10. Investment: Investment is usually variable on both sides of the interaction: both protection effort and food production can be adjusted (with the possible exception of Homoptera). In the case of possible ant predation of partners, investment is all or nothing. 11. Payoff symmetry: The effects of reduced investment in the partner are asymmetrical in ant–insect protection. If ants do not defend, predators/parasites can kill the insect, whereas ants just lose a bit of food if cheated. The payoffs are more symmetrical in ant–plant mutualisms, since the cost of lack of defense by a single ant individual is a marginally higher rate of herbivory for the plant rather than a matter of life and death. 12. Control over interaction: Ants usually have full control over their own behavior: they can presumably choose how much protection they give, and they can quickly adjust to environmental conditions. Tended partners could in principle control to some extent how much food they offer; Homoptera may have limited control over what they offer, although there is some evidence that they control its chemistry (Fischer and Shingleton, 2001). A major problem might be that tended partners cannot respond instantaneously to changes in their most important variable in the environment, namely predator attack.

80

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

However, the ability of lycaenids to alter secretion rates according to their perceived need for protection, and of some lycaenids and plants to increase secretion on attack, may reduce subsequent attack rates (Axe´n et al., 1996; Leimar and Axe´n, 1993; Ness, 2003). Important features of ant protection mutualisms not covered by our game structure assessment: (1) Most importantly, the magnitude of the benefits ants can offer to their partners depends crucially on the identity and density of the partners’ natural enemies. If there are no enemies, ant tending may confer no benefit (but see Morales, 2000); and (2) the benefits that the partner species provide depend on the identity and quality of alternative food sources available to the ants. Under some circumstances, partner species may not be able to provide food of high enough quality to make tending a profitable option for ants. 3. Mixed Species Aggregations Individuals or groups of different species of vertebrates aggregate for some or all of their lives. Detailed studies have investigated mixed species associations in forest primates (Ho¨ner et al., 1997), ungulates (FitzGibbon, 1990), a large variety of bird species (Moynihan, 1962), and fish (Ehrlich and Ehrlich, 1973). For most of these interspecific associations, it is assumed that a reduction in predation risk is the primary advantage. There is good evidence for this claim in birds and mammals (Bshary and Noe¨, 1997; Fitzgibbon, 1990; Greig-Smith, 1981; Noe¨ and Bshary, 1997). Potential mechanisms facilitating increased protection include dilution effects, confusion effects, increased early warning, and improved defense. Increasing group size with allospecifics rather than conspecifics may reduce competition over food or mating partners. It may also lead to the joining of complementary skills in predator avoidance, or may reduce predation risk if the partner species is ‘‘preferred’’ by predators (references in Noe¨ and Bshary, 1997). Increased foraging efficiency for individuals of one or both partner species is an alternative hypothesis to explain mixed species associations (references in Cords, 1987). 1. Dependency: Dependency in these associations is probably low, and each species could survive without its partner. Nevertheless, there will often be an asymmetry, in that one partner species benefits more from the association than the other partner species. 2. Specificity: Associations can be quite partner specific (e.g., olive colobus seek Diana monkeys at Tiwai Island; Whitesides, 1989). Often, however, several partner species are involved in the association. Still, some species may be better partner species than others because of their vigilance abilities, active predator defense, or passive food provisioning.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

81

3. N interactions: Interactions are usually repeated between the same individuals or groups. In some bird flocks or fish schools, individuals may only meet once, although this has rarely been documented. 4. Offer produced: Offers here mean that (a) the presence of each individual adds to the safety for all other individuals through early warning, dilution, or confusion effects, and (b) one partner species facilitates access to a food source for the other one in association when foraging benefits exist. Offers are produced during the interaction. 5. Moves: This term is difficult to apply to interspecific associations. Moves may be termed simultaneous in that everybody responds to the presence of a predator. If access to food is traded for increased protection (as possibly in some marine associations), food may be accessed continuously while predatory attacks are unpredictable. 6. Mobility: Partners are generally mobile. 7. Active choice: Because of their mobility, individuals may often be able to actively choose with whom to associate. Territoriality and living in stable groups may constrain active choice options in some species, however. 8. Partner recognition: Partner recognition below the species or sometimes group level appears not to be necessary unless it can be shown that cheating the partner species is a problem in interspecific interactions (see below). 9. Behavioral options: Cheating appears not to be an option within these interactions. The benefits of mixed species associations are usually an emergent property of the increase in group size: improved early detection of predators due to ‘‘many eyes and ears,’’ dilution effects, and confusion effects. Note, however, that there may be individual differences with respect to the likelihood of giving alarm calls, although these differences are seen as intraspecific strategies (Sherman, 1977). 10. Investment: There is no investment that benefits the partner directly, although there are costs of staying together and synchronizing activity patterns. 11. Payoff symmetry: As there is no cheating, the payoff symmetry is of no concern. 12. Control over interaction: As mentioned above, the benefits are a result of simply associating together. No active exchange between partners

82

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

occurs. Hence, every participating partner gets the increased protection by default. Important feature of mixed species associations not covered by our assessment of game structures: The costs and benefits of mixed species associations depend on group size. Costs of being in a group larger than some optimal level cannot be attributed to the behavior of individuals. Individuals do not cheat, but competition for food increases with the number of individuals present. Population densities may hence become a crucial parameter. 4. Mu¨llerian Mimicry In Mu¨llerian mimicry, two or more partner species (1) strongly resemble each other in colors and morphology, and (2) are to some degree unpalatable or poisonous (Ruxton et al., 2004; Speed, 1999). Individuals of partner species do not interact directly with each other. On exposure to the same potential predator species, look-alikes benefit from the dilution effects associated with individual predators learning to avoid prey with certain characteristics (color, smell). The best-studied Mu¨llerian mimicry complexes involve Lepidoptera. The crucial question is what prevents individuals of partner species from reducing the unpalatable or poisonous products in their body and becoming Batesian mimics that resemble unpalatable species but are harmless. 1. Dependency: Dependency is probably mutually low: each species can survive without the partner species, although fitness may be higher in its presence. 2. Specificity: All species that resemble each other, are unpalatable, and occur in the same place should be good partners. The number of partner species may therefore range from one to many. 3. N interactions: Individuals do not interact directly with each other. Potential predators, however, may repeatedly interact with (i.e., encounter) individuals of all species involved. 4. Offer produced: Nothing is exchanged between mimicry species. Hence, the ‘‘offer,’’ if it can be said to exist at all, is the degree of unpalatability, which is produced continuously. 5. Moves: This parameter is difficult to apply to Mu¨llerian mimicry. Moves can be termed simultaneous in that all individuals involved may be encountered by a predator at any time. 6. Mobility: Partner species are mobile.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

83

7. Active choice: There is usually no potential for active choice in Mu¨llerian mimicry, since mimics do not appear to intentionally flock together. 8. Partner recognition: Partners do not need to recognize each other in order to share benefits, and it is unlikely that they do. 9. Behavioral options: The crucial questions are whether (a) the unpalatable substances are costly to produce, and (b) if they are costly, whether the exact investment of individuals depends critically on what individuals of partner species do rather than on what conspecifics do. For example, kin selection and low dispersal rates may stabilize unpalatability independently of the existence of a Mu¨llerian look-alike. If the answer to both questions is ‘‘yes,’’ then cheating is possible. With respect to production, there is evidence in some systems that sequestering secondary compounds from food is costly (Ruxton et al., 2004). 10. Investment: If unpalatability is costly, investment is variable in that the production or incorporation of compounds is variable. But as mentioned above, it is critical to know whether investment is a game between conspecifics or between all look-alikes. 11. Payoff symmetry: The payoffs should be symmetrical, as costs and benefits to each partner are similar. 12. Control over interaction: Individuals do not directly interact with each other. Each individual may control to some extent its production or sequestration of toxic compounds; however, it has no control over what other individuals are doing. Important features of Mullerian mimicry not covered by our assessment of game structures: (1) Population densities of partners may be crucial to understand how a reduction in unpalatability in one species would affect predator behavior and hence the fitness of all Mu¨llerian mimic species; and (2) as mentioned above, dispersal patterns might be another important variable. If dispersal rates are low, then kin selection advantages associated with predator defense may override any advantages of reduced investment. In contrast, if dispersal rates are high, a reduction in investment may be advantageous. C. Nutrition Mutualisms Until this point, we have focused on nonsymbiotic mutualisms, simply because more relevant ecological data are available to infer their game structures. In this section, we deal relatively briefly with a suite of

84

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

mutualistic symbioses based on the benefit of nutrition (reviewed by Douglas, 1994): nitrogen-fixing symbioses, plant–mycorrhizal mutualisms, coral–zooxanthellae associations, and lichens (mutualistic symbioses between fungus and algae). In these interactions, one partner acts as host to a microbial associate that lives in or on it. The two partners trade substances that are either of nutritional value or essential for physiological activities. Shelter offered by the host may be an additional offer. The crucial question in symbioses is how the nutrient transfer between partners is regulated. We stress that our parameter descriptions below are somewhat conjectural, in light of limited study of some of the critical features of these interactions. 1. Dependency: In most cases, hosts are highly dependent for survival and/or growth on their symbionts. Certain symbiont species are known to occur only in association with hosts, but in general, information is lacking on this point. 2. Specificity: While it was once assumed that nutritional symbioses exhibit low partner specificity, more recent molecular evidence suggests higher or even complete specificity, at least in vertically transmitted symbioses (Herre et al., 1999). 3. N interactions: There is usually one prolonged interaction between individual partners during which partners may alter their behavior repeatedly. From a theoretical point of view, the interaction can therefore be seen as repeated. On the brief end, some corals readjust the number of algae on a daily basis; on the long end, the interactions last a lifetime in lichens. Things become even more complex as there can be long-term associations between cell lineages, not just individuals. 4. Offer produced: Offers are produced continuously by both partners. Hosts sometimes must invest first, however, by producing a structure in which the symbiont can live. 5. Moves: Moves are simultaneous, and both partners may continuously provide nutrients to each other. 6. Mobility: In many symbioses, both partners are sessile and cannot move freely. Zooxanthellae, however, move in the water column. 7. Active choice: There is no choice possible in vertically transmitted symbioses (i.e., those passed internally between host generations). With regard to horizontally transmitted symbioses, at least some marine symbionts actively locate hosts, usually via chemical detection (Douglas, 1994).

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

85

8. Partner recognition: In this context, partner recognition refers to being able to assess the amount of nutrient received from different partners. Partner recognition of some form would be advantageous in all cases in which one individual interacts with several partner individuals, especially hosts associated with several symbiont lineages that may confer greater or lesser benefits. In mycorrhizae, an individual fungal genotype may be in contact simultaneously with several plants; in this case, both partners could benefit from partner recognition. 9. Behavioral options: Both sides invest in their partner, so both sides can in principle cooperate or cheat. A variety of forms of cheating have in fact been detected in both hosts and symbionts (e.g., Kiers et al., 2003; Smith and Smith, 1996). 10. Investment: The investment into the partner is variable rather than all or nothing. 11. Payoff symmetry: The payoffs may be relatively symmetrical in most systems, as being cheated usually means not receiving nutrients from the partner. However, being cheated may have greater consequences for a unicellular symbiont compared with its multicellular host. The payoffs in corals can also be asymmetric, as some polyps can eat their zooxanthellae. 12. Control over interaction: This point is of crucial importance for the outcome of the game, but is still unresolved. In corals, it is hypothesized that polyps have some control over nutrient flow, as the algae are placed into a host vacuole. Polyps may also expel the algae (Titlyanov et al., 1996). In nitrogen-fixing symbioses, plants must offer an initial investment that may or may not be reciprocated later on. During later stages of the interactions, evidence suggests that there is a complex amino acid transfer between plant and rhizobia that prevents either partner from dominating the interaction (Kiers et al., 2003; Lodwig et al., 2003). Such information is essential for all symbioses. We need to know how well each partner can control how much it gives and how much its partner has to give in return. Important features of nutritional symbioses not covered by our game structure assessment: (1) There is one important variable that exists only in symbioses: do partners associate horizontally or vertically? Vertical transmission of symbionts has been shown empirically and theoretically to reduce the potential for conflict among partners; and (2) Hosts and symbionts usually have radically different generation times. How

86

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

generation times affect the outcome of interactions at evolutionary time scales has barely been addressed (Bergstrom and Lachmann, 2003).

VII. How Similar/Different Are Mutualisms? In the previous section, we reviewed the best-studied mutualistic systems in order to extract their game structures. A summary of our results is presented in Table II. We now attempt to assess the similarities between these interactions in the context of game structure. The importance of ecology for our understanding of behavioral strategies and their consequences on the fitness of each partner are evaluated later. One might expect that the huge diversity mutualisms exhibit with respect to evolutionary origin, species identity, and ecology would also be reflected in game structures. If so, it is likely that general principles cannot be found. We promote a more optimistic view, based on the game theoretic concepts that we presented in Section I. These concepts focus on the question of how mutualistic systems may be stable against erosion through the evolution of cheating strategies. Below, we review the main features of game structure that mutualisms share, even when differing greatly in natural history. A. Investment In the majority of mutualistic systems investigated in this article, one or both partners invest in each other. In most cases, investment is variable; for example, plants may produce greater or lesser amounts of nectar per flower, and pollinators may deposit and carry a variable number of pollen grains between plants. Exceptions are protectors and clients that can act as predators of their partners. In these cases, individual investment may be all or nothing (i.e., the partner is either allowed to live or is eaten), as the original prisoner’s dilemma game assumed (Axelrod and Hamilton, 1981). However, cheating by predators terminates the game and hence violates the iterated prisoner’s dilemma game in another important way (Hammerstein and Hoekstra, 1995). For all mutualisms in which investment occurs, one can ask what factors may stabilize cooperative behavior by disfavoring individuals that play a strategy of reduced investment in the partner. In other mutualisms, however, there appears to be no investment whatsoever. In particular, in mixed-species foraging associations, mutual benefits appear to be a simple by-product of group enlargement. Many other, less well-studied mutualisms similarly involve no investment. Examples include gabar goshawks placing webs with social spiders into their nest

TABLE II Summary of Evaluated Parameters That May Have Important Implications for the Outcome of Interspecific Interactionsa

Parameter

Plant– pollinator

Plant–seed parasites

Plant–seed dispersers

Cleaner– client

Ants– partner

Mixed species

Mimicry

Host– symbiont

Dependency Specificity N interactions Offer produced Moves Mobility Active choice Partner recognition Behavioral options Investment

High/div Low/low Variable Before/ during Seq No/yes No/yes No/var

High/high High/high One-off Before/ during Seq No/yes No/var No-pos/no

Low/low Low/low Variable Before/ after Seq No/yes No/yes No/pos

Div/low Low/low Variable During/ during Simult Var/var Var/var Var/var

Div/div Div/div Repeated During/ during Simult Var/no Var/no Pos/no

Low/low Variable Repeated During/ during Simult Yes/yes Both Pos or no

Low/low Variable dna Constant/ constant Simult Yes/yes dna No/no

High/high Often high Lasting During/ during Simult No/var No/no Pos/no

cd/cd or c

cd/cd

cd/c

cd/cd

c/c

Cont/no

Sym

dna

Life–din

dna

Sym

Var/full

No/var

No/full

Full/var

Cont, a–n/cont Sym or life–din Full/var

No/no

Payoff symmetry Control

Cont, a–n/cont Life-din

cd or c/cd or c Cont/cont

cd/cd

Cont/a–n

cd or c/cd or c Cont/a–n

Full/full

Full/full

a

Cont/cont Sym or life–din ?/?

If two parameter states are given, separated by ‘‘/,’’ the first information applies to the mutualist named first in row 1, and the second information applies to the mutualist named second in row 1. For example, in plant–pollinator mutualisms, dependency is high for the plants and diverse for the pollinators (some species are highly dependent whereas others are not). Abbreviations: div, diverse parameter states occur in the system; d n a, does not apply; simult, simultaneous; seq, sequential; pos, possible (the ability would yield fitness advantages); c, to cooperate is the only behavioral option; cd, individuals could cooperate or cheat, at least in theory; cont, continuous; a–n, all or nothing; life–din: one partner would lose little if cheated whereas the other partner would lose its life; sym, symmetrical; var, variable.

88

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

to reduce parasite densities on young (Henschel et al., 1992) and cooperative hunting between groupers and moray eels (Bshary et al., 2001). Since there is no (or minimal) investment, there is presumably nothing for these species to gain by cheating in the interaction; in fact, it is difficult even to define what cheating would consist of. There can be opportunity costs in these systems, that is groupers may spend a considerable amount of time (up to 60 min; R. Bshary, unpublished observation) signalling to the moray eel, trying to elicit joint hunting. But these costs do not translate into benefits for the moray eel. We do not deal with these mutualisms further in this article, as game theory does not appear to have much to offer as a tool to study their stability. Furthermore, we cannot discuss Mu¨llerian mimicry examples properly without knowing whether or not there is selection on partner species to become palatable Batesian mimics. In any case, individuals of partner species do not invest in each other in interactions and hence our game structures do not yield important insights. B. Investment at the Outset of the Interaction In a diverse range of mutualisms, one partner may have to produce an initial investment for the interaction to begin. Some species must produce a reward in order to attract a partner, including plants (floral nectar, extrafloral nectar, and fruits) and lycaenid caterpillars (secretions). Other mutualisms are only initiated once one partner has produced a shelter for the other to inhabit (e.g., myrmecophytic plants and reef-building corals). Finally, the cleaner wrasse L. dimidiatus must invest initially in its resident clients by providing tactile stimulation before it is allowed to search each client’s surface for parasites (Bshary, 2002). In all these cases, any potential benefit for the investing partner is delayed until after substantial investment has been made. Hence, the question of when the other partners will cheat by taking the offer without returning any benefits becomes particularly interesting. Such behaviors are well documented in most of these mutualisms. For example, nectar robbers take floral nectar while bypassing the pollen and stigmas; it is difficult to see how plants could sanction these behaviors in a behavioral sense. C. Partner Choice A common feature of mutualisms (with the exception of mutualistic symbioses involving vertical transmission) is that individuals of one species may have, at least theoretically, the option to choose among several individuals of their partner species. Pollinators and seed dispersers are usually mobile and visit several plants, many clients are mobile and visit different cleaner individuals, and ants tend a variety of individual partners, often belonging to several

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

89

partner species. In all these systems, the predictions of biological market theory (Hoeksema and Schwartz, 2001, 2003; Noe¨, 2001; Noe¨ and Hammerstein, 1994; Noe¨ et al., 1991) could be tested. Biological market theory treats mutualistic interactions as an exchange of goods between partners that differ in the degree of control they have over the goods they trade. Predictions based on market theory would be as follows. (1) The chosen partner species adjust their offers according to supply and demand ratios. If they have few competitors for access to partners, they should offer less than if there are many competitors, because in the latter situation, individuals of the partner species can be choosy; and (2) individuals of the choosy partner species should stop interacting with cheating individuals and search for more cooperative partners. Such active choice would stabilize the occurrence of cooperative behavior in the partner species (Ferrie`re et al., 2002). In one marine cleaning mutualism, Bshary and Scha¨ffer (2002) found that client species with large home ranges do choose between cleaners on the basis of the quality of the service that cleaners provide. More specifically, clients switch to a new cleaner if they are bitten by their current one, and return with high probability if the cleaner behaves cooperatively (Bshary and Scha¨ffer, 2002). In addition, partner choice options may affect an array of other parameters, at least in the L. dimidiatus cleaning mutualism (Bshary and Noe¨, 2003). However, evidence that choosiness of clients stabilizes cooperative behavior of cleaners is still lacking. While individuals need to move freely in order to be able to actively choose between potential partners, there is also a way of choosing that may be called ‘‘passive.’’ For example, in most mutualistic symbioses, each host interacts simultaneously with several genetically distinct symbiont lineages. If these symbionts vary in quality, any ability of the host to sanction unproductive symbionts would be favored by natural selection. Evidence in nitrogen-fixing symbioses between plants and Rhizobium bacteria shows that plants do in fact discriminate between nodules with different N2 production (Kiers et al., 2003; Lodwig et al., 2003). Similarly, some yucca plants selectively abort fruits that contain high numbers of pollinator eggs (Pellmyr and Huth, 1994). Like active choice, this passive choice arguably stabilizes cooperative behavior of the symbionts. Similar ‘‘sanctions’’ have been argued to exist in some other mutualisms as well (e.g., West et al., 2002). D. Dependency and Specificity Dependency on the partner is typically high in nutritional symbioses, pollinating seed parasite mutualisms, and interactions between ants and myrmecophytic plants. In addition, a few cleaner species and many plants

90

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

can survive or reproduce only when clients/pollinators visit them. Interestingly, high dependency does not necessarily translate into high partner specificity. In particular, cleaner fish and shrimps access many client species, and many self-incompatible plants can be pollinated by a wide range of small insects. On the other hand, partner specificity may be high in some mixed species and mimicry associations, although dependency is probably relatively low in these cases. Dependency appears to be a parameter that may affect mutualistic interactions in ways that go beyond the game structures we presented. With respect to game structures, dependency may simply affect payoff values: if one partner is very dependent on the other, the payoff values for receiving a cooperative act from the partner are higher if an individual is critically dependent on the partner’s cooperation than if the individual can survive well without the partner. But the degree of dependency may also affect interactions more deeply, through increased selection pressure favoring mutants with greater access to resources provided by the partner. The result can be seen as increased sophistication of behavioral strategies, but also as potential threats to the evolutionary persistence of cooperative behavior, as the following examples indicate: (1) only obligate cleaners of the genus Labroides are known to manipulate client decisions through tactile stimulation of the client with the pelvic and pectoral fins (Bshary and Wu¨rth, 2001); and (2) flower mimicry of female pollinators instead of nectar provisioning is found in systems with high pollinator specificity (Dafni, 1984). E. Similarities within Different Mutualistic Classes It is important to note that the similarities in game structures across mutualisms mentioned above are not necessarily linked to the particular commodities that are exchanged by partners, but are more general. As a consequence, for example, some ant-tending mutualisms may share more features with pollination mutualisms than with mixed species associations or Mu¨llerian mimicry, although the latter two are protection mutualisms as well. As an example, interactions between plants with extrafloral nectar and ants have a game structure similar to interactions between plants and pollinators that can rob nectar (Table III). On the other hand, there can be remarkable similarities between specific systems in which similar commodities are exchanged. A striking case is the comparison between cleaner wrasse L. dimidiatus–predatory client interactions and aphid–ant interactions (Table III). These specific examples emphasize that insights gained from one mutualistic system may shed light on other mutualisms that at first sight appear to be very different.

91

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

Generally, transportation mutualisms appear to be more coherent in game structure than protection mutualisms. The transportation mutualisms discussed here generally share the following features (Table II): they involve one sessile and one highly mobile partner, the sessile partner invests prior to the interaction and moves are therefore sequential, the mobile partner may be able to actively choose while the sessile partner cannot, and the sessile partner has little immediate control over the course of interaction, which is determined mainly by the behavior of the mobile partner. Among protection mutualisms, mimicry and mixed-species associations are very peculiar. Mimicry differs from all other examples because partners do not interact directly with each other. This generates a suite of consequences for the game structure; even the game’s terminology is difficult to apply (Table II). Mixed-species associations differ from all other examples in that partners do not invest in each other, with another suite of consequences for the game structure (Table II). TABLE III Specific Examples of Strong Similarity in Game Structures Between Mutualisms That May Initially Appear Different

Parameter Dependency Specificity N interactions Offer produced Moves Mobility Active choice Partner recognition Behavioral options Investment Payoff symmetry Control a

L. dimidiatus– predator

Aphid–ant

Plant–ant

Plant– bumblebee

High/low Low/low Repeated During/ during Simult No/yes No/yes Pos/pos

High/low Low/low Repeated During/ during Simult No/yes No/yes No/pos

High/low Low/low Repeated Before/ during Seq No/yes No/yes No/pos

High/high Low/low Variable Before/ during Seq No/yes No/yes No/pos

cd/cd

cd/cd

cd/cd

cd/cd

Cont/a–n Life–din

Cont/a–n Life–din

Cont/cont Sym

Cont/a–n Sym

Full/high

Low/full

No/full

No/full

The first comparison is between the mutualism of the cleaner wrasse L. dimidiatus and a predatory client with access to several cleaning stations (e.g., a giant moray eel, Gymnothorax javanicus), and an aphid–ant mutualism (both examples are of protection mutualisms). The second comparison is between the mutualism of a plant with extrafloral nectar and ants (protection mutualism), and the mutualism between a plant and bumble bees as pollinator (transport mutualism). Bumble bees have been chosen as a representative of a pollinator that can rob nectar. For abbreviations, see Table II.

92

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

The different types of mutualism that we have reviewed are quite diverse with respect to the degree of variation they exhibit in the states of individual parameters (Table II). Plant–seed disperser mutualisms appear to be very uniform, as are mixed-species associations. In Mu¨llerian mimicry cases, crucial information is lacking for a proper assessment. Other types of mutualisms show considerably more variation. For example, pollination mutualisms range from species-specific and obligate to highly generalized and facultative. Nevertheless, the number of game structures is not overwhelming. Cleaning mutualisms appear to be the most diverse system with respect to the number of game structures possible. But even here, the number of different game structures is limited to about 15 different combinations of parameter states. In each of the other systems, free combinations of observed parameter states yield less than 10 potential game structures. Hence, the diversity in game structures is not so large that it could not be tackled by theoreticians.

VIII. The Importance of Ecology The approach to mutualisms that we have taken here focuses on the question of what factors may promote the persistence of cooperative behavior and keep cheating at levels that do not threaten the overall mutualistic outcome. Ideally, to address this question theoretically, the behavioral options of individuals are identified, the payoffs for each behavioral option are determined, and the interaction is formalized as a game. If the game yields a solution in which both partners show cooperative behavior, the mutualism is assumed to be evolutionarily stable against cheating. However, both short-term and evolutionary stability may depend on additional, ecological variables that are not easily covered by game theoretic assessments. Some of these are as follows. A. The Influence of Population Densities We identified several mutualistic systems in which population dynamics of partner species are almost certainly of major importance in determining payoff values. For example, in pollinating seed parasite mutualisms, high pollinator densities mean that flowers/inflorescences will become heavily laden with eggs, raising the possibility that the larvae will eat most of the developing seeds. Pollinator females arriving after all flowers have been pollinated also lay eggs, contributing to the cost of the mutualism but adding no extra benefits. The investment of a pollinator in its partner, as measured by the ratio of pollination acts to eggs laid, is therefore determined strongly

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

93

by sequence effects (being the first or second or third visitor), not only by female strategies. The problem for the plants is that they appear to have little control over the course of the interaction; various mechanisms have been invoked that would give them control (e.g., Bao and Addicott, 1998; Ganeshaiah et al., 1995), although most of these have been challenged on empirical or theoretical grounds. Theoreticians are beginning to explore how pollinator densities may be kept low enough to yield a net benefit to the plants, focusing on fruit abortion mechanisms that have been reported in some systems (e.g., Holland and DeAngelis, 2002; Holland et al., 2001). In cleaning mutualisms, a high cleaner-to-client ratio means that cleaners will not find enough parasites on clients to meet their daily food intake. Some cleaner species, in particular facultative cleaner species, may switch to other food sources under such circumstances. Full-time cleaners, however, may behave more parasitically under these conditions and increase feeding on healthy client tissue, resulting in conflicts between client and cleaner. Finally, in Mu¨llerian mimicry systems, relative population sizes of species within the complex may determine whether individuals of a rarer species can benefit from a reduction in the production of toxic material (assuming that it is costly to produce). That is, the balance between cooperation and conflict in these interactions may be frequency dependent (Ruxton et al., 2004). A common view, however, is that systems do not shift back and forth between Mu¨llerian and Batesian mimicry (Mallet, 1999). B. Population Densities of Third Species In protection mutualisms, the benefits for the protected partner generally depend critically on the population dynamics of third species, specifically, the natural enemies that the protector consumes or deters (Bronstein and Barbosa, 2002). Cleaners and ants generally do not improve the fitness of their partners if the partners’ enemies are rare or absent. In cleaning mutualisms, low parasite densities mean that clients simply seek cleaning less frequently. Hence, as long as clients can control number and duration of interactions with cleaners, no negative payoffs are to be expected, although the benefits would shift toward zero under circumstances of extreme parasite shortage (Johnstone and Bshary, 2002). The concept of ‘‘power’’ in the economic literature is very similar to what we have called ‘‘control over interaction’’ (Bowles and Hammerstein, 2003). Ant mutualisms differ from cleaning mutualisms in that partners are usually producing at least some reward for ants even when they do not require attendance. Therefore, partner species can incur net fitness costs from their interactions with ants during periods in which attack risk is low. These costs depend in part on the costliness of the reward. For example, lycaenids involved in

94

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

relatively obligate and species-specific ant-tending mutualisms produce particularly high-nutrient secretions, and experience reduced success when tended in the absence of a predation risk (Pierce et al., 1987). Untended, secreting lycaenids engaged in more facultative ant-tending mutualisms have not been found to experience these costs (e.g., Wagner, 1993). C. The Environment as ‘‘Competitor’’ Any partner in mutualisms involving food offers must outcompete alternative food sources that are available and acceptable to individuals of the partner species. These alternative food sources could be other potential mutualist species. For example, plants and Homoptera may compete for access to ant defenders (Sakata and Hashimoto, 2000), and several client species may compete among each other over access to cleaners (Bshary, 2001). In most cases, however, alternative foods are not associated with potential mutualists. Cleaners may feed from free-swimming invertebrates, ants must feed on protein as well as sugar-rich secretions and excretions, and vertebrate pollinators feed on insects as well as nectar. In addition, in some nutritional symbioses, critical nutrients may at some times be obtained directly from the environment (e.g., Johnson et al., 1997). Interacting with the partner is therefore at the expense of looking for alternative options in the environment, thus there are opportunity costs. If these costs are higher than the benefits of the mutualistic interactions, because of high abundances of alternative commodities in the environment, a (temporary) breakdown of mutualistic relationships may be observed, not because one partner starts cheating, but because it terminates the interaction. D. Mode of Transmission In symbioses, a critical question appears to be whether symbionts are transmitted vertically or horizontally. Any vertical transmission of symbionts into the next host generation may greatly reduce potential for conflicts, as it is then in the interest of both sides that the partner fares well (but see Frank, 1996). In contrast, if both sides reproduce independently and partners must locate each other at later stages of their life cycle, it is theoretically possible that exploitation of the partner to such a point that the partner is no longer able to reproduce (but is still able to offer the commodity in demand) may be favored in the short term (in the long term, of course, this may lead the mutualism to break down). A variation of this question exists in seed parasite pollination mutualisms. Fig wasp females collect the pollen within the inflorescence from which they emerge and use it to pollinate the inflorescence that they enter. Hence, the male reproductive success of the tree is positively coupled with the reproductive success

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

95

of the wasp. In contrast, female yucca and senita moths collect pollen elsewhere and the male reproductive success of their host plant is therefore not coupled to the reproductive success of the moths.

IX. Future Avenues with Respect to Evaluation of Game Structures There are three major gaps in our knowledge that must be filled if we are to reach a deeper understanding of how mutualistic behaviors persist. These points are crucial issues for future empirical studies to address. 1. We need to know more about the exact payoffs associated with different individual behaviors. There are two main problems to solve: (a) we need to know whether a reduction in investment into the partner yields a higher or lower (short-term) benefit to the actor. This distinction is crucial because it influences the game in the most fundamental way. If a reduction in investment yields a higher payoff, then cheating is a profitable option, and partner control mechanisms must reinforce cooperative behavior if the mutualism is not to be driven to extinction. If reduced investment in fact carries no net benefit because it predictably lowers the partner’s investment in return, then investment into the partner would be a form of pseudoreciprocity (Connor, 1986), and cheating is highly unlikely to be an issue; and (b) we must put all potential costs into the equation. Costs arise if an individual invests in the partner, if the partner performs a discrete cheating act (nectar robbing, eating a cleaner, etc.) or if it reduces the benefits through decreased investment, and if there are opportunity costs. These latter costs may not influence the course of interactions, but interactions may simply not take place if opportunity costs are too high. 2. We need to learn more about how well each player can control how much it offers to its partner, and the extent to which it can force its partner to provide commodities in return. Such information is particularly crucial to assess in mutualistic symbioses. Research on mechanisms of nutrient transfer in nitrogen-fixing associations by Lodwig et al. (2003) sets a prime example of what should be investigated in other symbioses as well. As long as each player has control over what it gives but not over what it receives, cooperative behavior may be more stable against cheating strategies than in situations in which one player has some power to rob/steal/purloin commodities from its partner.

96

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

3. We need to understand generalized mutualisms, that is, mutualisms that involve several potential partner species, in more detail. For example, it needs to be established whether or not a suboptimal partner species (e.g., an ant species that does not protect its partner species as efficiently as does another ant species) has a higher gain from the interaction than the partner species that provides the optimal service. Alternatively, suboptimal partners also do not gain high benefits from the interaction. In other words, suboptimal partners may be better exploiters, or they may be adapted to other partners themselves, and perform suboptimally in the association as well.

X. Conclusions We still have a long way to go to understand how cooperative behavior persists in mutualisms. To achieve this goal, we need more empirical information that is structured in a way that it is of direct value for theoreticians. An interactive approach, which involves developing theoretical concepts based on empirical information that are then amenable to further empirical tests, seems to be most promising at this stage. We hope that our review provides a step in this direction. While it seems impossible at this point to develop a single unifying concept for behavioral strategies that may explain the persistence of cooperative behavior in mutualisms, the diversity of game structures is not so large that it could not be tackled by theoreticians. Detailed theoretical and empirical studies of single parameters may yield building blocks for more sophisticated integrative models (Bshary and Noe¨, 2003). Such building blocks must not be confined to behavior and game theory, but we must incorporate ecological information and population dynamics as well.

XI. Summary Currently, there is little information transfer between empiricists working on cooperative interactions between species (mutualism) and theoreticians who model possible scenarios for the evolution and maintenance of cooperation between unrelated individuals. Furthermore, both theoretical and behavioral approaches often fail to consider ecological parameters that influence behavior. Our goal is to present the wealth of empirical knowledge (both behavioral and ecological) on mutualistic systems in a structure that may facilitate communication between empiricists and theoreticians.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

97

We have chosen eight broad categories of mutualisms that have been intensely studied and that are relatively well understood. For each system, we assess possible states of 12 parameters that can help theoreticians to construct game structures of mutualisms that are built on current empirical knowledge. We point out how ecological variables may influence behavioral decisions in ways not identified by our parameters. Finally, we elucidate similarities between mutualistic systems with respect to game structures that may not be expected given the diversity of mutualisms with respect to ecological and evolutionary background. On the basis of these results, we promote an interactive approach with models based on empirical knowledge, amenable to further testing. Acknowledgments We thank the editors for inviting us to write this article. We particularly thank Jane Brockmann and Marc Naguib for perceptive comments on a previous version. Funding was provided by the University of Liverpool.

References Addicott, J. F., Bronstein, J. L., and Kjellberg, F. (1990). Evolution of mutualistic life cycles: Yucca moths and fig wasps. In ‘‘Genetics, Evolution, and Coordination of Insect Life Cycles’’ (F. Gilbert, Ed.), pp. 143–161. Springer-Verlag, London. Axelrod, R., and Hamilton, W. D. (1981). On the evolution of cooperation. Science 211, 1390–1396. Axe´n, A. H., Leimar, O., and Hoffman, V. (1996). Signalling in a mutualistic interaction. Anim. Behav. 52, 321–333. Bao, T., and Addicott, J. F. (1998). Cheating in mutualism: Defection of Yucca baccata against its yucca moths. Ecol. Lett. 1, 155–159. Bergstrom, C., and Lachmann, M. (2003). The red king effect: Evolutionary rates and the division of surpluses in mutualisms. In ‘‘Genetic and Cultural Evolution of Cooperation: Report of the 90th Dahlem Workshop, Berlin, June 23–28, 2002’’ (P. Hammerstein, Ed.), pp. 223–240. MIT Press, Cambridge, MA. Bergstrom, C., Bronstein, J. L., Bshary, R., Connor, R. C., Daly, M., Frank, S. A., Gintis, H., Keller, L., Leimar, O., No¨e, N., and Queller, D. C. (2003). Interspecific mutualism— puzzles and predictions. In ‘‘Genetic and Cultural Evolution of Cooperation: Report of the 90th Dahlem Workshop, Berlin, June 23–28, 2002’’ (P. Hammerstein, Ed.), pp. 241–256. MIT Press, Cambridge, MA. Boucher, D. H., James, S., and Keeler, K. H. (1982). The ecology of mutualism. Annu. Rev. Ecol. Syst. 13, 315–347. Bowles, S., and Hammerstein, P. (2003). Does market theory apply to biology? In ‘‘Genetic and Cultural Evolution of Cooperation: Report of the 90th Dahlem Workshop, Berlin, June 23–28, 2002’’ (P. Hammerstein, Ed.), pp. 153–165. MIT Press, Cambridge, MA. Bronstein, J. L. (1994). Our current understanding of mutualism. Q. Rev. Biol. 69, 31–51. Bronstein, J. L. (2001a). Mutualisms. In ‘‘Evolutionary Ecology: Perspectives and Synthesis’’ (C. Fox, D. Fairbairn, and D. Roff, Eds.), pp. 315–330. Oxford University Press, Oxford.

98

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

Bronstein, J. L. (2001b). The exploitation of mutualisms. Ecol. Lett. 4, 277–287. Bronstein, J. L. (2001c). The costs of mutualism. Am. Zool. 41, 127–141. Bronstein, J. L., and Barbosa, P. (2002). Multitrophic/multispecies mutualistic interactions: The role of non-mutualists in shaping and mediating mutualisms. In ‘‘Multitrophic Level Interactions’’ (B. Hawkins and T. Tsharntke, Eds.), pp. 44–65. Cambridge University Press, Cambridge. Brown, J. L. (1983). Cooperation—a biologist’s dilemma. In ‘‘Advances in the Study of Behavior’’ (J. S. Rosenblatt, Ed.), pp. 1–37. Academic Press, New York. Bshary, R. (2001). The cleaner fish market. In ‘‘Economics in Nature’’ (R. Noe¨, J. A. R. A. M. van Hooff, and P. Hammerstein, Eds.), pp. 146–172. Cambridge University Press, Cambridge. Bshary, R. (2002). Building up relationships in asymmetric cooperation games between the cleaner wrasse Labroides dimidiatus and client reef fish. Behav. Ecol. Sociobiol. 52, 365–371. Bshary, R., and Noe¨, R. (1997). Red colobus and diana monkeys provide mutual protection against predators. Anim. Behav. 54, 1461–1474. Bshary, R., and Noe¨, R. (2003). Biological markets: The ubiquitous influence of partner choice on cooperation and mutualism. In ‘‘Genetic and Cultural Evolution of Cooperation: Report of the 90th Dahlem Workshop, Berlin, June 23–28, 2002’’ (P. Hammerstein, Ed.), pp. 167–184. MIT Press, Cambridge, MA. Bshary, R., and Scha¨ffer, D. (2002). Choosy reef fish select cleaner fish that provide high service quality. Anim. Behav. 63, 557–564. Bshary, R., and Wu¨rth, M. (2001). Cleaner fish Labroides dimidiatus manipulate client reef fish by providing tactile stimulation. Proc. R. Soc. Lond. B 268, 1495–1501. Bshary, R., Wickler, W., and Fricke, H. (2002). Fish cognition: A primate’s eye view. Anim. Cogn. 5, 1–13. Buckley, R. (1987). Interactions involving plants, Homoptera, and ants. Annu. Rev. Ecol. Syst. 18, 111–135. Bull, J. J., and Rice, W. R. (1991). Distinguishing mechanisms for the evolution of co-operation. J. Theor. Biol. 149, 63–74. Castellanos, M. C., Wilson, P., and Thomson, J. D. (2002). Dynamic nectar replenishment in flowers of Penstemon (Scrophulariaceae). Am. J. Bot. 89, 111–118. Cheney, K. L., and Coˆte´, I. M. (2001). Are Caribbean cleaning symbioses mutualistic: Costs and benefits of visiting cleaning stations to longfin damselfish. Anim. Behav. 62, 927–933. Connor, R. C. (1986). Pseudo-reciprocity: Investing in altruism. Anim. Behav. 34, 1562–1566. Cords, M. (1987). ‘‘Mixed-species Association of Cercopithecus Monkeys in the Kagamega Forest, Kenya.’’ University of California Press, Berkeley. Coˆte´, I. M. (2000). Evolution and ecology of cleaning symbioses in the sea. Oceanogr. Mar. Biol. Annu. Rev. 38, 311–355. Dafni, A. (1984). Mimicry and deception in pollination. Annu. Rev. Ecol. Syst. 15, 259–278. Darwin, C. (1859). ‘‘The Origin of Species.’’ John Murray, London. Despre´s, L. (2003). Sex and pollen: The role of males in stabilising a plant–seed eater pollinating mutualism. Oecologia 135, 60–66. Doebeli, M., and Knowlton, N. (1998). The evolution of interspecific mutualisms. Proc. Natl. Acad. Sci. USA 95, 8676–8680. Douglas, A. E. (1994). ‘‘Symbiotic Interactions.’’ Oxford University Press, Oxford. Dufay¨, M., and Anstett, M.-C. (2003). Conflicts between plants and pollinators that reproduce within inflorescences: Evolutionary variations on a theme. Oikos 100, 3–14.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

99

Dugatkin, L. A. (1997). ‘‘Cooperation among Animals: An Evolutionary Perspective.’’ Oxford University Press, Oxford. Ehrlich, P. R., and Ehrlich, A. H. (1973). Coevolution: Heterotypic schooling in Caribbean reef fishes. Am. Nat. 107, 157–160. Ferrie`re, R., Bronstein, J. L., Rinaldi, S., Law, R., and Gauduchon, M. (2002). Cheating and the evolutionary stability of mutualism. Proc. R. Soc. Lond. B 269, 773–780. Fischer, M. K., and Shingleton, A. W. (2001). Host plant and ants influence the honeydew sugar composition of aphids. Funct. Ecol. 15, 544–550. FitzGibbon, C. D. (1990). Mixed-species grouping in Thomson’s and Grant’s gazelles: The antipredator benefits. Anim. Behav. 39, 1116–1126. Frank, S. A. (1996). Host control of symbiont transmission: The separation of symbionts into germs and soma. Am. Nat. 148, 1113–1124. Ganeshaiah, K. N., Kathuria, P., Shaanker, R. U., and Vasudeva, R. (1995). Evolution of stylelength variability in figs and optimization of ovipositor length in their pollinator wasps: A coevolutionary model. J. Genet. 74, 25–39. Giurfa, M., and Nunez, J. A. (1992). Honeybees mark with scent and reject recently visited flowers. Oecologia 89, 113–117. Greig-Smith, P. W. (1981). The role of alarm responses in the formation of mixed-species flocks of heathland birds. Behav. Ecol. Sociobiol. 8, 7–10. Grutter, A. S. (1997). Spatio-temporal variation and feeding selectivity in the diet of the cleaner fish Labroides dimidiatus. Copeia 1997, 346–355. Hamilton, W. D. (1964). The genetical evolution of social behaviour. I and II. J. Theor. Biol. 7, 1–52. Hammerstein, P., and Hoekstra, R. F. (1995). Mutualism on the move. Nature 376, 121–122. Heil, M., and McKey, D. (2003). Protective ant–plant interactions as model systems in ecological and evolutionary research. Annu. Rev. Ecol. Evol. Syst. 34, 425–453. Henschel, J. R., Simmons, R. E., and Mendelsohn, J. M. (1992). Gabar goshawks and social spiders revisited: Untangling the web. Gabar 7, 49–50. Herre, E. A., Knowlton, N., Mueller, U. G., and Rehner, S. A. (1999). The evolution of mutualisms: Exploring the paths between conflict and cooperation. Trends Ecol. Evol. 14, 49–53. Herrera, C. M. (2002). Seed dispersal by vertebrates. In ‘‘Plant–Animal Interactions’’ (C. M. Herrera and O. Pellmyr, Eds.), pp. 185–210. Blackwell, Oxford. Hoeksema, J. D., and Schwartz, M. W. (2001). Modeling interspecific mutualisms as biological markets. In ‘‘Economics in Nature’’ (R. Noe¨, J. A. R. A. M. van Hooff, and P. Hammerstein, Eds.), pp. 173–183. Cambridge University Press, Cambridge. Hoeksema, J. D., and Schwartz, M. W. (2003). Expanding comparative-advantage biological market models: Contingency of mutualism on partners’ resource requirements and acquisition trade-offs. Proc. R. Soc. Lond. B 270, 913–919. Holland, J. N., and DeAngelis, D. L. (2001). Population dynamics and the ecological stability of obligate pollination mutualisms. Oecologia 126, 575–586. Holland, J. N., DeAngelis, D. L., and Bronstein, J. L. (2002). Population dynamics of mutualism: Functional responses of benefits and costs. Am. Nat. 159, 231–244. Holland, J. N., Bronstein, J. L., and DeAngelis, D. L. (2004). Testing hypotheses for excess flower production and fruit-to-flower ratios in a pollinating seed-consuming mutualism. Oikos 105, 633–640. Honer, O., Leumann, L., and Noe¨, R. (1997). Dyadic associations of red colobus and Diana monkey groups in the Tai National Park, Ivory Coast. Primates 38, 281–291.

100

REDOUAN BSHARY AND JUDITH L. BRONSTEIN

Huth, C. J., and Pellmyr, O. (1999). Yucca moth oviposition and pollination behavior is affected by past flower visitors: Evidence for a host-marking pheromone. Oecologia 119, 593–599. Irwin, R. E., Brody, A. K., and Waser, N. M. (2001). The impact of floral larceny on individuals, populations, and communities. Oecologia 129, 161–168. Johnson, N. C., Graham, J. H., and Smith, F. A. (1997). Functioning of mycorrhizal associations along the mutualism–parasitism continuum. New Phytol. 135, 575–585. Johnstone, R. A., and Bshary, R. (2002). From parasitism to mutualism: Partner control in asymmetric interactions. Ecol. Lett. 5, 634–639. Kiers, E. T., Rousseau, R. A., West, S. A., and Denison, R. F. (2003). Host sanctions and the legume–rhizobium mutualism. Nature 425, 78–81. Kjellberg, F., Gouyon, P. H., Ibrahim, M., Raymond, M., and Valdeyron, G. (1987). The stability of the symbiosis between dioecious figs and their pollinators: A study of Ficus carica L. and Blastophaga psenes L. Evolution 41, 693–704. Law, R., Bronstein, J. L., and Ferriere, R. G. (2001). On mutualists and exploiters: Plant– insect coevolution in pollinating seed–parasite systems. J. Theor. Biol. 212, 373–389. Leimar, O., and Axe´n, A. H. (1993). Strategic behavior in an interspecific mutualism: Interactions between lycaenid larvae and ants. Anim. Behav. 46, 1177–1182. Lodwig, E. M., Hosie, A. H. F., Bordes, A., Findlay, K., Allaway, D., Karunakaran, R., Downie, J. A., and Poole, P. S. (2003). Amino-acid cycling drives nitrogen fixation in the legume–Rhizobium symbiosis. Nature 422, 722–726. Losey, G. S., Grutter, A. S., Rosenquist, G., Mahon, J. L., and Zamzow, P. (1999). Cleaning symbiosis: A review. In ‘‘Behaviour and Conservation of Littoral Fishes’’ (V. C. Almada, R. F. Oliveira, and E. J. Goncalves, Eds.), pp. 379–395. Instituto Superior de Psichologia Aplicada, Lisbon, Portugal. Mallet, J. (1999). Causes and consequences of a lack of coevolution in mullerian mimicry. Evol. Ecol. 13, 777–806. Maloof, J. E., and Inouye, D. W. (2000). Are nectar robbers cheaters or mutualists? Ecology 81, 2651–2661. Molbo, D., Machado, C. A., Sevenster, J. G., Keller, L., and Herre, E. A. (2003). Cryptic species of fig-pollinating wasps: Implications for the evolution of the fig–wasp mutualism, sex allocation, and precision of adaptation. Proc. Natl. Acad. Sci. USA 100, 5867–5872. Morales, M. A. (2000). Mechanisms and density dependence of benefit in an ant–membracid mutualism. Ecology 81, 482–489. Moynihan, M. (1962). The organisation and probable evolution of some mixed species flocks of neotropical birds. Smithsonian Misc. Collect. 143, 1–140. Murray, K. G., Russell, S., Picone, C. M., Winnett-Murray, K., Sherwood, W., and Kuhlman, M. L. (1994). Fruit laxatives and seed passage rates in frugivores: Consequences for plant reproductive success. Ecology 75, 989–994. Ness, J. H. (2003). Catalpa bignonioides alters extrafloral nectar production after herbivory and attracts ant bodyguards. Oecologia 134, 210–218. Noe¨, R. (2001). Biological markets: Partner choice as the driving force behind the evolution of cooperation. In ‘‘Economics in Nature’’ (R. Noe¨, J. A. R. A. M. van Hooff, and P. Hammerstein, Eds.), pp. 93–118. Cambridge University Press, Cambridge. Noe¨, R., and Bshary, R. (1997). The formation of red colobus–Diana monkey associations under predation pressure from chimpanzees. Proc. R. Soc. Lond. B 264, 253–259. Noe¨, R., and Hammerstein, P. (1994). Biological markets: Supply and demand determine the effect of partner choice in cooperation, mutualism and mating. Behav. Ecol. Sociobiol. 35, 1–11.

GAME STRUCTURES IN MUTUALISTIC INTERACTIONS

101

Noe¨, R., van Schaik, C. P., and van Hooff, J. A. R. A. M. (1991). The market effect: An explanation for pay-off asymmetries among collaborating animals. Ethology 87, 97–118. Pellmyr, O., and Huth, C. J. (1994). Evolutionary stability of mutualism between yuccas and yucca moths. Nature 372, 257–260. Pierce, N. E., Kitching, R. L., Buckley, R. C., Taylor, M. F. L., and Benbow, K. F. (1987). The costs and benefits of cooperation between the Australian lycaenid butterfly, Jalmenus evagoras, and its attendant ants. Behav. Ecol. Sociobiol. 21, 237–248. Pierce, N. E., Braby, M. F., Heath, A., Lohman, D. J., Mathew, J., Rand, D. B., and Travassos, M. A. (2002). The ecology and evolution of ant association in the Lycaenidae (Lepidoptera). Annu. Rev. Entomol. 47, 733–771. Randall, J. E. (1958). A review of the labrid fish genus Labroides, with descriptions of two new species and notes on ecology. Pac. Sci. 12, 327–347. Roberts, G., and Sherratt, T. N. (1998). Development of cooperative relationships through increasing investment. Nature 394, 175–179. Ruxton, G., Sherratt, T. N., and Speed, M. (2004). ‘‘Avoiding attack: The Evolutionary Ecology of Crypsis, Warning Signals and Mimicry.’’ Oxford University Press, Oxford (in press). Sakata, H., and Hashimoto, Y. (2000). Should aphids attract or repel ants? Effect of rival aphids and extrafloral nectaries on ant–aphid interactions. Popular Ecol. 42, 171–178. Schwartz, M. W., and Hoeksema, J. D. (1998). Specialization and resource trade: Biological markets as a model of mutualisms. Ecology 79, 1029–1038. Sherman, P. W. (1977). Nepotism and the evolution of alarm calls. Science 197, 1246–1253. Sherratt, T. N., and Roberts, G. (2001). The role of phenotypic defectors in stabilizing reciprocal altruism. Behav. Ecol. 12, 313–317. Smith, F. A., and Smith, S. E. (1996). Mutualism and parasitism: Diversity in function and structure in the ‘‘arbuscular’’ (VA) mycorrhizal symbiosis Adv. Bot. Res. 22, 1–43. Speed, M. P. (1999). Batesian, quasi-Batesian or Mullerian mimicry? Theory and data in mimicry research Evol. Ecol. 13, 755–776. Tebbich, S., Bshary, R., and Grutter, A. S. (2002). Cleaner fish Labroides dimidiatus recognise familiar clients. Anim. Cogn. 5, 139–145. Titlyanov, E. A., Titlyanova, T. V., Leletkin, V. A., Tsukahara, J., van Woesik, R., and Yamazato, K. (1996). Degradation of zooxanthellae and regulation of their density in hermatypic corals. Mar. Ecol. Prog. Ser. 139, 167–178. Trivers, R. L. (1971). The evolution of reciprocal altruism. Q. Rev. Biol. 46, 35–57. Va´zquez, D. P., and Simberloff, D. (2002). Ecological specialization and susceptibility to disturbance: Conjectures and refutations. Am. Nat. 159, 606–623. Wagner, D. (1993). Species-specific effects of tending ants on the development of lycaenid butterfly larvae. Oecologia 96, 276–281. Weeks, P. (2000). Red-billed oxpeckers: Vampires or tickbirds? Behav. Ecol. 11, 154–160. West, S. A., Kiers, E. T., Simms, E. L., and Denison, R. F. (2002). Sanctions and mutualism stability: Why do rhizobia fix nitrogen? Proc. R. Soc. Lond. B 269, 685–694. Whitesides, G. H. (1989). Interspecific associations of Diana Monkeys, Cercopithecus diana, in Sierra Leone, West Africa, biological significance or chance? Anim. Behav. 37, 760–776.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Neurobehavioral Development of Infant Learning and Memory: Implications for Infant Attachment Tania L. Roth, Donald A. Wilson, and Regina M. Sullivan department of zoology university of oklahoma norman, oklahoma 73019, usa

I. Introduction The environmental demands on an altricial newborn are simple: procure food, warmth, and protection from a caregiver, and rapidly form an attachment to the caregiver. While pheromones mediate this attachment in some species (reviewed in Bartoshuk and Beauchamp, 1994; Hudson and Distel, 1999; Schaal et al., 1995), other species use learning, with avian imprinting being one of the most widely known examples of the latter. Similar learning during infancy has also been documented in mammals such as the human (DeCasper and Fifer, 1980; Sullivan et al., 1991), sheep (Nowak and Lindsay, 1992; Nowak et al., 1997), and rat (reviewed in Hofer and Sullivan, 2001). While this infant learning shares many characteristics with adult learning, some of the neural structures supporting learning and memory consolidation in the adult are not yet fully developed in the infant, suggesting that the neural basis for learning and memory differs between infants and adults. In this article we present evidence that the infant’s behavior and its brain are specifically designed to meet the demands of infancy and to ensure attachment to the caregiver. John Bowlby (1965) originally documented attachment in human children, and the basic characteristics of human attachment and behavior that he described are briefly reviewed here. First, children form rapid, strong attachments to their primary caregiver and seek proximity to the caregiver. Second, as highlighted in clinical reports, children will undergo considerable abuse yet still remain in contact with the abusive caregiver (Helfer et al., 1997) although, clinically, these children show some disturbed attachment 103 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

104

TANIA L. ROTH et al.

characteristics (Hesse and Main, 2000; O’Connor and Rutter, 2000; reviewed in Morton and Browne, 1998). Finally, Bowlby’s claim that the infant–caregiver relationship defines subsequent adult relationships can be documented most clearly in cases in which these abused children later form adult relationships that often involve insecure attachments and abusive behavior (Mullen et al., 1994; Smallbone and McCabe, 2003; Styron and Janoff-Bulman, 1997). Bowlby’s description of attachment applies to other species. Rapid attachment occurs in a variety of species, with imprinting as the classic example. Moreover, attachment despite abuse is widespread in the animal kingdom. For example, during the critical period for imprinting, chicks can be shocked while following the surrogate, yet following still occurs (Hess, 1962). This same treatment of a chick just hours after the critical period has closed produces an aversion to the surrogate. Similarly, infant dogs neglected, mishandled, or shocked by the caregiver also form a strong caregiver attachment, while the same situation produces quickly learned aversions in older dogs (reviewed in Rajecki et al., 1978). Perhaps one of the more dramatic examples of infant attachment in an abusive caregiver relationship was documented by H. Harlow in nonhuman primates (Harlow and Harlow, 1965). Specifically, maternally deprived infants were permitted to mature, mate, and give birth. Lacking mothering skills compounded with having a disturbed nature, these animals severely mistreated their young, yet the abused infants still developed a strong attachment to their caregivers. More recently a model of abuse in nonhuman primates at Yerkes National Primate Research Center has documented attachment after far more moderate abuse by presumably healthy, normal mothers (e.g., Maestripieri et al., 1997, 1999). While it may appear maladaptive to attach to an abusive caregiver, when one considers that survival in altricial species is greatly compromised without a caregiver (no food, protection, or warmth), evolution may have carved a learning attachment system in some species that functions regardless of the quality of parental care.

II. Unique Characteristics of Infant Learning For many altricial species, at least some learning is required for attachment to the caregiver. Using our rat model of attachment, we suggest that the unique infant learning characteristics documented below support this attachment system and ensure that the animal forms a repertoire of proximity-seeking behaviors directed toward the caregiver, regardless of the quality of care received. Indeed, as is illustrated in this review of the developmental learning literature on rats, the infant system seems to

DEVELOPMENT OF LEARNING AND MEMORY

105

potentiate learned approach responses and attenuate learned avoidance responses, ultimately supporting attachment under a variety of conditions. This period of unique infant learning ends at postnatal day (PN) 10 and is referred to as the sensitive period for learning. A. Acquisition of Infant Learning Learning involves the acquisition of a change in behavior to an environmental stimulus (reviewed in Abel and Lattal, 2001). A large body of literature on the development of learning in rodents suggests that acquiring new information appears quite similar in infant and adult rats (reviewed in Campbell, 1984; Fanselow and Rudy, 1998; Hudson et al., 1998; Robinson and Smotherman, 1995; Spear and Rudy, 1991; Stanton, 2000). Nevertheless, unique learning characteristics have been documented in pups. Rat pup learning seems to have two major periods of developmental change. The first occurs as pups make the transition from crawling to walking, which expands their environment to outside the nest. The second occurs around weaning, when parental assistance ends. We hypothesize that evolution has worked on each developmental transition to accommodate the pups’ changing environment. While both transitions are described below, our emphasis will be the first transition, which has been the focus of our research efforts. Also, this review focuses on classic conditioning with a strong emphasis on olfaction, because of its importance in mediating a range of neonatal behaviors, from suckling and huddling, to the formation of incentive-seeking or avoidance behaviors (e.g., Alberts, 1976; Rosenblatt, 1983). Because of the large amount of literature on infant learning, this is not a comprehensive review, but highlights the literature that demonstrates the specialization of the infant brain to support mother-infant attachment. 1. Behavioral Studies on Infant Acquisition Before walking emerges, pups are usually confined to the nest and appear to have a learning system that predisposes them to preference learning. Both olfactory and somatosensory learning occur rapidly in the neonate, with robust learning resulting from as little as 10 min of conditioned stimulus (CS)–reward pairings (reviewed in Hofer and Sullivan, 2001). Stimuli that function as a reward are broadly defined. Traditional rewards that support odor preference learning include warmth and milk (Brake, 1981; Gue´naire et al., 1982; Thoman et al., 1968; reviewed in Hofer and Sullivan, 2001); but unconventional rewards, such as tactile stimulation (called stroking—McLean et al., 1993, 1999; Pedersen et al., 1982; reviewed in Hofer and Sullivan, 2001), and paradoxical rewards, such as a 0.5-mA

106

TANIA L. ROTH et al.

shock and tail pinch, also produce a subsequent odor preference (Camp and Rudy, 1988; reviewed in Hofer and Sullivan, 2001). This changing reward value during development is illustrated in Fig. 1. The stimuli that function as reward become more narrowly defined as the learning system matures, coinciding with the emergence of walking around PN10 (Bolles and Woods, 1965). While traditional rewards (such as milk) retain their rewarding value throughout development, others (stroking) lose their rewarding value, and still others (shock) change their rewarding value from supporting an odor preference to supporting an odor aversion (Camp and Rudy, 1988; Johanson and Hall, 1982; Woo and Leon, 1987; reviewed in Hofer and Sullivan, 2001). Overall, during the sensitive period, the reward system appears to be designed to support rapid learning about the mother, at a time when pups are confined to the nest and the learning demands of the pup are limited to learning a preference for

Fig. 1. During the first postnatal week, a wide range of stimuli support odor preference learning. Milk, stroking, and shock each can produce an odor preference in PN6 pups. However, after the sensitive period, tactile stimulation no longer serves as a rewarding stimulus, and thus does not produce an odor preference. Conversely, shock in PN12 pups produces an odor aversion, illustrating a change in the hedonic value of shock. Modified from Sullivan and Wilson (1994); Hofer and Sullivan (2001).

DEVELOPMENT OF LEARNING AND MEMORY

107

maternal, nest, and peer odors. As the pup matures and begins to walk at PN10, a more adult-like learning system emerges that provides the pup with a more complex learning system to accommodate a complicated extranest environment. In addition, as compared with the adult or older pup, which focuses on a select group of salient stimuli, the sensitive period neonate seems to learn about a wide range of stimuli (Spear et al., 1989). Moreover, while preexposure to a CS, such as an odor, hinders subsequent conditioning to that CS in adults, CS preexposure facilitates subsequent learning in young pups (Hoffmann and Spear, 1989; Misanin et al., 1983). This broad sensitive period learning may help pups to learn about the complex features of the dam, siblings, and nest. Perhaps the most dramatic neonate learning characteristic is the limitations put on learning passive avoidance, active avoidance, and inhibitory conditioning during the sensitive period (Blozovski and Cudennec, 1980; Camp and Rudy, 1988; Collier et al., 1979; Roth and Sullivan, 2001, 2003; Sullivan et al., 2000a; reviewed in Myslivecek, 1997). Specifically, odor paired with moderate pain (0.5-mA shock or tail pinch) results in pups learning a preference for that odor as indicated by orienting toward the odor and even climbing an obstacle to approach the odor (Camp and Rudy, 1988; Roth and Sullivan, 2001; Sullivan et al., 2000a). The 0.5-mA shock is an intensity similar to that used in adult fear conditioning experiments, and it should be noted that pups feel pain. Threshold to shock does not appear to change developmentally, and a 0.5-mA shock elicits both broadband vocalizations (indicative of pain; White et al., 1992) and escape responses associated with pain (Emerich et al., 1985; Stehouwer and Campbell, 1978; Sullivan et al., 2000a). The end of the sensitive period coincides with the emergence of walking (Bolles and Woods, 1965), suggesting that a more adult-like learning system may be needed as pups begin to venture outside the nest. However, it should be noted that neonatal rats can learn an odor aversion following odor– malaise pairings. Specifically, odor paired with illness-inducing LiCl or very strong shock (1.0–1.5 mA) results in a subsequent aversion for that odor (e.g., Haroutunian and Campbell, 1979; Martin and Alberts, 1979; Rudy and Cheatle, 1977; reviewed in Sullivan, 2001). Shock levels of 1.0 mA and above exceed those typically used in adult fear conditioning experiments (e.g., LaLumiere et al., 2003; Paschall and Davis, 2002; Wilensky et al., 1999). Work from the laboratories of J. Rudy and B. Campbell suggests that until PN9–10, pups easily learn aversions based on interoceptive cues (malaise or internal shock) but not exteroceptive cues, such as a moderate intensity external shock (Camp and Rudy, 1988; Haroutunian and Campbell, 1979). Indeed, Rudy suggests that changes in the categorization of appetitive and

108

TANIA L. ROTH et al.

aversive events occur in pups some time between PN8 and PN12, coinciding with the end of the sensitive period (Camp and Rudy, 1988). While dramatic changes in learning occur in the 10-day-old pup, not all developmental changes in learning occur at this age. For example, conditioned fear emerges at different ages depending on the sensory system: freezing (immobility) first emerges at PN10, PN16, and PN18, respectively, for the olfactory, auditory, and visual systems (Hunt et al., 1994, 1997; Sullivan et al., 2000a), with ear opening at PN12–13 and eye opening around PN15. Additional learning abilities seem to emerge at weaning. For example, potentiated startle, which is the enhancement of a startle response by an innate (i.e., loud noise) or learned fear (odor previously paired with shock), does not emerge until PN21–PN23 (Richardson et al., 2000, 2003; reviewed in Hunt and Campbell, 1999). Similarly, contextual fear conditioning, in which an animal displays a fear response to a context associated with an aversive stimulus, does not emerge until approximately PN23–25 (Rudy and Morledge, 1994; reviewed in Stanton, 2000). Overall, the learning system of pups seems to change at the time of developmental landmarks, such as walking and weaning (Hassmannova´ et al., 1985). 2. Neural Correlates of Infant Acquisition One of the perplexing issues concerning the neural basis of odor learning in neonatal pups during the sensitive period is that they show excellent learning ability, yet brain areas identified as important in adult odor conditioning (amygdala, hippocampus, and frontal cortex) appear not to participate, and major neural connections, such as amygdaloid–hippocampal and hippocampal–entorhinal cortical connections, are immature (Alvarez et al., 2002; Astic and Saucier, 1982; Crain et al., 1973; Hall, 1987; Landers and Sullivan, 1999; Litaudon et al., 1997; Nair and Gonzalez-Lima, 1999; Ramus and Eichenbaum, 2000; Rudy and Morledge, 1994; Rudy et al., 1987; Saar et al., 2002; Tronel and Sara, 2002; Verwer et al., 1996; reviewed in Fanselow and Rudy, 1998; Stanton, 2000). This suggests that the neural structures supporting classic conditioning may be different in pups, and the literature suggests that the brain structures supporting sensitive period learning are the olfactory bulb, the noradrenergic locus coeruleus (LC), and the amygdala (see Fig. 2). The piriform cortex and the anterior olfactory nucleus also appear to influence how olfactory information is handled in the neonatal brain, and other brain areas will certainly be added to this list as more information about the developing brain emerges. The classic conditioning learning circuit is diagramed within the olfactory system in Fig. 2. Relative to other sensory systems, the olfactory system

DEVELOPMENT OF LEARNING AND MEMORY

109

Fig. 2. Schematic representation of the olfactory system (not drawn to scale). Information progresses from the olfactory bulb to either the piriform cortex or amygdala. A small percentage of neurons also make connections in the anterior olfactory nucleus. Locus coeruleus fibers terminate in the olfactory bulb, providing rich noradrenergic input that is both necessary and sufficient for neonate learning.

is simplistic: the information does not pass through the thalamus before going on to the primary sensory (piriform) cortex, and there is only one synapse between the olfactory receptors and the piriform cortex. Information progresses from the olfactory bulb to either the cortex or amygdala, with additional connections in the anterior olfactory nucleus (AON) (reviewed in Brunjes and Frazier, 1986; Shipley and Ennis, 1996; Wilson and Sullivan, 2003). Information about reward appears to reach the olfactory circuit via the LC directly into the olfactory bulb. a. The olfactory bulb and the source of norepinephrine, the locus coeruleus As illustrated in Fig. 3, during acquisition the olfactory bulb primary output neurons (mitral cells) exhibit a heightened excitatory response in experimental pups (receive paired presentations of odor and reward) as compared with odor control groups (Wilson and Sullivan, 1992). Indeed, while mitral cells normally quickly habituate to repeated odor presentations, this habituation is prevented when that odor is paired with a reward. This heightened mitral cell response during training may be a critical factor in induction of the behavioral and neural changes discussed below.

110

TANIA L. ROTH et al.

Fig. 3. During acquisition the olfactory bulb primary output neurons (mitral cells) exhibit a heightened excitatory response in experimental pups (odor–reward groups) as compared with control groups. As indicated by mitral/tufted cell single unit responses, control pups (odor only) habituate to the odor, whereas pups receiving contiguous odor presentations and stimulation of the medial forebrain bundle/lateral hypothalamus as a reward ( paired) maintain odor responsiveness. Modified from Wilson and Sullivan (1992).

Norepinephrine (NE) is both necessary and sufficient for the enhanced mitral cell response to the odor and the acquisition of the learning induced neurobehavioral effects (reviewed in Sullivan and Wilson, 1994, 2003). Specifically, either blocking NE receptors in the bulb or destroying the LC (source of NE) prevents odor learning. More importantly, NE is sufficient to support neonatal odor learning; activation of olfactory bulb NE -receptors with isoproterenol paired with odor stimulation produces a learned approach response in rat pups (Langdon et al., 1997; Moriceau and Sullivan, 2004b; Sullivan et al., 2000b; Yuan et al., 2003; reviewed in Sullivan and Wilson, 2003). Moreover, increasing olfactory bulb NE by stimulating the LC during an odor presentation is sufficient to support odor learning and the learning-induced changes in the olfactory bulb that occur during the sensitive period (Moriceau and Sullivan, 2004b; Sullivan et al., 2000b). This is in sharp contrast to the effects of NE in the adult, where blocking/activating NE generally has only a modulatory effect on adult acquisition of such tasks as inhibitory avoidance or escape from a water maze (e.g., Harris and Fitzgerald, 1991; Liang, 1998; Sara et al., 1995). However, adult learning critical for survival, such as mating and infant care, requires NE (reviewed in Brennan and Keverne, 1997; Fleming et al., 1999; Insel and Young, 2001; Levy, 2002).

DEVELOPMENT OF LEARNING AND MEMORY

111

There is no endogenous NE in the olfactory bulb; its sole source of NE is the LC (McLean and Shipley, 1991; Shipley et al., 1985), which releases copious amounts of NE into the neonatal bulb with almost any moderate intensity sensory stimulus (Rangel and Leon, 1995; reviewed in Nakamura and Sakaguchi, 1990). The neonatal LC also releases substantially more NE into the bulb than the adult LC (Rangel and Leon, 1995). As is illustrated in Table I, there is a sharp contrast between the functioning of the neonatal and older LC, with LC developmental changes that dramatically reduce NE release coinciding with the termination of the sensitive period. The emergence of neonatal LC autoreceptors, receptors that are activated by the neurotransmitter released from LC neurons (NE), is the primary cause of developmental changes in the LC. During the second postnatal week, excitatory 1-autoreceptor function becomes limited and no longer temporally extends the heightened response of the LC to sensory stimuli, and inhibitory 2-autoreceptor function emerges, shutting down the LC within milliseconds of a response onset (Winzer-Serhan and Leslie, 1999; reviewed in Marshall et al., 1991; Nakamura and Sakaguchi, 1990). These findings suggest that the infant LC is responsible for enhanced odor preference learning, and that maturation of the LC (via emergence of 2autoreceptor function) signals the termination of the sensitive period for odor–preference learning in rat pups. Although NE appears particularly important in neural plasticity during early development, many neurotransmitters have a role in olfactory learning in neonatal rats [cholecystokinin—Shayit and Weller, 2001; GABA—Okutani et al., 2002; glutamate—Lincoln et al., 1988; Mickley et al., 1988; opioids—e.g., Barr and Rossi, 1992; Kehoe and Blass, 1986; Panksepp et al., 1994; Roth and Sullivan, 2001, 2003; serotonin (5-HT)—MCLean et al., 1993]. Nitric oxide, an intracellular messenger, also has a role in early olfactory learning (Samama and Boehm, 1999). It has also been shown that the interaction of NE and 5-HT within the bulb mediates acquisition. The NE effect on learning displays an inverted U-shaped dose–response curve (reviewed in Sullivan and Wilson, 1994) that can be shifted with manipulations of olfactory bulb 5-HT activity (McLean et al., 1993, 1999). Specifically, more NE is required for learning if 5-HT is depleted, while less NE is required if a 5-HT receptor agonist is added to the bulb. It should be noted that 5-HT without NE is not sufficient to support learning, whereas NE alone is sufficient to support learning (Yuan et al., 2003). b. The anterior olfactory nucleus Although there are learning-induced changes within the olfactory bulb, the mitral cell signal also leaves the bulb, suggesting that learning-induced changes are likely to occur in other brain

TABLE I Evidence That Developmental Changes in the Functioning of the Locus Coeruleus Underlie Termination of the Unique Learning Abilities Displayed During the Sensitive Period Physiological and cellular characteristic

Infant LC

Adult LC

Response to sensory stimuli

More responsive to sensory stimuli, including both noxious and nonnoxious stimuli

Less responsive, particularly to nonnoxious stimuli

Response to repeated sensory stimulation Response to sensory stimulation

Fails to habituate

Habituates

20 to 30-s response

Few millisecond response

Electronically coupled LC Large output of NE due to excitatory autoreceptors

Less electronically coupled LC Smaller output of NE due to emergence of inhibitory autoreceptors Produces no increase in NE olfactory bulb levels Decreased levels

Tyrosine hydroxylase levels

Produces 200–300% increase in olfactory bulb NE levels Greater levels, with a transient peak on postnatal day 10

Abbreviations: LC, locus coeruleus; NE, norepinephrine.

Ref. Aston-Jones and Bloom (1981); Aston-Jones et al. (1994, 1999); Harley and Sara (1992); Kimura and Nakamura (1985); Nakamura and Sakaguchi (1990); Sara et al. (1995) Kimura and Nakamura (1985); Nakamura and Sakaguchi (1990); Vankov et al. (1995) Kimura and Nakamura (1985); Nakamura and Sakaguchi (1990) Christie and Jelinek (1993); Marshall et al. (1991) Kimura and Nakamura (1985); Nakamura and Sakaguchi (1990)

Rangel and Leon (1995)

Bezin et al. (1994a,b)

DEVELOPMENT OF LEARNING AND MEMORY

113

areas, such as the AON. The AON serves as a commissural relay, or a connection between the olfactory bulbs, and receives input from both the olfactory bulbs and olfactory cortex (reviewed in Shipley and Ennis, 1996; Wilson and Sullivan, 2003). As in the adult rat, olfactory stimulation via presentations of a novel odor increases activity in the AON of pups (Astic and Saucier, 1982; Hall, 1987). In adult rats, presentations of a conditioned odor enhance metabolic activity in the AON (Hamrick et al., 1993); however, following an appetitive learning paradigm in 6-day-old pups, there appears to be little metabolic activity in the AON in conditioned pups, but there is an increase within the piriform cortex, suggesting learning-induced neural changes downstream of the olfactory bulb (Hall, 1987). c. The piriform cortex In adults, the piriform cortex appears to serve at least three functions in odor perception: (1) it allows rapid filtering of background or irrelevant stimuli while maintaining responsiveness to novel stimuli (Wilson, 2000); (2) it allows experience-dependent synthetic processing of multiple odorant features and odorant mixtures into single perceptual odor objects (reviewed in Wilson and Stevenson, 2003); and (3) it serves as one of perhaps many sites involved in odor associative and contextual memory (Litaudon et al., 1997; Saar et al., 1999; Schoenbaum and Eichenbaum, 1995). However, relatively little is known about the functional ontogeny of this structure. Mitral cell afferent fibers are present throughout the piriform cortex (Schwob and Price, 1984) and can evoke responses in piriform cortical neurons by birth (Schwob et al., 1984). Preliminary evidence from our laboratory suggests that odor-evoked responses (i.e., odors with which the pup has had no prior experience) can be seen in piriform single units by at least the second postnatal week, the earliest time point examined (Wilson, unpublished observations). Furthermore, mechanisms responsible for cortical habituation and odor filtering (Best and Wilson, 2004) are expressed by at least PN7, as determined in in vitro slices (Best and Wilson, unpublished observations). Using a unilateral odor conditioning paradigm combined with directed lesions of the anterior commissure, Kucharski and Hall (1987, 1988) argue that memory for learned odor preferences is stored in the AON or piriform cortex. Learned odor preferences could be expressed by pups, using an untrained naris following conditioning of the contralateral naris, but only if the anterior commissure was intact. The anterior commissure provides strong, direct connections between the bilateral AON and piriform cortices. In line with this hypothesis, in a large 2-deoxyglucose (2-DG) autoradiography mapping study of neonatal associative memory by Hall (1987), PN6 pups exposed to learned odors had enhanced 2-DG uptake in

114

TANIA L. ROTH et al.

both the olfactory bulb and the piriform cortex compared with control pups. These 2-DG data suggest that activity within the piriform cortex may reflect learned changes in neonates, although whether these changes are intrinsic to the piriform cortex or simply reflect the modified output of the olfactory bulb is unclear (Johnson et al., 1995; Sullivan and Leon, 1986; Wilson and Leon, 1988b; Wilson and Sullivan, 1991; Wilson et al., 1987). d. The amygdala Immature and/or limited amygdala function during the sensitive period appears to underlie the inability of pups to learn conditioned fear from odor–shock (0.5 mA) conditioning during the sensitive period. As is illustrated in Fig. 1, pups learn to approach an odor even after that odor has been paired with a painful stimulus. As noted above, this shock-induced odor preference is not due to the inability of pups to feel pain since shock threshold varies little during development (Emerich et al., 1985; Stehouwer and Campbell, 1978). As shown in Fig. 4, our assessment of the amygdala during acquisition (using 2-DG autoradiography) shows that the amygdala is not significantly activated by odor–shock conditioning during the sensitive period, but is activated by postsensitive period conditioning when pups readily learn an aversion to an odor paired

Fig. 4. Amygdala activity as measured by 2-DG autoradiography. The amygdala of sensitive period pups (PN8) does not appear to participate in odor shock conditioning and may underlie pups’ difficulty in learning odor aversions. Older pups, past the sensitive period, have an amygdala that participates in learning, and odor aversions are easily learned. Modified from Sullivan et al. (2000a).

DEVELOPMENT OF LEARNING AND MEMORY

115

with a 0.5-mA shock (Sullivan et al., 2000a). In addition, lesioning the amygdala only slightly retards odor learning in PN6 rats (Sullivan and Wilson, 1993), while this procedure dramatically and permanently disrupts fear conditioning in adult rats (reviewed in Maren, 2001). Furthermore, passive avoidance learning can be greatly potentiated in infant rats by facilitating amygdala activity (Dumery et al., 1988). The amygdala is a limbic structure involved with emotions, especially innate and learned fear (reviewed in Eichenbaum and Cohen, 2001; Fanselow and LeDoux, 1999; Maren, 2001; McGaugh, 2002; Walker and Davis, 2002). While most sensory systems send input to the amygdala via the thalamus and/or sensory cortices, olfactory bulb mitral cells synapse directly within the amygdala (cortical nucleus), with additional olfactory input via the piriform (olfactory) cortex (cortical and lateral nuclei; Schwob and Price, 1984), although it is unclear how effective these connections are during the sensitive period. Amygdala development begins during the midembryonic period, with subdivision of the major nuclei occurring around PN7 and stabilizing around PN14 (Bayer, 1980; Berdel et al., 1997). Olfactory bulb afferent fibers are present in the cortical nucleus of the amygdala at birth, as are piriform cortex afferents to the amygdala (Schwob and Price, 1984). Development of synaptic terminals begins by PN5, with the most prolific increase between PN10 and PN20, and adult levels are reached by PN30 (Mizukawa et al., 1989). As shown in Fig. 4, our 2-DG data suggest that the amygdala can be significanlty activated by paired odor–shock stimulation as the sensitive period ends at PN10, but not before. Preliminary singleunit recordings from amygdala neurons of developing rats suggest dramatic changes in spontaneous activity and response latency to olfactory bulb stimulation from PN11 to adult, although younger ages have not yet been examined (Wilson, 2003). Our interpretation of amygdala function is supported by research on the developmental emergence of natural (unlearned) fear. Coinciding with the developmental emergence of learned fear, fear (freezing) to a natural predator odor emerges at PN10 (e.g., Moriceau et al., 2004; Takahashi, 1994; Wiedenmayer and Barr, 2001), as do cellular changes (activation of immediate early gene products, such as c-fos) within the amygdala in response to a predator odor (Moriceau et al., 2004; Wiedenmayer and Barr, 2001). B. Memory Consolidation During Infancy Consolidation represents a postacquisition period when a cascade of neural and molecular events, involving several transcription factors (cellular proteins that regulate gene expression) and changes in both gene expression and protein synthesis, transfers learned information into a less

116

TANIA L. ROTH et al.

labile neural and molecular representation (reviewed in Abel and Lattal, 2001; Davis and Squire, 1984; Dudai, 2002; Eichenbaum and Cohen, 2001; Stork and Pape, 2002). Despite the relative immaturity of the central nervous system in neonates, molecular processes similar to those that mediate consolidation in the adult are present and functional early in an infant’s life, when learning an odor preference for the caretaker enhances survival (Yuan et al., 2003; Zhang et al., 2003; reviewed in Sullivan and Wilson, 2003). However, while the molecular events appear similar in infants and adults, the circuitry and the role of neurotransmitters appear to reflect how the infant’s brain is optimized to facilitate rapid attachment to the caretaker. Unfortunately, consolidation has not been assessed for the transition occurring at weaning, thus the mechanisms associated with learning and consolidation in neonates are addressed here. 1. Behavioral Studies on Infant Consolidation In the adult rat, several neurotransmitters have been shown to participate in memory consolidation. Both dopamine (reviewed in Jay, 2003) and glutamate (reviewed in Riedel et al., 2003) enhance memory consolidation, while evidence suggests that serotonin impairs consolidation processes (reviewed in Meneses, 2003). In adult behavioral studies, particular attention has been given to the role of NE, glucocorticoids, and endogenous opioids in the memory of arousing or emotional events. In general, posttraining systemic or central administration of noradrenergic receptor agonists, glucocorticoid receptor agonists, or opioid receptor antagonists enhances memory, while noradrenergic receptor antagonists, glucocorticoid receptor antagonists, and opioid receptor agonists impair memory (reviewed in McGaugh, 2002; McGaugh and Roozendaal, 2002; McGaugh et al., 1993; Roozendaal, 2002). In neonatal rats during the sensitive period, behavioral studies indicate that memory consolidation processes emerge early in development. Pharmacological exploration of the mechanisms responsible for infant memory consolidation of odor preferences has demonstrated a role for the same neurotransmitters/hormones involved in adult consolidation. As in the adult, neonatal odor conditioning is impaired by posttraining administration of a dopaminergic (Weldon et al., 1991), glutamatergic (Weldon et al., 1997), or noradrenergic receptor antagonist (Wilson et al., 1994). However, in contrast to the adult, posttraining administration of a noradrenergic receptor agonist impairs memory, even at very low concentrations (Wilson et al., 1994). We have shown that opioids are necessary for consolidation of odor preferences in neonates. As shown in Fig. 5, following our preferenceinducing odor–shock conditioning in neonates, posttraining injection of

DEVELOPMENT OF LEARNING AND MEMORY

117

Fig. 5. The effects of opioid receptor antagonism after odor–shock conditioning. Naltrexone (NTX), an opioid receptor antagonist, delivered immediately after odor–shock conditioning in neonates (PN7–8) disrupts consolidation of an odor preference and yields an odor aversion. NTX delivered after the same conditioning in older pups (PN12–13) does not disrupt consolidation of an odor aversion. It should be noted that backward odor–shock presentations (backward) represent a control similar to random odor–shock presentations (random), as were used in experiments summarized in Figs. 1 and 4. SAL, saline. Modified from Roth and Sullivan (2001, 2003).

an opioid receptor antagonist (naltrexone) prevents the subsequent formation of an odor preference, and yields an odor aversion (Roth and Sullivan, 2001, 2003). The same manipulation in older rats past the sensitive period shows results that are more consistent with adult consolidation: posttraining injection of naltrexone does not disrupt memory formation of the conditioned aversion (Roth and Sullivan, 2003; for a review of the role of opioids in consolidation, see McGaugh et al., 1993). In adults, opioids affect memory processes by altering noradrenergic activity within the amygdala;

118

TANIA L. ROTH et al.

opioid agonists decrease noradrenergic input and thus disrupt consolidation, while opioid antagonists increase noradrenergic input and facilitate consolidation (reviewed in McGaugh et al., 1993). The interaction of NE and opioids in memory consolidation in the neonate has not been assessed. Thus, while the overall behavioral consolidation process appears similar in adult and neonatal rats, different mechanisms may be involved because of the prominent role of NE in infant learning, the facilitatory role of endogenous opioids in consolidation of odor preferences, and functional immaturity of certain brain areas. Indeed, these differences appear to reflect how the neonatal brain is designed to support rapid learning and memory of odor preferences that enhance attachment behaviors and, ultimately, survival. 2. Neural Correlates of Infant Memory Consolidation In the neonate’s brain, areas normally associated with consolidation in the adult do not appear functionally mature. Specifically, the amygdala, hippocampus, and neocortex are considered to be key loci of drug and hormonal modulation and/or encoding of adult memory consolidation (reviewed in Abel and Lattal, 2001; Eichenbaum and Cohen, 2001; Fanselow and Gale, 2003; McGaugh, 2002; Packard and Cahill, 2001; Roozendaal, 2002; Schafe et al., 2001). However, as discussed earlier, these structures are not fully functional in infants. In the neonate, molecular events in the olfactory bulb are necessary for the association of an odor and a stimulus (McLean et al., 1999; Yuan et al., 2003; Zhang et al., 2003). The infant’s cascade of learning-induced molecular events is consistent with that in adults. The binding of a neurotransmitter to a receptor activates a cascade of events that in turn activates cyclic adenosine monophosphate (cAMP), a secondary messenger. cAMP then stimulates another enzyme, which causes phosphorylation of the cAMP response element-binding protein (CREB). CREB is another transcription factor that regulates expression of genes and ultimately proteins required for memory consolidation. Thus, changes in protein synthesis allow a longterm trace of the CS–unconditioned stimulus (UCS) association. Following odor–stroke conditioning in PN6 pups, there is a greater increase in phosphorylated CREB (pCREB) levels in pups that learn an odor preference in comparison with pups that do not learn, and these levels are highest 10–30 min following training (McLean et al., 1999). Similarly, there are marked increases in CREB levels 10–360 min following odor–shock pairings in PN11 pups (Zhang et al., 2003). In addition, CREB-deficient PN11 rats do not demonstrate memory of an odor aversion when tested 24 h following odor–shock training; however, CREB-deficient pups are able to demonstrate an odor aversion 1 h following the training, demonstrating

DEVELOPMENT OF LEARNING AND MEMORY

119

the crucial role of CREB in long-term memory consolidation (Zhang et al., 2003). More recent work from the Harley and McLean research team demonstrated that a cAMP and pCREB response follows the shifted role of NE through manipulations of olfactory bulb 5-HT (mitral cells have both NE and 5-HT receptors) (Yuan et al., 2003). Specifically, they were able to confirm the effectiveness of the UCS at producing odor learning (manipulating the strength of the reward and/or NE levels) through elevated olfactory bulb cAMP levels. Moreover, they demonstrated that ineffective UCSs (too-low or too-high levels of NE or 5-HT lesions) do not elevate cAMP levels and are directly correlated with the inverted U-shaped performance curve seen in pup learning data. This cascade of molecular events associated with learning and memory consolidation has been identified in a wide variety of species at different stages of development, suggesting that the molecular biology underlying memory storage is highly conserved across both development and species (e.g., Carew, 1996; Carew and Sutton, 2001). Nevertheless, while learninginduced molecular events appear conserved, the neural circuitry involved in memory consolidation shows marked changes with development. Overall, comparison of adult and infant rats shows that during the first postnatal week similar molecular events that support adult learning and memory are already present to mediate early learning experiences. However, similarly to acquisition, neonatal consolidation has features unique to the neonate, especially with respect to the critical role of the opioid system in consolidation of odor preferences. C. Expression of Infant Learning Expression involves retrieval of the established memory (reviewed in Abel and Lattal, 2001; Szapiro et al., 2002). Similar to acquisition and consolidation, behavioral studies have indicated that no single neurotransmitter or brain area is responsible for the expression of a memory. Neurotransmitters involved in expression in the adult include dopamine, norepinephrine, glucocorticoids, and opioids (reviewed in Barros et al., 2003; Roozendaal, 2002). Brain systems implicated in expression of a memory include the amygdala (reviewed in Barros et al., 2003; Eichenbaum and Cohen, 2001) and the hippocampus (reviewed in Abel and Lattal, 2001; Szapiro et al., 2002), with the frontal cortex modulating expression through extinction (Milad and Quirk, 2002). Expression of a memory also appears to require molecular mechanisms similar to those used in memory consolidation (reviewed in Abel and Lattal, 2001; Miller and Matzel, 2000; Nader et al., 2000; Szapiro et al., 2002).

120

TANIA L. ROTH et al.

1. Behavioral Correlates of Expression Despite advancing knowledge on both the behavioral and neural correlates of learning and memory consolidation in infants, far less is known of the neurobiology underlying memory expression. Unlike its necessary role in infant learning and memory consolidation, NE is not necessary for the expression of an odor preference following odor conditioning (Sullivan and Wilson, 1991). We have also demonstrated that endogenous opioids are necessary for the expression of a shock-induced odor preference in neonates; however, opioids are not necessary for the expression of a shockinduced odor aversion in older pups (Roth and Sullivan, 2003). Similarly, Shide and Blass (1991) demonstrated that opioids are necessary for the expression of a sucrose-conditioned odor preference in neonates. Isolation from the home cage and mother prior to testing has been shown to disrupt olfactory memories in 18-day-old rats (Arnold and Spear, 1995), suggesting that stimuli from the home environment (from the mother and siblings) affect physiological/biochemical processes necessary for maintaining and expressing memories. Finally, Sandstrom et al. (1998) have shown that expression of a conditioned odor aversion in preweanling rats (after odor–footshock training on PN12) is impaired by scopolamine, suggesting that the cholinergic system has a role in the retrieval and expression of odor memories in young rats. The age of training can have an enormous impact on what is expressed. This is illustrated by the fact that odor–shock-conditioned pups trained during the sensitive period (who learn an odor preference) continue to express the odor preference even after the sensitive period (Sullivan et al., 2000a). Thus, animals express a behavior consistent with the age of training (during the sensitive period) rather than the age of testing (postsensitive period). A similar example can be seen in research on conditioned odor potentiation of startle, which emerges around PN23 (Richardson et al., 2000). As noted above, the startle response can be potentiated by presentation of an odor previously paired with shock. However, odor–shock conditioning cannot potentiate the startle response in pups younger than PN23 (Richardson et al., 2000, 2003). These results illustrate the importance of the age of learning on the expression of the learned response. Attachment odors learned in infancy retain value into adulthood, although the role of the odor in modifying behavior changes. Research from the laboratories of C. Moore (Moore et al., 1996) and E. Blass (Fillion and Blass, 1986) demonstrates that adult male rats exhibit enhanced sexual performance when exposed to natural and artificial odors learned in infancy. These results support observations in other species of the role of early experience on adult mate preference, such as the parental and social influences on avian sexual imprinting (Slagsvold et al., 2002; ten Cate and Vos, 1999).

DEVELOPMENT OF LEARNING AND MEMORY

121

2. Neural Correlates of Expression Neonates show a modified olfactory bulb response to presentation of a learned odor. This response is expressed both during the sensitive period and lasts into adulthood (Pager, 1974; Moriceau and Sullivan, 2004a; Woo and Leon, 1987). Moreover, this modified olfactory bulb response is expressed to both natural maternal and artificial odors experienced in the nest (Sullivan et al., 1990), as well as to odors in controlled learning experiments (Johnson et al., 1995; Sullivan and Leon, 1986; Wilson and Leon, 1988b; Wilson and Sullivan, 1991; Wilson et al., 1987). This learning-associated olfactory bulb response is characterized by enhanced immediate-early gene activity (c-fos; Johnson et al., 1995; Woo et al., 1996), enhanced 2-DG uptake in focal, odorspecific glomeruli in response to the conditioned odor, modified single-unit response patterns of mitral/tufted cells near the enhanced glomerular foci (Wilson and Leon, 1988a; Wilson and Sullivan, 1990; Wilson et al., 1987), odor-induced intrinsic optical signals (Yuan et al., 2002), and olfactory bulb anatomical changes reflected in enlarged glomeruli within these foci (Woo et al., 1987). As with the behavioral changes in attachment, these neural changes are retained into adulthood, with acquisition dependent on experiences during infancy. These changes have not been found in animals conditioned after the sensitive period, suggesting that the brain may be specifically designed to give special significance to neonatal odors with hedonic value. Additional experimental evidence indicates that the neonate’s access to stored memory of odor conditioning differs developmentally (Kucharski and Hall, 1987, 1988; Kucharski et al., 1990). Specifically, olfactory memories can be unilaterally stored by occluding one naris during training (yielding a trained and untrained olfactory bulb). During testing, a PN6 pup can access memory only if tested with the trained side, while testing with the untrained side yields no memory of conditioning. But, because of the development of the projections of the anterior limb of the anterior commissure to and from the AON and the anterior piriform cortex, PN12 pups can access the memory through either the trained or untrained side. Indeed, this suggests the inclusion of the AON and perhaps the piriform cortex as sites of memory encoding in the olfactory pathway after the sensitive period (see above). III. Early Experiences Affect Brain and Behavior Infant experiences produce long-lasting changes in behavior, and researchers have only begun to document the neural modifications that presumably underlie the behavior in both humans and other animals

122

TANIA L. ROTH et al.

(reviewed in Glaser, 2000; Grossman et al., 2003; Gunnar, 2001; Levine, 2001; Machado and Bachevalier 2003; Meaney et al., 2002; Sanchez et al., 2001; Schore, 2001, 2002; Teicher et al., 2003). On the basis of the animal literature, it is becoming increasingly clear that an early stressful environment canalizes brain development to prepare the infant to accommodate a lifetime within a stressful environment, while an enriched infant environment prepares the brain to successfully cope with any environment. Clinical data support such an interpretation. For example, an abused child is better at detecting angry faces, and is more likely to interpret a situation as negative or dangerous (Pollak and Kistler, 2002). This attitude appears to carry over into adulthood, and corresponding neural changes have been documented in adults abused as children, most notably in the temporal lobe including the amygdala, the frontal cortex, LC, hippocampus, and cerebellum (Perry et al., 1995; Teicher et al., 1997; Vythilingam et al., 2002). In addition to the attachment model reviewed here, other animal models such as ones of deprivation and maternal caregiving (reviewed in Caldji et al., 1998; Kuhn and Schanberg, 1998; Levine, 2001; Liu et al., 2000; Meaney et al., 2002; Sanchez et al., 2001) are helping to differentiate between causation and correlations, and they are providing a more systematic assessment of the specific maternal behaviors regulating infant physiology and behavior (for reviews see Hofer, 2002; Hofer and Sullivan, 2001; Levine, 2001). Work from these models suggests that the LC, amygdala, hippocampus, frontal cortex, and hypothalamic–pituitary–adrenal (HPA) axis are all affected by maternal behaviors, and thus offer potential sites to understand the damaging effects of infant stress and maltreatment on subsequent behavioral development (Dent et al., 2001; reviewed in Francis et al., 1999; Gutman and Nemeroff, 2002; Heim and Nemeroff, 2001; Levine, 2001; Sanchez et al., 2001). The overlap in these structures and those active in the attachment system suggests a possible mechanism for the uniquely powerful effects of early experiences seen in clinical settings.

IV. Summary There are three notable take-home messages. First, neonatal learning has unique characteristics that seem to ensure that the infant will attach to its caregiver. These characteristics include rapid and persistent seeking of proximity to the caregiver, and remaining attached to the caregiver despite abusive behavior. This basic phenomenon of the infant learning–attachment system appears to be common to a wide range of species including rats, birds, nonhuman primates, and humans. Second, the neonatal learning

DEVELOPMENT OF LEARNING AND MEMORY

123

abilities appear to be due to the unique underlying neural circuitry. Specifically, the olfactory bulb encodes the learned response, the noradrenergic LC ensures rapid, robust preference conditioning, and the lack of amygdala participation in aversive conditioning may prevent avoidance learning. Third, it is possible that this unique learning system may be, at least in part, responsible for the dramatic, long-term effects of early experiences on adult mental health. Clinical data have long suggested that experiences within the infant–caregiver attachment dyad have a special role in neurobehavioral development, and the data outlined in this article may provide insights into the mechanisms underlying the powerful effects of early attachment. Acknowledgments This work was supported by grants NICHD-HD33402 and NSF-IBN0117234 to R. M. S.; by grant NIDCD-DC03906 and a grant from the Oklahoma Center for the Advancement of Science and Technology to D. A. W.; and by grant HHS-PHS NRSA F31 DA06082 to T. L. R.

References Abel, T., and Lattal, K. M. (2001). Molecular mechanisms of memory acquisition, consolidation, and retrieval. Curr. Opin. Neurobiol. 11, 180–187. Alberts, J. R. (1976). Olfactory contribution to behavioral development in rodents. In ‘‘Mammalian Olfaction, Reproductive Processes, and Behavior’’ (R. L. Doty, Ed.). Academic Press, New York. Alvarez, P., Wendelken, L., and Eichenbaum, H. (2002). Hippocampal formation lesions impair performance in an odor–shock association task independently of spatial context. Neurobiol. Learn. Mem. 78, 470–476. Arnold, H. M., and Spear, N. E. (1995). Isolation disrupts retention in preweanling rat pups. Behav. Neurosci. 109, 744–758. Astic, L., and Saucier, D. (1982). Metabolic mapping of functional activity in the olfactory projections of the rat: Ontogenetic study. Dev. Brain Res. 2, 141–156. Aston-Jones, G., and Bloom, F. E. (1981). Norepinephrine-containing locus coeruleus neurons in behaving rats exhibit pronounced responses to non-noxious environmental stimuli. J. Neurosci. 1, 887–900. Aston-Jones, G., Rajkowski, J., Kubiak, P., and Alexinsky, R. (1994). Locus coeruleus neurons in the monkey are selectively activated by attended stimuli in a vigilance task. J. Neurosci. 14, 4467–4480. Aston-Jones, G., Rajkowski, J., and Cohen, J. (1999). Role of locus coeruleus in attention and behavioral flexibility. Biol. Psychiatry 46, 1309–1320. Barr, G. A., and Rossi, G. (1992). Conditioned place preference from ventral tegmental injections of morphine in neonatal rats. Dev. Brain Res. 66, 133–136. Barros, D. M., Izquierdo, L. A., Medina, J. H., and Izquierdo, I. (2003). Pharmacological findings contribute to the understanding of the main physiological mechanisms of memory retrieval. Curr. Drug Target CNS Neurol. Disord. 2, 81–94.

124

TANIA L. ROTH et al.

Bartoshuk, L. M., and Beauchamp, G. K. (1994). Chemical Senses. Annu. Rev. Psychol. 45, 419–449. Bayer, S. A. (1980). Quantitative 3H-thymidine radiographic analysis of neurogenesis in the rat amygdala. J. Comp. Neurol. 194, 845–875. Berdel, B., Morys, J., and Maciejewska, B. (1997). Neuronal changes in the basolateral complex during development of the amygdala of the rat. Int. J. Dev. Neurosci. 15, 755–765. Best, A. R., and Wilson, D. A. (2004). Coordinate synaptic mechanisms contribute to cortical olfactory adaptation. J. Neurosci. 24, 652–660. Bezin, L., Marcel, D., Rousset, C., Pujol, J. F., and Weissmann, D. (1994a). Quantitative study of tyrosine hydroxylase protein levels within the somatic area of the rat locus coeruleus during postnatal development. J. Neurosci. 14, 7502–7510. Bezin, L., Marcel, D., Rousset, C., Pujol, J. F., and Weissmann, D. (1994b). Ontogeny of tyrosine hydroxylase levels in the neuropil close to locus coeruleus. Neuroreport 5, 1809–1812. Blozovski, D., and Cudennec, A. (1980). Passive avoidance learning in the young rat. Dev. Psychobiol. 13, 513–518. Bolles, R. C., and Woods, P. J. (1965). The ontogeny of behavior in the albino rat. Anim. Behav. 12, 427–441. Bowlby, J. (1965). ‘‘Attachment.’’ Basic Books, New York. Brake, S. C. (1981). Suckling infant rats learn a preference for a novel olfactory stimulus paired with milk delivery. Science 211, 506–508. Brennan, P. A., and Keverne, E. B. (1997). Neural mechanisms of mammalian olfactory learning. Prog. Neurobiol. 51, 457–481. Brunjes, P. C., and Frazier, L. L. (1986). Maturation and plasticity in the olfactory system of vertebrates. Brain Res. Rev. 11, 1–45. Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P. M., and Meany, M. J. (1998). Maternal care during infancy regulates the development of neural systems mediating the expression of fearfulness in the rat. Proc. Natl. Acad. Sci. USA 95, 5335–5340. Camp, L. L., and Rudy, J. W. (1988). Changes in the categorization of appetitive and aversive events during postnatal development of the rat. Dev. Psychobiol. 21, 25–42. Campbell, B. A. (1984). Reflections on the ontogeny of learning and memory. In ‘‘Comparative Perspectives on the Development of Memory’’ (R. Kail and N. E. Spear, Eds.), pp. 23–35. Lawrence Erlbaum Associates, Hillsdale, NJ. Carew, T. J. (1996). Molecular enhancement of memory formation. Neuron 16, 5–8. Carew, T. J., and Sutton, M. A. (2001). Molecular stepping stones in memory consolidation. Nat. Neurosci. 4, 769–771. Christie, M. J., and Jelinek, H. F. (1993). Dye-coupling among neurons of the rat locus coeruleus during postnatal development. Neuroscience 56, 129–137. Collier, A. C., Mast, J., Meyer, D. R., and Jacobs, C. E. (1979). Approach–avoidance conflict in preweanling rats: A developmental study. Anim. Learn. Behav. 7, 514–520. Crain, B., Cotman, C., Taylor, D., and Lynch, G. (1973). A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat. Brain Res. 63, 195–204. Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memory: A review. Psychol. Bull. 96, 518–559. DeCasper, A. J., and Fifer, W. P. (1980). Of human bonding: Newborns prefer their mothers’ voices. Science 208, 1174–1176. Dent, G. W., Smith, M. A., and Levine, S. (2001). Stress-induced alterations in locus coeruleus gene expression during ontogeny. Dev. Brain Res. 127, 23–30.

DEVELOPMENT OF LEARNING AND MEMORY

125

Dudai, Y. (2002). Molecular basis of long-term memories: A question of persistence. Curr. Opin. Neurobiol. 12, 211–216. Dumery, V., Derer, P., and Blozovski, D. (1988). Enhancement of passive avoidance learning through small doses of intra-amygdaloid physostegmine in the young rat: Its relation to the development of acetylcholinesterase. Dev. Psychobiol. 21, 553–565. Eichenbaum, H., and Cohen, N. J. (2001). Emotional memory and memory modulation. In ‘‘Conditioning to Conscious Recollection: Memory Systems of the Brain,’’ pp. 344–367, 394–434. University Press, Oxford. Emerich, D. F., Scalzo, F. M., Enters, E. K., Spear, N., and Spear, L. (1985). Effects of 6-hydroxydopamine-induced catecholamine depletion on shock-precipitated wall climbing of infant rat pups. Dev. Psychobiol. 18, 215–227. Fanselow, M. S., and Gale, G. D. (2003). The amygdala, fear, and memory. Ann. N. Y. Acad. Sci. 985, 125–134. Fanselow, M. S., and LeDoux, J. E. (1999). Why we think plasticity underlying Pavlovian fear conditioning occurs in the basolateral amygdala. Neuron 23, 229–232. Fanselow, M. S., and Rudy, J. W. (1998). Convergence of experimental and developmental approaches to animal learning and memory processes. In ‘‘Mechanistic Relationships between Development and Learning’’ (T. J. Carew, R. Menzel, and C. J. Shatz, Eds.), pp. 15–28. John Wiley & Sons, New York. Fillion, T. J., and Blass, E. M. (1986). Infantile experience with suckling odors determined adult sexual behavior in male rats. Science 231, 729–731. Fleming, A. S., O’Day, D. H., and Kraemer, G. W. (1999). Neurobiology of mother–infant interactions: Experience and central nervous system plasticity across development and generations. Neurosci. Biobehav. Rev. 23, 673–685. Francis, D. D., Caldji, C., Champagne, F., Plotsky, P. M., and Meaney, M. J. (1999). The role of corticotropin-releasing factor–norepinephrine systems in mediating the effects of early experience on the development of behavioral and endocrine responses to stress. Biol. Psychiatry 46, 1153–1166. Glaser, D. (2000). Child abuse and neglect and the brain—a review. J. Child Psychol. Psychiatry 41, 97–116. Grossman, A. W., Churchill, J. D., McKinney, B. C., Kodish, I. M., Otte, S. L., and Greenough, W. T. (2003). Experience effects on brain development: Possible contributions to psychopathology. J. Child Psychol. Psychiatry 44, 33–63. Gue´naire, C., Costa, J. C., and Delacour, J. (1982). Conditionnement operant avec renforcement thermique chez le rat nouveau-ne´. Physiol. Behav. 29, 419–424. Gunnar, M. R. (2001). Effects of early deprivation: Findings from orphanage-reared infants and children. In ‘‘Handbook of Developmental Cognitive Neuroscience’’ (C. A. Nelson and M. Luciana, Eds.), pp. 617–629. MIT Press, Cambridge, MA. Gutman, D. A., and Nemeroff, C. B. (2002). Neurobiology of early life stress: Rodent studies. Semin. Clin. Neuropsychiatry 7, 89–95. Hall, W. G. (1987). Early motivation, reward, learning, and their neural bases: Developmental revelations and simplifications. In ‘‘Perinatal Development: A Psychobiological Perspective’’ (N. A. Krasnegor, E. M. Blass, M. A. Hofer, and W. P. Smotherman, Eds.), pp. 169–193. Academic Press, New York. Hamrick, W. D., Wilson, D. A., and Sullivan, R. M. (1993). Neural correlates of memory for odor detection conditioning in adult rats. Neurosci. Lett. 163, 36–40. Harley, C. W., and Sara, S. J. (1992). Locus coeruleus burst induced by glutamate trigger delayed perforant path spike amplitude potentiation in the dentate gyrus. Exp. Brain Res. 89, 581–587.

126

TANIA L. ROTH et al.

Harlow, H. F., and Harlow, M. K. (1965). The affectional systems. In ‘‘Behavior of Nonhuman Primates’’ (A. Schrier, H. F. Harlow, and F. Stollnitz, Eds.), Vol. 2. Academic Press, New York. Haroutunian, V., and Campbell, B. A. (1979). Emergence of interoceptive and exteroceptive control of behavior in rats. Science 205, 927–929. Harris, G. C., and Fitzgerald, R. D. (1991). Locus coeruleus involvement in the learning of classically conditioned bradycardia. J. Neurosci. 11, 2314–2320. Hassmannova´, J., Myslivecek, J., and Romoliniova, A. (1985). Changes of learning and memory between the 3rd and 4th postnatal week in rats. Physiol. Bohemoslov. 34, 422. Heim, C., and Nemeroff, C. B. (2001). The role of childhood trauma in the neurobiology of mood and anxiety disorders: Preclinical and clinical studies. Biol. Psychiatry 49, 1023–1039. Helfer, M. E., Kempe, R. S., and Krugman, R. D. (1997). ‘‘The Battered Child.’’ University of Chicago Press, Chicago, IL. Hess, E. H. (1962). Ethology: An approach to the complete analysis of behavior. In ‘‘New Directions in Psychology’’ (R. Brown, E. Galanter, E. H. Hess, and G. Mendler, Eds.), pp. 159–199. Holt, Rinehart and Winston, New York. Hesse, E., and Main, M. (2000). Disorganized infant, child, and adult attachment: Collapse in behavioral and attentional strategies. J. Am. Psychoanal. Assoc. 48, 1097–1127. Hofer, M. A. (2002). Unexplained infant crying: An evolutionary perspective. Acta Paediatr. 91, 491–496. Hofer, M. A., and Sullivan, R. M. (2001). Toward a neurobiology of attachment. In ‘‘Handbook of Developmental Cognitive Neuroscience’’ (C. A. Nelson and M. Luciana, Eds.), pp. 599–616. MIT Press, Cambridge, MA. Hoffman, H., and Spear, N. E. (1989). Facilitation and impairment of conditioning in the preweanling rat after prior exposure to the conditioned stimulus. Anim. Learn. Behav. 17, 63–69. Hudson, R., and Distel, H. (1999). The flavor of life: Perinatal development of odor and taste preferences. Schweiz Med. Wochenschr. 129, 176–181. Hudson, R. E., Bachevalier, J., Doupe, A. J., Fanselow, M. S., Kuhl, P. K., Menzel, R., Morris, R. G. M., Rudy, J. W., and Squire, L. R. (1998). Group report: What does behavior tell us about the relationship between development and learning? In ‘‘Mechanistic Relationships Between Development and Learning’’ (T. J. Carew, R. Menzel, and C. J. Shatz, Eds.), pp. 75–92. John Wiley & Sons, New York. Hunt, P., and Campbell, B. A. (1999). Developmental dissociation of the components of conditioned fear. In ‘‘Learning, Motivation, and Cognition: The Functional Behaviorism of Robert C. Bolles’’ (M. E. Bouton and M. S. Fanselow, Eds.). American Psychological Association, Washington, D.C. Hunt, P. S., Richardson, R., and Campbell, B. A. (1994). Delayed development of fearpotentiated startle in rats. Behav. Neurosci. 108, 61–80. Hunt, P. S., Richardson, R., Hess, M. F., and Campbell, B. A. (1997). Emergence of conditioned cardiac responses to an olfactory CS paired with an acoustic startle UCS during development: Form and autonomic origins. Dev. Psychobiol. 31, 151–163. Insel, T. R., and Young, L. J. (2001). The neurobiology of attachment. Nat. Rev. Neurosci. 2, 129–136. Jay, T. M. (2003). Dopamine: A potential substrate for synaptic plasticity and memory mechanisms. Prog. Neurobiol. 69, 375–390. Johanson, I. B., and Hall, W. G. (1982). Appetitive conditioning in neonatal rats: Conditioned orientation to a novel odor. Dev. Psychobiol. 15, 379–397.

DEVELOPMENT OF LEARNING AND MEMORY

127

Johnson, B. A., Woo, C. C., Duong, H., Nguyen, V., and Leon, M. (1995). A learned odor evokes an enhanced Fos-like glomerular response in the olfactory bulb of young rats. Brain Res. 699, 192–200. Kehoe, P., and Blass, E. (1986). Central nervous system mediation of positive and negative reinforcement in neonatal albino rats. Dev. Brain Res. 27, 69–75. Kimura, F., and Nakamura, S. (1985). Locus coeruleus neurons in the neonatal rat: Electrical activity and responses to sensory stimulation. Dev. Brain Res. 23, 301–305. Kucharski, D., and Hall, W. G. (1987). New routes to early memories. Science 238, 786–788. Kucharski, D., and Hall, W. G. (1988). Developmental changes in the access to olfactory memories. Behav. Neurosci. 102, 340–348. Kucharski, D., Burka, N., and Hall, W. G. (1990). The anterior limb of the anterior commissure is an access route to contralaterally stored olfactory preference memories. Psychobiology 18, 195–204. Kuhn, C. M., and Schanberg, S. M. (1998). Responses to maternal separation: Mechanisms and mediators. Int. J. Dev. Neurosci. 16, 261–270. LaLumiere, R. T., Buen, T. V., and McGaugh, J. L. (2003). Post-training intra-basolateral amygdala infusions of norepinephrine enhance consolidation of memory for contextual fear conditioning. J. Neurosci. 23, 6754–6758. Landers, M. S., and Sullivan, R. M. (1999). Vibrissae-evoked behavior and conditioning prior to functional ontogeny of the somatosensory vibrissae cortex. J. Neurosci. 19, 5131–5137. Langdon, P. E., Harley, C. W., and McLean, J. H. (1997). Increased  adrenoceptor activation overcomes conditioned olfactory learning induced by serotonin depletion. Dev. Brain Res. 114, 261–264. Levine, S. (2001). Primary social relationships influence the development of the hypothalamic– pituitary–adrenal axis in the rat. Physiol. Behav. 73, 255–260. Levy, F. (2002). Neurobiological mechanisms involved in recognition of olfactory signature of the young in sheep. J. Soc. Biol. 196, 77–83. Liang, K. C. (1998). Pretraining infusion of DSP-4 into the amygdala impaired retention in the inhibitory avoidance task: Involvement of norepinephrine but not serotonin in memory facilitation. Chinese J. Physiol. 41, 223–233. Lincoln, J., Coopersmith, R., Harris, E. W., Cotman, C. W., and Leon, M. (1988). NMDA receptor activation and early olfactory learning. Brain Res. 467, 309–312. Litaudon, P., Mouly, A. M., Sullivan, R., Gervais, R., and Cattarelli, M. (1997). Learninginduced changes in rat piriform cortex activity mapped using multisite recording with voltage sensitive dye. Eur. J. Neurosci. 9, 1593–1602. Liu, D., Diorio, J., Day, J. C., Francis, D. D., and Meany, M. J. (2000). Maternal care, hippocampal synaptogenesis and cognitive development in rats. Nat. Neurosci. 3, 799–806. Machado, C. J., and Bachevalier, J. (2003). Non-human primate models of childhood psychopathology: The promise and the limitations. J. Child Psychol. Psychiatry 44, 64–87. Maestripieri, D., Wallen, K., and Carroll, K. A. (1997). Infant abuse runs in families of groupliving pigtail macaques. Child Abuse Negl. 21, 465–471. Maestripieri, D., Tomaszycki, M., and Carroll, K. A. (1999). Consistency and change in the behavior of rhesus macaque abusive mothers with successive infants. Dev. Psychobiol. 34, 29–35. Maren, S. (2001). Neurobiology of Pavlovian fear conditioning. Annu. Rev. Neurosci. 24, 897–931. Marshall, K. C., Christi, M. M., Finlayson, P. G., and Williams, J. T. (1991). Developmental aspects of the locus coeruleus–noradrenaline system. Prog. Brain Res. 88, 173–185.

128

TANIA L. ROTH et al.

Martin, L. T., and Alberts, J. R. (1979). Taste aversion to mother’s milk: The age-related role of nursing in acquisition and expression of a learned association. J. Comp. Physiol. Psychol. 93, 430–445. McGaugh, J. L. (2002). Memory consolidation and the amygdala: A systems perspective. Trends Neurosci. 25, 456–461. McGaugh, J. L., and Roozendaal, B. (2002). The role of adrenal stress hormone in forming lasting memories in the brain. Curr. Opin. Neurobiol. 12, 205–210. McGaugh, J. L., Introini-Collison, I. B., et al. (1993). Involvement of opioid peptides in learning and memory. In ‘‘Handbook of Experimental Pharmacology: Opioids II’’ (A. Herz, Ed.), Vol. 104/II, pp. 429–447. Springer-Verlag, New York. McLean, J. H., and Shipley, M. T. (1991). Postnatal development of the noradrenergic projection from the locus coeruleus to the olfactory bulb in the rat. J. Comp. Neurol. 304, 469–477. McLean, J. H., Darby-King, A., Sullivan, R. M., and King, S. R. (1993). Serotonergic influences on olfactory learning in the neonatal rat. Behav. Neural Biol. 60, 152–162. McLean, J. H., Harley, C. W., Darby-King, A., and Yuan, Q. (1999). pCREB in the neonate rat olfactory bulb is selectively and transiently increased by odor preference-conditioned training. Learn. Mem. 6, 608–618. Meaney, M. J., Brake, W., and Gratton, A. (2002). Environmental regulation of the development of mesolimbic dopamine systems: A neurobiological mechanism for vulnerability to drug abuse? Psychoneuroendocrinology 27, 127–138. Meneses, A. (2003). A pharmacological analysis of an associative learning task: 5-HT1 to 5-HT7 receptor subtypes function on a pavlovian/instrumental autoshaped memory. Learn. Mem. 10, 363–372. Mickley, G. A., Schaldach, M. A., Snyder, K. J., Balagh, S. A., Len, T., Neimanis, K., Gaulis, P., Hug, J., Sauchak, K., Remmers-Roeber, D. R., Walker, C., and Yamamoto, B. K. (1998). Ketamine blocks a conditioned taste aversion in neonate rats. Physiol. Behav. 64, 381–390. Milad, M. R., and Quirk, G. J. (2002). Neurons in medial prefrontal cortex signal memory for fear extinction. Nature 420, 70–74. Miller, R. R., and Matzel, L. D. (2000). Memory involves far more than ‘‘consolidation.’’ Nat. Rev. 1, 214–216. Misanin, J. R., Guanowsky, V., and Riccio, D. C. (1983). The effects of CS-preexposure on conditioned taste aversion in young and adult rats. Physiol. Behav. 30, 859–862. Mizukawa, K., Tseng, I.-M., and Otsuka, N. (1989). Quantitative electron microscopic analysis of postnatal development of zinc-positive nerve endings in the rat amygdala using Timm’s sulphide silver technique. Dev. Brain Res. 50, 197–203. Moore, C. L., Jordan, L., and Wong, L. (1996). Early olfactory experience, novelty and choice of sexual partner by male rats. Physiol. Behav. 60, 1361–1367. Moriceau, S., and Sullivan, R. M. (2004a). Corticosterone influences on mammalian neonatal sensitive-period learning. Behav. Neurosci. 118, 274–281. Moriceau, S. A., and Sullivan, R. M. (2004b). Unique neural circuitry for neonatal olfactory learning. J. Neurosci. 24, 1182–1189. Moriceau, S., Roth, T. L., Okotoghaide, T., and Sullivan, R. M. (2004). Corticosterone controls the developmental emergence of fear and amygdala function to predator odor in infant rat pups. Int. J. Dev. Sci. 22, 415–422. Morton, N., and Browne, K. D. (1998). Theory and observation of attachment and its relation to child maltreatment: A review. Child Abuse Negl. 22, 1093–1104. Mullen, P. E., Martin, J. L., Anderson, J. C., Romans, S. E., and Herbison, G. P. (1994). The effect of child sexual abuse on social, interpersonal and sexual function in adult life. Br. J. Psychiatry 165, 35–47.

DEVELOPMENT OF LEARNING AND MEMORY

129

Myslivecek, J. (1997). Inhibitory learning and memory in newborn rats. Prog. Neurobiol. 53, 399–430. Nader, K., Schafe, G. E., and LeDoux, J. E. (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 406, 722–726. Nair, H. P., and Gonzalez-Lima, F. (1999). Extinction of behavior in infant rats: Development of functional coupling between septal, hippocampal, and ventral tegmental regions. J. Neurosci. 19, 8646–8655. Nakamura, S. T., and Sakaguchi, T. (1990). Development and plasticity of the locus coeruleus: A review of recent physiological and pharmacological experimentation. Prog. Neurobiol. 34, 505–526. Nowak, R., and Lindsay, D. R. (1992). Discrimination of merino ewes by their newborn lambs: Important for survival? Appl. Anim. Behav. Sci. 34, 61–74. Nowak, R., Murphy, T. M., Lindsay, D. R., Alster, P., Anderson, R., and Uvnas-Moberg, K. (1997). Development of a preferential relationship with the mother by the newborn lamb: Importance of the sucking activity. Physiol. Behav. 62, 681–688. O’Connor, T. G., and Rutter, M. (2000). Attachment disorder behavior following early severe deprivation: Extension and longitudinal follow-up. English and Romanian Adoptees Study Team. J. Am. Acad. Child Adolesc. Psychiatry 39, 703–712. Okutani, F., Zhang, J. J., Yagi, F., and Kaba, H. (2002). Non-specific olfactory aversion induced by intrabulbar infusion of the GABAA receptor antagonist bicuculline in young rats. Neuroscience 112, 901–906. Packard, M. G., and Cahill, L. (2001). Affective modulation of multiple memory systems. Curr. Opin. Neruobiol. 11, 752–756. Pager, J. (1974). A selective modulation of olfactory bulb electrical activity in relation to the learning of palatability in hungry and satiated rats. Physiol. Behav. 12, 189–195. Panksepp, J., Nelson, E., and Siviy, S. (1994). Brain opioids and mother–infant social motivation. Acta Pediatr. Suppl. 397, 40–46. Paschall, G. Y., and Davis, M. (2002). Second-order olfactory-mediated fear-potentiated startle. Leam. Mem. 9, 395–401. Pedersen, P., Williams, C. L., and Blass, E. M. (1982). Activation and odor conditioning of suckling behavior in 3-day-old albino rats. J. Exp. Psychol. Anim. Behav. Proc. 8, 329–341. Perry, B. D., Pollard, R., Blakely, T., Baker, W., and Vigilante, D. (1995). Childhood trauma, the neurobiology of adaptation and ‘‘use-dependent’’ development of the brain: How ‘‘states’’ become ‘‘traits.’’ Infant Ment. Health J. 16, 271–291. Pollak, S. D., and Kistler, D. J. (2002). Early experience is associated with the development of categorical representations for facial expressions of emotion. Proc. Natl. Acad. Sci. USA 99, 9072–9076. Rajecki, D. W., Lamb, M. E., and Obmascher, P. (1978). Towards a general theory of infantile attachment: A comparative review of aspects of the social bond. Behav. Brain Sci. 3, 417–464. Ramus, S. J., and Eichenbaum, H. (2000). Neural correlates of olfactory recognition memory in the rat orbitofrontal cortex. J. Neurosci. 20, 8199–8208. Rangel, S., and Leon, M. (1995). Early odor preference training increases olfactory bulb norepinephrine. Dev. Brain. Res. 85, 187–191. Richardson, R., Paxinos, G., and Lee, J. (2000). The ontogeny of conditioned odor potentiation of startle. Behav. Neurosci. 114, 1167–1173. Richardson, R., Fan, M., and Parnas, A. S. (2003). Latent inhibition of conditioned odor potentiation of startle: A developmental analysis. Dev. Psychobiol. 42, 261–268. Riedel, G., Platt, B., and Micheau, J. (2003). Glutamate receptor function in learning and memory. Behav. Brain Res. 140, 1–47.

130

TANIA L. ROTH et al.

Robinson, S. R., and Smotherman, W. P. (1995). Habituation and classical conditioning in the rat fetus: Opioid involvements. In ‘‘Fetal Development—a Psychobiological Perspective’’ (J. P. Lecanuet, W. P. Fifer, N. A. Krasnegor, and W. P. Smotherman, Eds.), pp. 295–314. Lawrence Erlbaum Associates, Hillsdale, NJ. Roozendaal, B. (2002). Stress and memory: Opposing effects of glucocorticoids on memory consolidation and memory retrieval. Neurobiol. Leam. Mem. 78, 578–595. Rosenblatt, J. S. (1983). Olfaction mediates developmental transition in the altricial newborn of selected species of mammals. Dev. Psychobiol. 16, 347–375. Roth, T. L., and Sullivan, R. M. (2001). Endogenous opioids and their role in odor preference acquisition and consolidation following odor–shock conditioning in infant rats. Dev. Psychobiol. 39, 188–198. Roth, T. L., and Sullivan, R. M. (2003). Consolidation and expression of a shock-induced odor preference in rat pups is facilitated by opioids. Physiol. Behav. 78, 135–142. Rudy, J. W., and Cheatle, M. D. (1977). Odor aversion learning in neonatal rats. Science 198, 845–846. Rudy, J. W., and Morledge, P. (1994). Ontogeny of contextual fear conditioning: Implications for consolidation, infantile amnesia, and hippocampal system function. Behav. Neurosci. 108, 227–234. Rudy, J. W., Stadler-Morris, S., and Alberts, P. A. (1987). Ontogeny of spatial navigation behaviors in the rat: Dissociation of ‘‘proximal-’’ and ‘‘distal-cue’’ based behaviors Behav. Neurosci. 101, 62–73. Saar, D., Grossman, Y., and Barkai, E. (1999). Reduced synaptic facilitation between pyramidal neurons in the piriform cortex after odor learning. J. Neurosci. 19, 8616–8622. Saar, D., Grossman, Y., and Barkai, E. (2002). Learning-induced enhancement of postsynaptic potentials in pyramidal neurons. J. Neurophysiol. 87, 2358–2363. Samama, B., and Boehm, N. (1999). Inhibition of nitric oxide synthase impairs early olfactory associative learning in newborn rats. Neurobiol. Learn. Mem. 71, 219–231. Sanchez, M. M., Ladd, C. O., and Plotsky, P. M. (2001). Early adverse experience as a developmental risk factor for later psychopathology: Evidence from rodent and primate models. Dev. Psychopathol. 13, 419–449. Sandstrom, N. J., Arnold, H. M., and Williams, C. L. (1998). Reactivation treatment prevents the memory-impairing effects of scopolamine in preweanling rats. Behav. Neurosci. 112, 909–919. Sara, S. J., Dyon-Laurent, D., and Herve, A. (1995). Novelty seeking behavior in the rat is dependent upon the integrity of the noradrenergic system. Cogs. Brain Res. 2, 181–187. Schaal, B., Orgeur, P., and Arnould, C. (1995). Olfactory preferences in newborn lambs: Possible influence of prenatal experience. Behavior 132, 351–365. Schafe, G. E., Nader, K., Blair, H. T., and LeDoux, J. E. (2001). Memory consolidation of Pavlovian fear conditioning: A cellular and molecular perspective. Trends Neurosci. 24, 540–546. Schoenbaum, G., and Eichenbaum, H. (1995). Information coding in the rodent prefrontal cortex. I. Single-neuron activity in orbitofrontal cortex compared with that in pyriform cortex. J. Neurophysiol. 74, 733–750. Schore, A. N. (2001). The effects of early relational trauma on right brain development, affect regulation, and infant mental health. Infant Ment. Health J. 22, 201–269. Schore, A. N. (2002). Dysregulation of the right brain: A fundamental mechanism of traumatic attachment and the psychopathogenesis of posttraumatic stress disorder. Aust. N.Z. J. Psychiatry 36, 9–30. Schwob, J. E., and Price, J. L. (1984). The development of axonal connections in the central olfactory system of rats. J. Comp. Neurol. 223, 177–202.

DEVELOPMENT OF LEARNING AND MEMORY

131

Schwob, J. E., Haberly, L. B., and Price, J. L. (1984). The development of physiological responses of the piriform cortex in rats to stimulation of the lateral olfactory tract. J. Comp. Neurol. 223, 223–237. Shayit, M., and Weller, A. (2001). Cholecystokinin receptor antagonists increase the rat pup’s preference toward maternal-odor and rug texture. Dev. Psychobiol. 38, 164–173. Shide, D. J., and Blass, E. M. (1991). Opioid mediation of odor preferences induced by sugar and fat in 6-day-old rats. Physiol. Behav. 50, 961–966. Shipley, M. T., and Ennis, M. (1996). Functional organization of olfactory system. J. Neurobiol. 30, 123–176. Shipley, M. T., Halloran, F. J., and De la Torre, J. (1985). Surprisingly rich projection from locus coeruleus to the olfactory bulb in the rat. Brain Res. 239, 294–299. Slagsvold, T., Hansen, B. T., Johannessen, L. E., and Lifjeld, J. T. (2002). Mate choice and imprinting in birds studied by cross-fostering in the wild. Proc. R. Soc. Lond B. 269, 1449–1455. Smallbone, S. W., and McCabe, B. A. (2003). Childhood attachment, childhood sexual abuse, and onset of masturbation among adult sexual offenders. Sex Abuse. 15, 1–9. Spear, N. E., and Rudy, J. W. (1991). Tests of the ontogeny of learning and memory: Issues, methods, and results. In ‘‘Developmental Psychobiology: New Methods and Changing Concepts’’ (H. N. Shair, G. A. Barr, and M. A. Hofer, Eds.), pp. 84–113. Oxford University Press, New York. Spear, N. E., Kucharski, D., and Miller, J. S. (1989). The CS-effect in simple conditioning and stimulus selection during development. Anim. Learn. Behav. 17, 70–82. Stanton, M. E. (2000). Multiple memory systems, development and conditioning. Behav. Brain Res. 110, 25–37. Stehouwer, D. J., and Campbell, B. A. (1978). Habituation of the forelimb-withdrawal response in neonatal rats. J. Exp. Psychol. Anim. Behav. Processes 4, 104–119. Stork, O., and Pape, H. (2002). Fear memory and the amygdala: Insights from a molecular perspective. Cell Tissue Res. 310, 271–277. Styron, T., and Janoff-Bulman, R. (1997). Childhood attachment and abuse: Long-term effects on adult attachment, depression, and conflict resolution. Child Abuse Negl. 21, 989–991. Sullivan, R. M. (2001). Unique characteristics of neonatal classical conditioning: The role of the amygdala and locus coeruleus. Integr. Physiol. Behav. Sci. 36, 293–307. Sullivan, R. M., and Leon, M. (1986). Early olfactory learning induces an enhanced olfactory bulb response in young rats. Dev. Brain Res. 27, 278–282. Sullivan, R. M., and Wilson, D. A. (1991). The role of norepinephrine in the expression of learned olfactory neurobehavioral responses in infant rats. Psychobiology 19, 308–312. Sullivan, R. M., and Wilson, D. A. (1993). Role of the amygdala complex in early olfactory associative learning. Behav. Neurosci. 107, 254–263. Sullivan, R. M., and Wilson, D. A. (1994). The locus coeruleus, norepinephrine, and memory in newborns. Brain Res. Bull. 35, 467–472. Sullivan, R. M., and Wilson, D. A. (2003). Molecular biology of early olfactory memory. Leam. Mem. 10, 1–4. Sullivan, R. M., Wilson, D. A., Wong, R., Correa, A., and Leon, M. (1990). Modified behavioral olfactory bulb responses to maternal odors in preweanling rats. Dev. Brain Res. 53, 243–247. Sullivan, R. M., Taborsky-Barba, S., Mendoze, R., Itano, A., Leon, M., Cotman, C. W., Payne, T. F., and Lott, I. (1991). Olfactory classical conditioning in neonates. Pediatrics 87, 511–518. Sullivan, R. M., Landers, M., Yeaman, B., and Wilson, D. A. (2000a). Good memories of bad events in infancy: Ontogeny of conditioned fear and the amygdala. Nature 407, 38–39.

132

TANIA L. ROTH et al.

Sullivan, R. M., Stackenwalt, G., Nasr, F., Lemon, C., and Wilson, D. A. (2000b). Association of an odor with activation of olfactory bulb noradrenergic -receptors or locus coeruleus stimulation is sufficient to produce learned approach response to that odor in neonatal rats. Behav. Neurosci. 114, 957–962. Szapiro, G., Galante, J. M., Barros, D. M., Levi de Stein, M., Vianna, M. R. M., Izquierdo, L. A., Izquierdo, I., and Medina, J. H. (2002). Molecular mechanisms of memory retrieval. Neurochem. Res. 27, 1491–1498. Takahashi, L. K. (1994). Organizing action of corticosterone on the development of behavioral inhibition in the preweanling rat. Dev. Brain Res. 81, 121–127. Teicher, M. H., Ito, Y., Gold, C. A., Andersen, S. L., Dumont, N., and Ackerman, E. (1997). Preliminary evidence for abnormal cortical development in physically and sexually abused children using EEG coherence and MRI. Ann. N. Y. Acad. Sci. 821, 160–175. Teicher, M. N., Andersen, S. L., Polcari, A., Anderson, C. M., Navalta, C. P., and Kim, D. M. (2003). The neurobiological consequences of early stress and childhood maltreatment. Neurosci. Biobehav. Rev. 27, 33–44. ten Cate, C., and Vos, D. R. (1999). Sexual imprinting and evolutionary processes in birds: A reassessment. Adv. Study Behav. 28, 1–31. Thoman, E., Wetzel, A., and Levine, S. (1968). Learning in the neonatal rat. Anim. Behav. 16, 54–57. Tronel, S., and Sara, S. J. (2002). Mapping of olfactory memory circuits: Region specific c-fos activation after odor-reward associative learning or after its retrieval. Learn. Mem. 9, 105–111. Vankov, A., Herve-Minvielle, A., and Sara, S. J. (1995). Response to novelty and its rapid habituation in locus coeruleus neurons of the freely exploring rat. Eur. J. Neurosci. 7, 1180–1187. Verwer, R. W., Van Vulpen, E. H., and Van Uum, J. F. (1996). Postnatal development of amygdaloid projections to the prefrontal cortex in the rat studied with retrograde and anterograde tracers. J. Comp. Neurol. 376, 75–96. Vythilingam, M., Heim, C., Newport, J., Miller, A. H., Anderson, E., Bronen, R., Brummer, M., Staib, L., Vermetten, E., Charney, D. S., Nemeroff, C. B., and Bremme, J. D. (2002). Childhood trauma associated with smaller hippocampal volume in women with major depression. Am. J. Psychiatry 159, 2072–2080. Walker, D. L., and Davis, M. (2002). The role of amygdala glutamate receptors in fear learning, fear-potentiated startle, and extinction. Pharmacol. Biochem. Behav. 71, 379–392. Weldon, D. A., Travis, M. L., and Kennedy, D. A. (1991). Post-training D1 receptor blockade impairs odor conditioning in neonatal rats. Behav. Neurosci. 105, 450–458. Weldon, D. A., Fedorcik, G. C., LoRusso, C. M., Tiburzi, M. J., and Lenoci, J. M. (1997). Olfactory conditioning impairment following posttraining NMDA receptor blockade in neonatal rats. Neurobiol. Learn. Mem. 67, 34–42. White, N. R., Adox, R., Reddy, A., and Barfield, R. J. (1992). Regulation of rat maternal behavior by broadband pup vocalizations. Behav. Neural Biol. 58, 131–137. Wiedenmayer, C. P., and Barr, G. A. (2001). Developmental changes in c-fos expression to an age-specific social stressor in infant rats. Behav. Brain Res. 126, 147–157. Wilensky, A. E., Schafe, G. E., and LeDoux, J. E. (1999). Functional inactivation of the amygdala before but not after auditory fear conditioning prevents memory formation. J. Neurosci. 19, RC48. Wilson, D. A. (2000). A comparison of odor receptive field plasticity in the rat olfactory bulb and anterior piriform cortex. J. Neurophysiol. 84, 3036–3042.

DEVELOPMENT OF LEARNING AND MEMORY

133

Wilson, D. A. (2003). Ontogeny of sensory-evoked single-unit responses in rat amygdala [Abstract 11]. In ‘‘International Society for Developmental Psychobiology,’’ 36th Annual Meeting, November 5–8, 2003, New Orleans, LA. Wilson, D. A., and Leon, M. (1988a). Noradrenergic modulation of olfactory bulb excitability in the postnatal rat. Dev. Brain Res. 42, 69–75. Wilson, D. A., and Leon, M. (1988b). Spatial patterns of olfactory bulb single-unit responses to learned olfactory cues in young rats. J. Neurophysiol. 59, 1770–1782. Wilson, D. A., and Stevenson, R. J. (2003). The fundamental role of memory in olfactory reception. Trends Neurosci. 26, 243–247. Wilson, D. A., and Sullivan, R. M. (1990). Olfactory associative conditioning in infant rats with brain stimulation as a reward. I. Neurobehavioral consequences. Dev. Brain Res. 53, 215–221. Wilson, D. A., and Sullivan, R. M. (1991). Olfactory associative conditioning in infant rats with brain stimulation as reward. II. Norepinephrine mediates a specific component of the bulb response to reward. Behav. Neurosci. 105, 843–849. Wilson, D. A., and Sullivan, R. M. (1992). Blockade of mitral/tufted cell habituation to odors by association with reward: A preliminary note. Brain Res. 594, 143–145. Wilson, D. A., and Sullivan, R. M. (2003). Sensory physiology of central olfactory pathways. In ‘‘Handbook of Olfaction and Gustation’’ (R. L. Doty, Ed.), pp. 181–201. Marcel Dekker, New York. Wilson, D. A., Sullivan, R. M., and Leon, M. (1987). Single-unit analysis of postnatal olfactory learning: Modified olfactory bulb output response patterns to learned attractive odors. J. Neurosci. 7, 3154–3162. Wilson, D. A., Pham, T., and Sullivan, R. M. (1994). Norepinephrine and posttraining memory consolidation in neonatal rats. Behav. Neurosci. 108, 1053–1058. Winzer-Serhan, U. H., and Leslie, F. M. (1999). Expression of 2A adrenoceptors during rat neocortical development. J. Neurobiol. 38, 259–269. Woo, C. C., and Leon, M. (1987). Sensitive period for neural and behavioral responses to learned odors. Dev. Brain Res. 36, 309–313. Woo, C. C., Coopersmith, R., and Leon, M. (1987). Localized changes in olfactory bulb morphology associated with early olfactory learning. J. Comp. Neurol. 263, 113–125. Woo, C. C., Oshita, M. H., and Leon, M. (1996). A learned odor response decreases the number of Fos-immunopositive granule cells in the olfactory bulb of young rats. Brain Res. 716, 149–156. Yuan, Q., Harley, C. W., McLean, J. H., and Kno¨pfel, T. (2002). Optical imaging of odor preference memory in the rat olfactory bulb. J. Neurophysiol. 87, 3156–3159. Yuan, Q., Harley, C. W., and McLean, J. H. (2003). Mitral cell  1 and 5-HT2A receptor colocalization and cAMP coregulation: A new model of norepinephrine-induced learning in the olfactory bulb. Learn. Mem. 10, 5–15. Zhang, J. J., Okutani, F., Inoue, S., and Kaba, H. (2003). Activation of the cyclic AMP response element-binding protein signaling pathway in the olfactory bulb is required for the acquisition of olfactory aversive learning in young rats. Neuroscience 117, 707–713.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Evolutionary Significance of Sexual Cannibalism Mark A. Elgar* and Jutta M. Schneider{ *department of zoology university of melbourne 3010 victoria, australia { institute of evolutionary biology and ecology university of bonn d-53121 bonn, germany

I. Introduction Several animal taxa have gained notoriety in the public imagination through their sexually cannibalistic behavior, in which the female attacks and consumes her male mate at some stage during courtship or mating. Historically, sexual cannibalism was usually regarded as anomalous behavior, offending the notion that the sexes are best served by a relatively harmonious courtship and copulation. Nevertheless, it was at the same time suggested that the risk of sexual cannibalism selected for male morphological or behavioral traits that provided some defense against female attack (see Elgar, 1992). More typically, male courtship behavior was thought to reduce the female’s predatory instincts and induce sexual receptivity. Now, sexual cannibalism is usually regarded as a particularly dramatic manifestation of sexual conflict, thereby fitting comfortably within the contemporary view of reproductive behavior. However, remarkably few studies have investigated the evolutionary significance of sexual cannibalism, perhaps because it does not have a broad taxonomic distribution, and thus its explanation may not be thought to provide more general insights into animal mating systems. As a result, the evolutionary significance of the behavior is still poorly understood. Given that males of sexually reproducing species are usually under strong selection to maximize fertilization success, typically by securing as many mates as possible, dying before or during mating may not appear to be in the victim’s best interest. Consequently, research on sexual cannibalism has focused on identifying potential benefits for the aggressor and/or 135 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

136

ELGAR AND SCHNEIDER

the victim that could compensate for the cost of losing future reproductive value. Importantly, the costs and benefits of sexual cannibalism to males and females depend critically on whether it occurs before or after insemination (Elgar, 1992). Accordingly, the timing of sexual cannibalism has profound implications for the extent to which the behavior can be regarded as a male and/or female mating strategy. Sexual cannibalism before mating is unambiguously not in the interest of the male, but postinsemination sexual cannibalism may be a strategy of either or both sexes. In this article, we discuss sexual cannibalism in the context of both sexual and natural selection. First, we describe briefly the natural history and taxonomic distribution of sexual cannibalism, drawing on earlier reviews and subsequent publications. Then we outline the various explanations of sexual cannibalism, emphasizing how the timing of sexual cannibalism changes the costs and benefits to males and females. A central theme here is that there is unlikely to be a single explanation of sexual cannibalism, but rather the behavior is likely to have several origins and, in certain circumstances, may have evolved as an integral part of a monogynous mating system. The main focus of this article is to examine the various explanations of sexual cannibalism and to evaluate their evidence. Our review concentrates on spiders, primarily because most studies of the evolutionary significance of sexual cannibalism have been on this taxon.

II. Natural History and Taxonomic Distribution Sexual cannibalism refers to females killing and consuming their male partner at some stage during courtship and mating (Elgar, 1992). While males have been observed to kill females, this reversed form of sexual cannibalism most probably reflects opportunistic foraging. For example, male amphipod Gammarus are more likely to attack females when there are no alternative prey items (Dick, 1995). Sexual cannibalism implies that the male perishes, but female attacks can result in loss of limbs only (see Elgar, 1992), suggesting that the definition could be relaxed to include the consumption of body parts and/or hemolymph. For example, female sagebrush crickets Cyphoderris strepitans feed on the fleshy hind wings of the male during copulation (Eggert and Sakaluk, 1994), and the female linyphiid spider Baryphyma pratense pierces the cuticle of her male mate and imbibes his hemolymph (Blest, 1987). Such a broad definition might be problematic because it could include species in which females feed on male secretions (see Vahed, 1998), in which case, male Zeus bugs Phoreticovelia disparata that feed on material secreted by the female (Arnqvist et al., 2003) would be included as sex-reversed cannibals. Furthermore, a broader

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

137

definition masks the important point that males confronted with a sexually cannibalistic female may severely compromise future reproduction. Thus, it is helpful to distinguish the provision of somatic gifts that refers to the consumption of parts of the male’s body but does not necessarily cause his death, from sexual cannibalism that more typically results in the death of the male (see also Simmons and Parker, 1989; Vahed, 1998). While the victims of sexual cannibalism clearly cannot mate again, the evolutionary significance of sexual cannibalism will vary according to the probability of female attack and the ability of males to escape. For example, most females of the golden orb-web spider Nephila plumipes attempt to attack and cannibalize males both before and during copulation (Elgar and Fahey, 1996; Schneider and Elgar, 2001, 2002) (see Fig. 1). However, this behavior is considerably less frequent among females of the sympatric species Nephila edulis (Elgar et al., 2003a; Schneider et al., 2000). Some aspects of male behavior may reduce the likelihood or success of female attacks. For example, male praying mantids Iris oratoria approach females from a specific direction that reduces the probability of attack (Maxwell, 1999). Males of some orb-web spiders court or mate with the female only when she is busy processing a recently captured prey item (e.g., Elgar and Fahey, 1996; Prenter et al., 1994; Schneider and Elgar, 2001). The prey must be sufficiently large to keep the female busy for more than a few seconds because once it is immobilized, the female switches attention toward the copulating male. Males may also attempt to mate with a female when she is molting to maturity. Females are highly vulnerable during the process of molting, and cannot move, which may allow males to copulate in comparative safety (Foellmer and Fairbairn, 2003; Robinson and Robinson, 1980). This strategic choice of mating opportunities may be confined only to sexually cannibalistic species; males of the less aggressive N. edulis mate repeatedly with the same female and capturing a prey item is not necessary to initiate courtship and copulation (Elgar et al., 2003a). Females of both N. plumipes and N. edulis regularly attempt to terminate copulation by attacking the male, but N. edulis males often escape such an attack. In contrast, males of N. plumipes escape less often, but may remain attached to the female even when dead (see below). In some species, there is considerable variation in the motivation of females to attack males, and in the ability of males to avoid being a victim. Males of Nephila that survive a female’s attack may lose one or two legs and their ability to copulate again may be limited by the number of legs that remain. In contrast, males of the theridiid spider Tidarren argo copulate once only (Knoflach and van Harten, 2001) and males of the orb-weaving spiders Argiope aemula (Sasaki and Iwahashi, 1995), A. aurantia (Foellmer and Fairbairn, 2003), and A. keyserlingi (Gaskett et al., 2004) never survive

138

ELGAR AND SCHNEIDER

Fig. 1. A dead male Nephila plumipes remains attached to the female by his pedipalp, which has become lodged within her reproductive tract. Photograph ß David Paul.

two palpal insertions with the same female (see below). Thus, all of these species are effectively monogynous (Fromhage et al., manuscript). Elgar (1992) provides an extensive, but not definitive, survey of the records of sexual cannibalism. This survey reveals that sexual cannibalism is apparently widely distributed among invertebrates, including gastropods, copepods, insects, and especially arachnids, thereby suggesting that it has evolved several times independently. However, this distribution is not even and sexual cannibalism is often concentrated among particular taxa within these groups. Elgar (1992) cautions that many of the records are anecdotal and thus may either under- or overrepresent the frequency of sexual

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

139

cannibalism for that taxon. For example, records of sexual cannibalism in orb-weavers of the genus Argiope usually reported that it occurs before insemination, whereas more recent observations of these species clearly indicate that sexual cannibalism can occur both before and during copulation (Elgar et al., 2000; Fromhage et al., 2003; Knoflach and van Harten, 2001; Sasaki and Iwahashi, 1995). Indeed, females of A. bruennichi never attack males before copulation, but rather assume a characteristic mating position as soon as a male commences courtship (Fromhage et al., 2003). Detailed observations of the courtship and mating behavior of certain scorpions indicates that Elgar’s (1992) cautionary comment was well placed. Following Polis and Sissom (1990), the bothriurids Bothriurus bonariensis and Urophonius jheringii and the buthid Leiurus quinquestriatus were classified in Elgar (1992) as postinsemination sexual cannibals. However, Peretti et al. (1999) did not observe sexual cannibalism in 89, 10, and 84 pairings of these species, respectively. Furthermore, sexual cannibalism was not observed in pairings of 19 other species of scorpions, including two (Urophonius brachycentrus and Buthus occitanus) that Elgar (1992) and Polis and Sissom (1990) recorded as sexual cannibals (Peretti et al., 1999). It seems likely that sexual cannibalism is less prevalent among scorpions than has been generally assumed (see also Benton, 2001).

III. The Timing of Sexual Cannibalism While sexual cannibalism is a relatively simple and unambiguous behavior, its evolutionary significance depends crucially on whether it occurs before or after insemination. The distinction is important simply because the timing influences the costs of the behavior to the male and female. Males forfeit further reproduction if they are cannibalized before insemination, and although females obtain a nutritious meal they also risk reproductive failure. On the other hand, a male may benefit from sexual cannibalism if it takes place after insemination, through paternal investment and/or mating effort (in the sense of Simmons and Parker, 1989). Clearly, females sustain no costs by cannibalizing males after insemination, and may benefit nutritionally and also by removing males from within their vicinity. The timing of sexual cannibalism may vary both within and between species. For example, sexual cannibalism in the orb-web spider Argiope bruennichi occurs only after the male has inseminated the female. In other species, such as the garden orb-web spider Araneus diadematus, sexual cannibalism may occur either before or after insemination. While postinsemination cannibalism can occur with or without preinsemination sexual cannibalism, the latter rarely, if ever, occurs alone.

140

ELGAR AND SCHNEIDER

For spiders, the association between the timing of sexual cannibalism and insemination is slightly complicated by their reproductive anatomy. Male spiders possess paired secondary mating organs, pedipalps, that are both filled with sperm. In most species, only one will be inserted at a time and another courtship sequence must precede transfer of sperm from the second pedipalp. Correspondingly, females possess paired genital openings that lead to independent receptacles for sperm storage. Thus, the mating options of male spiders can include: empty one pedipalp with one female and keep the second for another female; empty both pedipalps with the same female; or do either and recharge the empty pedipalp. Only the first two options are available for males of species that cannot recharge their pedipalps. Sexual cannibalism can occur at any time during the entire sequence, namely, before the first insertion, between the first and second insertion, or after the second insertion. Clearly, the mating history of the male and female is also important in elucidating the costs of sexual cannibalism. There may be little difference in the costs of pre- and postinsemination sexual cannibalism for a female that has already mated and is being courted by another male. Similarly, an already mated male that falls victim to preinsemination sexual cannibalism by another female suffers a loss of reproduction, but may nonetheless sire some offspring. Models that attempt to explain the evolution and maintenance of sexual cannibalism typically distinguish between pre- and postinsemination sexual cannibalism (see below). However, it is interesting to note that models investigating preinsemination sexual cannibalism implicitly assume that the probability of capturing a male successfully is greater before than after copulation. Such an assumption is necessary for the existence of a trade-off between mating and other needs (e.g., Newman and Elgar, 1991). However, this assumption has not received any empirical attention.

IV. Natural Selection of Sexual Cannibalism A. Female Foraging Strategies Sexually cannibalistic behavior may provide the female with nutrients that increase her survival and/or fecundity. In this case, sexual cannibalism that does not prevent fertilization attracts the same costs and benefits as cannibalism more generally (see Elgar and Crespi, 1992). The difficulty lies in distinguishing between cannibalism as a foraging strategy and cannibalism as indiscriminate foraging, perhaps through misidentification of the

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

141

victim. The former implies that the consumption of a male demonstrably increases the survival or fecundity of the cannibal, while the latter assumes that there is no significant cost of cannibalism. Clearly, there is no fertilization cost of postinsemination sexual cannibalism, but virgin females that attack males before mating risk reproductive failure. This cost suggests that if preinsemination sexual cannibalism is common then it is unlikely to occur through indiscriminate foraging. Newman and Elgar (1991) used stochastic dynamic programming to investigate the conditions under which preinsemination sexual cannibalism could evolve as a foraging strategy, using parameters derived from Elgar and Nash (1988). More particularly, the model investigated whether these conditions could predict preinsemination sexual cannibalism by virgin females. The two most important factors to influence the evolution of this behavior were the expected number of males encountered during the mating season, and the size of the variance in the mass gained from other food items. The model predicted that females would be less likely to attempt to cannibalize a courting male before mating if they expect to encounter only a few males, expect a high mean capture rate of other prey items, and expect the capture rate to be fairly constant. While the model identified conditions under which a virgin female might attempt to attack a courting male, it also showed that postinsemination sexual cannibalism should be more widespread than preinsemination sexual cannibalism. A crucial assumption of the Newman and Elgar (1991) model is that the consumption of males increases female fecundity. Arnqvist and Henriksson (1997) explicitly tested the predictions of the Newman and Elgar (1991) model, using the fishing spider Dolomedes fimbriatus (Pisauridae: Araneae). These large spiders feed on arthropods trapped at the water surface, and sometimes also prey on small fish and tadpoles. Courting males approach females cautiously, waving their legs and creating vibratory signals, only attempting to mount the female if she remains passive (Arnqvist and Henriksson, 1997). In the field, males are common victims of cannibalism by females and in the laboratory, males have a 5% risk of preinsemination sexual cannibalism (Arnqvist, 1992). Arnqvist and Henriksson (1997) found that the prevalence of sexual cannibalism was not influenced by female mating status, the size of the courting male, female foraging history, female size, or season. These data provide little support for the Newman and Elgar (1991) model, although it is important to recognize that the act of sexual cannibalism includes both the female’s attempts to cannibalize the male and the male’s attempts to avoid cannibalism. The Newman and Elgar (1991) model was concerned with the former only, and a more appropriate test might consider both successful and unsuccessful attempts to cannibalize the male.

142

ELGAR AND SCHNEIDER

This lack of empirical support for an economic foraging explanation of sexual cannibalism prompted Arnqvist and Henriksson (1997) to argue that selection may not favor preinsemination sexual cannibalism per se, but rather that it occurs as an indirect result of a foraging strategy that is adaptive in earlier life history stages. There is widespread evidence in spiders that female fecundity is influenced more by food consumption in the juvenile than in the adult stage (Arnqvist and Henriksson, 1997). This suggests that there will be strong selection for any trait that increases food intake rate, such as aggressively attacking any potential prey items that present themselves. Arnqvist and Henriksson (1997) argue that while the costs of this indiscriminate behavior to adult females may be high, in terms of failing to mate, genetic ‘‘constraints’’ may allow only its reduction rather than elimination. Thus, the sexually cannibalistic behavior of adult females is simply a nonadaptive ‘‘spillover’’ of the highly adaptive aggressive behavior of juveniles. Significantly, this explanation does not assume that the consumption of a single male increases female fecundity. Nevertheless, the model predicts that larger females are more likely to be sexually cannibalistic than smaller females, since their large size is directly attributed to a highly aggressive foraging behavior as juveniles. Elgar and Nash (1988) found that the size of the female did not influence her ability to successfully attack the male, but these data do not tell us whether female motivation to sexual cannibalism is influenced by her size. Johnson (2001) tested both models using the American fishing spider Dolomedes triton. By creating different juvenile and adult feeding regimens, he found strong fecundity selection maximizing foraging vigor of female spiders in all life history stages while consumption of a male did not increase fecundity. Hence, a critical assumption of the Newman and Elgar (1991) model was not met. A perplexing result was that failure to mate was not related to female aggressive behavior. Thus, the cost of aggression found for adult female D. fimbriatus could not be found in its congener. Of 15 female D. triton (20% of the observed females) that did not accept a mate, only 3 attacked a male, compared with 17 of 38 females that behaved aggressively toward their mates but still copulated. The causes for the failure to copulate have not been identified. Female D. fimbriatus typically attack males before mating (75% of females attack before mating, of which 5–10% were successful), whereas D. triton females mostly attack their mating partners after the male has made an initial palpal insertion, which is likely to be optimal for the female. Unfortunately, data on the success of such attacks are not available. While sexual cannibalism does not influence fecundity in D. triton, it does improve the hatching success of spiderlings (Johnson, 2001). This suggests that future studies of the benefits of sexual cannibalism for females should not be restricted to measures of fecundity

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

143

alone, but should also consider other reproductive parameters (Johnson, 2001). B. Postinsemination Sexual Cannibalism and Paternal Investment Sexual cannibalism should be more prevalent after than before insemination if it evolved directly, or indirectly, as a foraging strategy. This is because it is no longer constrained by the risk of reproductive failure. The comparative data are not sufficiently robust to address this question. Nevertheless, it is interesting to note that sexual cannibalism more typically occurs after insemination in species in which it is apparently common (e.g., Andrade, 1996; Elgar et al., 2000; Fromhage et al., 2003; Sasaki and Iwahashi, 1995; Schneider and Elgar, 2001, 2002; but see Elgar and Nash, 1988). However, this argument restricts the evolution of sexual cannibalism to a female perspective only. Even if females benefit from a foraging strategy that regularly includes the consumption of mating partners, the strength of counterselection on the side of the victim will critically determine whether such a female strategy can evolve. In theory, the victim of sexual cannibalism can benefit from his consumed soma if his sperm is used to fertilize the cannibal’s eggs (Buskirk et al., 1984; Parker, 1974). The Buskirk et al. (1984) model predicts that postinsemination sexual cannibalism with male complicity will evolve if the number of additional offspring produced as a result of sexual cannibalism exceeds the number of offspring the male might expect to sire if he avoided cannibalism and searched for additional females. While Buskirk et al. (1984) were primarily concerned with the benefits of sexual cannibalism to the victim, it is clear that if the male benefits then so must the female. The crucial assumption of this model is that the consumption of males increases female fecundity, although from the male perspective, his soma alone must increase her fecundity. However, there is little evidence for such an assumption (Table I): only two studies report an increase in female fecundity (or a surrogate variable), while seven studies failed to detect an effect. Andrade (1998) suggests that the benefit of sexual cannibalism to female red-back spiders Latrodectus hasselti depends on their foraging history (see also Liske and Davis, 1987). Female red-back spiders construct a horizontal sheet web, remaining in a retreat off the web during the day and venturing onto the web at night. Females hang from the web with their ventral surface up, and males initiate courtship from the edge of the web. During copulation, males may turn 180 , bringing their abdomens into the vicinity of the fangs of their female mate. Females may then attempt to consume the male (Andrade, 1996; Forster, 1992). Andrade (1998) compared the

TABLE I Effect of Sexual Cannibalism on Female Fecundity Species Araneidae Araneus diadematus Argiope bruennichi Argiope keyserlingii Phonognatha graeffei Theridiidae Latrodectus hasselti Pisauridae Dolomedes fimbriatus Dolomedes triton Dolomedes triton Pisaura mirabilis Mantodea Hierodula membranacrea Iris oratoria

Treatment Experimental addition of a male Experimental addition of a male Natural variation in sexual cannibalism Experimental addition of a male

Test variable

Effect

Ref.

Female mass

Increase

Elgar and Nash (1988)

Female fecundity

None

Fromhage et al. (2003)

Female fecundity

None

Elgar et al. (2000)

Female fecundity

None

Fahey and Elgar (1997)

Natural variation in sexual cannibalism

Female fecundity

None

Andrade (1996)

Natural variation in sexual cannibalism Natural variation in sexual cannibalism Natural variation in sexual cannibalism Experimental provision of nuptial gift

Female fecundity

None

Female fecundity

None

Hatching success

Increase

Arnqvist and Henriksson (1997) Spence et al. (1996); Johnson (2001) Johnson (2001)

Female fecundity

None

Sta¨lhandske (2001)

Natural variation in sexual cannibalism Natural variation in sexual cannibalism

Female fecundity

Increase

Birkhead et al. (1988)

Female fecundity

None

Maxwell (2000)

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

145

sexually cannibalistic behavior of red-backs in the field, whose diet had or had not been experimentally augmented. The experiment failed to reveal a significant effect, although the body condition of noncannibalistic females was significantly better than that of cannibalistic females. However, these data do not provide a convincing case that female redbacks are sexually cannibalistic in order to compensate for poor body condition and thus increased fecundity. First, consuming a single male does not increase female fecundity (Andrade, 1996), although there may be some benefit to multiply mating females. Second, it is not clear why sexual cannibalism should be confined to females in poor condition, especially since there is no obvious cost of the behavior to females. Third, sexual cannibalism in redbacks appears to be a fixed part of the male mating behavior without females playing any active role in the initiation of the cannibalistic act. Females of the golden orb-web spider Nephila plumipes attempt to cannibalize courting males both before and during copulation. The webs of these comparatively large spiders form large aggregations, and while a single female is found on each web, there may be up to 10 males located within and on the periphery of the orb-web (Elgar, 1989; Elgar and Fahey, 1996; Schneider and Elgar, 2001). Not all females attempt to cannibalize their mating partners, and Schneider and Elgar (2002) compared various life history characteristics of cannibalistic and noncannibalistic females. An important feature of this study is that the male victim was removed immediately after the female killed him, thereby preventing her from consuming him. The mass gain from maturation to oviposition was greater for cannibalistic than noncannibalistic females and the former also produced larger clutches. However, these cannibalistic females typically matured at a smaller size and mass than their noncannibalistic counterparts. There was no discernible difference in the behavior of males that did or did not survive their encounter with a female. These data suggest that in N. plumipes, selection has not favored sexual cannibalism per se, but rather it is a consequence of female foraging vigor; females that mature in poorer condition are more aggressive foragers and as a result include males in their diet. The data obtained by Schneider and Elgar (2002) have important implications for studies that measure the fecundity consequences of natural variation in female cannibalism. Any difference in fecundity between cannibalistic and noncannibalistic females could be due to either the consumption of the male or the foraging vigor of the female (Schneider and Elgar, 2002). This confounding factor was controlled explicitly in studies of two species of orb-web spiders, Araneus diadematus and Argiope bruennichi. The garden spider Araneus diadematus is often common where it occurs in both Europe and North America. Females are heavier than

146

ELGAR AND SCHNEIDER

males, although the body length of a mature male is only slightly less than that of a mature female. Sexual cannibalism occurs before or after copulation, and females that consumed a male were significantly heavier than females prevented from consuming a male (Elgar and Nash, 1988). Females of the highly sexually dimorphic Argiope bruennichi consume males only after copulation, and there was no difference in either the fecundity or the fertility of females that did or did not consume one or two males (Fromhage et al., 2003). The evidence that sexual cannibalism evolved as an adaptive component of female foraging strategies is equivocal. There is little evidence that the consumption of a single male increases female fecundity in a broad range of species (Table I), so paternal investment is unlikely to be a general explanation. Nevertheless, sexual cannibalism may well be a nonadaptive consequence of female foraging strategies. However, it would be premature to conclude that Gould (1984; see also Jamieson, 1986) is right that sexual cannibalism is of little evolutionary significance per se, since there is compelling evidence that sexual cannibalism may affect both sexes through the process of sexual selection.

V. Sexual Selection and Sexual Cannibalism Sexual cannibalism could have evolved through the process of sexual (Andersson, 1994; Darwin, 1871) rather than natural selection. Thus, instead of providing males or females with a benefit in terms of increasing survival or reproductive output, sexual cannibalism may provide a mechanism by which individuals gain a competitive advantage over their samesex rivals. There are two mechanisms of sexual selection, which may operate both before and after insemination. One refers to the direct competition between individuals of one sex for fertilization success, and typically involves competition between males. It is manifested as male–male competition in the context of preinsemination sexual selection (Andersson, 1994), and sperm competition in the context of postinsemination sexual selection (e.g., Birkhead and Møller, 1998; Simmons, 2001). Sexual cannibalism is unlikely to be relevant in the former, but there is clear evidence that it is of significance in terms of sperm competition. The other mechanism of sexual selection is mate choice, in which individuals of one sex choose mating partners based on various traits in the other sex. Females are usually the choosy sex, and may prefer males according to features of their secondary sexual characteristics, such as color, odor, song, or behavior (Andersson, 1994). Female choice may persist after insemination has taken place, although the preferred traits may not be immediately obvious

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

147

and for this reason the term cryptic female choice was coined (Eberhard, 1996; Thornhill, 1983). There is evidence that sexual cannibalism facilitates both female choice and cryptic female choice. A. Female Choice via Preinsemination Sexual Cannibalism Preinsemination sexual cannibalism may provide females with an opportunity to assess aspects of male quality (Elgar and Nash, 1988), by challenging them with a radical form of mate choice. Males that succeed in avoiding sexual cannibalism may be of better quality, and males of lesser quality are nonetheless nutritious fodder. Males of the garden spider Araneus diadematus court the female on a specially constructed mating thread that is suspended from the edge of the orb-web to either the vegetation or surrounding web support threads. The male drums and plucks the mating thread with his legs and the vibrations attract the female, who must hang from this thread in order to mate. As the female approaches the male, he lightly touches her body with his legs before attempting to copulate. Females may capture and cannibalize males before copulation takes place, but larger males are less likely to be caught than smaller males (Elgar and Nash, 1988). Thus, females may benefit from preinsemination sexual cannibalism because it allows them to distinguish between males on the basis of both their size (which might reflect their foraging ability as juveniles) and their competence at avoiding cannibalism (Elgar and Nash, 1988). In this context, the cost of sexual cannibalism to females, in terms of risking remaining unmated, is no greater than that experienced in other species with discriminating females. The only difference is that sexually cannibalistic females do not have the opportunity to mate with males that were previously rejected. Further evidence that sexual cannibalism allows females to exercise mate choice based on a secondary sexual trait is provided by Hebet’s (2003) study of the wolf spider Schizocosa uetzi (Lycosidae). Courtship in this species is highly visual, with males waving their legs repeatedly in front of the female. The legs are ornamented with varying degrees of black pigmentation on a portion of the tibia of their forelegs. Hebets (2003) showed that the mating preference of adult female S. uetzi was influenced by their history of interactions with males as juveniles. In particular, females were more likely to mate with males of a familiar phenotype, and more likely to attack and cannibalize males with an unfamiliar phenotype. It is possible that the cannibalistic spiders did not recognize the unfamiliar males as conspecifics (even though some were able to mate) and thus it might not be appropriate to consider sexual cannibalism to be a mechanism of mate choice. If it is such a mechanism, then it is curious that

148

ELGAR AND SCHNEIDER

sexual cannibalism is not more common among lycosid spiders (Elgar, 1992). Females that exercise mate choice by preferentially mating with particular males may be confronted by the unwanted courtship of rejected males. Sexual cannibalism may provide females with an efficient means of reducing the costs of this sexual harassment. The presence of males on the webs of the orb-web spider Argiope keyserlingi reduced the foraging efficiency of the female and also attracted the attention of potential predators (Herberstein et al., 2002). Sexual cannibalism in this species may allow the female to remove these unwanted males from her web. B. Postinsemination Sexual Cannibalism: Sperm Competition and Cryptic Female Choice Postinsemination sexual cannibalism may represent a form of male mating effort (in the sense of Simmons and Parker, 1989) by providing a means of preventing sperm competition or increasing paternity when engaging in sperm competition (Elgar, 1998). Male ceratopogonid biting midges form dense mating swarms in which females may encounter and potentially mate with numerous males. However, once the pair is in copula, the female pierces the cerebral cuticle of the male with her biting mouth parts and proceeds to imbibe the male’s body fluids. The male remains attached to the female and eventually his desiccated body breaks from the terminal segment that contains his reproductive organs (Downes, 1978). Further mating by the female may be prevented by the presence of the remains of the male, although it is not clear how long his remains remain attached. Perhaps sexual cannibalism in biting midges represents mating effort, in which the remains of the cannibalized male acts as a mating plug that prevents rivals from mating with the female (Elgar, 1992; Thornhill and Alcock, 1983). Andrade (1996) showed that sexual cannibalism in red-back spiders is a mechanism of increasing paternity share in a sperm competition environment. All males of L. hasselti perform a somersault during copulation that places their abdomen directly onto the female mouthparts. Females excrete digestive fluids onto the male that slowly digest his body (Andrade, 1996; Forster, 1992). However, prior to mating, the male probably removes vital parts from the posterior to the anterior abdomen which may help him to survive a first copulation and attempt the transfer of the sperm stored in the second pedipalp after a short break (Forster, 1992). After two insertions, the damage to the male’s copulatory organs, the pedipalps, renders him functionally sterile (Andrade and Banta, 2002). Males that are cannibalized more than double the duration of copulation (from 11 to 25 min),

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

149

which effectively doubles the paternity share of that particular male under sperm competition (Andrade, 1996). Hence self-sacrifice provides a competitive advantage to males by extending the duration of sperm transfer. In addition, sexually cannibalistic females are less receptive to other mating partners, thereby reducing the risk of sperm competition. The precise role that cannibalism plays in prolonging copulation and reducing female receptivity in spiders is not clear. Studies of cannibalistic and noncannibalistic spider species have demonstrated a positive correlation between the duration of copulation and paternity, and females are less receptive following mating in many species of spiders (Elgar, 1998). A broad survey of the copulation duration of spiders suggests that sexually cannibalistic species tend to copulate for a shorter time than species that are not sexually cannibalistic (Elgar, 1995), although this study did not control for phylogenetic effects. For example, the duration of copulation in the sexually cannibalistic orb-weaver N. plumipes ranges between 2 and 300 s (average, 55 s; Schneider and Elgar, 2001), which is considerably shorter than that of the less frequently cannibalistic N. edulis (range, 8 to 1710 s; average, 233 s; Schneider et al., 2000). Nevertheless, males of N. plumipes that escape a female attack mate for a shorter time than males that are killed (Schneider and Elgar, 2001). Sexual cannibalism does not provide a straightforward benefit to males of the golden orb-web spider N. plumipes (Elgar et al., 2003b; Schneider and Elgar, 2001, 2002; Schneider et al., 2001). In double-mating trials, sexual cannibalism did not improve the fertilization success of males mating with virgin females. A male mating with a virgin female can expect to fertilize 50% of her eggs on average, independently of whether he dies or survives the copulation (Schneider and Elgar, 2001). However, males that mate with previously mated females can increase their share of paternity by sacrificing their lives, since this prolongs copulation (Schneider and Elgar, 2001). Interestingly, these patterns do not change qualitatively when females copulate with three males; the first male can still expect to fertilize 50% of the brood and later rivals share the rest (Elgar et al., 2003b). Thus, there is no evidence that sexual cannibalism in N. plumipes has evolved through a mating effort function. Indeed, males appear to attempt to avoid cannibalism at least after their first insertion and there is some evidence that sexual cannibalism reflects a conflict of interest over the duration of copulation in this species (Schneider et al., 2001; see below). Males of the New World orb-web spider Argiope argentata spontaneously die during their second insertion (Foellmer and Fairbairn, 2003). However, unlike other species of this genus, the dead male remains attached to the female. Foellmer and Fairbairn (2003) suggest that the male’s corpse becomes a short-term mating plug, thereby reducing the likelihood of

150

ELGAR AND SCHNEIDER

further mating in a manner analogous to the ceratopogonid biting midges (Elgar, 1992; Thornhill and Alcock, 1983). Schneider et al. (2001) tested this possibility in the sexually cannibalistic golden orb-web spider N. plumipes. Like many spiders, females of N. plumipes control the duration of copulation by aggressively attacking the male, which usually results in the male terminating copulation and leaving. A female that kills a copulating male typically attempts to move his remains away from her genital opening, but females are not always successful because males can remain firmly attached by the aid of a morphological hooklike process on their pedipalp (Schneider et al., 2001; see Fig. 1). However, there was no evidence that the presence of the male’s pedipalp in the female’s reproductive opening prevented rival males from copulating successfully with her. Indeed, in the event that it did prevent a male from inserting his palp, a cannibalized male would have had to leave his palpal remains in both genital openings of the female. Males may remain attached to the female simply to prolong the duration of copulation beyond the interests of the female. Unlike red-back spiders, there is little evidence of male complicity in sexual cannibalism in the St. Andrew’s cross orb-web spider Argiope keyserlingi. The paternity share of cannibalized males was not significantly different from that of males that survived copulation (Elgar et al., 2000). Instead, sexual cannibalism in this species may allow females to control paternity. Females of A. keyserlingi terminate copulation by attempting to wrap and subsequently consume the male while he is copulating. Copulation ceases as soon as the female attacks the male. Double-mating experiments revealed that females delayed sexual cannibalism when copulating with relatively smaller males, thereby prolonging the duration of copulation. The result was that these males copulated for relatively longer and thus fertilized relatively more eggs (Elgar et al., 2000). It is not clear why females should prefer the offspring of smaller males, but perhaps they require less time to reach sexual maturity and thus experience less risk of predation during this time. In summary, multiple mating is typical of the females of those sexually cannibalistic spiders that have been studied in some detail (see also Elgar, 1998), and thus the risk of sperm competition (in the sense of Parker, 1998) is probably high in these species. The accumulating evidence suggests that sexual cannibalism prolongs the duration of copulation, either absolutely or relatively, thereby increasing the paternity share of the victims. Interestingly, this does not translate into an interspecific pattern of relatively longer copulations in sexually cannibalistic species (Elgar, 1992). This may be due to a sexual conflict of interest over paternity (see below).

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

151

VI. Postinsemination Sexual Cannibalism, Self-Sacrifice, and Monogyny Why would males sacrifice any future reproductive activity by allowing or even provoking cannibalism by their mating partner? More generally, what factors might be responsible for the evolution of monogyny, or single mating by males? In some respects, the question represents a classic life history trade-off between current and future reproductive potential, and the closest behavioral analogy is with mate guarding. Males of many species guard their mating partners from rival males in order to reduce the risk of sperm competition (Birkhead and Møller, 1998), even though this behavior reduces further mating opportunities. Thus, the duration of mate guarding is determined by balancing the benefits of remaining with the female against the costs of not securing matings with other females (Parker, 1974; Simmons, 2001). The probability of a male securing further matings depends on the number of receptive females in the population and the mortality associated with locating those females. The expected reproductive success of a male through the search for additional females must be low for postinsemination sexual cannibalism with male complicity to evolve or at least it would explain why counterselection by males would be weak (Elgar, 1992; Parker, 1979). It is widely held that the main impediment to future mating in these species is the high mortality rate associated with mate searching (e.g., Buskirk et al., 1984; Maxwell, 1998; Vollrath, 1998). Andrade (2003) proposed a formal model with exactly this logic for redback spiders, L. hasselti. Field data revealed that only 20% of males found a female, suggesting that there are very high search costs for these spiders. Andrade (2003) concluded that the paternity benefits of sexual cannibalism (Andrade, 1996), together with the high mortality rate during mate search, are sufficient to select for self-sacrifice in male red-back spiders. Fromhage et al. (manuscript) identified a fundamental problem with the above argument, and questioned whether monogyny would evolve under the conditions suggested by Andrade (2003). The problem is that the high mortality suffered by males during mate search will reduce the number of males and thus the intensity of sexual selection (Vollrath and Parker, 1992). As a consequence, there will be little selection for the paternity protection arising from sexual cannibalism because the absence of competing rivals ensures little risk of sperm competition. As a consequence, males should always attempt to seek further matings, irrespective of the probability of finding another female. The paradox was solved by incorporating the novel idea of the effective sex ratio, which is the ratio of the number of males and females that

152

ELGAR AND SCHNEIDER

copulate at least once in their lifetime (Fromhage et al., manuscript). The effective sex ratio might be male biased through high juvenile female mortality, a male-biased primary sex ratio, or through a high proportion of females that are either inaccessible or simply fail to attract the attention of any male. Field studies of the reproductive success of nursery web spiders revealed that many females failed to mate (Austad and Thornhill, 1986), suggesting that the latter explanation of a male-biased effective sex ratio may not be infrequent among web-building spiders. If ecological or demographic conditions ensured that there was a surplus of males competing for fertilization success, selection will favor any male strategy that increases his share of paternity with individual females. Thus, male selfsacrifice, or monogyny, will be favored if it increases the paternity of the sacrificial male above the average of a polygynous male. Interestingly, reports of cannibalistic spiders, particularly orb-web spiders, frequently describe several males attending a female (e.g., Elgar, 1998). Andrade (1996, 2003) also reported that several male red-back spiders congregate on the webs of females, despite the high mortality associated with mate search, and that this creates a risk of sperm competition. Monogynous males are expected to exhaust their entire reproductive potential with one female. The secondary mating organs of male spiders, the pedipalps, are paired and are typically inserted separately into paired insemination ducts leading to two independent spermathecae. Thus, monogynous spiders should transfer the entire sperm load in each of the paired pedipalps, which can be achieved only if the male survives his first insertion and can successfully inseminate the same female with a second insertion. A successful insertion may require a long copulation, which may be facilitated by allowing sexual cannibalism. If sexual cannibalism is common, we expect it to occur in concert with a monogynous mating strategy that allows males to maximize their relative paternity success. This can be manifested in different ways. For example, males of Nephila fenestrata often survive two insertions, but remain on the web apparently guarding the female from rivals (L. Fromhage and J. M. Schneider, unpublished data). Males of A. keyserlingi never survive two copulations (Gaskett et al., 2004), but may guard their females from rival males after the first copulation. In summary, explanations of self-sacrifice and postinsemination sexual cannibalism should be couched in terms of components of a monogynous mating strategy, rather than treating the behavior as an independent trait.

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

153

VII. Sexual Conflict over Mating Rate and the Duration of Copulation It is now widely accepted in contemporary evolutionary biology that the ‘‘interests’’ of males and females differ over a raft of issues related to reproduction, including the choice of mating partner and the frequency of mating (e.g., Lessells, 1999; Parker, 1979; Rowe et al., 1994; Schneider and Lubin, 1998; Stockley, 1997). As a consequence, sexual selection may favor various adaptations that allow one sex to overcome the interests of the other. Selection of this kind is especially evident in polyandrous species: while both males and females benefit from mating with several partners, male fertilization success may be compromised if his mate copulates with rival males. However, selection will favor each sex to overcome any reproductive disadvantage arising from an adaptation favoring the other sex. The result is an antagonistic coevolution of adaptation and counteradaptation (e.g., Arnqvist and Rowe, 2002; Rice, 1996). Sexual cannibalism has been described as a particularly dramatic manifestation of this conflict (e.g., Elgar, 1992) and this is certainly true of preinsemination sexual cannibalism. Given that females benefit from consuming a male, female rapaciousness will select for male strategies that reduce the risk of being killed before sperm can be transferred. Mating while the female is molting or feeding may be examples of male mating strategies that evolved as counteradaptations to female aggressiveness. However, there are few experimental data to support these ideas and they may be very difficult to obtain. Sexual cannibalism might not be observed, even though the female may have attempted to capture her mate. Clearly, the issue can be resolved only by obtaining detailed observations of female offensive and male defensive (or complicit) behaviors. Small body size has been discussed as another potential counteradaptation (see below). How sexual conflict may have influenced postinsemination sexual cannibalism is less straightforward. The act of killing and consuming the male sexual partner may or may not be a part of the male mating strategy, depending on whether copulation can continue after the female attacks. As already described for the red-back spider, males actively provoke sexual cannibalism, thereby increasing the duration of copulation and reducing female receptivity to future rivals (Andrade, 1996). Andrade (1998) argues that sexual cannibalism in red-back spiders occurs when male and female interests coincide rather than conflict, but the possibility that males manipulate female reproductive behavior through their sacrifice has not been explored. A reduction in female receptivity following cannibalism clearly

154

ELGAR AND SCHNEIDER

benefits the male by decreasing the likelihood of sperm competition. This may not necessarily be in the interest of the female, and she may have the option of not attacking the male. Males do not necessarily invite cannibalism in other postinsemination sexually cannibalistic spiders. However, sexual cannibalism at some stage of the mating sequence may be a fixed component of the male’s mating strategy occurring, for example, because he does not attempt to escape the female’s attack. Either active self-sacrifice or simply an absence of an escape response that precedes the act of cannibalism is expected if this maximizes male fertilization success, and especially if it results in exclusive paternity. However, selection may favor multiple mating in females (e.g., Arnqvist and Nilsson, 2000; Jennions and Petrie, 2000). Thus, females may benefit by preventing a second insertion if that ensures mixed paternity while at the same time ensuring that there are sufficient sperm to fertilize all her eggs. Under these conditions, female and male interests will differ over the number of insertions per male. Indeed, in Dolomedes triton most females successfully attack their mates after the first insertion (Johnson, 2001) while L. hasselti males rarely die after their first insertion (Andrade, 1996). In the very cannibalistic species of the genera Argiope and Nephila, females regularly attack males during the first insertion and males will generally struggle to escape. The survival probability of singly mated males differs between species: in N. plumipes 56%, in A. keyserlingi 61%, and in A. bruennichi 80% of the males do not survive their first copulation with a virgin female (Elgar et al., 2000; Fromhage et al., 2003; Schneider and Elgar, 2001). Accumulating evidence suggests that after two successful insertions, males will never escape a female attack. While no study has measured the significance of two insertions for male fitness, the total duration of copulation will be increased by additional insertions. In many invertebrates, including spiders and insects, absolute and relative fertilization success is positively correlated with the duration of copulation (Elgar, 1998; Simmons, 2001). Consequently, selection may favor males that can copulate for longer, since that will increase their paternity share, but favor females that can copulate more briefly since that facilitates equitable sperm mixing (e.g., Elgar, 1998). By attacking and killing a male during copulation, females not only prevent further insertions by the same male but may also control the duration of copulation. The duration of copulation among spiders varies from a few seconds to many minutes (Elgar, 1995; Stratton, 1996) but is relatively short in the orbweb spiders, particularly in the genus Argiope. Orb-weavers show a high frequency of postinsemination sexual cannibalism and in a comparative analysis, Elgar (1995) found that, within the Araneidae, sexually cannibalistic genera have shorter copulation durations than other taxa. In Argiope

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

155

bruennichi, copulations of less than 10 s are sufficient to fertilize all clutches laid by a female (Schneider et al., 2004). Females attack the male immediately after his genitalia become attached to the female. Male genitalia have to unfold in a complicated way and can apparently allow the male to maintain a hold on the female; males that do not jump off after about 8 s are killed by the female, but can nevertheless copulate for about 40 s (Fromhage et al., 2003). Copulation in the golden orb-web spider N. plumipes is also short and the male pedipalp can remain stuck in her genital opening, despite the female’s attempt to dislodge it (Fig. 1). Often, part of the pedipalp will break off, but the presence of these remains will not prevent rival males from mating with this female nor influence the duration or success of a rival’s copulation (Schneider et al., 2001). The pedipalps of N. plumipes are characterized by a small process that may facilitate the firm attachment of male genitalia. Three other species of Nephila (N. tetraganthoides, N. fenestrata, and N. constricta) possess similar structures, while the majority of species in this genus do not. We are investigating whether male pedipalp morphology has evolved in response to female control of copulation. Such an adaptation will further limit male mating frequency because males cannot use damaged genitalia, suggesting that it is closely associated with a monogynous mating strategy. Sexual conflict over the frequency of mating and the duration of copulation has been reported for various insect species where male adaptations to sperm competition induce costs on females (e.g., Crudginton and Siva-Jothy, 2000; Sakaluk et al., 1995) and sexual selection including antagonistic coevolution is likely responsible for the rapid evolution of genitalia in insects (Arnqvist, 1998, Arnqvist and Rowe, 2002). Analogously, male genitalia of some orb-web spiders may have evolved to counteract female control of copulation by sexual cannibalism. Then, within species, tolerating cannibalism in concert with genitalia that prevent females from dislodging the male, will allow males to prolong copulation. However, at a higher taxonomic level, the antagonistic coevolution will result in shorter copulations in taxa with strong conflicts of interest.

VIII. Sexual Cannibalism and Male Mate Choice Theory predicts that males should exercise some choice of mating partner because they sustain costs in terms of courtship and copulation (Bonduriansky, 2001). Accordingly, male mate choice might be more overt in sexually cannibalistic species, since they have a limited number of mating opportunities. Virgin females are commonly preferred as mates as they offer a reduced probability of sperm competition and thus the highest

156

ELGAR AND SCHNEIDER

immediate returns for male investment even when first male sperm precedence is low (Bonduriansky, 2001; Simmons, 2001). Thus, males of sexually cannibalistic species are predicted to prefer virgin over mated females since, on average, fertilization success of this strategy is likely to be higher. Few studies have investigated male mating preferences in sexually cannibalistic species, and so there are no clear patterns at this stage. Males of A. keyserlingi most likely mate with one female only, and males clearly prefer to court virgin over mated females (Gaskett et al., 2004; Herberstein et al., 2002). Sexual cannibalism is infrequent in Micrathena gracilis and there is no evidence of differences in male behavior toward virgin or mated females (Bukowski and Christenson, 2000). There is also no evidence of male mate choice in the sexually cannibalistic N. plumipes (Elgar et al., 2003), but males of N. edulis apparently prefer virgin over mated females, even though sexual cannibalism is uncommon. Most likely, male choice is more strongly influenced by patterns of paternity share in multiply mating females (Bonduriansky, 2001), than by sexual cannibalism alone.

IX. Sexual Cannibalism and Sexual Size Dimorphism Typically, the degree and direction of sexual size dimorphism in sexually reproducing organisms are influenced by the relative strengths of fecundity selection on females and sexual selection on males. For most insects and spiders, females are generally larger than males since fecundity frequently depends on body size. The smaller or lighter males may nonetheless have enlarged secondary sexual characteristics that have been favored through sexual selection. In some species, such as certain genera of orb-weaving spiders, males are so much smaller than females that they have been termed ‘‘dwarf males’’ (e.g., Vollrath, 1998). For many years, it was thought that the extreme sexual size dimorphism of the orb-web spiders allows males to avoid sexual cannibalism, either because they are not detected by the female or offer insufficient nutritional rewards to warrant attack (e.g., Cambridge, 1871; Darwin, 1871; Robinson and Robinson, 1980). The evidence for sexual cannibalism selecting for small males is equivocal. Elgar and Fahey (1996) showed that, consistent with this idea, females of the sexually cannibalistic and highly sexually size dimorphic orb-web spider Nephila plumipes were more likely to respond aggressively to larger than smaller courting males. However, there is a similarly high degree of sexual size dimorphism in N. edulis, despite the comparatively low frequency of preor postinsemination sexual cannibalism (Elgar et al., 2003; Schneider et al., 2000; Uhl and Vollrath, 1998). Elgar (1992) found no covariation between

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

157

sexual cannibalism and sexual size dimorphism across spiders generally, or within either orb-weaving (Araneidae) or jumping (Salticidae) spiders. Elgar et al. (2000) showed that sexual selection through sexual cannibalism favored smaller males because they were able to copulate for longer, but again the generality of this pattern is not clear. Nevertheless, Elgar (1991) suggests that an association between the degree of sexual size dimorphism and courtship behavior among orb-web spiders may be due to sexual cannibalism. Males of these spiders that mate on a mating thread may be less at risk of sexual cannibalism if they are larger (e.g., Elgar and Nash, 1988), while males that traverse the orb-web to mate with the female may be at less risk if they are smaller (e.g., Elgar and Fahey, 1996). Accordingly, sexual dimorphism should be more extreme in the latter than former species, a pattern that emerged in an interspecific comparative analysis (Elgar, 1991). While intriguing, the study was unable to control convincingly for phylogeny, so it is possible that the patterns may be due to other factors. Given the high cost of sexual cannibalism for males, one might expect strong selection to favor male traits that reduce the likelihood that a male falls victim to a female cannibal. Body size is an important predictor of the outcome of conflicts between conspecific spiders, both within and between sexes (e.g., Elgar, 1998; Schneider and Lubin, 1997). Accordingly, sexual size dimorphism may be less pronounced in species in which preinsemination sexual cannibalism is common. Elgar (1992) provides little evidence of such a pattern across spiders, although that analysis did not control for phylogeny. Elgar et al. (1990) found a significant negative correlation between relative male leg length and sexual size dimorphism within certain taxa of orb-weaving spiders. The pattern was attributed to preinsemination sexual cannibalism; relatively smaller males are more vulnerable to attack, and relatively longer legs allow the courting male to caress the female at a safe distance (Elgar et al., 1990).

X. Outlook and Summary Sexual cannibalism can occur before or during mating, which will determine potential explanations for its occurrence. Preinsemination sexual cannibalism does not appear to be taxonomically widespread, and reports of this behavior are mostly anecdotal (Elgar, 1992). Nevertheless, three explanations have been suggested. First, preinsemination sexual cannibalism is a nonadaptive consequence of selection on aggressive foraging in juvenile females (Arnqvist and Henriksson, 1997). Second, it is an adaptive component of female foraging, where females trade off foraging and

158

ELGAR AND SCHNEIDER

mating requirements (Newman and Elgar, 1991). Third, it is a radical form of female mate choice (Elgar, 1992; Elgar and Nash, 1988). Being killed before inseminating a potential mate is obviously costly for the male victim, but it is not clear how this dramatic mechanism of removing males from the population compares with more conventional means by which males are excluded from the mating pool. In the laboratory, one in four males of the orb-weaving spider Araneus diadematus was killed during courtship, but this high rate of preinsemination cannibalism has not been recorded in other species. Hence the notion that preinsemination cannibalism is a rare side product of other factors (Gould, 1984; Jamieson, 1986) cannot be convincingly refuted with the available data. Arnqvist and Henriksson (1997) propose that killing males is costly for virgin female fishing spiders Dolomedes fimbriatus because they risk remaining unmated. However, there was no evidence that female aggression affected the probability of mating in the congener D. triton (Johnson, 2001). While preinsemination sexual cannibalism has evolved against the interests of males, postinsemination sexual cannibalism appears to be predominantly a male mating strategy. The benefit to males appears to be in the context of mating effort rather than paternal investment, since there is little evidence that the consumption of male soma increases female reproductive output (Table I). Postinsemination cannibalism is, if it occurs, usually a regular, sometimes stereotyped component of the mating behavior, where a significant proportion of the males do not survive a copulation. In some species, males generally die during mating (e.g., Foellmer and Fairbairn, 2003; Knoflach and van Harten, 2001; Sasaki and Iwahashi, 1995), while in others species males are functionally sterile after two insertions (Andrade and Banta, 2003; Gaskett et al., 2004). It has been suggested that post insemination sexual cannibalism forms part of a monogynous mating strategy (Fromhage et al., ms). A monogynous mating strategy can evolve under a male-biased effective sex ratio (Fromhage et al., manuscript), and monogynous males will be under strong selection to maximize their paternity success with a single female. This may be achieved through self-sacrifice, genital damage, or failure to escape from females. It is not obvious in all species how cannibalism increases paternity share, although in many cases cannibalism extends the duration of copulation. It is possible that the male soma contains substances that inhibit female receptivity, as male accessory products do in insects (e.g., Simmons, 2001). Clearly, the connection between sexual cannibalism, male mating strategies, and monogyny needs to be explored. Irrespective of whether postinsemination sexual cannibalism represents a component of a monogynous mating strategy, it is still necessary to explain why females are cannibalistic in the first place. In many species,

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

159

there is variation in both female aggression and the ability (and/or motivation) of males to escape, yet this variation is poorly understood. One study suggests that females that reach sexual maturity in such good condition that they require no further food for egg production will not attack their mates (Schneider and Elgar, 2002). These data raise the possibility that female aggression against mates is determined by how quickly a female can change from a mating to a foraging mode. Comparisons between closely related taxa may provide insights. For example, the two species of Australian golden orb-web spiders Nephila inhabit apparently similar habitats and may occur in sympatry, yet our studies indicate that the species differ in the frequency of sexual cannibalism, sexual size dimorphism, male size variation, courtship behavior and certain aspects of their foraging behavior (e.g., Elgar et al., 2003; Griffiths et al., 2003; Schneider and Elgar, 2001; Schneider et al., 2000). Female spiders are generally food limited (Wise, 1993) and there is strong selection for traits that improve fecundity (e.g., Head, 1995). Thus, it is not unreasonable to assume that natural selection will favor rapacious foragers that do not reject conspecific prey. However, female foraging strategies that include sexual cannibalism (either directly or serendipitously) will be strongly opposed by counterselection on male mating success. Perhaps the strength of the latter explains why sexual cannibalism is not more widespread among spiders and other predacious invertebrates. Acknowledgments We thank Lutz Fromhage and John Prentas for helpful discussions. M.A.E. thanks the Australian Research Council for its long-term support of his research on sexual cannibalism. J.M.S. thanks the DFG (Deutsche Forschungsgemeinschaft) for funding her position and research. References Andersson, M. (1994). ‘‘Sexual Selection.’’ Princeton University Press, Princeton, NJ. Andrade, M. C. B. (1996). Sexual selection for male sacrifice in the Australian redback spider. Science 271, 70–72. Andrade, M. C. B. (1998). Female hunger can explain variation in cannibalistic behavior despite male sacrifice in redback spiders. Behav. Ecol. 9, 33–42. Andrade, M. C. B. (2003). Risky mate search and male self-sacrifice in redback spiders. Behav. Ecol. 14, 531–538. Andrade, M. C. B., and Banta, E. M. (2002). Value of remating and male functional sterility in redback spiders. Anim. Behav. 63, 857–870. Arnqvist, G. (1992). Courtship behaviour and sexual cannibalism in the fishing spider, Dolomedes fimbriatus (Clerck) (Aranea: Pisauridae). J. Arachnal. 20, 222–226. Arnqvist, G. (1998). Comparative evidence for the evolution of genitalia by sexual selection. Nature 393, 784–786.

160

ELGAR AND SCHNEIDER

Arnqvist, G., and Henriksson, S. (1997). Sexual cannibalism in the fishing spider and a model for the evolution of sexual cannibalism based on genetic constraints. Evol. Ecol. 11, 255–273. Arnqvist, G., and Nilsson, T. (2000). The evolution of polyandry: Multiple mating and female fitness in insects. Anim. Behav. 60, 145–164. Arnqvist, G., and Rowe, L. (2002). Antagonistic coevolution between the sexes in a group of insects. Nature 415, 787–789. Arnqvist, G., Jones, T. M., and Elgar, M. A. (2003). Sex-role reversed nuptial feeding in Zeus bugs. Nature 424, 387. Austad, S. N., and Thornhill, R. (1986). Female reproductive variation in a nuptial-feeding spider, Pisaura mirabilis. Bull. Br. Arachnol. Soc. 7, 48–52. Benton, T. (2001). Reproductive ecology. In ‘‘Scorpion Biology and Research’’ (P. Brownell and G. Polis, Eds.), pp. 278–301. Oxford University Press, Oxford. Birkhead, T. R., Lee, K. E., and Young, P. (1988). Sexual cannibalism in the praying mantis Hierodula membranacea. Behaviour 106, 112–118. Birkhead, T. R., and Møller, A. P. (1998). ‘‘Sperm Competition and Sexual Selection.’’ Academic Press, San Diego, CA. Blest, A. D. (1987). The copulation of a linyphiid spider, Baryphyma pratense: Does a female receive a blood meal from her mate? J. Zool. Lond. 213, 189–191. Bonduriansky, R. (2001). The evolution of male mate choice in insects: A synthesis of ideas and evidence. Biol. Rev. 76, 305–339. Bukowski, T. C., and Christenson, T. E. (2000). Determinants of mating frequency in the spiny orbweaving spider, Micrathena gracilis (Araneae: Araneidae). J. Insect Behav. 13, 331–352. Buskirk, R. E., Frohlich, C., and Ross, K. G. (1984). The natural selection of sexual cannibalism. Am. Nat. 123, 612–625. Cambridge, O.-P. (1871). Notes on some arachnida collected by Cuthbert Collingwood esq. M.D., during rambles in the China Sea, etc. Proc. Zool. Soc. Lond. (1871) 617–622. Crudgington, H. S., and Siva-Jothy, M. T. (2000). Genital damage, kicking and early death: The battle of the sexes takes a sinister turn in the bean weevil. Nature 407, 855–856. Darwin, C. (1871). ‘‘Sexual Selection and the Descent of Man.’’ Murray, London. Dick, J. T. A. (1995). The cannibalistic behavior of 2 Gammarus species (Crustacea, Amphipoda). J. Zool. Lond. 236, 697–706. Downes, J. A. (1978). Feeding and mating in the insectivorous ceratopogoninae (Diptera). Mem. Entomol. Soc. Canada 104, 1–62. Eberhard, W. G. (1996). ‘‘Female Control: Sexual Selection by Cryptic Female Choice.’’ Princeton University Press, Princeton, NJ. Eggert, A. K., and Sakaluk, S. K. (1994). Sexual cannibalism and its relation to male mating success in sagebrush crickets, Cyphoderris strepitans (Haglidae, Orthoptera). Anim. Behav. 47, 1171–1177. Elgar, M. A. (1989). Kleptoparasitism: A cost of aggregating for the orb-weaving spider Nephila edulis.. Anim. Behav. 37, 1052–1055. Elgar, M. A. (1991). Sexual cannibalism, size dimorphism and courtship behavior in orbweaving spiders (Araneae). Evolution 45, 444–448. Elgar, M. A. (1992). Sexual cannibalism in spiders and other invertebrates. In ‘‘Cannibalism: Ecology and Evolution among Diverse Taxa’’ (M. A. Elgar and B. J. Crespi, Eds.), pp. 129–156. Oxford University Press, Oxford. Elgar, M. A. (1995). Duration of copulation in spiders: Comparative patterns. Rec. West. Aust. Mus. (Suppl.) 51, 1–11.

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

161

Elgar, M. A. (1998). Sexual selection and sperm competition in arachnids. In ‘‘Sperm Competition and Sexual Selection’’ (T. R. Birkhead and A. P. Møller, Eds.), pp. 307–337. Academic Press, San Diego, CA. Elgar, M. A., and Crespi, B. J. (1992). Ecology and evolution of cannibalism. In ‘‘Cannibalism: Ecology and Evolution among Diverse Taxa’’ (M. A. Elgar and B. J. Crespi, Eds.), pp. 1–12. Oxford University Press, Oxford. Elgar, M. A., and Fahey, B. F. (1996). Sexual cannibalism, male–male competition and sexual size dimorphism in the orb-weaving spider Nephila plumipes. Behav. Ecol. 7, 195–198. Elgar, M. A., and Nash, D. R. (1988). Sexual cannibalism in the garden spider, Araneus diadematus (Araneae: Araneidae). Anim. Behav. 36, 1511–1517. Elgar, M. A., Ghaffer, N., and Read, A. (1990). Sexual dimorphism in leg length among orbweaving spiders: A possible role for sexual cannibalism. J. Zool. Lond. 222, 455–470. Elgar, M. A., Schneider, J. M., and Herberstein, M. E. (2000). Females control paternity in a sexually cannibalistic spider. Proc. R. Soc. Lond. Ser. B 267, 2439–2443. Elgar, M. A., Champion de Crespigny, F. E., and Ramamurthy, S. (2003a). Male copulation behaviour and the risk of sperm competition. Anim. Behav. 66, 211–216. Elgar, M. A., Bruce, M., Champion de Crespigny, F. E., Cutler, A. R., Cutler, C. L., Gaskett, A. C., Herberstein, M. E., Ramamurthy, S., and Schneider, J. M. (2003b). Male mate choice and patterns of paternity in the polyandrous, sexually cannibalistic orb-web spider, Nephila plumipes. Aust. J. Zool. 51, 357–365. Fahey, B. F., and Elgar, M. A. (1997). Sexual cohabitation as mate-guarding in the leaf-curling spider Phonognatha graeffi Kerserling (Araneoidea, Araneae). Behav. Ecol. Sociobiol. 40, 127–133. Foellmer, M. W., and Fairbairn, D. J. (2003). Spontaneous male death during copulation in an orb-weaving spider. Proc. R. Soc. Lond. B 270, S183–S185. Forster, L. M. (1992). The stereotyped behaviour of sexual cannibalism in Latrodectus hasselti Thorell (Araneae: Theridiidae): The Australian redback spider. Aust. J. Zool. 40, 1–11. Fromhage, L., Elgar, M. A., Schneider, J. M. Faithful without care: The evolution of monogyny. (Unpublished manuscript.) Fromhage, L., Uhl, G., and Schneider, J. M. (2003). Fitness consequences of sexual cannibalism in female Argiene bruennichi. Behav. Ecol. Sociobiol. 55, 60–64. Gaskett, A. C., Herberstein, M. E., Downes, B. J., and Elgar, M. A. (2004). Life-time male mating preferences in a sexually cannibalistic orb-web spider (Araneae: Araneidae). Behaviour. In press. Gould, S. J. (1984). Only his wings remained. Nat. Hist. 93, 10–18. Griffiths, B. V., Holwell, G. I., Herberstein, M. E., and Elgar, M. A. (2003). Frequency, composition and variation in external food stores constructed by orb-web spiders: Nephila edulis and Nephila plumipes (Araneae: Araneoidea). Aust. J. Zool. 51, 119–128. Hebets, E. A. (2003). Subadult experience influences adult mate choice in an arthropod: Exposed female wolf spiders prefer males of a familiar phenotype. Proc. Natl. Acad. Sci. USA 100, 13390–13395. Head, G. (1995). Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (class Araneae). Evolution 49, 776–781. Herberstein, M. E., Schneider, J. M., and Elgar, M. A. (2002). Costs of courtship and mating in a sexually cannibalistic orb-web spider: Female and male strategies. Behav. Ecol. Sociobiol. 51, 440–446. Jamieson, I. G. (1986). The functional approach to behaviour: Is it useful? Am. Nat. 127, 195–208. Jennions, M. D., and Petrie, M. (2000). Why do females mate multiply? A review of the genetic benefits Biol. Rev. 75, 21–64.

162

ELGAR AND SCHNEIDER

Johns, P. M., and Maxwell, M. R. (1997). Sexual cannibalism: Who benefits? Trends Ecol. Evol. 12, 127–128. Johnson, J. C. (2001). Sexual cannibalism in fishing spiders (Dolomedes triton): An evaluation of two explanations for female aggression towards potential mates. Anim. Behav. 61, 905–914. Knoflach, B., and van Harten, A. (2001). Tidarren argo sp. nov. (Araneae: Theridiidae) and its exceptional copulatory behaviour: Emasculation, male palpal organ as a mating plug and sexual cannibalism. J. Zool. Lond. 254, 449–459. Lessells, C. M. (1999). Sexual conflict in animals. In ‘‘Levels of Selection in Evolution’’ (L. Keller, Ed.), pp. 75–99. Princeton University Press, Princeton, NJ. Liske, E., and Davis, W. J. (1987). Coutship and mating behaviour of the chinese praying mantis, Tenodera aridifolia sinensis.. Anim. Behav. 35, 1524–1537. Maxwell, M. R. (1999). The risk of cannibalism and male mating behaviour in the Mediterranian praying mantid Iris oratoria. Behaviour 136, 205–219. Maxwell, M. R. (2000). Does a single meal affect reproductive output in the sexually cannibalistic praying mantid Iris oratoria? Ecol. Entomol. 25, 54–62. Newman, J. A., and Elgar, M. A. (1991). Sexual cannibalism in orb-weaving spiders: An economic model. Am. Nat. 138, 1372–1395. Parker, G. A. (1974). Courtship persistence and female-guarding as male time investment strategies. Behaviour 48, 157–194. Parker, G. A. (1979). Sexual selection and sexual conflict. In ‘‘Sexual Selection and Reproductive Competition in Insects’’ (M. S. Blum and N. A. Blum, Eds.), pp. 123–166. Academic Press, London. Parker, G. A. (1998). Sperm competition and the evolution of ejaculates: Towards a theory base. In ‘‘Sperm Competition and Sexual Selection’’ (T. R. Birkhead and A. P. Møller, Eds.), pp. 3–54. Academic Press, San Diego, CA. Peretti, A. V., Acosta, L. E., and Benton, T. G. (1999). Sexual cannibalism in scorpions: Fact or fiction. Biol. J. Lin. Soc. 68, 485–496. Polis, G. A., and Sissom, W. D. (1990). Life history. In ‘‘The Biology of Scorpions’’ (G. A. Polis, Ed.), pp. 161–223. Stanford University Press, Stanford, CA. Prenter, J., Elwood, R. W., and Montgomery, W. I. (1994). Male exploitation of female predatory behaviour reduces sexual cannibalism in male autumn spiders, Metellina segmentata. Anim. Behav. 47, 235–236. Rice, W. R. (1996). Sexually antagonistic male adaptation triggered by experimental arrest of female evolution. Nature 381, 232–234. Robinson, R. H., and Robinson, B. (1980). Comparative studies of the courtship and mating behaviour of tropical araneid spiders. Pac. Insects Monogr. 36, 1–218. Rowe, R., Arnqvist, G., Sih, A., and Krupa, J. J. (1994). Sexual conflict and the evolutionary ecology of mating patterns: Water striders as a model system. Trends Ecol. Evol. 9, 289–293. Sakaluk, S. K., Bangert, P. J., Eggert, A. K., Gack, C., and Swanson, L. V. (1995). The gin trap as a device facilitating coercive mating in sagebrush crickets. Proc. R. Soc. Lond. B 261, 65–71. Sasaki, T., and Iwahashi, O. (1995). Sexual cannibalism in an orb-weaving spider Argiope aemula. Anim. Behav. 49, 1119–1121. Schneider, J. M., and Elgar, M. A. (2001). Sexual cannibalism and sperm competition in the golden orb-web spider Nephila plumipes (Araneoidea): Female and male perspectives. Behav. Ecol. 12, 547–552. Schneider, J. M., and Elgar, M. A. (2002). Sexual cannibalism in Nephila plumipes as a consequence of female life history strategies. J. Evol. Biol. 15, 84–91.

EVOLUTIONARY SIGNIFICANCE OF SEXUAL CANNIBALISM

163

Schneider, J. M., and Lubin, Y. (1997). Infanticide by males in a spider with suicidal maternal care, Stegodyphus lineatus. Anim. Behav. 54, 305–312. Schneider, J. M., and Lubin, Y. (1998). Intersexual conflict in spiders. Oikos 83, 496–506. Schneider, J. M., Herberstein, M. E., Champion de Crespigny, F., Ramamurthy, S., and Elgar, M. A. (2000). Sperm competition and small size advantage for males of the golden orbweb spider Nephila edulis. J. Evol. Biol. 13, 939–946. Schneider, J. M., Thomas, M. L., and Elgar, M. A. (2001). Ectomised conductors in the golden orb-web spider Nephila plumipes (Araneoidea): A male adaptation to sexual conflict. Behav. Ecol. Sociobiol. 49, 410–415. Schneider, J. M., Fromhage, L., and Uhl, G. (2004). Extremely short copulations do not affect hatching success in Argiope bruennichi SCOPOLI, 1772 (Araneidae). J. Arachnol. In press. Simmons, L. W. (2001). ‘‘Sperm Competition and its Evolutionary Consequences in the Insects.’’ Princeton University Press, Princeton, NJ. Simmons, L. W., and Parker, G. A. (1989). Nuptial feeding in insects: Mating effort versus paternal investment. Ethology 81, 332–343. Spence, J. R., Zimmerman, M., and Wojcicki, J. P. (1996). Effects of food limitation and sexual cannibalism on reproductive output of the fishing spider Dolomedes triton (Araneae: Pisauridae). Oikos 75, 373–382. Sta¨lhandske, P. (2001). Nuptial gift in the spider Pisaura mirabilis maintained by sexual selection. Behav. Ecol. 12, 691–697. Stratton, G. E., Hebets, E. A., Miller, P. R., and Miller, G. L. (1996). Patterns and duration of copulation in wolf spiders (Araneae, Lycosidae). J. Arachnol. 24, 186–200. Stockley, P. (1997). Sexual conflict resulting from adaptations to sperm competition. Trends Ecol. Evol. 12, 154–159. Thornhill, R. (1983). Cryptic female choice and its implications in the scorpionfly Harpobittacus nigriceps. Am. Nat. 122, 763–788. Thornhill, R., and Alcock, J. (1983). ‘‘The Evolution of Insect Mating Systems.’’ Harvard University Press, Cambridge, MA. Uhl, G., and Vollrath, F. (1998). Little evidence of size-selective sexual cannibalism in two species of Nephila (Araneae). Zoology 101, 101–106. Vahed, K. (1998). The function of nuptial feeding in insects: review of empirical studies. Biol. Rev. 73, 43–78. Vollrath, F. (1998). Dwarf males. Trends Ecol. Evol. 13, 159–163. Vollrath, F., and Parker, G. A. (1992). Sexual dimorphism and distorted sex ratios in spiders. Nature 360, 156–159. Wise, D. H. (1993). ‘‘Spiders in Ecological Webs.’’ Cambridge University Press, Cambridge.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Social Modulation of Androgens in Vertebrates: Mechanisms and Function Rui F. Oliveira instituto superior de psicologia aplicada 1149-041 lisbon, portugal

I. Introduction The approaches of the social and the biological sciences to the study of behavior have been seen as almost mutually exclusive. The nature versus nurture debate has been almost permanently present in the history of the behavioral sciences. As examples one can mention the Cartesian dualism in the neurosciences, the Lorenz versus Lehrman debate on innate behaviors and on the nature of instinct in classic ethology in the 1950s (Lehrman, 1953, 1970; Lorenz, 1939, 1965), or the antagonistic views of social constructivism versus genetic (biological) determinism in psychology (e.g., Lewontin et al., 1984). However, more recently, a growing body of literature has documented social influences on genetic constitution and gene expression, functioning of the endocrine and nervous systems, and immune activity (Cacioppo et al., 2000). Thus, the effects of social factors on the expression of behavior may involve underlying biological processes. Therefore, the classic dichotomy of nature versus nurture should be abandoned and it should be recognized that for most behavioral traits there is a nonadditive contribution of both biological and social factors, and that the latter are expressed through biological mechanisms. This conciliatory view has emerged in psychology and has been labeled social neuroscience (Cacioppo and Berntson, 2002). The central topic of this review, the social modulation of androgens, its mechanisms and function, can be seen as another contribution toward expanding this new view of nature and nurture, as being complementary rather than mutually exclusive, to the field of behavioral endocrinology. The responsiveness of the endocrine system to social stimuli has been well established in behavioral endocrinology for many years (e.g., Cannon, 1929; Selye, 1976). However, most studies have concentrated on the response of the hypothalamus–pituitary–adrenocortical (HPA) axis 165 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

166

RUI F. OLIVEIRA

to stressors, and glucocorticoids came to be known as stress hormones (Sapolsky, 2002). On the other hand, less research has been carried out on the response of the reproductive axis (i.e., hypothalamus– pituitary–gonadal axis, HPG) to the social environment, and on its behavioral significance. Sex steroids, as the name implies, have been classically viewed as hormones directly involved in reproduction, and other potential roles for these hormones have only started to be hypothesized. Gonadal steroids, being secreted by cell populations intimately associated with gamete-producing cells, are especially well suited as coordinating agents between gonadal maturation and the expression of displaying traits (Oliveira and Almada, 1999). Displaying traits, both behavioral and morphological, have associated costs that are compensated for only when the individual is effectively capable of mating. In other words, it would make no sense from an evolutionary perspective for an individual to express reproductive behaviors and secondary sex characters if its gonads are not ready for gamete release and if there is no partner available. Thus, androgen responsiveness to the social environment can be expected in adult males in a breeding context, as a way to coordinate an integrative response of the organism to the environmental conditions (Wingfield et al., 1990, 1999, 2000). More specifically, the social modulation of androgens can be viewed as a mechanism for adjusting androgen-dependent behaviors to the current social environment of the individual. According to this hypothesis, the social interactions in which an individual participates influence its androgen levels, which in turn will modulate perceptive, motivational, and cognitive mechanisms as well as somatic releasers, which in turn may affect its subsequent behavior in social interactions (Fig. 1). In the current review, the mechanisms of androgen action on behavior are discussed first. The evidence for social modulation of androgen levels are then presented. Finally, androgen responsiveness to the social environment is discussed using an integrative approach, that is, by exploring both its proximate (i.e., physiological/developmental) and ultimate (i.e., functional/evolutionary) causes. The article ends with a discussion on the social modulation of androgens and its behavioral consequences in humans.

II. Androgens as Causal Agents of Behavior A. Historical Background In the words of one of its founders ‘‘behavioral endocrinology has a short history but a long past’’ (Beach, 1974). With these words Frank Beach meant that although the discipline was formally founded in 1948 with the

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

167

Fig. 1. Model for the interplay between androgens and social behavior. Androgens can affect the expression of an individual’s behavior either by acting on the neural mechanisms underlying behavior (i.e., perception, motivation, and cognition) or by changing social releasers at the periphery of the organism. The social interactions in which an individual participates may feedback on its androgen levels, which in turn will affect its behavior in subsequent social interactions (gray area, nervous system).

publication of its first textbook (Beach, 1948), knowledge about the effects of hormones on behavior was present even in ancient societies. A vivid view of ancient endocrinology is provided by Sapolsky (1997) in his essay ‘‘The Trouble with Testosterone,’’ where he describes the probable scenario of the discovery of the effects of testosterone (T) on behavior as follows: A dozen millennia years ago or so, an adventurous soul managed to lop off a surly bull’s testicles and thus invented behavioral endocrinology. It is unclear from historical records whether this individual received either a grant or tenure as a result of this experiment, but it certainly generated an influential finding— something or other comes out of the testes that helps to make males such aggressive pains in the ass. That something or other is testosterone.

In fact, since mammals have their testes located in a scrotum outside the body cavity, which facilitates their removal, it is not hard to imagine that during the process of animal domestication our neolithic farmer ancestors accidentally discovered the benefits of castration. Thus, the link between something produced by the testes and behavior may have been implicitly established for some millennia (Freeman et al., 2001). The castration of boys before puberty to serve as eunuchs (i.e., harem attendants) started before 700 b.c. in Asian courts. The recruitment of these loyal servants, who would not be tempted to challenge the paternity of their lords, to guard and serve in harems became a common practice in the courts of Egypt, Byzantium, and China (Scholz, 2001). In China they were used as

168

RUI F. OLIVEIRA

guards of the Emperor’s inner court and their numbers reached a peak of 70,000 at the end of the Ming Dynasty, when they gained immense political power (Tsai, 1996). Human castration has also been used as a form of punishment in some societies (e.g., from the Sui Dynasty in China to some current Western societies in the case of sexual offenders), and as an act of rejection of sexual identity in some religious traditions (e.g., ancient paganism, Christianity, and Buddhism; Scholz, 2001). In the sixteenth century the effects of castration on the maintenance of a higher pitched voice in boys were used to produce male sopranos to sing in church, where women were not allowed to sing. These sopranos were known as castrati and became very popular in the seventeenth and eighteenth centuries, singing in church choirs (e.g., the Sistine Chapel in the Vatican) and as male leaders in opera, mainly in Italy. This practice was banned in the late nineteenth century, when females or countertenors started to take their roles in operas (Jenkins, 2000). The Sistine Chapel stopped using castrati in 1903; the last of the castrati, Alessandro Moreschi, was recent enough to have made gramophone recordings and died in 1924 (Jenkins, 2000). In all these cases the knowledge of the link between prepubertal castration and physical and behavioral changes was implicit: castration was known to prevent the emergence of male sexual characters such as the development of muscle mass, the growth of the penis, the masculine pattern of fat accumulation, deepening of the voice, and a lack of interest in sexual behaviors. Explicit knowledge of the effects of androgens on behavior was clearly present in written form in classical antiquity. In his Historia Animalium (ca. 350 b.c.) Aristotle describes the effects of castration in birds and in men, and subsequently a humoral basis of biological functioning was established in the Western world with the theory of the four humors to explain the equilibrium between health and disease (Freeman et al., 2001). Only in the eighteenth century did the dawn of the experimental approach to behavioral endocrinology start to emerge with the first attempts at testicular transplantation. John Hunter (1728–1793) was the first to transfer the testis of a cock into the peritoneal cavity of a hen, but he was more interested in the surgical technique of tissue transplantation than in its effects (Schultheiss et al., 2000). Thus, it was not until 1849 that the first experiment in behavioral endocrinology was performed. Arnold Berthold (1801–1863) reimplanted or transplanted testes in castrated cockerels and showed that, by returning the testis to the abdominal cavity, the expression of male behaviors such as mounting, fighting, and crowing, was restored, while castrated cockerels did not develop either male behaviors or male secondary sex characters, such as the development of a comb. Since the implanted testis lacked neural connectivity to the animal, Berthold

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

169

concluded that the testes must secrete a substance into the bloodstream that affects the expression of these behavioral and morphological traits (Freeman et al., 2001; Nelson, 2000). Berthold’s study is now credited as a landmark in endocrinology since it established the concept of hormone, that is, the ‘‘secretory blood-borne product’’ in his own words (Nelson, 2000). The term hormone was later introduced into the scientific vocabulary by Ernest Starling in 1905 (Freeman et al., 2001). Meanwhile, testis-borne hormones started to be seen as causal agents of intellectual and sexual performance, which led to their use in physical and sexual rejuvenation. The first proponent of the rejuvenation hypothesis was Brown-Se´quard (1817–1894) who, at the age of 72 years, experimented on himself with an injection of animal testicular extract (Schultheiss et al., 1997). Subsequently Steinach (1861–1944) proposed vasectomy as a way to increase male hormone production, on the ground that blocking of the secretory output would increase the endogenous levels. This rejuvenation treatment became very popular and notable figures such as Freud (1856–1939) and Yeats (1865–1939) underwent the ‘‘Steinach operation’’ (Freeman et al., 2001; Schultheiss et al., 1997; Wyndham, 2003). B. Androgens: From Deterministic Factors to Neuromodulators Following Berthold’s study the ablation and replacement experiment became the classic paradigm in behavioral endocrinology, to establish a link between a given hormone and a given behavior. First, a tissue suspected of being the source of a particular hormone was removed and a decrease in the expression of a hormone-dependent behavior was predicted. If so, exogenous administration of the suspected active hormone should restore the expression of the affected behavior. Using this approach, androgens have been implicated in the expression of agonistic and sexual behaviors in a number of vertebrate species (for reviews, see Ball and Balthazart, 2002; Baum, 2002; Nelson, 2000). Berthold’s study can also be used as an example to introduce a classic dichotomy in behavioral endocrinology: the division of influences of androgens on behavior into activational versus organizational effects. In fact, Berthold manipulated (e.g., castrated) the individuals at a young age and looked for effects of the treatment several months later in adult animals. Therefore, his experiment is an example of how a manipulation at an early stage of the life of the individual might have a permanent effect on its adult behavior, that is, an organizational effect (Arnold and Breedlove, 1985). On the other hand, in adulthood hormones may affect behavior in a transient way, by activating proximate mechanisms underlying behavior, that is, by having an activational effect on behavior (Arnold and

170

RUI F. OLIVEIRA

Breedlove, 1985). This dichotomy was formally introduced by Phoenix and associates while studying the sexual behavior of female guinea pigs (Phoenix et al., 1959). Studies in which the behavioral effects of castration were reversed by androgen replacement, either at the organizational level or at the activational level, led to the view of hormones as deterministic causal agents of behavior, acting as ‘‘pushbuttons’’ on the display of particular behaviors. This view was challenged by experiments in which androgen replacement therapy was not on its own sufficient to activate a behavior, but would be effective in the presence of the right stimuli (e.g., Albert et al., 1993). These results suggest that androgens are necessary for the persistence of the behavior but are not sufficient to activate the expression of the behavior. Another type of evidence that supports this view comes from electrophysiological studies of the stria terminalis, a key component in the neural circuits of aggressive behavior in mammals, which connects the amygdala and the hypothalamus. In castrated Wistar rats T microinjections into the stria terminalis increase its neural activity by reducing the refractory period of action potentials (Kendrick and Drewett, 1979). However, the administration of T to the stria terminalis in the absence of preexisting neural activity has no effect. These results support the idea that T per se does not promote aggression but that it may exaggerate a preexisting pattern of aggression. The accumulation of these kinds of data shifted the conceptual paradigm of behavioral endocrinology from hormones as deterministic agents of behavior toward a more probabilistic view, according to which androgens started to be seen as facilitators of behavior. That is, the hormones would not activate per se the expression of the behavior but would increase the probability of its expression, by acting as modulators of the neural pathways underlying that behavioral pattern (Simon, 2002). Many studies have implicated androgens in the modulation of different neurotransmitter pathways, for example, in the serotonergic and vasopressinergic systems involved in the control of social behaviors in mammals (e.g., Simon, 2002). C. Mechanisms of Androgen Action on Behavior The above-mentioned neurochemical pathways modulated by androgens can be part of one of three major functional compartments of the nervous system: sensory systems (i.e., information input systems), central processors (i.e., the central nervous system) and effectors (i.e., output systems) (Nelson, 2000; Fig. 2). Androgens can also affect behavior by acting peripherally on somatic structures (Hinde, 1970) that have a role as sign

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

171

Fig. 2. Routes for the effects of androgens on behavior (gray area, nervous system).

stimuli or releasers (in the sense of Tinbergen, 1951), thus evoking a behavioral response in conspecifics (Fig. 2; see Table I for examples). 1. Central Actions of Androgens: Motivational Systems The localization of sex steroid receptors, both androgen receptors (ARs) and estrogen receptors (ERs), in the brain was a first indication that the brain is a major target of androgen action related to behavioral expression. Furthermore, identification of the specific brain areas that express AR and/ or ER led to detection of neuronal circuits involved in the control of behavior that could be influenced by gonadal steroids. It should be noted here that some of the behavioral effects of androgens are mediated by estradiol (E2) after it is formed by the aromatization of T (e.g., Balthazart et al., 1995). In fact, in most cases T acts as a prohormone that needs to be metabolized either into another biologically active androgen (e.g., 5-dihydrotestosterone [DHT] in mammals or 11-ketotestosterone [KT] in fish), or into E2 (through aromatization) to exert its effects using either AR or ER, respectively. In terms of the neural localization of these circuits, after the earlier maps of brain AR and ER were compiled, it became clear that the preoptic area/anterior hypothalamic region and a few other limbic areas were the main targets of androgens in the vertebrate brain (Ball and Balthazart, 2002; Schulkin, 2002), which suggested that motivational systems were one of the main target circuits. By acting on neural systems

TABLE I Examples of Different Modes of Action of Androgens on Peripheral Mechanisms Underlying Behavior Mode of action A. Sensory systems

B. Effectors

172 C. Somatic releasers

Example 1. Testosterone-treated juvenile cyprinid fish show an increased electro-olfactogram response to female sex pheromone and increased sexual behavior toward females 2. Male stingrays (Dasyatis sabina) treated with dihydrotestosterone shift the frequency response of their ampullary electroreceptors, increasing their mate electrolocation efficiency 3. Androgens increase penile sensitivity in rats, promoting ejaculation 1. Testosterone treatment increases syringeal muscle mass and the density of acetylcholine receptors in syringeal muscles of male zebra finches Taeniopygia guttata) and inhibits the activity of cholinesterase in the neuromuscular junctions of the syrinx 2. Androgen treatment induces the differentiation of the laryngeal motor neurons, muscle fibers, and laryngeal cartilage of the male African frog (Xenopus laevis) vocal system used to produce mate calls 3. Castration reduces and testosterone replacement therapy restores the sex-pheromone content in the abdominal glands of red-bellied newt (Cynops pyrrrhogaster) males 4. Castration reduces and testosterone reinstates pheromone production and scent marking in meadow voles (Microtus pennsylvanicus), tree shrews (Tupaia belangeri), and Wistar rats 1. Nuptial coloration is suppressed in castrated males and promoted by androgen treatment in African cichlids and sticklebacks (Gasterosteus aculeatus)

2. Testosterone induces the development of the sword in male swordtail (Xiphophorus helleri) caudal fin, which is used by females in mate choice 3. Castration suppresses and testosterone induces the plumage ornamental coloration in charadriform birds

Ref. Cardwell et al. (1995) Sisneros and Tricas (2000)

Beach and Levinson (1950); Larsson et al. (1973) Bleisch et al. (1984); Luine et al. (1980) Kelley (2002)

Yamamoto et al. (1996) Ferkin and Johnston (1993); Gawienowsky et al. (1976); Holst and Eichman (1998); Manzo et al. (2002) Fernald (1976); Ikeda (1933); Levy and Aronson (1955); Rouse et al. (1977); Wapler-Leong and Reinboth (1974) Baldwin and Goldin (1939); Rosenthal and Evans (1998) Kimball and Ligon (1999)

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

173

underlying motivation, androgens may affect both appetitive (e.g., courtship) and consummatory (e.g., copulation) aspects of behavior (Ball and Balthazart, 2002). As an example of an androgen effect on a motivational system underlying appetitive behaviors, one can mention the effects of T on territorial aggression and flank-marking behavior in rats and hamsters (Mesocricetus auratus), by preventing apoptosis of arginine vasopressin (AVP) neurons and by activating these neurons in the medial preoptic area, the anterior hypothalamus, and the medial amygdala (De Vries, 1995; Ferris et al., 1997). On the other hand, the induction of copulatory behavior in male birds by implants of T placed in the preoptic area is a classic example of an androgen effect on a motivational system underlying consummatory behaviors (Barfield, 1969, 1971; Hutchison, 1971; for a detailed review see Ball and Balthazart, 2002). However, central actions of androgens are not restricted only to motivational systems, since they can also affect central neural circuits underlying perception and cognition [for reviews on these topics, see Becker (2002) and Dohanich (2002), respectively]. 2. Central Actions of Androgens: Cognitive Performance The effects of androgens on cognitive function and on perception have been investigated using two complementary approaches: (1) documenting the occurrence of AR and ER in brain areas known to be involved in these processes, and (2) testing hormone-treated subjects in perceptual and cognitive tasks. The presence of AR in the hippocampus of mammals and birds (Kerr et al., 1995; Saldanha et al., 1999), a brain area known to be involved in relational memory processes, namely in spatial memory (Eichenbaum et al., 1992; Squire, 1992), and in cortical pyramidal cells in rats, monkeys, and humans (Kerr et al., 1995; Pomerantz and Scholl, 1987; Tohgi et al., 1995) supports the potential role of androgens as modulators of cognitive mechanisms in birds and mammals. ARs have also been found in the teleost lateral telencephalic pallium (Gelinas and Callard, 1997) and in the reptilian medial cortex (Rosen et al., 2002), which are neural mechanisms homologous to the mammalian/avian hippocampus that are selectively involved in spatial cognition both in fish and in reptiles (Salas et al., 2003). Also, cognitive performance can be influenced by androgens. Castration of male rodents and birds reduces, and systemic androgen replacement restores, their performance in a number of cognitive tasks. T facilitates conspecific song discrimination in zebra finches, Taeniopygia guttata (Cynx and Nottebhom, 1992), and in the case of rodents castration reduces and androgen replacement restores their performance in the following tasks: object recognition, radial arm maze, T-maze, inhibitory avoidance, and social memory (Ceccarelli et al., 2001; Frye and Seliga,

174

RUI F. OLIVEIRA

2001; Harrel et al., 1990; Havens and Rose, 1992; Kritzer et al., 2001; Sawyer et al., 1984; Vazquez-Pereyra et al., 1995). Furthermore, DHT intrahippocampal implants enhance cognitive performance of male rats (Frye et al., 2004), probably because of neuroprotective actions of androgens in the hippocampus (Frye and Reed, 1998; Mizoguchi et al., 1992). 3. Central Actions of Androgens: Perception and Action Apart from affecting brain areas directly involved in motivational states (e.g., hypothalamus and limbic system) or cognitive processes (e.g., hippocampus), androgens can also modulate the functioning of brain areas involved in the processing of sensorimotor information. The fact that there are often sex differences in perceptual thresholds suggests a potential role for sex steroids in the processing of sensory information (Becker, 2002). For example, when exposed to an electric shock, female rats have shorter latencies in an escape response and lower shock thresholds (Beatty, 1979). Such differences are found even when nociceptive reactivity is similar for both sexes, which indicates that the sex difference is located in the brain mechanisms mediating pain perception and not at the level of peripheral sensitivity (Ryan and Maier, 1988). This conclusion is further supported by the fact that shock sensitivity can be manipulated by organizational androgen treatments (Beatty, 1979). Motor circuits in the brain can also be sexually dimorphic. The vocal pattern generator in Xenopus is more active in males than in females (Wetzel et al., 1985) and males have a larger number of laryngeal motoneurons and of vocal interneurons (Kelley and Dennison, 1990; Kelley et al., 1988). These sex differences suggest that androgens may have organizational effects on the development of the vocal circuitry in amphibia. Androgens can also prevent the apoptosis of motoneurons in the spinal nucleus of the bulbocavernosa system that regulates penile erection in mammals (Breedlove and Arnold, 1981, 1983). Finally, the presence of ARs has been documented in the descending pathways of the song-control system in songbirds (Sohrabji et al., 1989; for a review, see Schlinger and Brenowitz, 2002). 4. Peripheral Actions of Androgens There is a growing body of evidence for the peripheral actions of androgens both on sensory systems and on effector pathways. ARs have been identified in several effector pathways, involved in a wide range of behaviors in different vertebrate taxa, ranging from electrocytes that produce an electric organ discharge which is used as a social signal in weakly electric fish (Dunlap and Zakon, 1998), to the perineal muscles (i.e., bulbocavernosus and levator ani) that innervate the penis in mammals and thus are

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

175

involved in copulatory behavior (Fishman et al., 1990). In the same vein, there is compelling evidence that sensory systems are the target of sex steroids including androgens. For example, ARs (and ERs) have been identified in a number of ocular structures of mammals including humans (Wickham et al., 2000), and 5--reductase and aromatase activity have been detected in the retina of humans and teleosts respectively (Callard et al., 1993; Rocha et al., 2000), suggesting that androgens might interfere in the peripheral processing of visual stimuli in vertebrates. ARs have also been found in the auditory system of a vocal lizard (Gekko gecko), namely in the cochlear nuclei and in the torus semicircularis (Tang et al., 2001), again suggesting a role for androgens in the peripheral processing of sensory (acoustic) stimuli. Apart from the localization of AR in sensory systems there are several experimental studies that demonstrate a peripheral effect of androgens on behavior. In Table I a few examples have been selected from the literature to illustrate these effects, trying to cover different behavioral modalities (i.e., acoustic, visual, chemical) in different vertebrate taxa.

III. Behavioral Feedback on Endocrine Function A. Historical Background Contrary to the long history of implicit knowledge about the effects of androgens on behavior, the idea that hormones can be affected by behavior is a relatively recent concept. Apart from seasonal variations relating to reproductive function, levels of sex steroids were thought to be relatively stable over time. The first indications suggesting that social stimuli could influence sex steroid secretion came from a series of studies during the 1950s on the effects of the social environment on female reproductive cycles in mammals. First, Van der Lee and Boot (1955) showed that female mice (Mus musculus) housed together in the absence of conspecific males extended the length of their cycles by increasing the diestrous stage. This condition, considered similar to pseudopregnancy, came to be known as the Lee–Boot effect. Another effect of the social environment on the female resproductive cycle was described shortly afterward by Whitten (1956). This consisted in estrus induction by the presence of a familiar male and estrus synchronization among females within the same colony (the Whitten effect). Finally, Bruce (1959) found that exposure of pregnant females to an unfamiliar male would induce pregnancy failure, that was subsequently followed by mating with the ‘‘alien male’’ within 3–6 days (the Bruce effect). All these effects are mediated by chemical signals

176

RUI F. OLIVEIRA

emitted by conspecifics. By this time, social influences on male sexual behavior had also started to be described. The most famous effect was the rearousal of male rats after mating, induced by the presence of a novel female (Wilson et al., 1963), which was termed the ‘‘Coolidge effect’’ by Frank Beach (Bermant, 1976) in an allusion to the following story involving the former U.S. President Calvin Coolidge and the First Lady: One day the President and Mrs. Coolidge were visiting a government farm. Soon after their arrival they were taken off on separate tours. When Mrs. Coolidge passed the chicken pens she paused to ask the man in charge if the rooster copulates more than once each day. ‘‘Dozens of times,’’ was the reply. ‘‘Please tell that to the President,’’ Mrs. Coolidge requested. When the President passed the pens and was told about the roosters, he asked, ‘‘Same hen every time?’’ ‘‘Oh no, Mr. President, a different one each time.’’ The President nodded slowly, then said, ‘‘Tell that to Mrs. Coolidge.’’ (Bermant, 1976, pp. 76–77)

The Coolidge effect may be explained by transient variations in androgen levels. When male mice are first exposed to a female they show a rapid increase in luteinizing hormone (LH) levels, while successive presentations of the same female progressively elicit smaller increases, suggesting an habituation effect. However, the presentation of a new female to these unresponsive males induces an acute burst of LH secretion (Coquelin and Bronson, 1979). These LH variations may mediate variations in androgen responsiveness to the presence of the female, which in turn may influence the activity of the mesolimbic dopaminergic system, involved in the initiation and maintenance of sexual behavior. In agreement with this hypothesis, an increase in dopamine transmission in the nucleus accumbens in satiated males, when exposed to a novel female, that peaks during copulation with the new female, has been demonstrated (Fiorino et al., 1997). In the early 1940s the suggestion that agonistic interactions and social dominance could affect sex hormone secretion was proposed for male mice (Ginsburg and Allee, 1942). Later on, contradictory results were collected with some studies suggesting an effect of social challenges on the reproductive axis (e.g., Bronson and Eleftheriou, 1964; Bronson and Marsden, 1973; Bronson et al., 1973) whereas others found no effect (e.g., Vale et al., 1970, 1971). It was only with the development of immunoassays (e.g., radioimmunoassays) that a more precise and sensitive method became available to measure rapid changes in circulating levels of hormones in response to social interactions. With this technique a new experimental paradigm in behavioral endocrinology emerged: to manipulate the social environment or to expose the subject to a behavioral/social stimulus while measuring pre- and postexposure levels of the hormone in the plasma to assess its responsiveness. Using this new approach, Bernstein and associates conducted a series of studies, using rhesus monkeys, that showed that

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

177

both male–male fighting behavior (e.g., Rose et al., 1971) and access to females (e.g., Rose et al., 1972) influence T levels in males (for a review see Bernstein et al., 1983). Data of this type led Leshner (1983, p. 397) to propose that ‘‘hormonal responses to competition    not only provide a mechanism for physiological adaptation to the stresses imposed by that experience but also feed back and affect ongoing and future agonistic response to patterns.’’ This hypothesis, describing the feedback effect of hormones on behavior, was termed by Leshner the ‘‘behavioral feedback hypothesis’’ (Leshner, 1979), and was the first formal conceptualisation of the interplay between hormones and behavior. However, two studies published during the 1970s (Maruniak et al., 1977; Nock and Leshner, 1976), both using mice, raised some difficulties for this hypothesis. To test whether hormonal responses to competition were behaviorally effective they manipulated the postexperience endocrine levels by castrating the subjects and subsequently replacing the target hormone with the dose expected to be induced by the behavioral interaction. In both studies, there was no effect of preventing the behaviorally induced changes in androgen levels on either ongoing or future agonistic behavior. Meanwhile data on a wider variety of vertebrate taxa and in field conditions started to be collected (e.g., Harding and Follett, 1979; Wingfield, 1984b, 1985) and evidence accumulated suggesting that both exposure to and interaction with potential mates, and male–male competition, influence circulating levels of androgens (Harding, 1981; Oliveira et al., 2002; Wingfield, 1999, 2000, 2001). In male–male contests a general pattern has been described for the variation of circulating T concentrations. At an early stage of the encounter there is a short-term increase in T levels in both participants, which might contribute to the escalation of the fight (Huntingford and Turner, 1986). After the fight is over and the status of each animal has been established, winners keep T levels high while losers experience a sharp decrease in circulating T (Harding, 1981). These endocrine changes tend to last for longer periods in subordinates than in dominants (Harding, 1981), and in subordinates they can be reversed by manipulating their social status (e.g., mice: Machida et al., 1978; rats: Schuurman, 1980) or by giving them access to females (e.g., rhesus monkeys: Rose et al., 1975). B. Social Interactions Affect Androgen Levels in Males: The Challenge Hypothesis As shown above, androgens can be viewed as causal agents of appetitive and consumatory aspects of sexual and aggressive behavior while, on the other hand, the endocrine system is responsive to social stimuli and to

178

RUI F. OLIVEIRA

social interactions in which the animal is involved. Several studies have shown rapid effects of social interactions on transient changes in androgen levels in a wide variety of vertebrate taxa, ranging from fish to primates including humans. This set of results led to the proposal of the ‘‘challenge hypothesis’’ by Wingfield and associates (Wingfield et al., 1987, 1990), according to which the social interactions in which the subject is involved would determine its androgen levels. The challenge hypothesis postulates the existence of three levels of circulating androgens: (1) a constitutive level (level A) inherent in the baseline activity of secretory (Leydig) cells; (2) a breeding level (level B), which is sufficient for successful reproduction: that is, for spermatogenesis, for the full expression of secondary sex characters and for the complete expression of male reproductive behaviors; and (3) a physiological maximum (level C) that can be reached by social stimulation provided by male–male aggression or by interactions with receptive females (Wingfield et al., 1990, 1997, 2000; Fig. 3). From the measure of these three levels a level of androgen responsiveness can be expressed as the ratio (C  A)/(B  A), which allows androgen responsiveness to social challenges to be compared, independent of individual variation in nonbreeding baseline levels. The use of this ratio also overcomes the potential problem of variation in responsiveness levels that would be present if absolute androgen levels were used. Finally, this ratio may be used not only at the intraspecific level, but also for comparisons of androgen responsiveness between different species. The challenge hypothesis has the added value of providing a conceptual framework for the study of the interplay between social factors and endocrine responses, by generating a number of testable predictions.

Fig. 3. The three levels of endocrine states postulated by the ‘‘challenge hypothesis’’: level A indicates constitutive levels at homeostasis; level B represents the increase to a breeding baseline needed for successful reproduction; level C represents a further increase up to a physiological maximum induced by male–male competition or by interactions with receptive females.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

179

1. First Prediction: Androgen Responsiveness and Mating System Androgen patterns during the breeding season are predicted to vary between species as a function of the number of social interactions to which individuals are exposed. In monogamous species with high levels of paternal care androgen levels should increase above the breeding baseline only when males are challenged by other males or by mating. At other times androgens should remain at the breeding baseline so that they do not interfere with paternal care. Conversely, androgen levels in polygynous males should be near the physiological maximum throughout the breeding season because of high levels of male–male competition in this type of breeding system (Fig. 4). Wingfield et al. (1990) reviewed the available literature on T and aggression in free-living birds and the results support these predictions: male androgen responsiveness was higher in monogamous and polyandrous species, which have monogamous males, than in polygynous species (Wingfield et al., 1990). Later on the number of species included in the analyses was expanded from the initial 20 covered by Wingfield et al. (1990) to 60, and the prediction was reevaluated, reaching the same conclusions (Wingfield et al., 2000). However, the observed relationship between androgen responsiveness and mating system could have been confounded by phylogenetic biases in the database (Harvey and Pagel, 1991), prompting a third analysis of this prediction, expanding further the number of avian species included in the analysis to 84 and using

Fig. 4. The effect of the mating system on androgen responsiveness to social challenges. (A) During the breeding season males of polygynous species (Mp) are expected to have androgen levels closer to the physiological maximum than males from monogamous species (Mm); thus, the androgen responsiveness (
) to social challenges is predicted to be lower in polygynous males than in monogamous males (
m >>
p). Androgen levels: A, constitutive baseline; B, breeding baseline; C, physiological maximum. (B) Predicted relationship between androgen responsiveness and mating system, based on (A).

180

RUI F. OLIVEIRA

comparative methods to control for phylogenetic bias (Hirschenhauser et al., 2003). In general, the same results appear to hold after controlling for phylogenetic relationships among the studied species (Hirschenhauser et al., 2003; see further details in Section VII.B). This prediction of the challenge hypothesis has also been tested in teleosts (Oliveira et al., 2001f, 2002). Using a literature survey, the constitutive baseline, breeding baseline, and physiological maximum androgen levels of male teleosts were compiled for 59 species with different mating systems, and then androgen responsiveness rates were computed for each species (Oliveira et al., 2002). A preliminary analysis of these data, in which the effects of mating system were tested separately from the effects of parenting type (i.e., the male contribution to parental care), revealed an effect of parenting but failed to detect significant differences in male androgen responsiveness among species with different mating systems (Oliveira et al., 2002). However, this result should be taken with caution since no control for potential phylogenetic bias was included. One way to overcome this potential flaw is to restrict the analysis to teleost families for which data are available for more than one species and variation in mating system also occurs. Contrary to the above-described study, a pairwise comparison of male androgen responsiveness between closely related teleost species with different mating systems suggests an effect of the mating system (Oliveira et al., 2001f; Fig. 5). Since using published data increases the probability of potential errors due both to variations in assays between different laboratories and to different blood sampling protocols based on which the A, B, and C circulating androgen levels are calculated, it was decided to experimentally test this prediction of the challenge hypothesis in a group of closely related species with different mating systems, using a standard experimental protocol. The family Cichlidae is a taxon in which a great variety of breeding systems is present (Fryer and Iles, 1972; Barlow, 1991), making it an ideal group for such a comparative study. Two extreme breeding patterns are present in cichlids: (1) monogamous substrate-brooding species, lacking sexual dimorphism, with a prolonged pair bond, and with parental care provided by both members of the pair (most neotropical species; Barlow, 1991); (2) polygamous mouthbrooding species, with strong sexual dimorphism and/or dichromatism, with short courtship episodes and no pair bond (most African species from the Great Lakes; Barlow, 1991; Fryer and Iles, 1972). Mouthbrooding among African cichlids is mainly maternal (or biparental), with paternal mouthbrooding occurring in only two cases (Sarotherodon melanotheron and S. occidentalis) (Trewavas, 1983). Thus, closely related cichlid species that differed in their prevalent mating system were selected to conduct a simulated territorial intrusion experiment to

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

181

Fig. 5. Male androgen responsiveness between closely related teleost species with different mating systems. Cyprinidae, open hexagons: c1, Cyprinus carpio; c2, Tinca tinca. Salmonidae, open triangles up: s1, Salmo trutta; s2, Salmo salar, Salmonidae, open triangles down: sv1, Salvelinus alpinus; sv2, Salvelinus fontinalis. Serranidae, solid hexagons: S1, Serranus subligarius; S2, Epinephelus morio; S3, Lates calcarifer. Cichlidae, solid squares: C1, Sarotherodon melanotheron; C2, Oreochromus mossambicus; C3, O. aureus; C4, O. niloticus. Pomacentridae, solid triangles: P1, Chromis dispilus; P2, Hypsypops rubicundus; P3, Acanthochromis polyacanthus. Data compiled from Oliveira et al. (2002).

assess the responsiveness of the breeding males of each species to behavioral interactions (Hirschenhauser et al., 2004). Two pairs of species of haplochromine cichlids were selected. 1. Neolamprologus pulcher versus Lamprologus callipterus. Neolamprologus pulcher is a monogamous species forming territorial pairs with helpers of both sexes, while L. callipterus is polygynous with small parasitic (sneaker) males; both species are biparental substrate brooders (Balshine et al., 2001; Sato, 1994; Taborsky, 1994; Taborsky and Limberger, 1981) 2. Tropheus moorii versus Pseudosimochromis curvifrons. Both species are polygynous maternal mouthbrooders but in T. moorii temporary pair formation occurs, while P. curvifrons breeds in high-density leks (Kuwamura, 1987; Nishida and Yanagisawa, 1991; Wickler, 1969). A fifth species, the Mozambique tilapia (Oreochromis mossambicus), was also included in this study, although a pair for it was lacking, since it represents an extreme situation for this prediction: a lekking species with exclusive female mouthbrooding (Trewavas, 1983). For each species we

182

RUI F. OLIVEIRA

assessed the constitutive nonbreeding androgen level (A), the male’s androgen response to the presence of and interaction with an ovulating female (level B), and the male’s response to an additional challenge by a conspecific intruder male (level C). Androgens [i.e., 11-ketotestosterone (KT) and T] were measured using a noninvasive method (i.e., fish-holding water; for details of this method see Hirschenhauser et al., 2002b, 2004). In all species sampled the KT circulating levels were highly responsive to territorial intrusions, while the KT responses to interactions with ovulating females were observed only in maternal mouthbrooders and not in biparental species (e.g., Lamprologini). At the interspecific level, the results not only confirmed expectations of higher androgen responsiveness among males from monogamous species, but the response was more pronounced in species with more intense pair bonding (i.e., Tropheus moorii; Fig. 6). Thus, this study confirms the predictions of the challenge hypothesis in cichlid fish at both an intra- and an interspecific level (Hirschenhauser et al., 2004). However, the question remained whether the variation in androgen responsiveness to social stimuli in relation to the mating system is a result of varying breeding baseline levels (B), or a consequence of different maximum physiological levels (C). To disentangle these two possibilities the mean and standard error of levels B and C for teleost species with different mating systems were compared (Fig. 7). The results suggest that variation in androgen responsiveness is related mostly to lower breeding

Fig. 6. Androgen responsiveness in cichlid species with various mating systems. Lines link closely related species. Data points are means of all sampled individuals within each species. Adapted from Hirschenhauser et al. (2004).

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

183

Fig. 7. Breeding baseline (A) and physiological maximum (B) androgen levels (means  SEM) for 59 teleost species collected from the literature. Adapted from Oliveira et al. (2001f).

baseline levels among monogamous and polyandrous male teleosts, rather than to a difference in maximum physiological levels among fish with different mating strategies (Oliveira et al., 2001f) as was originally proposed by Wingfield et al. (1990). In summary, the first prediction of the challenge hypothesis, that androgen responsiveness should be higher in monogamous than in polygynous species due to different regimes of intrasexual selection (i.e., male–male competition), has been confirmed both in teleost and in avian species. On the other hand, androgens may also play a role in shaping mating systems by facilitating social behaviors that promote polygyny. This could be due either to effects on motivational systems that promote sexual appetitive behaviors and a subsequent shift from monogamous to polygynous behaviors, or to an increase in territory size that might then include several females, thus promoting a transition from pair breeding to holding an harem (Beletsky et al., 1995; Ketterson and Nolan 1992; Wingfield et al., 1990). Experimental elevation of T levels of monogamous males resulted in a behavioral shift of these males toward polygyny [e.g., white-crowned sparrow (Zonotrichia leucophrys) and song sparrow (Melospiza melodia): Wingfield, 1984c; European starlings (Sturnus vulgaris): De Ridder et al., 2000; see Beletsky et al., 1995, for a review]. However, this effect is not universal (see Ketterson et al., 1996, and Oliveira et al., 2001f, for a discussion of this topic in birds and fish, respectively). For example, in the St. Peter’s fish (Sarotherodon galilaeus; Cichlidae), a species that shows

184

RUI F. OLIVEIRA

great flexibility in its mating system with facultative monogamous/polygynous males that may or may not contribute to parental care of the offspring (Balshine-Earn, 1996; Fishelson and Hilzerman, 2002), we tested whether males that were more attached to their partners (i.e., with more intense pair bonds) would have lower androgen levels. Paired males were offered access to a novel female out of sight from their female partner and the time spent with each female was measured (Fig. 8). At the end of the behavioral test the male-holding water was collected to assay 11-KT levels. A significant negative correlation between KT levels and partner preference occurred, suggesting that the more polygynous males had higher androgen levels (Oliveira et al., 2001f). In a second experiment, the effects of exogenous administration of T were assessed by giving male St. Peter’s fish intraperitoneal silastic implants of either T or vehicle only. Since in the first experiment a relationship between higher KT levels and a higher propensity to become polygynous had been found, it was predicted that T-treated males would have a lower partner preference than controls. Contrary to expectations, T treatment did not affect partner preference significantly (Oliveira et al., 2001f). The results from these two experiments taken together suggest that the association between partner preference and androgen levels in St. Peter’s fish is not due to a causal effect of androgens on partner preference but, since the androgen levels were measured at the end of the experiment, it is more parsimonious to consider that that they probably reflect variation in partner preference behavior observed among

Fig. 8. Relationship between the intensity of pair-bonding measured as partner preference and the levels of 11-ketotestosterone in male St. Peter’s fish (Sarotherodon galilaeus). Adapted from Oliveira et al. (2001f).

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

185

males. This interpretation is supported by data collected in seminatural conditions in Lake Kinneret, Israel, where KT levels of polygynous males did not differ from those of monogamous males (Ros et al., 2003). This result, together with the fact that T treatment had no effect on male mating strategy, apparently contradicts the effects of androgens on mating systems in wild male birds referred to earlier (e.g., Beletsky et al., 1995). One possible explanation for this difference may be the fact that St. Peter’s fish males are sequentially polygynous, that is, they do not establish harems but desert females after spawning to court other receptive females. In some other species, by contrast, T treatment increases home range/ territory size and as a consequence the number of females present in the territory is larger, thus, promoting simultaneous polygyny (Beletsky et al., 1995). 2. Second Prediction: Androgen Levels and Social Stability Androgen levels should be higher during periods of social instability when social interactions are more frequent and more intense. This prediction is supported by the fact that T levels are higher during the period of territory establishment than when territories are already established in both bird and fish species [see Oliveira et al. (2002) and Wingfield et al., (1999, 2000) for reviews on birds and teleosts, respectively]. For example, the effects of experimental territorial intrusions on androgen levels have been tested in a natural population of the stoplight parrotfish (Sparisoma viride), where it was found that peaks of androgens could be induced in established territorial males by experimental intrusions of other males (Cardwell and Liley, 1991). Also, in the above-mentioned comparative study on simulated territorial intrusions in cichlid fish, for the five species studied the resident always responded to the presence of the intruder male with a significant increase in KT level (Hirschenhauser et al., 2004). Furthermore, in dyadic interactions of the swordtail fish (Xiphophorus helleri) there was an association between some aspects of aggressive behavior and high levels of androgens (Hannes, 1986), whereas when androgen levels of dominant and of subordinate males from a socially stable community tank were compared no relationship between dominance and androgens was found (Hannes, 1984). Taken together, these two latter results are consistent with the suggestion that a causal relationship between androgens and the expression of aggression is present only in periods of social challenge. In conclusion, the available data support the prediction that the association between androgen levels and social status should emerge only during periods of social instability and that in stable social groups androgen levels become dissociated from social status.

186

RUI F. OLIVEIRA

3. Third Prediction: Androgens, Territoriality, and Social Status Territorial and/or dominant males are expected to show higher androgen levels than nonterritorial and/or subordinate males. This is because territorial males must defend their territories from intruders and dominant males must actively keep their status, thus having to engage in a higher proportion of aggressive interactions than either nonterritorial males or subordinates [see Wingfield et al. (1999, 2000) and Oliveira et al. (2002) for reviews on birds and teleosts, respectively]. The causal relationship between androgen levels and territoriality/dominance may be viewed in two ways: (1) androgen levels are predictors of social status; or (2) social status is itself the cause, and not the consequence, of higher androgen levels. To disentangle these two hypotheses, Oliveira et al. (1996) computed correlations between androgen levels and a social dominance index before and after group formation in the cichlid fish O. mossambicus. The rationale behind this experiment was that if androgen levels are the determining factors of social status acquisition, then androgen levels before group formation would be expected to be good predictors of the social status achieved by the individuals after group formation. Conversely, if androgen levels were a response to the acquired social status, it would be predicted that only after group formation would the correlation between androgen levels and social status be present. The latter hypothesis was the one supported by the data, which showed a lack of correlation between androgen levels prior to group formation and the social status achieved, but strong correlations between androgen levels measured after group formation and the acquired social status (Oliveira et al., 1996). However, it should be noted that these two hypotheses are by no means mutually exclusive. For example, a marginal advantage in terms of T levels might help an animal to become dominant, whereupon its T levels would increase further. Therefore, preceding and succeeding hormone levels might be linked in a positive feedback loop (see Oliveira and Almada, 1998b). Thus, the associations found between androgen levels and social status in male teleosts may potentially be explained by the challenge hypothesis, reflecting a more challenging social environment for territorial/ dominant males than for nonterritorial/subordinate ones. A paradoxical confirmation of this prediction comes from a study with another cichlid, Neolamprologus pulcher, with male helpers at the nest (Limberger, 1983; Taborsky, 1984; Taborsky and Limberger, 1981). In this species male helpers are subordinate to breeding males, but they also participate actively in the defense of the family’s breeding territory, and thus are exposed to high rates of territorial challenges. Notably, there is no difference between

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

187

breeding and helper males in androgen levels (Oliveira et al., 2003), which suggests that the differences observed between territorial and nonterritorial males of other species are due to the differential rates of social challenges to which they are exposed and not to differences in social status per se. 4. Fourth Prediction: Androgen Levels and Breeding Density Androgen levels of males breeding at different densities are predicted to be different because of a differential probability in territory intrusions. This prediction should be taken with caution since in a population with increased density, physiological and/or behavioral mechanisms may be present to avoid aggression. Nevertheless, positive correlations have been found between density of breeding territories in different populations and androgen levels both in fish and in birds (e.g., Ball and Wingfield, 1987; Beletsky et al., 1990, 1992; Pankhurst and Barnett, 1993). Moreover, in the peacock blenny (Salaria pavo), a species with alternative reproductive tactics, seasonal variations in sneaker (i.e., potential intruder) density within the same population are also directly associated with variations in androgen levels (Oliveira et al., 2001a). 5. Fifth Prediction: Androgens and Mating Success Males with higher reproductive (or mating) success should also have higher levels of androgens, since they should hold the best territories, being more challenged by other breeding males and visited by a larger number of females. This prediction has been tested in the satin bower bird (Ptilonorhynchus violaceus) (Borgia and Wingfield, 1991) and in red-wing blackbirds (Beletsky et al., 1989). In both species the levels of T during the mate guarding phase were well correlated with reproductive success. Moreover, T levels of male black grouse (Tetrao tetrix), which breed in leks, are positively correlated with mating success, providing further support for this prediction (Alatalo et al., 1996). In addition, T treatment of male dark-eyed juncos (Junco hyemalis) in the wild increased their attractiveness toward their mates (Enstrom et al., 1997), their mate-guarding behavior during the fertile period of their mates (Chandler et al., 1997), and the frequency of extrapair copulations (Raouf et al., 1997). All these effects potentially contribute to increased reproductive success. However, androgens may also have deleterious effects on male reproductive success because of their interference with parental care (see Section VII.A, below). This prediction may not hold in species with alternative mating tactics, in which males with lower competitive abilities sneak fertilizations using a variety of tactics. Male alternative reproductive tactics are especially

188

RUI F. OLIVEIRA

common among teleosts, where an association has been found between the courting morphotype (i.e., bourgeois male) and higher levels of KT (Brantley et al., 1993). However, sneaker males, with lower androgen levels, may have similar or even higher reproductive success than bourgeois males (Fu et al., 2001).

C. Levels of Analysis and the Social Modulation of Androgens: The Tinbergen Legacy Tinbergen (1963) pointed out that there are four fundamental levels of analysis in biology. Thus, when asking questions about behavior four different kinds of questions can be posed. What is the underlying mechanism for that behavior? How did the behavior develop during the ontogeny of the animal? What is the function of the behavior, that is, how does it contribute to the animal’s fitness? And finally, how has the behavior evolved and changed during phylogeny? The answers to these questions are complementary and not mutually exclusive, and necessitate the behavioral pattern being studied at each of the levels of mechanisms, development, function, and evolution. The first two are considered as proximate levels of analysis, because they ask how mechanisms operate within an animal in order to produce the observed behavior, that is, the proximate causes of behavior (Alcock, 1993). In the same way, the other two are considered as ultimate levels of analysis, since they deal with the question of why the animal has evolved over phylogenetic time the mechanisms that produce the observed behavior (Alcock, 1993). The analysis of any behavior is complete only if both proximate and ultimate explanations are achieved using an integrative approach. Therefore, in the following sections, Tinbergen’s ‘‘four why’s’’ will be applied to the study of the social modulation of androgens and behavior.

IV. Proximate Mechanisms for the Social Modulation of Androgens The question of the mechanisms by which the social environment feeds back onto the endocrine system, in order to modulate androgen levels, can be approached at two different levels: (1) what are the neuroendocrine mechanisms involved in the transduction of the relevant social stimuli into an endocrine response (i.e., androgen response)? And (2) what are the psychological (perceptual/cognitive) mechanisms that trigger a response from the neuroendocrine system? These two different levels of analysis are not mutually exclusive but complementary (Fig. 9).

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

189

Fig. 9. Mechanism for the social regulation of androgen levels through the activation of the hypothalamus–pituitary–gonadal axis by key social stimuli that are translated into a relevant neuroendocrine signal by the individual’s perception of the situation.

A. Neuroendocrine Mechanisms Since peripheral production of androgens is under the control of the hypothalamus–pituitary–gonadal (HPG) axis, the most parsimonious explanation for the social modulation of androgens would be social regulation of the activity of the HPG (Fig. 9). In the African cichlid fish Haplochromis burtoni, juveniles raised in the presence of adult territorial males show suppressed gonadal maturation together with smaller gonadotropin-releasing hormone (GnRH) neurons in the preoptic area (Davis and Fernald, 1990). Since this neuronal population projects to the pituitary, through which it regulates the activity of the HPG axis, these results indicate that the social environment during rearing can modulate sexual maturity by acting on the reproductive axis. In adults two male types occur: (1) territorial males, which express the full set of male displaying traits including nuptial coloration, defend a breeding territory, court females, and have mature testes; and (2) nonterritorial males, which lack the expression of male displaying traits, both morphological and behavioral, and have low gonadosomatic values (Fernald, 1977, 1984). A difference between the two male types also occurs at the level of forebrain neurochemistry: territorial males have larger GnRH neurons in the preoptic area than do nonterritorials (Francis et al., 1993). Since soma size is regarded as an

190

RUI F. OLIVEIRA

indicator of cellular activity, this result indicates that territorial males have a more active reproductive axis than subordinates. However, a causal relationship is not clear: the social status of territorial males may be due to a more active HPG that translates into higher androgen levels than would have granted them their dominance; or conversely, perhaps it is their social status that activates the HPG, resulting in higher androgen levels. To disentangle these two alternative explanations an experiment was performed in which the social status of adult males was manipulated in two opposite directions: (1) from territorial to nonterritorial (T ! NT), by moving territorial males into groups with larger territorial males that would dominate them; and (2) from nonterritorial to territorial (NT ! T), by placing NT males into groups consisting of smaller males that they could dominate (Francis et al., 1993). The T ! NT males had larger GnRH neurons in the preoptic area than did nonterritorials, but no differences were found in the size of other GnRH neuron populations, either in the terminal nerve of the telencephalon or in the midbrain. This suggests that changes in social status selectively modulate the activity of the HPG by regulating the activity of preoptic area GnRH neurons in adult individuals (Francis et al., 1993). More recently, it has been found that changes in the size of GnRH neurons are accompanied by changes in GnRH gene expression (White et al., 2002). Thus, in the NT ! T individuals the change in social status was found to activate GnRH expression only in the preoptic area. These studies give support to the hypothesis that the social modulation of sex steroids including androgens is mediated by the HPG, through the regulation of the activity of GnRH neurons. Therefore, the sensory systems processing the stimuli from the social environment to which androgens are responding must connect with the GnRH neurons in the preoptic area (Wilczynski et al., 1993). In frogs conspecific advertisement calls have been shown to influence androgen levels. Circulating concentrations of DHT are higher in males exposed to a conspecific chorus than in males exposed to manipulated (i.e., recording of the same chorus in which frequencies had been shifted to higher values) versions of the chorus (Chu and Wilczynski, 2001). Thus, auditory pathways in this group are good candidates to act as signal transduction mechanisms of relevant social stimuli for the HPG. In fact, both the preoptic area and the ventral hypothalamus receive projections from thalamic and midbrain auditory nuclei (Allison and Wilczynski, 1991; Wilczynski et al., 1993). It is expected that in other taxa, using different sensory modalities, similar circuits will link the relevant sensory system to the areas controlling HPG function and hence androgen responsiveness to social stimuli (e.g., for examples on birds see Ball and Balthazart, 2002).

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

191

In both mammals and birds, expression of immediate-early genes (e.g., c-fos) has been used as a marker of neuronal activation by social stimuli. Since these genes are rapidly transcribed in response to a change in stimulus conditions, they allow the study of the selective activation of different brain regions in response to specific social stimulation. For example, it has been shown both in birds and in mammals that both appetitive and consummatory aspects of sexual behavior induce the expression of these immediate early genes (for a review see Ball and Balthazart, 2001). In the Japanese quail (Coturnix japonica) the exposure of males to receptive females induces c-fos expression in the preoptic region of the hypothalamus and copulation activates c-fos in the ventromedial hypothalamus (Meddle et al., 1997). Also, in rodents, copulation induces c-fos expression in hypothalamic and limbic areas such as the medial amygdala, the bed nucleus of the stria terminalis, the central tegmental field, and the medial preoptic area (e.g., Baum and Everitt, 1992; Coolen et al., 1996; Robertson et al., 1991). However, it is not always clear whether the observed activation is related to perception of the stimuli or to activation of motor systems in response to it (Pfaus and Heeb, 1997). Nevertheless, this approach is a useful tool to unravel the circuits underlying the transduction of social stimuli into a neuroendocrine response. In songbirds another immediate-early gene named zenk has been found to respond to social context in brain areas involving song perception and production, in both canaries (Serinus canaria) and zebra finches (Mello et al., 1992). zenk induction in the auditory forebrain is more pronounced when listening to a conspecific song than to heterospecific songs, and no response is observed after exposure to simple tones (Mello et al., 1992). Moreover, zenk induction decreases with repeated presentations of the same conspecific song while a different conspecific song leads to a still higher induction level, suggesting habituation to the stimulus (Mello et al., 1995). zenk is also induced in several brain structures involved in song perception and production by territorial challenges (i.e., conspecific song playbacks) in breeding male song sparrows in the field (Jarvis et al., 1997). Finally, and most interestingly, in the European starling (Sturnus vulgaris) a socially relevant variation in conspecific song induced significant differences in zenk activation (Gentner et al., 2001). In this species males produce songs of variable length and females exhibit a preference for longer male songs (Gentner and Hulse, 2001). Concordantly, zenk induction in a subregion of the auditory telencephalon (i.e., caudal and medial neostriatum) of females was much higher in females exposed to longer songs than in females exposed to shorter ones (Gentner et al., 2001). Thus, this brain area could potentially play a major role in female mate choice in songbirds.

192

RUI F. OLIVEIRA

In summary, the studies with immediate-early genes in birds and mammals suggest that no induction is observed in peripheral systems that would involve a generalized response to the stimulus. Rather, activation is restricted to sensory integration areas in the telencephalon, as well as to limbic areas underlying motivational systems (e.g., amygdala, hypothalamus, preoptic area) and to premotor systems in the mesencephalic central gray (Ball and Balthazart, 2001). B. Psychological Mechanisms Independently of the neuroendocrine mechanisms that translate a social stimulus into an endocrine signal, another important question to be addressed in relation to social influences on endocrine function concerns the key element in the social interaction that triggers the endocrine response. In other words, what are the psychological (perception/cognition) mechanisms that translate a social stimulus into a signal for the HPG to respond? One of the central axioms of modern social cognition theory is the fact that what influences behavior is not the objective structure of the situation but the subject’s perception of the situation (Smith and Mackie, 1995). This principle has been extended to physiological responses to behavioral contexts. Mason and Brady (1956), by showing a corticosteroid response to a conditioned anxiety paradigm, were the first to propose that the physiological effects of stressors were due mainly to their psychological action and not to their physical characteristics. Later on, Mason (1968a,b) proposed that situations of uncertainty, unpredictability, or novelty were the most important psychological stressors, associated with an increased stress response. One of the first reports suggesting that the subject’s perception of the social situation could influence its androgen levels appeared as an anonymous communication to the journal Nature (Anonymous, 1970). The author, who had been living in isolation on a remote island for professional reasons, noticed a decrease in his beard growth. Interestingly, he also noticed an increase in beard growth in anticipation of his visits to the mainland, where he met with his fiance´e. He concludes that the stimulus for his increased beard growth when visiting the mainland was related to the expectation of sexual activity, which in turn would stimulate androgen production, promoting beard growth. Thus, the mere anticipation of sexual activity would have increased androgen levels in this subject. Although this is not robust evidence that a subject’s perception of the situation can activate an androgen response to that situation (e.g., the anonymous author could have shaved more closely when preparing himself to meet his fiance´e), subsequent studies do support this hypothesis. There is an association

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

193

between sexual activity and androgen levels in human males (Brown et al., 1970; Fox et al., 1972), with circulating T levels increasing during and after copulation when compared with resting conditions. However, no androgen response to masturbation was detected (Fox et al., 1972), which suggests that a stimulus from the female partner is needed to activate the response that hence is not a mere consequence of the ejaculation or of penile stimulation. It has been shown that brief social encounters with potential mating partners are enough to produce significant increases in salivary T levels in human males (Roney et al., 2003). In another study on the long-term relationship between T and social behaviors in men, it was found that prospective fathers (i.e., those who reported a wish to have children with the current partner) have a higher association between T peaks and sexual activity than do unpaired males or those who did not wish to have children with their current partner (Hirschenhauser et al., 2002a). Furthermore, only prospective fathers also displayed 28-day cycles of T, suggesting that men have the facultative potential to adjust their androgen responsiveness to their partner’s menstrual cycle (Hirschenhauser et al., 2002a). The perception of the situation seems to modulate androgen responsiveness not only in sexual but also in competitive contexts. In humans, sports contests can be seen as ritualized competition between participants, and thus offer an opportunity to study short-term endocrine responses to social challenges. As a general rule it was found that T levels increase in anticipation of the confrontation and that after the interaction they remain high in winners and drop in losers (e.g., Booth et al., 1989; Mazur and Booth, 1998). However, this transient response of T to competition was not present when a change in mood was not detected (i.e., when the subjects did not perceive a victory or a defeat, independently of the outcome of the interaction; Mazur and Lamb, 1980; McCaul et al., 1992). It should be added that this endocrine response is present in both physical (e.g., judo, tennis, hockey) and nonphysical contests (e.g., chess; Mazur et al., 1992), and that even individuals who do not participate directly in the contest but who identify themselves with one of the teams (i.e., fans) show the same androgen response and associated change in mood (Bernhardt et al., 1998). Together these results suggest that, at least in humans, social modulation of androgen levels is mediated by the individual’s perception of the social challenge they are facing and of its outcome. In fish there is anecdotal evidence that an individual’s perception of its status also activates its androgen response. In the Mozambique tilapia it was previously demonstrated that androgen levels were not good predictors of social status before a group formation experiment, whereas androgen levels at the end of the experiment were highly correlated with the

194

RUI F. OLIVEIRA

social status of each individual. This suggests that androgens are being modulated by the social interactions experienced by the individuals (Oliveira et al. 1996). Moreover, a significant positive correlation was found between the ratio of KT to T and social dominance expressed as the number of victories over the total number of victories plus defeats (Oliveira and Canario, 2001). Since T is a precursor of KT this result suggests that a rise in social status also promotes the conversion of T into KT, whereas a consequence of subordinance would be the blockage of KT production, possibly through the inhibition of 11-hydroxylase (Oliveira and Canario, 2001). Further support for this hypothesis comes from a study in the Siamese fighting fish (Betta splendens), where it was found that subordinate individuals have lower expression of male sex characters, reduced behavioral displays and the activity of 11-hydroxylase is blocked (Leitz, 1987). Most relevant to our discussion here is the fact that in our experiment all the individuals that were successful in establishing a territory had high KT-to-T ratios, while nonterritorials displayed low ratios. Interestingly, one individual that, despite having a high dominance index (it won 70% of the interactions in which it was involved), failed to establish a territory and had a low KT-to-T ratio that was within the range of subordinate individuals (Fig. 10; Oliveira and Canario, 2001). This suggests that it is the individual’s perception of its status, rather then an objective measure of its dominance behavior, that triggers KT production.

Fig. 10. Effect of holding a territory (T) on the relationship between metabolization of testosterone into 11-ketotestosterone (measured as the ratio of 11-ketotestosterone to testosterone) and social dominance (measured as the number of victories over the number of victories plus defeats). The arrow indicates an individual that, despite having a high dominance index, was not successful in establishing a territory. Data compiled from Oliveira et al. (1996).

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

195

To investigate this idea experimentally, the effect of mirror-elicited aggression on androgen levels was tested in the Mozambique tilapia (L. A. Carneiro and R. F. Oliveira, unpublished data). The mirror image stimulation test (MIS) is widely used to assess fish aggressiveness, because fish do not recognize their own image and so see the image as an intruder and attack it (Rowland, 1999). In O. mossambicus an increase in androgen levels in response to territorial intrusion by a live intruder has been shown (Hirschenhauser et al., 2004). With the MIS test we created a situation in which the perception of the outcome of the interaction (winning versus losing) is not available to the individual, since whatever the behavior expressed by the subject the ‘‘intruder’’ will reply with exactly the same behavior and so there will be no winner or loser. Therefore, if activation of the endocrine response to the social interaction is triggered by the behavioral (motor) output during the interaction (e.g., number of displays or time spent displaying) a variation in androgen levels is predicted. On the other hand, if the androgen response depends on behavioral feedback received from the opponent (i.e., perception of the situation) then an androgen variation is not predicted. In our experiment we found a strong behavioral response in fish toward their own mirror image but no androgen response (L. A. Carneiro and R. F. Oliveira, unpublished data), which suggests that if the endocrine system is to respond to a competitive interaction, that interaction must have a clear outcome. This result is also interesting because it shows that it is the dynamic component of the social interaction that may affect hormone levels, and not static social stimuli from the opponent. However, this result should not be taken as universal since, in songbirds for example, a simulated territorial intrusion using a loop-playback and a decoy is efficient in eliciting both a behavioral and an endocrine response (e.g., Wingfield, 1985). This suggests that activation of the motor circuit per se, that is, the proprioceptive stimuli from the subject’s behavioral performance or some sensory stimuli resulting from its own behavior, is activating the response. In birds the effect of behavioral self-feedback on endocrine responses has been demonstrated in female ring doves, in which female cooing autoinduces gonadal maturation of the female (Cheng, 1986, 1992).

V. Ontogeny of the Social Modulation of Androgens Although the challenge hypothesis was initially proposed to explain social influences on androgen levels in adult males, some studies on the endocrine response to social challenges have been performed in young animals, which in some species also perform significant amounts of

196

RUI F. OLIVEIRA

aggressive behavior. Aggression displayed by young animals may occur between siblings (e.g., bird chicks hatching from the same clutch; Drummond, 2001) or toward nonfamiliar conspecifics in the context of nonsexual territorial defense (e.g., Groothuis, 1989). Contrary to the predictions of the challenge hypothesis, T does not increase in response to sibling competition (Nun˜ez de la Mora et al., 1996; Ramos-Fernandez et al., 2000; Tarlow et al., 2001). T levels were undetectable even after brood starvation in blue-footed booby (Sula nebouxii) chicks in order to increase sibling competition (Nun˜ez de la Mora et al., 1996). Also, in another study with the same species, withinclutch T variation was not associated with chick status (i.e., dominant versus subordinate), and when unrelated chicks were paired no increase in T was detected although there was an increase in aggressiveness (Ramos-Fernandez et al., 2000). Finally, in the Gala´pagos Nazca booby (Sula granti) no correlation was found between T levels and chick status (i.e., dominant versus subordinate versus singleton). However, in this study one reversal in the dominance relationship was observed in a pair of chicks, and in this case the previously subordinate chick that was becoming dominant had a significantly higher T titer (Tarlow et al., 2001). This suggests that androgen responsiveness in young animals may be restricted to specific periods of time when behavioral persistence is needed. This would be a similar mechanism to that described for adult males of year-round territorial birds in the tropics (Wikelski et al., 1999a), and would be consistent with the high costs associated with keeping elevated androgen levels (Wingfield et al., 2001). Alternatively, sibling competition may not be regulated by androgens, at least in an activational fashion, if the developmental advantage of the older sibling (Mock and Parker, 1997) is sufficient to facilitate siblicide without increased androgen-dependent aggression. Interestingly, androgens present in the eggs are known to influence the social rank of juvenile birds, suggesting a maternal effect on the development of aggressive behavior in their offspring (e.g., canary, Serinus canaria; Schwabl, 1993). Among birds, yolk T either increases with laying sequence and so mitigates the effect of hatching asynchrony in younger chicks [redwinged blackbird (Agelaius phoeniceus): Lipar et al., 1999; American kestrel (Falco sparverius): Sockman and Schwabl, 2000; lesser black-backed gull (Larus fuscus): Royle et al., 2000; common tern (Sterna hirundo): French et al., 2001; black-headed gull (Larus ridibundus): Eising et al., 2001], or conversely it decreases with laying order, thus facilitating brood reduction when food conditions are poor [cattle egret (Bubulcus ibis): Schwabl et al., 1997; zebra finch: Gil et al., 1999]. Moreover, between-clutch variation in yolk androgen content is associated with higher T deposition at higher breeding densities [canary: Schwabl, 1996; house sparrow (Passer

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

197

domesticus): Schwabl (1997) and Mazuc et al. (2003); American coot (Fulica americana): Reed and Vleck, 2001; black-headed gull: Groothuis and Schwabl, 2002; European starling: Pilz and Smith, 2004], which may be a result of increased maternal T during the prelaying and laying period due to social interactions [e.g., tree swallow (Tachycineta bicolor): Whittingham and Schwabl, 2002; house sparrow: Mazuc et al., 2003]. Functionally this may be interpreted as a mechanism to prepare the offspring for higher levels of competition in adulthood, should they breed in their natal colonies. These data taken together suggest that the influence of maternal androgens on chick competitive behavior reflects organizational effects rather than activational effects of T on sibling competition (see also Ros et al., 2001). In contrast to the described lack of association between androgens and sibling competition, territorial defense by young animals has been associated with androgens. In the black-headed gull, chicks display very intense aggressive behavior toward unfamiliar conspecifics that is associated with T (Groothuis and Meeuwissen, 1992). Since androgen-dependent aggression has a number of associated costs for juvenile birds, such as growth reduction, decrease in begging behavior, and changes in plumage coloration (Groothuis and Meeuwissen, 1992; Ros, 1999), it is to be expected that androgen levels are under strict influence of social context so that unnecessary increases of circulating androgens are avoided. As predicted, gull chicks respond to short-term social challenges with a transient increase in T levels (Ros et al., 2002). Moreover, the association between androgen levels and breeding density described above, both for adults and for eggs, is also found in black-headed gull chicks: chicks raised in isolated families have lower T levels than ones raised in families kept together in large groups (Ros et al., 2002).

VI. Adaptive Significance of Social Modulation of Androgens As discussed above, social modulation of androgens allows individuals to adjust their agonistic behavior to a variable social environment according to their relative competitive ability. This flexibility has advantages over a fixed androgen level, due to the high costs associated with keeping high androgen levels (see Section VII.A, below). In social species this must be of the utmost importance since one of the key environmental factors to which individuals must respond is the social context. Animals interact with each other frequently and these interactions modulate subsequent interactions among them and with other group members (e.g., in dominance hierarchies, on territories). Thus, animals must fine-tune the expression of their

198

RUI F. OLIVEIRA

social behavior to the social environment in which they live, and androgens may be seen as endocrine mediators of the modulation of social behaviors by social context (Oliveira, 2004). This role of androgens can be played both at the activational level, by modulating the expression of behavior in the short term in response to social context, and at the organizational level, by influencing life-history stage decision-making processes (e.g., when and how to reproduce, depending on the social environment). For example, at the activational level this mechanism would allow subordinate individuals to downregulate the expression of their aggressive behavior, thus avoiding the costs associated with agonistic interactions that they have low probabilities of winning (Oliveira, 2005). Similarly, at the organizational level, individuals of lower competitive ability may adopt alternative conditiondependent tactics, such as breeding as parasitic males (Taborsky, 1994, 2001) or even a change of sex (Grober, 1998; Grober and Bass, 2002), and these transitions may be mediated by androgens (e.g., Oliveira et al., 2001b,d). A. Short-Term (Activational) Effects: Androgens and Social Context Animals from social species live in social networks, raising the possibility that dyadic social interactions can be both observed by and influenced by the presence of conspecifics (McGregor, 1993). This scenario potentially makes more complex the interplay between hormones (androgens) and behavior in all the individuals involved (i.e., the interacting pair plus other conspecifics exposed to the interaction). For example, the presence of a bystander may affect androgen levels in both the interacting individuals and the bystander, which could then affect the bystander’s subsequent social behavior (Oliveira, 2005). Thus, a number of phenomena that have been described in social ethology (e.g., territoriality, bystander effects, audience effects, winner-loser effects, dear enemy effects) may be physiologically mediated by transient changes in androgen levels. 1. Territoriality Effects Territoriality is widespread in vertebrates and its function seems to be in most cases an attempt to monopolize resources, especially food and access to mates (Huntingford and Turner, 1986). In territorial species there is a site-dependent advantage in social conflicts, such that prior residence in an area gives an individual a higher probability of aggressively dominating a conspecific. Individuals attack more readily and with greater intensity when defending their own territory than when they fight away from their home site (e.g., Braddock, 1949; Goulet and Beaugrand, 2004; Henderson

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

199

and Chiszar, 1977; Tinbergen, 1953; Zayan, 1975). Even when the territory owner is smaller than the intruder a site-advantage effect may still be observable in its fighting behavior (Beaugrand et al., 1996). Transient changes in androgen levels may mediate this prior-residence increase in competitive ability. In fact, the establishment of a territory is associated with increased androgen levels and territorial intrusions induce a rise in androgen level in territory owners (e.g., Cardwell and Liley, 1991; Hirschenhauser et al., 2004; Oliveira et al., 1996; Wingfield, 1985). As already mentioned (see Section IV.A), territory status reversal in an African cichlid fish is accompanied by a change in GnRH gene expression and activity in the preoptic area, which may lead to changes in constitutive androgen levels (Francis et al., 1993, White et al., 2002). However, these changes were detected on a time scale not compatible with short-term fluctuations in androgen levels. Endocrine data also became available on the effects of prior-residency reversal situations in human subjects. In humans, sports contests can be seen as a form of ritualized aggression. A home-advantage effect, equivalent to prior residency in animal contests, is well established in team games such as football, ice hockey, rugby, soccer, and basketball (Courneya and Carron, 1992; Neave and Wolfson, 2003; Nevill and Holder, 1999), and there is evidence of increased arousal and aggression at home venues (Kerr and Vanschaik, 1995; McGuire et al., 1992). Concomitantly, salivary T levels in soccer players were higher before home games than before away games (Neave and Wolfson, 2003). 2. Dear Enemy Effects Territorial males react less aggressively toward familiar intruders than to intrusions by strangers, a phenomenon called the ‘‘dear enemy effect’’ (Temeles, 1994; Ydenberg et al., 1988). At a functional level this process allows the resident to adjust its territorial behavior according to the threat posed by the intruder, thus reducing the costs of territorial defence (Temeles, 1994; Leiser and Itzkowitz, 1999; Whiting, 1999). At a proximate level this differential response to familiar versus unfamiliar intruders requires an ability of the resident individual to discriminate between the two intruder types, together with habituation to neighbors. Habituation to conspecific neighbors has been documented in several species, using different sensory modalities such as visual habituation in Siamese fighting fish (Bronstein, 1994) and auditory habituation to neighbor’s calls in frogs (Owen and Perrill, 1998). Since androgens are proposed to play a role as mediators of the dear enemy effect, it is predicted that the resident’s androgen responsiveness to an intrusion should be higher toward a stranger than toward a familiar intruder. We have run a pilot study to test this hypothesis in the Mozambique tilapia, in which intrusions were promoted

200

RUI F. OLIVEIRA

using either territorial neighbors, which were separated from the focal male by a transparent partition, or unfamiliar intruders (kept in separate tanks). For four consecutive days the residents were faced with two intrusions per day, one in the morning and the other in the afternoon. The order of presentation of the two intruder types was randomized. Resident tilapia males exhibited lower latencies to attack stranger intruders than neighbors, and the latency to attack the familiar intruder increased with the day of the experiment, suggesting a habituation process. Moreover, the androgen response to the intrusions showed an effect of intruder type, with strangers eliciting higher KT levels than neighbors; there was also a reduction in the response from the first to the fourth day, indicating habituation of the endocrine response (R. F. Aires, A. F. H. Ros, and R. F. Oliveira, unpublished data; Aires, 2003). Similarly, in humans it has been shown that in adult males from a rural Caribbean village competing at dominoes, T response was higher when playing against strangers from another village than when playing with familiar men from their own village (Wagner et al., 2002). Taken together, these results suggest a mediating role for androgens in the dear enemy effect. 3. Winner–Loser Effects Experiential factors may affect the outcome of social interactions. It is known that, in animal contests, individuals that have won in an interaction increase their probability of winning in a subsequent interaction and vice versa for losers. This effect of prior experience may last from minutes up to hours or even days and has been described for a variety of vertebrate taxa (e.g., fish: Beacham and Newman, 1987; Beaugrand and Zayan, 1985; Beaugrand et al., 1991, 1996; Francis, 1983; Frey and Miller, 1972; reptiles: Schuett, 1997; birds: Drummond and Osorio, 1992; mammals: Ginsburg and Allee, 1942). The winner effect is usually of shorter duration than the loser effect (Chase et al., 1994) and more recent experiences weigh more than former ones in this prior-experience effect (Hsu and Wolf, 1999). It is hypothesized that winner/loser effects may be mediated by shortterm changes in androgens. Ginsburg and Allee (1942) were the first to suggest a potential role for androgens as mediators of experiental effects on social behavior in mice. Data on mammals support a role for T in the loser effect, but not in the winner effect. In mice, seminal vesicle weight, a bioassay for androgen levels, was inversely correlated with the number of defeats (Bronson and Eleftheriou, 1964), and although levels of circulating gonadotropins decreased in both winners and losers after an agonistic interaction, the decrease was more marked and lasted longer in losers, suggesting lowered T secretion (Bronson et al., 1973). Rats that were defeated in a social interaction behaved less aggressively and showed

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

201

lowered levels of T, while no behavioral or endocrine effect was detected in winners (Schuurman, 1980). In rhesus monkeys (Macaca mulatta), defeat induced a long-term reduction in T levels, and transitory but significant elevations of androgens levels were reported after winning (Bernstein et al., 1974, 1983; Rose et al., 1972). In male golden hamsters (Mesocricetus auratus), repeated defeats significantly suppress T levels, and influence subsequent submissive behavior (Huhman et al., 1991). Two studies on male mice have addressed the behavioral significance of androgen responses to social competition in rodents. In both studies the normal androgen response to competition was prevented by castration and subsequent T administration, either by an injection regimen (Nock and Leshner, 1976) or by using Silastic implants (Maruniak et al., 1977). The agonistic behavior displayed by winners and losers over a series of winning and losing experiences was then recorded. No effects of the treatment were found on the subsequent aggressive behavior, suggesting that winner/loser effects were not being influenced by transient fluctuations in circulating androgens induced by the dominance/subordinance experience. Moreover, the profound behavioral changes observed in chronically defeated male rodents (i.e., conditioned defeat), which subsequently fail to defend their territories even against smaller nonaggressive intruders, have not been attributed to changes in T levels but to the hypothalamus– pituitary–adrenal (HPA) axis (Huhman et al., 1990, 1991, 1992, 2003; Leshner, 1983). We tested the hypothesis that androgens could mediate winner/loser effects in fish using male Mozambique tilapia (A. Silva and R. F. Oliveira, unpublished data). After staging a first fight between two males, 2 h later the winner and the loser fought two independent individuals that had not been involved in recent social interactions. As predicted, winners of the first encounter won the majority of the interactions with the naı¨ve fish and vice versa for losers. However, if winners were treated with an antiandrogen (cyproterone acetate) between the two interactions, the winner effect was no longer detectable in the second fight, suggesting an involvement of androgens in the winner effect. Contrary to predictions, the loser effect was not reduced in the second interaction by treating losers with exogenous androgens (i.e., KT), which suggests that, although a drop in androgens is observed in losers, it is not the underlying mechanism for the loser effect. This result is in agreement with the previously mentioned results for conditioned defeat in male rodents and suggests that other neuroendocrine mechanisms must be involved in the loser effect, namely the HPA axis and/ or the serotonergic system. Evidence from studies using different teleost species seems to support the involvement of the serotonergic system in the loser effect. First,

202

RUI F. OLIVEIRA

defeat increases brain levels of serotonin and subordinates have chronically elevated brain levels of serotonin (Winberg and Lepage, 1998; Winberg and Nilsson, 1993a,b; Winberg et al., 1997). Second, serotonin inhibits behavioral responsiveness in general and aggressive behavior in particular (Adams et al., 1996; Edwards and Kravitz, 1997; Winberg and Nilsson, 1993a,b). Thus, losers display a marked behavioral inhibition, with increased attack latencies in subsequent interactions, which prevents them from winning these interactions and reinforces their subordinate role. The evidence presented above suggests that the role of androgens in experiential effects on social interactions is not universal among vertebrates and that different neuroendocrine mechanisms may be involved, as well as cognitive processing of social information (e.g., Oliveira et al., 1998). 4. Bystander Effects The bystander effect consists of a priming effect on the aggressive motivation of spectators of agonistic interactions (Bronstein, 1989; Hogan and Bols, 1980). This priming response seems to be adaptive since it prepares the bystanders for forthcoming interactions in a context of social instability (i.e., where agonistic interactions are already present in the social environment). It was shown that bystanders increase their probability of winning their next social interaction (Clotfelter and Paolino, 2003). The priming of agonistic motivation in bystanders is another social phenomenon that might be mediated by androgens. To test this hypothesis we have conducted an experiment on Mozambique tilapia, in which a bystander male was able to observe two conspecific neighbors through a one-way glass. The neighboring males were separated by an opaque partition. In the experimental treatment, the opaque partition was removed and the bystander could observe an agonistic interaction between its neighbors, but in which it was prevented from participating. In the control treatment, the bystander could see its two neighbors resting or performing maintenance activities, each in its own compartment. Both KT and T increased in the observer in the experimental treatment while no significant variations in androgen levels were found in the control treatment (Oliveira et al., 2001e). As already discussed (see Section IV.B), sports fans also experience variations on their androgen levels depending on the outcome of the game they have attended, with supporters of the winning team experiencing an increase in circulating T whereas a reverse effect is observed in fans of the losing team (Bernhardt et al., 1998). These results suggest a mediating role for androgens in the priming of bystanders.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

203

B. Long-Term (Organizational) Effects: Androgens and Life Histories 1. Life History Trade-Offs and Phenotypic Plasticity Trade-offs are fitness costs that result from the linkage between a beneficial change in one trait and a detrimental change in another (Stearns, 1989). Life histories are shaped by such evolutionary trade-offs, in the form of current versus future reproduction, or the investment in offspring quality versus quantity. Negative genetic correlations and life-history trade-offs can be better understood if their proximate mechanisms are known (Sinervo and Svensson, 1998). Hormones are good candidates to play a major role as physiological mediators of life-history trade-offs, since they may have opposite effects on two or more traits (Ketterson and Nolan, 1992; Sinervo and Svensson, 1998). Phenotypic plasticity is a life-history trait that might have evolved to allow animals to shift resources from one life-history stage to another, for example, from reproduction into growth or vice versa, in a condition- or frequency-dependent fashion (West-Eberhard, 1989). These shifts between life-history stages may also be controlled by endocrine mechanisms, and it is expected that in some cases the life-history tradeoffs and the associated phenotypic plasticity have the same underlying physiological mechanism (Sinervo and Svensson, 1998). Since androgens are both involved in the animal’s investment in current reproduction and have multiple effects on different phenotypic traits, they are excellent candidates to orchestrate transitions between life history stages. In the Mozambique tilapia, the acquisition of dominant status is associated with an increase in androgen levels, which in turn mediate the expression of male display traits, both morphological (i.e., secondary sex characters) and behavioral (e.g., courtship behavior, nest building, nuptial coloration) (Oliveira et al., 1996). Subordinate males have lower androgen levels and display pseudo-female behavior so that they are not ejected from breeding colonies (leks) and obtain access to parasitic fertilizations when spawnings occur (Oliveira and Almada, 1998a). With time, an androgenization of dominant males occurs (i.e., they present exaggerated male displaying traits), which may reinforce their dominance status (Oliveira and Almada, 1998b). Thus, male investment in current reproduction seems to be mediated by androgens that are responding to social status. It should be noted that these two male types are interchangeable, and that individuals may switch between being territorial dominant males and nonterritorial subordinates several times during their life time in response to the social environment (Oliveira and Almada, 1996). Thus, phenotypic plasticity in this species seems to be driven by social stimuli and also to be mediated by androgens.

204

RUI F. OLIVEIRA

Socially modulated androgens can also impose constraints in life-history pathways. An example of such an organizational constraint on lifehistory pathways comes from Mongolian gerbils (Meriones unguiculatus). In this species, male fetuses vary in their intrauterine positions, and this variation is reflected in adult T levels: males gestated between two females (2F males) have lower T levels when adults than their brothers that were gestated between two males (2M males) (Clark et al., 1992b). As might be expected, the development of male sex characters and of sexual behavior is also affected by intrauterine position: 2F males have reduced bulbocavernosus muscle mass and alterations in their copulatory and scent-marking behavior, achieving a lower reproductive success than their 2M siblings (Clark et al., 1990, 1992a). On the other hand, 2F males express more paternal behavior than the 2M males (Clark et al., 1998). Thus, it has been suggested that 2F males that fail to reproduce and that are highly parental, trade direct reproduction by helping at the nest, thus increasing their inclusive fitness (Clark and Galef, 2000). This trade-off between direct reproduction and obligatory alloparenting is mediated by androgen levels modulated by the early (i.e., prenatal) social environment. 2. Sex Change An extreme example of a life-history transition is sex change. Sequential hermaphroditism is present in a number of teleost families, with protogynous sex change [female ! male; e.g., wrasses (Labridae) and parrotfishes (Scaridae)] being more common than protandry [male ! female; e.g., anemone-fish (Pomacentridae)] (Warner, 1984). In most cases sex change is functionally explained by the ‘‘size advantage model,’’ according to which if reproductive success varies with body size between the sexes, then an individual that changes sex at the right size (age) will have a higher lifetime reproductive success than one that remains exclusively male or female for its lifetime (Warner, 1975; but see St. Mary, 1997). At a proximate level of analysis, sex change has been shown to be induced by social factors, for example, either inhibition by the presence of males or social stimulation by other females (Shapiro, 1979). In protogynous species the critical cue to trigger sex change seems to be the perceived status of the fish, as indicated by the ratio of attacks given and received from other group members (Grober, 1998; Reavis and Grober, 1999; Shapiro, 1981). Removal of the male induces a behavioral change in the largest female that dramatically increases her aggressive behavior. This initial behavioral change is completely independent of gonadal steroids since it can be induced in ovariectomized females (Godwin et al., 1996): it is probably needed to inhibit sex change in the other females. Subsequently, however, a cascade of changes in neuroendocrine and morphological traits occurs as

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

205

the dominant female becomes male: the change in status is presumably transduced into a neuroendocrine signal by sensory pathways that project to the hypothalamus affecting GnRH activity, which in turn will affect the gonad anatomy and physiology, the levels of circulating sex steroids, and body coloration (Grober, 1998). A rapid decrease in the activity of brain aromatase (an enzyme that converts androgens into estrogens) was observed at the initial stages of sex change in the sex-changing goby Lythrypnus dalli (Black et al., 2003). This suggests that a change in brain steroidogenesis occurs early in the process and that this modifies the estrogen-to-androgen ratio toward androgens, since a lesser amount of androgens is being metabolized into estrogens. This androgen-biased neural environment would then trigger the cascade of changes described above. Thus, social modulation of brain androgen levels may mediate another life-history transition. 3. Alternative Reproductive Tactics Within the vertebrates, the highest incidence of species with male alternative reproductive tactics occurs in teleost fishes (Taborsky, 1994, 2001). Usually two male types are present: bourgeois males, which actively compete among themselves, investing in the acquisition of mates (e.g., by defending breeding territories); and parasitic males, which exploit the investment of bourgeois males to get access to females and fertilize eggs (e.g., sneakers, satellites; Taborsky, 1994, 2001). Moore (1991) proposed the ‘‘relative plasticity hypothesis’’ as a conceptual framework for the hormonal basis of alternative reproductive tactics. The rationale behind this hypothesis is that the effects of hormones in the differentiation of alternative reproductive tactics are equivalent to their effects in primary sex differentiation (Moore, 1991; Moore et al., 1998). Thus, by making a distinction between fixed alternative phenotypes (in which individuals adopt one of the tactics for their entire life) and flexible alternative phenotypes (in which individuals may switch tactics during their lifetime) this hypothesis predicts an organizational role of hormones in the former case and an activational role in the latter (Moore, 1991). Therefore, hormone profiles should differ in plastic adult morphs but not in fixed ones. In all the fish species with alternative reproductive tactics for which data are available (i.e., Salmonidae: Atlantic salmon, Salmo salar; Centrarchidae: bluegill sunfish, Lepomis macrochirus; Scaridae: stoplight parrotfish, Sparisoma viride; Labridae: saddleback wrasse, Thalassoma duperrey; corkwing wrasse, Symphodus melops; Batrachoididae: plainfin midshipman Porichthys notatus; Cichlidae: Mozambique tilapia; Blenniidae: rock-pool blenny, Parablennius sanguinolentus parvicornis; peacock blenny), bourgeois males have significantly higher KT levels

206

RUI F. OLIVEIRA

than parasitic males, whereas no clear pattern has been found for T (Brantley et al., 1993a; Oliveira et al., 1996, 2001b,c; Uglem et al., 2002). These results suggest either that KT plays a major role in the expression of the male bourgeois tactic, or that KT levels are highly responsive to the expression of the tactic itself (i.e., they are a consequence and not a cause of the expression of alternative mating tactics). However, the prediction concerning relative plasticity is not confirmed by this data set, since differences in KT levels between male types are present in all species irrespective of whether their tactics are fixed or flexible.

VII. Evolution of the Social Modulation of Androgens The evolution of a phenotypic trait requires that the observed phenotypic variation in that trait be partially dependent on genetic variability. Therefore, an implicit assumption for the evolution of a mechanism of social modulation of androgens, as a way of adjusting behavior to social context, is that androgen levels have some heritability. A study performed on human families revealed significant parent–offspring and sibling correlations in circulating T levels and estimated a heritability of approximately 70% (Hong et al., 2001). It should be pointed out, however, that in the same study significant spousal correlations in T levels were also found, suggesting that familial resemblance in androgens should be viewed as a result of both genetic resemblance and common familial environments (Hong et al., 2001). If one considers only the production rate of androgens, the heritability estimate decreases to 40%, with more than 50% of variation in androgen production being explained by environmental variables (Meikle et al., 1988). In nonhuman animals there is evidence for a genetic influence on both circulating androgen levels and androgen production patterns. In a well-known study on canid domestication, that on the effect of genetic selection for tameness in silver foxes (Vulpes vulpes), animals have been successfully selected for reduced aggression and fear toward humans (Belyaev, 1979; Trut, 1999). During the breeding season males in the selected line were found to show both lower plasma concentrations and testicular production of T than their wild counterparts (Osadchuk, 2001). Moreover, selected males showed a more pronounced androgen response to exposure to a receptive female (Osadchuk, 2001). This is in accordance with the first prediction of the challenge hypothesis (see Section II.B) since selected males have lower breeding baseline levels of T and thus a potentially higher androgen responsiveness to social stimuli. In summary, variation in androgen plasma levels and gonadal production seems

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

207

to be in part due to underlying genetic variability, and so is open to selection. For a phenotypic trait to be selected, the benefits it confers must outweigh the costs associated with its expression. Therefore, a cost–benefit analysis should be carried out to try to uncover all the beneficial and detrimental effects associated with the expression of the trait, and why the former outweigh the later. A. Cost–Benefit Analysis of Androgen Responsiveness to the Social Environment Despite the benefits of high androgen levels for the fitness of the individual, androgen levels are not fixed phenotypic traits selected for an optimal value. On the contrary, they represent a physiological state that is tightly regulated by environmental stimuli and by the internal state of the animal. This suggests that there should be costs of keeping high levels of androgens and that social modulation of circulating levels of androgens is a way to minimize these costs (Wingfield et al., 2001). From an adaptationist point of view, androgen-dependent mechanisms can be selected only when the associated benefits outweight the associated costs (Fig. 11).

Fig. 11. The selection of androgen-dependent mechanisms depends on the relationship between potential associated benefits and potential associated costs. For the mechanism to be adaptive the former must outweigh the latter.

208

RUI F. OLIVEIRA

1. Potential Benefits of High Androgen Levels In the preceding sections the potential fitness benefits of high androgen levels have already been discussed, such as an association of T with mating success (see Section III.B) and the effects of androgens on reproduction (i.e., spermatogenesis, expression of male traits, etc.). For example, in the Mozambique tilapia, a lek breeding cichlid fish, female mate choice is based mainly on the size of the spawning pits dug by territorial males (Nelson, 1995), a characteristic that is positively associated with male androgen levels (Oliveira and Almada, 1998b; Oliveira et al., 1996). At periods of social challenge, increased androgen levels may also be beneficial by acting on motivational systems underlying aggressive behavior so that competitive behavior is adjusted to the social environment (see Section II.C) and by acting on other mechanisms that enhance the success of the individual in competitive interactions such as social attention, social learning, memory, and risk taking (Andrew, 1991; Cynx and Nottebohm, 1992). Observing social interactions among conspecifics (eavesdropping) is a way of collecting information on the relative competitive ability of neighbors without paying the costs associated with fighting behavior, which would be an alternative way of collecting such information (McGregor, 1993). Thus, selective attention toward social interactions should be beneficial in territorial species. In the Siamese fighting fish it has been shown that, when given a choice, territorial males spent more time observing social interactions between pairs of conspecifics than observing pairs of conspecifics that were prevented from interacting (Oliveira et al., 1998). Moreover, territorial males of this species eavesdrop on agonistic interactions among conspecific neighbors, gathering information on relative fighting ability that they use in subsequent interactions with the previously observed individuals (Oliveira et al., 1998). The effect of androgens on selective attention to social interactions has been tested in the Siamese fighting fish, when it was found that androgen-treated males significantly increase the time spent observing conspecific interactions when compared with control males (R. F. Oliveira and L. A. Carneiro, unpublished data). This result suggests that androgens may promote selective attention to relevant social stimuli in the environment. Increased risk-taking behavior in a competitive situation may also confer an advantage to the individual. An experiment gives some support to this idea. Male mice were preexposed to the odor of an estrous female and subsequently exposed to the odors of predators. Mice in a situation of increased perception of predation risk, that is, mice that were exposed only to the predator’s odor, showed increased corticosterone and decreased

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

209

circulating levels of T (Kavaliers et al. 2001). Preexposure to female odor attenuated this response to the odor of the predator, which might reflect a greater tendency for risk-taking behaviors modulated by hormones (Kavaliers et al. 2001). 2. Potential Costs of High Androgen Levels A number of potential costs associated with high levels of androgens have been identified (Fig. 11), namely, increased energy consumption, interference with immunocompetence, increased risk of predation, higher incidence of injuries from agonistic interactions, trade-off with parental care, and carcinogenic effects (for reviews see Wingfield et al., 1999, 2001). Some of these costs are discussed here. The evidence on metabolic costs of high androgen levels is contradictory (Table II). The discrepancy in the results may be due to variations in the methods used in the different studies, such as the measures used, the duration of the experiment, and the season when it was performed. Nevertheless, there is evidence for a negative effect of androgens on metabolic rates in species from different taxa (teleosts, reptiles, and birds). Moreover, in the two bird species with discrepant results, explanations are available. In the dark-eyed junco the measure used was daily energy expenditure using doubly labeled water, which incorporates not only resting metabolism but also thermoregulatory and activity costs (Lynn et al., 2000). Furthermore, despite the absence of an effect of T on daily energy expenditure, androgen-treated birds increased their activity and reduced resting and self-maintenance behaviors. Thus, T may be reallocating the relative contributions of the different components of daily energy expenditure. This interpretation is in accordance with the fact that in the white-crowned sparrow T may also increase activity, and thus the observed reduction in the resting metabolic rate in T-treated individuals in this species can be interpreted as a way in which males compensate for increased activity metabolism (Wikelski et al., 1999b). In summary, androgens seem to affect metabolic activity in a nonlinear fashion and there seems to be both direct and indirect metabolic costs of high androgen levels. Another potential cost of high androgen levels is interference with immune function. Although initial studies suggested a negative effect of T on humoral immune function in mammals (Grossman, 1985, 1990), the currently available data are equivocal (see Table II for examples) and several hypotheses concerning the relationship between androgens and immunocompetence have been advanced (Braude et al., 1999; Folstad and Karter, 1992; Hillgarth et al., 1997; Wedekind and Folstad, 1994). The ‘‘immunocompetence handicap hypothesis’’ (Folstad and Karter, 1992) predicts a trade-off between androgens and immunocompetence,

TABLE II Examples of Potential Costs Associated with High Androgen Levels Examples

A. Increased energy expenditure

1. Increased resting metabolic rate and metabolic scope in cichlid fish (O. mossambicus) treated with 11-ketotestosterone 2. Androgen treatment induced higher maximal metabolic rate but had no effect on basal metabolic rate in the lizard Sceloporus jarrovi 3. Testosterone treatment increased the basal metabolic rate in house sparrows (Passer domesticus) 4. Testosterone implantation decreased resting metabolic rate in both intact and castrated white-crowned sparrows (Zonotrichia leucophrys) 5. Testosterone treatment had no effect either on basal metabolic rate or in daily energy expenditure in dark-eyed juncos (Junco hyemalis) 6. Testosterone implants decrease fat stores in song sparrows (Melospiza melodia) 1. Testosterone treatment has a negative effect both on leucocyte numbers and on antibody production in salmonid fish 2. Lymphocyte titers and antibody response to SRBCa are lower in satellite males than in bourgeois males, and negatively correlated with levels of 11-ketotestosterone in the rock-pool blenny (Parablennius parvicornis) 3. No relationship was found between androgen divergent alternative reproductive male morphs and leukocyte count in the corkwing wrasse (Symphodus melops) 4. Testosterone treatment enhanced the antibody response to immunization with SRBC in chicks of the black-headed gull (Larus ridibundus) 5. Higher testosterone levels in a selected line of domestic fowl (Gallus domesticus) selected for antibody responsiveness to sheep red blood cells

210

Type of cost

B. Interference with immune function

Ref. Ros et al. (2004a) Marler et al. (1995) Buchanan et al. (2001) Wikelski et al. (1999b) Deviche (1992); Lynn et al. (2000) Wingfield (1984c) Slater and Schreck (1993, 1997); Slater et al. (1995) A. F. H. Ros, N. Bouton, R. S. Santos, and R. F. Oliveria (unpublished data) Uglem et al. (2001) Ros et al. (1997) Verhulst et al. (1999)

C. Paternal care trade-off

211

6. No relationship between plasma testosterone levels and immunocompetencea in free-living male red-winged blackbirds (Agelaius phoeniceus) 7. Testosterone treatment is immunosuppressive, but in free-ranging males testosterone levels are positively correlated with immunocompetencea in superb fairywrens (Malurus cyaneus) 8. Viremiab decreases early and increases late in the course of infection of testosterone-treated male greenfinches (Carduelis chloris) infected with Sindbis virus, and for total viremia or antibody response there was no difference between treated and control males 1. Androgen levels are lower during the brooding phase than during the mating phase in some male teleost fish; e.g., demoiselles, Chromis dispilus; garibaldi, Hypsypops rubicundus; buegill, Lepomis macrochirus 2. 11-Ketotestosterone levels are lower in parental male midshipman (Porichthys notatus) guarding embryos than in males guarding eggs 3. Mouth-brooding eggs in the paternal mouthbrooder cichlid Sarotherodon melanotheron induces a reduction in circulating levels of testosterone 4. Androgen levels are lower during the brooding phase than during the mating phase in male birds; e.g., song sparrow, Melospiza melodia; white-crowned sparrows, Zonotrichia leucophrys pugetensis, Z.l. gambelli 5. Birth of offspring is associated with a decrease in testosterone levels in male mammals displaying paternal behaviors; e.g., Mongolian gerbils, Meriones unguiculatus; Djungarian hamsters, Phodopus campbelli; common marmosets, Callithrix kuhlii; cotton-top tamarins, Saguinus oedipus; humans, Homo sapiens

Hasselquist et al. (1999) Peters (2000)

Lindstro¨m et al. (2001)

Kindler et al. (1989); Pankhurst and Barnett (1993); Sikkel (1993) Knapp et al. (1999) Specker and Kishida (2000) Wingfield (1984a); Wingfield and Farner (1978a,b) Brown et al. (1995); Nunes et al. (2001); Reburn and Wynne-Edwards (1999); Storey et al. (2000); Ziegler et al. (1996)

a Immunocompetence was measured as secondary antibody production to a nonpathogenic antigen (e.g., sheep red blood cells; keyhole limpet hemocyanin). b Viremia, blood virus concentration.

212

RUI F. OLIVEIRA

and is an extension of the handicap principle proposed by Zahavi (1975) to explain the evolution of costly secondary sex characters. The rationale of this hypothesis is that one cost of honest androgen-dependent signals is that androgens have a immunosuppressive effect. However, this hypothesis is flawed in a Popperian sense as it cannot be disproved, since both positive and negative correlations, and even the lack of a relationship between T levels and arbitrary indices of immunity, all find support from it (Braude et al., 1999). Positive correlations between androgen levels and immunocompetence measures are interpreted as males having high-quality immune function that can resist the immunosuppressive effect of T. Negative correlations are interpreted as indicating that males are of such high quality that they can still display and court females despite being immunosuppressed and exposed to higher parasite loads. Finally, if no correlation is found it is argued that high-quality males reliably signal their resistance to parasites since they manage to resist infection even with high T levels (Braude et al., 1999). Subsequently Wedekind and Folstad (1994) have proposed that androgen immunosupression would allow the individual to reallocate resources from the immune system to sexual display characters, both morphological (e.g., horns) and behavioral (e.g., courtship song). This explanation supports the assumption of the immunocompetence handicap hypothesis that there is a trade-off between androgens and immunity. The weakness of this hypothesis is that the risk of infection due to immunosupression most probably outweighs the metabolic resources saved by it (Hillgarth and Wingfield, 1997). An alternative explanation for the association between T, leukocyte titers, and male display traits has been advanced by Braude et al. (1999) and is termed the ‘‘immunoredistribution hypothesis.’’ According to this hypothesis, androgens may have a similar effect to corticosteroids in promoting a redistribution of leukocytes to different compartments of the immune system, and this may suggest an immunosuppressive effect if only one measure of immune function is taken at a time (Braude et al., 1999). Thus, the main distinction between this hypothesis and that of Wedekind and Folstad is that the immunocompetence handicap hypothesis implies a reduction of immune function associated with T while the immunoredistribution hypothesis merely involves a reversible relocation of immune cells to tissues where they are temporarily needed. Moreover, it has also been shown that T may have both positive and negative effects on immune function and that the direction of the effect may depend on when it is measured (Lindstro¨m et al., 2001; Marsh, 1992; Olsen and Kovacs, 1996). Finally, suppression of parental care has also been proposed as a potential cost of high T levels in species in which males provide parental care

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

213

(Wingfield et al., 1990). If androgen levels increase due to social challenges, males will invest less time in paternal activities, and thus a trade-off will occur between social interactions and paternal care mediated by androgens (Wingfield et al., 1990). The reduction in time available for paternal care may emerge for two reasons: (1) because T results in more time spent on other behaviors (e.g., territorial defense), or (2) because T directly suppresses parental care. Thus, a decrease in androgen levels during parenting has been interpreted as a way of avoiding infanticide, reducing male distraction by courtship toward other females, and facilitating affiliative behaviors toward the young (Clark and Galef, 1999, 2000; Gubernick et al., 1994). Several studies on the temporal variation of androgen levels in fish, bird, and mammal species with male parental care show that during the breeding season male androgen levels are higher during the mating phase than during the parental phase (for reviews, see Wingfield et al., 1987, on birds; Ziegler, 2000, on mammals; and Oliveira et al., 2002, on fish; selected examples are given in Table II). Among birds the trade-off between parental care and androgens is larger during incubation than during later brooding phases (Hirschenhauser et al., 2003). In many bird species an experimental increase in circulating T in parental males suppresses paternal behavior and promotes agonistic interactions (e.g., Beletsky et al., 1995; Hegner and Wingfield, 1987; Ketterson et al., 1992; Silverin, 1980). However, the use of castration to reduce T levels has yielded inconsistent results in mammals. Castration increases paternal behaviors in Mongolian gerbils (Clark and Galef, 1999), has produced mixed results in prairie voles (Microtus ochrogaster) (Lonstein and De Vries, 1999; Wang and De Vries, 1993), and in the biparental California mouse (Peromyscus californicus) it decreases parental care while androgen replacement restores it (Trainor and Marler, 2001). These results are in accordance with the fact that, in some species, males keep high T levels after the birth of offspring despite the fact that they still display high levels of paternal care [e.g., Djungarian hamsters (Phodopus campbelli): Jones and Wynne-Edwards, 2000; cottontop tamarins (Saguinus oedipus): Ziegler and Snowdon, 2000]. This apparent contradiction may result from the fact that T action on parental behavior is mediated by its conversion to E2, which is likely to be the active steroid promoting paternal behaviors, as was shown in the California mouse (Trainor and Marler, 2002). In fish species in which the mating and the parental phase overlap, a trade-off between androgen levels and paternal care has not been found. For example, in a facultative biparental cichlid, the St. Peter’s fish (Sarotherodon galilaeus), there were no differences in androgen levels between brooding and nonbrooding males (Ros et al., 2003). Also in the

214

RUI F. OLIVEIRA

rock-pool blenny, a promiscuous species with exclusive male parental care, although an incompatibility between defense of a breeding territory and high levels of parental behavior exists, KT implants failed to suppress paternal behavior (Ros et al., 2004b). In this case the trade-off between androgen-dependent behaviors (territorial defense/mating) and parental behaviors may not be regulated by androgen hormones but may instead result from a time constraint in the individual’s activity budget (Ros et al., 2004b). Interestingly, there is an inverse pattern on the variation of prolactin levels with androgen levels during periods of male parental care (for a review see Ziegler, 2000), and prolactin has been found to be involved in the expression of paternal care in a wide range of vertebrates, from fish to mammals (for a review, see Schradin and Anzenberger, 1999). It is also known that prolactin is involved in the regulation of T secretion (e.g., Huang et al., 1999), and that in humans there is an orgasm-induced prolactin peak, which inhibits sexual arousal following an orgasm (Kru¨ger et al., 2002). Moreover, multiorgasmic males, lacking a refractory period after ejaculation, also lack an orgasm-induced prolactin peak (Haake et al., 2002). Thus, prolactin seems to have a negative effect on androgen production and on the expression of androgen-dependent sexual behaviors. Therefore, the pattern of a T decrease during the parental phase may be due to a physiological constraint imposed by a rise in prolactin associated with the expression of paternal behaviors. In summary, there are a number of potential costs, associated with keeping high levels of androgens for long periods of time, that may be avoided by the social modulation of androgens. B. Evolutionary Scenarios for the Selection of Androgen Responsiveness to the Social Environment Since hormones act on different target tissues, many traits may have a common underlying physiological mechanism and thus be phenotypically linked (Ketterson and Nolan, 1999; Ketterson et al., 2001), as is the case with androgen-dependent traits such as the expression of male morphological characters, muscle hypertrophy, and the expression of aggressive and sexual behavior. Thus, it is likely that selection acting on any one of these traits will affect the others, so that beneficial traits may evolve indirectly as exaptations (in the sense of Gould and Vrba, 1982). According to Ketterson and Nolan (1999) one way to distinguish between adaptations and exaptations in hormone-dependent traits would be to assess whether these traits arose either in response to selection on circulating hormone levels (in which case selection probably did not act on all correlated traits and thus

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

215

the ones that subsequently conferred an advantage to its carriers are exaptations), or whether they arose in response to variation in the responsiveness of the target tissues to invariant hormone levels (in which case selection probably acted independently on target tissue sensitivity to constant hormone levels). Although this dichotomy is simplistic (i.e., a mixed scenario may occur in which both circulating levels and target tissue sensitivity are under selection) it provides us with a framework to address the issue of endocrine-mediated adaptive traits. In this respect, androgen responsiveness to the social environment may be seen as a physiological mechanism that allows the organism to avoid the costs of high androgen levels at times when they are not crucial, thus suggesting that it might have evolved as an exaptation in the context of fixed target tissue sensitivity to circulating androgens. Independently of the question of whether androgen responsiveness to the social environment emerged as an adaptation or as an exaptation it is important to try to identify the selective pressures on this mechanism. Since a large data set on androgen levels in wild avian populations is available, together with an established phylogeny, which allows for the control of phylogenetic bias in the comparative analysis, the approaches to this question have been restricted to birds. We tested the effects of the mating system (specifically, the intensity of social challenges that the males are exposed to and the investment in paternal care) on the evolution of androgen responsiveness to the social environment (Hirschenhauser et al., 2003; Fig. 12). When controlling for phylogenetic relatedness among the sampled species, an effect of the degree of male–male competition on

Fig. 12. The evolution of androgen responsiveness to social challenges in male birds. Scatterplots of how androgen responsiveness varies with the degree of male–male aggression (right) and the degree of male investment in parental care (left); the lines represent a linear regression with phylogenetic distances.

216

RUI F. OLIVEIRA

androgen responsiveness was detected, whereas the degree of male investment in paternal care had no significant effect (Hirschenhauser et al., 2003; Fig. 12). To further investigate this lack of effect of paternal care on shaping androgen responsiveness, the analysis was repeated on a subset of altricial passerine species, using a quantitative scale of the degree of male incubation and feeding of offspring. This reanalysis showed an effect of the male contribution to incubation but no effect of the degree of offspring feeding (Hirschenhauser et al., 2003). Thus, the degree of male–male competition, rather than changes in parental investment, emerged as the most relevant factor for an evolutionary change in androgen responsiveness in birds. Further comparative studies will allow us to test which of these factors shaped the androgen response to social stimuli in other vertebrate taxa.

VIII. Social Modulation of Androgens in Men During the evolution of the primate brain, different regions have developed differentially. Over primate phylogeny, the neocortex and the striatum, which correspond to the ‘‘executive’’ brain, have increased considerably in relative size, at the cost of a reduction in the relative size of the ‘‘emotional’’ brain, that is, the hypothalamus and the septum (Keverne et al., 1996). Since the ‘‘emotional’’ brain is the major neural target tissue for sex steroids, this differential development of different areas of the primate forebrain has progressively emancipated primate sexual and social behaviors from the influences of gonadal hormones over phylogenetic time (Keverne et al., 1996). A major selective pressure for this trade-off between the relative development of the ‘‘executive’’ versus the ‘‘emotional’’ brain may have stemmed from strong social constraints imposed on reproductive decision-making in large-brained anthropoids, which requires the evolution of more flexible behavioral strategies (Keverne et al., 1996). Emancipation of sexual behavior from hormonal regulation is confirmed in behavioral studies of sexual behavior in humans and other primates (Wallen, 2001). For example, eliminating T in adult male primates, including humans, decreases sexual motivation but does not eliminate sexual behavior. Moreover, the effects of androgens in nonhuman primates vary with social context, and cognition is shown to play a major role in primate sexual behavior (Wallen, 1999, 2001). Thus, androgen responsiveness to social stimuli, if present in humans, seems to have reduced or no behavioral significance. Nevertheless, a consistent body of literature on social modulation of androgen levels in humans, already presented in other parts of

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

217

this review (see Sections IV.B and V.A for numerous examples; and Mazur and Booth, 1998, for a general review), suggests that it is a common phenomenon in our species. Adaptive androgen responsiveness to the social environment has also been described for other primates. For example, increases in male circulating T levels correlated with male–male aggression have been documented in ringtailed lemurs (Lemur catta; Cavigelli and Pereira, 2000), in rhesus monkeys (Macaca mulatta; Higley et al., 1996; Mehlman et al., 1997), and more notably also in free-living common chimpanzees (Pan troglodytes; Muller and Wrangham, 2004). In humans, the androgen response to success in a social challenge has been interpreted as an adaptive mechanism that reinforces the instrumental behavior(s) associated with winning the interaction (Mazur, 1985). The finding that T administration has rewarding and mood-enhancing effects is in agreement with this hypothesis (Packard et al., 1998; Rabkin et al., 1996). However, the evidence for winning/losing effects on T in humans is equivocal, with some studies failing to detect such a pattern (e.g., Gonzalez-Bono et al., 1999; Salvador et al., 1987; Suay et al., 1999; R. F. Oliveira, M. J. Gouveia, P. Almeida, and T. Oliveira, unpublished data on soccer players). It has been proposed that individual variations in motivational dispositions would explain these contrasting results (Schultheiss et al., 1999). According to the theory of implicit motives of human motivation, individuals may have unconscious enduring preferences for power or affiliation (McClelland, 1987). These implicit motives may influence behavior if activated by relevant situational cues. Individuals high in implicit power motivation, which is an indicator of the subject’s need for dominance or status, responded with a T increase to winning a dominance contest, while individuals with a low implicit power motive did not show an androgen response to either winning or losing the contest (Schultheiss et al., 1999). Furthermore, postvictory increases in T facilitated implicit learning of the instrumental behaviors associated with winning the contest, thus having a reinforcing role in winners with a high power motivation (Schultheiss and Rhode, 2002). Thus, androgen responsiveness to social challenges seems to be an adaptive trait in humans in individuals with a particular social motivation style. Androgen responsiveness per se may represent an ancestral state in our evolutionary history since it is widespread in other vertebrate taxa, and it may have proved beneficial in the environment of our hominid ancestors, as it would have reinforced behaviors leading to dominance status, which is associated with reproductive success in most animal species (Ellis, 1995). However, it may have become maladaptive in the settings of modern life. As mentioned above, an increase in T levels is seen in fans of sports teams when their team wins a game (Bernhardt et al., 1998). An independent

218

RUI F. OLIVEIRA

study on the relationship between the outcome of American football games and the frequency of violent assaults on women has found a positive association between the victory of the favorite team and the frequency of admissions of women victims of violent assaults in the local hospital (White et al., 1993). Taken together these studies suggest that experiencing a T surge in response to perceived dominance success in men with high power motivation might act as a trigger for violent behavior. In conclusion, social modulation of androgens is clearly also present in humans although its current adaptive value is questionable.

IX. Summary The social modulation of androgen levels seems to be a widespread phenomenon in vertebrates. It allows the individual to adjust its behavioral output to the context-dependent condition imposed by social constraints. The perception that the individual has of its social environment is needed to activate the androgen response. Animals can use different sensory channels to perceive the key social signals that trigger the response. The transduction of these stimuli into a neuroendocrine response involves the efferent projections from sensory areas of the central nervous system to the preoptic area/hypothalamus that controls the androgen response through the HPG axis. The development of this response during ontogeny can be primed by early exposure to androgens through maternal effects, and the androgen response to social challenges can be present early in the ontogeny of altricial species. The adaptive value (i.e., function) of having circulating androgen levels open to the influence of the social environment is to allow the individuals to adjust their competitive behavior to the social context according to their relative competitive ability, and this flexibility has advantages over an optimal fixed value of androgen levels, because of the high costs associated with high androgen levels (e.g., mismatched expression of behavior, trade-off with parental care; metabolic costs; immunosupression; survival; etc.). A phylogenetic analysis in birds suggests that the evolutionary scenario in which the social modulation of androgens has evolved was characterized for selective pressures imposed by male–male competition regimes that vary with mating system. In future the social modulation of hormones as an adaptive mechanism to adjust female behavior to social context should also be investigated. So far investigations using the conceptual framework provided by the challenge hypothesis have been mainly limited to androgen responses in males. In females, hormones other than androgens may be relevant in such a role, as is suggested by the less clear relationship between T and agonistic

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

219

behavior in females (e.g., Floody, 1983) and by the known effects of E2 on female aggression (e.g., Toda et al., 2001). A rise in progesterone has been described in female California mice as a response to a social challenge in a resident–intruder paradigm, while no changes were observed in T levels (Davis and Marler, 2003). These results suggest that in females different hormonal mechanisms may mediate behavioral responses to social challenges. The study of endocrine responses in females to social challenges is therefore a promising avenue for future research in this area. Acknowledgments I thank the following people with whom I have discussed over the years some of the ideas presented in this chapter: Katharina Hirschenhauser, Albert Ros, Luis Carneiro, David Goncalves, Adelino Cana´rio, Vitor Almada, and John Wingfield. Tim Roper, Peter Slater, and Anahita Kazem provided very helpful comments on an early version of this manuscript. Thanks are also due to the staff of ISPA’s library for help in finding the most obscure primary sources that I have requested, and to Teresa Garcia Marques for input on social psychology sources. Unpublished data reported in this review are part of an ongoing research project funded by the Portuguese Foundation for Science and Technology (FCT; grant ref. POCTI/BSE/38484/01). The laboratory of R.F.O. is funded by the FCT Pluriannual Program (Research Unit no. 331/94). I would like to express my gratitude to my lifetime partner Alexandra Lopes for her unconditional support during the writing of this chapter and to my two young children, Joa˜o and Catarina, who are now naively convinced that work is daddy’s favorite hobby.

References Adams, C. F., Liley, N. R., and Gorzalka, B. B. (1996). PCPA increases aggression in male firemouth cichlids. Pharmacology 53, 328–330. Aires, R. F. (2003). Androge´nios como mediadores endo´crinos no efeito de intruso˜es territoriais no ciclı´deo Africano Oreochromis mossambicus (Peters 1852): O efeito do querido inimigo. Honor’s thesis. Institute of Biomedical Science Abel Salazar, University of Porto, Porto, Portugal. Alatalo, R. V., Ho¨glund, J., Lundberg, A., Rintama¨ki, P. T., and Silverin, B. (1996). Testosterone and male mating success on the black grouse leks. Proc. R. Soc. Lond. B Biol. Sci. 263, 1697–1702. Albert, D. J., Jonik, R. H., and Walsh, M. L. (1993). Influence of combined oestradiol and testosterone implants on the aggressiveness and nonaggressive female rats. Physiol. Behav. 53, 709–713. Alcock, J. (1993). ‘‘Animal Behavior,’’ 5th ed. Sinauer Associates, Sunderland, MA. Allison, J. D., and Wilczynski, W. (1991). Thalamic and midbrain auditory projections to the preoptic area and ventral hypothalamus in the green treefrog (Hyla cinerea). Brain Behav. Evol. 38, 322–331. Anonymous. (1970). Effects of sexual activity on beard growth in man. Nature 226, 869–870. Arnold, A. P., and Breedlove, S. M. (1985). Organizational and activational effects of sex steroids on brain and behavior: A reanalysis. Horm. Behav. 19, 469–498.

220

RUI F. OLIVEIRA

Baldwin, F. M., and Goldin, H. S. (1939). Effects of testosterone propionate on the female viviparous teleost, Xiphophorus helleri Heckel. Proc. Soc. Exp. Biol. Med. 42, 813–819. Ball, G. F., and Balthazart, J. (2001). Ethological concepts revisited: Immediate early gene induction in response to sexual stimuli in birds. Brain Behav. Evol. 57, 252–270. Ball, G. F., and Balthazart, J. (2002). Neuroendocrine mechanisms regulating reproductive cycles and reproductive behavior in birds. In ‘‘Hormones, Brain, and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach, and R. T. Rubin, Eds.), pp. 649–798. Academic Press, New York. Ball, G. F., and Wingfield, J. C. (1987). Changes in plasma luteinizing hormone and sex steroid hormones in relation to multiple-broodness and nest-site density in male starlings. Physiol. Zool. 602, 191–199. Balshine-Earn, S. (1996). Reproductive rates, operational sex ratios and mate choice in St. Peter’s fish. Behav. Ecol. Sociobiol. 39, 107–116. Balshine, S., Leach, B., Neat, F., Reid, H., Taborsky, M., and Werner, N. (2001). Correlates of group size in a cooperatively breeding cichlid fish (Neolamprologus pulcher). Behav. Ecol. Sociobiol. 50, 134–140. Balthazart, J., Reid, J., Absil, P., Foidart, A., and Ball, G. F. (1995). Appetitive as well as consummatory aspects of male sexual behavior in quail are activated by androgens and estrogens. Behav. Neurosci. 109, 485–501. Barfield, R. J. (1969). Activation of copulatory behavior by androgen implanted into the preoptic area of the male fowl. Horm. Behav. 1, 37–52. Barfield, R. J. (1971). Activation of sexual and aggressive behavior by androgen implantation into the male ring dove brain. Endocrinology 89, 1470–1476. Barlow, G. W. (1991). Mating systems among cichlid fishes. In ‘‘Cichlid Fishes: Behavior, Ecology and Evolution’’ (M. H. A. Keenleyside, Ed.), pp. 173–190. Chapman & Hall, London. Baum, M. J. (2002). Neuroendocrinology of sexual behavior in the male. In ‘‘Behavioral Endocrinology’’ (J. B. Becker, S. M. Breedlove, D. Crews, and M. McCarthy, Eds.), 2nd ed., pp. 153–203. MIT Press, Cambridge, MA. Baum, M. J., and Everitt, B. J. (1992). Increased expression of c-fos in the medial preoptic area after mating in male rats: Role of afferent inputs from the medial amygdala and midbrain central tegmental field. Neuroscience 50, 627–646. Beach, F. A. (1948). ‘‘Hormones and Behavior.’’ Paul B. Hoeber, New York. Beach, F. A. (1974). Behavioral endocrinology and the study of reproduction. Biol. Reprod. 10, 2–18. Beach, F. A., and Levinson, G. (1950). Effects of androgen on the glans penis and mating behavior of castrated male rats. J. Exp. Zool. 114, 159–168. Beacham, J. L., and Newman, J. A. (1987). Social experience and the formation of dominance relationships in the pumpkinseed sunfish, Lepomis gibbosus. Anim. Behav. 35, 1560–1562. Beatty, W. W. (1979). Gonadal hormones and sex differences in nonreproductive behaviors in rodents: Organizational and activational influences. Horm. Behav. 12, 112–163. Beaugrand, J. P., and Zayan, R. (1985). An experimental model of aggressive dominance in Xiphophorus helleri (Pisces, Poeciliidae). Behav. Proc. 10, 1–52. Beaugrand, J. P., Goulet, C., and Payette, D. (1991). Outcome of dyadic conflict in male green swordtail fish, Xiphophorus helleri: Effects of body size and prior dominance. Anim. Behav. 41, 417–424.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

221

Beaugrand, J. P., Payette, D., and Goulet, C. (1996). Conflict outcome in male green swordtail fish dyads (Xiphophorus helleri): Interaction of body size, prior dominance/subordination experience, and prior residency. Behavior 133, 1–17. Becker, J. B. (2002). Hormonal influences on sensorimotor function. In ‘‘Behavioral Endocrinology’’ (J. B. Becker, S. M. Breedlove, D. Crews, and M. McCarthy, Eds.), 2nd ed., pp. 497–525. MIT Press, Cambridge, MA. Beletsky, L. D., Wingfield, J. C., and Orians, G. H. (1989). Relationships of hormones and polygyny to territorial status, breeding experience and reproductive success in male red winged blackbirds. Auk 106, 107–117. Beletsky, L. D., Orians, G. H., and Wingfield, J. C. (1990). Steroid hormones in relation to territoriality, breeding density, and parental behavior in male yellow-headed blackbirds. Auk 107, 60–68. Beletsky, L. D., Orians, G. H., and Wingfield, J. C. (1992). Year-to-year patterns of circulating levels of testosterone and corticosterone in relation to breeding density, experience, and reproductive success of the polygynous red-winged blackbird. Horm. Behav. 26, 420–432. Beletsky, L. D., Gori, D. F., Freeman, S., and Wingfield, J. C. (1995). Testosterone and polygyny in birds. In ‘‘Current Ornithology’’ (D. M. Power, Ed.), Vol. 10, pp. 1–41. Plenum Press, New York. Belyaev, D. K. (1979). Destabilizing selection as a factor in domestication. J. Hered. 70, 301–308. Bermant, G. (1976). Sexual behavior: Hard times with the Coolidge effect. In ‘‘Psychological Research: The Inside Story’’ (M. H. Siegel and H. P. Zeigler, Eds.), pp. 76–103. Harper & Row, New York. Bernhardt, P. C., Dabbs, J. M., Fielden, J. A., and Lutter, C. D. (1998). Testosterone changes during vicarious experiences of winning and losing among fans at sports events. Physiol. Behav. 65, 59–62. Bernstein, I. S., Rose, R. M., and Gordon, T. P. (1974). Behavioral and environmental events influencing primate testosterone levels. J. Hum. Evol. 3, 517–525. Bernstein, I. S., Gordon, T. P., and Rose, R. M. (1983). The interaction of hormones, behavior, and social context in nonhuman primates. In ‘‘Hormones and Aggressive Behavior’’ (B. B. Svare, Ed.), pp. 535–561. Plenum Press, New York. Black, M. P., Baillien, M., Balthazart, J., and Grober, M. S. (2003). Rapid decreases in brain aromatase activity at the onset of vertebrate sex change. Program 2004. ‘‘2003 Abstract Viewer/Itinerary Planner’’. Society for Neuroscience 33rd Annual Meeting. Society for Neurosciences, Washington, D.C. Online at http://sfn.scholarone.com/itin2003/ index.html. Accessed August 2004. Bleisch, W., Luine, V. N., and Nottebohm, F. (1984). Modification of synapses in androgensensitive muscle. I. Hormonal regulation of acetylcholine receptor number in the songbird syrinx. J. Neurosci. 4, 786–792. Booth, A., Shelley, G., Mazur, A., Tharp, G., and Kittok, R. (1989). Testosterone and winning and losing in human competition. Horm. Behav. 23, 556–571. Borgia, G., and Wingfield, J. C. (1991). Hormonal correlates of bower decoration and sexual display in the satin bowerbird (Ptilonorhynchus violaceus). Condor 93, 935–942. Braddock, J. C. (1949). The effect of prior residence upon the dominance in the fish Platypoecilus maculatus. Physiol. Zool. 22, 161–169. Brantley, R. K., Wingfield, J. C., and Bass, A. H. (1993). Sex steroid levels in Porichthys notatus, a fish with alternative reproductive tactics, and a review of the hormonal bases for male dimorphism among teleost fishes. Horm. Behav. 27, 332–347. Braude, S., Tang-Martinez, Z., and Taylor, G. T. (1999). Stress, testosterone, and the immunoredistribution hypothesis. Behav. Ecol. 10, 345–350.

222

RUI F. OLIVEIRA

Breedlove, S. M., and Arnold, A. P. (1981). Sexually dimorphic motor nucleus in the rat lumbar spinal cord: Response to adult hormone manipulation, absence in androgeninsensitive rats. Brain Res. 225, 297–307. Breedlove, S. M., and Arnold, A. P. (1983). Hormonal control of a developing neuromuscular system. I. Complete demasculinization of the spinal nucleus of the bulbocavernosus in male rats using the anti-androgen flutamide. J. Neurosci. 3, 417–423. Bronson, F. H., and Eleftheriou, B. E. (1964). Chronic physiological effects of fighting in mice. Gen. Comp. Endocrinol. 4, 9–14. Bronson, F. H., and Marsden, H. M. (1973). The preputial gland as an indicator of social dominance in male mice. Behav. Biol. 9, 625–628. Bronson, F. H., Stetson, M. H., and Stiff, M. E. (1973). Serum FSH and LH in male mice following aggressive and nonaggressive interactions. Physiol. Behav. 10, 369–372. Bronstein, P. M. (1989). The priming and retention of agonistic motivation in male Siamese fighting fish, Betta splendens. Anim. Behav. 37, 165–166. Bronstein, P. M. (1994). On the predictability, sensitisation, and habituation of aggression in male bettas (Betta splendens). J. Comp. Psychol. 98, 421–431. Brown, R. E., Murdoch, T., Murphy, P. R., and Moger, W. H. (1995). Hormonal responses of male gerbils to stimuli from their mate and pups. Horm. Behav. 29, 474–491. Brown, W. A., Monti, P. M., and Corriveau, D. P. (1970). Serum testosterone and sexual activity and interest in men. Arch. Sex. Behav. 7, 97–103. Bruce, H. M. (1959). An exteroceptive block to pregnancy in the mouse. Nature 184, 105. Buchanan, K. L., Evans, M. R., Goldsmith, A. R., Bryant, D. M., and Rowe, L. V. (2001). Testosterone influences basal metabolic rate in male house sparrows: A new cost of dominance signalling? Proc. R. Soc. Lond. B Biol. Sci. 268, 1337–1344. Cacioppo, J. T., and Berntson, G. G. (2002). Social neuroscience. In ‘‘Foundations in Social Neuroscience’’ (J. T. Cacioppo, G. G. Berntson, R. Adolphs, C. Sue Carter, R. J. Davidson, M. K. McClintock, B. S. McEwen, M. J. Meaney, D. L. Schacter, E. M. Strenberg, S. S. Suomi, and S. E. Taylor, Eds.), pp. 3–10. MIT Press, Cambridge, MA. Cacioppo, J. T., Berntson, G. G., Sheridan, J. F., and McClintock, M. K. (2000). Multilevel integrative analyses of human behavior: Social neuroscience and the complementing nature of social and biological approaches. Psychol. Bull. 126, 829–843. Callard, G. V., Drygas, M., and Gelinas, D. (1993). Molecular and cellular physiology of aromatase in the brain and retina. J. Steroid Biochem. Mol. Biol. 44, 541–547. Cannon, W. B. (1929). ‘‘Bodily Changes in Pain, Fear, and Rage.’’ Appleton, New York. Cardwell, J. R., and Liley, N. R. (1991). Androgen control of social status in males of a wild population of stoplight parrotfish, Sparisoma viride (Scaridae). Horm. Behav. 25, 1–18. Cardwell, J. R., Stacey, N. E., Tan, E. S. P., McAdam, D. S. O., and Lang, S. L. C. (1995). Androgen increases olfactory receptor response to a vertebrate sex pheromone. J. Comp. Physiol. A. 176, 55–61. Cavigelli, S. A., and Pereira, M. E. (2000). Mating season aggression and fecal testosterone levels in male ring-tailed lemurs (Lemur catta). Horm. Behav. 37, 246–255. Ceccarelli, I., Scaramuzzino, A., and Aloisi, A. M. (2001). Effects of gonadal hormones and persistent pain on non-spatial working memory in male and female rats. Behav. Brain Res. 123, 65–76. Chandler, C. R., Ketterson, E. D., and Nolan, V. Jr. (1997). Effects of testosterone on use of space by male dark-eyed juncos when their mates are fertile. Anim. Behav. 54, 543–549. Chase, I. D., Bartolomeo, C., and Dugatkin, L. A. (1994). Aggressive interactions and intercontest interval: How long do winners keep winning? Anim. Behav. 48, 393–400. Cheng, M.-F. (1986). Female cooing promotes ovarian development in ring doves. Physiol. Behav. 37, 371–374.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

223

Cheng, M.-F. (1992). For whom does the female dove coo? A case for the role of vocal selfstimulation Anim. Behav. 43, 1035–1044. Chu, J., and Wilczynski, W. (2001). Social influences on androgen levels in the southern leopard frog, Rana sphenocephala. Gen. Comp. Endocrinol. 121, 66–73. Clark, M. M., and Galef, B. G. Jr. (1999). A testosterone-mediated trade-off between parental care and sexual effort in male Mongolian gerbils (Meriones unguiculatus). J. Comp. Psychol. 113, 388–395. Clark, M. M., and Galef, B. G. Jr. (2000). Why some male Mongolian gerbils may help at the nest: Testosterone, asexuality and alloparenting. Anim. Behav. 59, 801–806. Clark, M. M., Malenfant, S. A., Winter, D. A., and Galef, B. G. Jr. (1990). Fetal uterine position affects copulation and scent marking by adult gerbils. Physiol. Behav. 47, 301–305. Clark, M. M., Tucker, L., and Galef, B. G. Jr. (1992a). Stud males and dud males: Intrauterine position effects on the success of male gerbils. Anim. Behav. 43, 215–221. Clark, M. M., vom Saal, F. S., and Galef, B. G. Jr. (1992b). Fetal intrauterine position correlates with endogenous testosterone levels in adult male Mongolian gerbils. Physiol. Behav. 51, 957–960. Clark, M. M., Vonk, J. M., and Galef, B. G. Jr. (1998). Intrauterine position, parenting and nest-site attachment in male Mongolian gerbils. Dev. Psychobiol. 32, 177–181. Clotfelter, E. D., and Paolino, A. D. (2003). Bystanders to contests between conspecifics are primed for increased aggression in male fighting fish. Anim. Behav. 66, 343–347. Coolen, L. M., Peters, H. J. P. W., and Veening, J. G. (1996). Fos immunoreactivity in the rat brain following consummatory elements of sexual behavior: A sex comparison. Brain Res. 738, 67–82. Coquelin, A., and Bronson, F. H. (1979). Release of luteinizing hormone in male mice during exposure to females: Habituation of the response. Science 206, 1099–1101. Courneya, K. S., and Carron, A. V. (1992). The home advantage in sport competitions: A literature review. J. Sport Exerc. Psychol. 14, 13–27. Cynx, J., and Nottebohm, F. (1992). Testosterone facilitates some conspecific song discriminations in castrated zebra finches Taeniopygia guttata. Proc. Natl. Acad. Sci. USA 89, 1376–1378. Davis, E. S., and Marler, C. A. (2003). The progesteroine challenge: Steroid hormone changes following a simulated territorial intrusion in female Peromyscus californicus. Horm. Behav. 44, 185–198. Davis, M. R., and Fernald, R. D. (1990). Social control of neuronal soma size. J. Neurobiol. 21, 1180–1188. De Ridder, E., Pinxten, R., and Eens, M. (2000). Experimental evidence of a testosteroneinduced shift from paternal to mating behavior in a facultatively polygynous songbird. Behav. Ecol. Sociobiol. 49, 24–30. Deviche, P. (1992). Testosterone and opioids interact to regulate feeding in a male migratory songbird. Horm. Behav. 26, 394–405. De Vries, G. J. (1995). Studying neurotransmitter systems to understand the development and function of sex differences in the brain: The case of vasopressin. In ‘‘Neurobiological Effects of Sex Steroid Hormones’’ (P. E. Micevych and R. P. Hamer Jr., Eds.), Cambridge University Press, Cambridge. Dohanich, G. (2002). Gonadal steroids, learning and memory. In ‘‘Hormones, Brain, and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach, and R. T. Rubin, Eds.), Vol. 2, pp. 265–327. Academic Press, New York. Drummond, H. (2001). The control and function of agonism in avian broodmates. In ‘‘Advances in the Study of Behavior’’ (P. J. B. Slater, J. S. Rosenblatt, C. T. Snowdon, and T. Roper, Eds.), Vol. 30, pp. 261–301. Academic Press, New York.

224

RUI F. OLIVEIRA

Drummond, H., and Osorio, J. L. (1992). Training siblings to be submissive losers: Dominance between booby nestlings. Anim. Behav. 44, 881–893. Dunlap, K. D., and Zakon, H. H. (1998). Behavioral actions of androgens and androgen receptor expression in the electrocommunication system of an electric fish, Eigenmannia virescens. Horm. Behav. 34, 30–38. Edwards, D. H., and Kravitz, E. A. (1997). Serotonin, social status and aggression. Curr. Opin. Neurobiol. 7, 812–819. Eichenbaum, H., Otto, T., and Cohen, N. J. (1992). The hippocampus: What does it do? Behav. Neural Biol. 57, 2–36. Eising, C. M., Eikenaar, C., Schwabl, H., and Groothuis, T. G. G. (2001). Maternal androgens in black-headed gull (Larus ridibundus) eggs: Consequences for chick development. Proc. R. Soc. Lond. B Biol. Sci. 268, 839–846. Ellis, L. (1995). Dominance and reproductive success among non-human animals: A crossspecies comparison. Ethol. Sociobiol. 16, 257–333. Enstrom, D. A., Ketterson, E. D., and Nolan, V. Jr. (1997). Testosterone and mate choice in the dark-eyed junco. Anim. Behav. 54, 1135–1146. Ferkin, M. H., and Johnston, R. E. (1993). Roles of gonadal hormones in control of five sexually attractive odors in meadow voles (Microtus pennsylvanicus). Horm. Behav. 27, 523–538. Fernald, R. D. (1976). The effect of testosterone on the behavior and coloration of adult male cichlid fish, Haplochromis burtoni. Horm. Res. 7, 172–178. Fernald, R. D. (1977). Quantitative behavioral observations of Haplochromis burtoni under semi-natural conditions. Anim. Behav. 25, 643–653. Fernald, R. D. (1984). Vision and behavior in an African cichlid fish. Am. Sci. 72, 58–65. Ferris, C. F., Melloni, R. H., Koppel, G., Perry, K. W., Fuller, R. W., and Delville, Y. (1997). Vasopressin/serotonin interactions in the anterior hypothalamus control aggressive behavior in golden hamsters. J. Neurosci. 17, 4331–4340. Fiorino, D. F., Coury, A., and Phillips, A. G. (1997). Dynamic changes in nucleus accumbens dopamine efflux during the Coolidge effect in male rats. J. Neurosci. 17, 4849–4855. Fishelson, L., and Hilzerman, F. (2002). Flexibility in reproductive styles of male St. Peter’s tilapia, Sarotherodon galilaeus (Cichlidae). Environ. Biol. Fishes 63, 173–182. Fishman, R. B., Chism, L., Firestone, G. L., and Breedlove, S. M. (1990). Evidence for androgen receptors in sexually dimorphic perineal muscles of neonatal male rats: Absence of androgen accumulation by perineal motoneurons. J. Neurobiol. 21, 694–705. Floody, O. R. (1983). Hormones and aggression in female mammals. In ‘‘Hormones and Aggressive Behavior’’ (B. B. Svare, Ed.), pp. 39–89. Plenum Press, New York. Folstad, I., and Karter, A. J. (1992). Parasites, bright males, and the immunocompetence handicap. Am. Nat. 139, 603–622. Fox, C., Ismail, A., Love, D., Kirkham, K., and Loraine, J. (1972). Studies on the relation of plasma testosterone levels and human sexual activity. J. Endocrinol. 52, 51–58. Francis, R. C. (1983). Experiential effects on agonistic behavior in the paradise fish, Macropodus opercularis. Behavior 85, 292–313. Francis, R. C., Soma, K. K., and Fernald, R. D. (1993). Social regulation of the brain– pituitary–gonadal axis. Proc. Natl. Acad. Sci. USA 90, 7794–7798. Freeman, E. R., Bloom, D. A., and McGuire, E. J. (2001). A brief history of testosterone. J. Urol. 165, 371–373. French, J. B. Jr., Nisbet, I. C. T., and Schwabl, H. (2001). Maternal steroids and contaminants in common tern eggs: A mechanism of endocrine disruption? Comp. Biochem. Physiol. 128, 91–98.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

225

Frey, D. F., and Miller, R. J. (1972). The establishment of dominance relationships in the blue gourami Trichogaster trichopterus (Pallas). Behavior 42, 8–62. Frye, C. A., and Reed, T. A. (1998). Androgenic neurosteroids: Anti-seizure effects in an animal model of epilepsy. Psychoneuroendocrinology 21, 431–439. Frye, C. A., and Seliga, A. M. (2001). Testosterone increases analgesia, anxiolysis, and cognitive performance of male rats. Cogn. Affect. Behav. Neurosci. 1, 371–381. Frye, C. A., Edinger, K. L., Seliga, A. M., and Wawrzycki, J. M. (2004). 5--reduced androgens may have actions in the hippocampus to enhance cognitive performance of male rats. Psychoneuroendocrinology 29, 1019–1027. Fryer, G., and Iles, T. D. (1972). ‘‘The Cichlid Fishes of the Great Lakes of Africa: Their Biology and Evolution.’’ Oliver and Boyd, Edinburgh, UK. Fu, P., Neff, B. D., and Gross, M. (2001). Tactic-specific success in sperm competition. Proc. R. Soc. Lond. B Biol. Sci. 268, 1105–1112. Gawienowsky, A. M., DeNicola, D. B., and Stacewicz-Sapuntzakis, M. (1976). Androgen dependence of a marking pheromone in rat urine. Horm. Behav. 7, 401–405. Gelinas, D., and Callard, G. V. (1997). Immunolocalization of aromatase and androgen receptor-positive neurons in the goldfish brain. Gen. Comp. Endocrinol. 106, 155–168. Gentner, T. Q., and Hulse, S. H. (2001). Female European starling preference and choice for variation in conspecific male song. Anim. Behav. 59, 443–458. Gentner, T. Q., Hulse, S. H., Duffy, D., and Ball, G. F. (2001). Response biases in auditory forebrain regions of female songbirds following exposure to sexually relevant variation in male song. J. Neurobiol. 46, 48–58. Gil, D., Graves, J., Hazon, N., and Wells, A. (1999). Male attractiveness and differential testosterone investment in zebra finch eggs. Science 286, 126–128. Ginsburg, B., and Allee, W. C. (1942). Some effects of conditioning of social dominance and subordination in inbred strains of mice. Physiol. Zool. 15, 485–506. Godwin, J., Warner, R. R., and Crews, D. (1996). Behavioral sex-change in the absence of gonads in a coral reef fish. Proc. R. Soc. Lond. B Biol. Sci. 263, 1683–1688. Gonzalez-Bono, E., Salvador, A., Serrano, M. A., and Ricarte, J. (1999). Testosterone, cortisol and mood in a sports team competition. Horm. Behav. 35, 55–62. Gould, S. J., and Vrba, E. S. (1982). Exaptation: A missing term in the science of form. Paleobiology 8, 4–15. Goulet, C., and Beaugrand, J. P. (2004). Inversion of initial dominance relationships following the interchange of roles of resident and intruder within pairs of male swordtail fish (Xiphophorus helleri). Behav. Process., in press. Grober, M. S. (1998). Socially controlled sex-change: Integrating ultimate and proximate levels of analysis. Acta Ethol. 1, 3–17. Grober, M. S., and Bass, A. H. (2002). Life history, neuroendocrinology, and behavior in fish. In ‘‘Hormones, Brain, and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach, and R. T. Rubin, Eds.), Vol. 2, pp. 331–347. Academic Press, New York. Groothuis, T. G. G. (1989). On the ontogeny of display behavior in the black-headed gull. II. Causal links between the development of aggression, fear and display behavior: Emancipation reconsidered. Behavior 110, 161–204. Groothuis, T. G. G., and Meeuwissen, G. (1992). The influence of testosterone on the development and fixation of the form of displays in two age classes of young black-headed gulls. Anim. Behav. 43, 189–208. Groothuis, T. G. G., and Schwabl, H. (2002). Determinants of within- and among-clutch variation in levels of maternal hormones in black-headed gull eggs. Funct. Ecol. 16, 281–289.

226

RUI F. OLIVEIRA

Grossman, C. J. (1985). Interactions between the gonadal steroids and the immune system. Science 227, 257–261. Grossman, C. J. (1990). Are there underlying immune–neuroendocrine interactions responsible for immunological sexual dimorphism? Prog. Neuroendocrinol. Immunol. 3, 75–82. Gubernick, D. J., Schneider, K. A., and Jeannotte, L. A. (1994). Individual differences in the mechanisms underlying the onset and maintenance of paternal behavior and the inhibition of infanticide in the monogamous biparental California mouse, Peromyscus californicus. Behav. Ecol. Sociobiol. 34, 225–231. Haake, P., Exton, M. S., Haverkamp, J., Kra¨mer, M., Leygraf, N., Hartmann, U., Schedlowski, M., and Kru¨ger, T. H. C. (2002). Absence of orgasm-induced prolactin secretion in a healthy multi-orgasmic male subject. Int. J. Impot. Res. 14, 133–135. Hannes, R. P. (1984). Androgen and corticoid levels in blood and body extracts of high and low ranking swordtail males (Xiphophorus helleri) before and after social isolation. Z. Tierpsychol. 66, 70–76. Hannes, R. P. (1986). Blood and whole-body androgen levels of male swordtails correlated with aggression measures in a standard-opponent test. Aggress. Behav. 12, 249–254. Harding, C. F. (1981). Social modulation of circulating hormone levels in the male. Am. Zool. 21, 223–231. Harding, C. F., and Follet, B. K. (1979). Hormone changes triggered by aggression in a natural population of blackbirds. Science 203, 918–920. Harrell, L. E., Goyal, M., Parsons, D. S., and Peagler, A. (1990). The effect of gonadal steroids on the behavioral and biochemical effects of hippocampal sympathetic in growth. Physiol. Behav. 48, 507–513. Harvey, P. H., and Pagel, M. D. (1991). ‘‘The Comparative Method in Evolutionary Biology.’’ Oxford University Press, Oxford. Hasselquist, D., Marsh, J. A., Sherman, P. W., and Wingfield, J. C. (1999). Is avian humoral immunocompetence suppressed by testosterone? Behav. Ecol. Sociobiol. 45, 167–175. Havens, M. D., and Rose, J. D. (1992). Investigation of familiar and novel chemosensory stimuli by golden hamsters: Effects of castration and testosterone replacement. Horm. Behav. 26, 505–511. Hegner, R. E., and Wingfield, J. C. (1987). Effects of experimental manipulations of testosterone levels on parental investment and breeding success in male house sparrows. Auk 104, 462–469. Henderson, D. L., and Chiszar, D. A. (1977). Analysis of aggressive behavior in the bluegil sunfish Lepomis macrochirus Rafinesque: Effects of sex and size. Anim. Behav. 25, 122–130. Higley, J. D., Mehlman, P. T., Poland, R. E., Taub, D. M., Vickers, J., Suomi, S. J., and Linnoila, M. (1996). CSF testosterone and 5-HIAA correlate with different types of aggressive behaviors. Biol. Psychiatry 40, 1067–1082. Hillgarth, N., and Wingfield, J. C. (1997). Parasite mediated sexual selection: Endocrine aspects. In ‘‘Host–Parasite Evolution: General Principles and Avian Models’’ (D. Clayton and J. Moore, Eds.), pp. 78–104. Oxford University Press, Oxford. Hillgarth, N., Ramenofsky, M., and Wingfield, J. C. (1997). Testosterone and sexual selection. Behav. Ecol. 8, 108–109. Hinde, R. (1970). ‘‘Animal Behavior: A Synthesis of Ethology and Comparative Psychology,’’ 2nd ed. John Wiley & Sons, New York. Hirschenhauser, K., Frigerio, D., Grammer, K., and Magnusson, M. S. (2002a). Monthly patterns of testosterone and behavior in prospective fathers. Horm. Behav. 42, 172–181.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

227

Hirschenhauser, K., Ros, A. F. H., Carneiro, L. A., Oliveira, T., Silva, A., Canario, A. V. M., and Oliveira, R. F. (2002b). Non-invasive hormone assessment from fish-holding water. Adv. Ethol. 37, 139. Hirschenhauser, K., Winkler, H., and Oliveira, R. F. (2003). Comparative analyses of male androgen responsiveness to social environment in birds: The effects of mating system and paternal incubation. Horm. Behav. 43, 508–519. Hirschenhauser, K., Taborsky, M., Oliveira, T., Canario, A. V. M., and Oliveira, R. F. (2004). A test of the ‘‘challenge hypothesis’’ in cichlid fish: Simulated partner and territory intruder experiments Anim. Behav. 68, 741–750. Hogan, J. A., and Bols, R. J. (1980). Priming of aggressive motivation in Betta splendens. Anim. Behav. 28, 135–142. Holst, D. V., and Eichmann, F. (1998). Sex-specific regulation of marking behavior by sex hormones and conspecifics scent in tree shrews (Tupaia belangeri). Physiol. Behav. 63, 157–164. Hong, Y., Gagnon, J., Rice, T., Pe´russe, L., Leon, A. S., Skinner, J. S., Wilmore, J. H., Bouchard, C., and Rao, D. C. (2001). Familial resemblance for free androgens and androgen glucorunides in sedentary black and white individuals: The HERITAGE Family Study. Helath, Risk Factors, Exercise Training and Genetics. J. Endocrinol. 170, 485–492. Hsu, Y., and Wolf, L. L. (1999). The winner and loser effect: Integrating multiple experiences. Anim. Behav. 57, 903–910. Huang, W. J., Yeh, J. Y., Tsai, S. C., Lin, H., Chiao, Y. C., Chen, J. J., Lu, C. C., Hwang, S. W., Wang, S. W., Chang, L. S., and Wang, P. S. (1999). Regulation of testosterone secretion by prolactin in male rats. J. Cell. Biochem. 74, 111–118. Huhman, K. L., Bunnell, B. N., Mougey, E. H., and Meyerhoff, J. L. (1990). Effects of social conflict on POMC-derived peptides and glucocorticoids in male golden hamsters. Physiol. Behav. 47, 949–956. Huhman, K. L., Moore, T. O., Ferris, C. F., Mougey, E. H., and Meyerhoff, J. L. (1991). Acute and repeated exposure to social conflict in male golden hamsters: Increases in plasma POMC-derived peptides and cortisol and decreases in plasma testosterone. Horm. Behav. 25, 206–216. Huhman, K. L., Moore, T. O., Mougey, E. H., and Meyerhoff, J. L. (1992). Hormonal responses to fighting in hamsters: Separation of physical and psychological causes. Physiol. Behav. 51, 1083–1086. Huhman, K. L., Solomon, M. B., Janicki, M., Harmon, A. C., Lin, S. M., Israel, J. E., and Jasnow, A. M. (2003). Conditioned defeat in male and female Syrian hamsters. Horm. Behav. 44, 293–299. Huntingford, F., and Turner, A. (1986). ‘‘Animal Conflict.’’ Chapman & Hall, London. Hutchison, J. B. (1971). Effects of hypothalamic implants of gonadal steroids on courtship behavior in Barbary doves (Streptopelia risoria). J. Endocrinol. 50, 97–113. Ikeda, K. (1933). Effect of castration on the sexual characters of anadromous three-spined stickleback, Gasterosteus aculeatus aculeatus (L.). Jpn. J. Zool. 5, 135–157. Jarvis, E. D., Schwabl, H., Ribeiro, S., and Mello, C. V. (1997). Brain gene regulation by territorial singing behavior in freely ranging songbirds. Neuroreport 8, 2073–2077. Jenkins, J. S. (2000). The lost voice: A history of the castrato. J. Pediatr. Endocrinol. Metab. 13(Suppl. 6), 1503–1508. Jones, J. S., and Wynne-Edwards, K. E. (2000). Paternal hamsters mechanically assist the delivery, consume amniotic fluid and placenta, remove fetal membranes and provide parental care during the birth process. Horm. Behav. 37, 116–125.

228

RUI F. OLIVEIRA

Kavaliers, M., Choleris, E., and Colwell, D. D. (2001). Brief exposure to female odors ‘‘emboldens’’ male mice by reducing predator-induced behavioral and hormonal responses Horm. Behav. 40, 497–509. Kelley, D. C. (2002). Hormonal regulation of motor output in amphibians: Xenopus laevis vocalizations as a model system. In ‘‘Hormones, Brain, and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach, and R. T. Rubin, Eds.), Vol. 2, pp. 445–468. Academic Press, New York. Kelley, D. C., and Dennison, J. (1990). The vocal motor neurons of Xenopus laevis: Development of sex differences in axon number. J. Neurobiol. 21, 869–882. Kelley, D. C., Fenstemaker, S., Hannigan, P., and Shih, S. (1988). Sex differences in the motor nucleus of cranial nerve IX–X in Xenopus laevis: A quantitative study. J. Neurobiol. 19, 413–429. Kendrick, K., and Drewett, R. (1979). Testosterone reduces refractory period of stria terminalis neurons in the rat brain. Science 204, 877–879. Kerr, J. E., Allore, R. J., Beck, S. G., and Handa, R. J. (1995). Distribution and hormonal regulation of androgen receptors (AR) and AR messenger ribonucleic acid in the rat hippocampus. Endocrinology 136, 3213–3221. Kerr, J. H., and Vanschaik, P. (1995). Effects of game venue and outcome on psychological mood states in rugby. Pers. Individ. Differ 19, 407–410. Ketterson, E. D., and Nolan, V. Jr. (1992). Hormones and life histories: An integrative approach. Am. Nat. 140, S33–S62. Ketterson, E. D., and Nolan, V. Jr. (1999). Adaptation, exaptation, and constraint: A hormonal perspective. Am. Nat. 153, S4–S25. Ketterson, E. D., Nolan, V. Jr., Wolf, L., and Ziegenfus, C. (1992). Testosterone and avian life histories: Effects of experimentally elevated testosterone on behavior and correlates of fitness in the dark-eyed junco (Junco hyemalis). Am. Nat. 140, 980–999. Ketterson, E. D., Nolan, V. Jr., Cawthorn, M. J., Parker, P. G., and Ziegenfus, C. (1996). Phenotypic engineering: Using hormones to explore the mechanistic and functional bases of phenotypic variation in nature. Ibis 138, 1–17. Ketterson, E. D., Nolan, V. Jr., Casto, J. M., Buerkle, C. A., Clotfelter, E., Grindstaff, J. L., Jones, K. J., Lipar, J. L., McNabb, F. M. A., Neudorf, D. L., Parker-Renga, I., Schoech, S. J., and Snajdr, E. (2001). Testosterone, phenotype and fitness: A research program in evolutionary behavioral endocrinology. In ‘‘Avian Endocrinology’’ (A. Dawson and C. M. Chaturvedi, Eds.), pp. 19–40. Narosa Publishing House, New Delhi, India. Keverne, E. B., Martel, F. L., and Nevison, C. M. (1996). Primate brain evolution: Genetic and functional considerations. Proc. R. Soc. Lond. Biol. Sci. 263, 689–696. Kimball, R. T., and Ligon, J. D. (1999). Evolution of avian plumage dichromatism from a proximate perspective. Am. Nat. 154, 182–193. Kindler, P. M., Philipp, D. P., Gross, M. R., and Bahr, J. M. (1989). Serum 11-ketotestosterone and testosterone concentrations associated with reproduction in male bluegill (Lepomis macrochirus: Centrarchidae). Gen. Comp. Endocrinol. 75, 446–453. Knapp, R., Wingfield, J. C., and Bass, A. H. (1999). Steroid hormones and paternal care in the plainfin midshipman fish (Porichthys notatus). Horm. Behav. 35, 81–89. Kritzer, M. F., McLaughlin, P. J., Smirlis, T., and Robinson, J. K. (2001). Gonadectomy impairs T-maze acquisition in adult male rats. Horm. Behav. 39, 167–174. Kru¨ger, T. H. C., Haake, P., Hartmann, U., Schedlowski, M., and Exton, M. S. (2002). Orgasminduced prolactin secretion: Feedback control of sexual drive? Neurosci. Biobehav. Rev. 26, 31–44. Kuwamura, T. (1987). Male mating territory and sneaking in a maternal mouthbrooder, Pseudosimochromis curvifrons (Pisces; Cichlidae). J. Ethol. 5, 203–206.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

229

Larsson, K., So¨dersten, P., and Beyer, C. (1973). Induction of male sexual behavior by oestradiol benzoate in combination with dihydrotestosterone. J. Endocrinol. 57, 563–564. Lehrman, D. S. (1953). A critique of Konrad Lorenz’a theory of instinctive behavior. Q. Rev. Biol. 28, 337–363. Lehrman, D. S. (1970). Semantic and conceptual issues in the nature–nurture problem. In ‘‘Development and Evolution of Behavior’’ (L. R. Aronson, E. Tobach, D. S. Lehrman, and J. S. Rosenblatt, Eds.), pp. 17–52. W. H. Freeman, New York. Leiser, J. K., and Itzkowitz, M. (1999). The benefits of dear enemy recognition in threecontender convict cichlid (Cichlasoma nigrofasciatum) contests. Behavior 136, 983–1003. Leitz, T. (1987). Social control of testicular steroidogenic capacities in the Siamese fighting fish Betta splendens Regan. J. Exp. Zool. 244, 473–478. Leshner, A. I. (1979). Kinds of hormonal effects on behavior: A new view. Neurosci. Biobehav. Rev. 3, 69–73. Leshner, A. I. (1983). The hormonal responses to competition and their behavioral significance. In ‘‘Hormones and Aggressive Behavior’’ (B. B. Svare, Ed.), pp. 393–404. Plenum Press, New York. Levy, M., and Aronson, L. R. (1955). Morphological effects of castration and hormone administration in the male cichlid fish Tilapia macrocephala. Anat. Rec. 122, 450–451. Lewontin, R. C., Rose, S., and Kamin, L. J. (1984). ‘‘Not in Our Genes.’’ Pantheon, New York. Limberger, D. (1983). Pairs and harems in a cichlid fish Lamprologus brichardi. Z. Tierpsychol. 62, 115–144. Lindstro¨m, K. M., Krakower, D., Lundstro¨m, J. O., and Silverin, B. (2001). The effects of testosterone on a viral infection in greenfinches (Carduelis chloris): An experimental test of the immunocompetence-handicap hypothesis. Proc. R. Soc. Lond. B Biol. Sci. 268, 207–211. Lipar, J. L., Ketterson, E. D., and Nolan, V. Jr. (1999). Intraclutch variation in testosterone content of red-winged blackbird eggs. Auk 116, 231–235. Lonstein, J. S., and De Vries, G. J. (1999). Sex differences in the parental behavior of adult virgin prairie voles: Independence from gonadal hormones and vasopressin. J. Neuroendocrinol. 11, 441–449. Lorenz, K. (1939). Vergleichende Verhaltensforschung. Zool. Anz. Suppl. 12, 69–102. Lorenz, K. (1965). ‘‘Evolution and Modification of Behavior.’’ University of Chicago Press, Chicago, IL. Luine, V., Nottebohm, F., Harding, C., and McEwen, B. S. (1980). Androgen affects cholinergic enzymes in syringeal motor neurons and muscle. Brain Res. 192, 89–107. Lynn, S. E., Houtman, A. M., Weathers, W. W., Keterson, E. D., and Nolan, V. Jr. (2000). Testosterone increases activity but not daily energy expenditure in captive male darkeyed juncos. Anim. Behav. 60, 581–587. Manzo, J., Garcia, L. I., Hernandez, M. E., Carrillo, P., and Pacheco, P. (2002). Neuroendocrine control of urine-marking behavior in male rats. Physiol. Behav. 75, 25–32. Marler, C. A., Walsberg, G., White, M. I., and Moore, M. (1995). Increased energy expenditure due to increased territorial defense in male lizards after phenotypic manipulation. Behav. Ecol. Sociobiol. 37, 225–231. Marsh, J. (1992). Neuroendocrine–immune interactions in the avian species: A review. Poultry Sci. Rev. 4, 129–167. Maruniak, J. A., Desjardins, C., and Bronson, F. H. (1977). Dominant–subordinate relationships in castrated male mice bearing testosterone implants. Am. J. Physiol. 233, E495–E499. Mason, J. W. (1968a). Organisation of the psychoendocrine mechanisms. Psychosom. Med. 30, 565–808.

230

RUI F. OLIVEIRA

Mason, J. W. (1968b). A review of psychoendocrine research on the pituitary–adrenal cortical system. Psychosom. Med. 30, 791–808. Mason, J. W., and Brady, J. V. (1956). Plasma 17-hydroxycorticosteroid changes related to reserpine effects on emotional behavior. Science 124, 983–984. Mazuc, J., Bonneaud, C., Chastel, O., and Sorci, G. (2003). Social environment affects female and egg testosterone levels in the house sparrow (Passer domesticus). Ecol. Lett. 6, 1084–1090. Mazur, A. (1985). A biosocial model of status in face-to-face primate groups. Social Forces 64, 377–402. Mazur, A., and Booth, A. (1998). Testosterone and dominance in men. Behav. Brain Sci. 21, 353–397. Mazur, A., and Lamb, T. (1980). Testosterone, status, and mood in human males. Horm. Behav. 14, 236–246. Mazur, A., Booth, A., and Dabbs, J. Jr. (1992). Testosterone and chess competition. Social Psychol. Q. 55, 70–77. McCaul, K., Gladue, B., and Joppa, M. (1992). Winning, losing, mood, and testosterone. Horm. Behav. 26, 486–506. McClelland, D. C. (1987). ‘‘Human Motivation.’’ Cambridge University Press, Cambridge. McGregor, P. K. (1993). Signalling in territorial systems: A context for individual identification, ranging and eavesdropping. Philos. Trans. R. Soc. Lond. B Biol. Sci. 340, 237–244. McGuire, E. J., Courneya, R. S., Widmeyer, W. N., and Carron, A. V. (1992). Aggression as a potential mediator of the home advantage in professional ice hockey. J. Sport Exerc. Psychol. 14, 148–158. Meddle, S. L., King, V. M., Follet, B. K., Wingfield, J. C., Ramenofsky, M., Foidart, A., and Balthazart, J. (1997). Copulation activates Fos-like immunoreactivity in the male quail forebrain. Behav. Brain Res. 85, 143–159. Mehlman, P. T., Highley, J. D., Fernald, B. J., Sallee, F. R., Suomi, S. J., and Linnoila, M. (1997). CSF 5-HIAA, testosterone and sociosexual behaviors in free-ranging male rhesus macaques in the mating season. Psychiatry Res. 72, 89–102. Meikle, A. W., Stringham, J. D., Bishop, D. T., and West, D. W. (1988). Quantitating genetic and nongenetic factors influencing androgen production and clearance rates in men. J. Clin. Endocrinol. Metab. 67, 104–109. Mello, C., Vicario, D. S., and Clayton, D. F. (1992). Song presentation induces gene expression in the songbird forebrain. Proc. Natl. Acad. Sci. USA 89, 6818–6822. Mello, C., Nottebohm, F., and Clayton, D. (1995). Repeated exposure to one song leads to a rapid and persistent decline in an immediate early gene’s response to that song in zebra finch telencephalon. J. Neurosci. 15, 6919–6925. Mizoguchi, K., Kunishita, T., Chui, D. H., and Tabira, T. (1992). Stress induces neuronal death in the hippocampus. Neurosci. Lett. 138, 157–160. Mock, D. W., and Parker, G. A. (1997). ‘‘The Evolution of Sibling Rivalry.’’ Oxford University Press, Oxford. Moore, M. C. (1991). Application of organization–activation theory to alternative male reproductive strategies: A review. Horm. Behav. 25, 154–179. Moore, M. C., Hews, D. K., and Knapp, R. (1998). Hormonal control and evolution of alternative male phenotypes: Generalizations of models for sexual differentiation. Am. Zool. 38, 133–151. Muller, M. N., and Wrangham, R. W. (2004). Dominance, aggression and testosterone in wild chimpanzees: A test of the ‘‘challenge hypothesis.’’ Anim. Behav. 67, 113–123.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

231

Neave, N., and Wolfson, S. (2003). Testosterone, territoriality, and the ‘‘home advantage.’’ Physiol. Behav. 78, 269–275. Nelson, C. M. (1995). Male size, spawning pit size and female mate choice in a lekking cichlid fish. Anim. Behav. 50, 1587–1599. Nelson, R. J. (2000). ‘‘An Introduction to Behavioral Endocrinology,’’ 2nd ed. Sinauer Associates, Sunderland, MA. Nevill, A., and Holder, R. L. (1999). Home advantage in sport: An overview of studies on the advantage of playing at home. Sports Med. 28, 221–236. Nishida, M., and Yanagisawa, Y. (1991). The social and mating system of the maternal mouthbrooder Tropheus moorii (Cichlidae) in Lake Tanganyika. Jap. J. Ichthyol. 38, 271–282. Nock, B. L., and Leshner, A. I. (1976). Hormonal mediation of the effects of defeat on agonistic responding in mice. Physiol. Behav. 17, 111–119. Nunes, S., Fite, J. E., Patera, J., and French, J. A. (2001). Interactions among paternal behavior, steroid hormones, and parental experience in male marmosets (Callithrix kuhilii). Horm. Behav. 39, 70–82. Nun˜ez de la Mora, A., Drummond, H., and Wingfield, J. C. (1996). Hormonal correlates of dominance and starvation-induced aggression in chicks of the blue-footed booby. Ethology 102, 748–761. Oliveira, R. F. (2005). Hormones, social context and animal communication. In ‘‘Communication Networks’’ (P. K. McGregor, Ed.), pp. 481–520. Cambridge University Press, Cambridge. Oliveira, R. F., and Almada, V. C. (1996). On the (in)stability of dominance hierarchies in the cichlid fish Oreochromis mossambicus. Aggress. Behav. 22, 37–45. Oliveira, R. F., and Almada, V. C. (1998a). Mating tactics and male–male courtship in the lekbreeding cichlid Oreochromis mossambicus. J. Fish Biol. 52, 1115–1129. Oliveira, R. F., and Almada, V. C. (1998b). Androgenization of dominant males in a cichlid fish: Androgens mediate the social modulation of sexually dimorphic traits. Ethology 104, 841–858. Oliveira, R. F., and Almada, V. C. (1999). Male displaying characters, gonadal maturation and androgens in the cichlid fish, Oreochromis mossambicus. Acta Ethol. 2, 67–70. Oliveira, R. F., and Canario, A. V. M. (2001). Hormones and social behavior of cichlid fishes: A case study in the Mozambique tilapia. J. Aquaricult. Aquat. Sci. 9, 187–207. Oliveira, R. F., Almada, V. C., and Canario, A. V. M. (1996). Social modulation of sex steroid concentrations in the urine of male cichlid fish Oreochromis mossambicus. Horm. Behav. 30, 2–12. Oliveira, R. F., McGregor, P. K., and Latruffe, C. (1998). Know thine enemy: Fighting fish gather information from observing conspecific interactions. Proc. R. Soc. Lond. B Biol. Sci. 265, 1045–1049. Oliveira, R. F., Almada, V. C., Goncalves, E. J., Forsgren, E., and Canario, A. V. M. (2001a). Androgen levels and social interactions in breeding males of the peacock blenny. J. Fish Biol. 58, 897–908. Oliveira, R. F., Canario, A. V. M., and Grober, M. S. (2001b). Male sexual polymorphism, alternative reproductive tactics and androgens in combtooth blennies (Pisces: Blenniidae). Horm. Behav. 40, 266–275. Oliveira, R. F., Canario, A. V. M., Grober, M. S., and Santos, R. S. (2001c). Endocrine correlates of alternative reproductive tactics and male polymorphism in the Azorean rock-pool blenny, Parablennius sanguinolentus parvicornis. Gen. Comp. Endocrinol. 121, 278–288.

232

RUI F. OLIVEIRA

Oliveira, R. F., Carneiro, L. A., Goncalves, D. M., Canario, A. V. M., and Grober, M. S. (2001d). 11-Ketotestosterone inhibits the alternative mating tactic in sneaker males of the peacock blenny, Salaria pavo. Brain Behav. Evol. 58, 28–37. Oliveira, R. F., Lopes, M., Carneiro, L. A., and Canario, A. V. M. (2001e). Watching fights raises fish hormone levels. Nature 409, 475. Oliveira, R. F., Ros, A. F. H., Hirschenhauser, K., and Canario, A. V. M. (2001f). Androgens and mating systems in fish: Intra- and interspecific analyses. In ‘‘Perspectives in Comparative Endocrinology: Unity and Diversity’’ (H. J. Goos, R. K. Rastogi, H. Vaudry, and R. Pierantoni, Eds.), pp. 203–215. Monduzzi Editore, Bologna, Italy. Oliveira, R. F., Hirschenhauser, K., Carneiro, L. A., and Canario, A. V. M. (2002). Social modulation of androgens in male teleost fish. Comp. Biochem. Physiol. 132B, 203–215. Oliveira, R. F., Hirschenhauser, K., Canario, A. V. M., and Taborsky, M. (2003). Androgen levels of reproductive competitors in a cooperatively breeding cichlid. J. Fish Biol. 63, 1615–1620. Olsen, N. J., and Kovacs, W. J. (1996). Gonadal steroids and immunity. Endocrinol. Rev. 17, 369–384. Osadchuk, L. V. (2001). The effect of genetic selection for lack of aggression towards humans on male reproductive physiology in the silver fox. In ‘‘Proceedings of the British Society of Animal Science’’ 2001. Available at http://www.bsas.org.uk/meetings/annlproc/pdf2001 1016.pdf (accessed August 2004). Owen, P. C., and Perrill, S. A. (1998). Habituation in the green frog, Rana clamitans. Behav. Ecol. Sociobiol. 44, 209–213. Packard, M. G., Schroeder, J. P., and Alexander, G. M. (1998). Expression of testosterone conditioned place preference is blocked by peripheral or intra-accumbens injection of -flupenthixol. Horm. Behav. 34, 39–47. Pankhurst, N. W., and Barnett, P. W. (1993). Relationship of population density, territorial interaction and plasma levels of gonadal steroids in spawning male demoiselles Chromis dispilus (Pisces: Pomacentridae). Gen. Comp. Endocrinol. 90, 168–176. Peters, A. (2000). Testosterone treatment is immunosuppressive in superb fairy-wrens, yet free-living males with high testosterone are more immunocompetent. Proc. R. Soc. Lond. B Biol. Sci. 267, 883–889. Pfaus, J. G., and Heeb, M. M. (1997). Implications of immediate-early gene induction in the brain following sexual stimulation of female and male rodents. Brain Res. Bull. 44, 397–407. Phoenix, C., Goy, R., Gerall, A., and Young, W. (1959). Organizing action of prenatallyadministered testosterone proprionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology 65, 369–382. Pilz, K. M., and Smith, H. G. (2004). Egg yolk androgen levels increase with breeding density in the European starling, Sturnus vulgaris. Funct. Ecol. 18, 58–66. Pomerantz, S. M., and Sholl, S. A. (1987). Analysis of sex and regional differences in androgen receptors in fetal rhesus monkey brain. Brain Res. 433, 151–154. Rabkin, J. G., Wagner, G., and Rabkin, R. (1996). Treatment of depression in HIV men: Literature review and report of an ongoing study of testosterone replacement therapy. Ann. Behav. Med. 18, 24–29. Ramos-Fernandez, G., Nun˜ez de la Mora, A., Wingfield, J. C., and Drummond, H. (2000). Endocrine correlates of dominance in chicks of the blue-footed booby (Sula nebouxii): Testing the challenge hypothesis. Ethol. Ecol. Evol. 12, 27–34. Raouf, S. A., Parker, P. G., Ketterson, E. D., Nolan, V. Jr., and Ziegenfus, C. (1997). Testosterone influences reproductive success by increasing extra-pair fertilizations in male dark-eyed juncos, Junco hyemalis. Proc. R. Soc. Lond. B Biol. Sci. 264, 1599–1603.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

233

Reavis, R. H., and Grober, M. S. (1999). An integrative approach to sex change: Social, behavioral and neurochemical changes in Lythrypnus dalli (Pisces). Acta Ethol. 2, 51–60. Reburn, C. J., and Wynne-Edwards, K. E. (1999). Hormonal changes in males of a naturally biparental and uniparental mammal. Horm. Behav. 35, 163–176. Reed, W. L., and Vleck, C. M. (2001). Functional significance of variation in egg-yolk androgens in the American coot. Oecologia 128, 164–171. Robertson, G. S., Pfaus, J. G., Atkinson, L. J., Matsumura, H., Phillips, A. G., and Fibiger, H. C. (1991). Sexual behavior increases c-fos expression in the forebrain of the male rat. Brain Res. 564, 352–357. Rocha, E. M., Wickham, L. A., da Silveira, L. A., Krenzer, K. L., Yu, F. S., Toda, I., Sullivan, B. D., and Sullivan, D. A. (2000). Identification of androgen receptor protein and 5--reductase mRNA in human ocular tissues. Br. J. Ophthalmol. 84, 76–84. Roney, J. R., Mahler, S. V., and Maestripieri, D. (2003). Behavioral and hormonal responses of men to brief interactions with women. Evol. Hum. Behav. 24, 365–375. Ros, A. F. H. (1999). Effects of testosterone on growth, plumage pigmentation, and mortality in black-headed gull chicks. Ibis 141, 451–459. Ros, A. F. H., Groothuis, T. G. G., and Apanius, V. (1997). The relationship among gonadal steroids, immunocompetence, body mass and behavior in young black-headed gulls (Larus ridibundus). Am. Nat. 150, 201–219. Ros, A. F. H., Hirschenhauser, K., and Oliveira, R. F. (2001). The interaction between organizational and activational effects of testosterone in the control of early aggression in birds: A comment on Sasva´ri, Hegyi and Pe´czely. Ethology 107, 851–853. Ros, A. F. H., Dieleman, S. J., and Groothuis, T. G. G. (2002). Social stimuli, testosterone, and aggression in gull chicks: Support for the challenge hypothesis. Horm. Behav. 41, 334–342. Ros, A. F. H., Canario, A. V. M., Couto, E., Zeilstra, I., and Oliveira, R. F. (2003). Endocrine correlates of intra-specific variation in the mating system of the St. Peter’s fish (Sarotherodon galilaeus). Horm. Behav. 44, 365–373. Ros, A. F. H., Becker, K., Canario, A. V. M., and Oliveira, R. F. (2004a). Androgen levels and energy metabolism in Oreochromis mossambicus. J. Fish Biol. 45, 895–905. Ros, A. F. H., Bruintjes, R., Santos, R. S., Canario, A. V. M., and Oliveira, R. F. (2004b). The role of androgens in the trade-off between territorial and parental behavior in the Azorean rock-pool blenny, Parablennius parvicornis. Horm. Behav. 46, 491–497. Rose, R. M., Holaday, J. W., and Bernstein, I. S. (1971). Plasma testosterone, dominance rank and aggressive behavior in male rhesus monkeys. Nature 231, 366–368. Rose, R. M., Gordon, T. P., and Bernstein, I. S. (1972). Plasma testosterone levels in the male rhesus: Influences of sexual and social stimuli. Science 178, 643–645. Rose, R. M., Bernstein, I. S., and Gordon, T. P. (1975). Consequences of social conflict on plasma testosterone levels in rhesus monkeys. Psychosom. Med. 37, 50–61. Rosen, G., O’Bryant, E., Matthews, J., Zacharewski, T., and Wade, J. (2002). Distribution of androgen receptor mRNA expression and immunoreactivity in the brain of the green anole lizard. J. Neuroendocrinol. 14, 19–28. Rosenthal, G. G., and Evans, C. S. (1998). Female preference for swords in Xiphophorus helleri reflects a bias for larger apparent size. Proc. Natl. Acad. Sci. USA 95, 4431–4436. Rouse, E. F., Coppinger, C. J., and Barnes, P. R. (1977). The effect of an androgen inhibitor on behavior and testicular morphology in the stickleback Gasterosteus aculeatus. Horm. Behav. 9, 8–18. Rowland, W. J. (1999). Studying visual cues in fish behavior: A review of ethological techniques. Environ. Biol. Fish 56, 285–305. Royle, N. J., Surai, P. F., and Hartley, I. R. (2000). Maternally derived androgens and antioxidants in bird eggs: Complementary but opposing effects? Behav. Ecol. 12, 381–385.

234

RUI F. OLIVEIRA

Ryan, S. M., and Maier, S. F. (1988). The estrous cycle and estrogen modulate stress-induced analgesia. Behav. Neurosci. 102, 371–380. Salas, C., Broglio, C., and Rodı´guez, F. (2003). Evolution of forebrain and spatial cognition in vertebrates: Conservation across diversity. Brain Behav. Evol. 62, 72–82. Saldanha, C. J., Clayton, N. S., and Schlinger, B. A. (1999). Androgen metabolism in the juvenile oscine forebrain: A cross-species analysis at neural sites implicated in memory function. J. Neurobiol. 40, 397–406. Salvador, A., Simon, V., Suay, F., and Llorens, L. (1987). Testosterone and cortisol responses to competitive fighting in human males: A pilot study. Aggress. Behav. 13, 9–13. Sapolsky, R. M. (1997). ‘‘The Trouble with Testosterone.’’ Scribner, New York. Sapolsky, R. M. (2002). Endocrinology of the stress-response. In ‘‘Behavioral Endocrinology’’ (J. B. Becker, S. M. Breedlove, D. Crews, and M. McCarthy, Eds.), 2nd ed. pp. 409–450. MIT Press, Cambridge, MA. Sato, T. (1994). Active accumulation of spawning substrate: a determinant of extreme polygyny in a shell-brooding cichlid fish. Anim. Behav. 48, 669–678. Sawyer, T. F., Hengehold, A. K., and Perez, W. A. (1984). Chemosensory and hormonal mediation of social memory in male rats. Behav. Neurosci. 98, 908–913. Schlinger, B. A., and Brenowitz, E. A. (2002). Neural and hormonal control of birdsong. In ‘‘Hormones, Brain and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach, and R. T. Rubin, Eds.), Vol. 2, pp. 799–839. Academic Press, New York. Scholz, P. O. (2001). ‘‘Eunuchs and Castrati: A Cultural History.’’ Markus Wiener, Princeton, NJ. Schradin, C., and Anzenberger, G. (1999). Prolactin, the hormone of paternity. News Physiol. Sci. 14, 223–231. Schuett, G. W. (1997). Body size and agonistic experience affect dominance and mating success in male copperheads. Anim. Behav. 54, 213–224. Schulkin, J. (2002). Hormonal modulation of central motivational states. In ‘‘Hormones, Brain and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach, and R. T. Rubin, Eds.), Vol. 1, pp. 633–657. Academic Press, New York. Schultheiss, D., Denil, J., and Jonas, U. (1997). Rejuvenation in the early 20th century. Andrologia 29, 351–355. Schultheiss, D., Bloom, D. A., Wefer, J., and Jonas, U. (2000). Tissue engineering from Adam to the zygote: Historical reflections. World J. Urol. 18, 84–90. Schultheiss, O. C., and Rhode, W. (2002). Implicit power motivation predicts men’s testosterone changes and implicit learning in a contest situation. Horm. Behav. 41, 195–202. Schultheiss, O. C., Campbell, K. L., and McClelland, D. C. (1999). Implicit power motivation moderates men’s testosterone response to imagined and real dominance success. Horm. Behav. 36, 234–241. Schuurman, T. (1980). Hormonal correlates of agonistic behavior in adult male rats. Prog. Brain Res. 53, 415–420. Schwabl, H. (1993). Yolk is a source of maternal testosterone for developing birds. Proc. Natl. Acad. Sci. USA 90, 11446–11450. Schwabl, H. (1996). Environment modifies the testosterone levels of a female bird and its eggs. J. Exp. Zool. 276, 157–163. Schwabl, H. (1997). The contents of maternal testosterone in house sparrow Passer domesticus eggs vary with breeding condition. Naturwissenschaften 84, 406–408. Schwabl, H., Mock, D., and Gieg, J. (1997). A hormonal mechanism for parental favouritism. Nature 386, 231. Selye, H. (1976). ‘‘The Stress of Life.’’ McGraw-Hill, New York.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

235

Shapiro, D. Y. (1979). Social behavior, group structure and the control of sex reversal in hermaphrodite fish. Adv. Study Behav. 10, 43–102. Shapiro, D. Y. (1981). Size, maturation, and the social control of sex reversal in the coral reef fish Anthias squamipinnis. J. Zool. 193, 105–128. Sikkel, P. C. (1993). Changes in plasma androgen levels associated with changes in male reproductive behavior in a brood cycling marine fish. Gen. Comp. Endocrinol. 89, 229–237. Silverin, B. (1980). Effects of long-acting testosterone treatment on free-living pied flycatchers, Ficedula hypoleuca, during the breeding period. Anim. Behav. 28, 906–912. Simon, N. G. (2002). Hormonal processes in the development and expression of aggressive behavior. In ‘‘Hormones, Brain and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach, and R. T. Rubin, Eds.), Vol. 1, pp. 339–392. Academic Press, New York. Sinervo, B., and Svensson, E. (1998). Mechanistic and selective causes of life-history tradeoffs and plasticity: Evolutionary physiology of the costs of reproduction. Oikos 83, 432–442. Sisneros, J. A., and Tricas, T. C. (2000). Androgen-induced changes in the response dynamics of ampullary electrosensory primary afferent neurons. J. Neurosci. 20, 8586–8595. Slater, C. H., and Schreck, C. B. (1993). Testosterone alters the immune response of chinook salmon (Oncorhynchus tshawytcha). Gen. Comp. Endocrinol. 89, 291–298. Slater, C. H., and Schreck, C. B. (1997). Physiological levels of testosterone kill salmonid leucocytes in vitro. Gen. Comp. Endocrinol. 106, 113–119. Slater, C. H., Fitzpatrick, M. S., and Schreck, C. B. (1995). Androgens and immunocompetence in salmonids: Specific binding of and reduced immunocompetence of salmonid lymphocytes exposed to natural and synthetic androgens. Aquaculture 136, 363–370. Smith, E. R., and Mackie, D. M. (1995). ‘‘Social Psychology.’’ Worth Publishers, New York. Sockman, K. W., and Schwabl, H. (2000). Yolk androgens reduce offspring survival. Proc. R. Soc. Lond. B Biol. Sci. 267, 1451–1456. Sohrabji, F., Nordeen, K. W., and Nordeen, E. J. (1989). Projections of androgen-accumulating neurons in a nucleus controlling avian song. Brain Res. 488, 253–259. Specker, J. L., and Kishida, M. (2000). Mouthbrooding in the black-chinned tilapia, Sarotherodon melanotheron (Pisces: Cichlidae): The presence of eggs reduces androgen and estradiol levels during paternal and maternal behavior. Horm. Behav. 38, 44–51. Squire, L. R. (1992). Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychol. Rev. 9, 195–231. Stearns, S. C. (1989). Trade-offs in life-history evolution. Funct. Ecol. 3, 259–268. St. Mary, C. M. (1997). Sequential patterns of sex allocation in simultaneous hermaphrodites: Do we need models that specifically incorporate this complexity? Am. Nat. 150, 73–97. Storey, A. E., Walsh, C. J., Quinton, R. L., and Wynne-Edwards, K. E. (2000). Hormonal correlates of paternal responsiveness in new and expectant fathers. Evol. Hum. Behav. 21, 79–95. Suay, F., Salvador, A., Gonzalez-Bono, E., Sanchis, C., Martinez, M., Martinez-Sanchis, S., Simon, V. M., and Montoro, J. B. (1999). Effects of competition and its outcome on serum testosterone, cortisol and prolactin. Psychoneuroendocrinology 24, 551–566. Taborsky, M. (1984). Broodcare helpers in the cichild fish Lamprologus brichardi: Their costs and benefits. Anim. Behav. 32, 1236–1252. Taborsky, M. (1994). Sneakers, satellites and helpers: Parasitic and cooperative behavior in fish reproduction. Adv. Study Behav. 23, 1–100. Taborsky, M. (2001). The evolution of bourgeois, parasitic, and cooperative reproductive behaviors in fishes. J. Hered. 92, 100–110. Taborsky, M., and Limberger, D. (1981). Helpers in fish. Behav. Ecol. Sociobiol. 8, 143–145.

236

RUI F. OLIVEIRA

Tang, Y. Z., Piao, Y. S., Zhuang, L. Z., and Wang, Z. W. (2001). Expression of androgen receptor mRNA in the brain of Gekko gecko: Implications for understanding the role of androgens in controlling auditory and vocal processes. J. Comp. Neurol. 438, 136–147. Tarlow, E. M., Wikelski, M., and Anderson, D. J. (2001). Hormonal correlates of siblicide in Gala´pagos Nazca boobies. Horm. Behav. 40, 14–20. Temeles, E. J. (1994). The role of neighbours in territorial systems: When are they ‘‘dear enemies’’? Anim. Behav. 47, 339–350. Tinbergen, N. (1951). ‘‘The Study of Instinct.’’ Clarendon Press, Oxford. Tinbergen, N. (1953). ‘‘Social Behavior in Animals.’’ Methuen, London. Tinbergen, N. (1963). On aims and methods in ethology. Z. Tierpsychol. 20, 410–433. Toda, K., Saibara, T., Okada, T., Onishi, S., and Shizuta, Y. (2001). A loss of aggressive behavior and its reinstatement by oestrogen in mice lacking the aromatase gene (Cyp19). J. Endocrinol. 168, 217–220. Tohgi, H., Utsugisawa, K., Yamagata, M., and Yoshimura, M. (1995). Effects of age on mRNA expression of glucocorticoid, thyroid hormone, androgen, and estrogen receptors in postmortem human hippocampus. Brain Res. 700, 245–253. Trainor, B. C., and Marler, C. A. (2001). Testosterone promotes paternal behavior in a monogamous mammal via conversion to oestrogen. Proc. R. Soc. Lond. Biol. Sci. 269, 823–829. Trainor, B. C., and Marler, C. A. (2002). Paternal behavior, aggression, and testosterone in the California mouse, Peromyscus californicus. Horm. Behav. 40, 32–42. Trewavas, E. (1983). ‘‘Tilapiine fishes of the genera Sarotherodon, Oreochromis and Danakilia.’’ British Museum (Natural History), London. Trut, L. N. (1999). Early canid domestication: The farm fox experiment. Am. Sci. 87, 160–169. Tsai, S.-S. H. (1996). ‘‘The Eunuchs in the Ming Dynasty.’’ SUNY Press, Albany, NY. Uglem, I., Galloway, T. F., Rosenqvist, G., and Folstad, I. (2001). Male dimorphism, sperm traits and immunology in the corkwing wrasse (Symphodus melops L.). Behav. Ecol. Sociobiol. 50, 511–518. Uglem, I., Mayer, I., and Rosenqvist, G. (2002). Variation in plasma steroids and reproductive traits in dimorphic males of corkwing wrasse (Symphodus melops L.). Horm. Behav. 41, 396–404. Vale, J. R., Lee, C. T., and Ray, D. (1970). The effect of defeat upon the reproductive behavior and upon adrenal, testis and seminal vesical weights of C57BL/6 mice. Behav. Biol. 5, 225–236. Vale, J. R., Vale, C. A., and Harley, J. P. (1971). Interaction of genotype and population number with regard to aggressive behavior, social grooming and adrenal and gonadal weight in male mice. Behav. Biol. 6, 209–221. van der Lee, S., and Boot, L. M. (1955). Spontaneous pseudopregancy in mice. Acta Physiol. Pharmacol. Neer. 4, 442–443. Vazquez-Pereyra, F., Rivas-Aranciba, S., Loaeza-Del Castillo, A., and Schneider-Rivas, S. (1995). Modulation of short term and long term memory by steroid sexual hormones. Life Sci. 56, 255–260. Verhulst, S., Dieleman, S. J., and Parmentier, H. K. (1999). A tradeoff between immunocompetence and sexual ornamentation in domestic fowl. Proc. Natl. Acad. Sci. USA 96, 4478–4481. Wagner, J. D., Flinn, M. V., and England, B. G. (2002). Hormonal responses to competition among male coalitions. Evol. Hum. Behav. 23, 437–442. Wallen, K. (1999). Risky business: Social context and hormonal modulation of primate sexual desire. In ‘‘Reproduction in Context’’ (K. Wallen and J. Schneider, Eds.), pp. 289–323. MIT Press, Cambridge, MA.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

237

Wallen, K. (2001). Sex and context: Hormones and primate sexual motivation. Horm. Behav. 40, 339–357. Wang, Z., and De Vries, G. J. (1993). Testosterone effects on paternal behavior and vasopressin immunoreactive projections in prairie voles (Microtus ochrogaster). Brain Res. 631, 156–160. Wapler-Leong, D. C. Y., and Reinboth, R. (1974). The influence of androgenic hormone on the behavior of Haplochromis burtoni (Cichlidae). Fortschr. Zool. 22, 334–339. Warner, R. R. (1975). The adaptive significance of sequential hermaphroditism in animals. Am. Nat. 109, 61–81. Warner, R. R. (1984). Mating behavior and hermaphroditism in coral reef fishes. Am. Sci. 72, 128–136. Wedekind, C., and Folstad, I. (1994). Adaptive and non-adaptive immunosupression by sex hormones. Am. Nat. 143, 936–938. West-Eberhard, M. J. (1989). Phenotypic plasticity and the origins of diversity. Annu. Rev. Ecol. Syst. 20, 249–278. Wetzel, D. M., Haerter, U. L., and Kelley, D. B. (1985). A proposed neural pathway for vocalization in South African frogs, Xenopus laevis. J. Comp. Physiol. A. 157, 749–761. White, G. F., Katz, J., and Scarborough, K. E. (1993). The impact of professional football games upon violent assaults on women. Violence Victims 7, 157–171. White, S. A., Nguyen, T., and Fernald, R. D. (2002). Social regulation of gonadotropinreleasing hormone. J. Exp. Biol. 205, 2567–2581. Whiting, M. J. (1999). When to be neighbourly: Differential agonistic responses in the lizard Platysaurus broadleyi. Behav. Ecol. Sociobiol. 46, 210–214. Whitten, W. K. (1956). Modification of the oestrus cycle of the mouse by external stimuli associated with the male. J. Endocrinol. 13, 399–404. Whittingham, L. A., and Schwabl, H. (2002). Maternal testosterone in tree swallow eggs varies with female aggression. Anim. Behav. 63, 63–67. Wickham, L. A., Gao, J., Toda, I., Rocha, E. M., Ono, M., and Sullivan, D. A. (2000). Identification of androgen, estrogen and progesterone receptor mRNAs in the eye. Acta Ophthalmol. Scand. 78, 146–153. Wickler, W. (1969). Zur Soziologie des Brabantbuntbarsches, Tropheus moorei (Pisces, Cichlidae). Z. Tierpsychol. 26, 966–987. Wikelski, M., Hau, M., and Wingfield, J. C. (1999a). Social instability increases plasma testosterone in a year-round territorial neotropical bird. Proc. R. Soc. Lond. B Biol. Sci. 266, 1–6. Wikelski, M., Lynn, S., Breumer, C., Wingfield, J. C., and Kenagy, G. J. (1999b). Energy metabolism, testosterone and corticosterone in white-crowned sparrows. J. Comp. Physiol. A 185, 463–470. Wilczynski, W., Allison, J. D., and Marler, C. A. (1993). Sensory pathways linking social and environmental cues to endocrine control regions of amphibian forebrains. Brain Behav. Evol. 42, 252–264. Wilson, J., Kuehn, R., and Beach, F. (1963). Modifications in the sexual behavior of male rats produced by changing the stimulus female. J. Comp. Physiol. Psychol. 56, 636–644. Winberg, S., and Lepage, O. (1998). Elevation of brain 5-HT activity, POMC expression, and plasma cortisol in socially subordinate rainbow trout. Am. J. Physiol. 274, R645–R654. Winberg, S., and Nilsson, G. E. (1993a). Roles of brain monoamine neurotransmitters in agonistic behavior and stress reactions, with particular reference to fish. Comp. Biochem. Physiol. C. 10, 597–614.

238

RUI F. OLIVEIRA

Winberg, S., and Nilsson, G. E. (1993b). Time course of changes in brain serotonergic activity and brain tryptophan levels in dominant and subordinate juvenile Arctic charr. J. Exp. Biol. 179, 181–195. Winberg, S., Winberg, Y., and Fernald, R. D. (1997). Effect of social rank on brain monoaminergic activity in a cichlid fish. Brain Behav. Evol. 49, 230–236. Wingfield, J. C. (1984a). Environmental and endocrine control of reproduction in the song sparrow, Melospiza melodia. I. Temporal organization of the breeding cycle. Gen. Comp. Endocrinol. 56, 406–416. Wingfield, J. C. (1984b). Environmental and endocrine control of reproduction in the song sparrow, Melospiza melodia. II. Agonistic interactions as environmental information stimulating secretion of testosterone. Gen. Comp. Endocrinol. 56, 417–424. Wingfield, J. C. (1984c). Androgens and mating systems: Testosterone-induced polygyny in normally monogamous birds. Auk 101, 665–671. Wingfield, J. C. (1985). Short-term changes in plasma levels of hormones during the establishment and defense of a breeding territory in male song sparrows, Melospiza melodia. Horm. Behav. 19, 174–187. Wingfield, J. C., and Farner, D. S. (1978a). The endocrinology of a naturally breeding population of the white-crowned sparrow, Zonotrichia leucophrys pugetensis. Physiol. Zool. 51, 188–205. Wingfield, J. C., and Farner, D. S. (1978b). The annual cycle in plasma irLH and steroid hormones in feral populations of the white-crowned sparrow, Zonotrichia leucophrys gambelli. Biol. Reprod. 19, 1046–1056. Wingfield, J. C., Ball, G. F., Dufty, A. M., Hegner, R. E., and Ramenofsky, M. (1987). Testosterone and agression in birds: Tests of the challenge hypothesis. Am. Sci. 75, 602–608. Wingfield, J. C., Hegner, R. E., Dufty, A. M., and Ball, G. F. (1990). The ‘‘challenge hypothesis’’: Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies Am. Nat. 136, 829–846. Wingfield, J. C., Jacobs, J., and Hillgarth, N. (1997). Ecological constraint and the evolution of hormone–behavior interrelationships. Ann. N.Y. Acad. Sci. 807, 22–41. Wingfield, J. C., Jacobs, J. D., Soma, K., Maney, D. L., Hunt, K., Wisti-Peterson, D., Meddle, S., Ramenofsky, M., and Sullivan, K. (1999). Testosterone, aggression, and communication: ecological bases of endocrine phenomena. In ‘‘The Design of Animal Communication’’ (M. D. Hauser and M. Konishi, Eds.), pp. 257–283. MIT Press, Cambridge, MA. Wingfield, J. C., Jacobs, J. D., Tramontin, A. D., Perfito, N., Meddle, S., Manney, D. L., and Soma, K. (2000). Toward an ecological basis of hormone–behavior interactions in reproduction of birds. In ‘‘Reproduction in Context’’ (K. Wallen and J. E. Schneider, Eds.), pp. 85–128. MIT Press, Cambridge, MA. Wingfield, J. C., Lynn, S. E., and Soma, K. K. (2001). Avoiding the ‘‘costs’’ of testosterone: Ecological bases of hormone–behavior interactions Brain Behav. Evol. 57, 239–251. Wyndham, D. (2003). Versemaking and lovemaking—W. B. Yeats’ ‘‘strange second puberty’’: Norman Haire and the Steinach rejuvenation operation J. Hist. Behav. Sci. 39, 25–50. Yamamoto, K., Toyoda, F., Tanaka, S., Hayashi, H., and Kikuyama, S. (1996). Radioimmunoassay of a newt sex pheromone, sodefrin, and the influence of hormones on its level in the abdominal gland. Gen. Comp. Endocrinol. 104, 356–363. Ydenberg, R. C., Giraldeau, L. A., and Falls, J. B. (1988). Neighbours, strangers and the asymmetric war of attrition. Anim. Behav. 36, 343–347. Zahavi, A. (1975). Mate selection: A selection for a handicap. J. Theor. Biol. 53, 205–214.

SOCIAL MODULATION OF ANDROGENS IN VERTEBRATES

239

Zayan, R. (1975). Modification des effets lie´s a` la priorite´ de re´sidence chez Xiphophorus (Pisces, Poecilidae): Le roˆle de l’expe´rience imme´diate de dominance et de soumission. Rev. Compt. Anim. 8, 296–311. Ziegler, T. E. (2000). Hormones associated with non-maternal infant care: A review of mammalian and avian studies. Folia Primatol. 71, 6–21. Ziegler, T. E., and Snowdon, C. T. (2000). Preparental hormone levels and parenting experience in male cotton-top tamarins, Saguinus oedipus. Horm. Behav. 30, 59–167. Ziegler, T. E., Wegner, F. H., and Snowdon, C. T. (1996). Hormonal responses to parental and non-parental conditions in male cotton-top tamarins, Saguinus oedipus, a New World primate. Horm. Behav. 30, 287–297.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Odor Processing in Honeybees: Is the Whole Equal to, More Than, or Different from the Sum of Its Parts? Harald Lachnit,* Martin Giurfa,{ and Randolf Menzel{ *department of psychology philipps-university marburg 35032 marburg, germany { centre de recherches sur la cognition animale cnrs, paul-sabatier-university, umr 5169 31062 toulouse cedex 4, france { department of neurobiology free university of berlin 14195 berlin, germany

I. Introduction Processing and learning compound stimuli is an important biological problem.1 Every organism, in nearly every environment, is faced with a continual, multifaceted stream of stimuli that may be related to specific effects or outcomes. How do animals process compound stimuli under such circumstances? Do they treat the components separately and associate them with a common outcome? Do they treat the compound stimulus as an entity that is drastically different from its components and associate this entity with a specific outcome? These questions reveal useful ways of thinking about how animals may process and respond to a compound stimulus, and they correspond to the following general question: What mechanisms govern the integration of that stream of events and how are these events represented in the nervous system? The answer to these 1 We use the term ‘‘compound stimulus’’ in the sense of a stimulus composed of two or more elements without making assumptions about the way in which the nervous system treats such a stimulus. The term nevertheless refers to a psychological perspective and thus to the question of how this kind of stimulus is processed.

241 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

242

HARALD LACHNIT et al.

questions may be provided by a combined approach in which Pavlovian conditioning, one of the oldest and most systematically studied phenomena in psychology, and neurobiological investigations on stimulus processing in model systems converge into a common strategy. In the present review, we describe studies of compound stimulus processing and learning that have focused on olfactory learning in an invertebrate, the honeybee Apis mellifera. Honeybees are a traditional model for studying learning and memory (Giurfa, 2003; Menzel, 1999, 2001; Menzel and Giurfa, 2001). Olfactory learning in bees has been well characterized and follows the Pavlovian conditioning scheme (Bitterman et al., 1983). Here we describe a series of conditioning experiments and derive conclusions on how bees treat olfactory compound stimuli. Since the neural pathways underlying olfactory learning in the bee brain are relatively well known (Menzel, 1999, 2001), we are able to interpret our findings from a neurobiological perspective in which we correlate our behavioral findings with findings on neural coding of olfactory compound stimuli in the olfactory neuropils of the bee brain. Our review may be taken as an advertising campaign for interdisciplinary studies in which approaches from experimental psychology (Pavlovian conditioning as a research tool in behavioral biology) and neurobiology (optophysiological studies on olfactory coding as a research tool in sensory neuroscience) are combined.

II. Pavlovian Conditioning and Models of Compound Stimulus Processing As early as 1897, the Russian physiologist Pavlov noted that, in dogs, stimuli preceding the appearance of food (e.g., the smell of food) elicited what he called ‘‘psychic’’ salivary secretion (Pavlov, 1902), or a conditional reflex. His empirical and theoretical elaboration of these ‘‘psychic’’ secretions and of cortical function founded a scientific endeavor distinguished by objective terminology, well-defined procedures, and empirically driven theoretical concepts (for an overview see Gormezano and Moore, 1969). Pavlov, working within the reflex tradition of physiology and early behaviorism, saw conditioning as a kind of low-level mechanical process in which control over responding is passed from one stimulus to another. However, in the late 1960s and early 1970s, after the ‘‘cognitive revolution,’’ conceptualization and knowledge of the associative processes underlying Pavlovian conditioning changed and expanded dramatically (see, e.g., Rescorla, 1988). Nowadays, Pavlovian conditioning deals with the mechanisms that enable organisms to represent the structure of their world, especially the relationships between environmental events. This new view may be traced

ODOR PROCESSING IN HONEYBEES

243

back to the associative tradition originating in philosophy (see Lachnit, 2003). This combination of properties, especially the fact that research in Pavlovian conditioning (1) allows for precise manipulations of independent variables and is primarily driven by (2) experimental designs and (3) explicit—in part formalized—theories, encouraged us to take advantage of this approach in the study of olfactory processing in honeybees. Let us now focus on the initial question from a theoretical viewpoint, namely, which mechanisms govern the integration of the stream of incoming events in compound stimulus processing and how are these events represented in the neural system? Three main positions should be distinguished, and these also lay the foundation for rival theories of associative learning. First, an elemental perspective, assuming that the whole (compound stimulus) equals the sum of its parts (elements). Second, the idea that the whole is more than the sum of its parts. And last, but not least, the assumption that the whole is (totally) different from the sum of its parts, a point of view that is typical of the so-called configural approaches. In Pavlovian conditioning many well-known effects have long been successfully explained in an elemental manner (Wagner, 1971), assuming that stimulus components are represented as separate entities and that the overall associative strength of a compound stimulus is based on the algebraic sum of the associative strength of its components (Rescorla and Wagner, 1972). If a compound stimulus AB consisting of the elements A and B is paired with an unconditioned stimulus (US), it is assumed that each element enters into a separate association with the US. This elemental summation principle, incorporated into many theories of associative learning (e.g., Mackintosh, 1975; Pearce and Hall, 1980; Rescorla and Wagner, 1972; Wagner, 1981), has important consequences. On the one hand, it may serve as a convenient but nevertheless powerful modular design principle for dealing with complex environmental events. The total associative strength of such an event is distributed among several components. In some situations this strength may be equally distributed, in others rather unequally, among the different components. On the other hand, when two separately trained stimuli are presented together, response to this compound stimulus will be more pronounced than to either element alone, a phenomenon that is called summation (Rescorla and Wagner, 1972). There is, however, considerable evidence that animals are able to successfully handle discrimination problems that cannot be solved in a purely elemental manner. In such nonlinear discriminations, the associative strength of a compound stimulus does not result from the simple sum of the elemental associative strengths. An example of this kind of discrimination is negative patterning (e.g., Rescorla, 1972; Whitlow and Wagner, 1972). In negative patterning, two stimuli are always reinforced when they are

244

HARALD LACHNIT et al.

presented on their own (Aþ, Bþ), but never when they are presented together as a compound stimulus (AB–). Another nonlinear discrimination problem is biconditional discrimination (Saavedra, 1975). In this discrimination problem, animals learn to discriminate two reinforced (ABþ, CDþ) from two nonreinforced (BC–, AD–) compound stimuli consisting of pairs of two out of four elements A, B, C, and D. In both discrimination problems, each element is reinforced and nonreinforced equally often so that the elemental summation principle incorrectly predicts intermediate levels of response to each compound stimulus. Animals are nevertheless able to both show anticipatory response to stimuli that are followed by an outcome, and to refrain from responding to stimuli that are not followed by an outcome. Solving these discrimination tasks implies that nonelemental processing of compound stimuli is possible. One of the simplest ways in which associative theories could allow for this would be to assume that stimulus components are always encoded elementally but that, in addition, configurations of those components may also be encoded. Such configural units might be treated as additional ‘‘elements’’ (Rescorla, 1972, 1973; Whitlow and Wagner, 1972). In other words, the joint presentation of elements in a compound stimulus would result in an additional unique cue, which would be specific to those elements and that could by itself be associated with a specific outcome. According to such a view, the overall associative strength of a compound stimulus (e.g., AB) is based on the summed associative strength of its elements plus the associative strength of the unique cue. By postulating the existence of such an additional unique cue, elemental models can successfully handle the acquisition of discrimination problems such as negative patterning or biconditional discrimination. Negative patterning (Aþ, Bþ, AB–), for example, can be explained by assuming that A and B will both develop excitatory associations to the outcome and that their unique cue will develop an inhibitory association with this outcome. When A or B is presented alone, each will be able to activate a representation of the outcome so that a conditioned response will be elicited. When A and B are presented together, however, their joint presence will activate the unique cue, which then compensates for the excitatory properties of the elements due to its inhibitory associative status. The unique cue model can be viewed as an extension of elemental models, one that allows for quite a large degree of nonelemental processing, but at the same time keeps the gist of elemental models, that is, the summation of elements. In this view, the whole is more than the sum of its parts. Configural models of compound stimulus processing propose another view of nonelemental processing. Some authors have suggested that compound stimuli are always processed configurally. The strongest view

ODOR PROCESSING IN HONEYBEES

245

on how such a configural representation might be formed is that a compound stimulus is an entity that is distinct from its elements and that it is this entity alone that enters into an association with the outcome. This means that a compound stimulus that consists of two components, say A and B, is processed and represented with no relation to its components but instead as an independent configuration, which we shall call G. This would be a purely configural view (for a review see Kehoe and Gormezano, 1980). A less extreme, and in the area of associative learning empirically far more successful position, however, assumes that stimuli are processed configurally but that generalization takes place between the compound stimulus and the elements. According to the most prominent configural theory that has been put forward (Pearce, 1987, 1994, 2002), the degree of generalization between two stimuli is based on their component similarity, that is, on the number of elements they share. Thereby, part of the associative strength of a compound stimulus generalizes to its elements and vice versa (e.g., there will be reciprocal generalization between AB and A as well as between AB and B, but not between A and B). In the following section we illustrate how it is possible to decide between these different models by choosing a model system, the honeybee Apis mellifera, which has allowed us to design Pavlovian conditioning experiments using olfactory compound stimuli.

III. Olfactory Pavlovian Conditioning and Olfactory Compound Stimulus Processing in the Honeybee Honeybees can learn to associate odor stimuli with a sucrose solution reward. Such an association is biologically relevant, since it allows the bee to learn that a particular flower species, characterized by a specific odor, is profitable and can be exploited because it yields a nectar reward essential for the survival of the individual and the colony as a whole. Olfactory learning in honeybees can be studied and reproduced in the laboratory under controlled conditions by using the paradigm of olfactory conditioning of the proboscis extension reflex (PER) (Bitterman et al., 1983; Takeda, 1961). In this paradigm, hungry, harnessed honeybees are presented with olfactory stimuli associated with a reward of sucrose solution (see Fig. 1). When the antennae of a hungry bee are touched with sucrose solution, the animal reflexively extends its proboscis to reach out to and suck the sucrose. Odors to the antennae do not usually release such a reflex in naive animals. If an odor is presented immediately before sucrose solution (forward pairing), an association is formed such that the mere presentation of the odor will subsequently release the PER in a following test. This

246

HARALD LACHNIT et al.

Fig. 1. Classic conditioning of the proboscis extension reflex (PER) in the honeybee. When the antennae of a hungry bee harnessed in a metal tube are touched with sucrose solution (toothpick, top left), the animal will reflexively extend its proboscis to reach out toward the sucrose and suck it. Odors or other stimuli to the antennae do not release such a reflex in naive animals (odor, bottom left). If, however, an odor is presented immediately before sucrose solution (forward pairing), an association is formed that enables the odor to release the proboscis extension response (PER) in a successive test (middle). The acquisition curves (right) show a typical differential conditioning experiment: one odor is paired with sucrose (CSþ) and the other odor is presented unpaired (CS–) between CSþ trials. The bees learn to respond to the CSþ and not to the CS–.

effect is clearly associative and involves classic conditioning (Bitterman et al., 1983). Thus the odor can be viewed as the conditioned stimulus (CS) and sucrose solution as an excitatory unconditioned stimulus (US). Using this paradigm it is possible to ask how bees process and learn olfactory compound stimuli, odor mixtures2 (Chandra and Smith, 1998; Deisig et al., 2001, 2002, 2003). Early results indicated that bees could solve nonelemental olfactory discriminations such as negative patterning (Deisig et al., 2001) and biconditional discrimination (Chandra and Smith, 1998; Hellstern et al., 1995). This finding allows us to dismiss a purely elemental approach as an explanation for odor processing in honeybees. In the case of negative patterning, for instance, an elemental approach predicts that if the odors A and B were rewarded, bees should react considerably more to the mixture stimulus than to each of the elements. Thus the fact that bees learn to extend their proboscis to the single odors but not the compound stimulus (see Fig. 2) cannot be explained from this perspective. In further 2

For such stimuli we use the term ‘‘odor mixture,’’ thus acknowledging their unimodal nature. The term ‘‘mixture’’ refers to a physical perspective and has no implications for the question of how this kind of stimulus is treated by the nervous system.

ODOR PROCESSING IN HONEYBEES

247

Fig. 2. Result of a negative patterning experiment. Bees were trained with odors A and B and with their mixture AB given according to a 6 Aþ, 6 Bþ, and 12 AB– pseudo-randomized order. Acquisition curves for the single odors Aþ and Bþ and for the mixture AB– are shown. In the case of responses to AB– data two consecutive trials were averaged to obtain 6 blocks of CS– trials. Adapted from Deisig et al. (2001).

experiments dealing with odor processing in honeybees (Deisig et al., 2002) we found a certain amount of generalization between elements and the mixture stimuli, a fact that also allows us to dismiss the extreme configural interpretation of odor processing. Clearly, if a mixture stimulus is an entity that differs drastically from its components, no generalization should occur. In addition, in these experiments we were able to show that a reduction of similarity between a mixture stimulus and the elements enhances the degree to which bees discriminate between both kinds of CS (Deisig et al., 2002). Discrimination was best in a CD–versus Aþ, Bþ task (no sharing between compound stimulus and elements), intermediate in an AC– versus Aþ, Bþ task (one out of two elements shared), and lowest in an AB– versus Aþ, Bþ task (two out of two elements shared). Solving the last task replicates the finding that bees can solve a negative patterning discrimination in the olfactory domain (Deisig et al., 2001) and allows us to again dismiss purely elemental interpretations of compound stimulus processing. Conversely, both the unique cue and Pearce’s configural theories can account for the acquisition of these three tasks. Furthermore, bees trained with the Aþ, Bþ versus AB– discrimination, and then tested with a novel mixture stimulus BC after training, showed a stronger response to BC than to Bþ. This finding is in contradiction to Pearce’s configural theory (for more details, see Deisig et al., 2002), which

248

HARALD LACHNIT et al.

would predict a decrement in responding due to generalization from B to BC. Although our results so far suggest that olfactory mixture stimulus processing in bees relies on the use of unique cues, we reconsidered this suggestion on the basis of two main lines of evidence. On the one hand, results from other species (e.g., pigeons, rabbits) support the idea that—as is the case with odor mixtures—compound stimuli consisting of elements in the same modality favor the emergence of configural rather than of elemental associations (e.g., Pearce and Redhead, 1993; Pearce et al., 1997; Redhead and Pearce, 1995; Rescorla and Coldwell, 1995), whereas compound stimuli consisting of elements in different modalities favor the emergence of elemental associations (e.g., Bahc¸ekapili, 1997; Myers et al., 2001). These findings would therefore predict that olfactory mixtures are treated as a configural entity and not as the sum of their elements plus a given unique cue. On the other hand, in the studies reviewed so far, we neglected the problem that in olfactory PER conditioning of honeybees the conditionability of odorants might be reduced because of interference between the components in an olfactory mixture. Neither the unique cue theory nor Pearce’s configural theory takes into account such interference. Hence, we additionally considered a modification of the unique cue theory initially suggested by Redhead and Pearce (1995). This modified unique cue theory predicts that conditioning a single reinforced element progresses in the manner predicted by the Rescorla–Wagner theory (Rescorla and Wagner, 1972), but in the case of compound stimulus trials it states that the presence of one element restricts learning about the other, thus yielding a slower rate of learning for the compound stimulus. Interestingly, it can be shown that our findings reviewed up to now (Deisig et al., 2001, 2002) are consistent with this modified unique cue theory. Hence, in order to further decide between the unique cue theory, the modified unique cue theory, and Pearce’s configural theory, we confronted the bees with an Aþ, BCþ, ABC– discrimination task, which requires them to learn three different stimuli (single odorants, binary and ternary odor mixture stimuli). Redhead and Pearce (1995) used this task successfully in a study on visual conditioning in pigeons to discriminate between different theoretical possibilities for compound stimulus processing. The task allows one to decide between the rival theories for the following reasons (for further details see Deisig et al., 2003): the unique cue theory expects a better differentiation between BCþ and ABC–than between Aþ and ABC–because of summation of the associative strengths of B and C on binary compound stimulus presentation. Conversely, both Pearce’s configural theory and the modified unique cue theory predict better differentiation between Aþ and ABC– than between BCþ and ABC–. For Pearce’s

ODOR PROCESSING IN HONEYBEES

249

configural theory this is due to the higher degree of generalization between the BCþ and the ABC– compound stimuli (two elements, B and C, in common), which renders this discrimination difficult. For the modified unique cue theory the presence of one stimulus restricts learning about the other in reinforced compound stimulus trials (BCþ) and this results in a slower rate of conditioning for this compound stimulus. Furthermore, the modified unique cue theory and Pearce’s configural theory differ in their prediction of summation at the beginning of training. According to the configural theory no summation should occur on compound stimulus presentation, while the modified unique cue theory predicts summation. Thus, on the basis of empirical results, the Aþ, BCþ, ABC– design enables us to decide between the three theories. Deisig et al. (2003) showed that bees differentiated faster between the reinforced single odor Aþ and the nonreinforced ternary mixture stimulus ABC– than between the reinforced binary mixture stimulus BCþ and the nonreinforced ternary mixture stimulus ABC– (see Fig. 3). Thus, the outcome of the experiment clearly rejected the predictions of the unique cue theory, but was consistent with the predictions of the modified

Fig. 3. Conditioned proboscis extension response (% PER) during trials in an Aþ, BCþ versus ABC– discrimination. The curve depicts the course of percent PER to the reinforced single odor Aþ (d), the reinforced binary mixture BCþ (j), and the nonreinforced ternary mixture ABC– () during acquisition along six blocks of training. For the single odor and the binary mixture, each block consists of one trial each, whereas for the ternary mixture, each block consists of the average of two consecutive trials. Half the training took place on day 1 (blocks 1 and 2), the other half (block 3 and 4) on day 2. Adapted from Deisig et al. (2003).

250

HARALD LACHNIT et al.

unique cue and Pearce’s configural theory. Focusing on summation at the outset of training, in the first block of training we found that responses to ABC– were significantly stronger than those to Aþ and BCþ. This finding clearly supports the modified unique cue theory, but contradicts Pearce’s configural theory (see also Rescorla, 1997). Taken together, all of our results reported so far are in accordance with a modified unique cue approach, which incorporates a correction for the reduced salience of components in a compound stimulus (in our case odor mixture stimuli) due to interference between them (James and Wagner, 1980; Mackintosh, 1971). Such interference might be related to the fact that the associative strength is limited such that all CS must compete for it (Revusky, 1971) or it might be related to limited attentional capacity. Paying attention to one CS will decrease the attention to, and learning about, another stimulus (e.g., Sutherland and Mackintosh, 1971). Last but not least, interference could occur at the sensory and neural level (e.g., odors competing for the same receptor type, or at a more central level, odors eliciting competing neural representations) or may be due to a storage problem (e.g., stimuli might compete for the maintenance of information in short-term memory even if they are easily perceived independently; see Atkinson and Shiffrin, 1968; Norman, 1968; Wagner, 1976).

IV. Physiological Correlates of Odor Processing and Element/Compound Interactions In general, odors as used in the studies described above are received by chemoreceptors, in the bee antennae, that are localized in groups of 20–30 within a particular type of sensilla (sensilla placodea). Several hundred of such placodes are distributed over each antenna, and are particularly dense at its tip. The receptors have broad and overlapping response profiles, indicating that odor identity must be extracted from the combination of multiple parallel receptor inputs (Akers and Getz, 1993; Getz and Akers, 1994; Vareschi, 1971). Stronger odors elicit higher spike activity in each receptor cell and activate more receptor cells. However, odors also compete at the peripheral level for receptor sites or inhibit each other—possibly via different receptor molecules connected to antagonizing second-messenger pathways. Extracellular recordings from honeybee placodes revealed in some cases only the same or even fewer responses to binary olfactory mixtures as compared with single odorants (Getz and Akers, 1995). Model calculations indicated that antagonistic secondmessenger pathways may lead to such competition and inhibition in the chemoreceptors (Malaka et al., 1995). However, the assumption that

ODOR PROCESSING IN HONEYBEES

251

one chemoreceptor expresses more than one chemoreceptor molecule is not in accordance with the current view as derived from molecular genetic results in Drosophila (Vosshall et al., 1999). In any case, it is possible that already in the periphery both cooperative and antagonizing phenomena may lead to a deviation from simple additive superposition of odor-induced activity when olfactory mixtures are used. Further studies are needed to determine to what extent nonlinearity and interference occur at a peripheral level in the case of olfactory stimulus compounds. At the central level, olfactory information is first processed in the antennal lobe, the primary olfactory neuropil of the insect brain. Odors are coded here by multiple activity patterns of glomeruli, the structural elements of the antennal lobe. Each of the 160 glomeruli in the bee antennal lobe is likely to receive input from only one receptor type. This assumption is made on the basis of findings in Drosophila, where molecular genetic experiments revealed that similar receptor genes are expressed in the receptor axons reaching the same glomerulus (Vosshall et al., 1999). Thus the antennal lobe can be considered as a landscape with receptor-specific locations (glomeruli) that are activated in combinations according to the rules defined by the overlap of the respective response profiles and the peripheral processing (Joerges et al., 1997). Intense synaptic processing occurs within and between the glomeruli. Local interneurons (approximately 4000) that are predominantly inhibitory, and projection neurons (approximately 800) conveying information to the higher order processing centers, the mushroom bodies, are the structural components in this network. Intracellular recordings from these elements revealed complex response patterns indicative of nonlinear processing that sharpens the response profiles, makes them less dependent on odor intensity, and leads to mixture-specific activation patterns of the output neurons (projection neurons) (Abel et al., 2001; Mu¨ller et al., 2002; Sachse and Galizia, 2002). These mixture-specific activations are of concern here because the rules behind the combination of activity patterns induced by the elements may govern how odor mixtures are coded and learned. The working of the antennal lobe network can be studied most effectively by optical imaging techniques. In these studies Ca2þ-sensitive dyes are infused into either the whole network or selectively into specific components (e.g., single or multiple projection neurons), and neuronal activity is read from the Ca2þ-dependent fluorescence signal when the antennae are stimulated with odor (Galizia et al., 1997; Joerges et al., 1997). Since an atlas exists for the bee antennal lobe, thus allowing identification of each glomerulus in recordings from different animals (Galizia et al., 1999), it is possible to establish the combinatorial physiological activity code for a

252

HARALD LACHNIT et al.

large number of odors (Sachse et al., 1999) and their mixtures (Joerges et al., 1997; see Fig. 4). For the question considered in this article, it is relevant to know how olfactory mixtures are encoded at the level of the antennal lobe and to relate these results to the rivals theories on compound processing. An important finding in this context is that activity patterns induced by mixtures of odors deviate more or less from the patterns that would be expected if simple linear superposition of the patterns elicited by the single odors governed neural processing (Joerges et al., 1997). Both the kind and the number of odors mixed together appear to play a role. Binary mixtures, for example, sometimes evoke patterns of glomerular activation that correspond to the simple addition of the patterns of the constituent odorants but that sometimes may also differ strongly from the pattern expected from such a linear addition. Ternary or even higher order olfactory mixtures always deviate strongly from the expected pattern (see Fig. 5; Joerges et al., 1997; Sachse and Galizia, 2002). The reason for such odor-specific suppression phenomena lies in the intrinsic organization of the antennal lobe networks. Two inhibitory networks can be found at the level of the local interneurons connecting glomeruli (see Fig. 5): a GABA-A receptor-dependent network and a histaminergic network (Sachse and Galizia, 2002). The first provides a widespread and more general inhibition, whereas the latter would be responsible for more localized specific inhibition. It is thus to be expected that the quality of the single odors—as reflected in their corresponding glomerular activity patterns and the number of odors in a mixture—should influence the recognition and possibly also the learning of the mixture. Nonlinear and inhibitory interactions within the antennal network, therefore, make the olfactory code of the olfactory mixture unique with respect to its constituting odors. Generalization between single odors and the mixture will depend on the amount of overlap between their activity patterns. In concluding this, we have therefore to refer processing of olfactory mixtures in the antennal lobe to the unique cue and the modified unique cue approaches postulated for compound stimulus processing. It is worth recalling at this stage that behavioral experiments supported the latter approach. The second processing stage in the insect olfactory pathway is constituted by the mushroom bodies, which are involved in olfactory learning phenomena, and play a particularly important role in the olfactory memory trace (Menzel, 2001; Menzel and Giurfa, 2001). It is in these structures that the multitude of sensory modalities converge, and where contextdependence, across-modality configuration and comparison with remote memories (both acquired and innate) may be performed (Menzel and

ODOR PROCESSING IN HONEYBEES

253

Fig. 4. Activity pattern of glomeruli in the antennal lobe of the bee when stimulated with 37 different odors. High activity as measured optophysiologically is indicated by red/dark colors. Top left: Schematic representation of the antennal lobe with the numbers of the glomeruli measured with the optophysiological technique (see text). The 13 pictures (top right) show the activity patterns of different olfactory stimuli ranging from single odors, mixtures of two or three compounds, or mixtures composed of many compounds such as in the case of floral odors (Cit, citral; Ger, geraniol; Iso, isoamylacetate; Pfm, peppermint oil; Org, orange; Car, carnation; Lnd, lime blossom, Lim, limonene; Cin, cineol; Eug, eugenol; Lio, linalool; Mnt, menthol; Cio, dl-citronellol. Bottom left: Twenty-four activity patterns showing

254

HARALD LACHNIT et al.

Fig. 5. Two inhibitory networks shape the odor representation of projection neuron (PN) responses. Top: Examples of time courses of PN responses to odors: gray, perfusion with picrotoxin (PTX); black, Ringer control. Bottom: Resulting model for the functional connectivity between olfactory glomeruli. Glomerulus c receives strong receptor neuron (RN) input and inhibits other glomeruli with PTX-insensitive synapses (solid circles). Glomeruli a, b, and e receive weak receptor neuron (RN) input; glomerulus d receives no RN input at all for this odor. All glomeruli feed into a global, PTX-sensitive inhibitory network (gray circles, GABA). Therefore application of PTX leads to an increase in the PN response of weakly activated glomeruli (a) and a prolongation of glomeruli with a strong RN input (c). The tonic increase in intracellular calcium and spontaneous activity due to PTX leads to the calcium decrease becoming more visible; thus the inhibitory response of inhibited glomeruli is enhanced (b), and inhibitory PN responses during odor stimulation are visible in previously nonresponding glomeruli (d), in particular if they are spontaneously active shortly before the stimulus onset. The reduction of type e glomeruli after PTX application may be due to PN desynchronization. Right: A homomeric local interneuron that diffusely innervates between 30 and 100 glomeruli. These local interneurons are immunoreactive to an antibody against GABA. Adapted from Sachse and Galizia (2002).

Giurfa, 2001). Odor learning leads to synaptic plasticity both at the input and the output side of this neuropil (Mauelshagen, 1993). At the input side the information from the antennal lobe, via the projection neurons, reaches the intrinsic neurons of the mushroom body, the Kenyon cells. Since each projection neuron contacts many Kenyon cells, the combinatorial activity representing an odor is distributed over a large part of each mushroom body. Indeed, imaging experiments document large fields of activation in the activity patterns of alkanes (first row), primary alcohols (second row), aldehydes (third row), and secondary ketones (fourth row). The number of carbon atoms in each of these four chemical classes is indicated above each column (C-5 to C-10). Bottom right: Images as they appear in the microscope for the four respective odors in the C-10 column. Adapted from data published by Sachse et al. (1999) and Galizia et al. (1999). (See Color Insert.)

ODOR PROCESSING IN HONEYBEES

255

Kenyon cells with complex patterns of thousands of small activation spots (called microglomeruli) (Faber and Menzel, 2001; Szyszka et al., 2002). These patterns are odor specific, representing another odor-coding spatial code, but one now based on a large number of microglomeruli. At the output side of the mushroom body, neurons connect the mushroom body with descending neurons acting as premotor neurons, and with neurons that project back to its input side. These recurrent neurons are thought to provide an inhibitory feedback, since they are immunoreactive to GABA (Ganeshina and Menzel, 2001). Both types of output neurons show associative plasticity in odor conditioning experiments (Gru¨newald, 1999; Mauelshagen, 1993).

V. Functional Model of the Olfactory System as a Neural Substrate for Elemental and Compound Processing Next we shall ask whether the results reported above on elementary versus configural learning can be compiled into a functional model of the olfactory system of the honeybee that takes into account these data from the neuroanatomical, electro- and optophysiological measurements. An additional piece of evidence is important in this context, namely the finding that the olfactory memory trace is distributed between the antennal lobes and the mushroom bodies. Each structure is able to form its own associative memory, and they can in turn control learned behavior independently of each other (Hammer and Menzel, 1998). So far, it is not known how these parallel memories may differ. We shall propose here that the memory trace formed in the antennal lobe results only from elemental associations, but that of the mushroom body includes mixture processing as a substrate for configural associations. Support for this view comes from Malun et al. (2002), who showed that an elemental discrimination (Aþ, B) could be learned by bees with a single antenna and with mushroom body lesions (but intact antennal lobe) on the side of the stimulated antenna. In addition, Komischke et al. (2003) found that a positive patterning task (A, B, ABþ) can be learned by a bee with a single antenna, but a negative patterning task only with both antennae. Since there are practically no connections between the two antennal lobes, but strong connections between the two sides of the brain at the level of the mushroom bodies, one may assume that only the stimulation of both antennae involves the mushroom bodies such that their contribution to olfactory coding and learning is normal. Figure 6 summarizes the structure of our model. The odors A and B are the elements of AB, the mixture stimulus used for olfactory stimulation (level I). At level II, these odors are coded in the antennal lobe (AL) in the

256

HARALD LACHNIT et al.

Fig. 6. Model of the olfactory pathway in the honeybee as it may apply for the elemental and configural forms of learning. Seven levels of integration are indicated (right) and correspond to the following processes: I, elementary odors A, B and the mixture AB, a binary compound stimulus of the elements; II, odor coding at the level of the antennal lobe. Odors are represented as combinatorial codes of glomerular activity. Glomeruli 3, 4, and 5 are considered to inhibit glomerulus 2 (shaded gray). This inhibition is uncovered when the compound stimulus AB is used as a stimulus. It is assumed that each glomerulus can be associated with the outcome of the behavior (elemental associations), and will be able to control learned behavior at an elemental level; III, projection neurons (PN) transmit the information from the antennal lobe to the mushroom bodies. Their activity is displayed in the height of the vertical bars. Each odor is represented as an overlapping across-fiber activity pattern; IV, the first neural representation in the form of the PN across-fiber pattern provides the input to the second stage (central processing, Z) of neural integration in the mushroom bodies; V, here the input signals are classified (ZA, ZB, and ZAB). Because of generalization processes the elements A and B will also activate the compound representation. These generalization processes are believed to be adaptive and learning dependent (see text); VI, the classified representations are associated with behavioral outcome; VII, learned behavior is a joint function of the integration processes at both the antennal lobe level (II) and the level of the mushroom body (VI).

form of activities in glomeruli whose combinatorial patterns overlap. For the purpose of illustration, we assume that each olfactory stimulus leads to the activation of five glomeruli. In the case of olfactory mixture stimuli, specific inhibitory interactions between glomeruli exist such that the activation of one glomerulus is suppressed (here any glomerulus 3, 4, and/or 5 inhibits glomerulus 2). Furthermore, a general inhibitory network (see

ODOR PROCESSING IN HONEYBEES

257

above) leads to a normalization of activity for each odor stimulation, be it an element or a mixture stimulus. To account for the fact that the antennal lobe may form elemental associations, we assume that each glomerulus is associated separately with the behavioral outcome and controls behavior (arrow to the right in Fig. 6). Thus odors inducing activity in more glomeruli are more salient, and the overlap of combinatorial activity defines discrimination between odors. This interpretation is supported by the anatomical condition in which the projection neurons have collaterals bypassing the mushroom body and reaching a neural area (the lateral horn) where descending fibers originate that may provide premotor signals. Learned discrimination between the elements A and B is possible but not between the mixture stimulus AB and the elements because of the larger overlap of activity patterns between each element with that of the compound stimulus than between those of the elements. The result of the odor processing in the antennal lobe is transmitted via the projection neurons (PN) in the form of an across-fiber activity pattern that carries the features of the neural representation at the sensory integration level (level IV: first neural representation). These distributed and combinatorial activity patterns (boxed A, B, and AB in Fig. 6) provide the input to the central integration level in the mushroom bodies (Z level). This input carries the information for both the elements and the compound stimulus. At the Z level (mushroom bodies, level V in Fig. 6) the stimuli are categorized and form the second level of representation (ZA, ZB, and ZAB), now including a specific representation of the mixture stimulus AB. Models about the function of the mushroom body assume that the high neuroanatomical divergence between the input (PN) and the intrinsic fibers (Kenyon cells) together with a neural strategy of sparse coding may lead to representations in single neurons or small clusters of few neurons that are selectively activated by very specific combinations of particular odors or odor mixtures at their particular intensities (Heisenberg, 2003; Laurent, 2003). Consequently the mixture stimulus AB will activate not only the central representation of the compound stimulus AB, ZAB, but also in part those of the elements (ZA and ZB). To reduce generalization between the compound stimulus and the elements one may assume specific inhibition between ZAB and the elements’ inputs (boxed A and B in Fig. 6). This specific inhibition can be provided by the recurrent neurons mentioned above that read out the categorization processes within the mushroom body that lead to the unique cue representation of the compound stimulus. Since these neurons are plastic in an associative manner they could make their recurrent inhibition dependent on former experience. As a consequence ZA and ZB will be reduced relative to the activation of ZAB.

258

HARALD LACHNIT et al.

ZA, ZB, and ZAB are associated with their respective outcomes leading to the second and independent olfactory memory trace in the mushroom bodies (see Fig. 6, level VI). This trace now includes a specific code for the unique cue resulting from the mixture stimulus now represented as a compound stimulus, and thus will support nonelemental forms of learning. Optical imaging during olfactory learning indicated in addition that Kenyon cells change their response properties not only to the reinforced stimulus but also to the nonreinforced stimulus (Gru¨newald, 1999). ZA, ZB, and ZAB are thus represented as stored traces in the mushroom body. The amount of cross-talk between the learned ZA, ZB, and ZAB will depend on recurrent inhibition induced by ZAB and thus should depend on schedules of training since generalization strongly depends on the kind of training. As pointed out above, the Aþ, BCþ, ABC– discrimination task allows testing which of the three theories discussed here might apply to olfactory compound stimulus learning in bees. It was found that the data are consistent with a modified unique cue theory (Deisig et al., 2003) because summation was found at the outset of training (i.e., responses to ABC–were significantly stronger than those to Aþ and BCþ in the first block of training; see Fig. 3). One can account for this finding by assuming that, initially during training, behavior is more strongly controlled by the contribution of the antennal lobe, where the summation of elemental associations dominates. Later, when learning progresses and the recurrent inhibitory properties in the mushroom bodies develop, the unique cuecontrolled associations control behavior more strongly. Such a view is supported by the results of experiments with free-flying bees trained with color compound stimuli in which a short training results in elemental associations while longer training results in nonelemental associations (Giurfa et al., 2003). Our combined antennal lobe/mushroom body model captures the features of the modified unique cue model. Additional features that are related more to the behavioral side, such as competition for attention or reduction in salience if elements appear in a compound stimulus, could be included in the model if one assumes some form of limitation in reinforcement per trial. Again one would have to keep in mind that reinforcement is represented both at the level of elemental processing (the antennal lobe) and compound stimulus processing (mushroom body) (Hammer, 1993, 1997). The consequence will be that elemental and configural tasks have different dependencies on such restrictions, and in the course of learning the balance between their contributions in controlling behavior may change.

ODOR PROCESSING IN HONEYBEES

259

VI. Conclusion Formal models of behavioral routines are usually neither motivated by nor aimed at the neural mechanisms that cause these behaviors. The reason for this lies in the complexity of neural structures and mechanisms, and in the fact that formal models capturing essential properties of behavior are not specified a priori to address neural functions. The lower complexity of the insect brain offers the opportunity to build a bridge between such formal models and their potential implementation in the brain. We study Pavlovian olfactory learning in the honeybee under the strict control of conditioning experiments in order to infer the mechanisms by which odor compound stimuli are encoded by the nervous system, and ask whether the configuration of odor processing found can be related to brain mechanisms. Such an attempt is facilitated by the possibility of recording spatially and temporally resolved brain activity in those neural nets that process and learn the odors in the honeybee. Our conditioning experiments are consistent in indicating that odor compound stimuli are treated in the bee brain according to the modified unique cue theory. According to this theory a compound stimulus (in our case a mixture of odors) is represented as the sum of the individual neural representations of the components plus a specific neural event, proper to the simultaneous occurrence of the compound stimulus components. Neurobiological studies of olfactory coding in the brain are in accordance with this theory. Mixture stimuli are not merely represented as the sum of individual neural representations of the elemental stimuli, but instead include specific odor suppression phenomena, proper to the joint occurrence of the mixture stimulus components. With distinct neural representations for the components and the unique cue at hand, the honeybee brain could extract the necessary information for solving elemental and nonelemental (e.g., configural) olfactory discriminations in different parts of its olfactory pathway, the antennal lobe and mushroom body, respectively. So far, combined efforts between psychological and neurobiological studies have led to a model, which presents a possible implementation of the processes involved in elemental and configural learning. This model makes specific predictions that can be tested in further experiments. For example, if elemental learning is indeed the function of the antennal lobes, while nonelemental (e.g., configural) learning requires the extraction and comparison of the neural representations of components and unique cues and occurs upstream of the antennal lobes, at the level of the mushroom bodies, it should be possible to selectively interfere with elemental and nonelemental processing. This research strategy has already provided some hints supporting the model (e.g., Malun et al., 2002). Even more attractive

260

HARALD LACHNIT et al.

is the possibility of changing the nature of the olfactory representations at the antennal lobe level to facilitate or block the resolution of a given olfactory discrimination in elemental or nonelemental terms. For instance, it is conceivable that a given mixture stimulus results in a more or less pronounced suppression phenomenon (glomerular inhibition) that can act as a unique cue, thus making it possible to solve a negative patterning problem. In such circumstances it would be theoretically possible to accentuate or suppress inhibition during mixture stimulus presentation by means of pharmacological experiments and thus to determine the impact of such interference in the choice of elemental or nonelemental discrimination strategies. Functional imaging studies in the course of elemental and configural learning tasks will also allow us to ask how the different memory traces are combined to lead to coherent behavior in retrieval situations and when memory is updated during new learning. We believe that the present article exemplifies a research strategy that proves to be rewarding and mutually enriching, one in which the power of different domains, experimental psychology and neurobiology, is exploited in working toward a common aim: the understanding of compound stimulus processing and learning by the nervous system. Such combined research is possible because the choice of an appropriate model system, the honeybee, allows us to address questions in both domains experimentally. With the results of the experiments at hand, future research must deepen the modeling approach in order to obtain a more precise physiological model of the olfactory circuit capable of accounting for both the behavioral performances measured in this project and other nonrelated behavioral findings. There are fair chances of attaining this goal using the honeybee, a model in which different domains converge and enrich each other in a multidisciplinary and productive way.

VII. Summary We address the question of whether the nervous system treats and learns a sensory compound stimulus as the simple sum of its components, or as an entity different from them by using olfactory discrimination and learning in honeybees. To study elemental and nonelemental forms of learning in honeybees, we use olfactory classic conditioning of the proboscis extension reflex. In our paradigms, bees had to learn to discriminate mixture stimuli from individual components. We find that a modified unique cue model best describes the results, indicating that both the elements of a compound stimulus as well as a unique cue that is specific to the compound stimulus are processed and learned in parallel. In a next step, we review

ODOR PROCESSING IN HONEYBEES

261

findings using optophysiological recordings from the antennal lobe of the bee. Special emphasis is given to the question of how the combinatorial activation pattern of glomeruli induced by an olfactory mixture stimulus corresponds to the simple sum of the activation patterns elicited by the single components of the mixture stimulus. We show that the local inhibitory network of the antennal lobe leads to nonlinear summation and suppression effects such that the neural representation of a mixture stimulus includes the representations of the components but also odor-specific inhibitory phenomena, which may correspond to the unique cue. Furthermore, it is noted that olfactory memory resides in both the antennal lobe and the mushroom bodies, the second-order neuropil. We compile these results in a model of olfactory processing and learning that assumes associative learning to be implemented both in the antennal lobe and the mushroom body. The model predicts that elemental forms of learning dominate associative processes in the antennal lobe and nonelemental forms of learning in the mushroom bodies. Preliminary data support these predictions. The aim of this article is to demonstrate the applicability of physiological interpretations of behavioral categories for a less complex nervous system. In the case of the honeybee, our original question on compound stimulus learning and perception can be studied at both psychological and neural levels in order to decide between different models of compound stimulus processing. Acknowledgments Preparation of this article was supported by Grants LA 564/10-1, LA 564/10-3, Me 365/23-1, and Me 365/23-3 from the German Research Foundation (Deutsche Forschungsgemeinschaft) to Harald Lachnit and Randolf Menzel. Martin Giurfa received support from the Programme Action Cognitique and the ACI Computational and Integrative Neurosciences of the French Research Ministry, the European Union’s Human Frontiers in Science Programme (Young Investigator Award), the Institut Universitaire de France, the Re´gion Midi Pyre´ne´es, and the Fondation pour la Recherche Me´dicale. We are grateful for most constructive discussion with our colleagues Nina Deisig, Monique Gauthier, Jean-Christophe Sandoz, and Bernhard Komischke. References Abel, R., Rybak, J., and Menzel, R. (2001). Structure and response patterns of olfactory interneurons in the honeybee, Apis mellifera. J. Comp. Neurol. 437, 363–383. Akers, R. P., and Getz, W. M. (1993). Response of olfactory receptor neurons in honey bees to odorants and their binary mixtures. J. Comp. Physiol. A. 173, 169–185. Atkinson, R., and Shriffin, R. (1968). Human memory: A proposed system and its control processes. In ‘‘The Psychology of Learning and Motivation: Advances in Research and Theory’’ (K. Spence and J. Spence, Eds.), Vol. 2, pp. 89–195. Academic Press, New York.

262

HARALD LACHNIT et al.

Bahc¸ekapili, H. G. (1997). An evaluation of Rescorla and Wagner’s elemenistic model versus Pearce’s configural model in discrimination learning. Unpublished doctoral dissertation. Yale University, New Haven, CT. Bitterman, M. E., Menzel, R., Fietz, A., and Scha¨fer, S. (1983). Classical conditioning of proboscis extension in honeybees (Apis mellifera). J. Comp. Psychol. 97, 107–119. Chandra, S., and Smith, B. H. (1998). An analysis of synthetic processing of odor mixtures in the honeybee (Apis mellifera). J. Exp. Biol. 201, 3113–3121. Deisig, N., Lachnit, H., Giurfa, M., and Hellstern, F. (2001). Configural olfactory learning in honeybees: Negative and positive patterning discrimination. Learn. Mem. 8, 70–78. Deisig, N., Lachnit, H., and Giurfa, M. (2002). The effect of similarity between elemental stimuli and compounds in olfactory patterning discriminations. Learn. Mem. 9, 112–121. Deisig, N., Lachnit, H., Sandoz, J.-C., Lober, K., and Giurfa, M. (2003). A modified version of the unique cue theory accounts for olfactory compound processing in honeybees. Learn. Mem. 10, 199–208. Faber, T., and Menzel, R. (2001). Visualizing mushroom body response to a conditioned odor in honeybees. Naturwissenschaften 88, 472–476. Galizia, C. G., Joerges, J., Ku¨ttner, A., Faber, T., and Menzel, R. (1997). A semi-in-vivo preparation for optical recording of the insect brain. J. Neurosci. Methods 76, 61–69. Galizia, C. G., McIlwrath, S. L., and Menzel, R. (1999). A digital 3D atlas of the honeybee antennal lobe based on optical sections acquired using confocal microscopy. Cell Tissue Res. 295, 383–394. Ganeshina, O. T., and Menzel, R. (2001). GABA-immunoreactive neurons in the mushroom bodies of the honeybee: An electron microscopic study. J. Comp. Neurol. 437, 335–349. Getz, W. M., and Akers, R. P. (1994). Honeybee olfactory sensilla behave as integrated processing units. Behav. Neural Biol. 61, 191–195. Getz, W. M., and Akers, R. P. (1995). Partitioning nonlinearities in the response of olfactory neurons to binary odors. Biosystems 34, 27–40. Giurfa, M., Schubert, M., Reisenmann, C., Gerber, B., and Lachnit, H. (2003). The effect of cumulative experience on the use of elemental and configural visual discrimination strategies in honeybees. Behav. Brain Res. 145, 161–169. Gormezano, I., and Moore, J. W. (1969). History and method. In ‘‘Learning Processes’’ (M. M. Marx, Ed.), pp. 121–203. Macmillan, Toronto, Canada. Gru¨newald, B. (1999). Physiological properties and response modulations of mushroom body feedback neurons during olfactory learning in the honeybee Apis mellifera. J. Comp. Physiol. A. 185, 565–576. Hammer, M. (1993). An identified neuron mediates the unconditioned stimulus in associative olfactory learning in honeybees. Nature 366, 59–63. Hammer, M. (1997). The neural basis of associative reward learning in honeybees. Trends Neurosci. 20, 245–252. Hammer, M., and Menzel, R. (1998). Multiple sites of associative odor learning as revealed by local brain microinjections of octopamine in honeybees. Learn. Mem. 5, 146–156. Heisenberg, M. (2003). Mushroom body memoir: From maps to models [Review]. Nat. Rev. Neurosci. 4, 266–275. Hellstern, F., Wu¨stenberg, D., and Hammer, M. (1995). Contextual learning in honeybees under laboratory conditions. In ‘‘Learning and Memory: Proceedings of the 25th Go¨ttingen Neurobiology Conference’’ (N. Elsner and R. Menzel, Eds.), Abstract 30. Thieme, Stuttgart, Germany. James, J. H., and Wagner, A. R. (1980). One-trial overshadowing: Evidence of distributive processing. J. Exp. Psychol. Anim. Behav. Process. 6, 188–205.

ODOR PROCESSING IN HONEYBEES

263

Joerges, J., Ku¨ttner, A., Galizia, C. G., and Menzel, R. (1997). Representation of odours and odour mixtures visualized in the honeybee brain. Nature 387, 285–288. Kehoe, E. J., and Gormezano, I. (1980). Configuration and combination laws in conditioning with compound stimuli. Psychol. Bull. 87, 351–378. Komischke, B., Sandoz, J.-C., Lachnit, H., and Giurfa, M. (2003). Non-elemental processing in olfactory discrimination tasks needs bilateral input in honeybees. Behav. Brain Res. 145, 135–143. Lachnit, H. (2003). The principle of contiguity. In ‘‘Principles of Learning and Memory’’ (R. H. Kluwe, G. Lu¨er, and F. Ro¨sler, Eds.), pp. 3–13. Birkha¨user, Basel, Switzerland. Laurent, G. J. (2003). Olfactory network dynamics and the coding of multidimensional signals [Review]. Nat. Rev. Neurosci. 3, 884–895. Mackintosh, N. J. (1971). An analysis of overshadowing and blocking. Quart. J. Exp. Psychol. 23, 118–125. Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychol. Rev. 82, 276–298. Malaka, R., Ragg, T., and Hammer, M. (1995). Kinetic models of odor transduction implemented as artificial neural networks. Biol. Cybern. 73, 195–207. Malun, D., Giurfa, M., Galizia, G., Plath, N., Brandt, R., Gerber, B., and Eisermann, B. (2002). Hydroxyurea-induced partial mushroom body ablation does not affect acquisition and retention of olfactory differential conditioning in honeybees. J. Neurobiol. 53, 343–360. Mauelshagen, J. (1993). Neural correlates of olfactory learning in an identified neuron in the honey bee brain. J. Neurophysiol. 69, 609–625. Menzel, R. (1999). Memory dynamics in the honeybee. J. Comp. Physiol. A. 185, 323–340. Menzel, R. (2001). Searching for the memory trace in a mini-brain, the honeybee. Learn. Mem. 8, 53–62. Menzel, R., and Giurfa, M. (2001). Cognitive architecture of a mini-brain: The honeybee. Trends Cognit. Sci. 5, 62–71. Mu¨ller, D., Abel, R., Brandt, R., Zo¨ckler, M., and Menzel, R. (2002). Differential parallel processing of olfactory information in the honeybee, Apis mellifera L. J. Comp. Physiol. A. 188, 359–370. Myers, K. M., Vogel, E. H., Shin, J., and Wagner, A. R. (2001). A comparison of the Rescorla– Wagner and Pearce models in a negative patterning and a summation problem. Anim. Learn. Behav. 19, 36–45. Norman, D. A. (1968). Toward a theory of memory and attention. Psychol. Rev. 75, 522–536. Pavlov, I. P. (1902). ‘‘The Work of the Digestive Glands’’ (W. H. Thompson, Trans.), Charles Griffen, London. Pearce, J. M. (1987). A model for stimulus generalization in pavlovian conditioning. Psychol. Rev. 94, 61–73. Pearce, J. M. (1994). Similarity and discrimination: A selective review and a connectionist model. Psychol. Rev. 101, 587–607. Pearce, J. M. (2002). Evaluation and development of a connectionist theory of configural learning. Anim. Learn. Behav. 30, 73–95. Pearce, J. M., and Hall, G. (1980). A model for pavlovian learning: Variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychol. Rev. 87, 532–552. Pearce, J. M., and Redhead, E. S. (1993). The influence of an irrelevant stimulus on two discriminations. J. Exp. Psychol. Anim. Learn. Process 19, 180–190. Pearce, J. M., Aydin, A., and Redhead, E. S. (1997). Configural analysis of summation in autoshaping. J. Exp. Psychol. Anim. Behav. Process 23, 84–94. Redhead, E. S., and Pearce, J. M. (1995). Stimulus salience and negative patterning. Q. J. Exp. Psychol. 48, 67–83.

264

HARALD LACHNIT et al.

Rescorla, R. A. (1972). ‘‘Configural’’ conditioning in discrete-trial bar pressing J. Comp. Physiol. Psychol. 79, 307–317. Rescorla, R. A. (1973). Evidence for unique stimulus account of configural conditioning. J. Comp. Physiol. Psychol. 85, 331–338. Rescorla, R. A. (1988). Pavlovian conditioning: It’s not what you think it is. Am. Psychol. 43, 151–160. Rescorla, R. A. (1997). Summation: Assessment of a configural theory. Anim. Learn. Behav. 25, 200–209. Rescorla, R. A., and Coldwell, S. E. (1995). Summation in autoshaping. Anim. Learn. Behav. 23, 314–326. Rescorla, R. A., and Wagner, A. R. (1972). A theory of pavlovian conditioning: Variations in the effectiveness of reinforcement and non-reinforcement. In ‘‘Classical Conditioning II: Current Research and Theory’’ (A. H. Black and W. F. Prokasy, Eds.), pp. 64–99. Appleton-Century-Crofts, New York. Revusky, S. (1971). The role of interference in association over delay. In ‘‘Animal Memory’’ (W. K. Honig and P. H. R. James, Eds.), Academic Press, New York. Saavedra, M. A. (1975). Pavlovian compound conditioning in the rabbit. Learning and Motivation 6, 314–326. Sachse, S., and Galizia, C. G. (2002). The role of inhibition for temporal and spatial odor representation in olfactory output neurons: A calcium imaging study. J. Neurophysiol. 87, 1106–1117. Sachse, S., Rappert, A., and Galizia, C. G. (1999). The spatial representation of chemical structures in the antennal lobe of honeybees: Steps towards the olfactory code. Eur. J. Neurosci. 11, 3970–3982. Sutherland, N. S., and Mackintosh, N. J. (1971). Mechanisms of animal discrimination learning. Academic Press, New York. Szyszka, P., Galkin, A., Galizia, C. G., and Menzel, R. (2002). Visualizing odor-induced mushroom body activity during differential conditioning in the honey bee, Apis mellifera. Abstract 149 presented at Berlin Neuroscience Forum 2002, Liebenwalde, Germany, April 18–20, 2002. Takeda, K. (1961). Classical conditioned response in the honey bee. J. Insect Physiol. 6, 168–179. Vareschi, E. (1971). Duftunterscheidung bei der Honigbiene: Einzelzell-Ableitungen und Verhaltensreaktionen. Z. Physiol. 75, 143–173. Vosshall, L. B., Amrein, H., Morozov, P. S., Rzhetsky, A., and Axel, D. R. (1999). A spatial map of olfactory receptor expression in the Drosophila antenna. Cell 96, 725–736. Wagner, A. R. (1971). Elementary associations. In ‘‘Essays in Neobehaviorism: A Memorial Volume to Kenneth W. Spence’’ (H. H. Kendler and J. T. Spence, Eds.), pp. 187–213. Appleton-Century-Crofs, New York. Wagner, A. R. (1976). Priming in STM: An information-processing mechanism for selfgenerated or retrieval-generated depression in performance. In ‘‘Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology’’ (T. J. Tighe and R. N. Leaton , Eds.), Erlbaum, Hillsdale, NJ. Wagner, A. R. (1981). SOP: A model of automatic memory processing in animal behavior. In ‘‘Information Processing in Animals: Memory Mechanisms’’ (N. E. Spear and R. R. Miller, Eds.), pp. 5–47. Erlbaum, Hillsdale, NJ. Whitlow, J. W., and Wagner, A. R. (1972). Negative patterning in classical conditioning: Summation of response tendencies to isolable and configural components. Psychonomic Sci. 27, 299–301.

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Begging, Stealing, and Offering: Food Transfer in Nonhuman Primates Gillian R. Brown,*,y Rosamunde E. A. Almond,z and Yfke van Bergeny *school of psychology university of st andrews st. andrews, fife ky16 9ju, united kingdom y sub-department of animal behaviour university of cambridge cambridge cb3 8aa, united kingdom z department of psychology university of wisconsin-madison madison, wisconsin 53706, usa I. Introduction The idea that eating and sharing meat played an important role in human evolution (Isaac, 1978) has led to considerable interest in reports of hunting and meat sharing in chimpanzees (Stanford, 1999). A detailed understanding of food-sharing behavior in chimpanzees and other nonhuman primates could arguably provide insights into the evolutionary history of food sharing in human beings. Although the role of meat eating and meat sharing in human evolution remains a highly debated issue (Stanford and Bunn, 2001), the pervasiveness of food transfer within contemporary human societies is beyond doubt (Gurven, 2004). What benefits might individuals obtain from allowing others to gain access to their food? Several evolutionary explanations have been suggested for the benefits of transferring food to other individuals, including kin selection, reciprocal altruism, and costly signaling. While the literature on food transfer in human beings has been compiled and evaluated recently (Gurven, 2004), the latest review of the subject in nonhuman primates was published more than a decade ago (Feistner and McGrew, 1989), The aim of this review is to bring together theoretical and empirical work on food transfer in nonhuman primates (hereafter referred to as ‘‘primates’’). While food transfer has also been found to occur in a broad range of species, including insects, birds, cetaceans, bats, and other mammals (reviewed by Stevens and Gilby, 2004), such behavior has been particularly well studied in primates. 265 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

266

GILLIAN R. BROWN et al.

Acquiring food captured or harvested by another individual has been described using several different terms, including ‘‘sharing’’ (Feistner and McGrew, 1989), ‘‘scrounging’’ (Barnard and Sibley, 1981), ‘‘kleptoparasitism’’ (Brockmann and Barnard, 1979), and ‘‘tolerated theft’’ (Blurton Jones, 1984, 1987). We use the term ‘‘food transfer’’ to avoid implying either willingness or reluctance on the part of the possessor to relinquish the food item (additional terms are defined in Table I). The transfer of food from one individual to another raises interesting questions regarding the costs and benefits that might be involved in (1) the decision of the food possessor to defend a food item or allow it to be taken, and (2) the decision of the food receiver to obtain a food item from another individual rather than forage independently. As the interests of the food possessor and food receiver may not coincide, the behavior of both individuals must be considered. Here, we first describe transfers among adult primates and investigate whether patterns of food transfer may be explained in terms of trade or reciprocity, or whether alternative explanations provide a better fit to the data. Although adult–adult food transfer in primates has gained considerable attention, the most common form of food transfer in primates appears to occur between mothers and infants. We move on to describe the transfer of solid food to infant primates from parents and alloparents, reviewing the inter- and intraspecific evidence that food transfer reduces time to weaning, increases the infant’s growth rate, or increases the infant’s chance of survival. The transfer of food to infants may ensure that infants (1) receive nutrients during the critical transition to independent foraging (nutritional hypotheses), and/or (2) learn about diet breadth or food-processing techniques (informational hypotheses). The evidence in favor of these hypotheses is discussed, with particular reference to whether adults or infants appear to be controlling patterns of food transfer. The key difference between this type of provisioning and, for example, the feeding of chicks at the nest, is that infants are able to locomote and have some control over which individuals they approach for food and the types of food for which they beg. The situations that we describe perhaps more closely match feeding patterns of fledgling birds. Finally, we investigate whether there is any evidence that adults actively use food transfer to direct offspring learning, and discuss potential directions for future research.

II. Adult–Adult Food Transfer Most interactions over food among adults are likely to be predicted by dominance relationships, with higher ranking individuals taking food items from lower ranking individuals or displacing them from feeding sites. For

TABLE I Definitions of Behavior Patterns Observed during Food Transfer Events Term

Definition

Exemplars

Interest

An individual looks at, touches, or sniffs a food item that is in the possession of another individual

Beg

An individual exhibits specific posture (e.g., extended upturned hand) or specific vocalization while showing interest An individual attempts to take a portion of a food item Any situation in which part or all of the food item changes possession from one individual to another An individual moves into the feeding position vacated by another individual An individual attempts to prevent transfer by moving or turning away, or by vocal or physical aggression or threat Food transfer occurs despite resistance by the possessor

de Waal et al. (1993); Perry and Rose (1994); Wrangham (1975) Feistner and Price (1990); Goodall (1968) Brown et al. (2004) Blurton Jones (1987); Brown and Mack (1978) Nishida and Turner (1996); Yamagiwa (1992) Brown and Mack (1978); Goodall (1986) Hoage (1982); Ruiz-Miranda et al. (1999) Brown and Mack (1978); Feistner and Chamove (1986); Goodall (1968); Hoage (1982) Boesch and Boesch (1989)

Attempted transfer Transfer/provision Displacement Resist Steal Offer

A food possessor passes food to another individual or adopts a specific posture and/or vocalizes

Retrieve

An individual takes food that another individual has dropped on the ground or placed there

268

GILLIAN R. BROWN et al.

low-ranking individuals, relinquishing food could be the least costly strategy in terms of time, energy, and risk of injury (Blurton Jones, 1986; Wrangham, 1975). However, the observation that adult male chimpanzees sometimes allow lower-ranking individuals to take a portion of their meat (Takahata et al., 1984; Teleki, 1973) led to the suggestion that meat transfer involves trade, either for grooming, enhanced alliance partnerships, or sexual interactions (Nishida et al., 1992; Teleki, 1973). In chimpanzee communities, adult males are generally the primary hunters of large prey, such as bush pigs or colobus monkeys (Boesch, 1994; Boesch and Boesch, 1989; Mitani and Watts, 1999; Takahata et al., 1984). Following a hunt, large items are frequently broken up into several pieces, usually involving a large amount of aggressive competition for a piece (Goodall, 1968, 1986; Takahata et al., 1984; Teleki, 1973; Wrangham, 1975). Grooming relationships and alliances during aggressive encounters tend to parallel food transfer relationships in both free-ranging and captive chimpanzee groups, leading to the hypothesis that food was being traded for social support (de Waal, 1989, 1997a; Mitani and Watts, 2001; Nishida et al., 1992). For example, a study of free-ranging adult male chimpanzees found a significant association between the number of times that meat was transferred within a dyad and the number of times that these males engaged in coalitionary support (Mitani and Watts, 2001). Although such findings are consistent with the hypothesis that food is traded for social services, there remains the possibility that food transfer is simply more likely to occur within affiliative relationships, or between closely ranked individuals, and are not causal in maintaining such relationships. Female chimpanzees exhibiting a sexual swelling have been reported to have higher success at begging for meat from males than do other females (Teleki, 1973). In addition, meat transfer has been seen to occur between a male and a female either just before, or soon after, a copulation has occurred, or during a consortship (Goodall, 1986; Nishida et al., 1992; Takahata et al., 1984), leading to the hypothesis that meat was being exchanged for sexual interactions. However, reports of meat-for-sex exchanges are relatively uncommon in the literature and account for only a small proportion of cases of food transfer. Also, studies have reported that male chimpanzees are not more likely to gain matings with females during cycles in which they transfer food compared with cycles in which they do not transfer food (Mitani and Watts, 2001), nor does food transfer appear to increase a male’s chance of siring offspring with a particular female (Hemelrijk et al., 1999). Therefore, there is currently no compelling evidence that male chimpanzees gain greater mating access, or significantly increase their chance of paternity, by transferring meat to females.

FOOD TRANSFER IN NONHUMAN PRIMATES

269

Food transfer has also been hypothesized to increase the likelihood that the recipient will provide food to the original possessor in the future. Whether food transfer involves reciprocal altruism has been specifically investigated in capuchins and tamarins. In captivity, capuchins have been observed to throw food, or push food through wire mesh partition, into a cage containing other capuchins, and to pass food directly to a group mate (de Waal et al., 1993; Westergaard et al., 1998, 1999). Experimental studies, in which pairs of female capuchins were separated by a wire mesh and provided with food at different times, indicated that rates of food exchange were similar in both directions within dyads (de Waal, 1997b, 2000). While the results of these experimental studies may be interpreted as evidence of a well-developed system of exchange or reciprocity, these patterns of food transfer may instead reflect affiliative relationships and tolerance of proximity (de Waal, 1997b). Most exchanges involved one individual picking up items that had been dropped by the other while sitting near the mesh, suggesting that excess food was available. When the food was of a preferred type, the capuchin with food sat further away from the mesh and appeared less willing to allow the other individual to obtain food items (de Waal, 2000). While data on capuchins may be explained most simply as demonstrating tolerated food taking, a study of captive tamarins suggests that these monkeys may have the ability to engage in reciprocal food transfer (Hauser et al., 2003). However, another study of food transfer in captive groups of tamarins failed to find significant correlations between frequencies of food given and food received within dyads (Rapaport, 2001). Cheney and Seyfarth (1990) have postulated that reciprocity in primates is more likely to involve social interactions, such as grooming, than the exchange of material goods, such as food. The possibility of reciprocal food transfer in primates ensures that this will remain an interesting area for future research. While most studies on food transfer among adult primates have focused on the role of trade or reciprocal altruism, the early idea that food transfer may result from the costs imposed by harassment, such that relinquishing food may be the least costly strategy in terms of time, energy, or risk of injury (Blurton Jones, 1986; Wrangham, 1975), has been gaining renewed attention (Caraco and Brown, 1986; Stevens and Stephens, 2002). Stevens and Stephens (2002) produced a game theoretical model of food transfer, which predicted that transfer would be most likely to occur when harassment costs on the owner are high and when the owner cannot easily defend the food. Stevens (2004) went on to test predictions from this model in captive chimpanzees and spider monkeys, and confirmed that food owners transferred food more frequently as begging intensity increased. If two individuals were to be repeatedly involved in bouts of harassment and

270

GILLIAN R. BROWN et al.

transfer, the resulting pattern of food transfer could resemble reciprocal exchange (Brown, 2004; Stevens and Cushman, 2004). Thus, Stevens (2004) points out that, while harassment does not exclude the possibility that other factors may be involved in food transfer, harassment must either be ruled out, or statistically controlled for, before more complex explanations such as trade or reciprocity are invoked. A number of additional theoretical perspectives may provide directions for further investigation of food transfer among adults. For example, producer–scrounger games (Giraldeau and Beauchamp, 1999; Giraldeau and Caraco, 2000) have already been used to investigate foraging behavior in a small foraging group of capuchins (Di Bitetti and Janson, 2001). Ideal free distribution models could be used to investigate how the occurrence of food transfer depends on attributes of the food, such as handling time, and on attributes of the competitors, such as competitor density and relative fighting ability (Hamilton, 2002). The predictions of these models could be tested experimentally by manipulating variables such as the feeding rate of individuals, the divisibility of the food resource, and the numbers of individuals present (Stevens and Stephens, 2002), providing considerable scope for future research. In addition, researchers studying the distribution of food within human populations have investigated the possibility that food transfer may be a form of costly signaling that provides information about phenotypic quality or about an individual’s willingness to cooperate (Gurven, 2004; see Winterhalder, 1996). The hypotheses and predictions formulated by these researchers may stimulate future lines of enquiry for those studying food transfer in nonhuman primates. The remainder of this review concentrates on food transfer between older and younger individuals. Although in some cases older individuals take food items from younger animals, in most cases food is transferred in the opposite direction. Where food is transferred from an older to a younger individual, dominance relations, trade, or reciprocity is unlikely to provide a convincing explanation.

III. Food Transfer to Infants from Parents and Helpers In mammals, infants may be most likely to obtain their first solid food items by taking a portion of the food that their mothers are eating. In primates, the most commonly reported occurrences of food transfer are from mothers to infants (Feistner and McGrew, 1989). Once infants are more mobile, they may move away from their mothers, approach other group members, and beg for food. Importantly, this mobility results in infants having the opportunity to determine which foods they are most

FOOD TRANSFER IN NONHUMAN PRIMATES

271

likely to receive by approaching individuals who are holding particular food items. In the following sections, we first review the evidence that food transfer influences infant growth or survival, and then discuss the two main sets of functional explanations for transfer of food to infants, the nutritional and informational hypotheses. However, first, the distribution of food transfer across primate species, and the involvement of different age and sex categories in provisioning of infants, is summarized. The transfer of solid food to infants has been reported in 27 species of primate, both in captivity and in free-ranging populations (Table II). The data in Table II suggest that the occurrence of food transfer from adults to infants is not evenly distributed across primate groups (1 species of tarsier; 13 species of callitrichids; 8 other New World primates; 5 apes; no Old World primates). One possible explanation for this pattern of data is that there has been a reporting bias in the literature. However, given the large amount of research that has been carried out on Old World primates, the lack of reporting suggests that transfer of food to infants from others is a genuinely rare occurrence. The most detailed studies of food transfer to infants have been carried out on callitrichid primates and chimpanzees. In callitrichids, which exhibit cooperative breeding systems in which twin offspring are reared communally, infants appear to be almost entirely provisioned by others when they begin to eat solid food, before gradually developing into independent feeders (e.g., Feistner and Price, 1990, 2000; Hoage, 1982; Ruiz-Miranda et al., 1999). The high level of food transfer in callitrichids may be related to their cooperative breeding system. Alternatively, the breadth of diet or importance of extractive foraging in the diets of chimpanzees and some callitrichid species may potentially explain the high frequency with which infants obtain solid food items from others if food transfer plays a role in infants learning diet choices or food-processing skills. Evidence described later suggests that the function of food transfer may differ between chimpanzees and callitrichids. In callitrichid primates, mothers, fathers, and alloparents may be predicted to differ in the frequency with which they transfer food to infants. For example, as breeding females are able to become pregnant while suckling the previous set of offspring, fathers may be selected to provision offspring with solid food if this allows their mate to allocate resources to the next set of offspring, while mothers may play a lesser role in provisioning as they bear the metabolic costs of gestation and lactation (Feistner and Price, 1990). Both parents may be predicted to transfer food to infants more frequently than do alloparents. Only a small number of studies have reported the relative contributions of mothers, fathers and alloparents to offspring provisioning in callitrichids, with variable results and generally small sample sizes (Table III). Six of the 10 studies reported that fathers

TABLE II Species in Which Infants or Juveniles Have Been Reported to Obtain Food via Transfer from Older Group Members 272

Species

Habitat

Ref.

Spectral tarsier, Tarsius spectrum Bare-ear marmoset, Callithrix argentata Buffy-headed marmoset, Callithrix flaviceps Geoffroy’s marmoset, Callithrix geoffroyi Common marmoset, Callithrix jacchus

FR C FR C C

Pied bare-faced tamarin, Saguinus bicolor Saddle-back tamarin, Saguinus fuscicollis Red-bellied tamarins, Saguinus labiatus

C C FR C

Moustached tamarin, Saguinus mystax Black-mantle tamarin, Saguinus nigricollis Cotton-top tamarin, Saguinus oedipus

FR FR C

Black lion tamarin, Leontopithecus chrysopygus

C

Gursky (2000) Feistner and Price (1991) Ferrari (1987, 1992) Feistner and Price (1991) Brown et al. (2004); Chalmers and Locke-Haydon (1984); Feistner and Price (1991); Vitale and Quayras (1997) Price and Feistner (2001) Cebul and Epple (1984) Goldizen (1989) Cebul and Epple (1984); Coates and Poole (1983); Feistner and Price (1991) Heymann (1996) Izawa (1978) Feistner and Chamove (1986); Feistner and Price (1990, 1991); Roush and Snowdon (2001) Feistner and Price (2000)

Golden lion tamarin, Leontopithecus rosalia

Pygmy marmoset, Cebuella pygmaea Goeldi’s monkey, Callimico goeldii Dusky titi monkey, Callicebus moloch Yellow-handed titi, Callicebus torquatus Owl monkey, Aotus trivirgatus Tufted capuchin, Cebus apella White-faced capuchin, Cebus capucinus Howler monkey, Alouatta palliata Spider monkey, Ateles geoffroyi White-handed gibbons, Hylobates lar

273

Orangutan, Pongo pygmaeus Gorilla, Gorilla gorilla Bonobo, Pan paniscus Chimpanzee, Pan troglodytes

Abbreviations: C, captive; FR, Free-ranging.

C FR C C FR FR C C FR FR FR C FR FR C FR C FR

Brown and Mack (1978); Hoage (1982); Price and Feistner (1993); Rapaport (1999) Ruiz-Miranda et al. (1999) Feistner and Price (1991) Feistner and Price (1991); Jurke and Pryce (1994) Wright (1984) Starin (1978) Wright (1984) Fragaszy et al. (1997) Rose (2001) Carpenter (1965) Dare (1974) Schessler and Nash (1977) Nettelbeck (1998) Russon (2003); Utami and van Hooff (1997) Maestripieri et al. (2002) Kuroda (1984) Silk (1979) Boesch and Boesch-Achermann (2000); Goodall (1968, 1986); Nishida and Turner (1996); Silk (1978)

274

GILLIAN R. BROWN et al.

TABLE III Relative Contributions of Mothers, Fathers, and Alloparents to Provisioning Infants with Food Species Spectral tarsier, Tarius spectrum Common marmoset, Callithrix jacchus Pied bare-faced tamarin, Saguinus bicolor Saddle-back tamarin, Saguinus fuscicollis Cotton-top tamarin, Saguinus oedipus

Owl monkey, Aotus trivirgatus Dusky titi monkey, Callicebus moloch Yellow-handed titi monkey, Callicebus torquatus

Relative contribution Mothers > subadult female > males Fathers ¼ mothers > siblings Fathers > mothers

na

Ref.

6, 1, 7

Gursky (2000)

4, 4, 5

Vitale and Queyras (1997) Price and Feistner (2001) Cebul and Epple (1984) Wolters (1978)

3, 3

Siblings > mothers > fathers Fathers > mothers

3, 5, 5 6, 6

Parents > siblings Fathers > others

16, 46 5, 17

Father > mother

1, 1

Price (1992a) Roush and Snowdon (2001) Wright (1984)

Father > mother

1, 1

Wright (1984)

Father > sibling > mother

1, 1, 1

Starin (1978)

a

N, Total number of provisioners in each class of individuals in study, listed in the same order as under relative contribution.

transfer food to offspring more frequently than do other group members (see also Washabaugh et al., 2002), and the 2 studies with the largest samples reported that fathers transfer more food than do others (Roush and Snowdon, 2001) or that both parents transfer more food than do siblings (Price, 1992a). These limited data suggest that callitrichid fathers play a particularly important role in providing offspring with solid food. Parents are expected to transfer solid food to offspring if this increases offspring growth, increases the chance of infant survival, or precipitates weaning. However, it is less clear why individuals other than parents would allow infants to take their food. The proposed benefits to helpers include (1) gaining experience in rearing offspring, so as to improve their chance of successfully breeding in the future, (2) increasing their chance of inheriting a breeding position by reducing the likelihood of being expelled from the group, (3) raising individuals that may serve as future helpers once in a breeding position, (4) gaining inclusive fitness through caring for and improving the survival chances of relatives, (5) increasing the chances of being chosen as mating partner, and (6) increasing their own survival

FOOD TRANSFER IN NONHUMAN PRIMATES

275

probability by increasing group size (reviewed by Cockburn, 1998; Emlen, 1991; Jennions and Macdonald, 1994; Kokko et al., 2001; Snowdon, 1996; Tardif, 1997). The costs and benefits of alloparental care may depend on the age, sex, and relatedness of the helper to the offspring. For example, helpers of the sex that remains in the natal group may be predicted to provide higher levels of care than the sex that disperses, as a result of being more closely related to infants or being more likely to benefit from an increase in future group size (Brotherton et al., 2001). The predicted relationship between dispersal patterns and levels of alloparental care has been upheld in a number of studies on cooperatively breeding mammals. For example, in species such as meerkats (Suricata suricatta) and red foxes (Vulpes vulpes), in which males generally disperse from their natal group to breed, females contribute more to rearing young than do males (Brotherton et al., 2001; Clutton-Brock et al., 2001, 2002; Moehlman and Hofer, 1997), while in other species, such as African wild dogs (Lycaon pictus), females disperse to breed and male helpers provide more care than do females (Malcolm and Marten, 1982). As yet, there are insufficient data on the provisioning levels of male and female nonparental helpers to test this relationship in callitrichid primates. Studies that investigate the extent to which different categories of carers contribute to the feeding of offspring need to control for the effects of individual differences in food acquisition (Clutton-Brock et al., 2001); for example, in meerkats, the number of food items given by helpers to pups was found to be approximately linearly related to the number of items located (Clutton-Brock et al., 2001).

IV. Does Food Transfer Influence Infant Growth and/or Survival? Before investigating the manner in which infants might benefit by receiving food from others, it is important to assess whether infants do in fact obtain fitness benefits as a result of obtaining solid food from other group members. The transfer of food to infants could (1) increase offspring growth rates, reduce the age of weaning, or reduce the age of nutritional independency; and/or (2) increase offspring survival chances during the transition to complete nutritional independence (Feistner and McGrew, 1989; Fragaszy and Bard, 1997; Lefebvre, 1985; McGrew, 1975; Price and Feistner, 1993). By weaning infants at an earlier age, mothers may be able to begin the next reproductive event sooner (Ross and MacLarnon, 2000). Here we use inter- and intraspecific data to evaluate key predictions

276

GILLIAN R. BROWN et al.

Prediction 1: Offspring that receive solid food from others have younger ages at weaning, higher growth rates, or younger ages at nutritional independency compared with offspring that obtain most of their solid food themselves

In an early comparative study, Lefebvre (1985) found that weaning did not occur at an earlier age than expected in those species in which food transfer has been reported most frequently (i.e., lion tamarins and chimpanzees). However, more recent cross-species comparative analyses have indicated that, among anthropoid primates, alloparenting (based on the occurrence of carrying rather than food transfer) does correlate with relatively high infant growth rates, a young age at weaning (but at the same weight relative to the mother), and high reproductive rates (Mitani and Watts, 1997; Ross and MacLarnon, 2000). The relative importance of provisioning as an aspect of alloparental care is not known and there are currently insufficient data on the levels of food transfer across a broad range of primate species to test this hypothesis further. However, in callitrichid primates, the transfer of solid food to infants potentially played a vital role in the evolution of twinning and the ability of females to return to reproductive condition and conceive soon after giving birth (Garber and Leigh, 1997). Cross-species comparisons are confounded by the possibility that food transfer may have evolved to compensate for otherwise slow growth rates or late weaning ages in particular species. A further complication is that an earlier age at weaning does not necessarily translate into a correspondingly early age at full nutritional independence. In species in which helpers provision offspring with solid food, the period of postweaning care may in fact be longer than in other species, as a result of helpers feeding infants while parents initiate the next breeding attempt. In birds, cooperative breeding is associated with unusually long periods of offspring dependency (Langen, 2000). Therefore, the transfer of solid food to infants may reduce the time to weaning but may be offset by an increased age at full independence. There is little evidence from within species to test whether those infants receiving food from others grow faster or are weaned earlier compared with other infants, or to test whether mothers that provide solid food to infants have shorter interbirth intervals than other mothers. Again, correlational data are problematic as there may be confounding factors; for example, high levels of food transfer may not correlate with infant growth rate if slow-growing infants obtain more food from others than do fast-growing infants. Prediction 2: Infant survival is positively correlated with the amount of food received from others

Although there are no empirical data with which to test directly the prediction that food transfer increases infant survival chances, it is possible

FOOD TRANSFER IN NONHUMAN PRIMATES

277

to evaluate whether infants that have many helpers in their social group have a higher chance of survival than other infants. A number of studies of free-ranging callitrichids report that infant survival is positively correlated with the number of helpers in the group (moustached tamarins: Garber et al., 1984; common marmosets: Koenig, 1995; cotton-top tamarins: Snowdon, 1996; lion tamarins: Bales et al., 2000). Studies of captive cottontop tamarins have shown that infants in large groups receive more food in total than those in smaller groups (Feistner and Price, 1990; Price, 1992b). The presence of helpers is also reported to significantly increase offspring survival in other mammalian species in which helpers provision offspring, including silverbacked and golden jackals (Canis mesomelas and Canis aureus: Moehlman and Hofer, 1997) and meerkats (Russell et al., 2002). However, in contrast, no relationship between group size and infant survival has been found in other callitrichid populations (free-ranging lion tamarins: Baker et al., 1993; Dietz and Baker, 1993; free-ranging pygmy marmosets: Heymann and Soini, 1999; captive cotton-top tamarins: Price and McGrew, 1990; captive common marmosets: Rothe and Darms, 1993). Unfortunately, correlations between group size and infant survival are generally confounded by other variables, such as territory quality, and provide no indication of the relative importance of food transfer compared with other aspects of care, such as defense against predators. In summary, the presence of alloparents appears to correlate with relatively high infant growth rates and early age at weaning in primates, although food transfer may not necessarily correlate with an early age at nutritional independence. Also, there is some evidence that the presence of helpers increases the survival chances of infants during early life. However, the relative importance of food transfer in comparison with other aspects of alloparental care in accounting for these findings is currently unknown.

V. Functional Explanations of Infant Food Transfer Two main sets of hypotheses have been put forward for the mechanisms by which infant growth or survival may be enhanced by food transfer (Lefebvre, 1985; McGrew, 1975; Price and Feistner, 1993; Rapaport, 1999; Ruiz-Miranda et al., 1999; Silk, 1978). The transfer of food may (1) provide nutrients to infants during the period of weaning when they are susceptible to food shortage or, more generally, while they develop as independent foragers—nutritional hypotheses, and (2) play a role in the acquisition of knowledge about diet choices and food processing skills— informational hypotheses. These sets of hypotheses are not necessarily

278

GILLIAN R. BROWN et al.

mutually exclusive. For example, by receiving a half-eaten insect, an infant primate may benefit by gaining important nutrients that it may not be able to obtain for itself, may learn that a particular insect species is edible, and may gain the opportunity to practice essential food processing skills. In addition, both sets of benefits may play a role but their relative importance may vary with infant age; for example, prior to weaning, food transfer may supplement infant nutrition, while after weaning it may play a role in increasing diet breadth (Rapaport, 1999). The two hypotheses can nonetheless be regarded as distinct in suggesting that the primary function of food transfer is the acquisition of nutritive or informational benefits, respectively, while any additional benefits are regarded as by-products of selection rather than traits directly favored by selection. Predictions and evidence in favor of these two sets of hypotheses are discussed in the following sections, with distinctions being made between the relative roles of beggars and food possessors in determining patterns of food transfer. At this point, it is important to remember that, as the transfer of food involves two individuals, there may be a number of circumstances under which conflicts of interest occur. A food possessor that is approached by a begging infant may not necessarily hand over a portion of their food, and may instead resist by moving away or turning the body away from the other individual or by exhibiting threat displays or physical aggression to the individual attempting to gain access to the food item. Food possessors may be less willing to transfer food if they have low energy levels, if there is little food currently available, if the item is highly preferred, or if it has taken substantial time or energy to obtain. In addition, adults and offspring may be in conflict over the amount of investment provided to offspring (Trivers, 1974). In the following sections, separate predictions are made regarding the behavior of the infant and the food possessor, as evidence in favor of the prediction for one individual does not necessarily mean that the prediction for the other individual is also supported. A. Nutritional Hypotheses Food transfer may simply provide infants with additional nutrients during the period of weaning or during a more extended transition to independent foraging. Observational data suggest that food transfers are not evenly distributed between food types. Below, we compare patterns of begging and food transfer to the prediction that food items that are nutritionally rich, or that infants are unable to obtain for themselves, are more likely to be transferred than other items, and the prediction that rates of food transfer will be highest during the period of weaning or the transition to independent foraging.

FOOD TRANSFER IN NONHUMAN PRIMATES

279

Prediction 3a: Infants beg for food items that are nutritionally rich, or that they are unable to obtain for themselves, more than for other items Prediction 3b: Adults transfer food items that are nutritionally rich, or that infants are unable to obtain for themselves, more than other items

Observational studies of free-ranging primates suggest that large insects and large plant items that infants find difficult to process are the items most frequently solicited by infants (buffy-headed marmosets: Ferrari, 1987; black-mantle tamarins: Izawa, 1978; moustached tamarins: Heymann, 1996; golden lion tamarins: Ruiz-Miranda et al., 1999; yellow-handed titi monkeys: Starin, 1978; lar gibbon: Nettelbeck, 1998; chimpanzees: Assersohn and Whiten, 1999; Corp and Byrne, 2002). For instance, free-ranging infant chimpanzees solicit those foods that are both nutritionally rich and difficult for infants to obtain by themselves, such as nuts and fruits that require processing, more frequently than expected by chance given the amount of time that mothers were feeding on these different food types (Hiraiwa-Hasegawa, 1990a; Nishida and Turner, 1996; Silk, 1978). Such items appear to require manual dexterity, strength, or skills not yet acquired by the infants. Chimpanzee infants do not differ in rates of begging for high-quality and low-quality food (Nishida and Turner, 1996), suggesting that the important variable for infants is whether they are able to obtain specific food resources themselves. The amount of food acquired by transfer was greatest for those food types that infants were not observed to obtain independently (defined as difficult-to-process items), although the success of begging was similar for these foods and items that infants were able to acquire by themselves (Hiraiwa-Hasegawa, 1990a; Nishida and Turner, 1996; Silk, 1978). At least in chimpanzees, infants appear to be responsible for the observed patterns of food transfer. These data suggest that food transfer may provide infants with an opportunity to gain nutrients, particularly protein, that they would not otherwise be able to obtain. However, the data are also consistent with the hypothesis that infant begging allows them to practice food-processing skills. In addition, food items such as insects, large fruits, and nuts may be relatively rare and therefore novel to some youngsters, supporting the suggestion that infants gain information about diet breadth through provisioning. Studies in captivity, in which particular characteristics of food items, such as novelty, rarity, and difficulty in processing, can be varied independently, will allow the relative importance of these characteristics to be unraveled. Prediction 4a: Rates of infant begging will be highest during the period of weaning Prediction 4b: Rates of food transfer from adults will be highest during the period of weaning

280

GILLIAN R. BROWN et al.

If food transfer provides infants with nutritional benefits, begging may be predicted to peak during the weaning period if infants are most susceptible to nutritional deficits during this transition to independent foraging while, if food transfer provides infants with informational benefits, no such peak is necessarily expected. In both cases, begging is predicted to decline as infants become better able to forage for themselves, although the time course may depend on the food items involved, with difficult-to-process items continuing to be transferred to offspring until a later age. In fact, surprisingly little is known about the development of independent feeding in primates (Feistner and Price, 2000; Nicolson, 1986). However, in captive callitrichid primates, both begging and provisioning peak around the time of weaning (10–14 weeks), while food items such as large insects continue to be transferred until a later age (Feistner and Price, 1990, 2000; Hoage, 1982; Price and Feistner, 2001). Begging success appears to remain relatively constant during the transition to independent feeding (Feistner and Price, 2000; Price and Feistner, 2001), suggesting that adults respond directly to the begging levels of infants. The prediction therefore appears to be upheld in callitrichids, suggesting that provisioning may play an important role in the successful transition of infants through weaning in these species. In free-ranging chimpanzees, no peak in food transfer occurs at the age of weaning (around 4 years; Ross and Jones, 1999), and most instances of food begging and transfer occur prior to weaning (Hiraiwa-Hasegawa, 1990a; Nishida and Turner, 1996; Silk, 1978). Begging success rates appear to be similar for offspring of all ages (Hiraiwa-Hasegawa, 1990a; Nishida and Turner, 1996). An exception to this time course is the transfer of nuts from mother to infant chimpanzees, which peaks around 4 years of infant age and continues to around 8 years (Boesch and Boesch-Achermann, 2000). The lack of a peak in begging and provisioning around the age of weaning suggests that the main function of food transfer in chimpanzees may be informational rather than nutritional, and highlights the possibility that the role of food transfer differs between species. B. Informational Hypotheses Here, we assess the evidence that young primates acquire information about dietary choice or food-processing skills by obtaining food from other individuals. 1. Learning Food Preferences Rather than simply acquiring nutrients, young primates may acquire food preferences through exposure to food obtained from other group members, and food transfer may provide a supplemental source of information that is

FOOD TRANSFER IN NONHUMAN PRIMATES

281

used in addition to that acquired by an infant during its own exploration (Schessler and Nash, 1977). Social learning may be an important means by which primates incorporate new foods into their diets, as it is for many other animal species (Galef and Giraldeau, 2001; Galef et al., 2001; Heyes and Galef, 1996; Zentall and Galef, 1988). If food transfer allows infants to learn about diet choices, a number of predictions follow. Prediction 5a: Infants beg more for food items that are novel than for food items that are familiar to them Prediction 5b: Adults are more likely to transfer food items that are novel to infants than ones that are familiar to them

Studies of captive callitrichid primates have reported that infants are relatively unwilling to take novel food items from food bowls themselves compared with familiar items (Price and Feistner, 1993) and exhibit higher levels of begging for novel than for familiar food items (Brown et al., 2004). In contrast, older immature callitrichids are willing to eat novel food items in the absence of experienced conspecifics (Vitale and Queyras, 1997; Queyras et al., 2000). These data suggest that infant callitrichids preferentially obtain novel food items from social group members, while older immatures perhaps rely on their own experiences with novel foods or acquire food preferences through observational learning. A study has reported that, when family groups of captive common marmosets were provided with either novel or familiar food, levels of begging by infants were higher with novel than with familiar food (Brown et al., 2004). These data support the hypothesis that infant marmosets actively attempt to obtain information about diet by soliciting food from adults. Two earlier studies did not find higher levels of begging with novel food (Price and Feistner, 1993; Rapaport, 1999), although the age differences in subjects used may explain the different patterns of results; see Brown et al., 2004). Few data are available to test these predictions in chimpanzees, although one study suggests that the amount of time spent by infants feeding on items transferred from the mother is greater for novel than for familiar foods (Hiraiwa-Hasegawa, 1990b). The study by Brown et al. (2004) also reported that adults exhibited higher levels of refusals with novel than with familiar food, and begging success with novel foods was lower than with familiar, leading to similar levels of food transfer during tests with novel and familiar food. These data do not support the suggestion that adults are actively involved in facilitating learning about diet in infants. Similarly, Price and Feister (1993) reported that refusals occurred more frequently, and begging success was lower, in tests with novel than with familiar food. In contrast, a study by Rapaport (1999) reported that frequencies of transfer per food item were higher in tests with novel than with familiar

282

GILLIAN R. BROWN et al.

food, which would suggest that adults did alter their behavior in a manner that would facilitate learning in immatures. The contradictory findings of these studies indicate that further data are required, and highlight the potential influence of offspring age on patterns of begging and food transfer. Prediction 6a: The variety of items that are begged for will decline with age and begging will decrease as diet is learned Prediction 6a: The variety of items that are transferred will decline with age and provisioning will decrease as diet is learned

While frequencies of food transfer generally decrease as infants grow older, there are currently insufficient data to assess whether the variety of food items begged for and transferred declines with age, or whether begging and transfer decrease specifically in line with increasing offspring diet breadth. Observational data from field and captive settings are required that monitor the range of items that are begged for and transferred over time, and that relate patterns of begging and provisioning to the knowledge levels of particular youngsters. Prediction 7: Offspring provisioning influences future dietary choices

Despite the relatively large number of studies on social influences on feeding, there is little direct evidence that obtaining food from others during early life influences subsequent dietary choices in mammals (Fragaszy and Visalberghi, 1996; Galef and Giraldeau, 2001), apart from one study on brown rats (Rattus norvegicus: Galef et al., 2001). Social influences on diet choice may involve processes other than food transfer, including the use of olfactory and visual cues. For instance, young free-ranging baboons sniff adult muzzles while they are eating (King, 1991, 1999) and may gain olfactory cues about diet through this behavior, in a similar manner to rodents (Galef, 1996). Also, in captive cotton-top tamarins, individuals emit alarm calls and exhibit head shaking, frothing at the mouth, and mouth rubbing when they encounter noxious foods, which appears to deter other animals from approaching the food (Snowdon, 2001; Snowdon and Boe, 2003). However, one study has experimentally investigated whether infants acquire a dietary preference as a result of food transfer. This study found that infant common marmosets given a choice between food that had been transferred from group members and food that had been experienced independently exhibited a strong preference for the food that had been obtained from others (Almond et al., 2004). These data, together with the data presented by Brown et al. (2004), strongly suggest that infant common marmosets actively seek and obtain information about diet via food transfer.

FOOD TRANSFER IN NONHUMAN PRIMATES

283

2. Learning Food-Processing Skills Social learning has been hypothesized to result in the transmission of complex food processing techniques within groups of primates, including the transmission of traditional food processing techniques that differ between communities (Fragaszy and Perry, 2003; Whiten et al., 1999). A number of studies of chimpanzees have presented data supporting the suggestion that social interactions are important in the transmission of specific tool-using techniques, particularly interactions between mothers and offspring (Biro et al., 2003; Hirata and Celli, 2003; Lonsdorf et al., 2004). When mothers are processing food items, the transfer of food to infants could potentially facilitate the transmission of information about processing techniques (Caldwell and Whiten, 2003; Russon, 2003). Where there is the potential to obtain a food reward, infants may be particularly likely to attend to the processing techniques of their mothers and learn by observation, compared with situations where no food reward is available. Alternatively, by obtaining half-processed items, infants may have the opportunity to learn specific techniques for themselves, as has been reported in young black rats (Rattus rattus: Terkel, 1996). An important study on common marmosets has shown that scrounging a food reward increases the probability that a naive individual will learn a novel foraging task that is being carried out by a trained demonstrator (Caldwell and Whiten, 2003). Scrounging food items that have been harvested by others has also been shown to increase the probability that foraging behavior will be learned in adult black-capped chickadees (Parus atricapillus: Sherry and Galef, 1984) and Florida scrub jays (Aphelocoma coerulescens: Midford et al., 2000), but has been reported to hinder such learning, for example, in pigeons (Columba livia: Giraldeau and Lefebvre, 1987). If food transfer involves learning about food processing skills, the following predictions can be made. Prediction 8a: Infants beg more for food items that they are unable to process themselves than for items that they can process themselves Prediction 8b: Adults transfer food items that infants cannot process themselves at a higher rate than food items that infants can process themselves

Earlier, we discussed whether begging and food transfer are more likely to involve food items that infants find difficult to process, and suggested that the data could be interpreted as supporting either nutritional or informational hypotheses. The key distinction between these hypotheses is whether food transfer results in infant learning. Adults may actively encourage learning of food processing techniques by preferentially transferring items that are difficult for infants to process. However, studies of free-ranging

284

GILLIAN R. BROWN et al.

chimpanzee infants reported that begging success is not greater with difficult-to-process compared with easy-to-process foods (Nishida and Turner, 1996; Silk, 1978). In fact, when chimpanzee infants are very young, mothers are less likely to respond to infant begging by relinquishing difficult-toprocess food compared with easy-to-process food (Byrne, 1999; Corp and Byrne, 2002; Hiraiwa-Hasegawa, 1990a; Silk, 1978). Mothers may be unwilling to transfer unprocessed food items that could be harmful to infants. Prediction 9a: Rates of infant begging will be related to the skill level of that individual and will decrease as skills are learned Prediction 9b: Rates of food transfer will be related to the skill level of that individual and will decrease as skills are learned

With difficult-to-process foods, the amount of food obtained via transfer from their mothers decreases in line with an increase in the amount of these food items that youngsters obtain by themselves (Boesch and Boesch-Achermann, 2000; Corp and Byrne, 2002). However, whether infant begging decreases with age, strength, or foraging ability is not known. Difficult-to-process food items may continue to be obtained via transfer until a later age because youngsters lack the physical strength or dentition to open these items, rather than because they lack a particular skill that requires time to learn. Future studies may ascertain whether levels of begging are related to the skill levels of individual youngsters. Studies in captivity would be particularly useful as skill levels could be manipulated independently of physical strength and dental development. There is currently no evidence that primate mothers are more willing to transfer food depending on the skill level of the infant. Captive adult capuchin monkeys did not differ in their willingness to hand over nuts to infants that were either able or unable to open nuts by themselves (Fragaszy et al., 1997). Prediction 10: Individuals that have had the opportunity to obtain food from other individuals acquire processing skills at a younger age than individuals without this opportunity and acquire skills similar to those of their demonstrators

While researchers have postulated that infants may learn processing skills through the transfer of food items from others, at present there is no direct evidence that food transfer influences the subsequent processing abilities of youngsters or reduces the age at which skills are learned. In species in which different food processing techniques exist across study sites (Fragaszy and Perry, 2003; Whiten et al., 1999) there is currently little direct evidence that infants socially learn the technical variants exhibited by their social group members (Galef, 2003). The role of food transfer could be assessed in captivity by manipulating the amount of food transferred

FOOD TRANSFER IN NONHUMAN PRIMATES

285

and recording the development of food processing abilities, or in the field by determining the correlation between these variables. Whether the learning of processing skills is differentially influenced by the transfer of unprocessed, semiprocessed, or processed food items also deserves further investigation.

VI. Information Donation and Teaching Adults may direct the learning of dietary choices and processing skills in younger individuals in a number of ways. One method would be to prevent infants from eating certain foods by removing food items from the hands or mouths of infants, by giving vocalizations when infants attempt to eat particular items, or by threatening infants when they approach specific food sources. This would be an example of ‘‘coaching’’ (Caro and Hauser, 1992). There are a number of descriptive accounts, particularly among apes, in which adult primates have apparently prevented younger individuals from eating certain food items (e.g., Boinski and Fragaszy, 1989; Fletemeyer, 1978; Hiraiwa-Hasegawa, 1990b; Nishida et al., 1983). However, researchers generally agree that there is a distinct absence of evidence that adults prevent youngsters from ingesting food items that adults have learned to avoid (Fragaszy and Visalberghi, 1996; Galef and Giraldeau, 2001; King, 1994; Snowdon, 2001). Rather than preventing youngsters from eating certain foods, adults might actively encourage them to eat other food types. Adults in several species of callitrichid primate have been reported to ‘‘offer’’ food to offspring by holding a food item in an outstretched hand and emitting a vocalization that is similar to the one used during begging and appears to result in an infant approaching and taking the food item (e.g., Brown and Mack, 1978; Feistner and Price, 1990, 1991; Ferrari, 1987; Hoage, 1982; Moody and Menzel, 1976). A number of primate species have been reported to give ‘‘food calls’’ on locating a patch of food (including spider monkeys: Chapman and Lefebvre, 1990; macaques: Dittus, 1984; Hauser and Marler, 1993a,b; chimpanzees: Hauser and Wrangham, 1987; Wrangham, 1975), which may benefit the caller by, for example, attracting the caller’s relatives, reducing predation risk, or enhancing social status (Evans and Evans, 1999; Hauser, 1996; Hauser and Marler, 1993b; Wrangham, 1975). Therefore, the fact that adults give food calls does not necessarily imply that coaching or teaching is occurring. However, adults could potentially enhance infant learning by giving food calls preferentially with food items that are novel to infants or that infants are not yet able to process themselves (King, 1999; Maestripieri and Call, 1996; Snowdon, 2001).

286

GILLIAN R. BROWN et al.

For food calling to infants to be seen as a form of ‘‘teaching’’ (Caro and Hauser, 1992), it would need to be shown that (1) adults produce more calls when infants are present than when they are absent (controlling for any effects of group size), (2) adults transfer food at a higher rate after calling than when holding food and not calling, (3) adults are sensitive to the state of the infant such that they give more calls for foods that are novel rather than familiar to infants, or for foods that infants do not have the skills to process for themselves rather than for other food items, and (4) the subsequent diet choices or food processing skills of infants are influenced by these interactions. Studies of captive cotton-top tamarins have revealed (1) that rates of food calling while eating are higher for adults with offspring than for adults without offspring (Roush and Snowdon, 2000) and (2) that attempts by infants to take food from adults are more successful when the adult vocalizes (Roush and Snowdon, 2001). Points 3 and 4 have yet to be tested. Current data suggest that calls are given when individuals are aroused and when competition for food is greatest, and that calls provide a signal of arousal that results in attraction of offspring to food sources that are preferred or rapidly consumed (Benz et al., 1992; Caine et al., 1995; Elowson et al., 1991; Roush and Snowdon, 2000). However, there is currently insufficient evidence that food calling functions to transmit information to offspring or that it directs learning in offspring. In summary, while diet choice and food-handling techniques may be passed from one generation to another via social learning processes, there is debate over whether adult primates provide active guidance to their offspring in a manner congruent with coaching or teaching (Caro and Hauser, 1992; Russon, 1997).

VII. Summary The main aim of this review was to collate data on food transfer in primates and to compare these data with predictions that stem from various functional explanations of food transfer. First, the data on adult–adult food transfer were summarized. In adults, although patterns of food transfer can often be predicted by dominance relationships, with dominant individuals obtaining food from subordinates, the data indicate that not all food transfer follows this pattern. Researchers have suggested that food transfer may therefore involve trade for other commodities, including grooming, social support, or sexual access, although these hypotheses are not strongly supported by the available data. Also, while studies of chimpanzees, capuchins, and tamarins have indicated that frequencies of food transfer between individuals are often correlated within dyads,

FOOD TRANSFER IN NONHUMAN PRIMATES

287

such patterns do not necessarily indicate that reciprocity had occurred. Analyzing the costs and benefits of defending a food item may reveal that the costs of harassment provides sufficient explanation for a notable proportion of food transfer events. A number of theoretical perspectives may prove useful in future investigations of food transfer among adults. Patterns of food transfer from older to younger individuals were then reviewed. Infants obtain food from older individuals in a number of primate species, in particular the callitrichid primates. Quantifying the relative contributions of individuals of different age and sex categories to offspring provisioning will potentially increase our understanding of parental and alloparental care in cooperatively breeding species. The current data on callitrichids and chimpanzees highlight the possibility that both nutritional and informational benefits may be gained by infants. In callitrichids, food transfer peaks around the age of weaning and is most likely to involve novel or difficult-to-process items, with evidence suggesting that infants learn dietary preferences via food transfer. In chimpanzees, most food transfer occurs prior to weaning and involve items that are difficult for infants to process by themselves. Studies suggest that interactions between mothers and infants may result in the transmission of tool-using skills. Obtaining a food reward during such interactions may enhance any social learning processes that are taking place. Whether adults actively direct the dietary choices or food-processing techniques of youngsters remains controversial. We hope to have shown that this area of research has undergone key developments and presents considerable scope for further empirical investigation both in primates and in other animal species.

Acknowledgments This research was funded by the MRC (G.R.B. and R.E.A.A.) and the BBSRC (Y.v.B.). We thank Richard Byrne, Christine Caldwell, Jeff Galef, Rebecca Kilner, Kevin Laland, Tim Roper, Joan Silk, Peter Slater, Charles Snowdon, and an anonymous referee for comments on the manuscript.

References Almond, R. E. A., van Bergen, Y., and Brown, G. R. (2004). Social influences on dietary preferences in infant common marmosets (Callithrix jacchus). (In preparation.) Assersohn, C., and Whiten, A. (1999). Food sharing between mother and infant chimpanzees in the wild: Beggars can be choosers. Folia Primatol. 70, 226–227. Baker, A. J., Dietz, J. M., and Kleiman, D. G. (1993). Behavioural evidence for monopolization of paternity in multi-male groups of golden lion tamarins. Anim. Behav. 46, 1091–1103.

288

GILLIAN R. BROWN et al.

Bales, K., Dietz, J., Baker, A., Miller, K., and Tardif, S. D. (2000). Effects of allocare-givers on fitness of infants and parents in Callitrichid primates. Folia Primatol. 71, 27–38. Barnard, C. J., and Sibly, R. M. (1981). Producers and scroungers: A general model and its application to captive flocks of house sparrows. Anim. Behav. 29, 543–550. Benz, J. J., Leger, D. W., and French, J. A. (1992). Relationship between food preferences and food-elicited vocalisations in golden lion tamarins (Leontopithecus rosalia). J. Comp. Psychol. 106, 142–149. Biro, D., Inoue-Nakamura, N., Tonooka, R., Yamakoshi, G., Sousa, C., and Matsuzawa, T. (2003). Cultural innovation and transmission of tool use in wild chimpanzees: Evidence from field experiments. Anim. Cogn. 6, 213–223. Blurton Jones, N. G. (1984). A selfish origin for human food sharing: Tolerated theft. Ethol. Sociobiol. 5, 1–3. Blurton Jones, N. G. (1986). Fitness returns from resources and the outcome of contests: Some implications for primatology and anthropology. In ‘‘Primate Ontogeny, Cognition and Social Behaviour’’ (J. G. Else and P. C. Lee, Eds.), pp. 393–406. Cambridge University Press, Cambridge. Blurton Jones, N. G. (1987). Tolerated theft, suggestions about the ecology and evolution of sharing, hoarding and scrounging. Soc. Sci. Info. 26, 31–54. Boesch, C. (1994). Hunting strategies of Gombe and Taı¨ chimpanzees. In ‘‘Chimpanzee Cultures’’ (R. W. Wrangham, W. C. McGrew, F. B. M. de Waal, and P. G. Heltne, Eds.), pp. 77–91. Harvard University Press, Boston, MA. Boesch, C., and Boesch, H. (1989). Hunting behavior of wild chimpanzees in the Taı¨ National Park. Am. J. Phys. Anthropol. 78, 547–573. Boesch, C., and Boesch-Achermann, H. (2000). ‘‘The Chimpanzees of the Taı¨ Forest: Behavioural Ecology and Evolution.’’ Oxford University Press, Oxford. Boinski, S., and Fragaszy, D. M. (1989). The ontogeny of foraging in squirrel monkeys, Saimiri oerstedi. Anim. Behav. 37, 415–428. Brockmann, H. J., and Barnard, C. J. (1979). Kleptoparasitism in birds. Anim. Behav. 27, 487–514. Brotherton, P. N. M., Clutton-Brock, T. H., O’Riain, M. J., Gaynor, D., Sharpe, L., Kansky, R., and McIlrath, G. M. (2001). Offspring food allocation by parents and helpers in a cooperative mammal. Behav. Ecol. 12, 590–599. Brown, G. R. (2004). Tolerated scrounging in non-human primates. Behav. Brain Sci. (in press). Brown, G. R., Almond, R. E. A., Bates, N. (2004). Adult-infant food transfer in common marmosets: An experimental study. Am. J. Primatol. (In press). Brown, K., and Mack, D. S. (1978). Food sharing among captive Leontopithecus rosalia. Folia Primatol. 29, 268–290. Byrne, R. W. (1999). Cognition in great ape ecology: Skill-learning ability opens up foraging opportunities. In ‘‘Mammalian Social Learning: Comparative and Ecological Perspectives’’ (H. O. Box and K. R. Gibson, Eds.), pp. 333–350. Cambridge University Press, Cambridge. Caine, N. G., Addington, R. L., and Windfelder, T. L. (1995). Factors affecting the rates of food calls given by red-bellied tamarins. Anim. Behav. 50, 53–60. Caldwell, C., and Whiten, A. (2003). Scrounging facilitates social learning in common marmosets, Callithrix jacchus. Anim. Behav. 65, 1085–1092. Caraco, T., and Brown, J. L. (1986). A game between communal breeders: When is foodsharing stable? J. Theor. Biol. 118, 379–393. Caro, T. M., and Hauser, M. D. (1992). Is there teaching in nonhuman animals? Q. Rev. Biol. 67, 151–174.

FOOD TRANSFER IN NONHUMAN PRIMATES

289

Carpenter, C. R. (1965). The howlers of Barro Colorado Island. In ‘‘Primate Behavior: Field Studies of Monkeys and Apes’’ (I. DeVore, Ed.), pp. 250–291. Holt, Rinehart and Winston, New York. Cebul, M. S., and Epple, G. (1984). Father–offspring relationships in laboratory families of saddle-back tamarins (Saguinus fuscicollis). In ‘‘Primate Paternalism’’ (D. M. Taub, Ed.), pp. 1–19. Van Nostrand Reinhold, New York. Chalmers, N. R., and Locke-Haydon, J. (1984). Correlations among measures of playfulness and skillfulness in captive common marmosets (Callithrix jacchus jacchus). Dev. Psychobiol. 17, 191–208. Chapman, C. A., and Lefebvre, L. (1990). Manipulating foraging group size: Spider monkey food calls at fruiting trees. Anim. Behav. 39, 891–896. Cheney, D. L., and Seyfarth, R. M. (1990). ‘‘How Monkeys See the World: Inside the Mind of Another Species.’’ University of Chicago Press, Chicago, IL. Clutton-Brock, T. H., Brotherton, P. N. M., O’Riain, M. J., Griffin, A. S., Gaynor, D., Kansky, R., Sharpe, L., and McIlrath, G. M. (2001). Contributions to cooperative rearing in meerkats. Anim. Behav. 61, 705–710. Clutton-Brock, T. H., Russell, A. F., Sharpe, L. L., Young, A. J., Balmforth, Z., and McIlrath, G. M. (2002). Evolution and development of sex differences in cooperative behavior in meerkats. Science 297, 253–256. Coates, A., and Poole, T. B. (1983). The behaviour of the callitrichid monkey, Saguinus labiatus labiatus, in the laboratory. Int. J. Primatol. 4, 339–371. Cockburn, A. (1998). Evolution of helping behavior in cooperatively breeding birds. Annu. Rev. Ecol. System 29, 141–177. Corp, N., and Byrne, R. W. (2002). The ontogeny of manual skill in wild chimpanzees: Evidence from feeding on the fruit of Saba florida. Behaviour 139, 137–168. Dare, R. (1974). Food-sharing in free-ranging Ateles geoffroyi (red spider monkeys). Lab. Primate Newslett. 13, 19–21. de Waal, F. B. M. (1989). Food sharing and reciprocal obligations among chimpazees. J. Hum. Evol. 18, 433–459. de Waal, F. B. M. (1997a). The chimpanzee’s service economy: Food for grooming. Evol. Hum. Behav. 18, 375–386. de Waal, F. B. M. (1997b). Food transfers through mesh in brown capuchins. J. Comp. Psychol. 111, 370–378. de Waal, F. B. M. (2000). Attitudinal reciprocity in food sharing among brown capuchin monkeys. Anim. Behav. 60, 253–261. de Waal, F. B. M., Luttrell, L. M., and Canfield, M. E. (1993). Preliminary data on voluntary food sharing in brown capuchin monkeys. Am. J. Primatol. 29, 73–78. Di Bitetti, M. S., and Janson, C. H. (2001). Social foraging and the finder’s share in capuchin monkeys, Cebus apella. Anim. Behav. 62, 47–56. Dietz, J. M., and Baker, A. J. (1993). Polygyny and female reproductive success in golden lion tamarins, Leontopithecus rosalia. Anim. Behav. 46, 1067–1078. Dittus, W. P. J. (1984). Toque macaque food calls: Semantic communication concerning food distribution in the environment. Anim. Behav. 32, 470–477. Elowson, A. M., Tannenbaum, P. L., and Snowdon, C. T. (1991). Food-associated calls correlate with food preferences in cotton-top tamarins. Anim. Behav. 42, 931–937. Emlen, S. T. (1991). Evolution of cooperative breeding in birds and mammals. In ‘‘Behavioural Ecology: An Evolutionary Approach’’ (J. R. Krebs and N. B. Davies, Eds.), 3rd ed., pp. 301–337. Blackwell Scientific, Oxford. Evans, C. S., and Evans, L. (1999). Chicken food calls are functionally referential. Anim. Behav. 58, 307–319.

290

GILLIAN R. BROWN et al.

Feistner, A. T. C., and Chamove, A. S. (1986). High motivation towards food increases foodsharing in cotton-top tamarins. Dev. Psychobiol. 19, 439–452. Feistner, A. T. C., and McGrew, W. C. (1989). Food-sharing in primates: A critical review. In ‘‘Perspectives in Primate Biology’’ (P. K. Seth and S. Seth, Eds.), Vol. 3, pp. 21–36. Today and Tomorrow’s Printers and Publishers, New Delhi, India. Feistner, A. T. C., and Price, E. C. (1990). Food-sharing in cotton-top tamarins (Saguinus oedipus). Folia Primatol. 54, 34–45. Feistner, A. T. C., and Price, E. C. (1991). Food offering in New World primates: Two species added. Folia Primatol. 57, 165–168. Feistner, A. T. C., and Price, E. C. (2000). Food sharing in black lion tamarins (Leontopithecus chrysopygus). Am. J. Primatol. 52, 47–54. Ferrari, S. F. (1987). Food transfers in a wild marmoset group. Folia Primatol. 48, 203–206. Ferrari, S. F. (1992). The care of infants in a wild marmoset (Callithrix flaviceps) group. Am. J. Primatol. 26, 109–118. Fletemeyer, J. R. (1978). Communication about potentially harmful foods in free-ranging chacma baboons, Papio ursinus. Primates 19, 223–226. Fragaszy, D. M., and Bard, K. (1997). Comparison of development and life history in Pan and Cebus. Int. J. Primatol. 18, 683–701. Fragaszy, D. M., and Perry, S. (2003). ‘‘The Biology of Traditions: Models and Evidence.’’ Cambridge University Press, Cambridge. Fragaszy, D. M., and Visalberghi, E. (1996). Social learning in monkeys: Primate ‘‘primacy’’ reconsidered. In ‘‘Social Learning in Animals: The Roots of Culture’’ (C. M. Heyes and B. G. Galef, Eds.), pp. 65–84. Academic Press, San Diego, CA. Fragaszy, D. M., Feuerstein, J. M., and Mitra, D. (1997). Transfers of food from adults to infants in tufted capuchins (Cebus apella). J. Comp. Psychol. 111, 194–200. Galef, B. G. (1996). Social enhancement of food preferences in Norway rats: A brief review. In ‘‘Social Learning in Animals: The Roots of Culture’’ (C. M. Heyes and B. G. Galef, Eds.), pp. 49–64. Academic Press, San Diego, CA. Galef, B. G. (2003). Social learning: Promotor or inhibitor of innovation? In ‘‘Animal Innovation’’ (S. M. Reader and K. N. Laland, Eds.), pp. 137–152. Oxford University Press, Oxford. Galef, B. G., and Giraldeau, L. (2001). Social influences on foraging in vertebrates: Causal mechanisms and adaptive functions. Anim. Behav. 61, 3–15. Galef, B. G., Marczinski, C. A., Murray, K. A., and Whiskin, E. E. (2001). Food stealing by young Norway rats. J. Comp. Psychol. 115, 16–21. Garber, P. A., and Leigh, S. R. (1997). Ontogenetic variation in small-bodied New World primates: Implications for patterns of reproduction and infant care. Folia Primatol. 68, 1–22. Garber, P. A., Moya, L., and Malaga, C. (1984). A preliminary field study of the moustached tamarin monkey (Saguinus mystax) in northeastern Peru: Questions concerned with the evolution of a communal breeding system. Folia Primatol. 42, 17–32. Giraldeau, L.-A., and Beauchamp, G. (1999). Food exploitation: Searching for the optimal joining policy. Trends Ecol. Evol. 14, 102–106. Giraldeau, L.-A., and Caraco, T. (2000). ‘‘Social Foraging Theory.’’ Princeton University Press, Princeton, NJ. Giraldeau, L.-A., and Lefebvre, L. (1987). Scrounging prevents cultural transmission of foodfinding behaviour in pigeons. Anim. Behav. 35, 387–394. Goldizen, A. W. (1989). Social relationships in a cooperatively polyandrous group of tamarins (Saguinus fuscicollis). Behav. Ecol. Sociobiol. 24, 79–89.

FOOD TRANSFER IN NONHUMAN PRIMATES

291

Goodall, J. (1968). The behaviour of free-living chimpanzees in the Gombe Stream Reserve. Anim. Behav. Monogr. 1, 161–311. Goodall, J. (1986). ‘‘The Chimpanzees of Gombe: Patterns of Behavior.’’ Harvard University Press, Boston, MA. Gursky, S. (2000). Allocare in a nocturnal primate: Data on the spectral tarsier, Tarsius spectrum. Folia Primatol. 71, 39–54. Gurven, M. (2004). To give and not to give: The behavioral ecology of human food transfers. Behav. Brain Sci. (in press). Hamilton, I. M. (2002). Kleptoparasitism and the distribution of unequal competitors. Behav. Ecol. 13, 260–267. Hauser, M. D. (1996). ‘‘The Evolution of Communication.’’ MIT Press, Cambridge, MA. Hauser, M. D., and Marler, P. (1993a). Food-associated calls in rhesus macaques (Macaca mulatta). I. Socioecological factors. Behav. Ecol. 4, 194–205. Hauser, M. D., and Marler, P. (1993b). Food-associated calls in rhesus macaques (Macaca mulatta). II. Costs and benefits of call production and suppression. Behav. Ecol. 4, 206–212. Hauser, M. D., and Wrangham, R. W. (1987). Manipulation of food calls in captive chimpanzees. Folia Primatol. 48, 207–210. Hauser, M. D., Chen, M. K., Chen, F., and Chuang, E. (2003). Give unto others: Genetically unrelated cotton-top tamarin monkeys preferentially give food to those who altruistically give food back. Proc. R. Soc. Lond. B Biol. Sci. 270, 2363–2370. Hemelrijk, C. K., Meier, C., and Martin, R. D. (1999). ‘‘Friendship’’ for fitness in chimpanzees? Anim. Behav. 58, 1223–1229. Heyes, C. M., and Galef, B. G. (1996). ‘‘Social Learning in Animals: The Roots of Culture.’’ Academic Press, San Diego, CA. Heymann, E. W. (1996). Social behavior of wild moustached tamarins, Saguinus mystax, at the Estacio´n Biolo´gica Quebrada Blanco, Peruvian Amazonia. Am. J. Primatol. 38, 101–113. Heymann, E. W., and Soini, P. (1999). Offspring number in pygmy marmosets, Cebuella pygmaea, in relation to group size and the number of adult males. Behav. Ecol. Sociobiol. 46, 400–404. Hiraiwa-Hasegawa, M. (1990a). Role of food sharing between mother and infant in the ontogeny of feeding behavior. In ‘‘The Chimpanzees of the Mahale Mountain: Sexual and Life History Strategies’’ (T. Nishida, Ed.), pp. 267–275. University of Tokyo Press, Tokyo. Hiraiwa-Hasegawa, M. (1990b). A note on the ontogeny of feeding. In ‘‘The Chimpanzees of the Mahale Mountain: Sexual and Life History Strategies’’ (T. Nishida, Ed.), pp. 277–283. University of Tokyo Press, Tokyo. Hirata, S., and Celli, M. L. (2003). Role of mothers in the acquisition of tool-use behaviours by captive infant chimpanzees. Anim. Cogn. 6, 235–244. Hoage, R. J. (1982). Social and physical maturation in captive lion tamarins, Leontopithecus rosalia rosalia (Primates: Callitrichidae). Smithsonian Contrib. Zool. 354, 1–56. Isaac, G. (1978). The food-sharing behavior of protohuman hominids. Sci. Am. 238, 90–108. Izawa, K. (1978). A field study of the ecology and behavior of the black-mantle tamarin (Saguinus nigricollis). Primates 19, 241–274. Jennions, M. D., and Macdonald, D. W. (1994). Cooperative breeding in mammals. Trends Ecol. Evol. 9, 89–93. Jurke, M. H., and Pryce, C. R. (1994). Parental and infant behaviour during early periods of infant care in Goeldi’s monkey, Callimico goeldii.. Anim. Behav. 48, 1095–1112. King, B. J. (1991). Social information transfer in monkeys, apes, and hominids. Yearb. Phys. Anthropol. 34, 97–115.

292

GILLIAN R. BROWN et al.

King, B. J. (1994). ‘‘The Information Continuum: Evolution of Social Information Transfer in Monkeys, Apes, and Hominids.’’ School of American Research Press, Santa Fe, NM. King, B. J. (1999). New directions in the study of primate learning. In ‘‘Mammalian Social Learning: Comparative and Ecological Perspectives’’ (H. O. Box and K. R. Gibson, Eds.), pp. 17–32. Cambridge University Press, Cambridge. Koenig, A. (1995). Group size, composition, and reproductive success in wild common marmosets (Callithrix jacchus). Am. J. Primatol. 35, 311–317. Kokko, H., Johnstone, R. A., and Clutton-Brock, T. H. (2001). The evolution of cooperative breeding through group augmentation. Proc. R. Soc. Lond. B Biol. Sci. 268, 187–196. Kuroda, S. (1984). Interactions over food among pygmy chimpanzees. In ‘‘The Pygmy Chimpanzee: Evolutionary Biology and Behavior’’ (R. L. Susman, Ed.), pp. 301–324. Plenum Press, New York. Langen, T. A. (2000). Prolonged offspring dependence and cooperative breeding in birds. Behav. Ecol. 11, 367–377. Lefebvre, L. (1985). Parent–offspring food sharing: A statistical test of the early weaning hypothesis. J. Hum. Evol. 14, 255–261. Lonsdorf, E. V., Eberly, L. E., and Pusey, A. E. (2004). Sex differences in learning in chimpanzees. Nature 428, 715–716. Maestripieri, D., and Call, J. (1996). Mother–infant communication in primates. Adv. Study Behav. 25, 613–642. Maestripieri, D., Ross, S. K., and Megna, N. L. (2002). Mother–infant interactions in western lowlands gorillas (Gorilla gorilla gorilla): Spatial relationships, communication, and opportunities for social learning. J. Comp. Psychol. 116, 219–227. Malcolm, J. R., and Marten, K. (1982). Natural selection and the communal rearing of pups in African wild dogs (Lycaon pictus). Behav. Ecol. Sociobiol. 10, 1–13. McGrew, W. C. (1975). Patterns of plant food sharing by wild chimpanzees. In ‘‘Contemporary Primatology: Proceedings of the 5th International Congress on Primatology, Nagoya August 1974’’ (S. Kondo, M. Kawai, and A. Ehara, Eds.), pp. 304–309. S. Karger, Basel, Switzerland. Midford, P. E., Hailman, J. P., and Woolfenden, G. E. (2000). Social learning of a novel foraging patch in families of free-living Florida scrub jays. Anim. Behav. 59, 1199–1207. Mitani, J. C., and Watts, D. (1997). The evolution of non-maternal caretaking among anthropoid primates: Do helpers help? Behav. Ecol. Sociobiol. 40, 213–220. Mitani, J. C., and Watts, D. P. (1999). Demographic influences on the hunting behavior of chimpanzees. Am. J. Phys. Anthropol. 109, 439–454. Mitani, J. C., and Watts, D. P. (2001). Why do chimpanzees hunt and share meat? Anim. Behav. 61, 915–924. Moehlman, P. D., and Hofer, H. (1997). Cooperative breeding, reproductive suppression, and body mass in canids. In ‘‘Cooperative Breeding in Mammals’’ (N. G. Solomon and J. A. French, Eds.), pp. 76–128. Cambridge University Press, Cambridge. Moody, M. I., and Menzel, E. W. (1976). Vocalizations and their behavioral contexts in the tamarin Saguinus fuscicollis. Folia Primatol. 25, 73–94. Nettelbeck, A. R. (1998). Observations on food sharing in wild lar gibbons (Hylobates lar). Folia Primatol. 69, 386–391. Nicolson, N. A. (1986). Infants, mothers, and other females. In ‘‘Primate Societies’’ (B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham, and T. T. Strusaker, Eds.), pp. 330–342. University of Chicago Press, Chicago, IL. Nishida, T., and Turner, L. A. (1996). Food transfer between mother and infant chimpanzees of the Mahale Mountains National Park, Tanzania. Int. J. Primatol. 17, 947–968.

FOOD TRANSFER IN NONHUMAN PRIMATES

293

Nishida, T., Wrangham, R. W., Goodall, J., and Uehara, S. (1983). Local differences in plantfeeding habits of chimpanzees between the Mahale Mountains and Gombe National Park, Tanzania. J. Hum. Evol. 12, 467–480. Nishida, T., Hasegawa, T., Hayaki, H., Takahata, Y., and Uehara, S. (1992). Meat-sharing as a coalition strategy by an alpha male chimpanzee? In ‘‘Topics in Primatology,’’ ‘‘Human Origins’’ (T. Nishida, W. C. McGrew, P. Marler, M. Pickford, and F. B. M. de Waal, Eds.), Vol. 1, pp. 159–174. University of Tokyo Press, Tokyo. Perry, S., and Rose, L. (1994). Begging and transfer of coati meat by white-faced capuchin monkeys, Cebus capuchinus. Primates 35, 409–415. Price, E. C. (1992a). Contributions to infant care in captive cotton-top tamarins (Saguinus oedipus): The influence of age, sex, and reproductive status. Int. J. Primatol. 13, 125–141. Price, E. C. (1992b). The benefit of helpers: Effects of group and litter size on infant care in tamarins (Saguinus oedipus). Am. J. Primatol. 26, 179–190. Price, E. C., and Feistner, A. T. C. (1993). Food sharing in lion tamarins: Tests of three hypotheses. Am. J. Primatol. 31, 211–221. Price, E. C., and Feistner, A. T. C. (2001). Food sharing in pied bare-faced tamarins (Saguinus bicolor bicolor): Development and individual differences. Int. J. Primatol. 22, 231–2241. Price, E. C., and McGrew, W. C. (1990). Cotton-top tamarins (Saguinus(o.) Oedipus) in a semi-naturalistic colony. Am. J. Primatol. 20, 1–12. Queyras, A., Scolavino, M., Puopolo, M., and Vitale, A. (2000). Social influence on induced food preference in common marmosets (Callithrix jacchus). Folia Primatol. 71, 367–374. Rapaport, L. G. (1999). Provisioning of young in golden lion tamarins (Callitrichidae, Leontopithecus rosalia): A test of the information hypothesis. Ethology 105, 619–636. Rapaport, L. G. (2001). Food transfer among adult lion tamarins: Mutualism, reciprocity or one-sided relationships? Int. J. Primatol. 22, 611–629. Rose, L. M. (2001). Meat and the early human diet: Insights from neotropical primate studies. In ‘‘Meat-Eating and Human Evolution’’ (C. B. Stanford and H. T. Bunn, Eds.), pp. 141–159. Oxford University Press, Oxford. Ross, C., and Jones, K. E. (1999). Socioecology and the evolution of primate reproductive rates. In ‘‘Comparative Primate Socioecology’’ (P. C. Lee, Ed.), pp. 73–110. Cambridge University Press, Cambridge. Ross, C., and MacLarnon, A. (2000). The evolution of non-maternal care in anthropoid primates: A test of the hypotheses. Folia Primatol. 71, 93–113. Rothe, H., and Darms, K. (1993). The social organisation of marmosets: A critical evaluation of recent concepts. In ‘‘Marmosets and Tamarins; Systematics, Behaviour, and Ecology’’ (A. B. Rylands, Ed.), pp. 11–77. Oxford University Press, Oxford. Roush, R. S., and Snowdon, C. T. (2000). Quality, quantity, distribution and audience effects on food calling in cotton-top tamarins. Ethology 106, 673–690. Roush, R. S., and Snowdon, C. T. (2001). Food transfer and development of feeding behavior and food-associated vocalizations in cotton-top tamarins. Ethology 107, 415–429. Ruiz-Miranda, C. R., Kleiman, D. G., Dietz, J. M., Moraes, E., Grativol, A. D., Baker, A. J., and Beck, B. B. (1999). Food transfers in wild and reintroduced golden lion tamarins, Leonotopithecus rosalia. Am. J. Primatol. 48, 305–320. Russell, A. F., Clutton-Brock, T. H., Brotherton, P. N. M., Sharpe, L. L., McIlrath, G. M., Dalerum, F. D., Cameron, E. Z., and Barnard, J. A. (2002). Factors affecting pup growth and survival in co-operatively breeding meerkats Suricata suricatta. J. Anim. Ecol. 71, 700–709. Russon, A. E. (1997). Exploiting the expertise of other. In ‘‘Machiavellian Intelligence II: Extensions and Evaluations’’ (A. Whiten and R. W. Byrne, Eds.), pp. 174–206. Cambridge University Press, Cambridge.

294

GILLIAN R. BROWN et al.

Russon, A. E. (2003). Developmental perspectives on great ape traditions. In ‘‘The Biology of Traditions: Models and Evidence’’ (D. Fragaszy and S. Perry, Eds.), Cambridge University Press, Cambridge. Schessler, T., and Nash, L. T. (1977). Food sharing among captive gibbons (Hylobates lar). Primates 18, 677–689. Sherry, D. F., and Galef, B. G. (1984). Cultural transmission without imitation: Milk bottle opening by birds. Anim. Behav. 32, 937–938. Silk, J. B. (1978). Patterns of food sharing among mother and infant chimpazees at Gombe National Park, Tanzania. Folia Primatol. 29, 129–141. Silk, J. B. (1979). Feeding, foraging, and food sharing behavior of immature chimpanzees. Folia Primatol. 31, 123–142. Snowdon, C. T. (1996). Infant care in cooperatively breeding species. Adv. Study Behav. 25, 643–689. Snowdon, C. T. (2001). Social processes in communication and cognition in callitrichid monkeys: A review. Anim. Cogn. 4, 247–257. Snowdon, C. T., and Boe, C. Y. (2003). Social communication about unpalatable foods in tamarins. J. Comp. Psychol. 117, 142–148. Stanford, C. B. (1999). ‘‘The Hunting Apes: Meat Eating and the Origins of Human Behavior.’’ Princeton University Press, Princeton, NJ. Stanford, C. B., and Bunn, H. T. (Eds.) (2001). ‘‘Meat-Eating and Human Evolution.’’ Oxford University Press, Oxford. Starin, E. D. (1978). Food transfer by wild titi monkeys (Callicebus torquatus torquatus). Folia Primatol. 30, 145–151. Stevens, J. R. (2004). The selfish nature of generosity: Harassment and food sharing in primates. Proc. R. Soc. Lond. B Biol. Sci. 271, 451–456. Stevens, J. R., and Cushman, F. A. (2004). Cognitive constraints on reciprocity and tolerated scrounging. Behav. Brain Sci. ( in press). Stevens, J. R., and Gilby, I. C. (2004). A conceptual framework for nonkin food sharing: Timing and currency of benefits. Anim. Behav. 67, 603–614. Stevens, J. R., and Stephens, D. W. (2002). Food sharing: A model of manipulation by harassment. Behav. Ecol. 13, 393–400. Takahata, Y., Hasegawa, T., and Nishida, T. (1984). Chimpanzee predation in the Mahale mountains from August 1979 to May 1982. Int. J. Primatol. 5, 213–233. Tardif, S. D. (1997). The bioenergetics of parental behavior and the evolution of alloparental care in marmosets and tamarins. In ‘‘Cooperative Breeding in Mammals’’ (N. G. Solomon and J. A. French, Eds.), pp. 11–33. Cambridge University Press, Cambridge. Teleki, G. (1973). ‘‘The Predatory Behavior of Wild Chimpanzees.’’ Bucknell University Press, Lewisburg, PA. Terkel, J. (1996). Cultural transmission of feeding behavior in the black ratRattus rattus). In ‘‘Social Learning in Animals: The Roots of Culture’’ (C. M. Heyes and B. G. Galef, Eds.), pp. 17–47. Academic Press, San Diego, CA. Trivers, R. L. (1974). Parent–offspring conflict. Am. Zool. 14, 249–264. Utami, S. S., and van Hooff, J. A. R. A. M. (1997). Meat-eating by adult female Sumatran orangutans (Pongo pygmaeus abelii). Am. J. Primatol. 43, 159–165. Vitale, A., and Queyras, A. (1997). The response to novel foods in common marmoset (Callithrix jacchus): The effects of different social contexts. Ethology 103, 395–403. Washabaugh, K. F., Snowdon, C. T., and Ziegler, T. E. (2002). Variations in care for cottontop tamarin, Saguinus oedipus, infants as a function of parental experience and group size. Anim. Behav. 63, 1163–1174.

FOOD TRANSFER IN NONHUMAN PRIMATES

295

Westergaard, G. C., Kuhn, H. E., Babitz, M. A., and Suomi, S. J. (1998). Aimed throwing as a means of food transfer between tufted capuchins (Cebus apella). Int. J. Primatol. 19, 123–131. Westergaard, G. C., Haynie, M. K., Lundquist, A. L., and Suomi, S. J. (1999). Carrying, sharing, and hand preference in tufted capuchins (Cebus apella). Int. J. Primatol. 20, 153–162. Whiten, A., Goodall, J., McGrew, W. C., Nishida, T., Reynolds, V., Sugiyama, Y., Tutin, C. E. G., Wrangham, R. W., and Boesch, C. (1999). Cultures in chimpanzees. Nature 399, 682–685. Winterhalder, B. (1996). Social foraging and the behavioral ecology of intragroup resource transfers. Evol. Anthropol. 5, 46–57. Wolters, H. J. (1978). Some aspects of role taking behaviour in captive family groups of the cotton-top tamarin (Saguinus oedipus oedipus). In ‘‘Biology and Behaviour of Marmosets’’ (H. Rothe, H. Wolters, and J. P. Hearn, Eds.), pp. 259–278. Eigenverlag, Go¨ttingen, Germany. Wrangham, R. W. (1975). ‘‘The behavioural ecology of chimpanzees in Gombe National Park, Tanzania.’’ Unpublished Ph.D. thesis. University of Cambridge, Cambridge. Wright, P. C. (1984). Biparental care in Aotus trivirgatus and Callicebus moloch. In ‘‘Female Primates: Studies by Women Primatologists’’ (M. F. Small, Ed.), pp. 59–75. Alan Liss, New York. Yamagiwa, J. (1992). Functional analysis of social staring behavior in an all-male group of mountain gorillas. Primates 33, 523–544. Zentall, T. R., and Galef, B. G. (1988). ‘‘Social Learning: Psychological and Biological Perspectives.’’ Lawrence Erlbaum Associates, Mahwah, NJ.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Song Syntax in Bengalese Finches: Proximate and Ultimate Analyses Kazuo Okanoya faculty of letters, chiba university chiba 263-8522, japan and precursory research for embryonic science and technology (presto) japan science and technology agency, tokyo 102-8666, japan

I. Introduction A. Bird Song and Tinbergen’s Four Questions The study of bird song is a favorite textbook example in ethology (Alcock, 2001; Manning and Dawkins, 1998; McFarland, 1999; Slater, 1999). Bird song is a behavior well suited to quantitative description, controlled by a defined neural circuitry and functioning in sexual selection. It also has variability that makes evolutionary comparisons possible. It naturally fits into the four questions of mechanism, development, function, and evolution, which Tinbergen (1963) defined as the subject matter of ethology. 1. Divergence of Bird Song Studies Tinbergen’s approach was the agenda for ethologists, but as the study of behavior advanced, as Willson (1975) predicted, the study of proximate causes became integrated into neuroethology (Ewert, 1976) while that of ultimate causes was integrated into behavioral ecology (Krebs and Davies, 1993). With the conceptual and technological advances of these sibling disciplines, it became more difficult to integrate the results into one perspective (Eens, 1997). One reason for this might be that it was difficult to maintain the techniques required to ask both ultimate and proximate questions within one research group. Cooperation between several groups to cover the whole range of Tinbergen’s questions is also often difficult because species used by researchers on ultimate questions and those on proximate causes are often different. In bird song neurobiology, most 297 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

298

KAZUO OKANOYA

studies have been conducted on zebra finches (Taeniopygia guttata) or canaries (Serinus canaria). Field workers focusing on the ultimate causes of bird song seldom use these species so that integrating knowledge from both sides is simply not possible. 2. Convergence of Bird Song Studies As ideas about the ultimate causes of behavior have developed and experimental evidence for its proximate causes have accumulated, the strength of Tinbergen’s integrative approach in this field has become appreciated again. For example, the predictions of sexual selection have now been examined as a cost of singing to the male brain (Airey and DeVoogd, 2000), and in the female brain as a cost of mate evaluation (Hamilton et al., 1997). The idea of honest signaling (Zahavi and Zahavi, 1997) has been examined at the level of immunology (Møller et al., 2000) and new theories supplementing this idea have been emerging (Nowicki et al., 2002). Advanced molecular techniques have been utilized for the study of song perception (Mello et al., 1992) and production (Jarvis et al., 1998). These advances have been further improved through collaboration between behavioral ecologists and neurobiologists (e.g., Catchpole, 2000; Freeberg et al., 2002). In our laboratory, we have employed an integrative approach to bird song, with studies ranging from molecular to evolutionary analyses. To keep the integrative approach manageable, we focus on only one aspect of song in a single species: song complexity in Bengalese finches (Lonchura striata var. domestica). B. Auditory Feedback in Bengalese Finches Bengalese finches are well adapted to the domestic environment and show the full range of natural behavior in the laboratory. The wild strain of the Bengalese finch, the white-backed munia (Lonchura striata), is widespread in Southeast Asia, making field experiments and evolutionary comparisons possible (Mizuta et al., 2003). Moreover, since the Bengalese finch belongs to the same family as the zebra finch, what is known about the neuroscience of zebra finch song is largely applicable to that of Bengalese finches. In this article we show the strength of the integrative approach put forward by Tinbergen to the full understanding of behavioral questions. Our choice of study species was partly determined by a chance finding that the male Bengalese finch relies critically on auditory feedback to maintain adult song (Okanoya and Yamaguchi, 1997). As soon as auditory feedback was removed, the bird began stuttering or jumbling its song

SONG SYNTAX IN BENGALESE FINCHES

299

Fig. 1. An example of song ‘‘stuttering’’ in Bengalese finches after deafening. Before surgery, the bird was singing the sequence ‘‘ABCDE’’ as an initial part of his song. After the bilateral removal of the cochleae, the bird stuttered after song note A or C. Modified from Okanoya and Yamaguchi (1997).

sequences and, later, acoustical deterioration of each song note followed (Fig. 1). 1. Deafening Studies in Bengalese Finches The necessity for auditory feedback in the maintenance of the crystallized song was previously shown in zebra finches by Nordeen and Nordeen (1992). Songs of deafened zebra finches gradually deteriorated over several months. But the song deterioration observed in Bengalese finches was much more rapid, observable on the day of surgery and, thus, more dramatic than that of the zebra finch. Canaries are known to learn new songs every year and so classified as open-ended learners (Nottebohm et al., 1986). On the other hand, zebra finches and Bengalese finches learn their song only once during their sensitive period, and are thus classified as close-ended learners (Clayton, 1987; Price, 1979; but see Jones et al., 1996), and these species are thought not to require auditory feedback to maintain crystallized song in adulthood. Our observations on Bengalese finches were a serious challenge to the traditional view of song maintenance, which states that auditory feedback is not necessary after song is crystallized (Konishi, 1964, 1965). Our original finding in Bengalese finches was duplicated by other laboratories (Brainard and Doupe, 2000a,b; Watanabe and Aoki, 1998; Woolley and Rubel, 1997). Furthermore, there have been several new extensions of the results. First, the degree of auditory dependence in adulthood has been shown to be correlated with the degree of neural

300

KAZUO OKANOYA

turnover in the central vocal–auditory nucleus (Scott et al., 2000); the rate of neurogenesis in the forebrain nucleus HVC was several times faster in adult Bengalese finches than in adult zebra finches. This rate in adult Bengalese finches rivals that of the canary, which continues to learn new song in adulthood. Second, reversible auditory deprivation by ototoxic drug and noise exposure in adult Bengalese finches was found to result in gross modification of crystallized song, often including the production of new song syllables, when recovering from temporary deafness (Woolley and Rubel, 2002). That is, by temporarily deafening them, adult Bengalese finches could be pushed back to the brain state before the sensitive period. These new findings provide a unique opportunity to study song learning and maintenance and question the distinction between open/close-ended learners. Studies by Scott et al. (2000) and Woolley and Rubel (2002) might be interpreted as showing that there are similar underlying processes for song maintenance and song learning. In Bengalese finches, at least, song learning might occur when the discrepancy between the auditory template and the motor output is large, but the timing of learning is not as restricted as once thought. Research on the relationship between auditory feedback and song maintenance has traditionally been focused mainly on species with simple, stereotyped repertoires (Konishi, 1964, 1965; Nordeen and Nordeen, 1992). The difference between the previous results and our results on Bengalese finches seemed to stem from the complexity of the song. Unlike the previously studied species, Bengalese finch songs could not be analyzed in terms of a simple linear song model in which note-to-note transitions were more or less determined and a song motif could be clearly defined. Bengalese finch songs have complex note-to-note sequences that do not permit such clear definitions. It is probably because this species modifies ongoing song in real time that auditory feedback is so crucial in Bengalese finches. 2. Effect of Helium Bilateral removal of the cochleae has produced convincing data on the necessity of real-time auditory feedback in Bengalese finches. However, this is an invasive and irreversible procedure that does not allow further neuroethological investigation. As a procedure to alter auditory feedback reversibly, we placed eight adult male Bengalese finches in a helium atmosphere that changed the resonance properties of the vocal tract (Yamada and Okanoya, 2001, 2003). This procedure was originally used by Nowicki (1987) to show that birds use vocal tract resonance to shape vocalizations. Rather than describing the resonance properties of the Bengalese finch vocal tract, we aimed to reversibly change auditory

SONG SYNTAX IN BENGALESE FINCHES

301

feedback from the self-produced song and see the effect of this on real-time production of the note sequence. Undirected songs were recorded before, during, and after this treatment and the phonological structure of the songs produced was analyzed. Songs were recorded for 30 min in a small chamber before the bird was placed in the helium. The chamber was then filled with helium within 3 min and songs in this helium were recorded for 30 min. After that, the helium was immediately removed and songs in normal air were again recorded for 30 min. Two types of phonological changes were observed during this experiment: the higher harmonics of those song elements that had harmonic structures changed in amplitude while some other song elements that had a single peak frequency and no obvious harmonic structures changed in their peak frequency. The former types of change are well-known results of resonant changes in helium, but the latter types of change could not be accounted for by such simple mechanical factors. Peak frequency change in the helium atmosphere has been found only in Bengalese finches and suggests that these are active processes of sound modulation possibly mediated by auditory feedback in this species (Yamada and Okanoya, 2001). Furthermore, we observed that six of eight birds changed their song syntax as described by transition diagrams when placed in the helium air. These birds added new types of note-to-note transitions that were not observed in the song sung in normal air (Yamada and Okanoya, 2003). This was exactly what we expected to see in Bengalese finches. We interpreted this phenomenon as follows: the spectral changes that occurred in the helium air changed auditory feedback from the self-produced songs; the bird’s sensory–motor integration mechanisms actively tried to compensate for these changes and this resulted in syntactical modifications. We think this preparation would be useful for study of the neurophysiological processes involved in real-time vocal control. This research has shown that ongoing auditory feedback is actively processed by male Bengalese finches when they are singing. In summary, our deafening and helium studies on Bengalese finches have established that the complex note-to-note transitions in the songs of this species are actively controlled by real-time sensory motor integration. As discussed below, Bengalese finches are a domesticated form of the whitebacked munia and their songs might also have been influenced by the process of domestication. This consideration involves not only proximate, but also ultimate, questions. This is the reason why we selected song complexity in Bengalese finches as our target for integrative behavioral study.

302

KAZUO OKANOYA

II. Origin of Bengalese Finches As a basis for our integrative behavioral study on the song complexity of Bengalese finches, we first needed to establish the origin of Bengalese finches. This step was very important for us to conduct evolutionary comparisons and to relate ultimate and proximate questions. A. Avicultural Records The Bengalese finch does not occur in the wild. This species was suspected to be a domesticated strain of the white-backed munia (Lonchura striata), but only on the basis of avicultural literature (Taka-Tsukasa, 1917). From the literature, several white-backed munias were imported into Japan in 1762. The birds were then domesticated in Japan because of their tameness and reproductive potency. In an avicultural note published in 1856, mutations with generally white plumage and some brown patches were described in the whitebacked munia and this strain became popular as Bengalese finches. In those days, one pair of Bengalese finches with white plumage cost as much as $4000! But there are no records in the avicultural literature to indicate that Bengalese finches were selected for their songs (Washio, 1996). Nevertheless, we found that the song syntax differs between the white-backed munia and the Bengalese finch (Honda and Okanoya, 1999). In Europe, white-backed munias were imported independently from East Asia to Belgium and England (Eisner, 1957, 1960; Hernalsteen and Iwens, 1995). Thus, we can identify at least three lines of ‘‘Bengalese’’ finches: wild white-backed munias, European Bengalese finches, and Japanese Bengalese finches. As mentioned in a later section, songs of the European Bengalese finches have an intermediate level of transition complexity between white-backed munias and Japanese Bengalese finches (Woolley and Rubel, 1997; Yodogawa, 2000). B. Comparison of Calls In an effort to establish the relationship between the white-backed munia and the Bengalese finch, we examined one of the most common calls in finches, the distance contact call (Okanoya, 1997). The distance call is emitted when a bird is visually separated from other birds (Zann, 1996). Distance calls are special in Bengalese finches and in zebra finches in that these calls are sexually dimorphic (Guettinger and Nicolai, 1973). In Bengalese finches, male calls are narrow banded, pure tonelike vocalizations

SONG SYNTAX IN BENGALESE FINCHES

303

with rising frequency modulation while female calls are wide-banded, pulse trainlike utterances (Okanoya and Kimura, 1993; Yoneda and Okanoya, 1991). The calls of wild white-backed munias have a pattern identical to that of Bengalese finches (Okanoya et al., 1995). Sex as determined by a molecular technique and that estimated from distance calls matched perfectly in wild white-backed munias (Mizuta et al., 2003). These results indicated that, at least from their calls, these two groups of birds are very close.

C. DNA Hybridization Studies To establish the genetic relationship between the Bengalese finch and white-backed munia, we utilized the RAPD (randomly amplified polymorphic DNA) molecular genetic method (Lessells and Mateman, 1998). This method relies on the chance matching of a random primer sequence and the DNA sequence of the animal. We used 5 random primers with 10 bases. The bands from polymerase chain reaction (PCR) were separated by gel electrophoresis, and we observed the pattern of bands that appeared for different species and processed these patterns using cluster analysis (Y. Yodogawa, unpublished data). In the first experiment, we tried to refute several existing hypotheses on the origin of the Bengalese finch. These include the silver bill and the chesnut-brested finch as candidates. We included DNA samples from these birds and samples from Bengalese finches, white-backed munias, and a zebra finch. We analyzed several patterns of electrophoresis that resulted from the RAPD method. Bengalese finches and white-backed munias were linked into one cluster while the silver bill, the chesnut-breasted finch, and the zebra finch formed an independent cluster. These results clearly demonstrate that white-backed munias and Bengalese finches are indeed closely related. In the next experiment, we tested genetic relatedness of wild whitebacked munias, European breeds of Bengalese finches, and Japanese breeds of Bengalese finches. The same RAPD procedures and cluster methods were used. The resulting dendrogram showed that both the Japanese and European breeds of Bengalese finch were related to subpopulations of white-backed munias (Fig. 2). These results, taken together, establish that both Japanese and European Bengalese finches were derived directly from wild white-backed munias from Southeast Asia. Having these data, we can now ask evolutionary questions by comparing songs of white-backed munias and Bengalese finches.

304

KAZUO OKANOYA

Fig. 2. Results of RAPD analyses of Bengalese finches and related species. The gel electrophoresis pattern was analyzed by the furthest neighbor clustering method. WBM, white-backed munias; Bengalese, Japanese Bengalese finches; Europe, European breeds of Bengalese finches; Zebra, zebra finch. The clustering pattern indicates that both Japanese and European breeds of Bengalese finches were derived from wild white-backed munias.

III. Analyses of Bengalese Finch Songs A. Traditional Terms Traditionally, bird song has been analyzed using terms, such as notes, syllables, and phrases (Catchpole and Slater, 1995). A song note is the smallest unit of utterance recognizable as a continuous trace on the sonogram. A syllable is a minimum unit of repetition within the song, usually a combination of a few notes, or sometimes equal to a particular note itself when such a note is itself a unit of repetition. A song phrase is built of a number of syllables. A motif refers to a stereotyped sequence within a song (Sossinka and Boehner, 1980) and a strophe refers to a bout of motifs (Price, 1979). In species whose songs are highly stereotyped, such as the swamp sparrow Melospiza georgiana (Clark et al., 1987), song sparrow Melospiza melodia (Podos et al., 1992), chaffinch Fringilla coelebs (Thorpe, 1961), white-crowned sparrow Zonotrichia leucophrys (Petrinovich et al., 1972), and zebra finch (Price, 1979), these definitions were suitable and functioned as a common framework within which to discuss song structures.

SONG SYNTAX IN BENGALESE FINCHES

305

B. Finite State Syntax in Bengalese Finches We found it difficult to analyze the song of Bengalese finches using these traditional terms. A stereotyped song motif could not be firmly identified in the song episodes of a Bengalese finch. This is because songs of Bengalese finches have complex note-to-note transition patterns. One note could be followed by several possible notes in a song. Multiple notes are organized into chunks, but these chunks do not have a stereotyped structure; a note could be used in several different chunks and some notes could be repeated several times. These chunks are in turn sung with probabilistic rules; one chunk could be followed by several different chunks. The most suitable notation to describe this complex organization turned out to be from linguistics. A finite state automaton, or syntax, describes several states connected by arrows (Fig. 3). One state can transit to several possible states, each associated with some probability. When a transition

Fig. 3. An example of finite state syntax derived from a Bengalese finch song. Top: A sonogram of part of the song. Each song note is denoted by a letter and song bouts were converted into strings of letters. These strings are statistically analyzed and song notes that occur together are chunked. Chunks are shown by brackets on top of the sonogram. Bottom: The transitive statistical relationships between the chunks are indicated by the finite state syntax. Transition from one state to another produced certain chunks shown in this diagram. There are several possible routes through this diagram. This syntax can produce all possible song sequences produced by this bird.

306

KAZUO OKANOYA

occurred, a string could be produced (Manning and Schutze, 1999). We utilized this notation to describe Bengalese finch songs by identifying strings associated with each state transition. To find the finite state syntax of a Bengalese finch song, we first denoted each unique note in the sequence with a letter of the alphabet. Thus, a song sequence can be expressed as a string of letters. Suppose a bird sings the following songs: S1: abcdeabcdeabdefgabcdefgabcdeabcdeab S2: abcdeabcdeabcdeabcdeabcdedecabcdefgabcdeabcdeabcdeabcdefgab S3: deabcdeabcdefgabcdeabcdeabcdeabcdeabcdefgabcdeabfabcdeab 

If we had more such data, we might rewrite the above strings by utilizing rules such as Rule 1: ab ! 1; Rule 2: cde ! 2; Rule 3: fg ! 3 Using these rules, we obtain S1: 12121de312312121 S2: 1212121212dec1231212121231 S3: de1212312121212123121f121  These rules enable us to rewrite the original strings into chunk-strings, but remaining are those small portions of the strings that could not be rewritten with the rules (usually less than 5% of the original song notes). Within these sequences, we can identify phrases, such as 1-2, 1-2-1, 1-2-3-1, and so on. We then have a finite state syntax based on this bird’s song such as that shown in Fig. 3. When one takes a different path in the finite state syntax, then one obtains a new phrase and these phrases are organized into a bout. To identify the finite state syntax, we need a procedure to detect chunks of strings that are produced when transition between states occurs. Most transition analyses in animal behavior rely on a simple Markov model in which the probability of one note being followed by the others is calculated and a simple first-order transition probability matrix constructed (Manning and Schutze, 1999). This model has been used efficiently to describe differences in directed and undirected songs (Sossinka and Boehner, 1980), and effects of central (Scharff and Nottebohm, 1991) or peripheral lesions (Okanoya and Yamaguchi, 1997). C. Song Linearity On the basis of the first-order transition matrix, we can also calculate an index of song linearity by dividing the number of note types in the song by the number of nonzero elements of the transition matrix (Scharff and

SONG SYNTAX IN BENGALESE FINCHES

307

Nottebohm, 1991). The value of this index will be close to 1 when the song is very deterministic so that the same note always follows a particular one. When song is completely random and any note could follow any other note, then the index will be small, close to 0. Thus, song linearity was defined as follows. Let N be the number of unique song notes and P be the number of observed note-to-note transition patterns; then song linearity ¼ N/P. Song linearity defined in this way varies from 1.0 (when the song is completely linear and deterministic) to N/N2 ¼ 1/N (when the song is completely random, so that each note can be followed by all note types, in this case P ¼ N2). For the results discussed in this chapter we used a simple Markov model (Chatfield and Lemon, 1970) to construct note-to-note transition matrices. When necessary, we used linearity measures to compare degree of song complexity quantitatively among two or more songs. D. Comparison with Other Species with Complex Songs Besides Bengalese finches, there are many other species that sing complex songs. Among these, several procedures have been developed to describe song complexity. Here we compare procedures developed for the nightingale (Luscinia megarhynchos) and the European starling (Sturnus vulgaris) with that we use for Bengalese finches. 1. Nightingales In nightingales, a song type (lasting about 4 s) is a fixed sequence of several song notes. An individual nightingale possesses 100–200 song types and these song types are clustered into 2–12 ‘‘packages.’’ The order of song types arranged into the package is unidirectional or multidirectional, depending on that individual. The order of singing of the packages is probabilistic (Todt and Hultsch, 1996). When hand-raised with artificial tutoring stimuli, nightingales spontaneously grouped three to seven song types into one package (Hultsch and Todt, 1989). In this fashion, nightingale song has a hierarchical structure of song type, package, and song bout. 2. Starlings A starling song bout may last up to 1 min and has a complex organization. It is composed of many distinct song types that occur once or repeatedly before the next song type. A song type contains song elements in a fixed order and lasts 0.5–1.5 s. A complete bout typically includes four distinct sections consisting of song types that occur at these particular points in the song. In this fashion, starling songs have a stereotyped gross structure common to nearly all bouts (Eens, 1997). Gentner and Hulse

308

KAZUO OKANOYA

(1998) used first-, second-, and third-order Markov models to analyze starling songs. They took each song type as a unit in their analyses. The uncertainty of the song dropped rapidly when the second-order Markov model was applied, indicating that triplet relations between song types is sufficient to predict the sequence within a song bout in starlings. Thus, like nightingale songs, starling songs also have a hierarchical structure with song types, clusters of song types, and song bouts. 3. Willow Warblers Gil and Slater (2000) provided an extensive analysis on songs of the willow warbler (Phylloscopus trochilus). The organization of song elements can be expressed by a hierarchically branching pattern. The transitional probabilities of one element to another can vary among the elements: one element could always be followed by another, but another element could be followed by several possible elements. Thus willow warbler songs could be expressed by dendrograms of branching patterns. 4. Uniqueness of Bengalese Finch Song Compared with the nightingale, starling, and the willow warbler, song complexity in the Bengalese finch has some similarities but also unique properties of its own. As was done on European starling song, we analyzed Bengalese songs by using different orders of Markov models. Entropy of the estimation did not drop until the third-order model was applied to Bengalese finch song (Hosino and Okanoya, 2000). When the song type was the unit of analysis, starling songs had second-order Markov complexity. Thus, starling songs have a hierarchy one step higher than that of Bengalese finch songs, but overall sequential complexity of the song may be comparable in these two species. In the nightingale, the order of song types arranged into the package could also vary considerably among the packages, adding more variation in the overall song structure. Thus, nightingale songs have one more hierarchy level than starling songs. Syntactical complexity in Bengalese finches could be equal to that of nightingales when the package in nightingales is treated as the same hierarchy level as the song element in Bengalese finches. Willow warbler songs are similar in organization to Bengalese finch songs, except that the recursive loops seen in Bengalese finch songs are missing in willow warbler songs and, more importantly, willow warbler songs seem to have more sequential variations than do Bengalese finch songs, since each of the song episodes has a different element sequence. Bengalese finch songs have a hierarchical nature similar to that of starling and nightingale songs but the level of the hierarchy is simpler than

SONG SYNTAX IN BENGALESE FINCHES

309

in these species. On the other hand, whether or not the probabilistic relationship between the song type chunks, or the packages in nightingale songs, could be expressed as finite state syntax, has not been examined in species other than Bengalese finches. E. Summary Bengalese finch songs have complex note-to-note transition patterns. To analyze this, we decided to use the notation of finite state syntax. Each song note is denoted by a unique letter and a song bout is expressed by a string of letters. These song bouts are then subjected to Markov analyses in which the probability that one note will be followed by another is calculated. Song notes that occur together are grouped into a chunk and transition patterns from one chunk to the others are further analyzed. Results are represented by finite state syntax in which different states are connected by arrows. The hierarchical and probabilistic nature of song syntax in Bengalese finches is a simplified version of nightingale or European starling. In these species, the song type (starlings) or the group of song types (nightingales) can be treated as a unit in analyses, while in Bengalese finches it is the song note, a discrete trace on the sonogram, that is the basic unit. Therefore, Bengalese finch song has the advantage of simplicity that allows us to extract rules of song organization in a more comprehensive way than can be done in species with extraordinary complex songs and song sequences.

IV. Tinbergen’s Four Questions Bengalese finch song has a hierarchical nature similar to that of starlings and nightingales, yet is conveniently handled by finite state syntax notations. This behavior is curious because it depends on continuous auditory feedback and because this degree of complexity is absent in the ancestral species. From here on, we analyze the song complexity of Bengalese finches in terms of the four questions of Tinbergen. A. Evolution The first question to be asked is that of evolution. Since we had established that Bengalese finches are the domesticated strain of white-backed munias, we first compared the song morphology and syntax of the two strains (Honda and Okanoya, 1999) in order to estimate the nature of changes that have occurred during 240 years of domestication. Figure 4 shows examples of a white-backed munia song and a Bengalese finch song.

310

KAZUO OKANOYA

Fig. 4. Song comparison between the white-backed munia and the Bengalese finch. An example sonogram from a white-backed munia (top) and that from a Bengalese finch (bottom) are shown along with the finite state syntax derived from each song. White-backed munia syntax is basically linear while that of the Bengalese finch is complex, having several states and transitions.

SONG SYNTAX IN BENGALESE FINCHES

311

Since Bengalese finches are domesticated, the complication of song syntax did not occur in nature, but rather was probably facilitated by intervention of humans. There are no records in Japanese avicultural literature that Bengalese finches were selected for their song (Washio, 1996), but it is possible that finch breeders selected pairs that were efficient in reproduction, and males in such pairs happened to be good singers. Although there may have been such interventions, it could still be said that the process occurred because of the perceptual bias of females, facilitated only incidentally by selection of good breeding pairs. In this context, the process of domestication is assumed to be a good model with which to understand evolution. 1. Comparisons of Song Note Morphology For this analysis, songs from four individuals from each strain were used (Honda and Okanoya, 1999). Since the sample size is small, the conclusion of this analysis should be considered with caution, but later observations generally support the results obtained here. From the recording of their song episodes, a song unit that contained all song elements identified in the entire recording of each bird was selected and the song unit was further divided into elements. Five types of elements were taken from the beginning of each bird’s song unit and were separately saved as a file. Altogether, 20 elements were obtained from four individuals from each of the two strains. In this analysis each element, rather than each individual, was treated as a unit and thus average values of these elements were not used to represent the individual. The availability of experimental animals limited the sample size to four individuals in each strain. We tried to obtain simple acoustic parameters that would differentiate song notes of Bengalese finches and white-backed munias, but no such parameters were found (Honda and Okanoya, 1999). Thus, we had to rely on rather complex procedures to describe differences in song notes between the two strains. The 40 song elements taken from each strain of birds were analyzed using the CORMAT program of the SIGNAL sound analysis system (Engineering Design) to look for spectrographic similarity. CORMAT slides two sound spectrograms over each other in the time domain to find the highest spectrotemporal correlation (Clark et al., 1987). The resulting similarity index was submitted to a multidimensional scaling (MDS) program (SYSTAT, Inc.) to see whether the strains were different at the level of song element morphology. The 40 elements were scaled in two dimensions. Dimensional coordinates were taken for Bengalese finches and for white-backed munias separately. These values were then submitted to a

312

KAZUO OKANOYA

multivariate analysis of variance to test whether the strains were different in the morphological characteristics of their song elements. Phonological structures of song elements (morphology) differed between the two strains of birds. This observation was confirmed by the spectrographic correlation analyses. Multivariate analysis of variance run on the coordinate values detected a significant difference between the two strains (F ¼ 3.824, df ¼ 2, 37, p ¼ 0.031). Post hoc univariate analyses of variance (ANOVAs) detected significant differences between the strains on dimension 1 (F ¼ 7.027, df ¼ 1, 38, p ¼ 0.012) but not on dimension 2 (F ¼ 0.368, df ¼ 1, p ¼ 0.547). Dimension 1 was correlated with the acoustic parameter of the frequency of maximum amplitude (r ¼ 0.57, p < 0.01). Bengalese finch song notes had higher peak frequencies than white-backed munia song notes. 2. Comparisons of Song Complexity Songs recorded from eight Bengalese finches and five white-backed munias were used in this analysis (Honda and Okanoya, 1999). Recorded songs were analyzed on a PC running the Avisoft SAS Lab (Specht, 1997) sound analysis system. The 2-min recordings were all sonogrammed and printed. Each of the song elements was categorized into several distinct types through visual inspection and a letter designating each category was assigned. Thus, song output of each individual was expressed as a string of letters. For each song, the number of unique element types was counted. The string was analyzed and transition probabilities from each note type to others were calculated, to give a matrix of transition probabilities. On the basis of this matrix, a transition diagram visually showing the pattern of song syntax was drawn. The linearity index (Scharff and Nottebohm, 1991) was also calculated for all songs. The number of song notes used by white-backed munias (average, 8.4; range, 7–10) and Bengalese finches is the same (average, 9.3; range, 6–15). However, the average linearity is significantly lower (p < 0.001) in Bengalese finches (average, 0.33; range, 0.24–0.43) than in white-backed munias (average, 0.61; range, 0.41– 0.69). Figure 4 shows examples of transition diagrams obtained from a white-backed munia and a Bengalese finch. European aviculturists independently domesticated white-backed munias and produced European Bengalese finches. European aviculturists put more emphasis on the appearance of the finches than on their reproductive potency (Hernalsteen and Iwens, 1995). Thus, this strain of whitebacked munias is not suited for use as foster parents. We have also examined songs of this strain (Y. Yodogawa and K. Okanoya, unpublished observations). The song of the European strain was simpler than the Japanese strain, but more complex than the wild strain. Song linearity

SONG SYNTAX IN BENGALESE FINCHES

313

calculated from six individuals of the European strain was 0.49, in between those of the Japanese and wild strains. We suspect this resulted from the lower emphasis on reproductive potency in the European strain than in the Japanese strain. 3. Overall Sound Density of the Song Songs recorded from eight Bengalese finches and five white-backed munias (all songs that had been used in previous syntactical analyses) were used in this analysis (Honda and Okanoya, 1999). A 5-s portion of continuous singing was taken from the recording of each bird and a rootmean-square (RMS) value was calculated on that portion. Values were expressed as decibels (arbitrary voltage but fixed for all measurements) and compared between the two strains. This value is referred to as the sound density because the value partially depends on how tightly the elements are organized into the unit time. One sound density value was taken from each bird. Average sound density was 25 dB for the eight Bengalese finch songs and 11 dB for the five white-backed munias. The difference between the two strains was significant by t test ( p < 0.001). Since measurements were taken with identical settings for the two strains, the difference obtained here should approximate that in sound pressure level in the songs of these two strains. Bengalese finches sang, on average, 14 dB louder than white-backed munias. 4. Summary We have shown that several song characteristics are different between white-backed munias and Bengalese finches. These include song note morphology, song syntax, and sound density. However, the number of song notes used by each individual is the same between the two. These changes in song characteristics occurred during 240 years of domestication and may reflect changes in learning environment, brain circuitry, and peripheral morphology for song production. Domestication resulted in evolutionary changes in song behavior. B. Mechanisms Here we ask about the anatomical and physiological representations of song complexity and compare the results with what is known, especially in zebra finches. The comparison of brain mechanisms for song production and learning in the zebra finch and Bengalese finch is particularly interesting because these two species represent different degrees of song complexity.

314

KAZUO OKANOYA

1. General Architecture of the Song Control System in Zebra Finches What mechanisms make the complex song of Bengalese finches possible? Central mechanisms for bird song production have been well studied using zebra finches, an estrildid species related to Bengalese finches (Margoliash, 1997). In discussing these findings, we use some of the acronyms for brain areas as proper names. This is because these acronyms were constructed on the basis of neuroanatomical terms that turned out to be wrong in the light of modern anatomical knowledge. Anatomical terms for the avian forebrain have been updated to reflect modern findings (Reiner et al., 2004) and we use these updated terms below (Fig. 5). Briefly, real-time control of song production is governed by a set of discrete brain nuclei. Song is produced by combined activity of the respiratory, syringeal, and resonating apparatus (trachea, tongue, and beak). The pathway that directly controls syringeal activity has been well described and is often referred to as the posterior pathway (Margoliash, 1997). The syringeal muscles are directly controlled by the tracheosyringeal branch of the hypoglossal nerve (NXIIts, or ts for short) from nucleus XII. This nucleus receives input from both the telencephalic nucleus, RA (robust nucleus of the arcopallium), and the mesencephalic motor nucleus, the dorsomedialis (DM). Going upstream, RA is innervated by the nucleus HVC of the nidopallium. This nucleus receives auditory input from the primary auditory center, field L, and from surrounding higher auditory structures. Among them, the NIf (interfacial nucleus of the nidopallium) sends auditory/motor input to the HVC. The NIf is further innervated by the thalamic nucleus, Uva (nucleus uvaformis). Connecting the two main motor nuclei (HVC and RA) indirectly is the anterior forebrain pathway, which starts from the HVC, passes to area X of the basal ganglia (Bottjer and Arnold, 1997), innervates the thalamic nucleus (DLM) (Okuhata and Saito, 1987), and then comes back to the telencephalic nucleus (LMAN). This nucleus then connects to the RA, completing the loop (Bottjer and Arnold, 1997). Several studies have indicated the hierarchical nature of the posterior motor pathway in the zebra finch. When the HVC and RA were stimulated independently during singing, zebra finches stopped singing at different hierarchical levels of the song. Stimulation of the HVC resulted in a sudden cessation of the entire song while RA stimulation caused deterioration in the acoustic morphology of a particular song note or eliminated a particular note from the song (Vu et al., 1994). Electrophysiological recordings made from the HVC and RA when the bird was singing showed that HVC neurons were active during the entire song episode while some RA neurons were active only when a

SONG SYNTAX IN BENGALESE FINCHES

315

Fig. 5. The Bengalese finch brain. Top left: Cytochrome oxidase staining of the medial part of the sagittal section of a male Bengalese finch brain. Nuclei of the anterior pathway (LMAN and area X) are darkly stained. Top right: The same brain sectioned more laterally. Nuclei of the posterior song control pathway (NIf, HVC, and RA) show up as dark staining. Bottom: Schematic drawing of the song control pathways. See text for details.

particular note was sung (Yu and Margoliash, 1996). These results in zebra finches clearly demonstrated the hierarchical relationship between the HVC and RA. The basic architecture described here for the zebra finch is applicable to the song system of the Bengalese finch (Okanoya, 1997).

316

KAZUO OKANOYA

2. Lesion Studies on the Posterior Pathway We extended these findings in zebra finches to the Bengalese finch system by a series of lesion studies, stimulated because Bengalese finches sing songs that have higher hierarchical structure. In Bengalese finches, the song notes are organized into chunks, and each chunk is further incorporated into phrases, while in zebra finches, each song element is arranged into the song strophe in a fixed order. We hypothesized that this behavioral hierarchy is mirrored in the hierarchically organized forebrain nuclei. We made lesions at the levels of the ts nerve, RA, HVC, and NIf. Results showed that this anatomical hierarchy did indeed correspond to the behavioral hierarchy. Results of lesions at each level are described below. a. ts nerve and RA When the left ts nerve was severed, the phonology of song notes that had a higher (>2.0 kHz) fundamental frequency deteriorated. On the other hand, when the right ts nerve was cut, song notes with a lower fundamental frequency (<1.5 kHz) degenerated (Fig. 6). Thus, at the level of the ts nerve, production of song notes is lateralized according to the fundamental frequency (Okanoya, 2004). This is in contrast with most songbirds studied to date, in which the left side is responsible for lower fundamentals (Suthers, 1997). The reason why the role of the two sides is reversed in Bengalese finches is not known, but the range of the fundamental frequency (2–4 kHz) generated by the left ts nerve in Bengalese finches falls within the range of that produced by the left ts nerve in most birds (Suthers, 1997) while that produced by the right ts nerve of the Bengalese

Fig. 6. Examples of songs produced in the ts nerve-cutting experiments. There were two birds that sang an identical sequence in their songs (left) and this portion was compared for the cases of left ts nerve section (middle) and right ts nerve section (right). In the left ts nerve-cut bird, song notes indicated by ‘‘H’’ with fundamental frequency higher than 2 kHz deteriorated and the song note indicated by ‘‘L’’ with fundamental frequency lower than 1.5 kHz remained intact. In the right ts nerve-cut bird, song notes labeled as ‘‘H’’ remained intact while the ‘‘L’’ note deteriorated.

SONG SYNTAX IN BENGALESE FINCHES

317

finches (<1.5 kHz) is extraordinary low for bird sounds, suggesting that it was the function of the right syrinx that underwent substantial modifications. To lesion the nucleus RA, we used a radiofrequency electrode that produced heat within the brain tissue. Thus lesions involved both cell bodies and fibers en route. RA is a large nucleus, so that a complete lesion was not possible. The degree of lesion averaged around 30% of the entire nucleus. We made unilateral lesions in eight Bengalese finches. Partial lesions made in the left RA resulted in partial loss of the high fundamental (>2.0 kHz) song notes while lesions in the right RA eliminated some of the lower fundamental notes (<1.5 kHz) (K. Okanoya and N. Hirata, unpublished observation). Results were similar to the ts nerve lesion cases as they concerned the side of dominance for each class of song notes, but different in that song notes did not simply deteriorate but were eliminated by this operation (Fig. 7). These results are in good agreement with the neuroanatomical connections from the RA to the nXIIts. In zebra finches, there are two branches of fiber bundle output from the RA (Wild, 1993, 1997). One of these is connected to the nXIIts while the other innervates the nucleus retroambiguus, which controls respiration. Lesioning the RA partially eliminated the motor program for the syringeal control for particular song notes, and also the motor program for the respiratory control of those notes. Cutting the ts nerve unilaterally, on the other hand, did not eliminate the motor command for the respiratory control, and thus those song notes simply deteriorated, but were not eliminated. In summary, at the level of the RA and below, the song control system is lateralized for the fundamental frequencies of the song notes. Furthermore, the motor command is divided into that for respiration and that for phonation. b. HVC The RA is innervated by the HVC and lateralization of function similar to the level of the RA was expected for the HVC. However, complete lesion of the left HVC resulted in severely damaged song in which scarcely any of the preoperative song notes could be identified. On the other hand, complete lesion of the right HVC resulted in transient deterioration of the song lasting about 2–4 weeks, while the preoperative song structures completely recovered after 1 month (Okanoya et al., 2001). Thus at the level of the HVC, the left side is responsible for real-time production of the song, but the role of the right HVC was not elucidated. Since the functional laterality of the HVC was so strong, we decided to concentrate on the study of the left side in producing the song. Instead of complete lesions, we tried to make partial lesions (N ¼ 7) in the left HVC (Uno and Okanoya, 1998). In three control birds, the electrode was inserted but no current was applied. The size of lesion varied from 6 to 76%, but its extent could be

318

KAZUO OKANOYA

Fig. 7. Examples of RA lesion studies. Top: A case with partial lesions in the left RA. High fundamental song notes indicated by asterisks disappeared after the surgery. Bottom: A case with partial lesions in the right RA. Many low fundamental song notes indicated by asterisks disappeared after surgery. Data from K. Okanoya and H. Hirata (unpublished data).

SONG SYNTAX IN BENGALESE FINCHES

319

roughly divided into two categories: four birds received lesions larger than 25% and three birds received lesions smaller than 10%. Results were evaluated in terms of degree of lesions; large (N ¼ 4, 29–76%), small (N ¼ 3, 6–9%), and none (N ¼ 3, control), each comprising three birds. None of the birds reduced the number of their song elements as a result of this operation. Patterns of note-to-note transition could be expressed in a first-order Markov matrix in which numbers of occurrences of transitions were tabulated. We compared the distributions of the number of transitions by 2 test, and used the resulting Cramer’s index as a measure of the degree of syntactical changes. Cramer’s index is equivalent to taking a correlation coefficient between preoperative and postoperative transition matrices. Cramer’s index was on average 0.63 for the large lesion group, 0.82 for the small lesion group, and 0.81 for the control group. Post hoc tests (Fisher’s least significant difference) detected significant differences between the large group and the other groups ( p < 0.05 for each comparison). In other words, the chunking of song notes differed before and after the surgery, and the number of chunks was decreased by partial lesion of the left HVC. An example of chunk skipping, observed in one of the large lesion birds, is shown in Fig. 8. c. NIf Since the NIf is anatomically at a higher level than the HVC, this nucleus is thought to govern the higher order organization of song syntax (Margoliash, 1997). Bilateral lesioning of the NIf was attempted in zebra finches but the effects were not detectable, except for some transient deterioration that lasted about 2 weeks after the surgery (Vu et al., 1994). We suspected that the zebra finch songs might be too simple to reveal effects of NIf lesions, because the zebra finch song is usually a repeated sequence of the identical syllable order (Zann, 1996). If the NIf is in fact governing higher order song organization, the song to be studied should have higher order syntactical organization, more complex than a mere string of fixed sequence. Each Bengalese finch song is organized into several different phrases. Each song also follows unique paths according to finite state syntax. We hypothesized that such hierarchical and complex structure of Bengalese finch songs may highlight the functioning of the NIf (Hosino and Okanoya, 2000). Thus, we attempted to make NIf lesions in 12 Bengalese finches. Because of the small volume of the NIf, all successful lesions eliminated the nucleus completely. Of 12 attempts, we achieved bilateral NIf lesion in only 3 instances. In two cases, the lesion was unilateral, while in other cases lesions were misplaced. Unilateral or misplaced lesions did not produce syntactical changes in song.

320

KAZUO OKANOYA

Fig. 8. An example of syntactical changes in song induced by partial HVC lesion. The sequence c-d-c disappeared by postoperative day 14 in this bird. Data from N. Uno and K. Okanoya (unpublished data).

SONG SYNTAX IN BENGALESE FINCHES

321

Fig. 9. An example of syntactical changes in song resulting from NIf lesion. Song syntax sung by this bird before surgery shows the complex finite state syntax of the preoperative song. But, bilateral NIf lesion eliminated most of the loops and recursions from the preoperative finite state syntax and the song became much simpler and almost linear. Data replotted from Hosino and Okanoya (2000).

Interestingly, the result of the complete lesions depended on the degree of song complexity. In two birds that were singing complex, multiphrased songs, bilateral NIf lesion eliminated phrase level complexity: the multiphrased organization of the song was reduced to a single-phrase song (Fig. 9). In one bird that was already singing a simple single-phrase song, we did not obtain any effects of bilateral NIf lesion. Thus we conclude that the NIf is responsible for phrase-to-phrase transitions (Hosino and Okanoya, 2000). 3. Lesion in the Anterior Forebrain Pathway In adult zebra finches, lesioning the nuclei of the anterior forebrain pathway, including area X (Scharff and Nottebohm, 1991) of the basal ganglia and the LMAN of the telencephalon, did not result in any noticeable effects on song production (Nordeen and Nordeen, 1993). When deafened as adults, male zebra finches did show deterioration of the song (Nordeen and Nordeen, 1992) but the effect was smaller than in Bengalese finches (Okanoya and Yamaguchi, 1997), indicating that auditory dependence is larger in Bengalese finches. We thus reasoned that lesions in the anterior pathway should produce detectable effects in Bengalese finches, which rely heavily on auditory feedback to maintain adult song.

322

KAZUO OKANOYA

Fig. 10. An example of the effect of area X lesion on song performance. The song sequence before surgery shows typical Bengalese finch song patterning. After the surgery, however, this bird stuttered at a particular song note. Data from Kobayashi et al. (2001).

When partial lesions were made in area X in Bengalese finches, a very specific type of deterioration of song syntax was observed: the number of repetitions of the repeated song notes increased significantly (Kobayashi and Okanoya, 2001; Fig. 10). We observed this symptom in all five birds lesioned, but not in sham-lesioned birds. Furthermore, song note morphology remained unchanged, as shown by the spectrotemporal crosscorrelation technique (Clark et al., 1987). The symptom observed here is reminiscent of that of Parkinson’s disease in human patients (Pastor et al., 1992), suggesting that a similar neurophysiological process underlies the control of song in birds and human motor movements. 4. Electrophysiological Studies The result of the partial HVC lesion study suggests that the HVC may be coding how to organize song notes into chunks in the motor domain. We examined whether this is true for the perceptual domain as well. If a neuron in the HVC is selective for a specific combination of song notes, self-produced song should evoke higher responses than order-modified versions or the reversal of the self-produced song. Five stimuli, each lasting 0.8–1.6 s, were used for this experiment: self-produced; bird’s own song in normal forward order (FOR); reversed FOR (REV); order-reversed FOR (spectrotemporal structure of each note retained but the order of the notes reversed; OREV); and a song recorded from another conspecific individual (CON); and the locally reversed FOR (each note reversed locally without changing global ordering of the song notes; LREV). A thick (200 m) coaxial electrode was inserted into the HVC of urethene-anesthetized

SONG SYNTAX IN BENGALESE FINCHES

323

Bengalese finches and multiunit neural responses to each of the stimuli were recorded (Nakamura and Okanoya, 1999, 2004). Multiunit responses of the HVC were highest for the FOR, followed by OREV, LREV, CON, and REV. The degree of OREV response was correlated with the complexity of the self-produced song; in a subject that had a complex song, response to OREV tended to be close to that to FOR. In a complex song, there are several possible combinations of song notes. The stimulus used was only a portion of the long, complex sequence of such complex song. Because of this, the probability of OREV sequences appearing in the normal sequence of song increased. In summary, neural response of HVC has a combination-selective property (Fig. 11). The same stimuli were used to examine response properties of the RA. From the lesion study, we hypothesized that the RA response would not differentiate between FOR and OREV because the RA is responsible for producing each discrete song note and not for combinations of song notes. The results indicated that FOR produced a larger response than OREV, but the difference was not as great as that in the HVC. Thus, combination selectivity decreases as one descends the hierarchy from the HVC to the RA. 5. Perception a. Self-generated songs In the songbird forebrain, neuronal selectivity for temporal properties of each bird’s self-generated song has been well described, but the behavioral and perceptual correlates of this selectivity are not known. By operant procedures, we trained Bengalese finches to discriminate between songs that were played normally and in reverse. Male Bengalese finches learned the discrimination quicker when their selfgenerated song was used as a stimulus than when a song of another conspecific was used. When the global note order was retained, but each note was locally reversed, the song was more likely not to be discriminated from a forward song by the singer himself, but was discriminated by other birds. These results provide psychophysical evidence that the special processing of the self-generated song observed at the neural level might reflect an individual’s perception of his self-produced song (Okanoya et al., 2000). b. Side of dominance Male Bengalese finches are left-side dominant for the motor control of song in the sensory–motor nucleus (HVC) of the telencephalon. We examined whether perceptual discrimination of songs might also be lateralized in this species (Okanoya et al., 2001). Twelve male Bengalese finches were trained by operant conditioning to discriminate between a Bengalese finch song and a zebra finch song. Before

324

KAZUO OKANOYA

Fig. 11. Multunit responses of the HVC neurons to edited versions of a bird’s own song. FOR, forward bird’s own song; REV, reversed song; LREV, each song note is locally reversed and note the morphology is changed, but the overall note ordering remains unchanged; OREV, order of song notes is reversed but the note morphology remains unchanged. Left column: Data from a bird with relatively complex, syntactical song. Right column: Data from a bird with relatively simple, linear song. HVC neurons did not respond to the OREV in the bird with simple song, but they did strongly in the bird with complex song, indicating that combination selectivity is stronger in the bird with simple linear song. Modified from Nakamura and Okanoya (2004).

SONG SYNTAX IN BENGALESE FINCHES

325

Fig. 12. Learning rate of left and right HVC-lesioned birds and of intact birds on a go–no go discrimination between a Bengalese finch song and a zebra finch song. The y axis indicates the average of hit rate and correct rejection rate. Thin lines, data from control birds; dotted lines, data from right HVC-lesioned birds; thick lines, data from left HVC-lesioned birds. Graph modified from Okanoya et al. (2001).

training, the left HVC was lesioned in four birds and the right HVC was lesioned in four other birds. The remaining four birds were used as controls with sham surgery. Birds with a left HVC lesion required significantly more time to learn to discriminate between the two songs than did birds with a right HVC lesion or intact control birds (Fig. 12). These results suggest that the left HVC is not only dominant for the motor control of song, but also for the perceptual discrimination of song. 6. Summary Lesion studies indicated that the posterior pathway of the Bengalese finch is organized in a hierarchical fashion reflecting the behavioral hierarchy (Fig. 13). At the same time, each hemisphere of the brain governs different aspects of song motor and perceptual functions. The anterior pathway, once thought to be the learning pathway, also affects real-time production of the song in the Bengalese finch. Electrophysiological studies showed that HVC neurons are tuned to respond to combinations of song notes while RA neurons are not, again showing the hierarchical nature of how song is analyzed by the brain. C. Development The previous section demonstrated that the anatomical hierarchy of the forebrain is responsible for the behavioral hierarchy of song complexity in Bengalese finches. We must now ask about the process by which this

326

KAZUO OKANOYA

Fig. 13. Schematic summary of lesion studies on the posterior pathway. The neuroanatomical hierarchy corresponded to the behavioral hierarchy of song production; NIf lesion eliminated multiphrase structure, HVC lesion eliminated some song note chunks, and RA lesion eliminated some of the song notes.

behavior emerges through development. We first review what is known of neural and behavioral development in zebra finches. We then introduce preliminary data on song development in Bengalese finches. We recorded several male Bengalese finches throughout development and then once the crystallized song stage was reached we then went back retrospectively to look at the structure of notes and of syntax in those same birds. On the basis of these, we hypothesized that song development in Bengalese finches would be a bottom-up process: that is, each song note is first established, the notes are organized into chunks, and then the chunks further organized into phrases. 1. Neural and Behavioral Development in Zebra Finches The process of neural development in Bengalese finch song system is largely unknown but, considering the similarity of the developmental process (Clayton, 1987) and brain anatomy (Okanoya, 1997), knowledge obtained in zebra finches should largely be applicable to Bengalese finches (Doupe and Solis, 1999; Fukushima et al., 2002). Thus, this section begins with an overview the anatomical and physiological development in the zebra finch song system. Male zebra finches begin to vocalize with songlike utterances at around posthatch day 25. At day 80 –90, the song pattern of the male zebra finch becomes stereotyped (Boehner, 1990; Eales, 1985). At 12 days, males and females have HVC and RA song control nuclei of comparable sizes. At 25

SONG SYNTAX IN BENGALESE FINCHES

327

days, corresponding to the emergence of song motor behavior in males, the size of the HVC and RA becomes significantly larger in males than in females. At 53 days, the size of these nuclei in males reaches the adult level (Bottjer and Arnold, 1997). Anatomical and functional connection between HVC and RA is absent on day 25, but projection from the LMAN to the RA exists. The HVC–RA projection is established at around day 50 when plastic song appears (Nordeen and Nordeen, 1988). In parallel with these anatomical developments, neurons in the anterior pathway begin to respond to species specific song on day 30, but the selectivity for the bird’s own song develops by day 60. At this stage, selectivity for the tutor song is also observed but it becomes stronger at adulthood (Doupe and Solis, 1997, 1999). Thus, neuroanatomical connections and neurophysiological responses in the song system of the zebra finch correlate well with behavioral development. 2. Developmental Processes in Bengalese Finch Song The process of song development in the Bengalese finch is slightly slower than that of zebra finches. In male Bengalese finches, primitive song begins at around day 35 after hatching. Vocalization at this stage is more or less a train of noises, referred to as ‘‘subsong.’’ Between days 50 and 100 or so, the acoustical morphology of each note stabilizes. After day 80, almost all of the notes used in the adult song are present, but the sequence, the finite state syntax, of the song is not stabilized. Nevertheless, several ‘‘chunks’’ that appear in the adult song can be identified at this stage. The song at this stage is referred to as ‘‘plastic song.’’ Full crystallization of the song is usually attained at around day 120 (Okanoya, 1997). In this fashion, phonological development precedes syntactical development in Bengalese finches (Fig. 14). 3. Development of Phonology We have examined the developmental process of song phonology by means of spectrogram correlations (Clark et al., 1987). In five birds from three families, song at the crystallized stage (i.e., older than 150 days from hatching) was used to establish a set of templates for spectrographic correlation. Each song note in the adult song was correlated with visually identified precedents of that note in the song during development. We repeated this procedure until we could no longer identify any notes in the juvenile song. For all birds and for all song notes, the developing note attained the maximum correlation with adult note morphology on day 75–85, after which the correlation value remained the same. From this analysis, we conclude that song notes reach their final stage of development at around day 80 (Okanoya, 1997).

328

KAZUO OKANOYA

Fig. 14. Song development in a male Bengalese finch shown by sonograms. Only a 3-s portion of the song at each stage is shown. Acoustic morphology of each song note is established by posthatch day 104, but syntactical organization did not stabilize until later dates.

4. Development of Syntax Development of song syntax follows a curious path. We have examined the developmental process of syntax in detail for four birds from two families. We calculated the entropy (Gentner and Hulse, 1998) of the song for each bird at several developmental stages. When the day on which minimum entropy was attained (on average, at day 100) was aligned for all four birds, all of them formed a U-shaped pattern of entropy change (Fig. 15). When the finite state syntax of the father’s song was compared with the syntax constructed from songs during development, the day when minimum entropy was attained showed maximum similarity with the

SONG SYNTAX IN BENGALESE FINCHES

329

Fig. 15. Developmental changes of song entropy in four male Bengalese finches. Song sequence entropy refers to the uncertainty of one song note followed by others. The day of minimum entropy was aligned at day 0. All four subjects follow a U-shaped pattern of entropy change.

father’s song. After that, song syntax began to deviate from the father until day 120. At that point, song syntax stabilized and no further change was observed (K. Okanoya, unpublished observations). 5. Summary Behavioral development of Bengalese finches follows a pattern similar to that of zebra finches, although the general process occurs more slowly in Bengalese finches. Bengalese finches begin subsong-like vocalization at day 35. The morphology of each song note is established at around day 80, but song syntax is unstable until day 120. Development of anatomical structures and connections, and that of the physiological properties of the song control nuclei, in Bengalese finches have not yet been reported, but the basic developmental path is suspected to be similar to that of zebra finches.

D. Function What are the functions, then, of song complexity in Bengalese finches? Although Bengalese finches are domesticated, we postulated that it evolved partially as a result of sexual selection by females (Andersson, 1994; Catchpole, 2000; Okanoya, 2002). Since there are no records in Japanese avicultural literature indicating that songs had been artificially selected by

330

KAZUO OKANOYA

breeders (Washio, 1996), we assume that breeders only selected successful pairs and that this resulted indirectly in selecting good singers. Therefore, we assumed that song perception must differ between males and females. This question was examined by heart rate measurement. We also assumed that song complexity should efficiently stimulate sexual behavior of females. We tested this prediction by several assays that supplemented information from each other (Searcy, 1992; Searcy and Yasukawa, 1996). First, we tested the receptive behavior of females using the copulation solicitation assay. Second, we measured reinforcing properties of complex song with a perch-hopping operant technique. Third, degrees of nest-building behavior by female Bengalese finches were measured as a function of stimulus songs. Finally, we measured serum estradiol level in females stimulated with complex or simple songs. 1. Heart Rate Bird song may be assessed and processed differently by the sexes because the production and functional use of this behavior are often sexually dimorphic. However, straightforward examination of this hypothesis has been difficult, since different behavioral measures have been used to describe the process of song assessment in the two sexes. We analyzed changes in heart rate as an index of song assessment in Bengalese finches (Ikebuchi et al., 2003). In this species, only males sing, and song is exclusively used for mate attraction. This species is not territorial and songs are not used in aggressive contexts. When a song stimulus was presented for the first time, heart rate increased in Bengalese finch subjects. Duration of heart rate increase was defined as the duration during which the heart rate increased to two standard deviations above the baseline interval 10 s before the song presentation. In both sexes, repeated presentation of one song resulted in a waning of the heart rate response. Presentation of heterospecific (zebra finch) songs did not increase the heart rate. When a novel conspecific song was presented, the heart rate increased only in female birds with each presentation of the stimulus, but not in males (Fig. 16). These results correspond to the sex differences in response to song in this species; males ignore other birds’ songs while females are attentive. This result is not due to sex differences in memory capacity: operant conditioning studies showed that males and females are not different in their memory capacity for song (Ikebuchi and Okanoya, 2000). Results on females suggest that syntactically complex songs may be more potent than simple songs in keeping female finches in an aroused state. Whether or not this can lead to reproductive preparedness in female’s side, and/or can lead to more precious assessment of male quality by females, should be addressed in other experimental settings.

SONG SYNTAX IN BENGALESE FINCHES

331

Fig. 16. Duration of heart rate increase in seconds was compared for males and females for each class of song stimulus. A sex difference appeared only for singly presented Bengalese finch songs (Beng-S) while conspecific song presented multiple times (Beng-M) or songs of other species, whether they were singly (Zeb-S) or multiply (Zeb-M) presented, did not increase the heart rate and the sexes did not differ. Data from Ikebuchi et al. (2003).

2. Reinforcing Properties of Song Complexity Next, we used an operant conditioning technique with song as a reinforcer to examine the preference of female Bengalese finches for song complexity (Morisaka and Okanoya, 2002). The procedure and apparatus used by Gentner and Hulse (2000) to test female starlings’ song preferences were modified for Bengalese finches. We prepared a large metal cage and put pot-shaped nests in two upper corners. Inside the pot-shaped nests, we put small speakers for song playback. We also placed a perch in front of each of the nest pots. A natural song sung by a male Bengalese finch was used to prepare a simple song (order of song notes fixed) and a complex song (the order varied according to a finite state rule) which were played back from the relevant speaker when the bird sat on the perch. A female bird was placed inside this apparatus. Four of the eight birds tested chose the complex song, one the simple song, and the remaining three chose both songs randomly. These results suggest that song preferences of female Bengalese finches vary individually, but somewhat more birds did prefer complex over simple songs. Since this experiment utilized only one song type, results should be interpreted only with caution. It is

332

KAZUO OKANOYA

nevertheless suggestive that these female preferences could lead male Bengalese finches to evolve complex songs by sexual selection (Morisaka and Okanoya, 2002). 3. Copulation Solicitation Assay Copulation solicitation assays, first demonstrated by King and West (1977; see also West et al., 1979) with brown-headed cowbirds (Molothrus ater), were used by Clayton and Proeve (1989) to examine whether females prefer songs with larger numbers of elements in Bengalese finches. Linearly arranged songs with four, six, or eight different notes were prepared and the number of copulation solicitation displays emitted when each of the songs was played to female Bengalese finches was counted. The results indicated that song with more notes produced a larger number of copulatory responses. However, Clayton and Proeve did not examine the effect of changing song syntax. We have also utilized the copulation solicitation assay to examine females’ responsiveness to the playback songs (Okanoya et al., 1998). In our preliminary research on two female Bengalese finches, forward song of a male Bengalese finch produced several copulatory responses while reversed song (REV) did not produce any responses at all. However, we found that using the copulation assay in Bengalese finches was generally difficult. Of eight females tested, only two responded. Hormonal treatment did not boost the females’ response as it does in some species (West et al., 1979). 4. Nest-Building Behavior As yet another demonstration of function, we examined the nest-building behavior of females (Eisner, 1961, 1963) when stimulated with complex or simple syntax songs (Okanoya and Takashima, 1997). This procedure was first used by Hinde and Steel (1976) and Kroodsma (1976). Hinde and Steel showed that female domesticated canaries engaged in more nest material carrying when stimulated with conspecific songs rather than songs of other species. Kroodsma found female canaries performed more nest building and laid more eggs when stimulated with large than with small repertoire songs. Song recordings obtained from a male Bengalese finch were analyzed, and four distinctive song phrases were identified. In this bird’s song, these four phrases were organized so that phrases A or B were repeated several times, and phrases C or D followed this repetition, but these phrases were never repeated. After phrase C or D was sung once, phrase A or B was again repeated. We wrote software that produced this sequence of song phrases (‘‘complex syntax song’’), or that repeated only phrase B (‘‘simple

SONG SYNTAX IN BENGALESE FINCHES

333

syntax song’’). Phrase B included most of the song notes included in phrases A, C, and D. Three groups of female Bengalese finches were studied. Each group consisted of four finches, separately caged, and kept together in a sound isolation box. The first group was stimulated with the complex syntax song, the second group was stimulated with the simple syntax song, and the third group was not stimulated by any song. The number of items of nesting material carried each day when stimulated with complex or simple syntax songs was counted and compared. Female finches carried more nesting material (Fig. 17). We examined further whether randomly generated note sequences were more effective than syntactically synthesized ones. Females stimulated with random note sequences were less responsive and were comparable to the females stimulated with the simple sequence. Although random sequences resulted in complex ordering of the song notes, randomness was not the same as the complexity produced by syntax for the females (Okanoya et al., 2004). 5. Estradiol Level Three groups of female Bengalese finches were used. Each group consisted of four finches, separately caged, and kept together in a sound isolation box. The first group was stimulated with the complex syntax song, the second group, was stimulated with the simple syntax song, and the third

Fig. 17. Results of the nest-building assay. When stimulated with the complex song, female Bengalese finches carried significantly more nest material than did the birds stimulated with simple song or with no song. Each line indicates the median of four birds. Data are smoothed by taking running averages for each 4 days.

334

KAZUO OKANOYA

group received no playback songs. The levels of serum estradiol were compared among groups both prior to and after the experiment so that any baseline level differences could be taken into account (Okanoya et al., 2004). Serum estradiol levels were on average 0.37 ng/mg before the experiment and 0.76 ng/mg after the experiment in the females stimulated with the complex song, 0.55 ng/mg before and 0.67 ng/mg after the experiment in the females stimulated with the simple song, and 0.46 and 0.52 ng/mg, respectively, with a blank tape. Thus, the complex song was more effective in stimulating female Bengalese Finches into reproductive condition (t ¼ 2.858, p < 0.05 by post hoc tests after two-way ANOVA comparing stimulus condition and experimental periods). 6. Summary Song functions only in the heterosexual context in Bengalese finches. Reflecting this, individual differences in song modified heart rate in females, but not in males. Temporal features of song were important for female perception, as shown by the copulation solicitation assay. Females had a tendency to approach complex songs rather than simple ones and nesting behavior of the females was stimulated more efficiently with the complex song. Females stimulated with the complex song increased serum estradiol levels significantly compared with the basal level, but not in those stimulated with the simple song. These results all support the notion that song complexity functions to prepare, attract, and stimulate the reproductive behavior of female Bengalese finches. They also support the hypothesis that song complexity in Bengalese finches evolved through intersexual selection.

V. Further Questions We have examined song complexity of Bengalese finches from the four perspectives proposed by Tinbergen. In this section, we deal with more integrated further questions that could not be asked or answered from a single perspective. A. Directed and Undirected Songs: Possible Mechanisms Directed songs in zebra finches are obviously sung to stimulate females, but the function of undirected song in zebra finches has been an enigma (Caryl, 1981; Sossinka and Boehner, 1980). However, a key finding to understand the function of undirected song came, unexpectedly, from the

SONG SYNTAX IN BENGALESE FINCHES

335

study of gene expression (Jarvis et al., 1998). When singing directed songs, the nuclei of the posterior pathway (HVC and RA) of zebra finches were found to express the immediate-early gene, zenk. When singing undirected songs, however, in addition to the posterior path, the anterior path was also active and, especially, area X of the basal ganglia expressed a high level of the zenk gene. A similar pattern of gene expression was reported also in Bengalese finches by Yakura and Takeuchi (1998). From these findings, some speculation arose as to the function of undirected songs. When a behavior requires real-time modification through feedback information, then that behavior will involve the basal ganglia– cortex pathway. Likewise, these undirected songs are probably sung to maintain and fine-tune motor patterns and timing necessary for directed songs, and this may be why area X was highly active when singing undirected songs. This speculation remains a hypothesis, but we tested it indirectly by a behavioral experiment. Bengalese finches also sing directed and undirected songs and, as with zebra finches, the acoustic structure of the two song types is the same (Clayton, 1987; Okanoya, 1997). Since the effect of deafening is stronger and more acute in Bengalese finches than in zebra finches, we postulated that real-time monitoring of a bird’s own song is more important when singing undirected songs than directed songs. This hypothesis was examined by introducing background noise in the bird’s environment and measuring the sound pressure level of the song. We reasoned that amplitude compensation should be stronger when the bird was carefully listening to his own song, thus undirected song would produce stronger compensation than directed songs. Directed and undirected songs were recorded from six adult male Bengalese finches and the level of background white noise was randomly changed from 40 to 70 dB in 10-dB steps. When singing undirected songs, all six subjects sang louder in response to noise and the slope of the regression line deviated significantly from 0. When singing directed songs toward females, however, five of six males did not change their song amplitude under noisy conditions. One of the males increased his song amplitude in relation to the background noise, but the slope in this condition was smaller than when the same subject was singing undirected songs in noisy conditions. A paired t test comparing directed and undirected conditions for all subjects indicated that the slope for undirected singing was steeper than that of directed singing (t ¼ 2.64, p < 0.05) (Kobayashi and Okanoya, 2003a). Male Bengalese finches thus adjusted their song amplitude when singing solo song, but not when they were singing female-directed songs. These data on context-dependent song amplitude compensation agree well with the hypothesis that the bird is listening to his own song more attentively

336

KAZUO OKANOYA

when singing solo, undirected song than when singing female-directed song. Of course, a more functional explanation based on consideration of the communication context could also explain this phenomenon. Males do not need to sing louder if the listener is nearby, as is the case in directed song, while it is necessary to sing louder if undirected song is to attract distant females. At any rate, a higher degree of gene expression in the anterior pathway (HVC–area X–DLM–LMAN) and more attentive listening to self-produced song both occurred when singing undirected songs in Bengalese finches. Interestingly, such amplitude compensation also occurred in distance calls in male Bengalese finches, but not in female Bengalese finches (Kobayashi and Okanoya, 2003b). Functional usage of the calls may be different between the sexes, but it may also be related to sex differences in the brain circuitry involved in the production of distance calls in this species (Fukushima and Aoki, 2000; Fukushima et al., 2002). B. Preference for Song Complexity in Wild Strains We hypothesized that sexual selection acted in the evolution of finite state syntax in Bengalese finches. That is, females of the ancestral species preferred complex songs, but the ability of males of the ancestor species to develop and sing complex songs was limited by constraints in nature. Such costs might include predation risk and foraging costs, although whether these can really be costs should be examined in nature. When munias were domesticated, these costs disappeared and female preference augmented by human intervention operated to cause the evolution of song complexity. We directly tested this hypothesis by examining reproductive behavior in the ancestral species. A simple, linear song of a white-backed munia male was divided into individual song notes, and the finite state syntax of a Bengalese finch that used the same number of song notes was applied to the song notes of the white-backed munia. Thus, we composed a new song that had the phonology of the white-backed munia, and the syntax of the Bengalese finch. We played either the original simple munia song, or the ‘‘hybrid’’ song to females. Two groups of four white-backed munias were used in this experiment using the nesting material assay of song potency. The first group heard the original white-backed munia songs for 20 days and then the stimulus was switched to the hybrid song. The second phase lasted for 14 days. The second group was stimulated with Bengalese finch song for 20 days and then stimulated with the original song. The maximum number of pieces of nesting material carried by the two groups in the first 20 days

SONG SYNTAX IN BENGALESE FINCHES

337

did not differ. However, during the next 14 days the group of birds stimulated with the hybrid song carried significantly more nesting material than the second group stimulated by the original song (Okanoya et al., 2004). Males of wild white-backed munias sing syntactically simple, linearly arranged songs (Honda and Okanoya, 1999). This experiment showed, however, that females of these birds responded more to the hybrid song than to the original song. Female white-backed munias did not respond more to the Bengalese finch song, although it had the same degree of syntactical complexity because the hybrid song was modeled on the Bengalese finch song. This may be due to phonological differences between the two strains of songs (Honda and Okanoya, 1999). Most white-backed munia song notes are broad-banded, noisy utterances. On the other hand, most Bengalese finch song notes are harmonically organized sounds. Since our wild-caught female white-backed munias were undoubtedly exposed to wild white-backed munia songs when they were young, they should have been familiar with munia phonology. That probably made the phonology of the Bengalese finch song very alien to them. C. Cognitive Cost of Song Complexity We showed that a song with complex syntax effectively stimulated the reproductive system of females. From this, we postulate that complex song patterning should be more attractive to female birds, and, therefore, the song syntax in Bengalese finches may have evolved through the process of sexual selection (Darwin, 1871). The ability to sing a complex song may be an honest signal that advertises the potency of the singer (Gil and Gahr, 2002; Nowicki et al., 2000; Zahavi and Zahavi, 1997), since singing a complex song may require (1) higher testosterone levels, (2) a greater cognitive load, and (3) more brain space. We tested one of the above hypotheses by disrupting an ongoing song by shining a flashlight at the bird. Birds that were singing more complex songs were less prone to stop singing (Nakamura and Okanoya, 2000). One interpretation of this result is that singing a complex song requires a greater cognitive load and makes the animal less careful about potentially dangerous situations. From this could follow that, although a complex song may be disadvantageous for the bird’s survival, the singer is preferred by females because being able to withstand the disadvantage of complexity reflects the potency of the singer.

338

KAZUO OKANOYA

VI. Remaining Questions A. Learnability Given the differences between the songs of white-backed munias and Bengalese finches, we have implicitly postulated that there should be some genetic differences that control learnability of songs in these two strains. Of course, the sexual selection argument we have put forward also depends on this postulate. However, the differences in the songs could also be attributed solely to cultural transmission. Nevertheless, purely learned behavioral differences could also be subject to sexual selection. One of the procedures allowing one to examine whether this behavior is partially restricted by genetic constraints is to cross-foster between whitebacked munias and Bengalese finches. We are now conducting such experiments and obtaining some interesting data. We have found some Bengalese song notes and Bengalese type note-to-note transition patterns that were never learned by white-backed munias (M. Takahashi and K. Okanoya, unpublished data). More replications are necessary to draw conclusions, but the comparisons between these two strains would be fruitful in examining the degree in which behavioral plasticity is governed by genetic constraints. B. Auditory Dependence of Song Control in White-Backed Munias A very important question is whether the white-backed munia, the ancestral species of the Bengalese finch, would also require real-time auditory feedback to maintain intact song. If they do, then we can infer that song complexity in Bengalese finches might be the consequence of auditory dependence. If they do not, on the other hand, then auditory dependence presumably arose as song complexity evolved by factors involving sexual selection and domestication. The reason we have been unable to pursue this question up to now was that the availability of wild white-backed munias was limited, but we have located a source in Taiwan from which we can obtain wild birds. C. Where Would the Change Be? One of the most conspicuous features of the Bengalese finch song is its syntactical complexity: chunks of song notes are organized into finite state syntax. Interestingly, this feature disappeared after lesions of one of the telencephalic vocal control nuclei, the NIf. The NIf-lesioned Bengalese finches sang simple, linear songs (Hosino and Okanoya, 2000) that were syntactically similar to the song of white-backed munia. We therefore suspect that the function of NIf changed during the

SONG SYNTAX IN BENGALESE FINCHES

339

process of domestication so that more song complexity could be added to note-to-note transitions. Neuroanatomical, electrophysiological, and molecular genetic comparisons between the white-backed munia and Bengalese finch NIf are necessary to elucidate the proximate cause of the song complexity in the Bengalese finch. D. How Do Females Appreciate Song Complexity? Syntactically organized hierarchical songs of male Bengalese finches are produced by hierarchically organized forebrain nuclei. Females can perceive and discriminate between syntactical songs and random sequences of song notes as shown by the nest-building experiment. Thus, females should have brain mechanisms that can discern syntactical organization of song notes. In fact, female preference for variability is probably the major driving force that has shaped finite state syntax in Bengalese finch songs. Female Bengalese finches have completely different brain organization from males (Okanoya, 1997). Most of the forebrain song control nuclei including the NIf, HVC, RA, and area X are absent in females. How then is the hierarchical nature of the song structure appreciated by the female auditory system? By phonotaxis (Miller, 1979a,b) and by operant conditioning (Riebel, 2000), it has been shown that female zebra fiches prefer their father’s song. The neural basis for this memory is not known, but we found that auditory neurons in the nidopallium of female Bengalese finches became selective for particular songs if they were kept with a particular singing mate or trained by operant conditioning (Nakamura and Okanoya, 2001). To further understand this issue, more knowledge of the neuroanatomy of the female song auditory system is required.

VII. Scenario for the Evolution of Complex Syntax On the basis of the experimental results reviewed in this chapter, we can postulate several steps to explain the evolution of complex song syntax in Bengalese finches. Songs of most estrildid finches are used solely for mating purposes and not in male–male interactions. Sexual selection is thus likely to have inflated the properties of songs on which females based their choice, giving rise to traits that are handicaps in the wild environment. The following is one of the possible ‘‘scenarios’’ that can explain the emergence of finite-state syntax in Bengalese finches. In Bengalese finches, complexity in song note transitions became the sexually selected trait. This trait was subject to individual variations due to

340

KAZUO OKANOYA

genetic differences in neural capabilities and cultural differences in song traditions. However, the wild environment restricted the degree of possible song complexity because of several costs associated with the maintenance of such traits. These may include predation cost, foraging time, immunological cost of testosterone, and metabolic cost to maintain brain space for the song system. Thus, mutations leading to more song complexity would not have become fixed in the population of wild white-backed munias. Once domesticated, though, many of these potential costs, especially those of predation and foraging time, were eliminated. Thus domestication relaxes restrictions imposed on the evolution of the trait (Okanoya, 2002). Mutations leading to song complexity, probably involving functioning of the NIf, were thus not fixed in the natural environment but were in the domesticated environment. Changes in brain structure now enabled more elaborated song to be learnt and improvisation of song syntax to arise. Genes to allow learning of complex songs were selected because of the preferences of females. Much work needs to be done to examine the validity of this scenario. However, it is clear from our results so far that Bengalese finch song syntax will reward further study from all the four perspectives that Niko Tinbergen originally outlined.

VIII. Summary Bengalese finch song has a unique type of complexity. We analyzed the syntactical rules underlying it by statistical methods and found that the songs of these birds could be expressed as finite state syntax. In general, two to five song notes are arranged in fixed order to form ‘‘chunks’’ and several chunks are arranged with probabilistic rules into variable song phrases. Our aim was to understand the song complexity of Bengalese finches in terms of the ‘four questions’ raised by Tinbergen and thus to show the strength of his integrative approach in the study of behavior. Through song development, we found that the phonological characteristics appeared first and syntactical characteristics were then established. In other words, song development follows the hierarchical path from song phonology to song syntax. To understand the evolution of the song complexity, we examined songs of the wild strain of the Bengalese finch, the white-backed munia. Their songs are linear strings of notes in a highly stereotyped order, neither hierarchical nor complex like those of Bengalese finches. Thus, song syntax must have been acquired at some point during the process of domestication. We then investigated the brain mechanisms by which this complex syntax was enabled. A series of lesion studies

SONG SYNTAX IN BENGALESE FINCHES

341

demonstrated that the behavioral hierarchy of the song is expressed in the anatomical hierarchy of the forebrain song control system. Auditory neurons in these hierarchical brain structures respond to the particular aspects of song structures that reflect the behavioral hierarchy. Finally, in a series of studies examining female preferences for song complexity, we found that complex songs would attract not only Bengalese finch females, but also white-backed munia females. On the basis of these data, we suggest that song complexity in Bengalese finches may have been initiated as a sexually selected trait in the wild, and then enhanced in a domesticated environment that relaxed several selection pressures in the wild. Further research is necessary to examine the hypotheses that have arisen from these studies so far. Acknowledgments I thank P. J. B. Slater, C. T. Snowdon, M. Naguib, and J. Rosenblatt for suggestions, encouragement, and patience with earlier revisions of the manuscript. Most of the experimental results reported here stem from collaboration with my students and fellow researchers at Chiba University: M. Ikebuchi, K. Nakamura, N. Hirata, H. Uno, E. Honda, T. Hosino, K. Kobayashi, S. Tsumaki, A. Takashima, M. Futamatsu, M. Takahashi, H. Yamada, and T. Kawamura, among others. Illustrations were prepared mostly by H. Yamada and T. Kawamura. The research environment is generously supported by faculty members of the Department of Cognitive and Information Science (Faculty of Letters, Chiba University) Funding was supplied for this research from PRESTO, Japan Science and Technology Agency with supervision and encouragement by R. Suzuki and Y. Sawada. References Airey, D. C., and DeVoogd, T. J. (2000). Greater song complexity is associated with augmented song system anatomy in zebra finches. Neuroreport 11, 2339–2344. Alcock, J. (2001). ‘‘Animal Behavior.’’ Sinauer Associates, Sunderland, MA. Andersson, M. (1994). ‘‘Sexual Selection.’’ Princeton University Press, Princeton, NJ. Boehner, J. (1990). Early acquisition of song in the zebra finch. Anim. Behav. 39, 369–374. Bottjer, S. W., and Arnold, A. P. (1997). Developmental plasticity in neural circuits for a learned behavior. Annu. Rev. Neurosci. 20, 459–481. Brainard, M. S., and Doupe, A. J. (2000a). Interruption of a basal ganglia–forebrain circuit prevents plasticity of learned vocalizations. Nature 404, 762–766. Brainard, M. S., and Doupe, A. J. (2000b). Auditory feedback in learning and maintenance of vocal behaviour. Nat. Rev. 1, 31–40. Caryl, P. G. (1981). The relationship between the motivation of directed and undirected song in the zebra finch. Z. Tierpsychol. 57, 37–50. Catchpole, C. K. (2000). Sexual selection and the evolution of song and brain structure in the Acrocephalus warblers. Adv. Study Behav. 29, 45–97. Catchpole, C. K., and Slater, P. J. B. (1995). ‘‘Bird Song: Biological Themes and Variations.’’ Cambridge University Press, Cambridge. Chatfield, C., and Lemon, R. E. (1970). Analysing sequences of behavioral events. J. Theor. Biol. 29, 427–445.

342

KAZUO OKANOYA

Clark, C., Marler, P., and Beeman, K. (1987). Quantitative analysis of animal vocal phonology: An application to swamp sparrow song. Ethology 76, 101–115. Clayton, N. S. (1987). Song learning in Bengalese finches: A comparison with zebra finches. Ethology 76, 247–255. Clayton, N. S., and Prove, E. (1989). Song discrimination in female zebra finches and Bengalese finches. Anim. Behav. 38, 352–362. Darwin, C. (1871). ‘‘The Descent of Man, and Selection in Relation to Sex,’’ 2nd ed. revised and augmented. Appleton, New York. Doupe, A. J., and Solis, M. M. (1997). Song- and order-selective neurons develop in the songbird anterior forebrain during vocal learning. J. Neurobiol. 33, 694–709. Doupe, A. L., and Solis, M. M. (1999). Song- and order-selective auditory responses emerge in neurons of the songbird anterior forebrain during vocal learning. In ‘‘The Design of Animal Communication’’ (M. D. Hauser and M. Konishi, Eds.), pp. 343–368. MIT Press, Cambridge, MA. Eales, L. A. (1985). Song learning in zebra finches: Some effects of song model availability on what is learnt and when. Anim. Behav. 33, 1293–1300. Eens, M. (1997). Understanding the complex song of European starling: An integrated ethological approach. Adv. Study Behav. 26, 355–434. Eisner, E. (1957). The Bengalese finch. Avicult. Mag. 63, 101–108. Eisner, E. (1960). The biology of the Bengalese finch. Auk. 77, 271–287. Eisner, E. (1961). The behaviour of the Bengalese finch in the nest. Ardea. 49, 51–69. Eisner, E. (1963). A quantitative study of parental behaviour in the Bengalese finch. Behaviour 20, 134–206. Ewert, J.-P. (1976). ‘‘Neuroethology.’’ Springer-Verlag, Berlin. Freeberg, T. M., West, M. J., King, A. P., Duncan, S. D., and Segelaub, D. R. (2002). Cultures, genes, and neurons in the development of song and singing in brown-headed cowbirds (Molothrus ater). J. Comp. Physiol. A. Neuroethol. Sens Newal Behav. Physiol. 188, 993–1002. Fukushima, Y., and Aoki, K. (2000). The role of the dorsomedial nucleus (DM) of intercollicular complex with regard to sexual difference of distance calls in Bengalese finches. Zool. Sci. 17, 1231–1238. Fukushima, Y., Takagi, H., and Aoki, K. (2002). Developmental differences of neural function in the song and call control system in Bengalese finches. Neurosci. Lett. 334, 140–144. Gentner, T. Q., and Hulse, S. H. (1998). Perceptual mechanisms for individual vocal recognition in European starlings, Sturnus vulgaris. Anim. Behav. 56, 579–594. Gentner, T. Q., and Hulse, S. H. (2000). Female European starling preference and choice for variation in conspecific male song. Anim. Behav. 59, 443–458. Gil, D., and Gahr, M. (2002). The honesty of bird song: Multiple constraints for multiple traits. Trends Ecol. Evol. 17, 133–141. Gil, D., and Slater, P. J. B. (2000). Song organization and singing patterns of the willow warbler, Phylloscopus trochilus. Behaviour 137, 759–782. Guettinger, H. R., and Nicolai, J. (1973). Struktur und funktion der Rufe bei Prachtfinken (Estrildidae). Z. Tierpsychol. 33, 319–334. Hamilton, K. S., King, A. P., Sengelaub, D. R., and West, M. J. (1997). A Brain of her own: A neural correlate of song assessment in a female songbird. Neurobiol. Learn. Mem. 68, 325–332. Hernalsteen, N., and Iwens, P. (1995). ‘‘Japanese meeuwen.’’ Uitgeverij Imprim ‘Tout, Moeskroen.

SONG SYNTAX IN BENGALESE FINCHES

343

Hinde, R. A., and Steel, E. (1976). The effect of male song on an oestrogen-dependent behaviour pattern in the female canary (Serinus canariensis). Horm. Behav. 7, 293–304. Honda, E., and Okanoya, K. (1999). Acoustical and syntactical comparisons between songs of the white-backed munia (Lonchura striata) and its domesticated strain, the Bengalese finch (Lonchura striata var. domestica). Zool. Sci. 16, 319–326. Hosino, T., and Okanoya, K. (2000). Lesion of a higher-order song nucleus disrupts phrase level complexity in Bengalese finches. Neuroreport 11, 2091–2095. Hultsch, H., and Todt, D. (1989). Memorization and reproduction of song in nightingales (Luscinia megarhynchos): Evidence for package formation. J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol. 165, 197–203. Ikebuchi, M., and Okanoya, K. (2000). Limited auditory memory for conspecific songs in a non-territorial songbird. Neuroreport 11, 3915–3939. Ikebuchi, M., Futamatsu, M., and Okanoya, K. (2003). Sex differences in song perception by Bengalese finches as measured by cardiac responses. Anim. Behav. 65, 123–130. Jarvis, E. D., Scharff, C., Grossman, M. R., Ramos, J. A., and Nottebohm, F. (1998). For whom the bird sings: Context-dependent gene expression. Neuron 21, 775–788. Jones, A. E., Cate, C. T., and Slater, P. J. B. (1996). Early experience and plasticity of song in adult male zebra finches. J. Comp. Psychol. 110, 354–369. King, A. P., and West, M. J. (1977). Species identification in the North American cowbird: Appropriate response to abnormal song. Science 195, 1002–1004. Kobayashi, K. I., and Okanoya, K. (2003a). Context-dependentsong amplitude control in Bengalese finches. Neuroreport 14, 521–524. Kobayashi, K. I., and Okanoya, K. (2003b). Sex differences in amplitude regulation of distance calls in Bengalese finches. Anim. Biol. 53, 173–182. Kobayashi, K., Uno, H., and Okanoya, K. (2001). Partial lesions in the anterior forebrain pathway affect song production in adult Bengalese finches. Neuroreport 12, 353–358. Konishi, M. (1964). Effects of deafening on song development in two species of juncos. Condor 66, 85–102. Konishi, M. (1965). The role of auditory feedback in the control of vocalization in the whitecrowned sparrow. Z. Tierpsychol. 22, 770–783. Krebs, J. R., and Davies, N. B. (1993). ‘‘Introduction to Behavioural Ecology.’’ Blackwell Scientific, Oxford. Kroodsma, D. E. (1976). Reproductive development in a female songbird: Differential stimulation by quality of male song. Science 192, 574–575. Lessells, C. M., and Mateman, A. C. (1998). Sexing birds using random amplified polymorphic DNA (RAPD) markers. Mol. Ecol. 7, 187–195. Manning, A., and Dawkins, M. S. (1998). ‘‘An Introduction to Animal Behaviour.’’ Cambridge University Press, Cambridge. Manning, C. D., and Schutze, H. (1999). ‘‘Foundations of Statistical Natural Language Processing.’’ MIT Press, Cambridge, MA. Margoliash, D. (1997). Functional organization of forebrain pathways for song production and perception. J. Neurobiol. 33, 671–693. McFarland, D. (1999). ‘‘Animal Behaviour: Psychobiology, Ethology and Evolution.’’ Longman, Harlow, New York. Mello, C. V., Vicario, D. S., and Clayton, D. F. (1992). Song presentation induces gene expression in the songbird forebrain. Proc. Natl. Acad. Sci. USA 89, 6818–6877. Miller, D. B. (1979a). Long-term recognition of father’s song by female zebra finches. Nature 280, 389–390.

344

KAZUO OKANOYA

Miller, D. B. (1979b). The acoustic basis of mate recognition by female zebra finches (Taeniopygia guttata). Anim. Behav. 27, 376–380. Mizuta, T., Yamada, H., Lin, R. S., Yodogawa, Y., and Oknaoya, K. (2003). Sexing whiterumped munias in Taiwan, using morphology, DNA and distance calls. Ornithol. Sci. 2, 97–102. Møller, A. P., Henry, P. Y., and Erritzoe, J. (2000). The evolution of song repertoire and immune defence in birds. Proc. R. Soc. Lond. B. Biol. Sci. 267, 165–169. Morisaka, T., and Okanoya, K. (2002). Do female Bengalese finches prefer complex conspecific songs? Ethological approach Trans. Tech. Comm. Psychol. Physiol. Acoust. 32, 253–258. Nakamura, K., and Okanoya, K. (1999). Song linearity reflects song selectivity of HVc in Bengalese finch (Lonchura striata var domestica). Trans. Tech. Comm. Psychol. Physiol. Acoust. 29, 1–6. Nakamura, K., and Okanoya, K. (2000). Measuring the cognitive cost of voco-auditory behavior: A case of Bengalese finch song. Trans. Tech. Comm. Psychol. Physiol. Acoust. 30, 1–6. Nakamura, K., and Okanoya, K. (2001). Auditory response to male songs in the auditory center of the female songbirds: The relationship with the auditory experience. Trans. Tech. Comm. Psychol. Physiol. Acoust. 31, 603–609. Nakamura, K. Z., and Okanoya, K. (2004). Neural combination selectivity of HVC correlates with song complexity in Bengalese finches. Neuroreport 15, 1359–1363. Nordeen, K. W., and Nordeen, E. J. (1988). Projection neurons within a vocal motor pathway are born during song learning in zebra finches. Nature 334, 149–151. Nordeen, K. W., and Nordeen, E. J. (1992). Auditory feedback is necessary for the maintenance of sterotyped song in adult zebra finches. Behav. Neural Biol. 57, 58–66. Nordeen, K. W., and Nordeen, E. J. (1993). Long-term maintenance of song in adult zebra finch is not affected by lesions in the forebrain region involved in song learning. Behav. Neural Biol. 58, 79–82. Nottebohm, F., Nottebohm, M., and Crane, L. (1986). Developmental and seasonal changes in canary song and their relation to changes in the anatomy of song-control nuclei. Behav. Neural Biol. 46, 445–471. Nowicki, S. (1987). Vocal tract resonances in oscine bird sound production: Evidence from birdsong in a helium atmosphere. Nature 325, 53–55. Nowicki, S., Hasselquist, D., Bensch, S., and Peters, S. (2000). Nestling growth and song repertoire size in great reed warblers: Evidence for song learning as an indicator mechanism in mate choice. Proc. R. Soc. Lond. B. Biol. Sci. 267, 2419–2424. Nowicki, S., Searcy, W. A., and Peters, S. (2002). Brain development, song learning, and mate choice in birds: A review and experimental test of the ‘‘nutritional stress hypothesis.’’ J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol. 188, 1003–1014. Okanoya, K. (1997). Voco-auditory behavior in the Bengalese finch: A comparison with the zebra finch. Biomed. Res. 18, 53–70. Okanoya, K. (2002). Sexual selection as a syntactical vehicle: The evolution of syntax in birdsong and human language through sexual selection. In ‘‘Transitions to Language’’ (A. Wray, Ed.), pp. 46–63. Oxford University Press, Oxford. Okanoya, K. (2004). Hypoglossal lateralization for song production in the Bengalese finch. Zool. Sci. (in press). Okanoya, K., and Kimura, T. (1993). Acoustical and perceptual structures of sexually dimorphic distance calls in Bengalese finches (Lonchura striata domestica). J. Comp. Psychol. 107, 386–394.

SONG SYNTAX IN BENGALESE FINCHES

345

Okanoya, K., and Takashima, A. (1997). Auditory preference of the female as a factor directing the evolution of Bengalese finch songs. Trans. Tech. Comm. Psychol. Physiol. Acoust. 27, 1–6. Okanoya, K., and Yamaguchi, A. (1997). Adult Bengalese finches require real-time auditory feedback to produce normal song syntax. J. Neurobiol. 33, 343–356. Okanoya, K., Yoneda, T., and Iseki, S. (1995). Distance calls of the wild white-backed munia in Kijoka, Okinawa. Jpn. J. Ornithol. 44, 231–233. Okanoya, K., Honda, E., and Nishikawa, N. (1998). Time-series analyses of species-specific songs in birds. Trans. Tech. Comm. Psychol. Physiol. Acoust. 28, 1–6. Okanoya, K., Tsumaki, S., and Honda, E. (2000). Perception of temporal structures in selfgenerated songs by Bengalese finches (Lonchura striata var domestica). J. Comp. Psychol. 114, 239–245. Okanoya, K., Ikebuchi, M., Uno, H., and Watanabe, S. (2001). Left-side dominance for song discrimination in Bengalese finches. Anim. Cogn. 4, 241–245. Okanoya, K., Takashima, A., Okuhara, A., Ikebuchi, M., Otsuka, R., and Wada, M. (2004). Syntactical complexity of male songs stimulates reproductive development in female Bengalese finches. Zool. Sci. (in press). Okuhata, S., and Saito, N. (1987). Synaptic connections of thalamo-cerebral vocal nuclei of the canary. Brain Res. Bull. 18, 35–44. Pastor, M. A., Artieda, J., Jahanshahi, M., and Obeso, J. A. (1992). Time estimation and production is abnormal in Parkinson’s disease. Brain 115, 211–225. Petrinovich, L., Yoder, R. M., and Dobson, C. W. (1972). The significance of white-crowned sparrow song. Behav. Biol. 7, 795–803. Podos, J., Peters, S., Rudnicky, T., Marler, P., and Nowicki, S. (1992). The organization of song repetoires in song sparrows: Themes and variations. Ethology 90, 89–106. Price, P. (1979). Developmental determinants of structure in zebra finch song. J. Comp. Physiol. Psychol. 93, 260–277. Reiner, A., et al. (2004). Revised nomenclature for avian telencephalon and some related brainstem nuclei. J. Comp. Neurol. 473, 377–414. Riebel, K. (2000). Early exposure leads to repeatable preferences for male song in female zebra finches. Proc. R. Soc. Lond. B. Biol. Sci. 22, 2553–2558. Scharff, C., and Nottebohm, F. (1991). A comparative study of the behavioral deficits following lesions of various parts of the zebra finch song system: Implications for vocal learning. J. Neurosci. 11, 2896–2913. Scott, L. L., Nordeen, E. J., and Nordeen, K. W. (2000). The relationship between rates of HVc neuron addition and vocal plasticity in adult songbirds. J. Neurobiol. 43, 79–88. Searcy, W. A. (1992). Measuring responses of female birds to male song. In ‘‘Playback and Studies of Animal Communication’’ (P. K. McGregor, Ed.), Plenum Press, New York. Searcy, W. A., and Yasukawa, K. (1996). Song and female choice. In ‘‘Ecology and Evolution of Acoustic Communication in Birds’’ (D. E. Kroodsma and E. H. Miller, Eds.), Cornell University Press, Ithaca, NY. Slater, P. J. B. (1999). ‘‘Essentials of Animal Behaviour.’’ Cambridge University Press, Cambridge, UK. Sossinka, R., and Boehner, J. (1980). Song types in the zebra finch Poephila guttata castanotis. Z. Tierpsychol. 53, 123–132. Specht, R. (1997). ‘‘SAS Lab. Version 3.’’ Avisoft, Berlin. Suthers, R. A. (1997). Peripheral control and lateralization of birdsong. J. Neurobiol. 33, 632–652. Taka-Tsukasa (1917). ‘‘Kahidori’’ [in Japanese] Shokabo, Tokyo.

346

KAZUO OKANOYA

Thorpe, W. H. (1961). ‘‘Bird-Song: The Biology of Vocal Communication and Expression in Birds.’’ Cambridge University Press, Cambridge. Tinbergen, N. (1963). On aims and methods of ethology. Z. Tierpsychol. 20, 410–433. Todt, D., and Hultsch, H. (1996). Acquisition and performance of song repertoires: Ways of coping with diversity and versatility. In ‘‘Ecology and Evolution of Acoustic Communication in Birds’’ (D. E. Kroodsma and E. H. Miller, Eds.), Cornell University Press, Ithaca, NY. Uno, H., and Okanoya, K. (1998). Neural basis for song production in Bengalese finches: Effects of partial lesions on the higher vocal center. Tech. Rep. IEICE SP97, 137. Vu, E. T., Mazurek, M. E., and Kuo, Y. C. (1994). Identification of a forebrain motor programming network for the learned song of zebra finches. J. Neurosci. 14, 6923–6934. Washio, K. (1996). ‘‘Enigma of Bengalese finches’’ Kindai Bungei-sha, Tokyo [in Japanese]. Watanabe, A., and Aoki, K. (1998). The role of auditory feedback in the maintenance of song in adult male Bengalese finches Lonchura striata var. domestica. Zool. Sci. 15, 837–841. West, M. J., King, A. P., Eastzer, D., and Staddon, J. E. R. (1979). A bioassay of isolate cowbird song. J. Comp. Physiol. Psychol. 93, 124–133. Willson, E. O. (1975). ‘‘Sociobiology: The Synthesis.’’ Harvard University Press, Cambridge, MA. Wild, J. M. (1993). Descending projections of the songbird nucleus robustus archistriatalis. J. Comp. Neurol. 338, 224–241. Wild, J. M. (1997). Nerual pathways for the control of birdsong production. J. Neurobiol. 33, 653–670. Woolley, S. M. N., and Rubel, E. W. (1997). Bengalese finches Lonchura striata domestica depend upon auditory feedback for the maintenance of adult song. J. Neurosci. 15, 6380–6390. Woolley, S. M. N., and Rubel, E. W. (2002). Vocal memory and learning in adult Bengalese finches with regenerated hair cells. J. Neurosci. 22, 774–7787. Yakura, T., and Takeuchi, H.-A. (1998). Intracerebral zenk expression related to singing and hearing in Bengalese finch. Zool. Sci. 15, S102. Yamada, H., and Okanoya, K. (2001). Phonological changes of song elements in Bengalese finches (Lonchura striata var. domestica) singing in a helium atmosphere: Passive or active processes? Trans. Tech. Comm. Psychol. Physiol. Acoust. 31, 249–256. Yamada, H., and Okanoya, K. (2003). Song syntax changes in Bengalese finches singing in a helium atmosphere. Neuroreport 14, 1725–1729. Yodogawa, Y. (2000). Genetic relatedness and song complexity in wild and domesticated Bengalese finches. Senior thesis submitted to the faculty of culture, Tokyo University, Tokyo. Yoneda, T., and Okanoya, K. (1991). Ontogeny of sexually dimorphic distance calls in Bengalese finches (Lonchura dormestica). J. Ethol. 9, 41–46. Yu, A. C., and Margoliash, D. (1996). Temporal hierarchical control of singing in birds. Science 273, 1871–1875. Zahavi, A., and Zahavi, A. (1996). ‘‘The Handicap Principle: A Missing Piece of Darwin’s Puzzle.’’ Oxford University Press, Oxford. Zann, R. A. (1996). ‘‘The Zebra Finch: A Synthesis of Field and Laboratory Studies.’’ Oxford University Press, Oxford.

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 34

Behavioral, Ecological, and Physiological Determinants of the Activity Patterns of Bees P. G. Willmer* and G. N. Stoney *school of biology university of st. andrews fife ky16 9ts, united kingdom y institute of evolutionary biology university of edinburgh edinburgh eh9 3jt, united kingdom

I. Introduction A. Why Study Activity Patterns in Bees? An activity pattern is the change in levels of a particular activity through time. Activity pattern data form the basis of much of behavioral ecology for two simple reasons. First, they tell us how animals structure their time among a range of alternative behaviors. Second, by observing changes in activity patterns that animals make in response to natural or experimental variation in their internal state, or environment, or both, we can gain insights into factors structuring animal behavior. These factors include both biotic interactions between species and sexes, and aspects of the abiotic environment (such as ambient temperature) that constrain activity through limitations in the animal’s physiological ability. Within such constraints, activity patterns have the potential to reveal the currencies animals use to choose among available alternatives. So, why bees? Bees (Hymenoptera, Apoidea; see Table I) are an excellent taxon within which to investigate both the issues of constraint and currency. Their physiological constraints are better understood than those of almost any other invertebrate, allowing an unparalleled opportunity to integrate physiology into behavioral ecology. Many bees require elevated body temperatures (Tb) to fly (Section II), and hence thermal properties of their environment significantly constrain their activity (Section IV.A); however, many species are heterothermic, being able to elevate their 347 Copyright 2004, Elsevier Inc. All rights reserved. 0065-3454/04 $35.00

348

P. G. WILLMER AND G. N. STONE

TABLE I Families, Major Subfamilies, and Tribes within the Superfamily Apoidea (Bees), Locating Genera Mentioned in Text a Family

Subfamily

Tribe

Genera

Stenotritidae Colletidae Colletinae

Colletes, Leioproctus (þ s.g. Tetraglossula) Diphaglossinae Ptiloglossa, Caupolicana Hylaeinae Hylaeus (þ s.g. Prosopis) Andrenidae Andreninae Andrena, Melittoides, Heterosarellus? Panurginae Panurgus, Perdita, Melitturga, Anthemurgus, Nomadopsis Halictidae Rophitinae Rophites, Dufourea Nomiinae Nomia, Dieunomia, Pseudapis Halictinae Halictus, Lasioglossum (þ s.g. Sphecodogastra) Sphecodes,b Augochlora, Megalopta, Agapostemon Melittidae Dasypodainae Dasypoda, Hesperapis Melittinae Macropis, Melitta, Rediviva Megachilidae Megachilinae Lithurgus, Chelostoma, Hoplitis, Heriades, Osmia, Anthidium, Stelis, Coelioxys,b Megachile (þ s.g. Chalicodoma, Creightonella) Apidae Xylocopinae Xylocopa (þ s.g. Proxylocopa), Ceratina (þ s.g. Pithitis), Allodape, Braunsapis, Exoneura Nomadinae Nomada,b Epeolus Apinae Exomalopsini: Exomalopsis Emphorini Diadasia, Melitoma Eucerini: Eucera, Melissodes, Peponapis, Tetralonia, Martinapis, Synhalonia, Svastra, Xenoglossa Anthophorini: Anthophora, Amegilla Centridini: Centris Melectini: Melecta,b Thyreus b Euglossini: Euglossa, Eulaema, Exaerete Bombini: Bombus (þ s.g. Psithyrusb) Meliponini: Trigona, Melipona, Lestrimelitta, Plebeia, Scaptotrigona Apini: Apis a

After Michener (2000). Kleptoparasitic genera.

b

Tb endothermically when necessary (Section II.A), which gives some degree of escape from the usual thermal constraints on other entirely ectothermic insects. The thermal physiology of bees has received a great deal of research attention, revealing general underlying patterns but also

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

349

surprisingly subtle variations, in terms of both warming up and keeping cool. This background makes work on bees relevant to understanding the impact of variation in abiotic factors on the behavior of a far broader array of animals, including other heterothermic insects and heterothermic vertebrates such as shrews and hummingbirds. In terms of currency, the lives of most female bees (most familiarly the workers of social species, but also each individual female in the far more speciose solitary bee genera) are dominated by the need to visit flowers and collect pollen and nectar to provision their nest cells. Unlike the vast majority of insects, female bees are required to collect resources far beyond their individual needs, and their foraging abilities are translated directly into the number and quality (primarily determined by size) of their offspring. They are therefore predicted to be highly sensitive to variation in the availability and quality of floral resources. This is indeed the case, and analysis of responses of bees to spatial and temporal variation in these resources, whether natural or artificial, has revealed much of what is known of the basis of insect foraging behavior. This work has contributed substantially to much wider debates on animal foraging patterns, and to our understanding of how much learning can be achieved with a small and simple insect brain. Bees make hundreds of fast and sequential decisions when foraging: which flowers to visit and which to avoid; in which sequence, and how far apart; when to persist and when to give up. Given the amenability to training and handling (especially for social bees), and the easy availability of large numbers of individuals, they are excellent subjects for analysis of individual variability in learning, decision-making, and foraging, and of how far genetic differences may underlie such variation (see Harrison and Fewell, 2002; Page et al., 1995; Waddington, 2001). Three further aspects of bee biology make their activity patterns particularly amenable to behavioral analysis. First, bees almost invariably rely on just two food types, one of which (nectar) is certainly the most easily quantified of all resources, with simple measurements of volume and concentration giving an immediate conversion to caloric intake. Second, bees show the full spectrum of sociality (Section V), which allows investigation both of the ways in which limited time is allocated to a necessary sequence of behaviors, and of the ways in which time is allocated by individuals able to devote much or all of their effort to a single task. Third, females and males may have very different but interacting activity patterns; in particular, male bees show an unusually wide diversity of behaviors associated with locating mates and matings (Section III), and freed from the need to provision nest cells their activity patterns are far more diverse, directly reflecting those factors important in maximising paternal investment in a given mating system.

350

P. G. WILLMER AND G. N. STONE

More generally, bees are of huge importance as pollen vectors in most terrestrial ecosystems, and there is considerable evidence that the activity patterns of different bee taxa have had a significant selective impact on many aspects of floral biology. Understanding bee activity patterns is therefore crucial to understanding when and how gamete transfer takes place in a significant proportion of angiosperm plants. Where plants provide floral rewards for bees, timing of release of these rewards may both structure bee activity and in turn be sensitive to the energetic needs and constrained physiological abilities of bees. Feedback of this type is a potent driver of coevolution, the results of which range from interactions between specific pairs of bees and plants to effects structuring flowering in whole communities (Section VI). Where several plant species share a single pollinator there is the potential for competition for pollination, and bee activity patterns can therefore become a ‘‘resource’’ in their own right, to be partitioned or manipulated by other organisms. Finally, the link between pollinator activity and seed set has obvious relevance for agricultural crop success and for native plant conservation. Understanding the activity patterns of bee pollinators is therefore important when we seek to minimize detrimental effects of pesticide applications, habitat destruction and fragmentation, and other human interventions, and to preserve intact functional mutualisms within ecosystems. This review discusses the factors structuring activity patterns in a wide range of bees (for convenience we avoid obtrusive taxonomic terms in text by locating all the genera discussed in a resume of the bee superfamily in Table I). We aim to illustrate some of the more general behavioral and ecological questions that can be investigated with activity pattern data. Observed activity patterns are generated by interactions between properties of the bees themselves (intrinsic factors) and properties of their environment (extrinsic factors). Intrinsic factors include physiological differences between species (Section II) and between the sexes (Section III). Extrinsic factors (Section IV) are the fluctuating properties of the environment, and are either abiotic (such as temperature, light levels, precipitation) or biotic (availability of nectar and pollen in alternative floral sources, and levels of activity of potential mates, of competitors, or of predators). The results of these interactions vary widely in complexity and the underlying causes for observed patterns may be easy, or very difficult, to identify. We first discuss general implications of intrinsic differences among bees. We then discuss ways in which bee traits and environmental characteristics interact to generate a number of commonly encountered activity patterns, and illustrate these with examples drawn from the literature. Additional complexities resulting from sociality, especially the role of social facilitation of foragers, are examined in some

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

351

detail. Last, we consider some of the broader implications of bee activity patterns. B. How to Study Activity Patterns in Bees: Different Types of Activity Pattern Data Activity pattern data in the literature are broadly of three types: 1. Resource-based data describe activity of individuals at a particular resource. 2. Nest site-based data describe the timing of arrival and departure of individuals (usually females) at nest entrances, or count the numbers of bees entering or leaving communal nest entrances. 3. Total activity pattern data describe the allocation of time to the entire range of activities carried out over a sampled time interval. The sampling unit may be an individual (e.g., following a female solitary bee through all phases of her nesting cycle (see Section III.A), or a colony of a social species, over a period of days or weeks. These three types of activity pattern allow different inferences about factors structuring behavior. Resource-based data usually concern activity at flowers or feeding stations (although other possible resources include water or nest-building materials), and can usually be related to temporal patterns in resources arising from pollen release (dehiscence) or nectar secretion (Section IV). Activity patterns can also reflect more subtle changes in resource quality, such as the changes in pollen stickiness or nectar concentration that occur once the resources have been exposed to the microclimate surrounding the flower (Section IV.D). A resource-based approach has been used with dramatic success in those social bees that can be trained to visit a feeder (Section V.B). However, a potential weakness of data gathered at a single resource is that absence of activity could mean either that there is no activity at all, or that activity is continuing at another unobserved resource. Resource-based data thus become far more informative when a range of alternative resources is observed simultaneously (e.g., Stone et al., 1998). Data collected in this way additionally allow assessment of relative resource value, and hence the identification of the currency of choice that the bees use. As the examples in the following sections show, currencies of choice are not limited to caloric reward, but can include ambient or microclimatic temperature, nectar water content, or the number of flowers in a conspecific male’s territory. Once these currencies have been identified, activity patterns can be used to identify more subtle effects on

352

P. G. WILLMER AND G. N. STONE

individual behavior, such as the sexual harassment and paternal investment in offspring quality described in Section III. Nest-site based data can show how the timing and extent of activity by foragers change with time in response to potential structuring variables such as ambient temperature or the daily and seasonal phenology of flowering. Although details of foraging behavior at particular resources will remain unknown, it is possible, by analyzing pollen and nectar loads of returning bees, to reconstruct what the bees are doing away from the nest. This may be the only option where the plants chosen by bees are highly dispersed or otherwise inaccessible to full observation. For social bees, nest site data can conveniently be gathered remotely with photodiode monitors at the nest entrance (e.g., Morandin et al., 2001), potentially with balance pads to weigh bees entering and leaving, so giving some information on weight of resources gathered on each foraging trip. Total activity pattern data sets are the only type to provide a full picture of time allocation by an individual bee (and so potentially allow a comprehensive analysis of the currencies used to allocate time to different activities). For colonies of social bees this normally requires study of individually marked bees; without marking, there are particular complexities in interpreting instantaneous counts of bees (discussed by Corbet et al., 1993). It is also becoming possible with larger bees to use attached tracking devices to monitor some aspects of activity remotely (Capaldi et al., 2000), although again the information gathered is only partial. Even with such technological assistance, the extra information inevitably comes at a cost of increased sampling effort and time. Although some questions can only be answered using total activity pattern data, we show below that a combination of resource-based and nest-site based approaches can often shed light on the temporal structuring of different aspects of bee behavior.

II. Intrinsic Factors Affecting Bee Activity Patterns A. Overview of Intrinsic Factors Some of the most important general factors structuring bee activity are intrinsic. Properties such as size, taxon, and sex are constant for a given bee, and level of sociality is also usually fixed; but other properties, such as flight muscle performance, change over an individual’s lifetime. Size and phylogeny are strong predictors of several fundamental physiological traits in bees, particularly rates of radiative heat gain, endothermic heat generation, and heat loss (see Section II. B). The interactions between

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

353

patterns of temperature variation in the environment and the endothermic abilities of resident bees have strong implications for when and where particular bees can fly, and so for resultant activity patterns. In addition, the flight abilities of all bees are strongly influenced by age, with progressive wing wear during the few weeks of adult life gradually reducing flight performance. Sex is also of crucial significance (Section III), because for most of their adult lives female and male bees have very different priorities in resource and/or time allocation. Females need to collect resources far beyond individual requirements so that they can lay down provisions for their offspring. Their activity is generally structured by the need to gather pollen and nectar from one or more floral species, and to construct nests containing a series of larval cells in which these resources are stored. In contrast, males never stock larval cells, and they only need to collect sufficient nectar to fuel their own immediate needs and allow successful mating. Male activity can nonetheless always be tied, directly or indirectly, to floral biology (Section III.C). The only exceptions to the normal contrasts in activity between the sexes occur in kleptoparasitic bees (often termed cuckoo bees), in which individual females lay their eggs on the provisions gathered by females of other species. Freed from the need to construct and provision cells, female kleptoparasites show activity patterns quite different from those of provisioning species (Section III.D). Among female bees, a major differentiation exists between solitary species (see Section III.A) and social species (see Section IV). Females of solitary species complete all of the behaviors associated with nest construction and provisioning individually. This involves locating a suitable nest site and the construction, provisioning and sealing of a series of nest cells. A solitary bee female must allocate time among the different behaviors involved in the nesting cycle, each with its own physiological demands and balance of costs and benefits, but must also carry out the behaviors in an appropriate sequence. In contrast, in social species, different aspects of provisioning behavior and cell construction are often carried out by different subsets of the female worker population, and can be carried out simultaneously in different parts of the nest. Hence, while each worker may carry out each component of the nesting cycle at some time during her life (often in an age-related sequence), she does not perform them sequentially for any one cell, and does not need to allocate time accordingly. A final intrinsic activity-structuring factor is the degree of flower specialization shown by a given bee species or population. Many flowers attract bees, but a particular melittophilous flower type can be recognized, with relatively long zygomorphic blue/purple/pink corollas, landing platforms, nectar guides on the petals, pleasant scents, and moderate nectar rewards

354

P. G. WILLMER AND G. N. STONE

[see the ‘‘bee-pollinated syndrome’’ described in Faegri and van der Pijl (1979); although it is now widely recognized that the idea of pollination syndromes was overstressed in earlier literature, and that most flowers are in fact rather generalist in relation to pollinators; e.g., Johnson and Steiner (2000); Waser et al. (1996)]. While most bees show oligolecty or polylecty, collecting pollen from a range of ‘‘bee flowers’’ (and also from more open flower morphologies, and flowers with white and yellow coloration), a few are highly specialist and visit a single pollen source (monolecty). Monolecty is quite rare, and most known examples are solitary bees (Batra, 1978, 1984; Linsley, 1958; Roubik, 1989; Wcislo and Cane, 1996). For example, Hoplitis anthocopoides forages only from Echium flowers (Strickler, 1979), Anthemurgus from just one species of Passiflora (Michener, 2000), Andrena chaparralensis from only two species of Asteraceae (Neff and Larkin, 2002), Macropis only from loosestrife (Lysimachia) (Cane et al., 1983b; Pekkarinen et al., 2003), and Rediviva albifasciata almost entirely from Colpias (Steiner and Whitehead, 2002); and we note that the last two examples involve collection of floral oils, perhaps a resource that drives the evolution of highly specialized interactions (Roubik, 1989). Not only does flower specialization have very strong influences on the activity pattern of the bee itself, but it also impacts on the behaviors of other bees visiting the preferred plant, since the specialist is commonly more efficient and visits more flowers more quickly, leading to competitive interactions that may disadvantage or even deter other bee visitors (see Section IV.F). B. Heterothermy and Thermoregulation in Bees Bees, like other insects, are essentially ectothermic most of the time, and their activities are inevitably strongly linked with ambient temperatures and affected by surface properties such as color (reflectance) and insulation (reviewed in Willmer, 1983; Willmer and Unwin, 1981). In common with most other terrestrial ectotherms bees can use behavioral strategies to moderate their thermal interactions with the environment and give a degree of thermal independence from ambient conditions (most commonly employing long periods of basking to warm up, but also seeking shade or returning to the nest to cool down). In the absence of solar radiation most insects will inevitably have a minimum thermal limit for any activity, and somewhat higher minimum temperatures for specific activities such as flight (see Willmer, 1983). However, bees differ from most ectotherms in that they can also use endothermic mechanisms for endogenous warm-up; they have a facultative ability to switch on ‘‘shivering thermogenesis’’ using rapid opposing contractions of the large sets of indirect flight muscles (Heinrich, 1993), uncoupled from the wing base so that no flapping occurs.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

355

The flight muscles have a very high metabolic rate, and even when flying only 10–30% of the stored chemical energy is converted into wing muscle contraction (Dudley, 2000; Ellington, 1985), with the remainder being released as heat. Thus in a shivering bee all this endogenous heat can rapidly warm the thorax, and hence the whole body. Endothermic warmup is primarily used as a preflight mechanism, raising the thoracic musculature to a temperature well above ambient; without a period of preflight warm-up, many bees are unable to take off. This condition of combined ectothermy and facultative endothermy is commonly described as heterothermy; and in many ways it parallels the heterothermy shown by small vertebrates such as shrews and hummingbirds that are essentially endothermic most of the time but can become torpid by switching off the internal heat generation for parts of a day or of a season when it becomes too expensive to maintain. Although first demonstrated in social bumblebees and honeybees (e.g., Cooper et al., 1985; Heinrich, 1972, 1993), facultative endothermy is not restricted to just a few temperate bee genera but appears to occur in most species above 35–50 mg in body mass (Bishop and Armbruster, 1999; Stone, 1993a,b; Stone and Willmer, 1989a). Some bees that are active at low temperatures can show mean warm-up rates as  high as 10–15 C min1 (Stone, 1993a), exceeding the rates found in most endothermic vertebrates. Endothermy in most bees is commonly associated with flight, but in social species is also used during control of nest or swarm temperatures (Heinrich, 1993), brood incubation by individual queens (Heinrich, 1993), communication of resource quality during the honeybee waggle dance (Stabentheiner, 1991, 2001), trophallactic feeding (Farina and Wainselboim, 2001), and defense against hornets (Ono et al., 1995). Animals that are solely ectothermic are of course not necessarily thermoconformers but may achieve reasonably stable body temperatures by behavioral thermoregulation; but where endothermy is also an option it becomes markedly easier to maintain a stable high temperature, and many bees are therefore extremely good thermoregulators over a wide range of ambient temperatures. They may use behavioral strategies to achieve a stable Tb when working in the nest, but employ a sophisticated combination of physiological and behavioral mechanisms to keep a constant Tb  when in flight. For example, at an ambient temperature of around 20 C, honeybees, most bumblebees, and many moderate to large anthophorine  solitary bees maintain thoracic temperatures of 35 C or more when flying and foraging (e.g., Bishop and Armbruster, 1999; Chappell, 1984; Cooper et al., 1985; Heinrich, 1993; Heinrich and Buchmann, 1986; Jungman et al., 1989; Moffatt, 2001; Nicolson and Louw, 1982; Roberts and Harrison, 1998,

356

P. G. WILLMER AND G. N. STONE

1999; Roberts et al., 1998; Stone, 1993a,b, 1994a; Stone and Willmer, 1989a; Underwood, 1991). The metabolic cost of endothermy is high, but elevated thoracic temperatures are essential for many bees to be able to fly at all, because a high power output is essential for take-off and the power generated by the flight muscles increases with muscle temperature (Coelho, 1991; Josephson, 1981; Josephson et al., 2000a,b; Stevenson and Josephson, 1990). Thus each bee has a minimum muscle temperature below which it is unable to carry its own body mass in flight. Particularly in cool temperate habitats the necessary Tb elevation can be achieved only by endogenous physiological mechanisms. However, the cost of warm-up, and the time taken to reach a given thoracic temperature, both depend inversely on ambient temperature (Heinrich, 1975; Stone, 1993a), so the problems are compounded for bees in cold habitats. Furthermore, at low ambient temperatures a very high proportion of any heat generated is lost (Stone, 1993a). As a result, each bee species (and each sex within a species) has a minimum ambient temperature below which it will not even attempt preflight warm-up, thus placing a specific physiological limit on flight activity. This limit is influenced by how well any heat generated in the thorax can be retained by insulating structures such as internal air sacs or more obviously the external coats of dense hair giving a ‘‘fur’’ (Heinrich, 1993; May and Casey, 1983; Stone, 1993a), and by the extent to which heat transport away from the thorax in the hemolymph can be controlled (see below). Many bees active in the spring at moderate to high latitudes are therefore both highly endothermic and extremely well insulated, and can warm up to flight  temperatures from ambient temperatures at or below 0 C (Heinrich, 1993; Stone, 1993a). At such temperatures, relatively small solitary bees in the genera Anthophora and Osmia, with body masses of 100–200 mg, are  able to generate minimum thoracic temperatures for flight of around 27 C (Stone and Willmer, 1989a), and rather larger queen bumblebees (body  mass, 300–500 mg) maintain a thoracic temperature above 30 C. The high cost of endothermy underlying this level of thermoregulation and flight performance must be paid if a bee is to be active in times and places that give access to appropriate flowers. Those flowers must in turn offer sufficiently large rewards to offset the visitor’s foraging costs if they are to continue to be attractive to bees. This economic relationship was first made explicit by Heinrich and Raven (1972), and it underlies much of what we are concerned with here since it should be evident that thermal relationships will affect bee foraging patterns. The pattern of usage of endothemy is particularly important here. Foraging is often restricted to periods of insolation, to avoid or reduce the costs of endothermic warm-up, and both the time of daily first flight and the duration of foraging flights are related

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

357

to ambient temperature in much the same fashion as for any ectothermic insect (e.g., Willmer, 1983, 1985a,b). However, the option of using endothermy usually allows the first flight of a bee to be markedly earlier in the day (or season) than is possible for a purely ectothermic insect of similar size. The willingness of a bee to pay the cost of endothermy is not fixed, though, and will depend on its internal state; a given bee may be able to warm up under certain ambient conditions when well fed, but unwilling to expend the energy required to do so when starved or otherwise stressed. This can affect not only the time of onset of flight activity, but also the behavior once at the foraging site. Bees can maintain very different thoracic temperatures on different flowers (Kovac and Schmaranzer, 1996), and for both solitary bees (Stone, 1993a) and honeybees (Farina and Wainselboim, 2001; Moffatt and Nu´n˜ez, 1997; Waddington, 1990), the Tb recorded can be positively correlated with the availability and quality of nectar. Where alternative foraging strategies differ in their endothermic costs, metabolically cheaper strategies are sometimes preferred. For example, some bumblebees harvest nectar from platform-like umbellifer inflorescences by walking from flower to flower, and since flight is not required they reduce their energy expenditure by allowing Tb to fall (Heinrich and Heinrich, 1983). Similarly, some smaller bees can take advantage of the warm microclimate occurring within many long enclosed floral corollas (see Corbet et al., 1979), using the flower as a warming-up site; this ‘‘intrafloral basking’’ is used by some Andrena bees in spring foraging on Narcissus, allowing them to take off from one flower to the next without endothermy (Herrera, 1995). But in most other situations, continuity of flight and hence of elevated thoracic temperature is a prerequisite for successful flower visiting. While minimum ambient temperatures for activity are determined by the ability of bees to warm up, the maximum ambient temperatures at which bees fly are sometimes dictated by the risk of critical overheating. With few exceptions (e.g., Stone, 1993a) most bees have upper critical temperatures (UCTs) of 45–50  C. Because heat is inevitably generated in flight, the maximum ambient temperature at which a bee can keep flying may in practice be somewhat below its UCT. In very hot environments, bees may therefore be precluded from flying during periods in which other less metabolically active insects, such as beetles and ants, can maintain activity (Willmer et al., 2000). The danger to bees that do approach their UCT is illustrated by the work of Chappell (1984) on the bee Centris pallida, in which males patrol open sandy areas of their semidesert habitat for emerging females. On hot days, males that fly too low encounter the layer of hot air just above the sand, overheat until they lose flight control, and fall to the ground where they quickly die.

358

P. G. WILLMER AND G. N. STONE

Potential overheating problems in female bees are exacerbated by their need to carry significant loads (commonly up to 30–40% of their own body weight; see Section IV.B). Pollen is the more expensive resource to carry, since pollen-bearing honeybees have hovering metabolic rates about 10% higher than nectar bearers (Feuerbacher et al., 2003). The Tb of pollen bearers is also slightly higher in flight, so the upper ambient temperature limit for pollen gathering may be a little below that for nectar gathering for a given species or indeed a given individual bee. Given the impact of temperature on bee activity, bees could potentially benefit by regulating the rate of heat generation associated with a given level of flight muscle activity. Flying Apis mellifera and Centris pallida do decrease their metabolic rates in flight as ambient temperature rises, primarily mediated by changes in wingbeat patterns (Roberts and Harrison, 1998, 1999; Roberts et al., 1998b), but it remains unclear how widespread taxonomically such abilities are, and Moffatt (2000, 2001) points out that for Apis this decrease in metabolic rate can be more than offset by an increasing metabolism as reward intake increases. Several other mechanisms could potentially allow avoidance of overheating in bees that need to fly in hot weather. They could (1) tolerate high thoracic temperatures through evolution of a high UCT, (2) intermittently stop flying when the heat load is too high, or (3) increase their rate of heat loss. To date there is little evidence for selective modification of UCTs in bees from different thermal environments, although some desert Megachile species show apparent selective variation in this trait (Barthell et al., 2002). The commonest mechanisms for avoidance of overheating are therefore mechanisms 2 and 3, both essentially behavioral, and extremely relevant here as they may compete for time with other activities. The simplest strategy is to avoid flight in the warmest hours of the day or in the warmest microclimates (Linsley, 1978; Stone et al., 1988; Willmer, 1983; Willmer and Stone, 1997a), or if active in such conditions to cease flight periodically and seek temporary shade just before upper critical temperatures are reached. Alternatively, heat loss can be increased, by either behavioral or physiological means. Shade seeking, in the nest or within the shelter of vegetation, is an easy option, and convective cooling can be enhanced by choosing a site with significant air movement. Some bees elevate their convective heat loss through increased flight speed (e.g., Chappell, 1982), although this is of limited use during active foraging since it may be incompatible with efficient flower visitation. Some euglossine bees show lateral wing buzzing when perched, the frequency increasing with ambient temperature, which may enhance heat loss by forced convective cooling (Stern and Dudley, 1991). A final behavioral mechanism, ‘‘tongue lashing,’’ involves the extrusion of a droplet of nectar onto the tongue; heat loss through evaporation

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

359

of water cools the head and, by conduction, the thorax (Heinrich, 1979a). Such evaporative cooling is probably not common in bees; honeybees use it to good effect in arid environments (Cooper et al., 1985), and in colony thermoregulation (Heinrich, 1993), but are unusual in that their social organization allows individual bees to recoup the water lost once back at the nest, since water is gathered for nest cooling (Schaffer et al., 1979; Schmaranzer, 2000). Solitary bees of similar size usually cannot afford to lose the milligrams of water required to maintain tolerable thoracic temperatures at high ambient temperatures (but see Chalicodoma in Section IV.D.3, and Cooper et al., 1985). The alternative of physiological regulation of heat loss is known only in bumblebees and some anthophorine solitary bees (Coelho, 1991; Heinrich, 1976; Heinrich and Buchmann, 1986; Stone, 1993a). It is achieved by controlling patterns of blood (hemolymph) flow in the petiole between the thorax (housing the heat-generating musculature) and the cooler, relatively uninsulated abdomen. When these bees need to conserve heat in the thorax during warm-up, petiolar countercurrent heat exchange occurs with the warm blood flowing back from the thorax, transferring its heat into the cool blood flowing forward from the abdomen. However, when there is a risk of overheating the steady countercurrent flow is replaced by alternating pulses of blood in each direction, allowing export of heat from the thorax. The hot blood entering the abdomen then flows close to the hairless ventral surface, allowing heat loss by radiation and convection (Coelho, 1991; Heinrich, 1976; Heinrich and Buchmann, 1986; Stone, 1993a). Clearly, bees are subject to conflicting thermal demands. Most environments in which they are active as flower visitors are characterized by strong diurnal fluctuations in ambient temperature, and adaptations associated with maintaining a high Tb in cold air (e.g., at dawn and dusk) are at least partially counteradaptive to maintaining activity at high ambient temperatures in the middle of the day. This means that in thermally fluctuating environments activity by bees is often constrained within a ‘‘window’’ of physiologically tolerated ambient temperatures (see Section IV.A). Thus the portion of the day during which a bee is physiologically capable of foraging activity imposes crucial limits, within which other considerations—such as resource availability in the flowers, and the relative value of alternative floral resources—may provide the finer grained cues for activity patterns. Bees show substantial interspecific and intraspecific variation in endothermic and thermoregulatory abilities, which in turn influences activity patterns. The major causes of such variation are now briefly discussed.

360

P. G. WILLMER AND G. N. STONE

1. Size-Related Variation in Thermal Relationships Warm-up rates using both external and internal heat sources show strong relationships with size that can be predicted from scaling relationships. For ectothermic warming, the high surface-to-volume ratio of small bees means that they rapidly absorb heat energy from the environment (e.g., when basking), but also cool rapidly when the heat source is removed. Large bees warm more slowly, but under the same conditions will equilibrate at a higher body temperature (Willmer and Unwin, 1981). During endothermic warm-up, the same considerations mean that small bees lose a higher proportion of the heat they generate, and for a given rate of heat generation will reach a lower equilibrium thoracic temperature. Hence, where bees of a range of sizes have the same heat-generating ability per unit mass of flight muscle, larger bees are predicted to show higher warm-up rates. It is generally true both within and among species that larger bees are able to warm up faster, and to initiate flight activity at lower ambient temperatures, than smaller bees (Heinrich and Heinrich, 1983; May and Casey, 1983; Stone, 1993b, 1994a; Stone and Willmer, 1989a; Willmer and Unwin, 1981). The ability of larger individuals to maintain higher thoracic temperatures means in turn that their thoracic flight muscles are able to generate more power per unit mass (Josephson, 1981). Larger female bees are therefore not only active at lower ambient temperatures, but are also able to carry relatively larger masses of pollen and nectar than smaller females flying under the same conditions (Stone, 1994b). Conversely, at high ambient temperatures their lower surface-to-volume ratio acts against larger bees, limiting their ability to lose heat quickly and making them more likely to overheat than small bees; thus smaller bees are commonly able to maintain activity longer at higher ambient temperatures (e.g., Herrera, 1990; Willmer and Stone, 1997a). Small bees also benefit from higher rates of heat loss per unit area of thorax, either through possession of shorter or less dense pubescence and hence less insulation (May and Casey, 1983; Underwood, 1991) or through higher thermal emissivity of the thorax linked with the exposed darker cuticle (Willmer and Unwin, 1981). The thermal effects of body size can operate within a single species. The tropical bee Amegilla sapiens is found in Papua New Guinea from altitudes of 2000 m down to sea level. High-altitude populations have larger body sizes and are more endothermic than populations at sea level. This in turn allows them to forage at lower minimum ambient temperatures, but also results in cessation of activity at lower maximum ambient temperatures (Stone, 1993b). Thermal effects combine to produce the general rule that very small bees are primarily ectothermic and relatively poor thermoregulators, while large

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

361

bees are highly endothermic and often very good thermoregulators. They also generate repeatedly observed patterns in the sizes and taxa (of insects in general, but of bees in particular) that are active in particular habitats under particular conditions (Pereboom and Biesmeijer, 2003; Willmer, 1983; Willmer and Stone, 1997a). Larger bees (and also species that are darker colored or more hairy, since these factors likewise affect rates of heat exchange) are active early and late in the day, giving a bimodal activity pattern (Fig. 1) and they are also generally active earlier in the season. Smaller bees (often relatively hairless or lighter colored) are commoner in the middle of a day and in the hotter months of the season (Potts

Fig. 1. A comparison of daily activity patterns in two species of Xylocopa foraging for nectar on Sodom apple (Calotropis procera) in Israel; ambient temperature through the day is also shown. Xylocopa pubescens is active at cooler temperatures, both earlier and later in the day, being larger and more densely hairy than X. sulcatipes. Data from Willmer (1988).

362

P. G. WILLMER AND G. N. STONE

et al., 2003a,b; Shmida and Dukas, 1990). Comparisons of activities of large and small bees foraging on a single plant species (lavender) are shown in Fig. 2. It follows that environments with given thermal characteristics tend to have bee faunas with similar combinations of size and endothermic ability; and where the environments are in similar biogeographic provinces, species active under similar conditions are often from phylogenetically close taxa. For example, in the Northern Hemisphere the spring bee faunas commonly include highly endothermic and relatively large-bodied Bombus, Anthophora, and Osmia species, while smaller bodied and less endothermic genera such as Andrena, Colletes, and Halictus are relatively more abundant in the summer. Turning the relationship around, demonstration of seasonal and/or daily temporal patterning in body size can be strong evidence that considerations of thermal physiology underlie observed activity patterns. 2. Phylogenetic Variation in Thermal Relationships Having taken into account the effects of body size, bee taxa vary substantially in a number of interlinked aspects of thermal physiology, including rates of heat generation and loss, minimum thoracic temperatures required for flight, and thermoregulatory ability. These issues are discussed in detail elsewhere (Bishop and Armbruster, 1999; Dyer and Seeley, 1987; Stone, 1994a; Stone and Willmer, 1989a) and are reviewed only very briefly here. Members of some taxa are routinely able to warm up more rapidly under standard conditions, and maintain activity at lower ambient temperatures, than others. Examples of highly endothermic genera include Osmia, Anthophora, Apis, and Bombus. Genera showing low rates of warm-up for their body size include Andrena, Colletes, Anthidium, Xylocopa, and many tropical euglossine bees. Endothermic ability is a good general predictor of the range of abiotic conditions (on daily or seasonal timescales) under which a given taxon is physiologically able to maintain flight activity. Highly endothermic bees are generally well represented in higher latitude spring bee faunas, and in these habitats are able to fly at low ambient temperatures. Poorly endothermic taxa are usually more abundant only in the summer at similar latitudes, or if active in spring (some temperate Andrena and Colletes, e.g.) they occur only patchily in warmer microclimates (e.g., Herrera, 1995; Michener and Rettenmeyer, 1956). Phylogenetic variation in endothermy also underlies the patterns described in Sections II.B.3 and II.B.5 below.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

363

Fig. 2. Activity patterns of four different bees of different mean body mass, foraging at lavender flowers (Lavendula stoechas) in southern Spain, with increasing bimodality for the larger bees. Data from Herrera (1990).

364

P. G. WILLMER AND G. N. STONE

3. Biogeographical and Environment-Related Variation in Thermal Relationships Bees active in routinely warm habitats, such as the humid tropics, generally have thoracic temperatures in flight that are close to those recorded in bees from temperate environments (Stone and Willmer, 1989a,b), but because warm-up is initiated from higher ambient temperatures, they need to raise their Tb by a smaller increment before flight, and thus face much lower average costs of endothermy. For example, Bombus and Anthophora active in the British spring routinely elevate thoracic temperatures more  than 25 C above ambient levels, whereas many bees active in tropical habitats only need an elevation of 5–10  C (Casey et al., 1985; Stone and Willmer, 1989a). This probably explains the generally lower warm-up rates achieved under standard conditions by tropical bee genera (Stone and Willmer, 1989a). However, this generalization applies only to tropical habitats that are consistently warm. Tropical deserts show the most dramatic ambient temperature fluctuations of all bee-occupied habitats, with temperatures during the night often approaching freezing (Section IV.A), and bees active early in the morning in such habitats include the most endothermic species known (Stone, 1994a). Indeed, deserts are dominated by matinal, crepuscular, or bimodally active bees avoiding the dry heat of the middle of the day and coping instead with the rather cool dawn and dusk conditions when rapid warm-up is appropriate (Willmer and Stone, 1997a). 4. Sexual Variation in Thermal Relationships Bees show enormous variation in mating system (Section III.B), and this is reflected in variation in the relative endothermic abilities of males and females and their resulting activity patterns. With the exception of kleptoparasites, and the queens of social species once their workers are reared, female bee reproductive success results from the ability of individuals to stock cells. The broader the range of environmental conditions over which a female can forage, the greater her reproductive success is likely to be (Stone, 1994b). The diversity of male mating tactics found in bees makes general patterns for males more difficult to identify, although it must still generally be true that greater independence of ambient conditions results in greater reproductive success. At one end of the spectrum, the males of many species of Andrenidae, Colletidae, and Megachilidae are significantly smaller than females and, where this has been quantified, they are also significantly less endothermic (Stone, 1993b). This predicts that males should have a narrower period for activity on a given day than conspecific females. In such cases, the males are often short-lived, courtship and

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

365

mating do not involve direct intermale competition, and male activity is often restricted to warmer microhabitats (such as nest sites or flowers) and warmer parts of the day (e.g., Stone, 1993b). At the other end of the spectrum are species in which male reproductive success is closely coupled to physiological performance, often through intraspecific or interspecific competition. In species such as the anthophorine solitary bees Centris pallida and Anthophora plumipes, males compete directly for access to females, and larger more endothermic males are better able to patrol encounter sites, and more likely to win contests for access to females (Chappell, 1984; Stone et al., 1995). In social swarming species, such as members of the genus Apis, males have a much enlarged thorax relative to queens, in response to selection for prolonged flight and intense competition with other males within the mating swarm (Roubik, 1989). Male endothermy is most elevated in bees such as Anthidium, in which the males guard floral resources not only against conspecific males, but against the foraging females of other species, and mate with visiting conspecific females (see Section III.B). Anthidium manicatum males in Britain commonly defend patches of labiate herbs (e.g., Lamium, Salvia, and Stachys) in which nectar secretion occurs in the morning several hours before the arrival of females (G. N. Stone, unpublished data). To maintain patch quality, the males must establish their territories before the early morning arrival of highly endothermic honeybees and bumblebees. In a reversal of the typical bee pattern, territorial male A. manicatum are therefore much larger and have far higher warm-up rates than conspecific females (G. N. Stone, unpublished data). 5. Endothermy in Kleptoparasitic Bees Because kleptoparasites are freed from the need to stock their own cells, their reproductive success is less tightly coupled with independence of flight activity from ambient conditions (Stone, 1993b; Stone and Willmer, 1989a), and more dependent on their ability to successfully enter the nests of other bees. Nest-provisioning species are commonly attacked by kleptoparasites within the same bee family (Roubik, 1989); for example, bumblebees (Bombus) are attacked by Psithyrus; Anthophora and Amegilla are attacked, respectively, by Melecta and Thyreus; and Creightonella is attacked by Coelioxys. In all cases, having controlled for differences in body size, these kleptoparasitic species have lower endothermic abilities than their nest-provisioning relatives (Stone and Willmer, 1989a); a striking example of convergent impacts of life history strategy on thermal physiology. To summarize this section, it should be evident that physiological factors relating to body temperature are critical determinants of bee activities, and

366

P. G. WILLMER AND G. N. STONE

interact on many fronts with weather conditions to constrain foraging times (Willmer, 1991a). Climatic variables affect both minimum and maximum ambient temperatures at which a given bee can warm up and then maintain flight, giving a ‘‘thermal window.’’ The limits of this window will depend on intrinsic factors such as bee size, insulation, and endothermic ability (all of which show biogeographic and habitat-related trends), and also on how much extra load a bee has taken on during foraging. The profitability of foraging can also vary with environmental parameters such as wind and rain, which can therefore also constrain the activity window: for example, bumblebees show a 3- to 4-fold greater range of metabolic rates (as measured with doubly labelled water) when free-flying in natural environments with varying windspeeds compared with flights in controlled greenhouses (Wolf et al., 1999). Wherever environmental factors narrow the window and make transport costs higher within that window, bees must require greater food inputs from more flowers per trip to offset the costs of any foraging trip they choose to make. C. Light Levels and Flight Activity in Bees While many bees can forage in low light intensities around dawn or dusk (Cazier and Linsley, 1963; Linsley, 1960, 1962; Linsley and Cazier, 1970; Roubik and Michener, 1984; Wolda and Roubik, 1986), the great majority of bee activity occurs during the day. In some tropical or semidesert habitats with wide daily fluctuations in temperature, a few large bee species specialize in matinal (at or just before dawn) or crepuscular (at dusk) foraging, and a very small number are genuinely nocturnal. Male Ptiloglossa guinnae are active only at dawn, with light levels of less than 1 lux, while females forage both at dawn and dusk, requiring light levels of about 10 lux when they are only just visible to a human observer (Roberts, 1971). In India, the giant honeybee Apis dorsata makes predawn visits to the nocturnally flowering leguminous tree Pterocarpus by moonlight (Rao et al., 2001). Several studies report nocturnal foraging in carpenter bees: Proxylocopa visit caper plants (Capparis) in Israel after dusk (Dafni et al., 1987), Xylocopa tranquebarica forages at late dusk and occasionally into the night in Thailand (Burgett and Sukumalanand, 2000); and in India X. tenuiscapa visits Heterophragma trees solely at night and independent  of any moonlight, at air temperatures of 2–14 C (Somanathan and Borges, 2001). The genus Sphecodogastra is equally specialist at nocturnal activity, synchronizing its peak foraging time with lunar periodicity (Kerfoot, 1967), and Megalopta bees gather pollen from Parkia trees in the Amazonian forest deep into the night (Hopkins et al., 2000). These cases all involve

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

367

visitation of flowers that only begin to open in the night, perhaps allowing both plant and pollinator to avoid high daytime temperatures, with the bees perhaps also benefiting from reduced predation and parasitism (see Section IV.F). The rarity of nocturnal activity in other bees may be related to at least two key constraints. First, bees rely strongly on photoreception and have good visual learning abilities (see Section V.B). Visual perception plays a huge part in their recognition of floral species and of floral traits such as flower age and flower color (e.g., Faegri and van der Pijl, 1979; Kevan, 1983; Lunau, 1996; Weiss, 1995) as well as in their sexual interactions (e.g., Raw, 1975; Severinghaus et al., 1981; Stone et al., 1995). The bee eye is of the compound apposition type associated with day-active species, far from ideal at night since individual ommatidia cannot gather enough light. Warrant et al. (1996) suggest that some bee eyes have better spatial resolution in dim light than predicted from simple optics, implying some higher neural summation mechanisms, but even so visual recognition at dusk or by moonlight is relatively poor (and most of the nocturnal bees referred to above have unusually large simple ocelli to enhance their general photoreceptivity). In practice, most nocturnally active insect pollinators, such as moths, locate their floral nectar sources mainly using olfactory cues (Raguso, 2001). While scents are of course extremely important in many aspects of bee behavior—including foraging associated with resource location (e.g., Dobson, 1987) and floral memory leading to fidelity (Menzel and Muller, 1996; and see Section V.B)—the majority of their long-range floral detection seems to be visual (Giurfa and Lehrer, 2001), with olfactory cues usually altering flower choices only over relatively short distances. The strong dependence on long-range visual cues from flowers may be a major reason why most bees seem not to locate or exploit nocturnally flowering plants. Second, perhaps because of their limited ability to see, bees are able to fly only slowly in very low light, and this may have a significant negative impact on the cost–benefit balance of foraging at night. The cost of flight is effectively independent of forward speed in bees (Ellington, 1990), so that slow flight is as metabolically expensive as faster flight. If the harvesting rate of floral resources is (at least initially) positively correlated with flight speed, activity at night may yield too poor a net benefit to support provisioning bees. Furthermore, longevity of bees is negatively correlated with energy consumption (Neukirch, 1982), and while the immediate metabolic costs of nocturnal flower visiting can potentially be met from the floral resources gathered, the decrease in future longevity per cell provisioned is likely to be greater for nocturnal foraging.

368

P. G. WILLMER AND G. N. STONE

D. Circadian Rhythms in Bee Activity Many species of bee show substantial nocturnal periods of complete inactivity. Several solitary bee genera (e.g., Epeolus, Anthophora, Chalicodoma, Melecta) exhibit ‘‘sleep-like’’ states at night (Kaiser, 1995), and male Pseudapis, Amegilla, and Anthophora show gregarious sleeping with a mass dispersion flight in the morning (e.g., Aldini, 1994). ‘‘Sleeping’’ bees often rest within vegetation rather than returning to a nest; for example, some Nomia species sleep in hedgerow vegetation, and the aptly named Chelostoma florisomne sleeps in flowers, particularly harebells. Usually we do not have any evidence that these phenomena have an endogenous circadian component. Indeed, there is relatively little information on endogenous circadian rhythms for most bees. However, for honeybees there is a growing literature on inherent rhythmicity (reviewed by Moore, 2001), and here sleep occurs only at certain times and is dependent on the task being performed (Kaiser, 1988; Kaiser and Steiner-Kaiser, 1983). Typically sleep develops only with maturity, and then it does show clear circadian patterning (Toma et al., 2000). Nurse honeybees looking after the brood must do so throughout the 24-h cycle and do not sleep (Moore et al., 1998), whereas foragers (usually the older workers, 2–3 weeks old) commonly go into a rest state for up to 7 h a day and show all the characteristics of circadian rhythmicity. Interestingly, when climatic conditions prevent foraging or when nurse bees are severely depleted, some foraging bees revert to nursing, and they then lose rhythmicity completely, uncoupling or overriding the cellular clock mechanism linked to the period gene (Bloch and Robinson, 2001; Bloch et al., 2001). This is the first clear demonstration of such plasticity in behavioral chronobiology, with social factors exerting an influence on regulation of ‘‘clock genes.’’ Since they do have a clock mechanism, foraging honeybees can make precise associations between food presence and time of day, dependent on their endogenous periodicity. These inbuilt rhythms are used for sun compass navigation as well as for timing visits to flowers, and must play a key role in determining activity patterns outside the hive. How far inherent timing systems are present in other bees is uncertain, although there are indications of circadian effects in the foraging of eusocial Scaptotrigona bees when kept under constant light (Bellusci and Marques, 2001). E. Aging and Flight Many insects show age-related changes in flight muscle performance, and this is expected to result in corresponding changes in individual behavior and activity patterns. There is a general deterioration of flight muscle

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

369

performance with age (Collatz and Sohal, 1986), a widespread phenomenon documented in 20 families in 8 insect orders (Johnson, 1980). Progressive physical ‘‘wing wear’’ in aging insects is also well documented, with the trailing edge of both fore and hindwings becoming progressively tattered. In practice, these changes serve to limit the foraging lives of flying insects. The best illustration of this effect is the demonstration that, regardless of the period over which she completes them, a worker honeybee dies due to senescence after flights totalling approximately 800 km (Neukirch, 1982). More generally, there is an obvious trade-off between foraging effort and longevity (e.g., Schmid-Hempel and Wolf, 1988), and many bees do live much longer as adults than the intenisvely foraging honey bees. These effects may underlie a number of known age-related changes in bee activity patterns. The behavioral polyethism of social bees is discussed below (Section V), and it is possible that wing muscle maturation underlies the early hive-based duties of worker honeybees. Deterioration of flight muscle performance may also explain why, in a range of bees in which males are territorial, younger males are generally more successful than older males (Roubik, 1989). III. Sexual Differences Affecting Bee Activity Patterns A. Nesting Cycle of Female Bees Bees (particularly solitary species) exhibit a diverse array of nesting strategies with respect to the part of the habitat they nest in, the type of substrate they use, and the materials required for nest construction, all of which are reviewed elsewhere (Batra, 1984; Cane, 1991; Linsley, 1958; Michener, 1974; Potts et al., 2003b; Roubik, 1989; Wuellner, 1999a). On the basis of their nesting behaviors, bees can usefully be partitioned into ‘‘miners,’’ ‘‘masons,’’ ‘‘carpenters,’’ and ‘‘social nesters’’ (O’Toole and Raw, 1991). 1. Miners (or diggers) excavate holes in the ground or in vertical surfaces such as earth banks. The cells are commonly lined with glandular secretions that may enhance waterproofing and suppress fungal growth (Cane et al., 1983a; Hefetz et al., 1979; Norden et al., 1980). All species of Andrenidae and Melittidae are miners, as are most Halictidae and Colletidae, some Megachilidae, and the anthophorine Apidae. Reuse of existing burrows over succeeding generations is common, and may represent an adaptation to accelerate nest construction. 2. Masons generally construct their nests in preexisting cavities, such as pithy plant stems, small rock cavities, abandoned insect burrows, or

370

P. G. WILLMER AND G. N. STONE

even snail shells; a few construct their own nests from sands and muds (potter bees). Most belong to the family Megachilidae, and line their nests with gathered materials (fibres, waxes, etc.), rather than with glandular secretions. Leaf-cutter bees (Megachile) are a subgroup of masons, lining their nest with freshly gathered soft leaf material. In these cases, additional activity outside the nest is needed to gather the lining materials, and completely separate trips (involving few or no flower visits) may be made. 3. Carpenters excavate their own nests in woody substrates; this behavior is found in three genera of Apidae (Xylocopa, Ceratina, and Pithitis) and one megachilid genus (Lithurgus). Xylocopa species first excavate a shaft, and then construct cells within it, filling the shaft back toward the entrance. Additional trips to provide celllining materials are often unnecessary, since chewed wood is a common cell-lining material. 4. Social nesters use bigger preexisting cavities to build large social nests; all are in the subfamily Apinae, including honeybees, bumblebees, and stingless bees. Nest construction and maintenance here form a substantial part of the overall activity budget, but may be restricted to some portion of the workforce or some part of the life cycle of individual bees. 5. Finally, kleptoparasites do not construct nests at all, but instead parasitize the nests of other bees. Until recently most female bees have been reported as monandrous, becoming nonreceptive after they have mated once, although some anthidiine and panurgine bees are clearly polyandrous and other taxa are turning out to be less restricted to single inseminations than was once thought. In all cases, however, the nesting cycle normally begins with mating, which commonly occurs either at the emergence site, or while foraging at flowers (see Section III.B). An inseminated female bee then starts to search for a specific nest site. Her choice will depend on suitable substratum availability, but is often also affected by a preference for gregariousness and a philopatric tendency leading to nesting adjacent to her own emergence site (Michener, 2000; Potts and Willmer, 1998; Wuellner, 1999a). This searching phase can last for several days, and the responses of the female to variation in ambient conditions during this period can be very different from that recorded once she starts to provision cells. For example, a female Anthophora plumipes digs multiple nest tunnels during her life, in each of which she builds 7–10 cells (Stone, 1994b). Location of new tunnel sites takes 2–3 days, during which

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

371

the searching female emerges from her natal or current nesting tunnel only during good weather, and ceases flight under conditions in which she will readily fly when provisioning cells. This recurrent change from risk-averse behavior while searching for a nest site to greater risk taking while provisioning (Fig. 3) probably reflects the difference in investment and potential value between a tunnel and a completed nest cell (Stone, 1994b). For our purposes we can use the nest activity cycle of a mason bee as a useful illustration (Fig. 4). Site selection is followed by excavation and then initiation of the first cell, which involves a considerable investment of time and energy (in one species of miner, excavation of a 7-cm shaft required almost 2 days of continuous labor through day and night; Anzenberger, 1977). During these initial phases, the female commonly makes occasional nectar-foraging trips to offset the cost of nest building. Once the first cell cavity has been constructed, it is usually coated internally with the speciesspecific cell-lining material. These ‘‘housekeeping’’ activities within and around the nest often constitute a considerable proportion of the total daily and/or seasonal activity period, and preclude productive floral foraging for at least some of the period over which the bees are active (Stone, 1994a; Willmer, 1986; Willmer and Stone, 1988). If construction activity around the nest characteristically occurs at a particular part of the day, it will necessarily affect the activity periods observed at forage sites. Only once a cell is finished and lined can the female begin to stock it with food for the larva. The vast majority of bees stock their cells only with pollen and nectar (although a few bee species also incorporate floral oils, elaiophore lipids from seeds, carrion, and even feces in their larval provisions; see Section IV.B). Individual visits to flowers involve collection of pollen only, nectar only, or both. Some bees use the same flowers for both resources, while others visit discrete pollen and nectar sources in different locations and/or at different times of day; this can be relatively independent of how specialist or generalist (polylectic) a bee species is. Again, the timing of foraging trips to each floral resource can only occur in the bees’ permitted thermal window (see Section II), and will be further constrained by the timing of nectar secretion and pollen dehiscence in the requisite flowers (see Section IV). If the availability of floral resources is highly structured in time (Section IV.B–D), female bees should prefer to construct their nest cells during periods of low resource availability. The number of trips required to fully stock a cell will depend on how quickly the bee can reach and harvest resources from appropriate flowers; ideally a provisioning female will be able to choose flowers close to her nest, and if her thermal window matches the dehiscence or nectar secretion

372

P. G. WILLMER AND G. N. STONE

Fig. 3. Patterns of male and female Anthophora plumipes behavior, scored as numbers active at the nest site, for males () and for females that are nest searching (&) and nest stocking (&). In (A) the ambient temperature profiles for the 2 days shown in (B) (upper trace) and (C) (lower trace) are given, the higher temperatures on day B giving markedly greater overall activity; but note that on both days females already provisioning their nests are active much earlier and often later in the day, while nest seekers are more risk averse and are  active only when ambient temperatures rise to about 10 C. Data from Stone (1994b).

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

373

Fig. 4. A typical bee nest cycle, in this case Chalicodoma sicula from Israel; the nesting and foraging period is late February–early May. At various points in the repeated nest-stocking cycle within these dates, females also make some trips to gather nectar for their own fuelling needs (shown by N). From Willmer (1986).

patterns of the plant she should visit the flowers soon after they open, before the supplies are depleted by other visitors. But the number of pollen grains needed per cell can run into millions (e.g., Cane et al., 1996; Westerkamp, 1996), and even in ideal conditions this can represent visits to several hundred (or sometimes over 1000) flowers. The time taken to

374

P. G. WILLMER AND G. N. STONE

Fig. 5. Diagrammatic representation of activity patterns recorded at the nest site for Creightonella frontalis in Papua New Guinea. (A) Times of first arrival and final departure

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

375

complete and stock a cell will thus reflect the local abundance of floral resources and the length of the diurnal activity window. For many bees, one new cell (or rarely two) can be constructed and filled each day. Therefore nest construction commonly takes place late in the evening and/or around dawn, while flower visits dominate during daylight. These foraging trips are often short and efficient early on but lengthen later in the day as resources dwindle; an example of this was reported by Cameron et al. (1996) for Melissodes bees, which made six or seven trips per day to Aster and Erigeron flowers, these trips becoming much longer in the afternoon as pollen was depleted from the flowers. Figure 5 illustrates the same phenomenon with the daily pattern of trips recorded for the megachilid bee Creightonella frontalis (Willmer and Stone, 1988). The importance of synchronizing nest construction with daily availability of resources is illustrated by the atypical female in Fig. 5, which constructed its cells during the day, missed the periods of optimum resource availability, and was thus forced to make many more foraging trips, each far less productive. Some bees operate much faster cycles even where only one cycle occurs each day: for example, Perdita texana requires only 72 min to complete a cell, with nectar trips averaging 17 min and pollen trips just 6 min (Neff and Danforth, 1991); rapid cell construction in this case is probably facilitated by the fact that both pollen and nectar are harvested from a locally superabundant resource, the prickly pear cactus Opuntia. At the other end of the spectrum, many species of bee (perhaps especially in cooler and/or low-resource habitats) take several days to collect adequate floral resources for a single cell, and diurnal synchronization with floral resources is less critical. Once a cell has been filled with the appropriate volume of pollen and nectar, the female may need to adjust the balance of these two components. If the mix is too liquid (as a result of too much dilute nectar), the egg may become submerged or overgrown by fungus (although in some tropical bees watery cell contents harbor yeasts that contribute to the larval diet; Roberts, 1971). Conversely, the pollen ball must be liquid

for six females are arrowed; other activities are keyed. Bees 1–4 finished cell construction and the lining of the cell with cut pieces of leaf in the early morning and could then gather nectar (until about 09:00) followed by pollen (09:00–11:30) before starting another cell. Bee 6 was digging most of the day, presumably working on an entirely new nest shaft. However, bee 5 took most of the morning to build a cell, thus having to gather food resources in the afternoon when pollen and nectar supplies were much lower and required longer trips; this effect is shown below, where the numbers (B) and durations (C) of trips for all the bees are compared against time of day. Data from Willmer and Stone (1988).

376

P. G. WILLMER AND G. N. STONE

enough to sustain the water balance needs of the resulting larva. Hence some bees make special flower-visiting trips before egg laying and cell closure to collect whichever resource is needed to get this balance right; an example is shown in Fig. 6, for a xeric mason bee (Chalicodoma sicula) in which water balance considerations are particularly crucial (see Section IV.D). Egg laying is the next stage, and represents a critical point in determining the activity patterns of both the current and the future generation. The haplodiploid sex determination system of Hymenoptera allows a mother bee to choose the sex of her egg—a fertilized egg for a daughter, and an unfertilized egg for a son—and thus to control the sex ratio of her offspring. It is commonly better, in terms of reproductive success, for relatively smaller Hymenoptera offspring to be male and larger offspring to be female (e.g., Alcock et al., 1978; Stone, 1994b; Willmer, 1985a,b; and see Section III.B below). Thus female bees can preselect the sex of their next offspring and build and stock a cell of appropriate relative size; and in practice cells commonly fall into large and small size classes, with small cells yielding sons and large cells usually yielding daughters. Deciding the sex of her offspring has implications for the mother bee’s activity patterns, since a smaller son can require as little as half the provisions, and hence foraging time, as a larger daughter (e.g., Alcock, 1999; Neff, 2003; Willmer and Stone, 1988). If there are strong selective advantages to completing a cell within a given time interval [dictated for example by a female’s intrinsic rate of egg production (Rosenheim et al., 1996) or by pressure from parasites], there will be days when ambient conditions dictate that a given female can complete a male cell, but not a female cell; hence occasionally large cells contain males, because favorable weather and/or resource conditions dictated large cell construction but then deteriorated so that the female adopted the fall-back position of laying a male egg in a relatively empty large cell. Decisions about egg sex and thus sex ratios can also be affected by local mate competition (Stubbelefield and Seger, 1990). The success of male offspring is frequency dependent (Charnov, 1982), while that of females may be partly frequency related but more clearly related to absolute density of other competing females, and of resources. For all these reasons the cell size for the switch between male and female offspring can change if the absolute availability of resources changes over time. The sex ratio from cells of a given size in a single nest can change as the season progresses, and overall sex ratios can vary substantially between seasons at any one nesting locality. A clear example of this is provided by Megachile apicalis, a solitary bee with spring and summer generations (Kim and Thorp, 2001). Spring-emerging females are large enough to cope with spring conditions,

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

377

Fig. 6. (A) Chalicodoma bees in the arid coastal habitats of Israel collect mainly nectar and pollen from trips to Lotus flowers in the morning, and sand for construction of the next cell in the afternoon. (B) Rarely, they make trips for nectar only, and the occurrence of such trips for 10 bees is shown, revealing the close link with timings of cell completions as shown by the 10 arrows, fresh nectar being needed to dilute the cell contents to an appropriate mix before laying an egg. From Willmer (1986).

and exploit the rich spring flora to produce large numbers of relatively smaller offspring. These smaller females of the summer generation are capable of foraging effectively under summer conditions, but in their turn produce fewer larger offspring, most of which emerge (and are better suited to

378

P. G. WILLMER AND G. N. STONE

fly in) the following spring. Thus a trade-off of numbers and size of offspring produces daughters best suited to the climatic and floristic environments they will encounter (Kim and Thorp, 2001). However, the interactions of size and sex for any one bee species are complex, and interspecific variation in adult size within each sex can be very marked. With both sex and size being in part ‘‘elective’’ this may be inevitable, since there is probably low heritability in body size, with far greater importance of provision size (volume of cell contents, or specifically of protein content) in determining final offspring size. In a study of 31 bee species Roulston and Cane (2000a) showed that size variability was unusually pronounced, and especially so in cavity nesters, suggesting a physical constraint (from the size of cavity selected) on how much provision could be supplied to each cell. Once the decision as to offspring sex is made and an egg is ready, in many bees there may be a further inspection of the completed cell contents to locate and eject any foreign eggs laid by nest parasites (particularly kleptoparasitic bees, chrysidid wasps, or anthomyiid and bombylid flies; Roubik, 1989). Only then does the female lay her own single egg on top of the pollen ball or on the cell wall. Finally, she (normally) seals the cell with further lining material and often a soil cap, before starting another cell cycle. A few solitary bee species do not seal their cells (e.g., Roberts, 1971), especially in very humid environments. In other species not only is a sealing cap made but an additional small empty cell may be made on top of the filled cell, probably as an additional protection against those parasitoids (such as some torymid wasps: Roubik, 1989) that can lay an egg through a cell cap. In the megachilid bee Chelostoma florisomne this empty cell is constructed in front of more than 60% of all filled cells, which then suffer only 5% parasitism as compared with 29% of cells without a covering empty cell (Munster-Swendsen and Calabuig, 2000). Where kleptoparasitic pressure is high, this kind of antiparasite measure must add significantly to the time spent in the nest, and so reduce the proportion of time spent in foraging activity. Furthermore, in some gregarious ground-nesting bees there can be intraspecific parasitism as an alternative female mating strategy (Field, 1992); a good example is provided by Dieunomia triangulifera where some proportion of females, having well developed oocytes but with pollen in their gut rather than on the scopa, search around nest sites for opportunities to lay in other females’ nests (Wuellner, 1999b). This may be commoner than currently realized, potentially affecting time within the nest for resident females while also leading to a subpopulation of floating females with far less intensive foraging activities. In addition to the activities described above, female bees can spend substantial periods of nonforaging rest within the nest. Poor weather can

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

379

produce nest-bound days for almost any temperate bee, and heavy rain may even preclude digging if soils become very water-logged. However, long-lived tropical bees may spend much longer periods nest-bound; Neotropical Eulaema bees are known to stay in their nest for 15–59 days through the hot wet season, and up to 78 days in the cooler drier season (Santos and Garofalo, 1994). Although they may have very low metabolic demands in these periods, they presumably rely on some consumption of their own stored nectar, so that provisioning for this eventuality must also impact on their overall activity budgets. It is evident that the factors affecting female bee activity change through time, often with a recurring cycle of different behaviors (Stone, 1994b). Knowledge of the structure of a female’s nesting cycle, and of her position in that cycle, are thus both central to interpreting her activity patterns at a given time. B. Diversity of Mating Systems and Male Strategies in Bees Bees show an enormous diversity of mating systems, associated male behaviors, and hence male activity patterns (Roubik, 1989; Thornhill and Alcock, 1983). Because male bees make no investment in their offspring, their activity patterns are largely directed toward maximizing the number of matings they achieve. Male bee activity is largely structured by three general needs: (1) to achieve and maintain thoracic temperatures required for flight activity, at minimal metabolic cost, (2) to collect enough nectar to fuel that activity, and (3) to locate and mate with females. Given these needs, the diversity of male mating tactics in bees is largely structured by the distribution of females in the environment, the ratio of males to sexually receptive females, and the variance in male size (Thornhill and Alcock, 1983). When not engaged in these activities (e.g., at night, or in bad weather) males rest singly or in groups in nest tunnels (Danforth and Neff, 1992; Stone et al., 1995), take refuge in flowers (e.g., Peponapis bees resting in pumpkin flowers; Willis and Kevan, 1995), or sometimes attached to plant stems by their mandibles. A general feature of bee biology is that most females are sexually receptive only for a brief period at the start of their lives (Mayer and Miliczky, 1998; Roubik, 1989; Wcislo, 1992). In most bees the males emerge earlier in the season than females (protandry), so that the sex ratio is initially strongly male-biased, allowing the males to locate sites or resources at which receptive females can subsequently be encountered, and to establish their position in territorial (or other) hierarchies. The major exceptions to this protandrous rule are the Anthidium species described below, and the social bees, in which male and female reproductives are

380

P. G. WILLMER AND G. N. STONE

produced simultaneously (either at the end of the colony cycle in annual species, or in species- and population-characteristic patterns in perennial species). However, in the majority of solitary and social species (and again Anthidium is an exception) receptive females are available for only a brief period and male reproductive fitness can be achieved only within this window. In most situations there is no selective advantage to male longevity (but see Seger, 1983); and, faced with limited resources, selection should favor investment in multiple (perhaps small) short-lived sons rather than fewer (perhaps larger) long-lived sons. This prediction is supported: although males vary across species in their longevity relative to females, females almost always live longer. The male imperative is to show activity patterns that maximize paternity during their short lives. We now consider the three functions of male activity in more detail. The first of these was establishment and maintenance of flight temperatures. As discussed above, endothermic ability is strongly correlated with body size in bees, and in most bees, females are larger than males. Even where males are as endothermic per unit body mass as females, their smaller body size generally results in a greater dependence of activity on ambient conditions. Male activity is thus commonly restricted to warmer parts of the day and/or to warmer microclimates, while females commonly have longer daily activity patterns and are active over a wider range of ambient conditions (Willmer, 1991a,b). Male bees therefore tend to spend more of their time basking than females, extending their thermal window and allowing greater overlap in time and space with female activity. Furthermore, basking maintains preparation for flight without metabolic cost. The significance of basking for males can be illustrated with data from Anthophora plumipes. A basking male in British spring sun at an ambient  temperature of 9 C receives up to 0.4 W of thermal energy per gram of bee (Stone, 1993a), which without any endothermic expenditure will raise  the male’s thoracic temperature from 9 to 20 C (Stone et al., 1995). At the same ambient temperature, males of this species can generate only 0.56 Wg1 body mass endothermically, and would warm up only very slowly. Basking thus boosts the thermal energy available for warm-up by  70%. Male A. plumipes have a preferred thoracic temperature of 30 C for take-off under these conditions and basking reduces the temperature   increment needed from 21 C to only 10 C, Regardless of the mating system, then, in environments with high diurnal fluctuations in ambient temperature most males bask extensively in morning and evening. Thermal consequences of body size also have major implications for variation in male activity among and within species. In temperate climates, males of small-bodied species (50 mg or less for many Andrena, Colletes, Halictus, Megachile, and Osmia) are active in the spring, but only in the

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

381

warmest part of the day. Males of larger bodied spring species (such as Anthophora plumipes, 130–180 mg) show much more extensive periods of activity (Fig. 7), although females (with a body mass commonly between 180 and 220 mg) are active at still lower temperatures in morning and evening (Stone, 1993a, 1994a; Stone et al., 1995). The impact of male body size on flight ability can also be significant within species; where there is substantial intraspecific variation in male body size, there will be conditions under which larger males are able to fly, and so obtain matings, while smaller males cannot (Stone et al., 1995). The second critical need for males was nectar foraging. In contrast to females, they forage for nectar only to meet their own immediate needs, but their ability to meet the costs of endothermy is still dependent on the availability of nectar fuel. However, nectar is expensive to transport in flight, and males with easy access to flowers tend to carry only enough to meet short-term requirements. Thus while female Anthophora plumipes commonly carry up to 30 l of nectar (Stone, 1994a), males pursuing them at flowers carry only 1–5 l (Stone et al., 1995). Male foraging tends to be

Fig. 7. Numbers of male (solid symbols; squares indicate patrollers and diamonds represent foragers) and female (open squares) Anthophora plumipes active on comfrey (Symphytum) flowers through a day in the southern United Kingdom. Crosses show variation in ambient air temperature. The females are larger bodied and bimodal, with a midday lull as temperatures  rise above about 15 C, but the smaller males are unimodal and show a switch from foraging earlier in the day to patrolling at the flower clumps from about midday onward. See also Fig. 10. Data from Stone et al. (1995).

382

P. G. WILLMER AND G. N. STONE

concentrated at the beginning of the day, and is usually the first flight activity once thoracic temperatures for flight have been achieved. Once a male carries an adequate reserve to meet anticipated costs, all subsequent activity is centered on achieving matings. This transition is clear for male Anthophora plumipes in Fig. 7. The third and perhaps most crucial issue is that of variation in male mating strategies and its effects on activity patterns. Bee reproductive strategies have been reviewed in depth elsewhere (Alcock et al., 1978; Eickwort and Ginsberg, 1980; Roubik, 1989; Thornhill and Alcock, 1983), and we provide only a brief outline here. As in many animal taxa, males seek multiple matings throughout adult life while females commonly prefer to avoid reproductive activities after they have been mated once. Within this framework, individual species can be broadly defined as showing one of a number of different reproductive strategies. 1. In most cases, male mating tactics involve flight and interception of females active outside their nests, but we can deal first with unusual exceptions. In Perdita portalis males are dimorphic: large-headed and completely flightless males fight and mate within the nest, and very little is known about their activity patterns, while smaller headed flying males compete for access to females on flowers (Danforth, 1991; Danforth and Neff, 1992; Danforth et al., 2003b). A similar male dimorphism occurs in Lasioglossum hemichalceum, where it has been linked with larval nutrition (Kukuk, 1996). In addition, there is a growing body of evidence from molecular markers that even winged males of some species achieve part of their reproductive success by mating with their sisters prior to emergence from their natal nest; for example, Andrena agilissima copulates repeatedly at flowers but in fact 97% of freshly emerging females already contain spermatozoa (Paxton et al., 1999) and 70% of A. jacobi have similarly been shown to have mated intranidally (Paxton and Tengo¨, 1996). From such work it is becoming clear that estimates of male tactics and total reproductive success based only on activity observed outside the bee’s nest must be treated with some caution. 2. Where female nest sites and the resources the female bees collect are widely dispersed in the environment, the most common strategy seen in males is scramble competition polygyny. Males do not compete with each other, or exclude each other from any resource visited by the females, but instead attempt to be the first to intercept a receptive female by searching around suitable habitats. This is exemplified by Osmia rufa (Seidelmann, 1999) and Andrena rudbeckiae (Neff and Simpson, 1997), in which males patrol rather freely around nest

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

383

sites and flowers. A similar strategy can occur where there are high bee densities but where individual males are unable to guard, or otherwise control access to, individual females. Here an example is provided by the genus Diadasia, where males scramble for access to receptive females emerging from dense aggregations of nests (Thornhill and Alcock, 1983), with no hierarchy among the males. Creightonella frontalis males also patrol up and down the ditches in whose walls the females nest, and show the effects of ambient temperatures on male activity clearly, with basking time decreasing and the time taken to complete a patrol circuit decreasing due to higher flight speeds as air temperature rises (Fig. 8). 3. In other situations where females are concentrated in space, mating opportunities are limited and competition intense, and then males may show female defense polygyny, which can occur in two forms. First, individual males may compete to dominate access to specific nesting sites (effectively establishing a ‘‘harem’’ of females therein); this is rare, but is known for some bumblebees and carpenter bees (Thornhill and Alcock, 1983). For example, in Xylocopa californica (Alcock, 1991) territorial males hover at specific positions within the nest site for up to 8 h a day, excluding other males and pursuing females both entering and leaving (although preferring those without pollen, which are more likely to be unmated). The arena for such competition can be very small: just a small patch of earth with one or a few nest holes. The second and more common form of female defense polygyny involves competition for access to specific females; males hover or patrol close to the soil surface, centering on a particular landmark such as a stone, while awaiting the emergence of each receptive virgin female. Males rarely interact aggressively with each other in the absence of a female, but do often try to dig out emerging females from their nests (detecting the mechanical vibrations and/or scent cues), and may then scrabble together in a frenzied mating ball as a female appears. Here male success is hierarchical and positively correlated with body size (Chappell, 1984). 4. An alternative competitive male mating system is resource defense polygyny. Here, females are concentrated in space as they visit localized resources (almost invariably patches of favored flowers). Dominant males defend territories containing the resource, and they thereby benefit from enhanced opportunities to mate with visiting females; but they must pay the costs of maintaining territory quality (so that the activity patterns of territory-holding males are strongly structured by the need to maintain adequate nectar standing crops to

384

P. G. WILLMER AND G. N. STONE

Fig. 8. The mean duration of basking sessions (A) and of patrol tours around the female nesting site (B), for male Creightonella frontalis in Papua, New Guinea; at higher ambient temperatures the need for basking is much reduced, and patrol circuits are shorter as flight speed increases, so allowing more frequent returns to the female nest openings. Data from Stone (1989); values of n and error bars representing 1 SEM are shown.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

385

attract females), and the costs of aggressively defending their territories against all comers. In essence, the better the floral resources in the territory, the more matings the holder obtains (Thornhill and Alcock, 1983; Villalobos and Shelley, 1991). Females may have to mate multiply with territory holders, but in return they get freedom to harvest high-quality guarded resources. The area controlled by each male may be one tree, a few bushes, or several discrete patches of herbaceous plants, rarely more than a few square meters in area (Alcock et al., 1978), and a particular patch is often held by the same bee for several days on end. Many bees use this kind of strategy, with males and females meeting at particular flowers in relatively distinct temporal windows when the floral resources are peaking. But in some examples of this mating system, patterns of temporal overlap between the sexes are very different; a classic case is the megachilid solitary genus Anthidium (Alcock et al., 1978; Jaycox, 1967; Severinghaus et al., 1981; Starks and Reeve, 1999; Wirtz et al., 1992), with unusually large males/smaller females and a female-biased sex ratio. Since their floral resources secrete nectar in the morning, territorial Anthidium males must be on hand to exclude visitation by other insects even though their own smaller females will not be able to visit the patch until later. The males use brute force and the pointed ‘‘carder’’ at the tip of their abdomen to intimidate (and sometimes kill) other insects (Nachtigall, 1997; Wirtz et al., 1988), and male dominance is strongly positively correlated with individual body size (Starks and Reeve, 1999; Sugiura, 1991, 1994). The large male size must also help in the elevation and maintenance of thoracic temperatures for flight early in the day; as shown in Fig. 9, territorial male Anthidium manicatum show prolonged daily activity patterns, while female activity is concentrated in the middle of the day. Early in the day, males spend a high proportion of their time basking, rising only to intercept foragers intruding on their patch; later in the day, they make increasing numbers of patrolling flights around their patch, and to neighboring resource patches. One other case of flower-based territoriality with different male and female activity periods occurs in Anthophora pauperata in the arid wadis of Sinai, where territorial males have specific pathways around their flower patches throughout most of the period of insolation (Willmer et al., 1994), with occasional inspection flights to adjacent nesting sites, from which females forage for just a few hours in the morning and midafternoon each day. Here size differences between the sexes are more limited, probably reflecting the warmer climate and high insolation levels.

386

P. G. WILLMER AND G. N. STONE

Fig. 9. (A) Activity patterns for male and female Anthidium manicatum observed foraging from lamb’s ears, Stachys byzantium, and (B) shade air temperature for the same site at Oxford University Botanical Gardens on July 13, 1987. Data from Stone (1989).

5. Yet another male bee strategy is to patrol at a site unrelated to defense of any resource. In some cases, males hover on ridges or hilltops (hence the term ‘‘hilltopping’’), or on kopjes in Africa, where they are highly visible. This may be accompanied by release of pheromones, as in the carpenter bee Xylocopa varipuncta, which waits on ridge tops in afternoons at times when females are most likely to be present (Alcock, 1996). In other cases, such as tropical South American Eulaema meriana and Euglossa imperialis, males occupy less exposed display sites on the bark of trees (Stern, 1991). The mating strategy of drone honeybees would also fall within this

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

387

category. These various kinds of male display sites can be regarded as leks, which females visit to select high-quality mates, and they may also represent sites from which males are best able to intercept passing females (Thornhill and Alcock, 1983). However, there is usually no size dependency of territory size, or of copulatory success. In most of the competitive reproductive strategies (Sections III.B.2–4) discussed above, individual males can also show alternative male tactics (see Austad, 1984; Brockmann, 2001; Hutchings and Myers, 1994), often based on relative male size but occasionally linked with male age; such tactics affect both male and female activity patterns. While most males are holding territories, others (variously referred to as ‘‘patrollers,’’ ‘‘sneakers,’’ ‘‘satellites,’’ or ‘‘transients’’) roam more widely, often skirting the edges of resident bees’ defended areas, and occasionally trying to usurp territory holders. These are usually smaller males (e.g., Larsson, 1991) with differing thermal problems from their larger conspecifics. For example, in the Anthidium bees small males cannot defend territories against larger males, and often show an alternative nonterritorial strategy by intercepting females on their way to or from the guarded resources (Villalobos and Shelley, 1991). In Centris pallida (reviewed by Thornhill and Alcock, 1983), larger individual males patrol nest sites and compete directly for opportunities to excavate and mate with emerging females, but the smallest males patrol the edge of the nest aggregation, while yet others (of small and intermediate sizes) hover at specific sites from which they intercept passing females. We should stress finally that the various male mating strategies described here are not fixed, or mutually exclusive, even within an individual bee; tactics can change with circumstance. Population density is a common influence, because the cost of territory defense is in part determined by density of competing males. When density is very low males may just wander around the mating areas and rarely interact with each other; but when it is very high, the effort required to exclude other males, and so provide the necessary payoff in matings, may be prohibitive, and territoriality may be abandoned in favor of very local competition for specific mating opportunities. The transition can be marked, as shown by Anthophora plumipes (Stone et al., 1995): at the highest male densities, males compete for position by flying in a ‘‘queue’’ behind a receptive female as she feeds on flowers, and only the foremost male will attempt to copulate (Stone et al., 1995). In the Sinai populations of Anthophora pauperata the upper end of this spectrum is absent because density is never very high; but again male territoriality is conspicuously present where males are at moderate densities, and is abandoned in favor of random searching in

388

P. G. WILLMER AND G. N. STONE

times and places of low male competition, so that male patrol frequency depends on bee density in this habitat (Fig. 10). Other possible causes of tactical switching in male bees are variations in the distribution of resources, or changes in the age and sex structure of a population (Leys, 2000; Stone et al., 1995). This is often manifested as a seasonal change: for example, Andrena rudbeckiae males switch between nest site patrolling, floral resource patrolling, and stationary waiting at nonflowering shrubs (Neff and Simpson, 1997); and Xylocopa males may switch from hilltopping strategies to female defense strategies later in the season (Alcock,

Fig. 10. The increase in Anthophora pauperata male patrol frequencies at sites of differing bee density at four sites in Sinai. Bees are rare in the plains, but occur at increasing density in the wadis that radiate from the plains, Wadi Dir, Wadi Arbaein, and Wadi Tofaha having progressively denser populations (see Willmer et al., 1994). Males switch from random searching to highly structured resource defence territoriality around Alkanna plants in the latter sites. Error bars represent 1 SEM.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

389

1993; Leys, 2000). Aging and wing wear are therefore often associated with changes in male mating tactics; in Eulaema younger males with less wing wear hold territories in sites with good diffuse sunlight, while more worn (older) males, regardless of body size, patrol the margins of territories (Stern, 1991), and likewise in Hoplitis younger bees hold better territories and can usually displace older bees from good sites (Eickwort, 1977). Indeed, as reviewed for other taxa by Brockmann (2001), it is very likely that male bees’ strategies vary—with time of day, time of season, location, and population density—far more than has yet been recorded, and observation of specific male behaviors in specific populations is always needed when interpreting the recorded activity patterns. C. From Paternal Investment to Sexual Harassment: The Impact of Males on Female Activity Patterns Thus far we have emphasized the diversity in male behaviors, but male activity can also have direct impacts on female patterning. For most bees, a consequence of the short window of opportunity for paternity and the male-biased sex ratios commonly recorded at mating sites is that males are rarely choosy. Males of most species will attempt to copulate with unreceptive as well as receptive females (and may even pursue males or females of other species that vaguely resemble their own females). Once females are mated, repeated male attentions on or around flowers may severely inhibit the gathering of floral resources, with reductions in the frequency, duration, or effectiveness of female flower visits. Where females are the larger and stronger sex, they are usually able to evade male attentions and forage reasonably normally. In extreme cases, large female Perdita and Nomadopsis can carry a male around in copula, while continuing to forage. Alternatively, some large female bees initiate rapid but brief zig–zag flights when pursued, perhaps a genuine attempt at evasion, but also a potential mechanism for female mate choice by estimation of male quality (Leys, 2000). Where females are more closely matched with male size, their foraging could potentially be greatly disturbed by male attentions, and other solutions may be needed. In Nomia melanderi receptive females are distinguished by a scent that attracts mates, and although they become nonreceptive after mating they do retain an attractive scent for a few hours; as a result, their activity on their first day of adult life is strongly disrupted by repeated mating attempts from males (Mayer and Miliczky, 1998). In Anthophora plumipes, males can so reduce the flowervisiting rates of nonreceptive, provisioning females (Fig. 11) that the females modify their foraging behavior (Stone, 1995), abandoning more rewarding flowers on the exposed exterior of forage plants and instead

390

P. G. WILLMER AND G. N. STONE

Fig. 11. The effects of male activity on female foraging in Anthophora plumipes. (A) Mean number of male interceptions, and copulations attempts per female, against time, which leads to a midday lull in female foraging (see Fig. 7); (B) attempted copulation rate varies with the

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

391

visiting less rewarding but sheltered flowers within a bush or patch, where they experience lower rates of sexual harassment. Such costs of mating attempts may explain the evolution of specific male-repellent pheromones released by some female bees after mating, as in Andrena nigroaena (Schiestl and Ayasse, 2000). At the other end of this spectrum, male behavior can have a direct positive impact on offspring quality. Where there is resource defense by males, the ability of females to forage from high-value flowers during the middle of the day is a direct result of the male investment in territory quality. Thus high-quality territories not only yield higher numbers of matings for the male, but the abundance of resources he maintains must produce more and/or higher quality offspring (and could potentially reduce the foraging time required) for the females within his territory. A second positive impact may accrue where the usual size dimorphism is reduced or reversed (particularly in Anthidium) and where females mate repeatedly, with no clear transition between receptive and unreceptive states; this may give females scope for assessing male quality and so improving their offspring quality, and for some bees this effect may be enhanced (or at least modified) by sperm competition (e.g., Simmons et al., 2000). D. Special Case: Kleptoparasites Female kleptoparasites are freed from the need to provision their own cells, and instead their fitness results from successful entry of, and oviposition within, nests of provisioning hosts. Where hosts and their ‘‘cuckoos’’ have been studied together, the kleptoparasites are usually less strongly endothermic for a given size, and so must restrict their activities to a narrower thermal window, in warmer parts of the day (Stone, 1989; Stone and Willmer, 1989a; Stone et al., 1988); an example is shown for Anthophora and Melecta in Fig. 12. The briefer periods of kleptoparasite activity must be assumed adequate to successfully attack sufficient host cells; selection for more endothermic parasites that could invade and take over more host cells in a given time period is presumably offset by the greater need such a parasite would have for flower visiting time to fuel its endothermic metabolism, at the expense of reduced searching time at nest sites.

proportion of flowers on the exterior of the plant visited by females; (C) females do progressively retreat to the inner flowers (even though these have lower resources) later in the day when males are present. Within the bushy plants the females are relatively free from sexual harassment by male copulation attempts. Data from Stone (1995); values of n and error bars representing 1 SEM are shown.

392

P. G. WILLMER AND G. N. STONE

Fig. 12. Activity patterns for the solitary bee Anthophora plumipes and its kleptoparasite Melecta albifrons, together with ambient shade air temperature at a nest aggregation in Oxford, United Kingdom on April 22, 1987. Data from Stone (1989).

Since female kleptoparasites have relatively narrow thermal windows and no need for high-quality floral resources, males are similarly restricted to short daily active periods, centered on nesting sites, and with little evidence of complex mating strategies.

IV. Extrinsic Factors Structuring Bee Activity A. Variation in Abiotic Conditions The crucial effects of ambient temperature and body temperature on bee activities have already been reviewed in Section II, and the resultant effects on pollinator effectiveness are clear (Cane and Payne, 1993; Corbet, 1990; Corbet et al., 1993; Herrera, 1990; Willmer, 1983). First, and somewhat trivially, there is an obvious annual effect, such that in strongly seasonal habitats adult forager bees do not fly from late autumn until early spring, and even in Mediterranean climates they are rare in winter; for example, Dafni (1996) cites just 2–4 species of solitary bee active in the Israeli winter compared with 100–200 species in spring. Second, even within the seasonal (spring to autumn) active periods, daily fluctuations in ambient temperature are characteristic of most terrestrial habitats; they can be marked at

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

393

any latitude, but are at their most extreme in low-latitude deserts (where clear skies at night allow the air temperature to plummet), in high-latitude summers, and at altitude in some tropical mountain ranges (Willmer et al., 2000). Inevitably, then, bees in common with most other animals show a fundamental diel structure to their activity patterns; their limited ability to warm up endothermically does not free them from the constraints of the weather, and this is especially true in the three kinds of habitat already mentioned. Of more interest here are the differing patterns of activity imposed within a day by climatic variation. Many studies have recorded weatherdependent activity, especially in crops where bee pollination is critical. For example, Abrol (1998) showed a dependence on solar radiation and (negatively) on humidity for the small bee Megachile nana on alfalfa crops; while Vicens and Bosch (2000) linked activities of apple pollinators to weather, with Apis mellifera dependent on temperature, radiation, and windspeed and Osmia cornuta dependent only on radiation and windspeed, and tolerating cooler temperatures and light rain [for similar reasons Batra (1984) recommended Anthophora species as apple pollinators]. Numbers of Bombus species foraging on foxgloves showed positive correlations with temperature, and more and longer flights were made downwind when windspeeds were higher (Comba, 1999). How far some of these climatic correlations are being directly determined by the weather, or indirectly by changes in reward, is of course always open to question. Where there are particular marked daily fluctuations in temperature, for reasons outlined in Section II there is selection for larger bees, from the more strongly endothermic taxa, to be active at the beginning and end of a day, with smaller, less endothermic taxa represented in the midday hours. In general the former tend to show U-shaped (bimodal) activity patterns, and the latter display a unimodal peak (see examples in Table II). Bumblebees are usually active early and late, honeybees more in the middle of the day; both Bombus and some Osmia are routinely active 1–2 h before Apis (e.g., Torchio, 1991; Willmer, 1983). This also results in daily shifts in the taxa, and in the body masses, of active species recorded on flowers (Willmer, 1983), a pattern repeated when records are taken from quite different kinds of flower as shown in Fig. 13. Again these effects are relevant to crop pollination: native Amegilla have activity patterns better suited to cardamom pollination than do introduced honeybees (Stone and Willmer, 1989c), and Creightonella are likewise temporally better matched with coffee flowering than are Apis (Willmer and Stone, 1988). These are all temperate and tropical examples; it should be noted that in the cooler alpine and high-latitude zones where midday temperatures are never particularly high all bees may be relatively large and all tend toward

TABLE II Examples of Bee Activity Patternsa Plant species

Bee species

Ref.

Unimodal Patterns I. Mainly resource limited A. Matinal bees, oligolectic. Flowers open before dawn, pollen collected by specialized bees and usually exhausted within 1–2 h of dawn Camissonia (Onagraceae) Cucurbita (Cucurbitaceae)

394

Ipomoea (Convolvulaceae) Datura (Solanaceae) Solanum (Solanaceae) Larrea (Zygophyllaceae) B. Crepuscular and nocturnal bees, oligolectic. Flowers open at dusk, visited by specialist bees Camissonia (Onagraceae) Capparis (Capparaceae) Oenothera (Onagraceae) Datura (Solanaceae) C. Morning or afternoon bees, oligo- or polylectic, peaking with discrete dehiscence pattern and floral reward availability Ranunculus (Ranunculaceae)

Andrena, Anthophora, Exylaeus, Synhalonia Agapostemon, Peponapis, Xenoglossa Melitoma Agapostemon, Caupolicana, Xylocopa Anthophora, Svastra, Ptiloglossa, Xylocopa Ptiloglossa

Linsley (1978); Linsley et al. (1963) Linsley (1960); Willis and Kevan (1995) Linsley (1960) Linsley (1978) Linsley (1962, 1978); Shelly and Villalobos (2000) Hurd and Linsley (1975); Linsley and Cazier (1970)

Andrena, Anthophora, Exylaeus, Perdita, Sphecodogastra, Synhalonia Proxylocopa Agapostemon, Anthophora, Caupolicana, Evylaeus, Synhalonia Agapostemon, Caupolicana

Linsley (1978); Linsley et al. (1963) Dafni et al. (1987) Linsley (1978); Linsley et al. (1963) Linsley (1978)

Andrena

Linsley and MacSwain (1959)

Acacia (Fabaceae) Mentzelia (Loasaceae) II. Mainly temperature limited A. Midday bees, small bodied. Floral rewards relatively unrestricted temporally Clarkia (Onagraceae) Various Heracleum (Umbelliferae) Lavendula (Labiatae) Larrea (Zygophyllaceae) Capparis (Capparaceae)

Apis, Megachile, Amegilla, Nomia, Braunsapis Perdita

Stone et al. (1996, 1998)

Diadasia, Megachile Perdita

MacSwain et al. (1973) Bennett and Bread (1985); Torchio (1975) Willmer (1983) Herrera (1990) Hurd and Linsley (1975) Dafni et al. (1987)

Andrena, Lasioglossum Anthidiellum, Ceratina Nomia Andrena, Halictus

Michener (2000)

Bimodal Patterns 395

I. Mainly resource limited A. Dawn and dusk bees, often large-bodied. Floral rewards first available at dusk, some remaining next morning Oenothera (Onagraceae)

Andrena, Anthophora, Xylocopa

Barthell and Knops (1997); Linsley and MacSwain (1959); Linsley et al. (1963)

B. Morning and afternoon bees, longer tongued, and requiring nectar that is not too concentrated Hybanthus (Violaceae)

Melipona

Roubik and Buchmann (1984)

II. Mainly temperature limited A. Dawn and dusk bees, often large-bodied. Floral rewards peaking around dawn or relatively unrestricted temporally Larrea (Zygophyllaceae)

Caupolicana, Martinapis, Ptiloglossa

Hurd and Linsley (1975); Linsley and Cazier (1970) Willmer (1983) Fussell (1992)

Tilia (Tiliaceae) Trifolium (Fabaceae)

Bombus Bombus

(Continued)

TABLE II (Continued) Plant species B. Morning and afternoon bees. Rewards peak at various times, bee activity peaks dependent on size Alkanna (Boraginaceae)

396

Calotropis (Asclepiadaceae) Tilia (Tiliaceae) Heracleum (Umbelliferae) Capparis (Capparaceae) Trifolium (Fabaceae) Lavendula (Labiatae) Pulmonaria (Boraginaceae) a

Bee species

Anthophora Xylocopa Apis, Andrena Apis Apis Apis Anthophora Anthophora

Ref.

Stone et al. (1998); Willmer et al. (1994) Willmer (1988) Willmer (1983) Willmer (1983) Dafni et al. (1987) Fussell (1992) Herrera (1990) Stone (1989)

In each case the main influence is identified, for foraging females on a day of typical weather; but resource and temperature will usually interact for most bees, and patterns will vary when ambient temperatures are lower or higher than normal and may vary for individuals of different sizes (see text).

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

Fig. 13.

397

398

P. G. WILLMER AND G. N. STONE

unimodality (e.g., Lundberg, 1980); indeed Malo and Baonza (2002) note an altitudinal gradient of increasing pollinator size up mountains, linked to thermal constraints and correlating with an increase in floral size. These varied temporal patterns of activity, and hence of flower visiting, can also occur within species (Stone, 1994b) where there is significant size variation. Intraspecific patterns may have significant implications for the reproductive success of either sex. For example, larger males are able to fly at lower temperatures and may secure more matings in cool conditions (Larsson and Tengo¨, 1989; Stone et al., 1995), while smaller individuals are able to sustain activity and thus mate more successfully at higher ambient temperatures (Stone, 1993b; Stone et al., 1995). Intraspecific variation in foraging success also occurs: larger females are able to fly at lower temperatures (Fig. 14A) and thus complete more provisioning trips in poor weather, and so to stock more cells in marginal conditions (a feature also shown in sphecid wasps; Willmer, 1985a,b). As ambient temperature rises individual bees are more likely to return with larger pollen loads (Fig. 14B and C) (Stone, 1994b); but since the Tb of a bee affects its power output, larger bees with a higher temperature excess can still carry proportionately larger payloads, and they therefore require proportionately fewer trips to fill a cell even in warm weather. Strong daily effects of ambient temperature are particularly to be expected in small bee species, where the presence of substantial physiological thermoregulatory ability is unlikely (see Section II). For such bees, activity may be primarily limited to periods of the day when levels of solar radiation and ambient temperature allow behavioral thermoregulation alone (basking, or alternation of basking and shade-seeking) to achieve the necessary thoracic temperatures for flight. Many small or very small bees (such as Andrena, Perdita, Lasioglossum, Halictus) characteristically show unimodal activity patterns centered on the warmer parts of the diurnal cycle; specific examples are given in Table II. Medium-sized euglossines are also commonly unimodal in tropical forests, with males foraging somewhat earlier than females and showing a direct dependence on ambient temperature through the morning hours (Armbruster and McCormick,

Fig. 13. (A) Patterns of bee visits to Tilia (lime) flowers through a day in Norfolk, United Kingdom; Bombus show a bimodal pattern, while Apis and Megachile (the other major bees making visits) are unimodal. (B) Activity patterns against mean visitor body weight for the same Tilia tree and also (C) for Heracleum (hogweed) flowers at the same site, showing the consistently observed pattern of larger bodied insects toward dawn and dusk with smaller insects in the hotter midday hours. The umbellifer Heracleum has small, exposed flowers with low volumes of concentrated or crystalline nectar, while Tilia has cup-shaped flowers with copious, more dilute nectar; all data from Willmer (1983).

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

Fig. 14.

399

400

P. G. WILLMER AND G. N. STONE

1990); essentially these bees (unable to bask in the low light of the forest) are limited by low air temperature in the early morning and limited by overheating in the late afternoon (Armbruster and Berg, 1994). In large and/or highly endothermic species that generate a high thoracic temperature in flight, high ambient temperatures during the middle of the day may lead to intolerable heat loads in flight, and abandoning of the foraging site, even where floral supplies are still high. If these food supplies persist into the evening (either because they are not depleted by other visitors, or because the flowers can refill with nectar) such species may show a bimodal activity pattern. Most of the species recorded as showing bimodal activity patterns are either known to be capable of substantial endothermy (Bombus, Xylocopa, Anthophora) or are good potential candidates for endothermy in which the presence of this phenomenon has yet to be investigated (Ptiloglossa, Caupolicana). Following the same kind of arguments, it is to be expected that within one species, seasonal effects will also come into play. Activity patterns may be unimodal at the beginning of the season when days are colder, but become more bimodal as the season progresses and the thermal window expands. For larger bees there should be a single activity peak in cooler weather but bimodality in hotter weather, and this is neatly displayed by the Indian carpenter bee Xylocopa fenestrata (Sihag, 1993), which forages from February to November and whose activity changes from unimodal to bimodal as the summer advances. Similar effects should occur in less predictable climates on a day-to-day basis, and Anthophora plumipes shows exactly this pattern of unimodal activity on a cool day and a bimodal activity pattern on a hot day in the United Kingdom (Stone, 1994a). A comparable but less extreme change in activity has been described for the large carpenter bee Xylocopa pubescens in Israel (Mordechai et al., 1978; Willmer, 1988). Since bees often have active adult lives of only a few weeks, these patterns may also be paralleled by seasonal shifts in the bee species active, and in hot Mediterranean climates there are well-documented shifts from larger/more endothermic bees restricted to springtime to smaller/less endothermic bees dominating in high summer, both within species (Stone et al., 1995) and across species (Potts et al., 2003a,b; Shmida and Dukas, 1990). However, as these last authors point

Fig. 14. Effects of ambient temperature on foraging behavior in Anthophora plumipes in Oxford, United Kingdom. (A) The mean body size of active foragers decreases in warmer  weather, with the largest females inactive above 15 C. (B) The percentage of females returning with pollen, and (C) the mean load (nectar plus pollen) brought back to the nest, both increase as air temperature rises. Data from Stone (1994b); values of n and error bars representing 1 SEM are shown.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

401

out, the shift may not be purely a thermal effect, but could also arise from seasonal fluctuations in floral resource availability. All these examples of intraspecific change are good evidence for the importance of thermal effects in structuring bee activity patterns, but also emphasize the point that for any given bee we need analysis of visitation levels over a range of climatic conditions and levels of floral abundance. In summary, abiotic conditions can therefore be seen to affect bee activity patterns by (1) defining a broad climatic window within which other factors may become important, or (2) themselves defining a narrow window within a broader potential active period delimited by other factors, such as availability of floral resources. The former pattern is probably most common in environments with defined peaks of floral resource availability during the day (see Section IV.B), while the latter pattern is more to be expected when nectar and/or pollen supplies vary less absolutely and set a very broad resource window. We now consider these aspects of floral resource in more detail. B. Variation in Availability of Floral Rewards Within any limits imposed by ambient abiotic conditions, the timings of floral resource availabilities in a particular flower species have obvious consequences for the bees that rely on them. Several resources are relevant here: .

Pollen is crucial to cell provisioning in almost all bees, and timing of pollen release (dehiscence) therefore has particular significance for females. Some pollen may also be ingested, especially by younger foragers, to provide the protein needed for egg maturation (Michener, 2000). Bees can routinely carry 20–25% of their body weight as a pollen load, and this figure may exceed 35–40% on occasion for larger bees.

.

Nectar is used both in cell provisioning by females, and as an energy source for both males and females; a starved female bee may consume up to 90% of its body weight in nectar at one time, and can fly with almost this same load.

.

A small proportion of plants offer oils as alternative rewards, used by some solitary bees as larval food.

.

A range of bees will scrape up and collect plant resins. These are used specifically by a few solitary bees for nest building (e.g., from Dalechampia; see Armbruster, 1984) but are also gathered more generally by bees, including Apis, for incorporation in the nest, perhaps because they often have antibacterial and antifungal properties (classic

402

P. G. WILLMER AND G. N. STONE

examples are gum Arabic resins from certain Acacia, and myrrh and frankincense from Commiphora species in Africa). Where these are carried on the scopae they preclude simultaneous pollen collection. .

Some orchids produce aromatic scents that are gathered by male euglossine bees, perhaps as pheromone components (Dressler, 1982; Kimsey, 1980) or more specifically (since they are collected and accumulated throughout adult life, and sometimes even stolen from dead males) to signal longevity and suitability as a mate (Eltz et al., 1999; Roubik, 1998).

.

Some stingless Trigona bees will gather fragments of meat from carcasses (necrophagy), probably as larval food (Roubik, 1989). Some tropical apids (Eulaema, Eufriesea, and some Meliponines) collect animal feces, either for use in nest construction, or possibly additionally as a source of nutrients (Roubik, 1989).

.

There are also instances of bees gathering honeydew, usually secreted by homopterans; this is equivalent in many respects to nectar as a fuel, but its production from groups of tree hoppers, aphids, and the like is more structured. Some Trigona tend homopterans in the same way that ants can do (e.g., Almeida-Neto et al., 2003), and can benefit from a highly focussed and also potentially ‘‘manageable’’ resource since homopterans can often be induced to secrete by tactile stimulation.

.

Some cynipid oak galls also secrete sugary secretions that are occasionally collected by Bombus (Larson, 1999).

In most cases, adult bees collecting the more unusual foodstuffs must also at least occasionally gather some nectar for their own fuelling needs (although a few necrophagus bees probably do not gather nectar at all but rely on fuel from fruits or from extrafloral nectaries; Noll, 1997). Normally taking on nectar or pollen loads will entail visits to many tens or even hundreds of flowers within a foraging trip, the number being generally less for pollen than for nectar. Solitary bees may make specific trips for single resources but often take both resources on any one trip; social bees more commonly forage for just one resource at a time. The number of trips made depends in part on the distance between nest and resource, the foraging range for most solitary bees being quite small; Gathmann and Tscharntke (2002) found a maximum range of just 150–600 m for 16 bee species examined in detail, and they also noted that foraging range was positively correlated with size. Both large size and sociality may contribute to the particularly long foraging ranges recorded for some bumblebees [with recaptures up to 1750 m from the nest

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

403

(Walther-Hellwig and Frankl, 2000) achieved with flight speeds of up to 15 ms1 (Osborne et al., 1999)] and for carpenter bees (with occasional recapture at 30 km according to Kapil and Dhaliwal, 1969). In general, bees make relatively few food-gathering trips per day, although this is again in part size related and can range from 1 or 2 in small stingless and halictid bees up to 50 or more in some larger and/or social bees (with some foragers of social species making repeat trips uninterruptedly for the same resource throughout their available daily activity window). Overall, there may be more pollen trips per day than nectar trips, probably largely because pollen is a more substantial component of the cell contents but also perhaps because it is usually more available in usable form all day compared to nectar (see Sections IV.C and IV.D). The pattern of nectar and pollen gathering can also vary intraspecifically; for example, in bumblebees, foraging trip length is inversely size related when gathering nectar but unrelated to body size for pollen gatherers (Goulson et al., 2002), so that the number of possible trips per day will also vary with the size–resource interaction. Although we distinguish here between abiotic and biotic aspects of a bee’s environment, and have thus far treated climatic windows and resource windows as separate issues, it now becomes important to appreciate that the two are tightly linked. There is considerable evidence that the timing of release of pollen, the timing of secretion and replenishment of nectar, and the properties of nectar once it has been secreted and starts to equilibrate with ambient conditions, are strongly influenced by daily fluctuations in abiotic conditions (especially the temperature and humidity of the air constituting the microclimate around the plant); these aspects are dealt with in some detail in Sections IV.C and IV.D. To this of course must be added the additional variation in reward imposed by the frequency of previous visitation to the flower, which will itself be partly determined by the weather. Indeed, visitors not only remove nectar (and pollen) directly, but nectar depletion can in turn directly affect nectar replenishment by the flower, and can indirectly affect equilibration rate (since reduced nectar volume means a smaller droplet with a higher surface area-to-volume ratio). Given these various interacting factors, it is particularly difficult to demonstrate whether the responses of flowers to ambient conditions, in terms of the rewards they offer, represent (1) exploitation of climatic changes simply as cues for the timing of events, or (2) selection acting on plant behaviors whose costs and benefits are sensitive to abiotic conditions. These issues are explored with specific examples in the following three sections.

404

P. G. WILLMER AND G. N. STONE

C. Shapes of Activity Patterns Structured by Resource Availability We discussed earlier the reasons why activity patterns structured by resource availability at a single site can be unimodal or bimodal, and given the discussion above it is no surprise that activity patterns primarily structured by availability of floral resources can take similar forms to those responding more obviously to changes in ambient temperature. Thus, when bimodal patterns are observed it can be tempting to attribute departure of bees from foraging sites as temperatures rise in the later morning to their avoidance of overheating. However, if the flowers visited release nectar (or pollen) as a single pulse early in the morning, the departure of bees may in fact occur due to exhaustion of food supplies, substantially before any microclimatic constraints become important. A resource-related explanation needs to be carefully justified: it is more likely to be true when departure from foraging sites occurs at only moderate ambient temperatures. This is seen in large matinal and vernal bees in deserts, such as Ptiloglossa arizonensis and P. jonesi foraging from creosote bush (Larrea tridentata) in the deserts of the southern United States (Hurd and Linsley, 1975); even though departure from the foraging site does correlate with microclimatic change, this change is not directly responsible for departure. Avoidance of overheating can be excluded as a structuring factor in such cases by estimating the heat load to which bees are exposed as they leave the site; if this load is low, then avoidance of high ambient temperatures is unlikely to be a significant structuring factor (e.g., Stone et al., 1999). Bimodality can also arise from resource limitation when the resource becomes available in the late evening, and evening foraging does not deplete the resource below minimum harvestable levels; bees may then return the following morning and generate a second peak in the activity pattern. Examples of this kind are provided by bees foraging from evening primrose, such as Andrena raveni feeding from Oenothera species (Linsley and MacSwain, 1959), Anthophora neglecta and A. affabilis foraging from Oenothera pallida (Linsley et al., 1963), and Xylocopa tabaniformis foraging from Oenothera elata (Barthell and Knops, 1997). Another such situation arises with Anthophora bees feeding on Alkanna orientalis in Sinai (Stone et al., 1999). In all these cases the floral resource may also be gathered by moths during the night, as certainly occurs for some evening primroses, and this will act as a competitive effect (see Section IV.F) altering the reward status for bees returning in the early morning. A further consideration relevant to resource-based activity patterns is that in many environments nectar and/or pollen are on offer at a range of sources that differ in their availability (and relative attractiveness to bees) over time. Under such conditions, bees may move from one source to

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

405

another through the day, and the activity patterns shown by provisioning females will depend on the site at which they are observed (showing the limitations of site-based data, as discussed in Section I.A). Changes in activity at a nectar or pollen site may correlate with changes in temperature, and it is tempting to conclude that activity levels are thermally dependent; but the possibility that the source visited by the bees may have changed must always be checked. For example, Linsley et al. (1963) described the mass movement of a population of Andrena omninigra clarkiae from their Clarkia pollen source from midmorning through midafternoon to their Brassica nectar source in the evening. A similar shift by an entire population from one forage source to another has also been observed in Anthophora (Stone, 1994a). These examples illustrate the point that the full range of nectar and pollen sources visited by a species must be observed before limitations imposed by climate can be inferred. Even at one site with only one floral species present it is crucial to distinguish the resource being gathered with care and note changes with time. Bees foraging only for nectar will often choose different flowers, or different inflorescences, from those seeking both nectar and pollen (e.g., Gonzalez et al., 1995), because different inflorescences may vary in the proportions of male and female flowers (or, more commonly, male and female phase flowers). The sexual phases are obviously dichotomous in terms of pollen reward, but may also contain different nectar rewards and may occur at different heights on a spike of flowers, often with older female phase flowers basally and younger male phase flowers at the top. Bees may therefore learn to associate spatial position, advertisement features such as color or scent, and reward status. Such differences may be readily visible, and are often advertised by differential flower morphology in each phase. Without close observation, bees may appear to be foraging at the same site for many hours irrespective of climatic changes, when in reality their behavior has changed substantially due to nectar concentration changes or pollen depletion. Finally, we should note that resource availability can be misleading as a structuring agent in another way. Many cases are known where bees will gather both nectar and pollen from one plant but later either alter or reject one of the resources, the floral rewards thus contributing in unexpected ways to overall activity budgets. It is not uncommon for nectar that is gathered to be ‘‘nonideal’’ in concentration for a particular bee. It may be too concentrated, in which case additional water-gathering trips (or trips to flowers with much more dilute nectar) may have to be added to the activity cycle; an example is Chalicodoma foraging on Lotus and Calycotome flowers in Israel (Willmer, 1986). Or it may be too dilute, and in that event some bees will gather it but then spend significant time at the

406

P. G. WILLMER AND G. N. STONE

nest entrance (or elsewhere, perhaps unobserved) regurgitating droplets on the tongue to allow evaporation and concentration, before storing the nectar in the cells; here an example occurs with Xylocopa bees foraging on Passiflora (Corbet and Willmer, 1981). For bees concentrating on a single pollen source, the nectar they can get from the pollen plant may again be nonideal, so requiring a few additional trips to another plant for suitable nectar (e.g., Creightonella bees foraging on Coffea; Willmer and Stone, 1988). Finally, even pollen that is gathered may later be rejected; here again Xylocopa bees working passionflowers serve as an example, the Passiflora pollen on their backs being groomed off onto the ground back at the nest, with pollen gathered instead from a leguminous source (Corbet and Willmer, 1981). All of these examples can lead to unimodal activity at one plant with intermittent activity at another. Many (but not all) of these examples involve introduced crop plants and/or nonnative bees, but the message remains clear: a full analysis of the behaviors at flowers and at nest is always advisable to properly understand bee activity patterns. D. Activity Patterns Structured by Nectar Availability Floral nectar is produced by nectaries, which can be derived from various different floral tissues and which therefore have no fixed shape or size or position in the flower; Pacini et al. (2003) provide a review. Nectaries are usually situated basally, so that a bee visitor must insert its head or whole body into the flower and will reliably contact the anthers in passing and so pick up pollen. Nectar-producing tissue commonly occur as disks of tissue around the ovary, or on septa between the locules of the ovary, and may discharge into a nectar ‘‘spur,’’ an elongate tube (often several centimeters long) from the base of the flower that acts as a nectar storage site. Nectar spurs help to keep the secreted nectar at a stable composition, and are common in moth and hummingbird flowers as well as in flowers visited by long-tongued bees. All nectaries are ultimately supplied by phloem vessels, and secretion of the nectar is an active (ATP-requiring) process, probably linked to ion pumps and subsequent osmosis, although details are still unclear (Nicolson, 1998; Pais and Figueredo, 1994; Pacini et al., 2003; Vesprini et al., 1999). Whatever its location in the flower, nectar is essentially a mixture of sugars and water and provides a highly suitable fuel supply for pollinating visitors, including bees. The sugars may be glucose, fructose, and sucrose in various proportions; the monosaccharides glucose and fructose are not normally found in plant phloem, so must either be made from phloem-derived sucrose within the nectary cells or by sucrose breakdown (inversion) after the nectar has been secreted. Baker and Baker (1983) suggested that the

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

407

ratio of sucrose to glucose plus fructose is adaptive, and that, while small bees prefer a monosaccharide-rich nectar, larger bees (together with hummingbirds and hawkmoths) prefer a sucrose-rich resource; if true, this could potentially act as a structuring agent for bee activities. However, these apparent preferences may be more to do with the degree of exposure of a nectary, since sucrose may remain dominant in deep hidden nectaries (as preferred by larger/longer tongued bees) while glucose and fructose levels may become progressively higher in shallow exposed sites where postsecretion inversion can occur more readily. Nectar may also contain very small amounts of other solutes in addition to sugars, but these are usually unimportant in structuring visitor activities. Lipids, antioxidants, and traces of some deterrents found in plants—alkaloids, phenolics, glycosides, and so on—have been detected, as also have amino acids. Most or all may be just passive and inevitable contaminants, as the plant delivers an extract of its own phloem across a leaky surface. Amino acids may possibly have some adaptive significance for some pollinators (Baker and Baker, 1983) although their significance has been doubted (Willmer, 1980) and there is little evidence that they are important to, or even detected by, visiting bees. In fact, in their relationships to nectar the great majority of bees are generalists and will sample from any flower to which they can gain access. This key issue of access may be largely determined by tongue dimensions in bees. If a bee is seeking nectar from flowers it should preferentially visit species with corolla depths that approximately match with its own tongue length. Short-tongued bees (especially Colletidae, Halictidae, and Andrenidae) can take nectar from open and shallow flowers, or from longer corollas at certain times of day when they are very full (although they sometimes visit illegitimately at long corollas, as discussed in Section IV.F). Long-tongued bees (especially most Megachilidae and Apidae) could theoretically access a wider range of flowers (Harder, 1986), but may operate best with fairly deep corollas (where they can harvest even the last dregs available, so gaining a reward over a much longer daily period than a shorter tongued visitor); on an open flower not only are their tongues unwieldy but also nectar in such flowers often becomes too concentrated to be sucked into a long tubular tongue as explained below. In addition to tongue length, the size and strength of bees affect flower access and thus flower choice, an example being the inability of any bees except large Xylocopa species to force a way into the lidded nectaries of certain Passiflora flowers. Some aspects of innate or learned behavior also of course affect flower access, especially where floral morphology is complex. Once nectar is accessed, the crucial parameters for bees, as for most flower visitors, are the concentration and volume of the nectar, both of

408

P. G. WILLMER AND G. N. STONE

which can be measured rather easily; and together (with a small correction for varying density) these give a direct assessment of the sugar and thus the calories that the foraging bee gains if it takes all of the nectar on offer at the moment of sampling. 1. Nectar Volume Nectar volume per flower is a major determinant of the interaction between a flower and any bee visitor, because of the essentially conflicting selective forces operating on plant and animal. The plant will benefit from keeping the volume as low as possible, minimizing its own costs and ensuring that the bee moves on to more flowers (taking with it the essential pollen that can ensure male success in the next plant generation). But the volume must not be so low that the visitor ‘‘gives up’’ on that floral species. In contrast, the bee will benefit from visiting as few flowers, each with as large a volume, as possible; ideally it would fill up all its crop capacity from one flower, remaining relatively sedentary, and incidentally transferring minimum pollen around the system. The ‘‘correct’’ volume dispensed per flower inevitably varies hugely with pollinator type (e.g., Cruden et al., 1983; Faegri and van der Pijl, 1979): a typically bat-pollinated plant may require several milliliters per flower, while a fly-pollinated generalist plant may dispense almost nonmeasurable amounts, less than 0.05 l. As a fairly crude generalization bee flowers come somewhere in the lower middle range of this spectrum, commonly with a mean reward of perhaps 0.1–10 l, reflecting both the fact that bees are generally somewhat larger and more expensive to fuel (given the demands of endothermy) than are flies, and that bees are also gathering resources for their progeny as well as for their own individual needs. With rewards of this volume, an individual bee may need to visit tens or hundreds of ‘full’ flowers per foraging trip to completely fill its crop [although bees do not always fill up in this manner, as discussed by Schmid-Hempel et al. (1985) and Moffatt (2000)]. This number of flower visits may require a foraging trip ranging from just a few minutes to well over 1 h where nectar volumes are particularly low. 2. Nectar Concentration Nectar concentration is almost invariably measured with refractometers, giving a reading as percent sucrose equivalents (w/w). The common range of concentrations encountered in temperate flowers is 20–50%, rising to 70% in some conditions. Mean concentrations are usually rather less in flowers from the moist tropics, and may be much higher in hot arid conditions. This is a consequence of two main factors, one relating to the visiting animal and the other to the environment itself.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

409

The first issue is that different visitors prefer or indeed require particular concentrations. Specific figures are often quoted as ‘‘means’’: for example, bats have been said to prefer 17%, hummingbirds 21%, hawkmoths 27%, and bees 46%. Such generalizations are certainly a great oversimplification, but do reflect the reality that animals with longer tongues cannot drink more concentrated nectars. Hence long-tongued hummingbirds, bats, and lepidopterans must seek dilute sources, often 15–20%; whereas flies, mostly with quite short lapping probosces rather than sucking tongues, can take nectar concentrations up to 70% (and even beyond, to the point where nectar is effectively crystalline, since they can then regurgitate some saliva into it and lap up the resultant fluid). Bees again lie in the middle of this spectrum, but vary considerably; longer tongued species may prefer 20–40%, and short-tongued ones may prefer 40–60%. For stingless bees (Melipona), Roubik and Buchmann (1984) report foraging on all concentrations from 20 to 75%, with a maximum imbibing rate at about 45% and maximum caloric intake at about 60%. In a later study on a variety of tropical bees, Roubik et al. (1995) recorded optimal nectars as being in the range of 35–65% overall, with means of 38% for euglossines, 44% for meliponines, and 48% for centridines. Only a minority of bees would take nectar at 10–15%, and very few would take it at 65–70%. However, a cautionary note is added by Biesmeijer et al. (1999), who point out that apparently different nectar preferences in closely related stingless bees can in fact arise from thermally determined preferences for sunlit or shady sites in which nectar is inevitably differentially concentrated (see below). Nectar preference effects are mainly due to the balancing of caloric reward and nectar viscosity, which rises exponentially with concentration (so that a 60% solution is 28 times as viscous as a 20% solution). Hence a very concentrated nectar becomes extremely thick and sticky, and cannot travel fast enough (or at all) up the long hollow tongue of many bees to give an adequate rate of gain of food. Each floral visitor is predicted to seek an optimal balance of the rate of taking in the fluid (higher if it is dilute) and the rate of taking in calories (lower if dilute); and this optimum will vary primarily with mouthpart morphology, but also with aspects of behavior (e.g., a bee can make only short individual flower visits if hovering, so may need quicker uptake than another bee that is foraging by crawling over the flowers). The discussion so far implies that flowers merely have to supply nectar at the right concentration for a particular visitor; bee-visited flowers could simply select their nectar concentration through the course of evolution according to the size and tongue length of their main bee visitors. However, this is difficult to achieve in practice because most plants cannot just secrete nectar at the ‘‘right’’ concentration and then leave it to be gathered, due to the second factor influencing nectar in varying habitats, that is the

410

P. G. WILLMER AND G. N. STONE

postsecretory changes that occur in nectar concentration (Bu´rquez and Corbet, 1991; Corbet et al., 1979; Willmer, 1983). Any fluid exposed to dry air will tend to equilibrate with it by losing water, and this will be more marked in hot arid habitats but less evident in the humid tropics. Thus after secretion from the nectary tissues, nectar will almost always tend to dry out and become gradually more concentrated and more viscous. It could be argued that the plant loses control of the situation once the nectar is released into the corolla. However, matters are still not that simple; for example, in temperate habitats nectar is commonly secreted at perhaps 20–40% sugar, but although it does then become more concentrated, especially on warmer and drier days, it does not change as much as expected for a free sugar solution. Furthermore, while nectar in open cup-shaped flowers equilibrates fairly rapidly and becomes viscously concentrated or even crystalline (>75% sugar), rendering it useless to longtongued bees, initially similar nectar in an enclosed elongate corolla (a morphology usually more attractive to bees) may remain relatively dilute throughout the day so that it is readily drawn up by a bee’s tongue. It is therefore appropriate and convenient that this elongate morphology is usually more attractive to bees and is commonly regarded as part of the ‘‘melittophily syndrome’’ (Faegri and van der Pijl, 1979). The explanation of these effects was given by Corbet et al. (1979), who showed that tubular flowers have their own internal microclimate of more humid air, slowing the rate of equilibration with ambient conditions and so helping the flowers control their own nectar. These authors specified a number of floral features that contribute to these effects: elongate corollas, relatively enclosed corollas, hairs within the corolla just above the nectary, constrictions of the corolla wall beyond the nectary, and even lipid layers over the nectar as protective ‘‘waterproofing’’ to delay evaporative loss. Hence a classic tubular ‘‘bee flower’’ such as Echium vulgare can keep its nectar below 50%, suitable for most bee visitors, through most of a temperate summer day, while nectar in an adjacent cup-shaped flower such as Crataegus may reach 60% between 11:00 h and 15:00 h and the nectar on an open umbelliferous flower such as Heracleum evaporates fairly freely and can become crystalline (hence only suitable for flies, beetles, etc.) for much of that same day (see Corbet et al., 1979; Willmer, 1983). Of course on cool, overcast, or wet days these differential effects will be very much reduced. Hence the shape of flowers helps to determine their temporal pattern of nectar concentration, irrespective of initial nectar concentration at the time of secretion, in concert with the climatic regime. It follows that the timing of initial secretion will in turn affect the concentration of the nectar encountered by a visiting bee at different times of day, and to differing extents in different kinds of flower.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

411

We should note here that such wide and differential variations in nectar concentrations also mean that ‘‘one-off’’ measurements of a flower’s nectar concentration cannot be trusted; many flowers must be sampled, from many plants, throughout the diurnal cycle, and in different weather conditions, to give a reasonably accurate picture of what a bee can expect to find in a ‘‘typical’’ flower at any given foraging time. 3. Nectar Sugar Reward It is relatively straightforward to combine nectar volume and concentration measurements (again taking into account temporal and climatic effects) to obtain the mean total sugar amount per flower for a particular species. For temperate plant species favored by bees, this generally gives values in the milligram range. Calculated sugar amounts can be particularly useful for comparisons between available floral resources in a habitat, and for estimating foraging budgets for particular bees. Furthermore, sugar reward is very easily recalculated as caloric reward, so giving a direct handle on the ‘‘currency’’ that a particular visitor might be seeking to optimize, whether by maximizing their net rate of energy gain, or maximizing their energy efficiency (rewards minus costs), or minimizing their risk by selecting the floral resource with lowest variance, all of these having been modeled (mainly for social bees; see Section V.B). This raises an additional problem, that there is very considerable variation in nectar reward between flowers within a plant (e.g., Pleasants and Zimmerman, 1983; Possingham, 1988; Real and Rathke, 1988). In fact, the distribution of nectar volumes is usually left-skewed and approximates to a negative exponential (see Bell, 1986; and references in Gilbert et al., 1991); some flowers are very rewarding and others nearby are almost empty (even when visitation has been precluded, so they have not just been emptied by another bee). It may be a general pattern, particularly where flowers are organized into complex inflorescences, that it pays to have ‘‘blanks’’ and ‘‘bonanzas’’ among the flower population (Brink, 1982; Thakar et al., 2003); this keeps a visitor somewhat uncertain, but prepared to go on trying to find the occasional bonanza. However, there are also alternative explanations relying on different bee responses: Biernaskie et al. (2002) show that high variance in nectar reward can lead to bees making fewer flower visits on a variable inflorescence, so again promoting out-crossing. In either case, spatial variation affects temporal aspects of visitation. A related complication arises in flowers that are protandrous or protogynous and have different nectar offerings in each sexual phase. Female flowers generally have lower nectar production than male flowers (reviewed by Dafni, 1984). But this is not always true in bee-visited flowers; for example, in squash plants (Cucurbita), female flowers have higher

412

P. G. WILLMER AND G. N. STONE

concentration and hence higher overall sugar reward (Nepi et al., 2001), and the same is true (with a 3-fold difference in production) in Alstroemeria flowers (Aizen and Basilio, 1998). In some cases the male and female flowers are in particular spatial relationships; for example, in Delphinium the older female flowers at the base of the flower spike have twice the sugar of the younger male ones with dehiscing anthers at the top (Waser, 1983). Waser suggests this is selected for by the usual pattern of foraging of bees, which tend to start at the bottom of a spike and work upward, so that incoming bees arrive at female flowers first and deposit ‘‘foreign’’ crossfertilizing pollen, and leave from the male flowers carrying new pollen away; thus by offering the ‘‘best’’ nectar flowers early on, the plant ensures that bees tend to keep working upward in hope of finding more bonanza flowers. Again, though, this means a necessary operation of caution in generalizing about floral rewards available to bees from just a few flower measurements. One final issue here is that the nectar, although commonly regarded merely as a sugar reward, can also serve as a ‘‘water reward’’ for some bees. This arises in dry climates due to potential water balance problems for a forager, which may affect activity and flower choices. Many bees do not collect free water (Apis is an exception), but rely on inputs from nectar; in deserts there is unlikely to be standing water anyway. Perhaps the clearest example of water as a constraint comes from work on Chalicodoma, a mason bee, which in Israel visits Lotus flowers. Here the plant must provide nectar that is sufficiently dilute to keep the bee in positive water balance while foraging in the hot dry climate. Measurements of the timing and duration of visits to different flowers, together with assessment of body fluid concentrations before and after foraging trips, showed that these bees were foraging throughout the day in ways that gave them a constant water input per trip from the nectar, with relatively little regard to the sugar intake (Willmer, 1986). On this point, it is also worth remembering that in tropical climates there is much more likely to be a problem for a bee of getting too much water from their flower visits (since they also generate very significant amounts of metabolic water in flight), and so having to unload some of this excess (Willmer and Stone, 1997a). This is especially true for bigger tropical and subtropical bee species, and even for large bumblebees in temperate habitats (Bertsch, 1984); it may provide a nonthermoregulatory explanation for their habits of regurgitating nectar onto the tongue to allow evaporation (tongue lashing), as well as causing the constant dribble of dilute excreta sometimes observed while in flight. To illustrate these conflicting water balance problems, carefully controlled studies by Roberts et al. (1998) showed that for flying Centris pallida there is strongly positive water balance at lower ambient

413

DETERMINANTS OF ACTIVITY PATTERNS OF BEES 

temperatures but that this switches into water deficit above 31 C as evaporative losses persist but metabolic heat production (and therefore metabolic water production) decrease. Thus, for many larger bees in warmer climates, water balance must be assessed in some detail before predicting how it might impact on flower choice and nectar requirements. 4. Nectar Secretion Pattern Temporal patterns of secretion vary widely, both in terms of timing of onset and rate of flow, between species and between individual flowers. Note that a flower could secrete in various different fashions: 1. A relatively fixed volume once only, on offer as the flower opens 2. Continuously during the life of a flower 3. With a fixed diurnal pattern, if a flower lasts more than 1 day 4. In relation to visitation, to keep a standard volume available by regular ‘‘topping up’’ (‘‘nectar homeostasis’’) All of these patterns do occur, and with no immediately obvious phylogenetic or biogeographic patterns. However, there may be some clear adaptive responses related to habitat or principal pollinator; for example, most tropical bat-pollinated flowers secrete at dusk, as do some mothvisited plants, presumably avoiding losses to diurnally active pollinators. Many desert species secrete nectar at dawn and dusk (an ‘‘eocrepuscular’’ pattern), which would allow these plants to invest water into nectar production at times when ambient relative humidity is higher and the plant is less water stressed (Bertsch, 1983; Nicolson, 1993). This release of floral resources at times of high relative humidity and low ambient temperature has probably had long-term implications for the physiology of bees in desert habitats; if nectar is a limiting resource, those bees able to collect it at the low temperatures associated with eocrepuscular secretion would be favored, leading to selection for increased endothermy. This might explain why deserts are home to the most endothermic bees yet studied (e.g., Stone, 1994a; Stone et al., 1999). However, it is also true that many nondesert plants secrete nectar mainly at night, perhaps because this allows accumulation of significant volumes in readiness for dawn and early morning visitors. Thus many temperate bee flowers have unimodal patterns of secretion, with the main production of nectar overnight or just predawn. Even plants that attract both nocturnal and diurnal visitors (such as the common milkweed Asclepias syriaca, visited by both bumblebees and moths) tend to secrete their nectar mostly at night, despite the fact that,

414

P. G. WILLMER AND G. N. STONE

for the milkweed, the nocturnal visitors are much less effective at moving pollinaria between flowers (Morse and Fritz, 1983). It remains unclear how far temperature and humidity are direct cues for the initial secretion of nectar; the observation that most plants show fairly specific diel peaks may indicate greater roles for photoperiodic triggers. However, temperature can affect overall production in different fashions that are assumed to be adaptive; for example, in the typical Mediterranean genus Thymus, nectar volume, concentration, and sugar content all in crease with temperature up to 38 C (as long as there is no water stress), whereas in the Mediterranean but shade-loving Ballota neither light nor temperature has significant effects (Petanidou and Smets, 1996), and in cool temperate flowers such high temperatures often lead to effectively nectarless flowers. Jakobsen and Kristjansson (1994) showed adaptive intraspecific effects for Trifolium, where clones from Iceland had optimum   nectar secretion at 10 C compared with Danish clones peaking at 18 C, the differences being strongly heritable. Such fine-tuning of the timing of secretion peaks by daily climatic variation may be more common than has yet been realized. Whatever their time of peak secretion, many flowers do ‘‘top up’’ homeostatically, so altering the costs to themselves in relation to the rate of visitation they are receiving [shown, e.g., in Penstemon flowers by Castellanos et al. (2002)]. There are also a few demonstrations (Nepi et al., 2001; Nicolson, 1995) that some plants can do the opposite and reabsorb nectar back into their tissues, even from storage sites in nectar spurs (Stpiczynska, 2003), giving these species an ability to regulate the volume and concentration on offer even more precisely. Nectar resorption especially occurs when a plant is unvisited for a substantial period, but it also occurs after pollination in the case of at least one orchid (Luyt and Johnson, 2002). To add even greater complexity, scientific sampling of nectar for experimental assessment may in itself promote subsequent further production (using the homeostatic mechanism); but removal could at least in theory alternatively trigger reduced secretion, in plants where just one or two visits normally effect pollination. Given the evidence just cited for an orchid, it would certainly seem possible that the two main floral rewards can interact, such that pollen removal or deposition could affect nectar production, probably via the various plant signaling systems that mediate postpollination abscission or color change. These kinds of effects have rarely been looked for (e.g., Aizen and Basilio, 1998) but would be appropriate adaptive responses for plants where resources were scarce, and could constitute an important factor in structuring bee visits. For all these reasons, what we can measure from a given flower or a population of flowers at any one time is potentially seriously misleading in

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

415

terms of assessing the natural ‘‘standing crop’’ of nectar. Not surprisingly, there is a considerable literature of inadequate data on the supposed ‘‘typical’’ nectar volumes and concentrations, and diurnal secretion patterns, for particular plant species. Indeed, there is variation in all these parameters even at the level of different parts of the same plant; nectar profiles from the sunlit and shady sides, or the outer flowers and middle flowers, of even a small bush may differ quite considerably (e.g., Stone, 1995). It should be evident that published data cannot be relied on when seeking to assess bee activity patterns in relation to floral nectar reward; each study of bee activities must include its own intensive assessment of the floral rewards available at the appropriate time(s) and place(s). The overall message here is that the plant does have a good measure of control over its nectar and sugar dispensing systems, since (1) nectar secretion is an active process with a potential on–off switch, and (2) postproduction changes can also be controlled (or at least limited) to varying degrees. Hence a plant has reasonable scope to produce exactly the ‘‘right’’ sugar reward to pay for bee visits and still encourage movement of the visitor onward to other flowers. Flowers can thus manipulate bee behaviors quite precisely in space and time, to obtain the best possible pollen movement and potential gene flow. 5. Overview of Nectar Effects Having discussed at some length the availability of nectar from the viewpoint of the plant, the question remains as to how far nectar availability in flowers determines bee activity patterns. Structuring by nectar could occur either before a flower visit (by assessment of what is on offer) or after it (by assessment of what has been received). Can bees assess nectar reward at a distance? If they could, it should affect activity patterns by making foraging more efficient and potentially reducing the ratio of feeding time to trip length. The evidence has been seen as somewhat equivocal (Corbet et al., 1984). However Gilbert et al. (1991) clearly demonstrated that Anthophora bees could distinguish between Cerinthe flowers on the basis of their nectar secretion rate (perhaps using flower age as a cue); and some flowers do signal their age and altered reward status rather clearly, by such cues as color change (e.g., Nuttman and Willmer, 2003). There are also at least two other likely scenarios: first, that where a particular nectar is strongly scented its abundance could be assessed at some distance from the flower, and second, that where petal tissue is reasonably transparent the level of nectar within a corolla tube could be assessed visually as a bee approached. This latter option is certainly sometimes available to the human eye, and Goulson et al. (2001)

416

P. G. WILLMER AND G. N. STONE

indicated that Apis could detect nectar levels in some elongate flowers. However, these same authors point out that such discrimination will not always be useful: where handling time is low, and/or flowers are scarce, it may make economic sense for a bee to visit even the relatively depleted flowers that it encounters. Bees undoubtedly do respond retrospectively to the nectar reward they have received from flowers, and on that basis they make decisions about when to persist (on a flower, a plant, a patch, or a species), and when to leave or switch to another resource. But it should now be evident that the relationships between reward and behavior are far from simple. This would be true for any flower forager, but it is especially so for bees given their particularly close relationship with flowers and their well-known abilities to learn and to modify their behavior after learning. There are two particularly important aspects of this learned behavior. One is the issue of learned preference contributing to floral constancy, and a second issue linked with this is the reduced floral handling time that occurs after learning. Both of these are best understood in social bees and are therefore covered in more detail in Section V.B. We note here, however, that even after learning to handle a complicated enclosed ‘‘bee-adapted’’ floral structure with considerable efficiency, a bee may still end up with a relatively low foraging success compared to what it could earn on an open ‘‘easy’’ flower (in all but the hottest parts of the day when nectar there became too viscous). Again, trade-offs between ease of flower handling and ease of nectar uptake must be made that are not straightforward, and that are further complicated by issues of travel time dependent on distance from nest, floral population density, and floral community structure. From the discussion so far, and recalling once again the interdependency of floral rewards and abiotic conditions, it should be apparent that clear evidence for bee activity depending on nectar can be difficult to find. The link should be most clear-cut where bees either forage at a plant only while it can maintain a rather narrow concentration range, or where it offers a certain minimum volume and/or sugar reward. Often this means looking for the conditions under which bees leave a particular floral resource. As an example of the first type, Melipona bees visit Hybanthus flowers only while nectar is below 60% sugar concentration (Roubik and Buchmann, 1984), which can tend to produce diurnal bimodality with no visitation in warmer midday hours. The second scenario is exemplified by Anthophora and Eucera both showing patterns of departure from Anchusa strigosa that depend on the amount of nectar received from the last two flowers visited (Kadmon and Shmida, 1992). Examples are noted elsewhere in this review where nectar is reported as a key structuring agent, and is certainly a part of the story; but in many of these cases there are potentially

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

417

confounding interactions with other intrinsic and extrinsic factors that remain unexplored. E. Activity Patterns Structured by Pollen Availability The majority of bees are polylectic, taking pollen from several different plants opportunistically. Only rarely are bees narrow pollen specialists, perhaps because most have an active adult life cycle longer than the flowering season of any one plant. However, some desert bees are pollen specialists (Minckley et al., 2000), relying on synchronous emergence with the blooming of rare and irregularly flowering plants, and this may in part explain why bees are at their most diverse in arid zones (Danforth, 1999). Where specialist bees do occur, it is seemingly random; species with narrow oligolecty are found in the same genus and subgenus as broad generalists. The clearest demonstration of this to date comes from the tribe Anthidiini, whose 72 Western Palearctic species were scored as 43% narrowly oligolectic, 18% moderately polylectic, and 35% strongly polylectic, with 4% unknown (Muller, 1996). A few examples of pollen specialists were given in Section I.C, and others are discussed by Michener (2000). While most bees are described as nonspecialist, they may still exercise a good deal of preference. Even honeybees, although highly polylectic, do seem to prefer some pollens over others (Schmidt, 1982), perhaps basing choices on pollen odor or taste, but we have no good evidence on this topic. Williams (2003) found that the apparently specialist forager Osmia californica in fact retained reasonable flexibility to accept novel pollens, and also showed that the specialism did not extend to the larvae, which grew as well on novel and normal pollens; the specialized foraging of the adult is not dictated by any particular feeding or nutritional specialization. Many other bees behave in a narrowly oligolectic fashion in practice, although they tend to show floral constancy on any one day or at least on any one trip when gathering pollen (Westerkamp, 1996). For example, Megachile rotundata visited only 21 species out of 209 offered (Small et al., 1997), those chosen being scattered across 14 genera and 7 families, but sharing the key feature of a particular corolla tube length. Many, perhaps most, bees have this kind of ‘‘polylecty,’’ which is in part imposed by nectar accessibility rather than by pollen, given that it is quicker to collect both resources from the same flower all other things being equal. Pollen quantity and quality have been extensively reviewed (e.g., Bernhardt, 1996; Cruden, 2000; Roulston and Cane, 2000b; Stanley and Linskens, 1974), often in an attempt to understand bee (and other pollinator) preferences. Pollen grains vary from 4 to 350 m in diameter, but most are in the range 15–60 m, readily picked up on the feathery hairs of bee

418

P. G. WILLMER AND G. N. STONE

bodies; and grain size tends to have a strong phylogenetic component. Pollen is rich in protein (usually 16–30% dry weight, and up to 60% in some cases) with 3–10% lipid, 0–15% sugar, and 1–7% starch; hence it provides a good balanced diet. However, the outer exine layer of a pollen grain is totally indigestible, and this may be more than half the grain volume, inevitably making up an even higher proportion of smaller grains. Roulston and Cane (2000b), in common with earlier authors, conclude there is no good evidence for adaptive relations between pollen food value and type of pollinator, and equally there is little or no evidence that bees assess pollen nutritional quality except perhaps in the crude sense of grain size. Pollen ‘‘longevity’’ could perhaps be an important aspect of quality of importance to bees; pollen viability declines continuously after dehiscence (Dafni and Firmage, 2000; Thomson and Thomson, 1992), accompanied by drying processes that sucrose-rich pollens survive better than very starchy pollen. Thus if bees detect when pollen becomes very desiccated and therefore perhaps less valuable to them, sucrose-rich pollens may have some benefits. Nevertheless it is primarily pollen quantity rather than quality that matters to a bee. Quantitative pollen availability within a plant, from a bee’s point of view, is compounded from several factors: 1. Patterns of anther dehiscence 2. Patterns of flower opening 3. Patterns of depletion, by visitors or by physical dislodgement Anthers (varying from one to several hundred per flower) are normally spherical, ovate, or elongate, and commonly contain four separate pollenfilled locules that undergo dehiscence. Within each locule the pollen grains are bathed in nutritive locular fluid, and as the anther matures the locular fluid reduces either by evaporation or resorption. It is the changes in this fluid that trigger anther opening, which can be simultaneous in one flower even to the extent of being ‘‘explosive,’’ simultaneous over a whole plant or a whole population of plants (Stone et al., 1998), or sequential between flowers or within one flower (e.g., in hellebore flowers the anthers ripen one at a time). In some plants (e.g., the rock rose Cistus, a major bee plant in Mediterranean habitats) each anther can open gradually along its length, so that only some of its pollen is exposed to a visitor at any one time. Once the anther is open, pollen may be passively released directly into the air, may be launched into the air by movements of the filament, or (most commonly in bee-visited plants) may be held on or in the anther chamber by its own sticky surface and/or by viscous threads. But it is worth mentioning that

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

419

pollen exposure by anther opening is not always an irreversible event, since some plants can close their anthers. Bees normally collect pollen both passively (grains stick to their branched hairs, and are groomed off later to the storage site), and actively with their legs and mouthparts or by rubbing the ventral body surface across a flower. Some bees moisten the pollen with nectar, others carry it dry. Once gathered it may be carried in the crop (for some very young adults, and a few very primitive bees), on the general body surface, or in a specific scopa on legs, thorax, or ventral abdomen. For some females there are substantial interruptions to foraging while the bee finds a resting site where pollen is packaged into the scopa; but many more advanced bees require only a brief hover between flower probings to achieve grooming and packaging, having little impact on the overall activity pattern. The special case of pollen collection by sonication (‘‘buzz pollination’’) is also relevant here, as it applies solely to bees and can greatly increase their foraging efficiency. It occurs where anthers have a pore at the tip or side (porose or poricidal anthers) rather than splitting lengthwise, and is characteristic of many (often nectarless) plants of the borage family, of Solanum and Dodecatheon, and of some poppies. Here a visiting bee grasps the anther with legs or mouthparts and vibrates her body, using the indirect wing muscles in a partially uncoupled state, and giving an audible buzz. Pollen in these anthers is usually 40–50 m in diameter, and generally very dry and powdery rather than sticky, emerging from the pore in puffs so that it is dusted over the bee’s body. It is possible that electrostatic forces between negatively charged pollen and the positively charged surface of a bee are involved here (reviewed by Vaknin et al., 2000). Buzz pollination is highly suited to bees since they have well-developed abilities to vibrate their thoracic muscles without flight for thermal reasons already considered. However, it is not available to all bees. King and Buchmann (2003) confirm that thoracic vibration, occurring at a higher frequency than that used in flight (and often the second harmonic of the flight frequency), dislodges pollen from poricidal anthers for Bombus and Xylocopa; but that these higher frequencies do not occur for Apis, which therefore does not dislodge pollen and cannot buzz pollinate flowers. How far sonication extends to other (solitary) bees remains unclear, although it may be reasonably common since examples are known for euglossines and for some megachilids (Neff and Simpson, 1988), anthophorines (Batra, 1994; Stone et al., 1999), and halictids (Shelly and Villalobos, 2000). We have repeatedly stressed that bee activities are structured by both biotic factors and by abiotic climatic variables, these being strongly interactive. Inevitably the same is true when considering pollen as a cue; that is, anther dehiscence and pollen display depend in part on the weather. In

420

P. G. WILLMER AND G. N. STONE

many plant species the daily cycles of declining relative humidity and rising ambient temperature will induce the structural changes in plant tissue that lead to dehiscence (Keijzer, 1999; Meakin and Roberts, 1990; Pacini, 1994, 2000). The effects on timing can be striking: for example, in the long-lived woodland plant dog’s mercury (Mercurialis annua) anthers may open around noon in February but as early as 07:00 in July when temperature is high and humidity low (Lisci et al., 1994). Similarly in hellebores, with anthers opening one after another, the number opening on any one day is temperature dependent (Vesprini and Pacini, 2000). In some cases, such as the structuring of dehiscence in Acacia communities (Stone et al., 1998), daily changes in relative humidity have been shown to act as a specific cue, with anther opening delayed on more humid mornings but to different extents in different species. Where the more drastic effect of anther closure occurs in certain plants, it is again weather dependent, usually occurring on rainy days, a sensible precaution since rain can irreversibly damage pollen grains and/or reduce its adherence to visiting bees (see Dafni, 1996). However, many flowers dehisce in the evening, when humidity is normally increasing, so abiotic cues are not necessarily the whole story (although lowered humidity through a morning could take time to induce sufficient drying for a later dehiscence). Evening opening means that pollen becomes available to crepuscular and nocturnal pollinators, usually bats and moths, but many plants serviced by these taxa are far from depleted overnight and their remaining pollen becomes a major resource for bees through the following day, so that diurnal pollination also occurs. Opening at dusk may be the best strategy for a plant that is ‘‘hedging its bets’’ across several kinds of pollinator syndrome in this way. For example, Miyake and Yahara (1999) suggest that for honeysuckle (Lonicera) the nocturnal pollinators (hawkmoths in their study) may remove less pollen per visit but show significantly greater pollen deposition efficiency than the later diurnal visitors (long-tongued bees), who are nevertheless attracted by the significant persistent pollen. In such a case, opening and dehiscing in the morning would give a surge of bee visits potentially leaving nothing for nocturnal visitors. For plants that have floral morphologies inviting visits from a range of long-tongued animals, then, selection for a dusk anthesis may require uncoupling of any relationship between dehiscence and humidity. Various factors can therefore alter pollen presentation, but how does this affect bee behavior? We know relatively little about bees’ assessment of pollen quantity or quality (Harder, 1990a; Rasheed and Harder, 1997; Robertson et al., 1999), but there is certainly evidence that bees can detect quantities of available pollen on at least some plants. This could be achieved visually or by using olfactory cues associated with dehiscence. Agapostemon bees can detect pollen levels visually in flowers with exposed

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

421

anthers (Goulson et al., 2001), and Bombus detect morphological change in Anemonopsis flowers as a cue to pollen availability (Pellmyr, 1988); for Alkanna the anthers are visible only at very close range but nevertheless give adequate visual cues to Anthophora females (Stone et al., 1999). In Acacia detection may involve both vision and scent (Stone et al., 1998). Beyond assessment of availability in advance of choosing flowers, bees must also make decisions on continuing to gather particular pollen rewards, and this is presumably based on volume or weight considerations such as the packing of the scopa or perhaps rate of increase of body weight. Some bees can assess pollen reward per flower (Harder, 1990a; Robertson et al., 1999), probably using sensitive setae around the scopa that are triggered by displacement. Similarly, Buchmann and Cane (1989) showed that both Bombus and Ptiloglossa can assess the pollen load they are receiving while sonicating Solanum flowers, presumably mechanically rather than from visual or olfactory cues. Certainly bumblebees make more visits per inflorescence when flowers are pollen rich, and return to these inflorescences more frequently; and they abandon pollen collecting on a plant when the rate of gain is too low, even if the anthers are not fully depleted (Harder, 1990a,b). Specific examples of pollen availability shaping bee activities are most readily to be found with pollen-only plants (lacking nectaries) and/or with monolectic bees. On a seasonal scale there is a clear effect for the desert bees already referred to, which can facultatively remain in larval diapause (Danforth, 1999) and do not begin any activity in years when the desert flora is sparse or absent; only when plants such as creosote bush are triggered into flowering by rains do specialist bees emerge synchronously and gather their pollen (Minckley et al., 2000). Similarly for bees facing the opposite problem of a very short season at high latitude in northern Sweden, there is a clear dependency on the pollen presentation patterns of the very limited available flora, with Bombus alpinus moving from unspecialized Saxifraga early on to Astragulus and then Bartsia (both typical of bee-pollinated flowers) as soon as they start flowering, so that on any one day thereafter the bees have 90–100% pollen from one of these species in their scopae (Stenstrom and Bergmann, 1998). Examples also occur for crop plants; the solitary eucerine bee Peponapis, an important pollinator of cucurbits in the United States, has a seasonal emergence and flight period closely tied in with the anthesis of pumpkins (Willis and Kevan, 1995), and crops such as kiwi fruit and pomegranate are also pollenonly plants where bee activity is tightly structured. Perhaps the best known example is the tomato, where Bombus nests are now routinely introduced and managed seasonally in greenhouses for improved pollination, showing

422

P. G. WILLMER AND G. N. STONE

Fig. 15. (A) Activities of male and female Anthophora pauperata on Alkanna plants in wadis in Sinai, with females showing early morning and midafternoon peaks but males present throughout the period of insolation (male visits are mainly due to one resident male patrolling a group of plants with ‘‘approaches’’ rarely involving any resource intake, but with some intruder males also recorded). (B) Pollen availability in Alkanna through the day; male-phase

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

423

tight temporal patterns of pollen gathering (e.g., Banda and Paxton, 1991; Morandin et al., 2001). On a diurnal scale the structuring effects of pollen are most clear when the dehiscence pattern is temporally sharp. Ludwigia elegans is related to evening primroses and shows this kind of brief and tightly defined anthesis and dehiscence; Gimenes et al. (1996) reported precise patterning of visits for its pollen (with some accompanying nectar gathering) in two Brazilian bees, Tetraglossula and Heterosarellus. The same is true for Perdita bees that specialize in collecting pollen from Mentzelia where the anthers open (unusually) in the late afternoon (Michener, 2000). Alkanna plants in Sinai likewise show very precise dehiscence in the arid wadis of Sinai, with associated tight patterning of female Anthophora pauperata visitation (Fig. 15) (Stone et al., 1999). Poppies are perhaps the classic case, producing only pollen and usually in the morning, when it is commonly gathered within a very few hours (Ribbands, 1953). Nansen and Korie (2000) reported clear pollen-related Apis activity on Cistus salvifolius, with a time lag of just 28–60 min between time of pollen release and honeybee activity over a 10-day period. There is also a good match between timing of visits, handling times, and pollen availability in Solanum flowers for three sonicating bee species; here early morning dehiscence gives high pollen reward in new flowers around 07:00–09:00 (Shelly and Villalobos, 2000). But perhaps the clearest evidence comes from Acacia flowers (Stone et al., 1996; Stone et al., 1998), where many species all very similar in flower morphology co-occur in African savanna habitats, with no barriers to heterospecific pollen deposition. All attract mass visitation since they may be the only plants in flower for parts of the season; and most species have no nectar, so we can be sure that pollen is the structuring agent. Flowers mostly last only one day, but crucially within that day different species dehisce at different times, as Fig. 16 shows; pollen availability for five species that coflower in Tanzania is spaced through the day, and can be tightly peaked for any one species, with a pollen window of just 1–2 h. Likewise the visitors (especially Apis and small megachilid bees) have peaks of activity on particular species of tree, and within a particular daily ‘‘window’’ their visits are all or mainly to one Acacia species; hence by the time an individual bee moves to another species it will have deposited and groomed off most of its pollen and so is fairly free of heterospecific pollen. Individual bees are clearly tracking pollen flowers increase in frequency from about 1400 as dehiscence occurs. (C) Resultant foraging behaviors of males and females collectively, with pollen collection (by females) peaking sharply from 15:00, but also occurring in the early morning (where the females are gathering residual pollen from flowers that dehisced the previous day). Data from Willmer et al. (1994) and Stone et al. (1999).

424

P. G. WILLMER AND G. N. STONE

Fig. 16. Patterns of pollen availability in five co-occurring species of Tanzanian Acacia. Different symbols represent different individual trees. Each species has its own ‘‘dehiscence window’’, with the early-opening species particularly tightly clumped; but the five species are statistically regularly dispersed in time, indicative of competitive interactions to avoid overlap and heterospecific pollen transfer. Foraging bees track the pollen availability across the successively opening plant species. Data from Stone et al. (1998).

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

425

availability closely, and their daily round of activities is largely dictated by the patterns of Acacia dehiscence. Even in the case of A.senegal, which does produce nectar (peaking between dawn and 09:30 h), bee activity follows the pollen reward which is commonly greatest from 11:00 h onward. Evidence for pollen structuring could also be derived from bees making pollen-only trips. Using this approach with native plants, Williams and Tepedino (2003) attempted to analyze the effects of resource distance on foraging for pollen-gathering Osmia bees. They showed only a weak relationship, with individual bees consistently mixing pollen from various near and distant patches, even when one was much more abundant than another; this was perhaps because there was always a confounding need to gather some nectar. The confounding effects of nectar can also be reduced or eliminated by working with social bees that genuinely do specialize on just one commodity on any one foraging trip, and only following the pollen gatherers. For example, in Melipona beecheii Biesmeijer and Toth (1998) showed that the pollen-gathering worker individuals were active only in the mornings, for 1–3 h per day that matched up with most local dehiscence times, in contrast with nectar gatherers that worked all day long (and died much younger). As a final caution, we note that highly structured daily activity patterns of pollen-collecting bees that may seem to be related to pollen availability are not necessarily so. El Shafie et al. (2002) found very different daily rhythms of pollen gathering in two species of Apis, even though they shared a habitat and had key pollen sources in common; here the different patterns were probably related to body size and thermal considerations (see Section V.B). F. Activity Patterns Structured by Competition and Predation Here we touch on two relatively ignored issues that deserve further attention. Many bees compete intra- and interspecifically with other bees, and also with other flower visitors, in ways that must affect their activity patterns. Furthermore, flowers are frequently used by predators as sites to locate and attack their prey, and the prey are sometimes bees. Interference competition through removal of resources is the most obvious effect, and undoubtedly occurs on most of the ‘‘generalist’’ flowers visited by bees as well as other pollinators. Even for flowers that can be seen as belonging to a more specialist melittophilous syndrome, many individuals and many species of bee may be visiting, so that a given bee’s activities on a particular plant may be quite different when it visits alone from when it visits as part of a community. Most strikingly, heavy visitation

426

P. G. WILLMER AND G. N. STONE

by a long-tongued bee can reduce the standing crop of nectar in flowers to the point where it is unreachable by shorter tongued visitors Indirect competition between bees can also occur, because of the habit of floral robbery that is common in shorter tongued bees. Here flower corollas are pierced basally to extract nectar, rather than being visited legitimately (i.e., in a manner that would effect pollination). The net effects of robbery can be a severe depletion of the floral resources (again rendering nectar that is present unreachable for some legitimate visitors), and also a significantly reduced floral longevity. The damage done by robbers such as short-tongued bumblebees and carpenter bees often allows later access for ‘‘secondary robbery,’’ commonly by honeybees and small solitary bees; Fig. 17 shows an example of the different timings of activity that can result for robbers and pollinators in a single plant species. In fact, robbery not only alters the activity patterns for pollinators but can sometimes turn them into secondary robbers themselves; for example, Villalobos and Shelly (1996) reported that honeybees visiting Chinese violets made legitimate (pollen-moving) visits only in the mornings, but made nectar-robbing visits (to holes made by carpenter bees) at all times of day. Robbery need not always be detrimental to the plant, sometimes increasing seed set through enhanced outcrossing when long-tongued bees are forced to move around more to find an adequate reward (e.g., Maloof, 2001; Maloof and Inouye, 2000; see also Irwin, 2003), or where the robbers themselves do contact anthers and move pollen around (Navarro, 2000). However, whether good or bad for the plant, floral robbery can certainly alter bee behaviors and in turn affect the community-level interactions (Irwin et al., 2001). Aggressive competition by interspecific defence of floral resources may be less common (except for some males, as described in Section III.B), and seems to be a species-specific trait in bees (Hubbell and Johnson, 1978; Johnson, 1982), correlated with strong and toothed mandibles (Roubik, 1989). Competition by aggressive foraging is particularly characteristic of stingless bees, and Nagamitsu and Inoue (1997) reported a competitive hierarchy amongst seven Trigona species foraging on Santiria flowers, where T. canifrons could exclude other species at times of peak nectar flow; the species coexisted at least partly because the more aggressive bees were less quick to find new resources, giving a window of access (07:00– 09:00 h) to the more timid but efficiently searching species when flowers first opened (see Fig. 18). Slaa et al. (2003) showed avoidance of landings in stingless bees when a larger and more aggressive resident was present on a flower site; however, they also noted that the presence of less aggressive species sometimes acted as an attraction cue.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

427

Fig. 17. Patterns of activity of (A) robbing short-tongued Bombus terrestris; (B) legitimate long-tongued Bombus visitors with a strongly bimodal pattern; and (C) both robbing and legitimate visits by Apis which are unimodal, on Trifolium (clover) in New Zealand. Data from Fussell (1992).

428

P. G. WILLMER AND G. N. STONE

Fig. 18. A competitive and hierarchical interaction among Trigona bees on Santiria flowers, where the aggressive T. canifrons (C) dominates the flowers at times of maximal nectar flow (A), but less aggressive species that search more efficiently (D and E) can locate and forage on the flowers in the early morning when pollen supplies are highest (B), and may be able to return again after midday when T. canifrons is absent again as rewards decline. From Nagamitsu and Inoue (1997).

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

429

It is often assumed that introduced bees can threaten native and usually solitary bees, either by aggressive or interference competition, and there are certainly some documented cases of this. Apis mellifera may have displaced native carpenter bees to less productive floral resources in Argentina (Telleria, 2000); and introduced bumblebees can usurp native queens from their nests and/or disrupt nectar availability (e.g., in Japan and Israel). However, data based on honeybee introductions are somewhat conflicting; Steffan-Dewenter and Tscharntke (2000) showed no significant negative relations between native bee and honeybee densities in central Europe, and Roubik and Wolda (2001), using a 17-year data set from Panama, showed no measurable declines in native bee populations while nonnative honeybees were invading. In a reversal of the normal introduction regime, EI Shafie et al. (2002) have shown coexistence between native Apis mellifera and introduced Apis florea (the ‘‘dwarf’’ honeybee) in Sudan; A. mellifera continued to collect pollen in the early morning and late afternoon, while the smaller A. florea foraged for pollen only in the middle of the day, giving little overlap on the flowers. However, there are specific problems with aggressive Africanized honeybees in the neotropics (Roubik, 2000), which can displace other social and solitary bees. Aggression also occurs between bees and other flower visitors. It might be expected that larger vertebrate visitors (primarily birds, as bat visits are usually at dusk and into the night) would be able to exclude bees from good nectar resources and so alter bee activity patterns. In practice, there are few records of this. Instead, bees can sometimes exclude birds by their aggressive activity around flowers, and may often exclude them by harvesting nectar quickly from bird-adapted flowers to leave only unprofitable and therefore useless flowers from the birds’ perspective. An example of the latter has occurred with introduced honeybees in Mauritius, which deplete nectar from previously preferred local trees (Sideroxylon) early in the morning and force the native white-eyes (nectarivorous birds) to switch to alternative and lower profit plants (Hanson et al., 2002). Honeybees are also blamed for declining numbers of buprestid beetles in Australia, which were formerly important pollinators of mallee scrub plants. Bees may be particularly prone to interact with ants on and around flowers. Ants are frequently present on plants, and are often encouraged into residency to act as a biotic defense (‘‘ant guards’’) by the offer of inducements from the plant, especially in the form of extrafloral nectaries (reviewed by Koptur, 1992). Hence ants that benefit the plant by deterring herbivores could also act as deterrents to bees and other pollinators. How this problem can be avoided in ant-defended plants has rarely been analyzed, except in the case of some Acacia species; here resident ants patrol the buds and flowers nearly all the time, but the flowers produce

430

P. G. WILLMER AND G. N. STONE

volatile scents just as they open (perhaps from the pollen itself) that are deterrent to the ants, so providing a brief window for bees to gather the pollen before ant guards return to the flowers later in the day (Raine et al., 2002; Willmer and Stone, 1997b). Here, the ants are enforcing a very highly structured activity regime on the visiting bees—and are perhaps making those bees’ match with pollen availability look more causally clear-cut than it really is for some Acacia species. Interactions between ants and bees can also occur where both are illegitimate visitors, robbing nectar from flowers; such effects were documented for Justicia flowers in the neotropics (Willmer and Corbet, 1981), where ants occurred mainly in shady sites and for most of the middle hours of the day could deter one of the stingless bees (Trigona fulviventris) that also preferred shady flowers, again enforcing a highly structured temporal activity pattern on the bee. Finally, bees may also interact with ants in one other scenario, where they both tend homopterans on plants for honeydew; Almeida-Neto et al. (2003) recorded reciprocal interference on Bauhinia and on mango trees, with the abundance of Trigona hyalinata directly proportional to homopteran abundance unless ants were also present. Predation of bees at flowers is also potentially important in eliciting antipredator behaviors that affect overall activity patterns (Dukas, 2001a,b). Bees may be captured on flowers by birds; by spiders; by mantids, bugs, and robber flies; and by some specialist bee-hunting sphecid wasps (‘‘bee wolves,’’ in the genus Philanthus). For example, individual bumblebees had a 14% chance each day of being attacked by a spider in a study on Asclepias flowers by Morse (1986), and bumblebees commonly have 10–15% parasitism by conopid flies as a result of flower visits (see Dukas, 2001b), so predation and attack rates on bees are certainly significant; Visscher and Dukas (1997) suggest that predation plays a strong role in selection on foraging ranges in social bees. Honeybees are evidently responsive to threats from predation, having been shown to prefer apparently ‘‘safe’’ flowers over more rewarding alternatives harboring either a spider or a dead bee, and avoiding revisiting sites where they had recently been threatened by predators (Dukas, 2001a); although this kind of effect is probably less in bees from honey-depleted colonies who are more prepared to take risks (Cartar, 1991). This in turn can impact on pollination success; for example, Suttle (2003) recorded lower seed set in Leucanthemum when the presence of crab spiders on flowers reduced the frequency and length of visits by bees and other insects, and suggests that the effect may be widespread and important. Finally, bee activity patterns can be altered in relation to predation/ parasitism at the nest site. Many bee nests suffer brood parasitism either by cuckoo bees (see Section III.D) or by other insects such as sapygid and

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

431

torymid wasps or bombylid flies, all of these seeking to lay eggs on the stored provisions in unsealed cells. Hence there is often strong selection to keep foraging trips as short as possible and protect the nest. Goodell (2003) has shown that Osmia bees suffer greater parasitism when forage resources are scarce, probably because the bees required more frequent and/or longer foraging trips. Bee activities may also be affected by avoidance behaviors required when leaving or entering the nest: for example, Wcislo and Schatz (2003) have shown that a halictid sweat bee (Lasioglossum) changes her behavior on return from a foraging trip if a predatory ant (Ectatomma) is waiting at her nest entrance, spending time on inspections, zig–zag approaches, or distant landing and a walk to the entrance from a different direction. G. Interaction of All These Cues: The Effects of Coevolution We have seen that abiotic components of the environment affect both bee foraging costs and the spatiotemporal structure of plant rewards. It must be cautioned that in many of the studies so far reported only one or other of these two aspects have specifically been considered, and weather effects are frequently not separated from resource effects. Rare exceptions include Willmer (1983), where climatic factors were shown to be more important than reward levels for smaller bees, but rewards were critical (and ambient temperature relatively unimportant) for large bumblebees having to pay the costs of endothermy early in the morning but benefiting from high nectar reserves; and Stone et al. (1999), where the balance of thermal and resource constraints on Anthophora pauperata were carefully dissected. Abrol (1998) similarly showed a dependence on solar radiation and (negatively) on humidity for the small bee Megachile nana, with only a minor and indirect effect from nectar concentration, while a study of ‘‘departure rules’’ for bees (Collevatti et al., 1997) revealed that for smaller bees quitting a patch of flowers was primarily dependent on time already spent there, whereas larger bees used a probabilistic rule related to the frequency of rewarding flowers and the likelihood of the next flower having an adequate reward. Thus in cases where both sides of this interaction have been studied, a common finding to emerge—all other things being equal— is that size matters, smaller bees being more dependent on the abiotic (climatic) factors, and the behavior of larger bees being more strongly structured by floral rewards. In practice other things are rarely equal and the confounding effects of other bees (especially conspecific males), other foragers, and other flower residents must also be factored in. Is it the flowers or the bees that are in control here? Bees can be active at flowers only when and where they do not get too warm (or too cold), and

432

P. G. WILLMER AND G. N. STONE

perhaps where they get enough water; in this sense the bee is setting the limits on the relationship. Within her broad physiological tolerance window an individual foraging bee can make behavioral choices to visit flowers (intra- or interspecifically, on one or more individual plants) according to their age or stage, their position on the plant and their local microclimate, and perhaps according to some distant reward assessment. The bee can choose rate of visitation, handling time per flower, movement pattern (hovering or landing, flying, or walking between flowers), and can vary floral fidelity, potentially abandoning a given plant species if its rewards fall too low. But since the pollination interaction is best seen as a coevolutionary system (mutualistic, although of course not altruistic), any flower that has a relatively narrow range of potential effectively pollinating bee visitors should therefore be selected to respond to the choices of these bees with an appropriate reward to keep those bees visiting and moving between flowers. Too little reward per flower risks the bees giving up, but too much reward will reduce the interflower movements of the bees; ideally the flower must offer just enough to keep the bees hungry but faithful, and so promote outcrossing. Selection should therefore operate to match the plant outputs (in that particular environment) to the temporal ‘‘bee window’’ (in that particular environment). It should ensure that the anther tissues of the plant desiccate at a rate commensurate with dehiscing at the appropriate time; and that its nectary tissues begin secreting nectar at a time, rate, and concentration that will (after interaction with the local environment) give a suitable reward at the appropriate time. However, in particular parts of the range these relationships may be fine-tuned by further selection to take account of competitive interactions occurring at the flowers, and, as Section VI shows, the ‘‘appropriate time’’ may be influenced by competitive effects from other plants in the community. Figure 19 summarizes all these interacting elements that may feed into bee pollinator energy budgets and hence bee activity patterns.

V. Special Effects of Sociality on Bee Activity Patterns As yet we have largely considered the activity patterns of bees collectively, making no distinction as to social system. But bee behavior varies substantially between solitary females foraging only for their own offspring, and various degrees of cooperating females where the agenda may be different. We therefore need to consider how far the constraints already discussed for all bees are important for social bees, and whether other constraints on activity pattern come into play for them. Certainly the majority of published studies on all aspects of bee behavior have focused

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

433

Fig. 19. A summary of the interactions between abiotic constraints (top) and floral resource levels (middle), and their combined effects (together with competitive interactions of other flowers and other visitors) on bee foraging trips.

on the common and familiar social bees, primarily the honeybees (Apis) and the bumblebees (Bombus), but in reality eusociality (despite having evolved separately at least eight times) is the minority condition for bees. A. Levels of Sociality in Bees There are various levels of sociality in animals as traditionally defined (e.g., Andersson, 1984; Brockmann, 1984; Holldobler and Wilson, 1990; Michener, 1974, 1985; Wilson, 1971) although definitions differ (e.g., Crespi and Yanega, 1995; Sherman et al., 1995). The bee superfamily (Apoidea) includes representatives at all levels, from strictly solitary species through to enormous eusocial colonies (Michener, 1974, 2000; Wilson, 1971). The possible degrees of sociality, and the underlying behavioral, genetic, and theoretical frameworks, have been reviewed in many of the sources just cited, so it is appropriate here only to consider how social structure specifically affects foraging patterns. The majority of bees are solitary, each female making and stocking her own nest and interacting only with her own progeny and then only as eggs laid in the nest (Batra, 1978, 1984). However, nests of one species are often grouped together as an aggregation, and individual bees may pick up cues

434

P. G. WILLMER AND G. N. STONE

on foraging opportunities from other incoming bees bearing pollen and floral scents, so affecting each others’ foraging activity. The great majority of solitary bees have only one generation per year, each adult flying for only a few weeks; however, a few temperate bees do have two (or several overlapping) generations per year, and some tropical species breed continuously and aseasonally, so that individuals can be encountered foraging at any time of the year. Parasocial bees (including communal, quasisocial, and semisocial groups) have adult bee populations consisting of a single generation at any one time, again with no interaction between a female and her offspring. In most respects activity patterns in these bees differ little from those of purely solitary species, since all females are stocking cells and most are laying their eggs on the resultant cell contents. In theory, where cells are stocked cooperatively there could be less demand on resources from each individual forager; but in practice the inputs are probably the same and the cooperation instead means that more cells can be made per unit time. However, there may be some scope for interactions between foragers in relation to information about available flowers, where bees pick up floral odors and may therefore be signaling the resource on which they have been foraging as they return to a nest. This could lead to some targeting of foraging bouts, and greater efficiency in the balance between travel and search time, and handling time. Eusocial bees live in nests with two generations of mature bees, a mother and her daughters. In primitively eusocial types, such as Bombus and some Lasioglossum, mother ‘‘queens’’ and daughter ‘‘workers’’ are usually physically similar although different in size, and queens can forage normally when required to do so (especially in the spring). In contrast, in highly eusocial bees (Apis, in colonies of up to 60,000–80,000, and the stingless bees such as Trigona with up to 180,000 individuals) the queens and workers are usually physically distinct, and very different in size. The workers do not mate, and if they do reproduce can only usually give rise to sons from unfertilized eggs (Roubik, 1989). Instead they are engaged on nest maintenance tasks and foraging, and feeding nonforaging nestmates by trophallaxis (Hart and Ratnieks 2002; Michener, 1974; Wainselboim et al., 2002); while the queens lack the normal structures for gathering or manipulating pollen and cannot survive long outside their colonies, so that new colonies must arise by swarming. Apis species are medium-sized and relatively nonaggressive, as are the stingless bees in the genus Melipona, whereas Trigona and related genera are usually much smaller, often highly aggressive to other bees, and commonly engage in nectar robbing on large flowers. We should note here that levels of sociality are not always fixed. Carpenter bees (the large Xylocopa and the smaller Ceratina) can show

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

435

facultative sociality, usually when the nest of a founding queen is usurped by a second female and the foundress becomes a guard and helper (Dunn and Richards, 2003; Stark et al., 1990; Watmough, 1983), so altering the activity patterns of both partners. The same probably applies to some euglossine bees (Cameron and Ramirez, 2001). Subsociality can also be facultative, particularly in some halictid bees (‘‘sweat bees’’) where the presence of adults helps to reduce ant predation on the larvae (Smith et al., 2003). Indeed, halictid bees are notoriously labile in their degrees of sociality, and may even have two different social morphs within one species (Richards et al., 2003); phylogenetic analysis suggests this diversity of habit has resulted from repeated reversals to solitary nesting or social polymorphism from a single eusocial root (Danforth et al., 2003a). In the present context of biotic versus abiotic influences on bee activity, it is particularly noteworthy that levels of sociality may be partly dependent on climatic factors. Halictus rubicundus is fully social in warmer climates and entirely solitary in cooler climates, with a mixture of behaviors in marginal zones (e.g., Potts and Willmer, 1997, 1998; Soucy, 2002; Soucy and Danforth, 2002). Almost opposite effects have been shown between years within a site for Halictus ligatus, with harsher conditions promoting more classical eusociality (Richards and Packer, 1996) perhaps because cool seasons produced small brood and greater queen–worker disparity. In both species, photoperiod, temperature, and resource availability all appear to have an effect on social interactions, and any or all of these factors may have contributed to loss of eusocial habits in these and other species. In allodapine bees, which also show varying levels of sociality related to climate and/ or latitude, Cronin (2001) has used nest translocations to show that this variation results both from behavioral plasticity within populations and from adaptive change between genetically distinct populations. The effects of climatic variables on social behavior certainly deserve greater attention. B. Activity Patterns in Eusocial Species Eusocial bees differ from other bees in several key features: 1. Some workers are not engaged in foraging, remaining nest-bound for whole days or sometimes for their whole adult lives, having an entirely different kind of activity pattern. 2. Much larger quantities of nectar and pollen may be gathered by a single foraging individual, who can in principle spend all day foraging if weather and resource availability permit, not necessarily having other tasks to perform.

436

P. G. WILLMER AND G. N. STONE

3. Each foraging bee can carry relatively very large loads of nectar and/ or pollen (figures of 92.5% body weight of nectar have been recorded for Apis workers), and will often forage over greater distances. 4. A longer seasonal sequence of forage plants is usually needed to support the activity of a species throughout a substantial part of each year. 5. Nests can be better defended against brood parasitism. 6. Nectar can be stored in open cells in the nest, for use by other nestbound workers; both nectar and pollen can be collected very intensively on days of plenty and/or when weather conditions are favorable but do not necessarily have to be gathered every day. 7. Queens can affect and control the behavior of workers. 8. There is particularly striking thermoregulation of the nest, and individuals may also derive substantial thermal benefit from the stable microclimate of the nest as a refuge. 9. Enhanced learning abilities lead to greater floral constancy for any one individual, and to other effects on efficiency of foraging. 10. Most crucially, social facilitation of foraging occurs due to complex interindividual communication. These last six points deserve particular attention for the way in which they can affect activity patterns. 1. Nest Defense Communal defense clearly adds an extra component into the overall activity patterns of social bees. Social bees may spend significant periods each day as guards at the nest entrance, and may take on more active defensive roles. In Apis florea acoustic warning signals from bees that detect an attack are taken up by others as a ‘‘hissing’’ response, spreading through the nest (Sen Sarma et al., 2002). This is noisy enough to deter small predators, and also leads to a reduction or total cessation of departures by foragers so reducing possible losses to larger predators such as birds. We should also note that effective defense, with no or lower brood losses, can also reduce the impact of nest parasitism and predation on the future activity demanded of the foraging individuals. 2. Nectar and Pollen Balance In contrast to solitary bees, individual social bees do not need to gather nectar and pollen in the ‘‘right’’ proportions at any one time, therefore are not forced into making compromises in collecting a balance of these two

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

437

resources within one trip or within a few trips. A solitary bee may have to switch between rich pollen and rich nectar sources, even when those are far apart, entailing much more flight time; a social bee can stick with one source, the problem being resolved at colony level by varying the proportions of nectar and pollen collectors. Using individually marked Melipona bees, Biesmeijer and Toth (1998) found that about half the individuals from one nest specialized as pollen or nectar or resin collectors for their entire adult life; other individuals switched daily, gathering pollen and/or resin in the morning, and nectar in the afternoons. This also affected longevity, with pollen foragers active only for 1–3 h in the morning and surviving for 12 days on average, but nectar foragers sometimes active all day and lasting for only 3 days. 3. Control of Worker Behavior and Task Allocation Following on from the last point, allocation of tasks in eusocial bee colonies is controlled both intrinsically (within the hive) and by extrinsic factors (reviewed by Gordon, 1996; see Sumpter and Pratt, 2003). Worker age and size both have some effect on the tasks they perform; for example, the larger bumblebees within a colony tend to be foragers with the smaller ones working within the nest (Goulson et al., 2002). Furthermore, some patrilines (i.e., offspring from particular fathers, after the multiple mating of the queen) are more predisposed to certain tasks than others. We now know something of the genes that control division of labor, notably the ‘‘foraging gene’’ Amfor in Apis; this is upregulated (partly under the influence of juvenile hormone; see Fahrbach, 1997) at the age-related transition point into being a forager, acting by modulating the bee’s phototaxis (Ben-Shahar et al., 2003). But this is not completely predetermined. Honeybees do progress through different tasks as they age (‘‘age polyethism’’), but the particular sequence is not rigidly fixed, and foragers can revert to nurse activities (see also Section II.C). Queens certainly ‘‘overrule’’ and direct tasks to some degree (Jaycox, 1970); for example, queen pheromones may stimulate more nectar collection by workers, and thus collection of pollen too, albeit indirectly. The presence of brood also appears to stimulate pollen collecting, probably via a further pheromonal effect; but workers are able to assess pollen status directly for themselves and adjust their pollen gathering appropriately (Vaughan and Calderone, 2002), and in honeybees individual workers inspect half-filled cells and thus directly assess the colony need for pollen inputs (Calderone and Johnson, 2002). Thus the number or proportion of workers active outside the nest engaged in particular foraging tasks is extremely labile according to colony need. Indeed, reallocation of labor also occurs in response to heat stress in honeybee nests (Johnson, 2002), providing a neat link to the next section.

438

P. G. WILLMER AND G. N. STONE

4. Thermoregulation and Climatic Dependence Bumblebees were among the first invertebrates shown to exhibit partial endothermy, and their physiological and behavioral mechanisms for maintaining a constant Tb before and during foraging flights are particularly well known (reviewed by Heinrich, 1979a, 1993; and see Section II). Honeybee thermoregulation is also relatively sophisticated (Coelho, 1991; Underwood, 1991), although lacking some of the elaboration of bumblebees. However, as members of social colonies both these genera derive substantial benefit from the emergent physiological properties (often termed ‘‘colony homeostasis’’) of a nest (Kronenberg and Heller, 1982; Roubik and Peralta, 1983; Southwick, 1990; Southwick and Moritz, 1987; Vogt, 1986a,b) or of a swarm (Heinrich, 1981a,b; Seeley et al., 2003); thus physiological constraints on activity are somewhat relaxed and individuals may be more inclined to take foraging risks at the margins of the ‘‘safe’’ thermal window. For honeybees the special ability to gather water and use it to cool the hive evaporatively (e.g., Kuhnholz and Seeley, 1997), and wing fanning by the worker bees (which together can keep hive conditions relatively cool and  comfortable even at air temperatures up to 39–40 C), represent additions to the behavioral repertoire that impact on overall activity patterns for many individual bees (see Heinrich, 1993). Nevertheless, temperature considerations remain important structuring agents for social bee diel activites. For example, in Sudan Apis mellifera collected pollen in the early morning and late afternoon while introduced Apis florea (the ‘‘dwarf’’ honeybee) foraged for pollen in the middle of the day, giving little overlap on the flowers (El Shafie et al., 2002). Likewise, Bombus species are usually active earlier and later than Apis and tend to reduce in numbers around midday (e.g., Stanghellini et al., 2002; Willmer, 1983), with precise patterning related in part to specific body size as discussed in Section II. Non-size-related physiological differences also occur between Bombus species, however; for example, of the common European bumblebees, the medium-sized B. lapidarius has a distinctly higher setting for its thermal window than most other species with which it is sympatric (Corbet et al., 1993). 5. Learning and Floral Constancy The ability to learn in social bees is a topic of intense research activity, from both the behavioral perspective (see reviews by Greggers and Menzel, 1993; Srinavasan and Zhang, 2003) and the neurological perspective (see Hammer, 1997; Hammer and Menzel, 1995; Menzel, 2001a,b; Menzel and Muller, 1996), especially using Apis and Bombus. Learned behavior is seemingly more marked and more complex in social

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

439

than in solitary bees (Campan and Lehrer, 2002; Dukas and Real, 1991). Therefore we focus here on aspects of learning that can affect (usually increase the efficiency of) social foraging activity, and so impact on overall activity patterns. a. Floral characters and constancy Eusocial bees can learn various features of flowers, including shape and size, orientation and symmetry (e.g., Lehrer, 1999; Moller, 1995), and also color (e.g., Giurfa et al., 1995; Gumbert, 2000; Hills et al., 1997), allowing them to learn a preference for a particular floral species at a given time (Menzel, 2001b). Solitary bees may also show constancy (Gross, 1992), but it is probably much less developed in most species. Bumblebee flower constancy has been particularly well documented, with both natural and artifical flowers, and the strategy of ‘‘majoring’’ on one species while having other flowers as minor reward sources is well known (Heinrich, 1979b). Honeybees are often even more constant in their behavior, with many successive trips to a single floral species. The efficiency of memory for particular species is certainly reinforced by learned motor patterns (handling behaviors); bees that learn a single task in the laboratory are more efficient than those that learn two (Chittka and Thomson, 1997), and individual bees often learn their own particular ways of handling complex flowers. In field studies handling time for a bumblebee on a variety of plants may be reduced 10-fold after a reasonable learning period (Laverty, 1994), greatly increasing the rate of intake of nectar. However, limitations from motor and sensory factors are not the whole story: persisting with the major plant may depend on its continuing to match a reward expectation (Wiegmann et al., 2003), with a bee sampling alternative flowers only when this expectation is no longer realized. Learning about this match to expectation can be remarkably fast: for example, honeybees appear to learn characteristics of nectar flow rate within one foraging bout (Wainselboim et al., 2002). Bumblebees show only limited intrinsic preferences for flower types (except perhaps for symmetry; Moller, 1995), but they can readily learn to associate reward size with flower size (Blarer et al., 2002) and may also learn to associate either of these with floral scents (Dornhaus and Chittka, 1999; Menzel and Muller, 1996). However, they do not transfer this learning to novel situations, thus limiting the overall efficiency of their foraging. Worker bees from eusocial colonies also tend to have temporary ‘‘constancy’’ to either pollen or nectar (Free, 1970), although they often have a full load of one and an additional small load of the other. There is no clear developmental precedence of one activity over another as the bee ages (indeed, some bees gather pollen all their working lives and others gather

440

P. G. WILLMER AND G. N. STONE

only nectar), but switching between the two is common and switches from pollen to nectar are commoner than reverse switches. The result of this relative constancy is that for Bombus 44–65% of pollen loads are monospecific (Free, 1970), and this figure is usually even higher for Apis. b. Landmarks, direction, and navigation Honeybees make ‘‘orientation flights’’ when they first mature into foragers, and make ‘‘learning flights’’ when they first encounter new food sources; in both cases they are memorizing visual landmarks to guide their return. The orientation flights have a fixed duration but there is increased flight speed with increasing experience, so that later trips cover larger areas (Capaldi et al., 2000). In contrast the learning flights increase in duration with landscape complexity, but decline in duration as familiarity with the environment increases; they also depend on the sugar concentration of the food source, suggesting that ‘‘learning effort’’ is adjustable (Wei et al., 2002). Both kinds of landmark memorizing can clearly affect subsequent activity patterns, potentially reducing trip durations and improving overall foraging efficiency. Honeybees store landmark information to locate sites they have previously visited, acquiring a series of stored visual images (‘‘general landscape memory’’; Menzel, 2001b); they also navigate using directional (sun and compass-related) cues, which feed into the landscape memory. Bumblebees can learn particular flight paths within a group of flowers (Thomson, 1996), allowing them to use ‘‘traplining’’ foraging, and honeybees too use a ‘‘specialized route memory’’ according to Menzel (2001b). Again, trip durations can be reduced—or at least the ratio of travel time to feeding time can be optimized—by such learning. Finally, experiments with trained honeybees indicate an intrinsic time sense in Apis, where information is learned differently according to time of day (Gould, 1987); thus a honeybee will return to the same patch of flowers at roughly the same time over several days. Some Trigona bees also show ‘‘time–place learning’’ with anticipatory visits to a daily food training site (Breed et al., 2002). However, for many other bees correct timing at a patch of flowers may be simply dependent on external stimuli such as floral odors. c. Conditioning and association Honeybees show classic conditioning responses, and can form complex associative links (e.g., between visual or olfactory floral advertisements and received floral profits). Hence they can apply learned flower features to distinguish between unfamiliar floral species, and use prior knowledge to detect poorly visible or cryptic flowers. All these attributes impact on their floral foraging efficiency and can reduce the length of time spent per unit reward, so affecting the overall activity patterns.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

441

d. Development and duration of memory Newly emerged and very young workers cannot achieve associative learning (e.g., linking odors and rewards), this ability normally appearing at about 5–9 days (Ichikawa and Sasaki, 2003). However, olfactory information can be gained during the pupal phase that affects behavior in mature bees (Sandoz et al., 2000), so that adult bee foraging activity could be affected by the stored pollens encountered within the nest as juveniles, perhaps giving ‘‘learned’’ continuity from one season to another. In the laboratory, memories relating to flower characters once acquired can persist for 3 weeks or more (Chittka, 1998), although they can decay a little each night and improve again each day (Keasar et al., 1996). Foraging success in honeybees increases through the first week of foraging activity, which corresponds with a median life span, so that these bees are spending a large part of their adult life in learning and improving their foraging skills (Dukas and Visscher, 1994). They perhaps only achieve maximally ‘‘efficient’’ activity patterns as they approach the end of their lives. Eusocial bees appear to have both long- and short-term memories; longterm retention is affected by experience of many flowers or many foraging trips, while short-term memory is related only to the last flower visited (Creswell, 1990; Real, 1992). In fact, Menzel (2001b) distinguishes three memory systems for honeybees: early short term (seconds only, for intrapatch movements), late short term (minutes, for interpatch movement), and mid- to long-term memory (over hours or days) that operates for successive foraging bouts and can even last for months with bees that overwinter. The first two will certainly moderate the activity pattern of an individual social bee as observed on any one day, and should be taken into account in seeking underlying causes of patterns. Linking together this and the preceding subsection, it has been demonstrated that thermal experience in honeybees can also affect their adult behavior (Tautz et al., 2003). Bees raised at higher brood temperature  (36 C) through the pupal stage were significantly faster learners than bees raised at 32 or 34  C, and also produced more energetic and consistent dances on return to the hive. As a result of their substantial learning capacities, eusocial bees are especially noteworthy for their ability to make complex decisions about flower foraging (although there are so few studies of solitary bees available that we cannot be sure their abilities are either lesser or different in kind). Because of this background knowledge, and the ease of training and management, social bees have been the chosen focus of many studies on food choice and discrimination, on optimization, and on the currency used in making foraging decisions, using both natural flower communities and artificial flower arrays. To summarize a great deal of this research, in

442

P. G. WILLMER AND G. N. STONE

general eusocial bees seem not to respond to nectar concentration or sugar rewards per se, and not to maximize their rate of energy gain except perhaps in the very short term, but rather they often prefer minimum variance flowers (‘‘risk-sensitive foraging’’), and will leave a highvariance inflorescence more quickly than one with constant resource (e.g., Biernaskie et al., 2002; Thiyagesan et al., 2001; see also reviews by Cartar and Abrahams, 1996; Perez and Waddington, 1996). Thus they evaluate floral rewards in a plant population in proportion to the mean, so that their choices are guided by the coefficient of variation of the distribution. Plants should then evolve strategies that maximize their perceived profitability to a bee with this approach, resulting in ‘‘cognition-mediated coevolution’’ between plant and bee (Shafir et al., 2003). However, it may be that the social bees have a tendency to be a little more ‘‘risky’’ in their behaviors than solitary bees, since the success of the nest (and an individual worker’s reproductive success in transferring its genes into the next generation) is not determined solely by the resource input from any one bee. As always, other variables come into play. An individual bee’s preferences may change with colony state (e.g., Thiyagesan et al., 2001), so that a bee from a nest that is short of nectar may go for gross intake of nectar volume, while bees from a well-endowed colony may prefer minimum risk and/or maximum profitability strategies. The foraging preferences of eusocial bees can also show some fixed intraspecific variation; individual honeybees and bumblebees from any one nest do vary greatly in size (often with a seasonal component), and small workers may make quite different choices from large workers (e.g., Goulson et al., 2002; Harder, 1988; Spaethe and Weidenmuller, 2002). Cox and Myerscough (2003) have developed models showing that honeybee colonies with more heterogeneous foragers do better than uniform colonies, which must be at least partly due to extension of the range of flowers and rewards that can be exploited. 6. Social Facilitation and Communication Social facilitation occurs when individuals influence and potentially enhance each others’ foraging. It can operate both at the level of recruitment to flower patches, and of flower choice within patches, and is perhaps the major factor altering the activity patterns of social as compared with solitary bees. a. Recruitment In a simple example, the arrival back at the nest of a Bombus well loaded with provisions increases the likelihood that another individual will leave on a foraging trip. There is some indication that this involves sound communication by buzzing in bumblebees (Schneider, 1972), with additional effects from irregular runs within the nest that seem

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

443

to spread a chemical signal (Dornhaus and Chittka, 2001), probably a specific pheromone from a tergal gland (Dornhaus et al., 2003). However, primitively eusocial bees such as these show no signs of being able to communicate any geographic details of food location. At best, they rely again on the odor of the provisions that they bring in, or odors temporarily carried on their bodies, which together with the general increase in activity prompted by their arrival and running behavior can direct new recruits to the source the incoming forager has been exploiting. Thus bumblebees fed in the nest with artificial nectar laced with mint, carnation, or aniseed scents will thereafter forage preferentially from nectar sources with the right odor (Dornhaus and Chittka, 1999). Similar effects seem to underlie recruitment in some stingless bees, where individuals can be recruited to artificial syrups only if the syrup has a scent added. However, simple forms of alerting behavior on return to the nest, using zig–zag runs that produce a general jostling effect, also occur (reviewed by Michener, 1974). Individual experienced bees may also lead new foragers for a short distance in the right direction from the nest toward a good forage site. Some stingless bees, especially when operating in the still air of tropical forests, can additionally use pheromone trails outside the nest to guide new foragers, and the scent marks can be polarized with heavier scents nearer the goal and sometimes an additional unique mark at the feeding site (Schmidt et al., 2003). Some returning Melipona show even greater sophistication, making sounds by wing vibration whose duration is proportional to distance from the food source. Aguilar and Briceno (2002) have shown that the pulse pattern of the sound also varies with concentration of the located food, and information about the height above ground of the resource can also be conveyed (Nieh et al., 2003). However, floral odor on the body of returning foragers continues to play a role for most of these species. Indeed, simple odor cues still play a role in the much more sophisticated communication systems of Apis, since changing the stored food in a hive can redirect subsequent foragers (Free, 1969). Honeybees may also augment the scent and hence attractiveness of floral sources once located, with secretions from their abdominal Nasonov glands (a habit more common in the subtropical Apis cerana and A. florea than in the familiar temperate A. mellifera). However, these recruitment behaviors are usually overridden by the much more precise individual-to-individual communication, by both chemical and mechanical signals, that are contained in the returning honeybee’s dances on the vertical walls within the nest, which are considerably more elaborate than the zig–zagging of stingless bees. These dances (the round dance, the tremble dance, and the tail-wagging dance) can convey vector information (i.e., range and direction) as well as identity of the

444

P. G. WILLMER AND G. N. STONE

flowers, and vary in complexity in relation to environmental variables (Steffan-Dewenter and Kuhn, 2003; Thom, 2003). Their effect can also be modified by acoustic signals (‘‘piping’’) that may retard recruitment to low-quality food sources (Thom et al., 2003). The end point of all these different kinds of communication is that the eusocial bees are able to recruit large numbers of individuals to newly opening flowers (or to artificial food supplies used to test or train them); inevitably such behaviors affect activity patterns for the colony as a whole, but also have substantial effects for individuals, who need waste much less time searching for suitable flowers. b. Flower choice Social facilitation of foraging may also affect choices between individual flowers within a foraging patch. ‘‘Local enhancement’’ can occur in stingless bees when individuals prefer to feed close to a nestmate (Slaa et al., 2003), although in some of these species the more experienced bees instead show ‘‘local inhibition’’ and space themselves away from existing foragers (so presumably avoiding recently emptied flowers). At a more sophisticated level, many eusocial bees can scent mark individual flowers after visiting them, essentially by leaving a ‘‘footprint’’ of volatile chemical, so that subsequent visitors avoid recently depleted flowers. This has been shown for bumblebees (Goulson et al., 1998, 2001; Schmidt and Bertsch, 1990; Stout et al., 1998), for honeybees (Ferguson and Free, 1979; Free and Williams, 1983; Wetherwax, 1986; Williams, 1998), and for stingless bees (Schmidt et al., 2003). A bee may avoid flowers visited only by conspecifics, or may also avoid those marked by congenerics; in practice honeybees and bumblebees both avoid flowers visited by each other, and seem to share the volatile 2-heptanone as a component of the repellent mark, although honeybees may also have a longer lasting attractant marker (Stout and Goulson, 2001). Marking and repellency are not automatic though; Goulson et al. (2001) concluded that many social bees rely on direct detection of reward when this is visible, and scent mark only when the rewards are hidden. Stout and Goulson (2002) also showed that the period of repellency was markedly different among flower species for the same bumblebee visitors, this period varying inversely with nectar secretion rates. Thus either signaling bees can alter the strength of the mark they deposit, or receiving bees can alter the duration of their response to repellent marks (or both may occur). Evidently there is considerable sophistication here geared to maximizing foraging returns and making activities as efficient as possible. However, flower footprinting is not unique to eusocial bee species, as some form of floral marking also operates in species of Xylocopa with varying levels of sociality (e.g., Frankie and Vinson, 1977) and even in

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

445

the entirely solitary genus Centris (Raw, 1975). In fact, the most detailed demonstrations are in Anthophora (Gilbert et al., 2001), where the marks left are complex multicomponent volatiles, individually recognizable and varying with foraging needs, and perhaps used as dominance/exclusion markers. Possibly flower marking originated as a side effect of territorial marking behaviors, and has merely taken on new roles in directing feeding activity and so enhancing colony foraging success in eusocial species.

VI. Summary: Why Bee Activity Patterns Matter Many plants, in nearly all ecosystems, share pollinators, and in community assemblages plants may therefore compete for pollen quantity or for pollen quality (essentially pollen purity, since heterospecific pollen deposition wastes male function directly and female function indirectly by blocking stigmatic surfaces). The former means receiving as many visits as possible, the latter means competing for the services of high-fidelity visitors. In general, the issue of pollen quality seems to be more important under most conditions (reviewed by Waser, 1983), and is especially relevant to bees and their activities since they are high-fidelity flower visitors and the dominant pollinators in most communities. This should impose selection on many plants to flower at different times, seasonally or even daily. Figure 16 showed an example; such cases provide powerful evidence not only for the ability of bees to track resource availability across multiple species, but also for the impact of bee activities on plant community structure. Understanding the array of factors that affect bee activities therefore has strong applied importance for habitat and community conservation, in the face of fragmentation, disturbance, and species introductions, as well as for applied management of crop pollinators. Pollination is a key driver in the maintenance of biodiversity and ecosystem function, and the sustainable management of pollinators is a crucial part of any conservation plan (e.g., Kremen and Ricketts, 2000; Kearns et al., 1998; Matheson et al., 1996). Only studies that recognize the diverse constraints on bee behavior, and take into account (and attempt to disentangle) both the intrinsic variables of particular bee species and the extrinsic influences in particular habitats, can have real predictive value. The common findings that smaller bees are more dependent on abiotic (climatic) factors, and that the behavior of larger bees is more strongly structured by floral rewards, may be generally useful; but in any given bee–plant interaction it must always be tempered by the confounding effects of other bees (especially conspecific males), other foragers, and other flower residents. We hope that our reviews of

446

P. G. WILLMER AND G. N. STONE

such influences will assist many future studies and further the cause of preserving bee diversity. References Abrol, D. P. (1998). Foraging ecology and behaviour of the alfalfa pollinating bee species Megachile nana (Hymenoptera: Megachilidae). Entomol. Gen. 22, 233–237. Aguilar, I., and Briceno, D. (2002). Sounds in Melipona costaricensis (Apidae; Meliponini): Effect of sugar concentration and nectar source distance. Apidologie 33, 375–388. Aizen, M. A., and Basilio, A. (1998). Sex differential nectar secretion in protandrous Alstroemeria aurea (Alstromeriaceae): Is production altered by pollen removal and receipt? Am. J. Bot. 85, 245–252. Alcock, J. (1991). Mate-locating behavior of Xylocopa californica arizonensis Cresson (Hymenoptera, Anthophoridae). J. Kans. Entomol. Soc. 64, 349–356. Alcock, J. (1993). Differences in site fidelity among territorial males of the carpenter bee Xylocopa varipuncta (Hymenoptera, Anthophoridae). Behaviour 125, 199–217. Alcock, J. (1996). Timing of mate-locating activities by males in relation to female activity in the carpenter bee Xylocopa varipuncta (Hymenoptera, Apidae). J. Insect Behav. 9, 321–328. Alcock, J. (1999). The nesting behavior of Dawson’s burrowing bee, Amegilla dawsonii (Hymenoptera: Anthophorini), and the production of offspring of different sizes. J. Insect Behav. 12, 363–384. Alcock, J., Barrows, E. M., Gordh, G., Hubbard, L. J., Kirkendall, L., Pyle, D., Ponder, T. L., and Zalom, F. G. (1978). The ecology and evolution of male reproductive behaviour in the bees and wasps. Biol. J. Linn. Soc. 64, 293–326. Aldini, R. N. (1994). Preliminary observations on gregarious sleeping in Amegilla (S.Str.) garrula (Rossi) (Hymenoptera, Apoidea, Anthophoridae). Ethol. Ecol. Evol. 3, 131–135. Almeida-Neto, M., Izzo, T. J., Raimundo, R. L. G., and Rossa-Feres, D. C. (2003). Reciprocal interference between ants and stingless bees attending the honeydew producing homopteran Aetalion reticulatum (Homoptera: Aetalionidae). Sociobiology 42, 369–380. Andersson, M. (1984). The evolution of eusociality. Annu. Rev. Ecol. Syst. 15, 165–189. Anzenberger, G. (1977). Ethological study of African carpenter bees of the genus Xylocopa (Hymenoptera, Anthophoridae). Z. Tierpsychol. 44, 337–374. Armbruster, W. S. (1984). The role of resin in angiosperm pollination: Ecological and chemical considerations. Am. J. Bot. 71, 1149–1160. Armbruster, W. S., and Berg, E. E. (1994). Thermal ecology of male euglossine bees in a tropical wet forest: Fragrance foraging in relation to operative temperature. Biotropica 26, 50–60. Armbruster, W. S., and McCormick, K. D. (1990). Diel foraging patterns of male euglossine bees: Ecological causes and evolutionary response by plants. Biotropica 22, 160–171. Austad, S. (1984). Alternative male reproductive behaviors. Am. Zool. 24, 309–319. Baker, H. G., and Baker, I. (1983). Floral nectar constituents and pollinator type. In ‘‘Handbook of Experimental Pollination Ecology’’ (C. E. Jones and R. J. Little, Eds.), pp. 117–141. S and AE Science Publications, New York. Banda, H. J., and Paxton, R. J. (1991). Pollination of greenhouse tomatoes by bees. Acta Hort. 288, 194–198. Barthell, J. F., and Knops, J. M. H. (1997). Visitation of evening primrose by carpenter bees: Evidence of a ‘‘mixed’’ pollination syndrome Southwest Nat. 42, 86–93.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

447

Barthell, J. F., Hranitz, J. M., Thorp, R. W., and Shue, M. K. (2002). High temperature responses in two exotic leafcutting bee species: Megachile apicalis and M. rotundata (Hymenoptera: Megachilidae). Pan Pac. Entomol. 78, 235–246. Batra, S. W. T. (1978). Aggression, territoriality, mating and nest aggregation of some solitary bees (Hymenoptera: Halictidae, Megachilidae, Colletidae, Anthophoridae). J. Kans. Entomol. Soc. 51, 547–559. Batra, S. W. T. (1984). Solitary bees. Sci. Am. 250, 86–93. Batra, S. W. T. (1994). Anthophora pilipes villulosa SM (Hymenoptera, Anthophoridae), a manageable Japanese bee that visits blueberries and apples during cool rainy spring weather. Proc. Entomol. Soc. Wash. 96, 98–119. Bell, G. (1986). The evolution of empty flowers. J. Theor. Biol. 118, 253–258. Bellusci, S., and Marques, M. D. (2001). Circadian activity rhythm of the foragers of a eusocial bee (Scaptotrigona aff. depilis Hymenoptera, Apidae, Meliponinae) outside the nest. Biol. Rhythms Res. 32, 117–124. Ben-Shahar, Y., Leung, H. T., Pak, W. L., Sokolowski, M. B., and Robinson, G. E. (2003). cGMP-dependent changes in phototaxis: a possible role for the foraging gene in honey bee division of labour. J. Exp. Biol. 206, 2507–2515. Bennett, B., and Breed, M. D. (1985). The nesting biology, mating behavior and foraging ecology of Perdita opuntiae (Hymenoptera, Andrenidae). J. Kans. Ent. Soc. 58, 185–194. Bernhardt, P. (1996). Anther adaptation in animal pollination. In ‘‘The Anther: Form, Function and Phylogeny’’ (W. G. D’Arcy and R. C. Keating, Eds.), pp. 192–220. Cambridge University Press, Cambridge. Bertsch, A. (1983). Nectar production of Epilobium angustifolium L. at different air humidities: Nectar sugar in individual flowers and the optimal foraging theory. Oecologia 59, 40–48. Bertsch, A. (1984). Foraging in male bumble bees (Bombus lucorum L.): Maximising energy or minimising water load? Oecologia 62, 325–336. Biernaskie, J. M., Cartar, R. V., and Hurly, T. A. (2002). Risk-averse inflorescence departure in hummingbirds and bumble bees: Could plants benefit from variable nectar volumes? Oikos 98, 98–104. Biesmeijer, J. C., and Toth, E. (1998). Individual foraging, activity level and longevity in the stingless bee Melipona beecheii in Costa Rica (Hymenoptera, Apidae, Meliponinae). Insect Soc. 45, 427–443. Biesmeijer, J. C., Richter, J. A. P., Smeets, M. A. J. P., and Sommeijer, M. J. (1999). Niche differentiation in nectar-collecting stingless bees: The influence of morphology, floral choice, and interference competition. Ecol. Entomol. 24, 380–388. Bishop, J. A., and Armbruster, W. S. (1999). Thermoregulatory abilities of Alaskan bees: Effects of size, phylogeny and ecology. Funct. Ecol. 13, 711–724. Blarer, A., Keasar, T., and Shmida, A. (2002). Possible mechanisms for the formation of flower size preferences by foraging bumblebees. Ethology 108, 341–351. Bloch, G., and Robinson, G. E. (2001). Reversal of honeybee behavioural rhythms. Nature 410, 1048. Bloch, G., Toma, D. P., and Robinson, G. E. (2001). Behavioral rhythmicity, age, division of labor and period expression in the honey bee brain. J. Biol. Rhythms 16, 444–456. Breed, M. D., Stocker, E. M., Baumgartner, L. K., and Vargas, S. A. (2002). Time–place learning and the ecology of recruitment in a stingless bee, Trigona amalthea (Hymenoptera, Apidae). Apidologie 33, 251–258. Brink, D. (1982). A bonanza–blank pollinator reward schedule in Delphinium nelsonii (Ranunculaceae). Oecologia 52, 292–294.

448

P. G. WILLMER AND G. N. STONE

Brockmann, H. J. (1984). The evolution of social behaviour in insects. In ‘‘Behavioural Ecology, an Evolutionary Approach’’ (J. R. Krebs and N. B. Davies, Eds.), 2nd ed. pp. 340–361. Blackwell Scientific, Oxford. Brockmann, H. J. (2001). The evolution of alternative strategies and tactics. Adv. Study Behav. 30, 1–51. Buchmann, S. L., and Cane, J. H. (1989). Bees assess pollen returns while sonicating Solanum flowers. Oecologia 81, 289–294. Burgett, D. M., and Sukumalanand, P. (2000). Flight activity of Xylocopa (Nyctomelitta) tranquebarica: A night flying carpenter bee. J. Apicult. Res. 39, 75–83. Bu´rquez, A., and Corbet, S. A. (1991). Do flowers reabsorb nectar? Funct. Ecol. 5, 369–379. Calderone, N. W., and Johnson, B. R. (2002). The within-nest behaviour of honeybee pollen foragers in colonies with a high or low need for pollen. Anim. Behav. 63, 749–758. Cameron, S. A., and Ramirez, S. (2001). Nest architecture and nesting ecology of the orchid bee Eulaema meriana (Hymenoptera: Apinae: Euglossini). J. Kans. Entomol. Soc. 74, 142–165. Cameron, S. A., Whitfield, J. B., Hulslander, C. L., Cresko, W. A., Isenberg, S. B., and King, R. W. (1996). Nesting biology and foraging patterns of the solitary bee Melissodes rustica (Hymenoptera, Apidae) in northwest Arkansas. J. Kans. Entomol. Soc. 69, 260–273. Campan, R., and Lehrer, M. (2002). Discrimination of closed shapes by two species of bee, Apis mellifera and Megachile rotundata. J. Exp. Biol. 205, 559–572. Cane, J. H. (1991). Soils of ground-nesting bees (Hymenoptera: Apoidea): Texture, moisture, cell depth and climate. J. Kans. Entomol. Soc. 64, 406–413. Cane, J. H., and Payne, J. A. (1993). Regional, annual and seasonal variation in pollinator guilds: Intrinsic traits of bees (Hymenoptera, Apoidea) underlie their patterns of abundance at Vaccinium ashei (Ericaceae). Ann. Entomol. Soc. Am. 86, 577–588. Cane, J. H., Gerdin, S., and Wife, G. (1983a). Mandibular gland secretions of solitary bees: Potential for nest cell disinfection. J. Kans. Entomol. Soc. 56, 199–204. Cane, J. H., Eickwort, G. C., Wesley, F. R., and Spielholz, J. (1983b). Foraging, grooming and mate-seeking behaviors of Macropis nuda (Hymenoptera, Melittidae) and use of Lysimachia ciliata (Primulaceae) oils in larval provisions and cell linings. Am. Midl. Nat. 110, 257–264. Cane, J. H., Schiffhauer, D., and Kervin, L. J. (1996). Pollination, foraging and nesting ecology of the leaf cutting bee Megachile (Delomegachile) addenda (Hymenoptera: Megachilidae) on cranberry beds. Ann. Entomol. Soc. Am. 89, 361–367. Capaldi, E. A., Smith, A. D., Osborne, J. L., Fahrbach, S. E., Farris, S. M., Reynolds, D. R., Edwards, A. S., Martin, A., Robinson, G. E., Poppy, G. M., and Riley, J. R. (2000). Ontogeny of orientation flight in the honeybee revealed by harmonic radar. Nature 403, 527–540. Cartar, R. V. (1991). Colony energy requirements affect response to predation risk in foraging bumble bees. Ethology 87, 90–96. Cartar, R. V., and Abrahams, M. V. (1996). Risk-sensitive foraging in a patch departure context: A test with worker bumblebees. Am. Zool. 36, 447–458. Casey, T. M., May, M. L., and Morgan, K. R. (1985). Flight energetics of euglossine bees in relation to morphology and wing stroke frequency. J. Exp. Biol. 116, 271–289. Castellanos, M. D., Wilson, P., and Thomson, J. D. (2002). Dynamic nectar replenishment in flowers of Penstemon (Scrophulariaceae). Am. J. Bot. 89, 111–118. Cazier, M. A., and Linsley, E. G. (1963). Territorial behaviour among males of Protoxaea gloriosa (Colletidae). Can. Entomol. 94, 547–556. Chappell, M. A. (1982). Temperature regulation of carpenter bees (Xylocopa californica) foraging in the Colorado Desert of southern California. Physiol. Zool. 55, 267–280.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

449

Chappell, M. A. (1984). Temperature regulation and energetics of the solitary bee Centris pallida during foraging and intermale mate competition. Physiol. Zool. 57, 215–225. Charnov, E. L. (1982). ‘‘The Theory of Sex Allocation.’’ Princeton University Press, Princeton, NJ. Chittka, L. (1998). Sensorimotor learning in bumblebees: Long term retention and reversal training. J. Exp. Biol. 201, 515–524. Chittka, L., and Thomson, J. D. (1997). Sensorimotor learning and its relevance to task specialization in bumble bees. Behav. Ecol. Sociobiol. 41, 385–398. Coelho, J. R. (1991). Heat transfer and body temperature in honeybee (Hymenoptera: Apidae) drones and workers. Environ. Entomol. 20, 1627–1635. Collatz, K. G., and Sohal, R. S. (eds.) (1986). ‘‘Insect Ageing.’’ Springer-Verlag, Berlin. Collatz, K. G., and Wilps, H. (1986). Ageing of the flight mechanism. In ‘‘Insect Ageing’’ (K. G. Collatz and R. S. Sohal, Eds.), pp. 55–72. Springer-Verlag, Berlin. Collevatti, R. G., Campos, L. A. O., and Schoereder, J. H. (1997). Foraging behaviour of bee pollinators on the tropical weed Triumfetta semitriloba: Departure rules from flower patches. Insect Soc. 44, 345–352. Comba, L. (1999). Patch use by bumblebees (Hymenoptera, Apidae): Temperature, wind, flower density and traplining. Ethol. Ecol. Evol. 11, 243–264. Cooper, P. D., Schaffer, W. M., and Buchmann, S. L. (1985). Temperature regulation of honeybees (Apis mellifera) foraging in the Sonoran Desert. J. Exp. Biol. 114, 1–15. Corbet, S. A. (1990). Pollination and the weather. Isrl. J. Bot. 39, 13–30. Corbet, S. A., and Willmer, P. G. (1981). Pollination of the yellow passionfruit: Nectar, pollen and carpenter bees. J. Agric. Sci. 95, 655–666. Corbet, S. A., Willmer, P. G., Beament, J. W. L., Unwin, D. M., and Prys Jones, O. E. (1979). Post-secretory determinants of sugar concentration in nectar. Plant Cell Environ. 2, 293–300. Corbet, S. A., Kerslake, C. J. C., Brown, D., and Morland, N. E. (1984). Can bees select nectarrich flowers in a patch? J. Apicult. Res. 23, 293–308. Corbet, S. A., Fussell, M., Ake, R., Fraser, A., Gunson, C., Savage, A., and Smith, K. (1993). Temperature and the pollinating activity of social bees. Ecol. Entomol. 18, 17–30. Cox, M. D., and Myerscough, M. R. (2003). A flexible model of foraging by a honey bee colony: The effects of individual behaviour on foraging success. J. Theor. Biol. 223, 179–197. Crespi, B. J., and Yanega, D. (1995). The definition of eusociality. Behav. Ecol. 6, 109–115. Creswell, J. E. (1990). How and why do nectar-foraging bumblebees initiate movements between inflorescences of wild bergamot Monarda fistulosa (Lamiaceae)? Oecologia 82, 450–460. Cronin, A. L. (2001). Social flexibility in a primitively social allodapine bee (Hymenoptera: Apidae): Results of a translocation experiment. Oikos 94, 337–343. Cruden, R. W. (2000). Pollen grains: Why so many? In ‘‘Pollen and Pollination’’ (A. Dafni, M. Hesse, and E. Pacini, Eds.), pp. 143–165. Springer, Vienna, Austria. Cruden, R. W., Hermann, S. M., and Peterson, S. (1983). Patterns of nectar production and plant-pollinator coevolution. In ‘‘The Biology of Nectaries’’ (B. Bentley and T. Elias, Eds.), pp. 80–125. Columbia University Press, New York. Dafni, A. (1984). Mimicry and deception in pollination. Annu. Rev. Ecol. Syst. 15, 259–278. Dafni, A. (1996). Autumnal and winter pollination adaptations under Mediterranean conditions. Bocconea 5, 171–181. Dafni, A., and Firmage, D. (2000). Pollen viability and longevity: Practical, ecological and evolutionary implications. In ‘‘Pollen and Pollination’’ (A. Dafni, M. Hesse, and E. Pacini, Eds.), pp. 113–132. Springer, Vienna, Austria.

450

P. G. WILLMER AND G. N. STONE

Dafni, A., Eisikowitch, D., and Ivri, Y. (1987). Nectar flow and pollinator’s efficiency in two co-occurring species of Capparis (Capparaceae) in Israel. Plant Syst. Evol. 157, 181–186. Danforth, B. N. (1991). The morphology and behavior of dimorphic males in Perdita portalis (Hymenoptera, Andrenidae). Behav. Ecol. Sociobiol. 29, 235–247. Danforth, B. N. (1999). Emergence dynamics and bet hedging in a desert bee, Perdita portalis. Proc. R. Soc. B Biol. Sci. 266, 1985–1994. Danforth, B. N., and Neff, J. L. (1992). Male polymorphism and polyethism in Perdita texana (Hymenoptera, Andrenidae). Ann. Entonol. Soc. Am. 85, 616–626. Danforth, B. N., Conway, L., and Ji, S. Q. (2003a). Phylogeny of eusocial Lasioglossum reveals multiple losses of eusociality within a primitively eusocial clade of bees (Hymenoptera: Halictidae). Syst. Biol. 52, 23–36. Danforth, B. N., Ji, S. Q., and Ballard, L. J. (2003b). Gene flow and population structure in an oligolectic desert bee, Macrotera (Macroteropsis) portalis (Hymenoptera: Andrenidae). J. Kans. Entomol. Soc. 76, 221–235. Dobson, H. E. M. (1987). Role of flower and pollen aromas in host–plant recognition by solitary bees. Oecologia 72, 618–623. Dornhaus, A., Brockman, A., and Chittka, L. (2003). Bumble bees alert to food with pheromone from tergal gland. J. Comp. Physiol. A. 189, 47–51. Dornhaus, A., and Chittka, L. (1999). Evolutionary origin of bee dances. Nature 401, 38. Dornhaus, A., and Chittka, L. (2001). Food alert in bumblebees (Bombus terrestris): Possible mechanisms and evolutionary implications. Behav. Ecol. Sociobiol. 50, 570–576. Dressler, R. L. (1982). Biology of the orchid bees (Euglossini). Annu. Rev. Ecol. Syst. 13, 373–394. Dudley, R. (2000). ‘‘The Biomechanics of Insect Flight: Form, Function, Evolution.’’ Princeton University Press, Princeton, NJ. Dukas, R. (2001a). Effects of perceived danger on flower choice by bees. Ecol. Lett. 4, 327–333. Dukas, R. (2001b). Effects of predation risk on pollinators and plants. In ‘‘Cognitive Ecology of Pollination: Animal Behaviour and Floral Evolution’’ (J. D. Thomson and L. Chittka, Eds.), pp. 214–236. Cambridge University Press, Cambridge. Dukas, R., and Real, L. (1991). Learning foraging tasks by bees: A comparison between social and solitary species. Anim. Behav. 42, 269–276. Dukas, R., and Visscher, P. K. (1994). Lifetime learning by foraging honey-bees. Anim. Behav. 48, 1007–1012. Dunn, T., and Richards, M. H. (2003). When to bee social: Interactions among environmental constraints, incentives, guarding, and relatedness in a facultatively social carpenter bee. Behav. Ecol. 14, 417–424. Dyer, F. C., and Seeley, T. D. (1987). Interspecific comparison of endothermy in honeybees (Apis): Deviations from the expected size-related patterns. J. Exp. Biol. 58, 123–135. Eickwort, G. C. (1977). Male territorial behaviour in the mason bee Hoplitis anthocopoides (Hymenoptera; Megachilidae). Anim. Behav. 25, 542–554. Eickwort, G. C., and Ginsberg, H. S. (1980). Foraging and mating in the Apoidea. Annu. Rev. Entomol. 25, 421–426. El Shafie, H. A. F., Mogga, J. B. B., and Basedow, T. (2002). Studies on the possible competition for pollen between the honey bee Apis mellifera sudanensis and the imported honey bee Apis florea (Hym., Apidae) in North Khartoum (Sudan). J. Appl. Entomol. 126, 557–562. Ellington, C. P. (1985). Power and efficiency of insect flight muscle. J. Exp. Biol. 115, 293–304. Ellington, C. P. (1990). Oxygen consumption of bumblebees in forward flight. Nature 347, 472–473.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

451

Eltz, T., Whitten, W. M., Roubik, D. W., and Linsenmair, K. E. (1999). Fragrance collection, storage and accumulation by individual male bees. J. Chem. Ecol. 25, 157–176. Faegri, K., and van der Pijl, L. (1979). ‘‘Principles of Pollination Ecology.’’ Pergamon, Oxford. Fahrbach, S. E. (1997). Regulation of age polyethism in bees and wasps by juvenile hormone. Adv. Study Behav. 26, 285–310. Farina, W. M., and Wainselboim, A. J. (2001). Changes in the thoracic temperature of honeybees while receiving nectar from foragers collecting at different reward rates. J. Exp. Biol. 204, 1653–1658. Ferguson, A. W., and Free, J. B. (1979). Production of a forage-marking pheromone by the honeybee. J. Apicult. Res. 18, 128–135. Feuerbacher, E., Fewell, J. H., Roberts, S. P., Smith, E. F., and Harrison, J. F. (2003). Effects of load type (pollen or nectar) and load mass on hovering metabolic rate and mechanical power output in the honey bee Apis mellifera. J. Exp. Biol. 206, 1855–1865. Field, J. (1992). Intraspecific parasitism as an alternative reproductive tactic in nest-building wasps and bees. Biol. Rev. 67, 79–126. Frankie, G. W., and Vinson, S. B. (1977). Scent-marking of passion flowers in Texas by females of Xylocopa virginica texana (Hymenoptera, Anthophoridae). J. Kans. Entomol. Soc. 50, 613–625. Free, J. B. (1969). Influence of the odour of a honeybee colony’s stores on the behavior of its foragers. Nature 222, 778. Free, J. B. (1970). The flower constancy of bumblebees. J. Anim. Ecol. 39, 395–402. Free, J. B., and Williams, I. H. (1983). Scent-marking of flowers by honeybees. J. Apicult. Res. 22, 86–90. Fussell, M. (1992). Diurnal patterns of bee activity, flowering, and nectar reward per flower in tetraploid clover. N. Z. J. Agric. Res. 35, 151–156. Gathmann, A., and Tscharntke, T. (2002). Foraging ranges of solitary bees. J. Anim. Ecol. 71, 757–764. Gilbert, F. S., Haines, N., and Dickson, K. (1991). Empty flowers. Funct. Ecol. 5, 29–39. Gilbert, F. S., Azmeh, S., Barnard, C., Behnke, J., Collins, S. A., Hurst, J., and Shuker, D., et al. (2001). Individually recognizable scent marks on flowers made by a solitary bee. Anim. Behav. 61, 217–229. Gimenes, M., Benedito-Silva, A. A., and Marques, M. D. (1996). Circadian rhythm of pollen and nectar collection by bees on the flowers of Ludwigia elegans (Onagraceae). Biol. Rhythms Res. 27, 281–290. Giurfa, M., and Lehrer, M. (2001). Honeybee vision and floral displays: From detection to close-up recognition. In ‘‘Cognitive Ecology of Pollination: Animal Behaviour and Floral Evolution’’ (J. D. Thomson and L. Chittka, Eds.), pp. 61–82. Cambridge University Press, Cambridge. Giurfa, M., Nunez, J., Chittka, L., and Menzel, R. (1995). Color preferences of flowernaı¨ve honeybees. J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol. 177, 247–259. Gonzalez, A., Rowe, C. L., Weeks, P. J., Whittle, D., Gilbert, F. S., and Barnard, C. J. (1995). Flower choice by honey bees (Apis mellifera L.): Sex phase of flowers and preferences among nectar and pollen foragers. Oecologia 101, 258–264. Goodell, K. (2003). Food availability affects Osmia pumila (Hymenoptera, Megachilidae) foraging, reproduction and brood parasitism. Oecologia 134, 518–527. Gordon, D. M. (1996). The organization of work in social insect colonies. Nature 380, 121–124. Gould, J. L. (1987). Honey bees store learned flower-landing behaviour according to time of day. Anim. Behav. 35, 1579–1580.

452

P. G. WILLMER AND G. N. STONE

Goulson, D., Hawson, S. A., and Stout, J. C. (1998). Foraging bumblebees avoid flowers already visited by conspecifics or by other bumblebee species. Anim. Behav. 55, 199–206. Goulson, D., Chapman, J. W., and Hughes, W. O. H. (2001). Discrimination of unrewarding flowers by bees: Direct detection of rewards and use of repellent scent marks. J. Insect Behav. 14, 669–678. Goulson, D., Peat, J., Stout, J. C., Tucker, J., Darvill, B., Derwent, L. C., and Hughes, W. O. H. (2002). Can alloethism in workers of the bumblebee, Bombus terrestris, be explained in terms of foraging efficiency? Anim. Behav. 64, 125–130. Greggers, U., and Menzel, R. (1993). Memory dynamics and foraging strategies of honeybees. Behav. Ecol. Sociobiol. 32, 17–29. Gross, C. L. (1992). Floral traits and pollinator constancy: Foraging by native bees among three sympatric legumes. Aust. J. Ecol. 17, 67–74. Gumbert, A. (2000). Color choices by bumble bees (Bombus terrestris): Innate preferences and generalization after learning. Behav. Ecol. Sociobiol. 48, 36–43. Hammer, M. (1997). The neural basis of associative reward learning in honeybees. Trends Neurosci. 20, 245–252. Hammer, M., and Menzel, R. (1995). Learning and memory in the honeybee. J. Neurosci. 15, 1617–1630. Hanson, D. M., Olesen, J. M., and Jones, C. G. (2002). Trees, birds and bees in Mauritius: Exploitative competition between introduced honeybees and endemic nectarivorous birds? J. Biogeogr. 29, 721–734. Harder, L. D. (1986). Effects of nectar concentration and flower depth on flower handling efficiency of bumblebees. Oecologia 69, 309–315. Harder, L. D. (1988). Choice of individual flowers by bumblebees: Interactions of morphology, time and energy. Behaviour 104, 60–77. Harder, L. D. (1990a). Behavioural responses by bumblebees to variation in pollen availability. Oecologia 85, 41–47. Harder, L. D. (1990b). Pollen removal by bumblebees and its implications for pollen dispersal. Ecology 71, 1110–1125. Harrison, J. F., and Fewell, J. H. (2002). Environmental and genetic influences on flight metabolic rate in the honey bee, Apis mellifera. Comp. Biochem. Physiol. A. 133, 323–333. Hart, A. G., and Ratnieks, F. L. W. (2002). Task-partitioned nectar transfer in stingless bees; work organisation in a phylogenetic context. Ecol. Entomol. 27, 163–168. Hefetz, A., Fales, H. M., and Batra, S. W. T. (1979). Natural polyesters: Dufour’s gland macrocyclic lactones form brood cell laminesters in Colletes bees. Science 204, 415–417. Heinrich, B. (1972). Temperature regulation in the bumblebee Bombus vagans: A field study. Science 175, 183–187. Heinrich, B. (1973). The energetics of the bumble bee. Sci. Am. 228, 97–102. Heinrich, B. (1975). Thermoregulation in bumblebees. 2. Energetics of warm-up and free flight. J. Comp. Physiol. 96, 155–166. Heinrich, B. (1976). Heat exchange in relation to blood flow between thorax and abdomen in bumblebees. J. Exp. Biol. 64, 561–585. Heinrich, B. (1979a). Keeping a cool head: Honeybee thermoregulation. Science 205, 1269–1271. Heinrich, B. (1979b). ‘‘Majoring’’ and ‘‘minoring’’ by foraging bumblebees, Bombus vagans: An experimental study. Ecology 60, 245–255. Heinrich, B. (1981a). The mechanisms and energetics of honeybee swarm temperature regulation. J. Exp. Biol. 91, 25–55.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

453

Heinrich, B. (1981b). The regulation of temperature in the honeybee swarm. Sci. Am. 244, 146–160. Heinrich, B. (1993). ‘‘The Hot-Blooded Insects: Strategies and Mechanisms of Thermoregulation.’’ Harvard University Press, Cambridge, MA. Heinrich, B., and Buchmann, S. L. (1986). Thermoregulatory physiology of the carpenter bee, Xylocopa varipuncta.. J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 156, 557–562. Heinrich, B., and Heinrich, M. J. E. (1983). Heterothermia in foraging workers and drones of the bumblebee Bombus terricola. Physiol. Zool. 56, 563–567. Heinrich, B., and Raven, P. H. (1972). Energetics and pollination ecology. Science 176, 597–602. Herrera, C. M. (1990). Daily patterns of pollinator activity, differential pollinating effectiveness, and floral resource availability, in a summer-flowering Mediterranean shrub. Oikos 58, 277–288. Herrera, C. M. (1995). Floral biology, microclimate and pollination by ectothermic bees in an early-blooming herb. Ecology 76, 218–228. Hills, P. S. M., Wells, P. H., and Wells, H. (1997). Spontaneous flower constancy and learning in honey bees as a function of colour. Anim. Behav. 54, 615–627. Holldobler, B., and Wilson, E. O. (1990). ‘‘The Ants.’’ Harvard University Press, Cambridge, MA. Hopkins, M. J. G., Hopkins, H. C. F., and Sothers, C. A. (2000). Nocturnal pollination of Parkia velutina by Megalopta bees in Amazonia and its possible significance in the evolution of chiropterophily. J. Trop. Ecol. 16, 733–746. Hubbell, S. P., and Johnson, L. K. (1978). Comparative foraging behavior of six stingless bee species exploiting a standardized resource. Ecology 59, 1123–1136. Hurd, P. D., and Linsley, E. G. (1975). The principle Larraea bees of the southwestern United States. Smiths. Contrib. Zool. 193, 1–74. Hutchings, J. A., and Myers, R. A. (1994). The evolution of alternative mating strategies in variable environments. Evol. Ecol. 8, 256–268. Ichikawa, N., and Sasaki, M. (2003). Importance of social stimuli for the development of learning capability in honeybees. Appl. Entomol. Zool. 38, 203–209. Irwin, R. E. (2003). Impact of nectar robbing on estimates of pollen flow: Conceptual prediction and empirical outcomes. Ecology 84, 485–495. Irwin, R. E., Brody, A. K., and Waser, N. M. (2001). The impact of floral larceny on individuals, populations and communities. Oecologia 129, 161–168. Jakobsen, H. B., and Kristjansson, K. (1994). Influence of temperature and floret age on nectar secretion in Trifolium repens L. Ann. Bot. Lond. 74, 327–334. Jaycox, E. R. (1967). Territorial behavior among males of Anthidium banningense (Hymenoptera: Megachilidae). J. Kans. Entomol. Soc. 40, 565–570. Jaycox, E. R. (1970). Honeybee foraging behavior: Responses to queens, larvae, and extracts of larvae. Ann. Entomol. Soc. Am. 63, 844–847. Johnson, B. (1980). An electron microscopic study of flight muscle breakdown in an aphid Megoura viciae.. Tissue Cell 12, 529–539. Johnson, B. R. (2002). Reallocation of labor in honeybee colonies during heat stress: The relative roles of task switching and the activation of reserve labor. Behav. Ecol. Sociobiol. 51, 188–196. Johnson, L. K. (1982). Foraging strategies and the structure of stingless bee communities in Costa Rica. In ‘‘Social Insects in the Tropics’’ (P. Jaisson, Ed.), pp. 31–58. Universite Paris-Nord, Paris.

454

P. G. WILLMER AND G. N. STONE

Johnson, S. D., and Steiner, K. E. (2000). Generalisation versus specialisation in plant pollination systems. Trends Ecol. Evol. 15, 140–143. Josephson, R. K. (1981). Temperature and the mechanical performance of insect muscle. In ‘‘Insect Thermoregulation’’ (B. Heinrich, Ed.), pp. 19–44. John Wiley & Sons, New York. Josephson, R. K., Malamud, J. G., and Stokes, D. R. (2000a). Asynchronous muscle: A primer. J. Exp. Biol. 203, 2713–2722. Josephson, R. K., Malamud, J. G., and Stokes, D. R. (2000b). Power output by an asynchronous flight muscle from a beetle. J. Exp. Biol. 203, 2667–2689. Jungmann, R., Rothe, U., and Nachtigall, W. (1989). Flight of the honey bee. I. Thorax surface temperature and thermoregulation during tethered flight. J. Comp. Physiol. B. 158, 711–718. Kadmon, R., and Shmida, A. (1992). Departure rules used by bees foraging for nectar: A field test. Evol. Ecol. 6, 142–151. Kaiser, W. (1988). Busy bees need rest, too: Behavioral and electromyographical sleep signs in honeybees. J. Comp. Physiol. 163, 565–584. Kaiser, W. (1995). Rest at night in some solitary bees: A comparison with the sleep-like state of honey-bees. Apidologie 26, 213–230. Kaiser, W., and Steiner-Kaiser, J. (1983). Neuronal correlates of sleep, wakefulness and arousal in a diurnal insect. Nature 301, 707–709. Kapil, R. P., and Dhaliwal, J. S. (1969). Biology of Xylocopa. 2. Field activities, flight range and trials on transportation of nests. J. Res. Ludhiana 6, 262–271. Kearns, C. A., Inouye, D. W., and Waser, N. M. (1998). Endangered mutualisms: The conservation of plant–pollinator interactions. Annu. Rev. Ecol. Syst. 29, 83–112. Keasar, T., Motro, U., Shur, Y., and Shmida, A. (1996). Overnight memory retention of foraging skills by bumblebees is imperfect. Anim. Behav. 52, 95–104. Keijzer, C. J. (1999). Mechanisms of angiosperm anther dehiscence: A historical review. In ‘‘Anther and Pollen: From Biology to Biotechnology’’ (C. Clement, E. Pacini, and J. C. Audran, Eds.), pp. 54–67. Springer-Verlag, Berlin. Kerfoot, W. B. (1967). The lunar periodicity of Sphecodogastra, a nocturnal bee. Anim. Behav. 15, 479–486. Kevan, P. G. (1983). Floral colors through the insect eye: What they are and what they mean. In ‘‘Handbook of Experimental Pollination Ecology’’ (C. E. Jones and R. J. Little, Eds.), pp. 3–30. Van Nostrand Reinhold, New York. Kim, J. Y., and Thorp, R. W. (2001). Maternal investment and size–number tradeoff in a bee, Megachile apicalis, in seasonal environments. Oecologia 126, 451–456. Kimsey, L. S. (1980). The behaviour of male orchid bees (Apidae, Hymenoptera, Insecta) and the question of leks. Anim. Behav. 28, 996–1004. King, M. J., and Buchmann, S. L. (2003). Floral sonication by bees: Mesosomal vibration by Bombus and Xylocopa, but not Apis (Hymenoptera, Apidae), ejects pollen from poricidal anthers. J. Kans. Entomol. Soc. 76, 295–305. Koptur, S. (1992). Extra-floral nectary-mediated interactions between insects and plants. In ‘‘Insect–Plant Interactions’’ (E. A. Bernays, Ed.), Vol. 4, pp. 81–129. CRC Press, Boca Raton, FL. Kovac, H., and Schmaranzer, S. (1996). Thermoregulation of honeybees (Apis mellifera) foraging in spring and summer at different plants. J. Insect Physiol. 42, 1071–1076. Kremen, C., and Ricketts, T. (2000). Global perspectives on pollination disruptions. Conserv. Biol. 14, 1226–1228. Kronenberg, F., and Heller, H. C. (1982). Colonial thermoregulation in honeybees (Apis mellifera). J. Comp. Physiol. 148, 65–76.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

455

Kuhnholz, S., and Seeley, T. D. (1997). The control of water collection in honey bee colonies. Behav. Ecol. Sociobiol. 41, 407–422. Kukuk, P. F. (1996). Male dimorphism in Lasioglossum (Chilalictus) hemichalceum: The role of larval nutrition. J. Kans. Entomol. Soc. 69, 147–157. Larson, B. M. H. (1999). Bumblebees, Bombus spp., foraging on red oak, Quercus rubra, acorn galls in southern Ontario. Can. Field Nat. 113, 285–286. Larsson, F. K. (1991). Some take it cool, some like it hot: A comparative study of male matesearching tactics in two species of Hymenoptera (Colletidae and Sphecidae). J. Therm. Biol. 16, 45–51. Larsson, F. K., and Tengo¨, J. (1989). The effects of temperature and body size on the mating pattern of a gregariously nesting bee, Colletes cunicularius (Hymenoptera: Colletidae). Ecol. Entomol. 14, 279–286. Laverty, T. M. (1994). Bumble bee learning and flower morphology. Anim. Behav. 47, 531–545. Lehrer, M. (1999). Shape perception in the honeybee: Symmetry as a global framework. Int. J. Plant Sci. 160, S51–S65. Leys, R. (2000). Mate locating strategies of the green carpenter bee Xylocopa (Lestis) aeratus and X. (L.) bombylans. J. Zool. 252, 453–462. Linsley, E. G. (1958). The ecology of solitary bees. Hilgardia 27, 543–599. Linsley, E. G. (1960). Observations on some matinal bees at flowers of Cucurbita, Ipomoea and Datura in desert areas of New Mexico and southeastern Arizona. J. N. Y. Entomol. Soc. 68, 13–20. Linsley, E. G. (1962). The colletid Ptiloglossa arizonensis, a matinal pollinator of Solanum. Pan Pac. Entomol. 38, 75–82. Linsley, E. G. (1978). Temporal patterns of flower visitation by solitary bees, with particular reference to the southwestern United States. J. Kans. Entomol. Soc. 51, 531–546. Linsley, E. G., and Cazier, M. A. (1970). Some competitive relationships among matinal and late afternoon foraging activities of caupolocanine bees in southeastern Arizona. J. Kans. Entomol. Soc. 43, 251–261. Linsley, E. G., and MacSwain, J. W. (1959). Ethology of some Ranunculus insects with emphasis on competition for pollen. Univ. Calif. Publ. Entomol. 16, 1–46. Linsley, E. G., MacSwain, J. W., and Raven, P. H. (1963). Comparative behavior of bees and Onagraceae. 1. Oenothera bees of the Colorado Desert. 2. Oenothera bees of the Great Basin. Univ. Calif. Publ. Entomol. 33, 1–58. Lisci, M., Tanda, C., and Pacini, E. (1994). Pollination ecophysiology of Mercurialis annua L. (Euphorbiaceae), an anemophilous species flowering all year round. Ann. Bot. Lond. 74, 125–135. Lunau, K. (1996). Signalling functions of floral colour patterns for insect flower visitors. Zool. Anz. 235, 11–30. Lundberg, H. (1980). Effects of weather on foraging flights of bumblebees (Hymenoptera, Apidae) in a subalpine/alpine area. Holorctic Ecol. 3, 104–110. Luyt, R., and Johnson, S. D. (2002). Post-pollination nectar reabsorption and its implications for fruit quality in an epiphytic orchid. Biotropica 34, 442–446. MacSwain, J. W., Raven, P. H., and Thorp, R. W. (1973). Comparative behaviour of bees and Onagraceae 4. Clarkia bees of the western United States. Univ. Calif. Publ. Ent. 70, 1–80. Malo, J. E., and Baonza, J. (2002). Are there predictable clines in plant–pollinator interactions along altitudinal gradients? The example of Cytisus scoparius (L) in the Sierra de Guadarrama (Central Spain) Divers. Distr. 8, 365–371. Maloof, J. E. (2001). The effects of a bumble bee nectar robber on plant reproductive success and pollinator behavior. Am. J. Bot. 88, 1960–1965.

456

P. G. WILLMER AND G. N. STONE

Maloof, J. E., and Inouye, D. W. (2000). Are nectar robbers cheaters or mutualists? Ecology 81, 2651–2661. Matheson, A. Buchmann, S. L. O’Toole, C. Westrich, P. and Williams, I. H. (Eds.), ‘‘The Conservation of Bees,’’ Linnean Society Symposium Series 18. Academic Press London. May, M. L., and Casey, T. M. (1983). Thermoregulation and heat exchange in euglossine bees. Physiol. Zool. 56, 541–551. Mayer, D. F., and Miliczky, E. R. (1998). Emergence, male behavior and mating in the alkali bee Nomia melanderi Cokerell (Hymenoptera, Halictidae). J. Kans. Entomol. Soc. 71, 61–68. Meakin, P. J., and Roberts, J. A. (1990). Dehiscence of fruit in oilseed rape (Brassica napus L.). 1. Anatomy of pod dehiscence. J. Exp. Bot. 41, 995–1002. Menzel, R. (2001a). Searching for the memory trace in a mini-brain, the honeybee. Learn. Mem. 8, 53–62. Menzel, R. (2001b). Behavioural and neural mechanisms of learning and memory as determinants of flower constancy. In ‘‘Cognitive Ecology of Pollination: Animal Behavior and Floral Evolution’’ (J. D. Thomson and L. Chittka, Eds.), pp. 21–40. Cambridge University Press, Cambridge. Menzel, R., and Muller, U. (1996). Learning and memory in honeybees: From behaviour to neural substrates. Annu. Rev. Neurosci. 19, 379–404. Menzel, R., Geiger, K., Chittka, L., Joerges, J., Kunze, J., and Muller, U. (1996). The knowledge base of bee navigation. J. Exp. Biol. 199, 141–146. Michener, C. D. (1974). ‘‘The Social Behavior of the Bees.’’ Harvard University Press, Cambridge, MA. Michener, C. D. (1985). From solitary to eusocial: Need there be a series of intervening species? In ‘‘Experimental Behavioral Ecology and Sociobiology’’ (B. Holldobler and M. Lindauer, Eds.), pp. 293–306. Gustav Fischer Verlag, Stuttgart, Germany. Michener, C. D. (2000). ‘‘The Bees of the World.’’ Johns Hopkins University Press, Baltimore, MD. Michener, C. D., and Rettenmeyer, C. W. (1956). The ethology of Andrena erythronii with comparative data on other species (Andrenidae). Univ. Kans. Sci. Bull. 37, 645–684. Minckley, R. L., Cane, J. H., and Kervin, L. (2000). Origins and ecological consequences of pollen specialization among desert bees. Proc. R. Soc. B Biol. Sci. 267, 265–271. Miyake, T., and Yahara, T. (1999). Theoretical evaluation of pollen transfer by nocturnal and diurnal pollinators: When should a flower open? Oikos 86, 233–240. Moffatt, L. (2000). Changes in the metabolic rate of foraging honeybee: Effect of the carried weight or the reward rate? J. Comp. Physiol. A. 186, 299–306. Moffatt, L. (2001). Metabolic rate and thermal stability during honeybee foraging at different reward rates. J. Exp. Biol. 204, 759–766. Moffatt, L., and Nu´n˜ez, J. A. (1997). Oxygen consumption in the foraging honeybee depends on the reward rate at the food source. J. Comp. Physiol. B Biochem. Syst. Environ. Phys. 167, 36–42. Moller, A. P. (1995). Bumblebee preference for symmetrical flowers. Proc. Natl. Acad. Sci. USA 92, 2288–2292. Moore, D. (2001). Honey bee circadian clocks: Behavioural control from individual workers to whole-colony rhythms. J. Insect Physiol. 47, 843–857. Moore, D., Angel, J. E., Cheeseman, I. M., Fahrbach, S. E., and Robinson, G. E. (1998). Timekeeping in the honey bee colony: Integration of circadian rhythms and division of labor. Behav. Ecol. Sociobiol. 43, 147–160.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

457

Morandin, L. A., Laverty, T. M., and Kevan, P. G. (2001). Bumble bee (Hymenoptera: Apidae) activity and pollination levels in commercial tomato greenhouses. J. Econ. Entomol. 94, 462–467. Mordechai, B., Cohen, Y. R. M., Gerling, D., and Moskovitz, E. (1978). The biology of Xylocopa pubescens (Hymenoptera, Anthophoridae) in Israel. Isr. J. Entomol. 12, 107–121. Morse, D. H. (1986). Predation risk to insect foraging at flowers. Oikos 46, 223–228. Morse, D. H., and Fritz, R. S. (1983). Contributions of diurnal and nocturnal insects to the pollination of common milkweed (Asclepias syriaca L.) in a pollen-limited system. Oecologia 60, 190–197. Muller, A. (1996). Host plant specialization in western palearctic anthidiine bees (Hymenoptera: Apoidea: Megachilidae). Ecol. Monogr. 66, 235–264. Munster-Swendsen, M., and Calabuig, I. (2000). Interaction between the solitary bee Chelostoma florisomne and its nest parasite Sapyga clavicornis: Empty cells reduce the impact of parasites. Ecol. Entomol. 25, 63–70. Nachtigall, W. (1997). Territorial defence by male wool carder bees, Anthidium septemdentatum (Hymenoptera, Megachilidae) directed towards giant carpenter bees, Xylocopa violacea (Hymenoptera, Anthophoridae). Entomol. Gen. 22, 119–127. Nagamitsu, T., and Inoue, T. (1997). Aggressive foraging of social bees as a mechanism of floral resource partitioning in an Asian tropical forest. Oecologia 110, 432–439. Nansen, C., and Korie, S. (2000). Determining the time delay of honey bee (Apis mellifera) foraging response to hourly pollen release in a typical pollen flower (Cistus salvifolius). J. Apicult. Res. 39, 93–101. Navarro, L. (2000). Pollination ecology of Anthyllis vulneraria subsp vulgaris (Fabaceae): Nectar robbers as pollinators. Am. J. Bot. 87, 980–985. Neff, J. L. (2003). Nest and provisioning biology of the bee Panurginus polytrichus Cockerell (Hymenoptera: Andrenidae). J. Kans. Entomol. Soc. 76, 203–216. Neff, J. L., and Danforth, B. N. (1991). The nesting and foraging behavior of Perdita texana (Cresson) (Hymenoptera, Andrenidae). J. Kans. Entomol. Soc. 64, 394–405. Neff, J. L., and Larkin, L. L. (2002). Andrena chaparralensis (new species), a new vernal bee associated with Asteraceae on the South Texas plains (Hymenoptera; Apoidea: Andrenidae). J. Kans. Entomol. Soc. 75, 268–273. Neff, J. L., and Simpson, B. B. (1988). Vibratile pollen-harvesting by Megachile mendica Cresson (Hymenoptera, Megachilidae). J. Kans. Entomol. Soc. 61, 242–244. Neff, J. L., and Simpson, B. B. (1997). Nesting and foraging behavior of Andrena (Callandrena) rudbeckiae Robertson (Hymenoptera: Apoidea: Andrenidae) in Texas. J. Kans. Entomol. Soc. 70, 100–113. Nepi, M., Guarnieri, M., and Pacini, E. (2001). Nectar secretion, reabsorption and sugar composition in male and female flowers of Cucurbita pepo. Int. J. Plant Sci. 162, 353–358. Neukirch, A. (1982). Dependence of the life span of the honeybee (Apis mellifera) upon flight performance and energy consumption. J. Comp. Physiol. 146, 35–40. Nicolson, S. W. (1993). Low nectar concentrations in a dry atmosphere: A study of Grevillea robusta (Proteaceae) and Callistemon viminalis (Myrtaceae). S. Afr. J. Sci. 89, 473–477. Nicolson, S. W. (1995). Direct demonstration of nectar reabsorption in the flowers of Grevillea robusta (Proteaceae). Funct. Ecol. 9, 584–588. Nicolson, S. W. (1998). The importance of osmosis in nectar secretion and its consumption by insects. Am. Zool. 38, 418–425. Nicolson, S. W., and Louw, G. N. (1982). Simultaneous measurement of evaporative water loss, oxygen consumption, and thoracic temperature during flight in a carpenter bee. J. Exp. Zool. 222, 287–296.

458

P. G. WILLMER AND G. N. STONE

Nieh, J. C., Contrera, F. A. L., Ramirez, S., and Imperatriz-Fonseca, V. L. (2003). Variation in the ability to communicate three-dimensional resource location by stingless bess from different habitats. Anim. Behav. 66, 1129–1139. Noll, F. B. (1997). Foraging behaviour on carcasses in the necrophagic bee Trigona hypogea (Hymenoptera, Apidae). J. Insect Behav. 10, 463–467. Norden, B., Batra, S., Fales, H., Hefetz, A., and Shaw, G. (1980). Anthophora bees: Unusual glycerides from maternal Dufour’s glands serve as larval food and cell lining. Science 207, 1095–1097. Nuttman, C. V., and Willmer, P. G. (2003). How does insect visitation trigger floral colour change? Ecol. Entomol. 28, 467–474. Ono, M., Igarashi, T., Ohno, E., and Sasaki, M. (1995). Unusual thermal defence by a honeybee against mass attack by hornets. Nature 377, 334–336. Osborne, J. L., Clark, S. J., Morris, R. J., Williams, I. H., Riley, J. R., Smith, A. D., Reynolds, D. R., and Edwards, A. S. (1999). A landscape-scale study of bumble bee foraging range and constancy, using harmonic radar. J. Appl. Ecol. 36, 519–533. O’Toole, C., and Raw, A. (1991). ‘‘Bees of the World.’’ Sterling Publishing, New York. Pacini, E. (1994). Cell biology of anther and pollen development. In ‘‘Genetic Control of Self Incompatibility and Reproductive Development in Flowers’’ (E. G. Williams, R. B. Knox, and A. E. Clarke, Eds.), pp. 289–308. Kluwer, Dordrecht, The Netherlands. Pacini, E. (2000). From anther and pollen ripening to pollen presentation. In ‘‘Pollen and Pollination’’ (A. Dafni, M. Hesse, and E. Pacini, Eds.), pp. 19–43. Springer, Vienna, Austria. Pacini, E., Nepi, M., and Vesprini, J. L. (2003). Nectar biodiversity: A short review. Plant Syst. Evol. 238, 7–21. Page, R. E., Waddington, K. D., Hunt, G. J., and Fondrk, M. K. (1995). Genetic determinants of honeybee foraging behavior. Anim. Behav. 50, 1617–1625. Pais, M. S., and Figueiredo, A. C. S. (1994). Floral nectaries from Limodorum abortivum (L) and Epipactis atropurpurea Rafin (Orchidaceae): Ultractructural changes in plastids during the secretory process. Apidologie 25, 615–626. Paxton, R. J., and Tengo¨, J. (1996). Intranidal mating, emergence and sex ration in a communal bee Andrena jacobi Perkins 1921 (Hymenoptera: Andrenidae). J. Insect Behav. 9, 421–440. Paxton, R. J., Giovanetti, M., Andrietti, F., Scamoni, E., and Scanni, B. (1999). Mating in a communal bee, Andrena agilissima (Hymenoptera, Andrenidae). Ethol. Ecol. Evol. 11, 371–382. Pekkarinen, A., Berg, O., Calabuig, I., Janzon, L. A., and Luig, J. (2003). Distribution and coexistence of the Macropis species and their cleptoparasite Epeoloides coecutiens (Fabr.) in NW Europe (Hymenoptera: Apoidea, Melittidae and Apidae). Entomol. Fenn. 14, 53–59. Pellmyr, O. (1988). Bumblebees (Hymenoptera: Apidae) assess pollen availability in Anemonopsis macrophylla (Ranunculaceae) through floral shape. Ann. Entomol. Soc. Am. 81, 792–797. Pereboom, J. J. M., and Biesmeijer, J. C. (2003). Thermal constraints for stingless bee foragers: The importance of body size and coloration. Oecologia 137, 42–50. Perez, S. M., and Waddington, K. D. (1996). Carpenter bee (Xylocopa micans) risk indifference and a review of nectarivore risk-sensitivity studies. Am. Zool. 36, 435–446. Petanidou, T., and Smets, E. (1996). Does temperature stress induce nectar secretion in Mediterranean plants? New Phytol. 133, 513–518. Pleasants, J. M., and Zimmerman, M. L. (1983). The distribution of standing crop of nectar: what does it really tell us? Oecologia 57, 412–414.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

459

Possingham, H. P. (1988). A model of resource renewal and depletion: Applications to the distribution and abundance of nectar in flowers. Theor. Popul. Biol. 33, 138–160. Potts, S. G., and Willmer, P. G. (1997). Abiotic and biotic factors influencing nest-site selection by Halictus rubicundus, a ground-nesting halictine bee. Ecol. Entomol. 22, 319–328. Potts, S. G., and Willmer, P. G. (1998). Compact housing in built-up areas: Spatial positioning of nests in aggregations of a ground-nesting bee. Ecol. Entomol. 23, 427–432. Potts, S. G., Vulliamy, B., Dafni, A., Ne’eman, G., O’Toole, C., Roberts, S., and Willmer, P. G. (2003a). Response of plant–pollinator communities following fire: Changes in diversity, abundance and reward structure. Oikos 101, 103–112. Potts, S. G., Vulliamy, B., Dafni, A., Ne’eman, G., and Willmer, P. G. (2003b). Linking bees and flowers: How do floral communities structure pollinator communities? Ecology 84, 2628–2642. Raguso, R. A. (2001). Floral scent, olfaction, and scent-driven foraging behaviour. In ‘‘Cognitive Ecology of Pollination: Animal Behavior and Floral Evolution’’ (J. D. Thomson and L. Chittka, Eds.), pp. 83–105. Cambridge University Press, Cambridge. Raine, N. E., Willmer, P. G., and Stone, G. N. (2002). Spatial structuring and floral repellence prevent ant–pollinator conflict a Mexican ant acacia. Ecology 83, 3086–3096. Rasheed, S. A., and Harder, L. D. (1997). Foraging currencies for non-energetic resources: Pollen collection by bumblebees. Anim. Behav. 54, 911–926. Rao, S. P., Atluri, J. B., and Reddi, C. S. (2001). Intermittent mass blooming, midnight anthesis and rockbee pollination in Pterocarpus santalinus (Fabaceae). Nordic J. Bot. 21, 271–276. Rasheed, S. A., and Harder, L. D. (1997). Foraging currencies for non-energetic resources: Pollen collection by bumblebees. Anim. Behav. 54, 911–926. Raw, A. (1975). Territoriality and scent marking by Centris males in Jamaica (Hymenoptera: Anthophoridae). Behavior 54, 311–321. Real, L. (1992). Information processing and the evolutionary ecology of cognitive architecture. Am. Nat. 140, 108–145. Real, L., and Rathke, B. J. (1988). Patterns of individual variability in floral resources. Ecology 69, 728–735. Ribbands, C. R. (1953). ‘‘The Behaviour and Social Life of Honeybees.’’ Bee Research Association, London. Richards, M. H., and Packer, L. (1996). The socioecology of body size variation in the primitively eusocial sweat bee Halictus ligatus (Hymenoptera: Halictidae). Oikos 77, 68–76. Richards, M. H., von Wettberg, E. J., and Rutgers, A. C. (2003). A novel social polymorphism in a primitively eusocial bee. Proc. Natl. Acad. Sci. USA 100, 7175–7180. Roberts, R. B. (1971). Biology of the crepuscular bee Ptiloglossa guinnae (Hymenoptera: Colletidae) with notes on associated bees, mites and yeast. J. Kans. Entomol. Soc. 44, 283–294. Roberts, S. P., and Harrison, J. F. (1998). Mechanisms of thermoregulation in flying bees. Am. Zool. 38, 492–502. Roberts, S. P., and Harrison, J. F. (1999). Mechanisms of thermal stability during flight in the honeybee Apis mellifera. J. Exp. Biol. 202, 1523–1533. Roberts, S. P., Harrison, J. F., and Hadley, N. F. (1998). Mechanisms of thermal balance in flying Centris pallida (Hymenoptera: Anthophoridae). J. Exp. Biol. 201, 2321–2331. Robertson, A. W., Mountjoy, C., Faulkner, B. E., Roberts, M. V., and Macnair, M. R. (1999). Bumble bee selection of Mimulus guttatus flowers: The effect of pollen quality and reward depletion. Ecology 80, 2594–2606. Rosenheim, J. A., Nonacs, P., and Mangel, M. (1996). Sex ratios and multifaceted parental investment. Am. Nat. 148, 501–535.

460

P. G. WILLMER AND G. N. STONE

Roubik, D. W. (1989). ‘‘Ecology and Natural History of Tropical Bees.’’ Cambridge University Press, New York. Roubik, D. W. (1998). Grave-robbing by male Eulaema (Hymneoptera, Apidae): Implications for euglossine biology. J. Kans. Entomol. Soc. 71, 188–191. Roubik, D. W. (2000). Pollination system stability in tropical Amazonia. Conserv. Biol. 14, 1235–1236. Roubik, D. W., and Buchmann, S. L. (1984). Nectar selection by Melipona and Apis (Hymenoptera; Apidae) and the ecology of nectar intake by bee colonies in a tropical forest. Oecologia 61, 1–10. Roubik, D. W., and Michener, C. D. (1984). Nesting biology of Crawfordapis in Panama (Hymenoptera: Colletidae). J. Kans. Entomol. Soc. 57, 662–671. Roubik, D. W., and Peralta, F. J. A. (1983). Thermodynamics in nests of two Melipona species in Brazil. Acta Amaz. 13, 453–466. Roubik, D. W., and Wolda, H. (2001). Do competing honey bees matter? Dynamics and abundance of native bees before and after honey bee invasion. Popul. Ecol. 43, 53–62. Roubik, D. W., Yanega, D., Aluja, M., Buchmann, S. L., and Inouye, D. W. (1995). On optimal nectar foraging by some tropical bees (Hymenoptera, Apidae). Apidologie 26, 197–211. Roulston, T. H., and Cane, J. H. (2000a). The effect of diet breadth and nesting ecology on body size variation in bees (Apiformes). J. Kans. Entomol. Soc. 73, 129–142. Roulston, T. H., and Cane, J. H. (2000b). Pollen nutritional quality and digestibility for animals. In ‘‘Pollen and Pollination’’ (A. Dafni, M. Hesse, and E. Pacini, Eds.), pp. 187–209. Springer, Vienna, Austria. Sandoz, J. C., Laloi, D., Odoux, J. F., and Pham-Delegue, M. H. (2000). Olfactory information transfer in the honeybee: Compared efficiency of classical conditioning and early exposure. Anim. Behav. 59, 1025–1034. Santos, M. L., and Garofalo, C. A. (1994). Nesting biology and nest reuse of Eulaema nigrita (Hymenoptera, Apidae, Euglossini). Insect Soc. 41, 99–110. Schaffer, W. M., Jensen, D. B., Hobbs, D. E., Gurevitch, J., Todd, J. R., and Schaffer, M. V. (1979). Competition, foraging energetics and the cost of sociality in three species of bees. Ecology 60, 976–987. Schiestl, F. P., and Ayasse, M. (2000). Post-mating odor in females of the solitary bee Andrena nigroaenea (Apoidea, Andrenidae) inhibits male mating behaviour. Behav. Ecol. Sociobiol. 48, 303–307. Schmaranzer, S. (2000). Thermoregulation of water collecting honey bees (Apis mellifera). J. Insect Physiol. 46, 1187–1194. Schmid-Hempel, P., Kacelnik, A., and Houston, A. J. (1985). Honeybees maximise efficiency by not filling their crop. Behav. Ecol. Sociobiol. 17, 61–66. Schmid-Hempel, P., and Wolf, T. (1988). Foraging effort and life-span of workers in a social insect. J. Anim. Ecol. 57, 509–521. Schmidt, J. O. (1982). Pollen foraging preferences of honeybees. Southwest Nat. 7, 255–259. Schmidt, V. M., Zucchi, R., and Barth, F. G. (2003). A stingless bee marks the feeding site in addition to the scent path. Apidologie 34, 237–248. Schmidt, U., and Bertsch, A. (1990). Do foraging bumblebees scent-mark food sources, and does it matter? Oecologia 82, 137–144. Seeley, T. D., Kleinhenz, M., Bujok, B., and Tautz, J. (2003). Thorough warm-up before takeoff in honey bee swarms. Naturwissenschaffer 90, 256–260. Seger, J. (1983). Partial bivoltinism may cause alternating sex ratio biases that favor eusociality. Nature 301, 59–62. Seidelmann, K. (1999). The race for females: The mating system of the red mason bee, Osmia rufa (L.) (Hymenoptera: Megachilidae). J. Insect Behav. 12, 13–25.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

461

Sen Sarma, M., Fuchs, S., Werber, C., and Tautz, R. (2002). Worker piping triggers hissing for coordinated colony defence in the dwarf honeybee Apis florea. Zoology 105, 215–223. Severinghaus, L., Kurtak, B. H., and Eickwort, G. C. (1981). The reproductive behaviour of Anthidium manicatum (Hymenoptera: Megachilidae) and the significance of size for territorial males. Behav. Ecol. Sociobiol. 9, 51–58. Shafir, S., Bechar, A., and Weber, E. U. (2003). Cognition-mediated coevolution: Contextdependent evaluations and sensitivity of pollinators to variability in nectar rewards. Plant Syst. Evol. 238, 195–209. Shelly, T. E., and Villalobos, E. (2000). Buzzing bees (Hymneoptera: Apidae, Halicitidae) on Solanum (Solanaceae): Floral choice and handling time track pollen availability. Florida Entomol. 83, 1801–1887. Sherman, P. W., Lacey, E. A., Reeve, H. K., and Keller, L. (1995). The eusociality continuum. Behav. Ecol. 6, 102–108. Shmida, A., and Dukas, R. (1990). Progressive reduction in the mean body sizes of the solitary bees active during the flowering season and its correlation with the sizes of bee flowers of the mint family. Isr. J. Bot. 39, 133–141. Sihag, R. C. (1993). Behavior and ecology of the subtropical carpenter bee Xylocopa fenestrata F. 6. Foraging dynamics, crop hosts and pollination potential. J. Apicult. Res. 32, 94–101. Simmons, L. W., Tomkins, J. L., and Alcock, J. (2000). Can minor males of Dawson’s burrowing bee, Amegilla dawsonii (Hymenoptera, Anthophorinii), compensate for reduced access to virgin females through sperm competition? Behav. Ecol. 11, 319–325. Slaa, E. J., Wassenberg, J., and Biesmeijer, J. C. (2003). The use of field-based social information in eusocial foragers: Local enhancement among nestmates and heterospecifics in stingless bees. Ecol. Entomol. 28, 369–379. Small, E., Brookes, B., Lefkovitch, L. P., and Fairey, D. T. (1997). A preliminary analysis of the floral preferences of the alfalfa leafcutting bee, Megachile rotundata. Can. Field Nat. 111, 445–453. Smith, A. R., Wcislo, W. T., and O’Donnell, S. (2003). Assured fitness returns favour sociality in a mass-provisioning sweat bee, Megalopta genalis (Hymenoptera, Halictidae). Behav. Ecol. Sociobiol. 54, 14–21. Somanathan, H., and Borges, R. M. (2001). Nocturnal pollination by the carpenter bee Xylocopa tenuiscapa (Apidae) and the effect of floral display on fruit set of Heterphragma quadriloculare (Bignoniaceae) in India. Biotropica 33, 78–89. Soucy, S. L. (2002). Nesting biology and socially polymorphic behavior of the sweat bee Halictus rubicundus (Hymenoptera: Halictidae). Ann. Entomol. Soc. Am. 95, 57–65. Soucy, S. L., and Danforth, B. N. (2002). Phylogeography of the socially polymorphic sweat bee Halictus rubicundus (Hymenoptera, Halictidae). Evolution 56, 330–341. Southwick, E. E. (1990). The colony as a thermoregulating superorganisms. In ‘‘The Behavior and Physiology of Bees’’ (L. J. Goodman and R. C. Fisher, Eds.), pp. 28–47. CAB International, Wallingford, UK. Southwick, E. E., and Moritz, R. F. A. (1987). Social control of air ventilation in colonies of honey bees, Apis mellifera. J. Insect Physiol. 33, 623–626. Spaethe, J., and Weidenmuller, A. (2002). Size variation and foraging rate in bumblebees. Insect Soc. 49, 142–146. Srinavasan, M., and Zhang, S. W. (2003). Small brains, smart minds: Vision, perception and cognition in honeybees. IETE J. Res. 49, 127–134. Stabentheiner, A. (1991). Thermographic monitoring of the thermal behaviour of dancing bees. In ‘‘The Behavior and Physiology of Bees’’ (L. J. Goodman and R. C. Fisher, Eds.), pp. 89–101. CAB International, Oxford.

462

P. G. WILLMER AND G. N. STONE

Stabentheiner, A. (2001). Thermoregulation of dancing bees: Thoracic temperature of pollen and nectar foragers in relation to profitability of foraging and colony need. J. Insect Physiol. 47, 385–392. Stanghellini, M. S., Ambrose, J. T., and Schultheis, J. R. (2002). Diurnal activity, floral visitation and pollen deposition by honey bees and bumble bees on field-grown cucumber and watermelon. J. Apicult. Res. 41, 27–34. Stanley, R. G., and Linskens, H. F. (1974). ‘‘Pollen; Biology, Biochemistry, Management.’’ Springer Verlag, New York. Stark, R. E., Hefetz, A., Gerling, D., and Velthuis, H. H. W. (1990). Reproductive competition involving oophagy in the socially nesting bee Xylocopa sulcatipes. Naturwissenschaften 77, 38–40. Starks, P. T., and Reeve, H. K. (1999). Condition-dependent alternative reproductive tactics in the wool-carder bee, Anthidium manicatum. Ethol. Ecol. Evol. 11, 71–75. Steffan-Dewenter, I., and Kuhn, A. (2003). Honeybee foraging in differentially structured landscapes. Proc. R. Soc. B Biol. Sci. 270, 569–575. Steffan-Dewenter, I., and Tscharntke, T. (2000). Resource overlap and possible competition between honey bees and wild bees in central Europe. Oecologia 122, 288–296. Steiner, K. E., and Whitehead, V. B. (2002). Oil secretion and the pollination of Colpias mollis (Scrophulariaceae). Plant Syst. Evol. 235, 53–66. Stenstrom, M., and Bergmann, P. (1998). Bumblebees at an alpine site in northern Sweden: Temporal development, population size and plant utilization. Ecography 21, 306–316. Stern, D. L. (1991). Male territoriality and alternative male behaviors in the euglossine bee Eulaema meriana (Hymenoptera, Apidae). J. Kans. Entomol. Soc. 64, 421–437. Stern, D. L., and Dudley, R. (1991). Wing buzzing by male orchid bees Eulaema meriana (Hymenoptera, Apidae). J. Kans. Entomol. Soc. 64, 88–94. Stevenson, R. D., and Josephson, R. K. (1990). Effects of operating frequency and temperature on mechanical power output from moth flight muscle. J. Exp. Biol. 149, 61–78. Stone, G. N., Amos, J. N., Stone, T. F., Knight, R. J., Gay, H., and Parrott, F. (1988). Thermal effects on activity patterns and behavioural switching in a concourse of foragers on Stachytarpheta mutabilis (Verbenaceae) in Papua New Guinea. Oecologia 77, 56–63. Stone, G. N. (1989). Endothermy and thermoregulation in solitary bees. D. Phil. thesis. University of Oxford, Oxford. Stone, G. N. (1993a). Endothermy in the solitary bee Anthophora plumipes (Hymenoptera; Anthophoridae): Independent measures of thermoregulatory ability, costs of warm-up and the role of body size. J. Exp. Biol. 174, 299–320. Stone, G. N. (1993b). Thermoregulation in four species of tropical solitary bees: The roles of size, sex and altitude. J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 163, 317–326. Stone, G. N. (1994a). Patterns of evolution of warm-up rates and body temperatures in flight in solitary bees of the genus Anthophora. Funct. Ecol. 8, 324–335. Stone, G. N. (1994b). The roles of temperature, nectar supplies and body size in activity patterns and nest provisioning in the solitary bee Anthophora plumipes. Ecol. Entomol. 19, 177–189. Stone, G. N. (1995). Female foraging responses to harassment in the solitary bee Anthophora plumipes. Anim. Behav. 50, 405–412. Stone, G. N., and Willmer, P. G. (1989a). Endothermy and temperature regulation in bees: A critique of ‘‘grab-and-stab’’ measurement of body temperature J. Exp. Biol. 143, 211–223. Stone, G. N., and Willmer, P. G. (1989b). Warm-up rates and body temperatures in bees: The importance of body size, thermal regime and phylogeny. J. Exp. Biol. 147, 303–328.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

463

Stone, G. N., and Willmer, P. G. (1989c). Pollination of cardamom in Papua New Guinea. J. Apicult. Res. 28, 228–237. Stone, G. N., Loder, P. M. J., and Blackburn, T. M. (1995). Foraging and courtship behaviour in males of the solitary bee Anthophora plumipes (Hymenoptera: Anthophoridae): Thermal physiology and the roles of body size. Ecol. Entomol. 20, 169–183. Stone, G. N., Willmer, P. G., and Nee, S. (1996). Daily partitioning of pollinators in an Acacia community. Proc. R. Soc. B.Biol. Sci. 263, 1389–1393. Stone, G. N., Willmer, P. G., and Rowe, J. A. (1998). Partitioning of pollinators during flowering in an African Acacia community. Ecology 79, 2808–2827. Stone, G. N., Gilbert, F., Willmer, P. G., Potts, S., Semida, F., and Zalat, S. (1999). Windows of opportunity and the temporal structuring of foraging activity in a desert solitary bee. Ecol. Entomol. 24, 208–221. Stout, J. C., and Goulson, D. (2001). The use of conspecific and interspecific scent marks by foraging bumblebees and honeybees. Anim. Behav. 62, 183–189. Stout, J. C., and Goulson, D. (2002). The influence of nectar secretion rates on the responses of bumblebees (Bombus spp.) to previously visited flowers. Behav. Ecol. Sociobiol. 52, 239–246. Stout, J. C., Goulson, D., and Allen, J. A. (1998). Repellent scent marking of flowers by a guild of foraging bumblebees (Bombus spp.). Behav. Ecol. Sociobiol. 43, 317–326. Stpiczynska, M. (2003). Incorporation of 3H sucrose after the resorption of nectar from the spur of Platanthera chlorantha (Custer) Rchb. Can. J. Bot. 81, 927–932. Strickler, K. (1979). Specialization and foraging efficiency of solitary bees. Ecology 60, 998–1009. Stubbelefield, J. W., and Seger, J. (1990). Local mate competition with variable fecundity: Dependence of offspring sex ratios on information utilization and mode of male production. Behav. Ecol. 1, 68–80. Sugiura, N. (1991). Male territoriality and mating tactics in the wool-carder bee Anthidium septemspinosum Lepeletier (Hymenoptera, Megachilidae). J. Ethol. 9, 95–103. Sugiura, N. (1994). Parental investment and offspring sex ratio in a solitary bee, Anthidium septemspinosum Lepeletier (Hymenoptera, Megachilidae). J. Ethol. 12, 131–139. Sumpter, D. J. T., and Pratt, S. C. (2003). A modelling framework for understanding social insect foraging. Behav. Ecol. Sociobiol. 53, 131–144. Suttle, K. B. (2003). Pollinators as mediators of top-down effects on plants. Ecol. Lett. 6, 688–694. Tautz, J., Maier, S., Groh, C., Rossler, W., and Brockmann, A. (2003). Behavioral performance in adult honey bees is influenced by the temperature experienced during their pupal development. Proc. Natl. Acad. Sci. USA 100, 7343–7347. Telleria, M. C. (2000). Exploitation of pollen resources by Xylocopa splendidula in the Argentine pampas. J. Apicult. Res. 39, 55–60. Thakar, J. D., Kunte, K., Chauhan, A. K., Watve, A. V., and Watve, M. G. (2003). Nectarless flowers: Ecological correlates and evolutionary stability. Oecologia 136, 565–570. Thiyagesan, K., Varadharajan, M., and Abirami, S. (2001). Studies on risk-sensitive foraging in the Indian honey bee (Apis cerana indica). J. Apicult. Res. 40, 16–20. Thom, C. (2003). The tremble dance of honey bees can be caused by hive-external foraging experience. J. Exp. Biol. 206, 2111–2116. Thom, C., Gilley, D. C., and Tautz, J. (2003). Worker piping in honey bees (Apis mellifera): The behaviour of piping nectar foragers. Behav. Ecol. Sociobiol. 53, 199–205. Thomson, J. D. (1996). Trapline foraging by bumblebees. 1. Persistence of flight-path memory. Behav. Ecol. 7, 158–164.

464

P. G. WILLMER AND G. N. STONE

Thomson, J. D., and Thomson, B. A. (1992). Pollen presentation and viability schedules in animal-pollinated plants: Consequences for reproductive success. In ‘‘Ecology and Evolution of Plant Reproduction: New Approaches’’ (R. Wyatt, Ed.), pp. 1–24. Chapman & Hill, New York. Thornhill, R., and Alcock, J. (1983). ‘‘The Evolution of Insect Mating Systems.’’ Harvard University Press, Cambridge Mass. Toma, D. P., Bloch, G., Moore, D., and Robinson, G. E. (2000). Changes in period mRNA levels in the brain and division of labor in honey bee colonies. Proc. Natl. Acad. Sci. USA 97, 6914–6919. Torchio, P. F. (1975). The biology of Perdita nuda and descriptions of its immature forms and those of its Sphecodes parasite. J. Kans. Ent. Soc. 48, 257–279. Torchio, P. F. (1991). Bees as crop pollinators and the role of solitary species in changing environments. Acta Hort. 288, 49–61. Underwood, B. A. (1991). Thermoregulation and energetic decision making by the honeybees Apis cerana, Apis dorsata and Apis laboriosa. J. Exp. Biol. 157, 19–34. Vaknin, Y., Gan-Mor, S., Bechar, A., Ronen, B., and Eisikowitch, D. (2000). The role of electrostatic forces in pollination. In ‘‘Pollen and Pollination’’ (A. Dafni, M. Hesse, and E. Pacini, Eds.), pp. 133–142. Springer, Vienna, Austria. Vaughan, D. M., and Calderone, N. W. (2002). Assessment of pollen stores by foragers in colonies of the honey bee, Apis mellifera L. Insect Soc. 49, 23–27. Vesprini, J. L., and Pacini, E. (2000). Breeding systems in two species of the genus Helleborus (Ranunculaceae). Plant Biosyst. 134, 193–197. Vesprini, J. L., Nepi, M., and Pacini, E. (1999). Nectary structure, nectar secretion patterns and nectar composition in two Helleborus species. Plant Biol. 1, 560–568. Vicens, N., and Bosch, J. (2000). Weather-dependent pollinator activity in an apple orchard, with special reference to Osmia cornuta and Apis mellifera (Hymenoptera: Megachilidae and Apidae). Environ. Entomol. 29, 413–420. Villalobos, E. M., and Shelley, T. E. (1991). Correlates of male mating success in two species of Anthidium bees (Hymenoptera, Megachilidae). Behav. Ecol. Sociobiol. 29, 47–53. Villalobos, E. M., and Shelley, T. E. (1996). Temporal and spatial variation in the foraging behavior of honey bees (Hymenoptera: Apidae) at Chinese violets. Florida Entomol. 79, 398–407. Visscher, P. K., and Dukas, R. (1997). Survivorship of foraging honey bees. Insect Soc. 44, 1–5. Vogt, F. D. (1986a). Thermoregulation in bumblebee colonies. 1. Thermoregulation versus brood maintenance behaviors during acute changes in ambient temperature. Physiol. Zool. 59, 55–59. Vogt, F. D. (1986b). Thermoregulation in bumblebee colonies. 2. Behavioural and demographic variation throughout the colony cycle. Physiol. Zool. 59, 60–68. Waddington, K. D. (1990). Foraging profits and thoracic temperature of honey bees (Apis mellifera). J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 160, 325–329. Waddington, K. D. (2001). Subjective evaluation and choice behaviour by nectar- and pollencollecting bees. In ‘‘Cognitive Ecology of Pollination: Animal Behaviour and Floral Evolution’’ (J. D. Thomson and L. Chittka, Eds.), pp. 41–60. Cambridge University Press, Cambridge. Wainselboim, A. J., Roces, F., and Farina, W. M. (2002). Honeybees assess changes in nectar flow within a single foraging bout. Anim. Behav. 63, 1–6. Walther-Hellwig, K., and Frankl, R. (2000). Foraging distances of Bombus muscorum, Bombus lapidarius, and Bombus terrestris (Hymenoptera, Apidae). J. Insect Behav. 13, 239–246.

DETERMINANTS OF ACTIVITY PATTERNS OF BEES

465

Warrant, E., Porombka, T., and Kirchner, W. H. (1996). Neural image enhancement allows honeybees to see at night. Proc. R. Soc. B Biol. Sci. 263, 1521–1526. Waser, N. M. (1983). Competition for pollination and floral character differences among sympatric plant species: A review of the evidence. In ‘‘Handbook of Experimental Pollination Biology’’ (C. E. Jones and R. J. Little, Eds.), pp. 277–293. Van Nostrand Reinhold, New York. Waser, N. M., Chittka, L., Price, M. V., Williams, N. M., and Ollerton, J. (1996). Generalization in pollination systems and why it matters. Ecology 77, 1043–1060. Watmough, R. H. (1983). Mortality, sex-ratio and fecundity in natural populations of large carpenter bees (Xycolopa spp.). J. Anim. Ecol. 52, 111–125. Wcislo, W. T. (1992). Attraction and learning in mate-finding by solitary bees, Lasioglossum figueresi Wcislo and Nomia triangulifera Vachal (Hymenoptera, Halictidae). Behav. Ecol. Sociobiol. 31, 139–148. Wcislo, W. T., and Cane, J. H. (1996). Floral resource utilization by solitary bees (Hymenoptera, Apoidea) and exploitation of their stored foods by natural enemies. Annu. Rev. Entomol. 41, 257–286. Wcislo, W. T., and Schatz, B. (2003). Predator recognition and evasive behavior by sweat bees, Lasioglossum umbripenne. (Hymenoptera: Halictidae), in response to predation by ants Ectatomma ruidum (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 53, 182–189. Wei, C. A., Rafalko, S. L., and Dyer, F. C. (2002). Deciding to learn: Modulation of learning flights in honeybees, Apis mellifera. J. Comp. Physiol. A. 188, 725–737. Weiss, M. (1995). Floral color change: A widespread functional convergence. Am. J. Bot. 82, 167–185. Westerkamp, C. (1996). Pollen in bee–flower relations: Some considerations on melittophily. Bot. Acta 109, 325–332. Wetherwax, P. B. (1986). Why do honey bees reject certain flowers? Oecologia 69, 567–570. Wiegmann, D. D., Wiegmann, D. A., and Waldron, F. A. (2003). Effects of a reward downshift on the consummatory behavior and flower choices of bumblebee foragers. Physiol. Behav. 79, 561–566. Williams, C. S. (1998). The identity of the previous visitor influences flower rejection by nectar-collecting bees. Anim. Behav. 56, 673–681. Williams, N. M. (2003). Use of novel pollen species by specialist and generalist solitary bees (Hymenoptera: Megachilidae). Oecologia 134, 228–237. Williams, N. M., and Tepedino, V. J. (2003). Consistent mixing of near and distant resources in foraging bouts by the solitary mason bee Osmia lignaria. Behav. Ecol. 14, 141–149. Willis, D. S., and Kevan, P. G. (1995). Foraging dynamics of Peponapis pruinosa (Hymenoptera, Anthophoridae) on pumpkin (Cucurbita pepo) in southern Ontario. Can. Entomol. 127, 167–175. Willmer, P. G. (1980). The effects of insect visitors on nectar constituents in temperate plants. Oecologia 47, 270–277. Willmer, P. G. (1983). Thermal constraints on activity patterns in nectar-feeding insects. Ecol. Entomol. 8, 455–469. Willmer, P. G. (1985a). Size effects on the hygrothermal balance and foraging patterns of a sphecid wasp, Cerceris arenaria. Ecol. Entomol. 10, 469–479. Willmer, P. G. (1985b). Thermal ecology, size effects and the origins of social behaviour in Cerceris wasps. Behav. Ecol. Sociobiol. 17, 151–180. Willmer, P. G. (1986). Foraging patterns and water balance: Problems of optimisation for a xerophilic bee, Chalicodoma sicula. J. Anim. Ecol. 55, 941–962. Willmer, P. G. (1988). The role of insect water balance in pollination ecology: Xylocopa and Calotropis. Oecologia 76, 430–438.

466

P. G. WILLMER AND G. N. STONE

Willmer, P. G. (1991a). Constraints on foraging by solitary bees. In ‘‘The Behaviour and Physiology of Bees’’ (L. J. Goodman and R. C. Fisher, Eds.), pp. 131–148. CAB International, Wallingford, UK. Willmer, P. G. (1991b). Thermal biology and mate acquisition in ectotherms. Trends Ecol. Evol. 6, 396–399. Willmer, P. G., and Corbet, S. A. (1981). Temporal and microclimatic partitioning of the floral resources of Justicia aurea amongst a concourse of pollen vectors and nectar robbers. Oecologia 51, 67–78. Willmer, P. G., and Stone, G. N. (1988). Pollination of lowland coffee (Coffea canephora): Costs and benefits of leaf cutter bees in Papua New Guinea. Entomol. Exp. Appl. 50, 113–124. Willmer, P. G., and Stone, G. N. (1997a). The physiological ecology of desert bees. J. Therm. Biol. 22, 453–465. Willmer, P. G., and Stone, G. N. (1997b). Ant deterrence in Acacia flowers: How aggressive ant-guards assist seed-set. Nature 388, 165–167. Willmer, P. G., and Unwin, D. (1981). Field analysis of insect heat budgets: Reflectance, size and heating rates. Oecologia 50, 250–255. Willmer, P. G., Gilbert, F. S., Ghazoul, J., Zalat, S., and Semida, F. (1994). A novel form of territoriality: Daily paternal investment in an anthophorid bee. Anim. Behav. 48, 535–549. Willmer, P. G., Stone, G. N., and Johnston, I. A. (2000). ‘‘Environmental Physiology of Animals.’’ Blackwell Scientific, Oxford. Wilson, E. O. (1971). ‘‘The Insect Societies.’’ Harvard University Press, Cambridge, MA. Wirtz, P., Szabados, M., Pethig, H., and Plant, J. (1988). An extreme case of interspecific territoriality: Male Anthidium manicatum (Hymenoptera, Megachilidae) wound and kill intruders. Ethology 78, 159–167. Wirtz, P., Kopka, S., and Schmoll, G. (1992). Phenology of two territorial solitary bees, Anthidium manicatum and A. florentinum (Hymenoptera, Megachilidae). J. Zool. 228, 641–651. Wolda, H., and Roubik, D. W. (1986). Nocturnal bee abundance and seasonal bee activity in a Panamanian forest. Ecology 67, 426–433. Wolf, T. J., Ellington, C. P., and Begley, I. S. (1999). Foraging costs in bumblebees: Field conditions cause large individual differences. Insect Soc. 46, 291–295. Wuellner, C. T. (1999a). Nest site preference and success in a gregarious, ground-nesting bee, Dieunomia triangulifera. Ecol. Entomol. 24, 471–479. Wuellner, C. T. (1999b). Alternative reproductive strategies of a gregarious ground-nesting bee, Dieunomia traingulifera (Hymenoptera, Halictidae). J. Insect Behav. 12, 845–863.

Index

A A. bruennichi. See Araneus bruennichi; Argiope bruennichi A. diadematus. See Araneus diadematus A. keyserlingi. See Argiope keyserlingi Acacia, 348, 421, 424, 425, 429–430 Acanthochromis polyacanthus, androgen responsiveness in, 181 Acanthomyrmex ferox, reproductive conflict in, 21 Activational effects, social modulation of androgens with, 198–203 Active avoidance, infant learning with, 107 Agapostemon, 348, 420 Agelaius phoeniceus, 196 Aggression challenge hypothesis on, 196 insect reproductive conflict with, 2, 10, 37–38 Aggressive competition, food transfer with, 268 AL. See Antennal lobe Alouatta palliata, food transfer with, 273 Amegilla, 348, 365, 368, 393, 395 Amegilla sapiens, thermal effects of body size of, 360 Amygdala, infant learning with, 109, 114–115, 122 Andrena activity patterns of, 348, 362, 380, 394–395, 398 male activity with, 380 Andrena agilissima, 382 Andrena omninigra clarkiae, 405

Andrena raveni, 404 Andrena rudbeckiae, 382, 388 Andrenidae, 348, 364 Androgen action with, 174 activational effects with, 198–203 adaptive significance of, 197–206 Agelaius phoeniceus in, 196 alternative reproductive tactics with, 205–206 androgen receptors with, 171 behavioral endocrinology in, 166–169 behavioral feedback background in, 175–177 behavioral feedback on endocrine function with, 175–188 benefits from high levels of, 207, 208–209 Betta splendens in, 194 breeding density with, 187 Bubulcus ibis in, 196 bystander effects with, 202 castration and, 167–168 causal agents of behavior in, 166–175 causal behavior background in, 175–177 central actions of androgens and, 171–174 challenge hypothesis with, 177–188 cichlids in, 180–181 cognitive performance with, 173–174 Coolidge effect with, 176 cost-benefit analysis for, 207–214 costs from high levels of, 207, 209–214 467

468 Androgen (continued ) Coturnix japonica in, 191 dear enemy effects with, 199–200 deterministic factors of, 169–170 estrogen receptors with, 171 eunuchs and, 167 evolution of, 206–216 evolutionary scenarios for selection with, 214–216 Falco sparverius with, 196 Fulica americana with, 197 Haplochromis burtoni with, 189 historical background on, 166–169, 175–177 HPA axis with, 201 hypothalamus-pituitary-gonadal axis with, 189–190 immunosuppression with, 207, 209–212 increased energy consumption with, 207, 209, 210 introduction to, 165–166, 167 Ketotestosterone/testosterone ratio with, 194 Lamprologus callipterus with, 181, 182 Larus fuscus with, 196 Larus ridibundus with, 196 Lemur catta with, 217 life history trade-offs with, 203–204 long-term effects with, 203–206 luteinizing hormone with, 171 Macaca mulatta with, 217 mating success with, 187–188 mating system with, 179–185 mechanisms of, 170–175 Melospiza melodia with, 183 men with, 216–218 Microtus ochrogaster with, 213 model for, 167 motivational systems with, 171–173 Neolamprologus pulcher with, 181, 182 neuroendocrine mechanisms for, 189–192 neuromodulators with, 169–170

INDEX

ontogeny of, 195–197 Oreochromis mossambicus with, 181 organizational effects with, 203–206 Pan troglodytes with, 217 parental care with, 207, 209, 211, 212–213 Passer domesticus with, 196–197 perception with, 174 peripheral actions of, 174–175 Peromyscus californicus with, 213 phenotypic plasticity with, 203–204 Phodopus campbelli with, 213 predation risk with, 207, 209 proximate mechanisms for, 188–195 Pseudosimochromis curvifrons with, 181, 182 psychological mechanisms for, 192–195 rejuvenation hypothesis with, 169 responsiveness with, 179–185 Saguinus oedipus with, 213 Sarotherodon galilaeus with, 183, 184 Sarotherodon melanotheron with, 180–181 Sarotherodon occidentalis with, 180–181 Serinus canaria with, 191, 196 sex change with, 204–205 short-term effects with, 198–203 social interactions influencing levels of, 177–188 social modulation of, 165–219 social stability with, 185 social status with, 186–187 spermatogenesis with, 207 Steinach operation with, 169 Sterna hirundo with, 196 Sturnus vulgaris with, 183, 191 Sula granti with, 196 Sula nebouxii with, 196 summary of, 218–219 Tachycineta bicolor with, 197 territoriality effects with, 198–199

INDEX

territoriality with, 186–187 Tinbergen legacy with, 188 Tropheus moorii with, 181, 182 Vulpes vulpes with, 206 winner-loser effects with, 200–202 Zonotrichia leucophrys with, 183 Androgen receptors (AR), social modulation with, 171 Ant protection mutualisms, game structures with, 78–80, 87 Antennal lobe (AL) honeybees with, 251–252, 253, 255, 257, 258 olfactory processing with, 251–252, 253, 255, 257, 258 Antennation, insect reproductive conflict with, 10–11, 17, 35–36 Anterior olfactory nucleus (AON), infant learning with, 109, 111–113, 121 Anthidiellum breviusculum, activity pattern of, 363 Anthidium, 348, 362, 365 Anthidium manicatum, 365, 385, 386 Anthophora, 348, 362, 364, 365, 368, 394–396, 400, 404–405, 415, 416, 421 Anthophora crassipes, activity pattern of, 363 Anthophora neglecta, 404 Anthophora pauperata, 385, 387–388, 422, 423, 431 Anthophora plumipes activity pattern of, 365, 370–371, 372, 380–382, 387, 389, 390, 392, 400 kleptoparasites influencing, 392 male strategies in, 380–382, 387, 389, 390 nectar foraging in, 381–382 nest behavior of, 370–371, 372 sexual harassment in, 389, 390 Anthophora quadrifasciata, activity pattern of, 363

469

Ants pheromones in, 13 reproductive conflict with, 1, 8, 9, 13, 22, 28 AON. See Anterior olfactory nucleus Aotus trivirgatus, food transfer with, 273, 274 Aphelocoma coerulescens, food transfer with, 283 Apis activity pattern in, 348, 362, 395–396, 401, 423, 425, 433, 434, 436, 440, 443 learning in, 440 odor cues with, 443 Apis cerana, 443 Apis dorsata, 366 Apis florea, 429, 436, 438, 443 Apis mellifera, 393, 429, 438, 443 queen polyandry with, 5 reproductive conflict in, 5, 7, 17, 19 AR. See Androgen receptors Araneus bruennichi, sexual cannibalism in, 145, 146 Araneus diadematus, sexual cannibalism in, 139, 144, 145, 147, 158 Argiope aemula, sexual cannibalism in, 137 Argiope argentata, sexual cannibalism in, 149 Argiope bruennichi, sexual cannibalism in, 139, 144, 154–155 Argiope keyserlingi, sexual cannibalism in, 137, 144, 148, 150, 152, 154, 156 Associative learning, Pavlovian conditioning in, 244 Associative theories, Pavlovian conditioning with, 243 Ateles geoffroyi, food transfer with, 273 Attachment, infant learning/memory with, 103–104

470

INDEX

B Baryphyma pratense, sexual cannibalism in, 136 Bees. See also Bumblebee; Honeybees abiotic conditions with, 392–401 Acacia genera of, 348, 421, 424, 425, 429–430 activity pattern shapes with, 404–406 activity patterns of, 347–446 Agapostemon genera of, 348, 420 age-related flight performance with, 368–369 Amegilla genera of, 348, 365, 368, 393, 395 Amegilla sapiens, 360 Andrena agilissima, 382 Andrena genera of, 348, 362, 380, 394–395, 398 Andrena omninigra clarkiae, 405 Andrena raveni, 404 Andrena rudbeckiae, 382, 388 Andrenidae species of, 348, 364 ant guards with, 429 Anthidiellum breviusculum, 363 Anthidium genera of, 348, 362, 365 Anthidium manicatum, 365, 385, 386 Anthophora crassipes, 363 Anthophora genera of, 348, 362, 364, 365, 368, 394–396, 400, 404–405, 415, 416, 421 Anthophora neglecta, 404 Anthophora pauperata, 385, 387–388, 388, 422, 423, 431 Anthophora plumipes, 365, 370–371, 372, 380–382, 387, 389, 390, 392, 400 Anthophora quadrifasciata, 363 Apis cerana, 443 Apis dorsata, 366 Apis florea, 429, 436, 438, 443 Apis genera of, 348, 362, 393, 395–396, 401, 423, 425, 433, 434, 436, 438, 440, 443 Apis mellifera, 393, 429, 438, 443

association with, 440 biogeographical variation in thermoregulation in, 364 Bombus alpinus, 421 Bombus genera of, 348, 362, 364, 365, 393, 395, 400, 402, 421, 427, 433, 434, 438, 442 Bombus lapidarius, 438 Bombus terrestris, 427 carpenter type of, 369, 370, 386, 434 Caupolicana genera of, 348, 394–395, 400 Centris pallida, 365, 387 Ceratina cyanea, 363 Ceratina genera of, 348, 370, 395, 434 Chalicodoma genera of, 348, 368, 405 Chalicodoma sicula, 373, 376, 377 Chelostoma florisomne, 348, 368, 378 circadian rhythms with, 368 climatic dependence with, 438 Coelioxys kleptoparasite of, 365 coevolution with, 431–432 Colletes genera of, 348, 362, 380 Colletidae species of, 348, 364 communication with, 442–445 competition with, 425–431 competitive mating tactics of, 382–383 conditioning of, 440 Creightonella frontalis, 374, 375, 383, 384 Creightonella genera of, 348, 365, 393, 406 defense polygyny mating tactics of, 383–385 Dieunomia triangulifera, 378 digger type of, 369 Echium vulgare for, 410 egg laying of, 376–378 environment-related variation in thermoregulation in, 364 Epeolus genera of, 348, 368 Eucera genera of, 348, 416

INDEX

Eulaema genera of, 348, 389 Eulaema imperialis, 386 Eulaema meriana, 386 eusocial bees as, 434, 435–445 eusocial bees v., 435–436 extrinsic factors with, 392–432, 433 female nesting cycle with, 369–379 flight activity in, 366–367 flight/interception mating tactics of, 382 flight muscle performance in, 352 floral constancy with, 438–442 floral resource availabilities with, 401–403 flower choice with, 444–445 flower specialization in, 353–354 Halictus genera of, 348, 380, 398 Halictus ligatus, 435 Halictus rubicundus, 435 Heterosarellus genera of, 348, 423 heterothermy in, 354–366 hilltopping of, 386 intrinsic factors of, 352–369 introduction to activity patterns of, 347–352 kleptoparasite of, 365, 370, 391–392 Lasioglossum genera of, 348, 395, 398, 434 Lasioglossum hemichalceum, 382 lateral wing buzzing in, 358 Lavendula stoechas with, 363 learning effect with, 440 learning with, 438–442 light levels with, 366–367 male strategies in, 379–389 males impact on female, 389–391 mason type of, 369–370, 371, 376 mating system diversity with, 379–389 Megachile apicalis, 376–377 Megachile genera of, 348, 370, 380, 395 Megachile nana, 393, 431

471 Megachile rotundata, 417 Megachilidae species of, 348, 364 Megalopta genera of, 348, 366 Melecta albifrons, 392 Melecta kleptoparasite of, 365, 391, 392 Melipona beecheii, 425 Melipona genera of, 348, 434 Melissodes genera of, 348, 375 melittophily syndrome with, 410 memory development/duration with, 441–442 miner type of, 369 navigation with, 440 nectar as water reward for, 412–413 nectar availabilities with, 406–417 nectar balance of, 436–437 nectar concentration for, 408–411 nectar effects overview for, 415–417 nectar homeostasis with, 413 nectar secretion pattern for, 413–415 nectar sugar reward for, 411–413 nectar volume for, 408 nest defense of, 436 Nomadopsis genera of, 348, 389 Nomia melanderi, 389 nonforaging rest of, 378–379 Osmia californica, 417 Osmia cornuta, 393 Osmia genera of, 348, 362, 380, 393, 425 Osmia rufa, 382 overview of intrinsic factors of, 352–354 parasocial bees as, 434 passive vs. active pollen collection by, 419 paternal investment in, 389–391 Peponapis genera of, 348, 379, 421 Perdita genera of, 348, 389, 394, 398 Perdita portalis, 382 phylogenetic variation in thermoregulation in, 362–263 phylogeny in, 352

472 Bees. See also Bumblebee; Honeybees (continued ) physiological regulation of heat loss in, 359 Pithitis genera of, 348, 370 pollen availabilities with, 417–425 pollen balance of, 436–437 pollen longevity for, 418 predation with, 425–431 Pseudapis genera of, 348, 368 Psithyrus genera of, 348, 365 Ptiloglossa arizonensis, 404 Ptiloglossa genera of, 348, 394–395, 400, 421 Ptiloglossa guinnae, 366 Ptiloglossa jonesi, 404 reasons to study activity patterns of, 347–351, 445–446 recruitment with, 442–444 sexual differences affecting, 369–392 sexual harassment among, 389–391 sexual variation in thermoregulation in, 364–365 size of, 352 size-related variation in thermoregulation in, 360–362, 363 social facilitation with, 442–445 social nester type of, 369, 370 sociality levels with, 433–435 sociality with, 432–445 solitary vs. social species in, 353 subfamilies of, 348 summary of activity patterns of, 445–446 task allocation with, 437 thermoregulation in, 354–366 thermoregulation with, 438 Thyreus kleptoparasite of, 365 time-place-learning in, 440 tongue lashing in, 358 tongue length of, 407 Trigona canifrons, 426, 428 Trigona fulviventris, 430 Trigona genera of, 348, 434, 440

INDEX

Trigona hyalinata, 430 upper critical temperatures of, 357, 358 wing wear with, 369 worker behavior control with, 437 Xylocopa californica, 383 Xylocopa fenestrata, 400 Xylocopa genera of, 348, 362, 370, 394–396, 400, 406, 407, 434 Xylocopa pubescens, 361, 400 Xylocopa sulcatipes, 361 Xylocopa tabaniformis, 404 Xylocopa tranquebarica, 366 Xylocopa varipuncta, 386 Begging definition of, 267 female sexual swelling and, 268 food transfer through, 267, 268, 269, 270, 278–280, 281, 282, 283, 284 infant age with, 282 infant food through, 270, 278–280, 281, 282, 283, 284 infant skill with, 284 informational hypotheses and, 281, 282, 283, 284 learning food preferences with, 281, 282 learning food-processing skills with, 283, 284 novel food items and, 281 nutritional hypotheses and, 278–280 Behavioral studies infant learning/memory in, 105–108, 116–118, 120–121 infant memory expression in, 120–121 learning acquisition in, 105–108 memory consolidation in, 116–118 Bengalese finch anterior forebrain pathway of, 321–322 auditory feedback with, 298–301 avicultural records with, 302 behavioral development with, 329 brain of, 315 call comparison with, 302–303

INDEX

cause of song complexity of, 338–339 cognitive cost of song complexity with, 337 complexity evolution scenario for, 339–340 copulation solicitation assay with, 332 deafening studies with, 299–300 developmental process with, 327, 328 DNA hybridization studies with, 303–304 electrophysiological studies on, 322–323, 324 estradiol level with, 333–334 female appreciation of song complexity with, 339 finite state in, 305–306 Fringilla coelebs with, 304 heart rate with, 330–331 helium studies with, 300–301 HVC of, 315, 317–319, 320, 322–323, 324, 325, 326–327 introduction to song syntax of, 297–301 lesion studies summary for, 325, 326 lesion studies with, 316–322 Lonchura striata with, 298, 302 Luscinia megarhynchos with, 307 Melospiza georgiana with, 304 Melospiza melodia with, 304 nest-building behavior with, 332–333 NIf of, 315, 319–321 nightingales with, 307 origins of songs of, 302–304 overall sound density with, 313 perceptual studies on, 323–325 Phylloscopus trochilus with, 308 posterior brain pathway of, 315, 316–321 RA of, 315, 316–317, 318, 323, 326–327 reinforcing properties of song complexity with, 331–332

473 self-generated songs with, 323 side of dominance with, 323–325 song analyses for, 304–309 song analyses summary for, 309 song control auditory dependence with, 338 song control system architecture with, 314–315 song development summary on, 329 song development with, 325–329 song evolution summary with, 313 song evolution with, 309–313 song function summary on, 334 song function with, 329–334 song learnability with, 338 song linearity in, 306–307 song mechanisms with, 313–325, 326 song note complexity comparisons within, 312–313 song note morphology comparisons within, 311–312 song phonology development with, 327 song syntax development with, 328–329 song syntax in, 297–341 species comparison with, 307–309 starlings with, 307–308 Sturnus vulgaris with, 307–308 stuttering with, 299 summary of song syntax with, 340–341 Tinbergen’s four questions with, 297–298, 309–334 traditional terms in song analyses of, 304 ts nerve of, 316–317 undirected/directed songs with, 334–336 uniqueness in, 308–309 wild strain birds song complexity and, 336–337 willow warblers with, 308 zebra finch with, 304, 314–315, 326–327 Zonotrichia leucophrys with, 304

474

INDEX

Betta splendens, 194 Biological market theory, game structures with, 61–62 Bombus activity patterns of, 348, 362, 364, 365, 395, 400, 402, 421, 433, 434, 438, 440 recruitment with, 440 sociality in, 433, 434 thermoregulation of, 364, 365, 438 Bombus alpinus, 421 Bombus hypnorum pheromones in, 13 reproductive conflict in, 13 Bombus lapidarius, 438 Bombus terrestris, 427 Bothriurus bonariensis, sexual cannibalism in, 139 Breeding density, androgen levels with, 187 Brown rats, food transfer with, 282 Bubulcus ibis, 196 Bumblebee, pheromones in, 13 Bumblebees, physiological regulation of heat loss in, 359 Buthus occitanus, sexual cannibalism in, 139 By-product mutualism, game structures with, 61, 62 Bystander effects, social modulation of androgens with, 202

C Callicebus moloch, food transfer with, 273, 274 Callicebus torquatus, food transfer with, 273, 274 Callimico goeldii, food transfer with, 273 Callithrix argentata, food transfer with, 272

Callithrix flaviceps, food transfer with, 272 Callithrix geoffroyi, food transfer with, 272 Callithrix jacchus, food transfer with, 272, 274 Camponotus floridanus, reproductive conflict in, 6–7 Camponotus planatus, reproductive conflict in, 28 Canis aureus, food transfer with, 277 Canis mesomelus, food transfer with, 277 Capuchins, food transfer with, 269 Cardiocondyla, reproductive conflict in, 33, 35 Carpenter bees, 369, 370, 386, 434 Caupolicana, 348, 394–395, 400 Cebuella pygmaea, food transfer with, 273 Cebus apella, food transfer with, 273 Cebus capucinus, food transfer with, 273 Centris pallida, 365, 387 Ceratina, 348, 370, 395, 434 Ceratina cyanea, activity pattern of, 363 Chalepoxenus muellerianus, reproductive conflict in, 36, 38 Chalicodoma, 348, 368, 405 Chalicodoma sicula, 376, 377 nesting of, 373 Challenge hypothesis aggressive behavior in, 196 androgen responsiveness with, 179–185 breeding density with, 187 breeding level in, 178, 180 cichlids in, 180–181 constitutive level in, 178, 180 Lamprologus callipterus in, 181, 182 mating success with, 187–188 mating system with, 179–185 Melospiza melodia in, 183 Neolamprologus pulcher in, 181, 182 Oreochromis mossambicus in, 181 physiological maximum in, 178, 180

475

INDEX

Pseudosimochromis curvifrons in, 181, 182 Sarotherodon galilaeus in, 183, 184, 213 Sarotherodon melanotheron in, 180–181 Sarotherodon occidentalis in, 180–181 social modulation of androgens and, 177–188 social stability in, 185 social status in, 186–187 Sturnus vulgaris in, 183 territoriality in, 186–187 three endocrine levels in, 178, 180 Tropheus moorii in, 181, 182 Zonotrichia leucophrys in, 183 Chelostoma florisomne, 348, 368, 378 Chimpanzees adult food transfer with, 268 food transfer with, 268, 279–280 infant food transfer with, 279–280 Chromis dispilus, androgen responsiveness in, 181 Cichlids, challenge hypothesis tested with, 180–181 Cleaning mutualisms, game structures with, 75–78, 87 Coelioxys, 365 Colletes activity patterns of, 348, 362, 380 male activity with, 380 Colletidae, 348, 364 Columba livia, food transfer with, 283 Conditioned stimulus (CS), Pavlovian conditioning with, 246, 247, 250 Configural theory, honeybees olfactory discrimination with, 247–249 Coolidge effect, 176 Coturnix japonica, 191 Creightonella, 348, 365, 406 Creightonella frontalis, nesting of, 374, 375 Crematogaster smithi, reproductive conflict in, 9 CS. See Conditioned stimulus

Cyphoderris strepitans, sexual cannibalism in, 136 Cyprinus carpio, androgen responsiveness in, 181

D D. fimbriatus. See Dolomedes fimbriatus D. triton. See Dolomedes triton Dear enemy effects, social modulation of androgens with, 199–200 Dependency ant protection mutualisms with, 78, 87 cleaning mutualisms with, 76, 87 game structures with, 63, 66, 69, 71, 74, 76, 78, 80, 82, 84, 87, 89–90, 91 mixed species aggregations with, 80, 87 Mu¨llerian mimicry with, 82, 87 mutualistic systems parameter of, 63, 66 nutrition mutualisms with, 84 pollinating seed parasite mutualisms with, 71, 87 pollination mutualisms with, 68, 87 protection mutualisms with, 76, 78, 80, 82, 87 seed dispersal mutualisms with, 74, 87 symbioses with, 84, 87 transportation mutualisms with, 68, 71, 74, 87 Diacamma australe pheromones in, 15 reproductive conflict in, 10, 15 Diacamma ceylonense, pheromones in, 13 Dieunomia triangulifera, 378 Digger bees, 369

476

INDEX

Dinoponera quadriceps pheromones in, 12, 13 reproductive conflict in, 11, 12, 13, 17, 18, 36 Discrimination problems, pavlovian conditioning and, 243–243 Dogs, food transfer with, 275 Dolichovespula arenaria, reproductive conflict in, 11 Dolichovespula saxonica, reproductive conflict in, 18 Dolomedes fimbriatus, sexual cannibalism in, 141, 142, 144, 158 Dolomedes triton, sexual cannibalism in, 142, 144, 154, 158 Dopamine, infant memory with, 116 Drosophila, genetic findings on, 251

E Echium vulgare, 410 Egg cannibalism, insect reproductive conflict with, 2, 11 Endogenous opioids, memory consolidation with, 116 Epeolus, 348, 368 Epinephelus morio, androgen responsiveness in, 181 ER. See Estrogen receptors Estrogen receptors (ER), social modulation with, 171 Eucera, 348, 416 Eulaema, 348, 389 Eulaema imperialis, 386 Eulaema meriana, 386 Eusocial bees association with, 440 climatic dependence with, 438 communication with, 442–445 conditioning of, 440 floral constancy with, 438–442 flower choice with, 444–445 learning effect with, 440 learning with, 438–442

memory development/duration with, 441–442 navigation with, 440 nectar balance of, 436–437 nest defense of, 436 other bees v., 435–436 pollen balance of, 436–437 recruitment with, 442–444 social facilitation with, 442–445 task allocation with, 437 thermoregulation with, 438 time-place-learning in, 440 worker behavior control with, 437 Eutetramorium mocquerysi, reproductive conflict in, 21, 36

F Falco sparverius, 196 Food transfer adult-adult, 266–270 adult chimpanzees in, 268 aggressive competition in, 268 Alouatta palliata in, 273 Aotus trivirgatus in, 273, 274 Aphelocoma coerulescens in, 283 Ateles geoffroyi in, 273 begging in, 267, 268, 269, 270, 278–280, 281, 282, 283, 284 behavior pattern definitions in, 266, 267 brown rats in, 282 Callicebus moloch in, 273, 274 Callicebus torquatus in, 273, 274 Callimico goeldii in, 273 Callithrix argentata in, 272 Callithrix flaviceps in, 272 Callithrix geoffroyi in, 272 Callithrix jacchus in, 272, 274 Canis aureus in, 277 Canis mesomelus in, 277 capuchins in, 269 Cebuella pygmaea in, 273 Cebus apella in, 273 Cebus capucinus in, 273

477

INDEX

Columba livia in, 283 conflict of interest with, 278 displacement in, 267 dogs in, 275 foxes in, 275 functional explanations of, 277 game theoretical model of, 269 Gorilla gorilla in, 273 Hylobates lar in, 273 infant age and, 282 infant chimpanzees in, 279–280 infant future diet from, 282 infant growth/survival with, 271, 275–277 infant helper benefits with, 274–275 infant mobility with, 271 infant weaning and, 276, 277, 278–279 infants in, 270–285 information donation through, 285–286 informational hypotheses with, 277, 280–285 interest in, 267 introduction to, 265–266, 267 jackals in, 277 learning food preferences through, 280–282 learning food-processing skills through, 283–285 Leontopithecus chrysopygus in, 272 Leontopithecus rosalia in, 273 Lycaon pictus in, 275 marmosets in, 272–273, 274, 279, 282 meerkats in, 277 monkeys in, 268, 269, 273, 274, 279–280 nonhuman primates with, 265–287 novel food items in, 281 nutritional hypotheses with, 277, 278–280 offering in, 267 Pan paniscus in, 273 Pan troglodytes in, 273 Parus atricapillus in, 283 Pongo pygmaeus in, 273

Rattus norvegicus in, 282 Rattus rattus in, 283 reciprocal altruism in, 269 resistance in, 267 retrieving in, 267 Saguinus bicolor in, 272, 274 Saguinus fuscicollis in, 272, 274 Saguinus labiatus in, 272 Saguinus mystax in, 272 Saguinus nigricollis in, 272 Saguinus oedipus in, 272, 274 spider monkeys in, 269, 273 stealing in, 267 summary of, 286–287 Suricata suricatta in, 275 tamarins in, 269, 272–273, 274, 279, 282 Tarsius spectrum in, 272, 274 teaching through, 285–286 trade in, 269 Vulpes vulpes in, 275 Formica fusca pheromones in, 13 reproductive conflict in, 32 Formica pergandei, reproductive conflict in, 21 Formica truncorum, reproductive conflict in, 23 Formicoxenini, reproductive conflict in, 8, 10, 11 Foxes, food transfer with, 275 Fringilla coelebs, song syntax in, 304 Frontal cortex, infant learning/memory with, 122 Fulica americana, 197

G Game structures active choice with, 63, 67, 70, 72, 74, 76, 79, 81, 83, 84, 87, 89–90, 91 ant protection mutualisms with, 78–80, 87 assessment parameters for, 65–68

478

INDEX

Game structures (continued ) behavioral options with, 63, 67, 70, 72–73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 biological market theory with, 61–62 by-product mutualism with, 61, 62 cleaning mutualisms with, 75–78, 87 conclusions on, 96 control over interaction with, 63, 68, 71, 73, 75, 77, 79–80, 81–82, 83, 85, 87, 89–90, 91 cooperative behavior and, 59–60 dependency with, 63, 66, 69, 71, 74, 76, 78, 80, 82, 84, 87, 89–90, 91 ecology’s importance with, 92–95 environment as competitor with, 94 future avenues with, 95–96 goals of article on, 63–64 initial investment with, 88 interspecific mutualism with, 60 intraspecific cooperation with, 60 investment with, 63, 67–68, 70, 73, 75, 77, 79, 81, 83, 85, 86–88, 87, 89–90, 91 literature evaluation for, 63, 66–86 mixed species aggregations with, 80–82, 87 mobility with, 63, 67, 70, 72, 74, 76, 78–79, 81, 82, 84, 87, 89–90, 91 moves with, 63, 67, 69–70, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 Mu¨llerian mimicry with, 82–83, 87, 88 mutualistic class similarity/difference with, 87, 90–92, 91 mutualistic interactions with, 59–97 n interactions with, 63, 67, 69, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 nectar-provisioning rate in, 62 nutrition mutualisms with, 83–86 offer produced with, 63, 67, 69, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 parameters evaluated with, 62, 63

partner choice with, 62, 88–89 partner recognition with, 63, 67, 70, 72, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 payoff symmetry with, 63, 68, 70, 73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 pollinating seed parasite mutualisms with, 71–74, 87 pollination mutualisms with, 66–71, 87 population density and, 92–94 prisoner’s dilemma game and, 65–66, 86 protection mutualisms with, 75–83 pseudoreciprocity with, 61, 62 seed dispersal mutualisms with, 74–75, 87 similar/different mutualisms and, 86–92 specificity with, 63, 66, 69, 72, 74, 76, 78, 81, 82, 84, 87, 91 symbioses with, 83–86, 87 terminology for, 64–65 theoretical approaches with, 60–63 third species population density with, 93–94 Tit-for-Tat strategy with, 60, 66 transmission mode with, 94–95 transportation mutualisms with, 66–75 Gaster curling, 36 Glutamate, infant memory with, 116 Gnamptogenys menadensis, reproductive conflict in, 11, 18 Gorilla gorilla, 273

H Halictus, 348, 380, 398 Halictus ligatus, 435 Halictus rubicundus, 435 Haplochromis burtoni, 189

INDEX

Harpagoxenus sublaevis, reproductive conflict in, 10 Harpegnathos saltator pheromones in, 13 reproductive conflict in, 13, 18, 21, 37 Heterosarellus, 348, 423 Hierodula membranacrea, sexual cannibalism in, 144 Hilltopping, bees doing, 386 Hippocampus, infant learning/memory with, 122 Honeybees anarchy in, 19–20 antennal lobe of, 251–252, 253, 255, 257, 258 Apis mellifera, 5, 7, 17, 19, 242, 245 compound stimulus processing models with, 242–245 conclusion to odor processing in, 259–260 Drosophila genetic findings and, 251 element/compound interactions in, 250–255 elemental/compound processing in, 255–259 forward pairing of PER in, 245, 246 genetic relationships with, 2–3, 5 glomerular inhibition in, 260 introduction to odor processing in, 241–242 Kenyon cells with, 255, 257, 258 modified unique cue theory with, 248, 258 mushroom bodies of, 252, 258 negative patterning with, 243, 246, 247 neural substrate for learning in, 255–259 odor processing in, 241–261 olfactory discrimination in, 246

479

olfactory system functional model for, 255–259 Pavlovian conditioning with, 242–250 Pearce’s configural theory with, 247–249 PER in, 245–250 pheromones in, 14 physiological correlates of odor processing in, 250–255 projection neuron with, 254, 256, 257 queen polyandry with, 5 receptor neuron with, 254 reproductive conflict in, 1, 2, 5, 14, 17, 19–20 summary on odor processing in, 260–261 Hoplitis, 348, 389 HPA. See Hypothalamic-pituitary adrenal axis HPG. See Hypothalamus-pituitarygonadal axis HVC Bengalese finch song syntax with, 315, 317–319, 320, 322–323, 324, 325, 326–327 zebra finch with, 314, 315 Hylobates lar, 273 Hymenoptera genetic relationships with, 2–3, 5 haplodiploid sex determination with, 3–4 queen polyandry with, 5 sex determination is, 3 Hypothalamic-pituitary adrenal axis (HPA) infant learning/memory with, 122 social modulation of androgens with, 201 Hypothalamus-pituitary-gonadal axis (HPG), social modulation of androgens with, 189–190 Hypsypops rubicundus, androgen responsiveness in, 181

480

INDEX

I I. oratoria. See Iris oratoria Infant learning/memory acquisition of, 105–115 active avoidance in, 107 amygdala in, 109, 114–115, 122 anterior olfactory nucleus in, 109, 111–113, 121 attachment in, 103–104 behavioral correlates of expressions in, 120–121 behavioral studies of acquisition with, 105–108 behavioral studies on consolidation in, 116–118 dopamine in, 116 early experiences influencing, 121–122 endogenous opioids in, 116 expressions of, 119–121 frontal cortex in, 122 glucocorticoids in, 116 glutamate in, 116 hippocampus in, 122 hypothalamic-pituitary adrenal axis in, 122 inhibitory conditioning in, 107 introduction to, 103–104 locus coeruleus in, 109–111, 112, 122 memory consolidation in, 115–119 milk in, 105–106 NE source in, 109–111 neural correlates of acquisition with, 108–115 neural correlates of consolidation in, 118–119 neural correlates of expressions in, 120–121 neurobehavioral development of, 103–123 neurotransmitters in, 116, 118 norepinephrine in, 116 odor preference in, 105–108, 116–117, 120

olfactory bulb in, 109–111 opioid receptor antagonist in, 117 passive avoidance in, 107 piriform cortex in, 109, 113–114 rats studied in, 104–121 shock training in, 105–106, 107, 114, 116–117, 120 stroking in, 105 summary on, 122–123 tactile stimulation in, 105 tail pinch in, 106, 107 unique characteristics of, 104–121 warmth in, 105 Inhibitory conditioning, infant learning with, 107 Insect societies Acanthomyrmex ferox in, 21 aggression with, 2, 10, 37–38 anarchic honeybees and, 19–20 antennation boxing in, 35–36 antennation in, 10–11, 17, 35–36 ants in, 1, 8, 9, 13, 22, 28 ants pheromones in, 13 Apis mellifera in, 5, 7, 17, 19 behavioral side of, 35–39 Bombus hypnorum pheromones in, 13 bumblebee pheromones in, 13 Camponotus floridanus in, 6–7 Camponotus planatus in, 28 Cardiocondyla in, 33, 35 caste discrimination in, 33–34 Chalepoxenus muellerianus in, 36, 38 conclusion to, 39–43 conflict resolution within, 35–39 control mechanisms within, 7–35 Crematogaster smithi in, 9 crouching in, 36 Diacamma australe in, 10, 15 Diacamma australe pheromones in, 15 Diacamma ceylonense pheromones in, 13 Dinoponera quadriceps in, 11, 12, 13, 17, 18, 36

INDEX

Dinoponera quadriceps pheromones in, 12, 13 Dolichovespula arenaria in, 11 Dolichovespula saxonica in, 18 dominance interactions in, 2, 11, 35–38 egg cannibalism in, 2, 11 escaping control in, 19–20 Eutetramorium mocquerysi in, 21, 36 Formica fusca in, 32 Formica fusca pheromones in, 13 Formica pergandei in, 21 Formica truncorum in, 23 Formicoxenini in, 8, 10, 11 foundress association conflict with, 25–26 gamergates in, 2 gaster curling in, 36 genetic lineages differences in, 32–33 genetic relationships shaping, 2–3, 5 Gnamptogenys menadensis in, 11, 18 haplodiploid sex determination within, 3–4 Harpagoxenus sublaevis in, 10 Harpegnathos saltator in, 13, 18, 21, 37 Harpegnathos saltator pheromones in, 13 honeybee pheromones in, 14 honeybees in, 1, 2, 5, 14, 17, 19–20 hymenopteran males in, 34–35 immobilization in, 17, 18 introduction to, 1–2 kin discrimination in, 32–33 Lasius niger in, 7 Leptothorax acervorum in, 11, 23, 28, 32, 40 Leptothorax allardycei in, 20 Leptothorax curvispinosus in, 26 Leptothorax gredleri in, 31, 36, 37 Leptothorax nylanderi in, 23, 33, 39 Leptothorax sphagnicolus in, 31 male offspring production in, 8–22

481 maturity in, 26–28, 29 multiple queens in, 2, 26–28, 29 Myrmecia gulosa pheromones in, 13 Myrmica tahoensis in, 9, 40 Nauphoeta cinerea pheromones in, 13 Odontomachus chelifer in, 28, 29 optimal skew theory with, 30–31 Pachycondyla inversa in, 13, 14, 15, 17, 19, 21, 25–26, 38 Pachycondyla inversa pheromones in, 13, 14, 15 Pachycondyla sublaevis in, 39 Paratrigona subnuda in, 9 perspectives on, 39–43 pheromones composition and, 12 Platythyrea punctata in, 13, 19, 36, 39 Platythyrea punctata pheromones in, 13 Polistes bellicosus in, 31 Polistes carolina in, 32 Polistes dominulus pheromones in, 13 Polistes fuscatus in, 38 Polyergus breviceps in, 21 Ponerinae in, 8, 10 Protomognathus americanus in, 9, 10, 20 punishment within, 16–19 pygidial display in, 36 queen control in, 9–11 queen pheromones in, 11–16 queen polyandry with, 4, 5, 6 queen-queen conflict in, 24–31 reconciliation with, 35–38 reproduction partitioning in, 24–31 reproductive conflict causes with, 2–7 reproductive conflict within, 1–43 Schwarziana quadripunctata in, 40 selfish larvae in, 33–34 sex allocation with, 22–24 sex ratio theory with, 22–23 shape of hierarchies with, 38–39 slave-making ants with, 22

482

INDEX

Insect societies (continued ) Solenopsis geminata in, 33 Solenopsis invicta in, 12, 24 Solenopsis invicta pheromones in, 12 subordination within, 35–38 summary of, 43 termite societies in, 35 types of, 7–35 Vespa crabro in, 40 Vespula vulgaris in, 18 virgin reproductives in, 29–30 wasps in, 8, 13 wasps pheromones in, 13 worker policing in, 16–19 worker-worker conflict in, 20–22 Interfacial nucleus of nidopallium (NIf), Bengalese finch song syntax with, 315, 319–321 Interspecific mutualism, game structures with, 60 Intraspecific cooperation, game structures with, 60 Investment ant protection mutualisms with, 79, 87 cleaning mutualisms with, 77, 87 game structures with, 63, 67–68, 70, 73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 mixed species aggregations with, 81, 87 Mu¨llerian mimicry with, 83, 87 mutualistic systems parameter of, 63, 67–68 nutrition mutualisms with, 85 pollinating seed parasite mutualisms with, 73, 87 pollination mutualisms with, 70, 87 protection mutualisms with, 77, 79, 81, 83, 87 seed dispersal mutualisms with, 75, 87 symbioses with, 85, 87 transportation mutualisms with, 70, 73, 75, 87 Iris oratoria, sexual cannibalism in, 137, 144

J Jackals, food transfer with, 277

K Kenyon cells, honeybees with, 255, 257, 258 Ketotestosterone/testosterone ratio, 194 Kleptoparasitic bees, activity patterns of, 365, 370, 391–392

L L. acervorum. See Leptothorax acervorum L. hasselti. See Latrodectus hasselti Lamprologus callipterus, 181, 182 Larus fuscus, 196 Larus ridibundus, 196 Lasioglossum, 348, 395, 398, 434 Lasioglossum hemichalceum, 382 Lasius niger, reproductive conflict in, 7 Lates calcarifer, androgen responsiveness in, 181 Latrodectus hasselti, sexual cannibalism in, 143, 144, 148, 151, 154 Lavendula stoechas, 363 Learning acquisition behavioral studies on, 105–108 infant learning and, 105–115 odor preference in, 105–108 shock training in, 105–106, 107, 114 tactile stimulation in, 105 Leiurus quinquestriatus, sexual cannibalism in, 139 Lemur catta, 217 Leontopithecus chrysopygus, 272 Leontopithecus rosalia, 273

INDEX

Leptothorax acervorum, reproductive conflict in, 11, 23, 28, 32, 40 Leptothorax allardycei, reproductive conflict in, 20 Leptothorax curvispinosus, reproductive conflict in, 26 Leptothorax gredleri, reproductive conflict in, 31, 36, 37 Leptothorax nylanderi, reproductive conflict in, 23, 33, 39 Leptothorax sphagnicolus, reproductive conflict in, 31 LH. See Luteinizing hormone Locus coeruleus, infant learning with, 109–111, 112, 122 Lonchura striata, song syntax of, 298–302 Luscinia megarhynchos, song syntax of, 307 Luteinizing hormone (LH), androgens with, 171 Lycaon pictus, 275

M Macaca mulatta, 217 Marmosets, food transfer with, 272–273, 274, 279, 282 Mason bees, 369–370, 371, 376 Mating success, androgen levels with, 187–188 Meerkats, food transfer with, 277 Megachile activity patterns of, 348, 370, 380, 395 male activity with, 380 Megachile apicalis, 376–377 Megachile nana, 393, 431 Megachile rotundata, 417 Megachilidae, 348, 364 Megalopta, 348, 366 Melecta, 348, 365, 368 Melipona, 348, 416, 434 Melipona beecheii, 425 Melissodes, 348, 375

483

Melittophily syndrome, 410 Melospiza georgiana, song syntax in, 304 Melospiza melodia, 183 song syntax in, 304 Memory consolidation behavioral studies on, 116–118 dopamine in, 116 endogenous opioids in, 116 glutamate in, 116 infant memory and, 115–119 neurotransmitters in, 116, 118 norepinephrine in, 116 odor preference study on, 116–117 opioid receptor antagonist in, 117 shock training study of, 116–117 Memory expression behavioral correlates in, 120–121 infant memory and, 119–121 neural correlates in, 120–121 Men, social modulation of androgens with, 216–218 Micrathena gracilis, sexual cannibalism in, 156 Microtus ochrogaster, 213 Miner bees, 369 Mixed species aggregations, game structures with, 80–82, 87 Mobility ant protection mutualisms with, 78–79, 87 cleaning mutualisms with, 76, 87 game structures with, 63, 66, 67, 69, 70, 72, 74, 76, 78–79, 81, 82, 84, 87, 89–90, 91 mixed species aggregations with, 81, 87 Mu¨llerian mimicry with, 82, 87 mutualistic systems parameter of, 63, 67 nutrition mutualisms with, 84 pollinating seed parasite mutualisms with, 72, 87 pollination mutualisms with, 70, 87 protection mutualisms with, 76, 78–79, 80, 82, 87

484

INDEX

Mobility (continued ) seed dispersal mutualisms with, 74, 87 symbioses with, 84, 87 transportation mutualisms with, 70, 72, 74, 87 Modified unique cue theory, honeybees olfactory discrimination in, 248, 258 Monkeys, food transfer with, 268, 269, 273, 274, 279–280 Monogyny, sexual cannibalism, 151–153 Mu¨llerian mimicry, game structures with, 82–83, 87, 88 Mushroom bodies honeybees with, 252, 258 olfactory processing with, 252, 258 Mutualistic interactions active choice with, 63, 67, 70, 72, 74, 76, 79, 81, 83, 84, 87, 89–90, 91 ant protection mutualisms with, 78–80, 87 assessment parameters for, 65–68 behavioral options with, 63, 67, 70, 72–73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 biological market theory with, 61–62 by-product mutualism with, 61, 62 cleaning mutualisms with, 75–78, 87 conclusions on, 96 control over interaction with, 63, 68, 71, 73, 75, 77, 79–80, 81–82, 83, 85, 87, 89–90, 91 cooperative behavior and, 59–60 dependency with, 63, 66, 69, 71, 74, 76, 78, 80, 82, 84, 87, 89–90, 91 ecology’s importance with, 92–95 environment as competitor with, 94 future avenues with, 95–96 game structures in, 59–97 goals of article on, 63–64 initial investment with, 88 interspecific mutualism with, 60 intraspecific cooperation with, 60

investment with, 63, 67–68, 70, 73, 75, 77, 79, 81, 83, 85, 86–88, 89–90, 91 literature evaluation for, 63, 66–86 mixed species aggregations with, 80–82, 87 mobility with, 63, 67, 70, 72, 74, 76, 78–79, 81, 82, 84, 87, 89–90, 91 moves with, 63, 67, 69–70, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 Mu¨llerian mimicry with, 82–83, 87, 88 mutualistic class similarity/difference with, 87, 90–92 n interactions with, 63, 67, 69, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 nectar-provisioning rate in, 62 nutrition mutualisms with, 83–86 offer produced with, 63, 67, 69, 72, 74, 76, 78, 81, 82, 84, 87, 89–90, 91 parameters evaluated with, 62, 63 partner choice with, 62, 88–89 partner recognition with, 63, 67, 70, 72, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 payoff symmetry with, 63, 68, 70, 73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 pollinating seed parasite mutualisms with, 71–74, 87 pollination mutualisms with, 66–71, 87 population density and, 92–94 prisoner’s dilemma game and, 65–66, 86 protection mutualisms with, 75–83 pseudoreciprocity with, 61, 62 seed dispersal mutualisms with, 74–75, 87 similar/different mutualisms and, 86–92 specificity with, 63, 66, 69, 72, 74, 76, 78, 81, 82, 84, 87, 91 symbioses with, 83–86, 87

485

INDEX

terminology for, 64–65 theoretical approaches with, 60–63 third species population density with, 93–94 Tit-for-Tat strategy with, 60, 66 transmission mode with, 94–95 transportation mutualisms with, 66–75 Myrmecia gulosa, pheromones in, 13 Myrmica tahoensis, reproductive conflict in, 9, 40

N N. edulis. See Nephila edulis N. plumipes. See Nephila plumipes Nauphoeta cinerea, pheromones in, 13 NE. See Norepinephrine Nectar-provisioning rate, 62 Negative patterning olfactory discrimination with, 246, 247 pavlovian conditioning and, 243, 246, 247 Neolamprologus pulcher, 181, 182 Nephila constricta, sexual cannibalism in, 155 Nephila edulis, sexual cannibalism in, 137, 149, 156 Nephila fenestrata, sexual cannibalism in, 152, 155 Nephila plumipes, sexual cannibalism in, 137, 138, 145, 149, 150, 154, 155, 156 Nephila tetraganthoides, sexual cannibalism in, 155 Neurotransmitters, memory consolidation with, 116, 118 NIf. See Interfacial nucleus of nidopallium Bengalese finch song syntax with, 315, 319–321 Nightingales, song syntax of, 307 Nomadopsis, 348, 389 Nomia melanderi, 389

Norepinephrine, memory consolidation with, 116 Norepinephrine (NE), infant learning with, 109–111 Nutrition mutualisms, game structures with, 83–86 Nutritional hypotheses, food transfer in, 277, 278–280

O Odontomachus chelifer, reproductive conflict in, 28, 29 Odor preference infant learning/memory with, 105–108, 116–117 learning acquisition studied with, 105–108 memory consolidation studied with, 116–117 Olfactory bulb, infant learning with, 109–111 Olfactory processing compound stimulus processing models with, 242–245 conclusion to honeybees with, 259–260 discrimination in, 246 Drosophila genetic findings and, 251 element/compound interactions in, 250–255 forward pairing of PER in, 245, 246 functional model for, 255–259 glomerular inhibition in, 260 honeybee antennal lobe in, 251–252, 253, 255, 257, 258 honeybee mushroom bodies in, 252, 258 honeybees with, 241–261 introduction to honeybees in, 241–242 Kenyon cells with, 255, 257, 258 modified unique cue theory with, 248, 258

486

INDEX

Olfactory processing (continued ) negative patterning with, 243, 246, 247 neural substrate for, 255–259 Pavlovian conditioning with, 242–250 Pearce’s configural theory with, 247–249 PER in, 245–250 physiological correlates of, 250–255 projection neuron with, 254, 256, 257 receptor neuron with, 254 summary on honeybees with, 260–261 Opioid receptor antagonist, memory consolidation with, 117 Optimal skew theory, insect reproductive conflict with, 30–31 Oreochromis mossambicus, androgen responsiveness in, 181, 182 Osmia activity patterns of, 348, 362, 380, 425 male activity with, 380 pollen-gathering of, 425 Osmia californica, 417 Osmia rufa, 382

P P. punctata. See Platythyrea punctata Pachycondyla inversa foundress associations with, 25–26 pheromones in, 13, 14, 15 reproductive conflict in, 13, 14, 15, 17, 19, 21, 25–26, 38 Pachycondyla sublaevis, reproductive conflict in, 39 Pan paniscus, 273 Pan troglodytes, 217, 273 Paratrigona subnuda, reproductive conflict in, 9

Partner recognition ant protection mutualisms with, 79, 87 cleaning mutualisms with, 77, 87 game structures with, 63, 67, 70, 72, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 mixed species aggregations with, 81, 87 Mu¨llerian mimicry with, 83, 87 mutualistic systems parameter of, 63, 67 nutrition mutualisms with, 85 pollinating seed parasite mutualisms with, 72, 87 pollination mutualisms with, 70, 87 protection mutualisms with, 77, 79, 80, 82, 87 seed dispersal mutualisms with, 75, 87 symbioses with, 85, 87 transportation mutualisms with, 70, 72, 75, 87 Parus atricapillus, 283 Passer domesticus, 196–197 Passive avoidance, infant learning with, 107 Pavlovian conditioning associative learning with, 244 associative theories with, 243 compound stimulus processing models with, 242–245 conditioned stimulus with, 246, 247, 250 configural models with, 244–245 discrimination problems and, 243 forward pairing of PER in, 245, 246 honeybees with, 242–250 modified unique cue theory with, 248 negative patterning and, 243, 246, 247 olfactory compound stimulus processing with, 245–250 Pearce’s configural theory with, 247–249

INDEX

PER in, 245–250 theoretical viewpoint on, 243 unconditioned stimulus with, 246 Payoff symmetry ant protection mutualisms with, 79, 87 cleaning mutualisms with, 77, 87 game structures with, 63, 68, 70, 73, 75, 77, 79, 81, 83, 85, 87, 89–90, 91 mixed species aggregations with, 81, 87 Mu¨llerian mimicry with, 83, 87 mutualistic systems parameter of, 63, 68 nutrition mutualisms with, 85 pollinating seed parasite mutualisms with, 73, 87 pollination mutualisms with, 70, 87 protection mutualisms with, 77, 79, 81, 83, 87 seed dispersal mutualisms with, 75, 87 symbioses with, 85, 87 transportation mutualisms with, 70, 73, 75, 87 Pearce’s configural theory, honeybees olfactory discrimination with, 247–249 Peponapis, 348, 421 PER. See Proboscis extension reflex Perdita, 348, 389, 394 Perdita portalis, 382 Peromyscus californicus, 213 Phenotypic plasticity, social modulation of androgens with, 203–204 Pheromones ants with, 13 Bombus hypnorum with, 13 bumblebees with, 13 composition of, 12 cuticular hydrocarbons associated with, 13–14 Diacamma australe with, 15 Diacamma ceylonense with, 13

487

Dinoponera quadriceps with, 12, 13 Formica fusca with, 13 Harpegnathos saltator with, 13 honest signaling hypothesis with, 15 honeybees with, 14 insect reproductive conflict with, 12–16 Myrmecia gulosa with, 13 Nauphoeta cinerea with, 13 origin of, 12 Pachycondyla inversa with, 13, 14, 15 Platythyrea punctata with, 13 Polistes dominulus with, 13 Solenopsis invicta with, 12 Vespa structor with, 12 wasps with, 13 Phodopus campbelli, 213 Phonognatha graeffei, sexual cannibalism in, 144 Phoreticovelia disparata, sexual cannibalism in, 136 Phylloscopus trochilus, song syntax of, 308 Piriform cortex, infant learning with, 109, 113–114 Pisaura mirabilis, sexual cannibalism in, 144 Pithitis, 348, 370 Platythyrea punctata pheromones in, 13 reproductive conflict in, 13, 19, 36, 39 PN. See Projection neuron Polistes bellicosus, reproductive conflict in, 31 Polistes carolina, reproductive conflict in, 32 Polistes dominulus, pheromones in, 13 Polistes fuscatus, reproductive conflict in, 38 Pollinating seed parasite mutualisms, game structures with, 71–74, 87 Pollination mutualisms, game structures with, 66–71, 87

488

INDEX

Polyergus breviceps, reproductive conflict in, 21 Ponerinae, reproductive conflict in, 8, 10 Pongo pygmaeus, 273 Population density, game structures and, 92–94 Primates, food transfer with, 265–287 Prisoner’s dilemma game, 65–66, 86 Proboscis extension reflex (PER) forward pairing with, 245, 246 honeybee learning with, 245–250 Projection neuron (PN), honeybees with, 254, 256, 257 Protection mutualisms, game structures with, 75–83 Protomognathus americanus, reproductive conflict in, 9, 10, 20 Pseudapis, 348, 368 Pseudoreciprocity, game structures with, 61, 62 Pseudosimochromis curvifrons, 181, 182 Psithyrus, 348, 365 Ptiloglossa, 348, 394–395, 400, 421 Ptiloglossa arizonensis, 404 Ptiloglossa guinnae, 366 Ptiloglossa jonesi, 404 Punishment, insect reproductive conflict in, 16–19

Q Queen polyandry, 4, 5, 6

R RA. See Robust nucleus of arcopallium Rats infant learning/memory studied in, 104–121 two developmental periods in, 104

Rattus norvegicus, 282 Rattus rattus, 283 Receptor neuron (RN), honeybees with, 254 Reciprocal altruism, food transfer with, 269 Reproductive conflict Acanthomyrmex ferox with, 21 aggression with, 2, 10, 37–38 anarchic honeybees and, 19–20 antennation boxing with, 35–36 antennation with, 10–11, 17, 35–36 ants pheromones in, 13 ants with, 1, 8, 9, 13, 22, 28 Apis mellifera having, 5, 7, 17, 19 behavioral side of, 35–39 Bombus hypnorum pheromones in, 13 bumblebee pheromones in, 13 Camponotus floridanus with, 6–7 Camponotus planatus with, 28 Cardiocondyla with, 33, 35 caste discrimination in, 33–34 causes of, 2–7 Chalepoxenus muellerianus with, 36, 38 conclusion to, 39–43 conflict resolution with, 35–39 control mechanisms with, 7–35 Crematogaster smithi with, 9 crouching in, 36 Diacamma australe pheromones in, 15 Diacamma australe with, 10, 15 Diacamma ceylonense pheromones in, 13 Dinoponera quadriceps pheromones in, 12, 13 Dinoponera quadriceps with, 11, 12, 13, 17, 18, 36 Dolichovespula arenaria with, 11 Dolichovespula saxonica with, 18 dominance interactions with, 2, 11, 35–38 egg cannibalism with, 2, 11 escaping control in, 19–20

INDEX

Eutetramorium mocquerysi with, 21, 36 Formica fusca pheromones in, 13 Formica fusca with, 32 Formica pergandei with, 21 Formica truncorum with, 23 Formicoxenini with, 8, 10, 11 foundress association conflict with, 25–26 gamergates leading to, 2 gaster curling in, 36 genetic lineages differences in, 32–33 genetic relationships shaping, 2–3, 5 Gnamptogenys menadensis with, 11, 18 haplodiploid sex determination with, 3–4 Harpagoxenus sublaevis with, 10 Harpegnathos saltator pheromones in, 13 Harpegnathos saltator with, 13, 18, 21, 37 honeybee pheromones in, 14 honeybees having, 1, 2, 5, 14, 17, 19–20 hymenopteran males in, 34–35 immobilization with, 17, 18 insect societies with, 1–43 introduction to, 1–2 kin discrimination in, 32–33 Lasius niger with, 7 Leptothorax acervorum with, 11, 23, 28, 32, 40 Leptothorax allardycei with, 20 Leptothorax curvispinosus with, 26 Leptothorax gredleri with, 31, 36, 37 Leptothorax nylanderi with, 23, 33, 39 Leptothorax sphagnicolus with, 31 male offspring production with, 8–22 mature societies with, 26–28, 29 multiple queens in, 2, 26–28, 29 Myrmecia gulosa pheromones in, 13 Myrmica tahoensis with, 9, 40

489 Nauphoeta cinerea pheromones in, 13 Odontomachus chelifer with, 28, 29 optimal skew theory in, 30–31 Pachycondyla inversa pheromones in, 13, 14, 15 Pachycondyla inversa with, 13, 14, 15, 17, 19, 21, 25–26, 38 Pachycondyla sublaevis with, 39 Paratrigona subnuda with, 9 perspectives on, 39–43 pheromones composition and, 12 Platythyrea punctata pheromones in, 13 Platythyrea punctata with, 13, 19, 36, 39 Polistes bellicosus with, 31 Polistes carolina with, 32 Polistes dominulus pheromones in, 13 Polistes fuscatus with, 38 Polyergus breviceps with, 21 Ponerinae with, 8, 10 Protomognathus americanus with, 9, 10, 20 punishment with, 16–19 pygidial display in, 36 queen control in, 9–11 queen pheromones in, 11–16 queen polyandry with, 4, 5, 6 queen-queen conflict in, 24–31 reconciliation with, 35–38 reproduction partitioning in, 24–31 Schwarziana quadripunctata with, 40 selfish larvae in, 33–34 sex allocation with, 22–24 sex ratio theory with, 22–23 shape of hierarchies with, 38–39 slave-making ants with, 22 Solenopsis geminata with, 33 Solenopsis invicta pheromones in, 12 Solenopsis invicta with, 12, 24 subordination with, 35–38 summary of, 43 termite societies in, 35

490

INDEX

Reproductive conflict (continued ) types of, 7–35 Vespa crabro with, 40 Vespula vulgaris with, 18 virgin reproductives in, 29–30 wasps pheromones in, 13 wasps with, 8, 13 worker policing in, 16–19 worker-worker conflict in, 20–22 RN. See Receptor neuron Robust nucleus of arcopallium (RA) Bengalese finch song syntax with, 315, 316–317, 318, 323, 326–327 zebra finch with, 314, 315

S Saguinus bicolor, 272, 274 Saguinus fuscicollis, 272, 274 Saguinus labiatus, 272 Saguinus mystax, 272 Saguinus nigricollis, 272 Saguinus oedipus, 213, 272, 274 Salmo salar, androgen responsiveness in, 181 Salmo trutta, androgen responsiveness in, 181 Salvelinus alpinus, androgen responsiveness in, 181 Salvelinus fontinalis, androgen responsiveness in, 181 Sarotherodon galilaeus, 183, 184, 213 Sarotherodon melanotheron androgen responsiveness in, 181 challenge hypothesis tested with, 180–181 Sarotherodon occidentalis, challenge hypothesis tested with, 180–181 Schizocosa uetzi, sexual cannibalism in, 147 Schwarziana quadripunctata, reproductive conflict in, 40 Seed dispersal mutualisms, game structures with, 74–75, 87 Selfish larvae, 33–34

Serinus canaria, 191, 196 Serranus subligarius, androgen responsiveness in, 181 Sex change, social modulation of androgens with, 204–205 Sex ratio theory, insect reproductive conflict with1, 22–23 Sexual cannibalism Araneus bruennichi with, 145, 146 Araneus diadematus with, 139, 144, 145, 147, 158 Argiope aemula with, 137 Argiope argentata with, 149 Argiope bruennichi with, 139, 144, 154–155 Argiope keyserlingi with, 137, 144, 148, 150, 152, 154, 156 Baryphyma pratense with, 136 Bothriurus bonariensis with, 139 Buthus occitanus with, 139 copulation duration with, 153–155 Cyphoderris strepitans with, 136 Dolomedes fimbriatus with, 141, 142, 144, 158 Dolomedes triton with, 142, 144, 154, 158 effective sex ration with, 151–152 evolutionary significance of, 135–159 female choice in, 147–148 female fecundity with, 143, 144 female foraging strategies with, 140–143 Gammarus with, 136 Hierodula membranacrea with, 144 introduction to, 135–136 Iris oratoria with, 137, 144 Latrodectus hasselti with, 143, 144, 148, 151, 154 Leiurus quinquestriatus with, 139 mate choice in, 147–148 mating rate with, 153–155 Micrathena gracilis with, 156 monogyny with, 151–153 natural history of, 136–139 natural selection of, 140–146

INDEX

Nephila constricta with, 155 Nephila edulis with, 137, 149, 156 Nephila fenestrata with, 152, 155 Nephila plumipes with, 137, 138, 145, 149, 150, 154, 155, 156 Nephila tetraganthoides with, 155 outlook on, 156–159 paternal investment in, 143–146 Phonognatha graeffei with, 144 Phoreticovelia disparata with, 136 Pisaura mirabilis with, 144 postinsemination cannibalism in, 143–146, 148–153 preinsemination with, 147–148 Schizocosa uetzi with, 147 self-sacrifice with, 151–153 sexual conflict with, 153–155 sexual selection and, 146–151 sexual size dimorphism and, 155–156 sperm competition with, 148–150 summary of, 156–159 taxonomic distribution with, 136–139 Tidarren argo with, 137 timing of, 139–140 Urophonius brachycentrus with, 139 Urophonius jheringii with, 139 Sexual harassment Anthophora plumipes, 389 bees, 389–391 Shock training infant learning/memory with, 105–106, 107, 114, 116–117 learning acquisition studied with, 105–106, 107, 114 memory consolidation studied with, 116–117 Slave-making ants, reproductive conflict with, 22 Social modulation of androgens action with, 174 activational effects with, 198–203 adaptive significance of, 197–206 Agelaius phoeniceus in, 196

491 alternative reproductive tactics with, 205–206 androgen cognitive performance and, 173–174 androgen motivational systems and, 171–173 androgen receptors with, 171 androgen responsiveness with, 179–185 androgens as deterministic factors and, 169–170 androgens as neuromodulators and, 169–170 behavioral endocrinology in, 166–169 behavioral feedback background in, 175–177 behavioral feedback on endocrine function with, 175–188 Betta splendens in, 194 breeding density with, 187 Bubulcus ibis in, 196 bystander effects with, 202 castration and, 167–168 causal agents of behavior in, 166–175 causal behavior background in, 175–177 central actions of androgens and, 171–174 challenge hypothesis with, 177–188 cichlids in, 180–181 Coolidge effect with, 176 cost-benefit analysis for, 207–214 Coturnix japonica in, 191 dear enemy effects with, 199–200 estrogen receptors with, 171 eunuchs and, 167 evolution of, 206–216 evolutionary scenarios for selection with, 214–216 Falco sparverius in, 196 Fulica americana in, 197 Haplochromis burtoni in, 189 high androgen level benefits with, 207, 208–209

492

INDEX

Social modulation of androgens (continued ) high androgen level costs with, 207, 209–214 historical background with, 166–169, 175–177 HPA axis with, 201 hypothalamus-pituitary-gonadal axis with, 189–190 immunosuppression with, 207, 209–212 increased energy consumption with, 207, 209, 210 introduction to, 165–166, 167 Ketotestosterone/testosterone ratio in, 194 Lamprologus callipterus in, 181, 182 Larus fuscus in, 196 Larus ridibundus in, 196 Lemur catta in, 217 life history trade-offs with, 203–204 long-term effects with, 203–206 luteinizing hormone with, 171 Macaca mulatta in, 217 mating success with, 187–188 mating system with, 179–185 mechanisms/functions in, 165–219 mechanisms of, 170–175 Melospiza melodia in, 183 men in, 216–218 Microtus ochrogaster in, 213 model for, 167 Neolamprologus pulcher in, 181, 182 neuroendocrine mechanisms for, 189–192 ontogeny of, 195–197 Oreochromis mossambicus in, 181 organizational effects with, 203–206 Pan troglodytes in, 217 parental care with, 207, 209, 211, 212–213 Passer domesticus in, 196–197 perception with, 174 peripheral actions of androgens and, 174–175 Peromyscus californicus in, 213

phenotypic plasticity with, 203–204 Phodopus campbelli in, 213 predation risk with, 207, 209 proximate mechanisms for, 188–195 Pseudosimochromis curvifrons in, 181, 182 psychological mechanisms for, 192–195 rejuvenation hypothesis in, 169 Saguinus oedipus in, 213 Sarotherodon galilaeus in, 183, 184 Sarotherodon melanotheron in, 180–181 Sarotherodon occidentalis in, 180–181 Serinus canaria in, 191, 196 sex change with, 204–205 short-term effects with, 198–203 social interactions with, 177–188 social stability with, 185 social status with, 186–187 spermatogenesis with, 207 Steinach operation in, 169 Sterna hirundo in, 196 Sturnus vulgaris in, 183, 191 Sula granti in, 196 Sula nebouxii in, 196 summary of, 218–219 Tachycineta bicolor in, 197 territoriality effects with, 198–199 territoriality with, 186–187 Tinbergen legacy with, 188 Tropheus moorii in, 181, 182 Vulpes vulpes in, 206 winner-loser effects with, 200–202 Zonotrichia leucophrys in, 183 Social nester bees, 369, 370 Social status, androgen levels with, 186–187 Solenopsis geminata, reproductive conflict in, 33 Solenopsis invicta pheromones in, 12 reproductive conflict in, 12, 24 Song syntax analyses of, 304–309 analyses summary on, 309

INDEX

Bengalese finch anterior forebrain pathway for, 321–322 Bengalese finch avicultural records with, 302 Bengalese finch behavioral development with, 329 Bengalese finch brain for, 315 Bengalese finch call comparison with, 302–303 Bengalese finch developmental process with, 327, 328 Bengalese finch DNA hybridization studies with, 303–304 Bengalese finch electrophysiological studies on, 322–323, 324 Bengalese finch finite state in, 305–306 Bengalese finch HVC for, 315, 317–319, 320, 322–323, 324, 325, 326–327 Bengalese finch NIf for, 315, 319–321 Bengalese finch origins with, 302–304 Bengalese finch perceptual studies on, 323–325 Bengalese finch posterior brain pathway for, 315, 316–321 Bengalese finch RA for, 315, 316–317, 318, 323, 326–327 Bengalese finch side of dominance with, 323–325 Bengalese finch stuttering with, 299 Bengalese finch ts nerve for, 316–317 Bengalese finch uniqueness in, 308–309 Bengalese finches auditory feedback with, 298–301 Bengalese finches deafening studies with, 299–300 Bengalese finches helium studies with, 300–301 Bengalese finches with, 297–341 cause of complexity in, 338–339 cognitive cost of complexity in, 337

493 complexity evolution scenario for, 339–340 copulation solicitation assay with, 332 development of, 325–329 development summary on, 329 estradiol level with, 333–334 evolution summary with, 313 evolution with, 309–313 female appreciation of complexity in, 339 Fringilla coelebs with, 304 function summary on, 334 function with, 329–334 heart rate with, 330–331 introduction to, 297–301 learnability with, 338 lesion studies summary for, 325, 326 lesion studies with, 316–322 linearity in, 306–307 Lonchura striata with, 298, 302 Luscinia megarhynchos with, 307 mechanisms with, 313–325, 326 Melospiza georgiana with, 304 Melospiza melodia with, 304 nest-building behavior with, 332–333 nightingales with, 307 overall sound density with, 313 phonology development with, 327 Phylloscopus trochilus with, 308 reinforcing properties of complexity in, 331–332 self-generated songs with, 323 song control auditory dependence with, 338 song control system architecture with, 314–315 song note complexity comparisons within, 312–313 song note morphology comparisons within, 311–312 species comparison with, 307–309 starlings with, 307–308 Sturnus vulgaris with, 307–308

494 Song syntax (continued ) summary of, 340–341 syntax development with, 328–329 Tinbergen’s four questions with, 297–298, 309–334 traditional terms in analyses of, 304 undirected/directed songs with, 334–336 wild strain birds complexity of, 336–337 willow warblers with, 308 zebra finch development with, 304, 314–315, 326–327 zebra finch with, 304, 314–315, 326–327 Zonotrichia leucophrys with, 304 Specificity ant protection mutualisms with, 78, 87 cleaning mutualisms with, 76, 87 game structures with, 63, 66, 69, 72, 74, 76, 78, 81, 82, 84, 87, 91 mixed species aggregations with, 80, 87 Mu¨llerian mimicry with, 82, 87 mutualistic systems parameter of, 63, 66 nutrition mutualisms with, 84 pollinating seed parasite mutualisms with, 72, 87 pollination mutualisms with, 69, 87 protection mutualisms with, 76, 78, 80, 82, 87 seed dispersal mutualisms with, 74, 87 symbioses with, 84, 87 transportation mutualisms with, 69, 72, 74, 87 Spider monkeys, food transfer with, 269, 273 Spiders. See Sexual cannibalism Starlings, song syntax of, 307–308 Stealing, food transfer with, 267 Sterna hirundo, 196 Sturnus vulgaris, 183, 191 song syntax of, 307–308

INDEX

Sula granti, 196 Sula nebouxii, 196 Suricata suricatta, 275 Symbioses, game structures with, 83–86, 87

T Tachycineta bicolor, 197 Tactile stimulation, infant learning with, 105 Tamarins, food transfer with, 269, 272–273, 274, 279, 282 Tarsius spectrum, 272, 274 Teaching, food transfer in, 285–286 Termite societies, reproductive conflict in, 35 Territoriality, androgens levels with, 186–187 Territoriality effects, social modulation of androgens with, 198–199 Thyreus, 365 Tidarren argo, sexual cannibalism in, 137 Tinbergen’s four questions Bengalese finch anterior forebrain pathway with, 321–322 Bengalese finch behavioral development with, 329 Bengalese finch brain in, 315 Bengalese finch developmental process with, 327, 328 Bengalese finch electrophysiological studies with, 322–323, 324 Bengalese finch HVC with, 315, 317–319, 320, 322–323, 324, 325, 326–327 Bengalese finch NIf with, 315, 319–321 Bengalese finch perceptual studies with, 323–325 Bengalese finch posterior brain pathway with, 315, 316–321 Bengalese finch RA with, 315, 316–317, 318, 323, 326–327

495

INDEX

Bengalese finch side of dominance with, 323–325 Bengalese finch ts nerve with, 316–317 bird song complexity reinforced with, 331–332 bird song control system architecture with, 314–315 bird song development summary for, 329 bird song development with, 325–329 bird song evolution in, 309–313 bird song evolution summary with, 313 bird song function summary for, 334 bird song function with, 329–334 bird song in, 297–298, 309–334 bird song mechanisms with, 313–325, 326 bird song morphology comparisons in, 311–312 bird song note complexity comparisons within, 312–313 bird song overall sound density with, 313 bird song phonology development with, 327 bird song study convergence with, 298 bird song study divergence with, 297–298 bird song syntax development with, 328–329 copulation solicitation assay with, 332 lesion studies with, 316–322 nest-building behavior with, 332–333 social modulation of androgens and, 188 Tinca tinca, androgen responsiveness in, 181 Tit-for-Tat strategy, 60, 66 Tracheosyringeal branch of hypoglossal nerve (ts nerve) Bengalese finch song syntax with, 316–317 zebra finch with, 314

Transportation mutualisms, game structures with, 66–75 Trigona, 348, 434, 440 Trigona canifrons, 426, 428 Trigona fulviventris, 430 Trigona hyalinata, 430 Tropheus moorii, 181, 182 ts nerve. See Tracheosyringeal branch of hypoglossal nerve

U UCT. See Upper critical temperatures Unconditioned stimulus (US), Pavlovian conditioning with, 246 Upper critical temperatures (UCT), bees with, 357, 358 Urophonius brachycentrus, sexual cannibalism in, 139 Urophonius jheringii, sexual cannibalism in, 139 US. See Unconditioned stimulus

V Vespa crabro, reproductive conflict in, 40 Vespula vulgaris, reproductive conflict in, 18 Vulpes vulpes, 206, 275

W Wasps pheromones in, 13 reproductive conflict with, 8, 13 Willow warblers, song syntax of, 308 Winner-loser effects, social modulation of androgens with, 200–202 Worker policing, insect reproductive conflict with, 16–19

496

INDEX

X Xylocopa activity patterns of, 348, 362, 370, 394–396, 400, 406, 407, 434 female nesting cycle of, 370 thermoregulation of, 362, 400 Xylocopa californica, 383 Xylocopa fenestrata, 400 Xylocopa pubescens activity patterns with, 361, 400 heat exchange of, 361 Xylocopa sulcatipes, heat exchange of, 361

Xylocopa tranquebarica, 366 Xylocopa varipuncta, 386

Z Zebra finch HVC of, 314, 315 RA of, 314, 315 song control system architecture with, 314–315 song syntax of, 304, 314–315 ts nerve of, 314 Zonotrichia leucophrys, 183 song syntax in, 304

Contents of Previous Volumes

Volume 18 Song Learning in Zebra Finches (Taeniopygia guttata): Progress and Prospects PETER J. B. SLATER, LUCY A. EALES, AND N. S. CLAYTON Behavioral Aspects of Sperm Competition in Birds T. R. BIRKHEAD Neural Mechanisms of Perception and Motor Control in a Weakly Electric Fish WALTER HEILIGENBERG Behavioral Adaptations of Aquatic Life in Insects: An Example ANN CLOAREC The Cicadian Organization of Behavior: Timekeeping in the Tsetse Fly, A Model System JOHN BRADY

The Evolution of Courtship Behavior in Newts and Salamanders T. R. HALLIDAY Ethopharmacology: A Biological Approach to the Study of Drug-Induced Changes in Behavior A. K. DIXON, H. U. FISCH, AND K. H. MCALLISTER Additive and Interactive Effects of Genotype and Maternal Environment PIERRE L. ROUBERTOUX, MARIKA NOSTEN-BERTRAND, AND MICHELE CARLIER Mode Selection and Mode Switching in Foraging Animals GENE S. HELFMAN Cricket Neuroethology: Neuronal Basis of Intraspecific Acoustic Communication FRANZ HUBER Some Cognitive Capacities of an African Grey Parrot (Psittacus erithacus) IRENE MAXINE PEPPERBERG

Volume 19 Volume 20 Polyterritorial Polygyny in the Pied Flycatcher P. V. ALATALO AND A. LUNDBERG Kin Recognition: Problems, Prospects, and the Evolution of Discrimination Systems C. J. BARNARD Maternal Responsiveness in Humans: Emotional, Cognitive, and Biological Factors CARL M. CORTER AND ALISON S. FLEMING

Social Behavior and Organization in the Macropodoidea PETER J. JARMAN The t Complex: A Story of Genes, Behavior, and Population SARAH LENINGTON The Ergonomics of Worker Behavior in Social Hymenoptera PAUL SCHMID-HEMPEL 497

498

CONTENTS OF PREVIOUS VOLUMES

‘‘Microsmatic Humans’’ Revisited: The Generation and Perception of Chemical Signals BENOIST SCHAAL AND RICHARD H. PORTER

Parasites and the Evolution of Host Social Behavior ANDERS PAPE MOLLER, REIJA DUFVA, AND KLAS ALLANDER

Lekking in Birds and Mammals: Behavioral and Evolutionary Issues R. HAVEN WILEY

The Evolution of Behavioral Phenotypes: Lessons Learned from Divergent Spider Populations SUSAN E. RIECHERT

Volume 21

Proximate and Developmental Aspects of Antipredator Behavior E. CURIO

Primate Social Relationships: Their Determinants and Consequences ERIC B. KEVERNE The Role of Parasites in Sexual Selection: Current Evidence and Future Directions MARLENE ZUK Conceptual Issues in Cognitive Ethology COLIN BEER Response in Warning Coloration in Avian Predators W. SCHULER AND T. J. ROPER Analysis and Interpretation of Orb Spider Exploration and Web-Building Behavior FRITZ VOLLRATH Motor Aspects of Masculine Sexual Behavior in Rats and Rabbits GABRIELA MORALI AND CARLOS BEYER On the Nature and Evolution of Imitation in the Animal Kingdom: Reappraisal of a Century of Research A. WHITEN AND R. HAM

Volume 22 Male Aggression and Sexual Coercion of Females in Nonhuman Primates and Other Mammals: Evidence and Theoretical Implications BARBARA B. SMUTS AND ROBERT W. SMUTS

Newborn Lambs and Their Dams: The Interaction That Leads to Sucking MARGARET A. VINCE The Ontogeny of Social Displays: Form Development, Form Fixation, and Change in Context T. G. GROOTHUIS

Volume 23 Sneakers, Satellites, and Helpers: Parasitic and Cooperative Behavior in Fish Reproduction MICHAEL TABORSKY Behavioral Ecology and Levels of Selection: Dissolving the Group Selection Controversy LEE ALAN DUGATKIN AND HUDSON KERN REEVE Genetic Correlations and the Control of Behavior, Exemplified by Aggressiveness in Sticklebacks THEO C. M. BAKKER Territorial Behavior: Testing the Assumptions JUDY STAMPS Communication Behavior and Sensory Mechanisms in Weakly Electric Fishes BERND KRAMER

CONTENTS OF PREVIOUS VOLUMES

Volume 24 Is the Information Center Hypothesis a Flop? HEINZ RICHNER AND PHILIPP HEEB Maternal Contributions to Mammalian Reproductive Development and the Divergence of Males and Females CELIA L. MOORE Cultural Transmission in the Black Rat: Pine Cone Feeding JOSEPH TERKEL The Behavioral Diversity and Evolution of Guppy, Poecilia reticulata, Populations in Trinidad A. E. MAGURRAN, B. H. SEGHERS, P. W. SHAW, AND G. R. CARVALHO Sociality, Group Size, and Reproductive Suppression among Carnivores SCOTT CREEL AND DAVID MACDONALD Development and Relationships: A Dynamic Model of Communication ALAN FOGEL Why Do Females Mate with Multiple Males? The Sexually Selected Sperm Hypothesis LAURENT KELLER AND HUDSON K. REEVE

499

An Overview of Parental Care among the Reptilia CARL GANS Neural and Hormonal Control of Parental Behavior in Birds JOHN D. BUNTIN Biochemical Basis of Parental Behavior in the Rat ROBERT S. BRIDGES Somatosensation and Maternal Care in Norway Rats JUDITH M. STERN Experiential Factors in Postpartum Regulation of Maternal Care ALISON S. FLEMING, HYWEL D. MORGAN, AND CAROLYN WALSH Maternal Behavior in Rabbits: A Historical and Multidisciplinary Perspective ´ LEZ-MARISCAL GABRIELA GONZA AND JAY S. ROSENBLATT Parental Behavior in Voles ZUOXIN WANG AND THOMAS R. INSEL Physiological, Sensory, and Experiential Factors of Parental Care in Sheep ´ VY, K. M. KENDRICK, F. LE E. B. KEVERNE, R. H. PORTER, AND A. ROMEYER

Cognition in Cephalopods JENNIFER A. MATHER

Socialization, Hormones, and the Regulation of Maternal Behavior in Nonhuman Simian Primates CHRISTOPHER R. PRYCE

Volume 25

Field Studies of Parental Care in Birds: New Data Focus Questions on Variation among Females PATRICIA ADAIR GOWATY

Parental Care in Invertebrates STEPHEN T. TRUMBO Cause and Effect of Parental Care in Fishes: An Epigenetic Perspective STEPHEN S. CRAWFORD AND EUGENE K. BALON Parental Care among the Amphibia MARTHA L. CRUMP

Parental Investment in Pinnipeds FRITZ TRILLMICH Individual Differences in Maternal Style: Causes and Consequences of Mothers and Offspring LYNN A. FAIRBANKS

500

CONTENTS OF PREVIOUS VOLUMES

Mother–Infant Communication in Primates DARIO MAESTRIPIERI AND JOSEP CALL Infant Care in Cooperatively Breeding Species CHARLES T. SNOWDON

Volume 27 The Concept of Stress and Its Relevance for Animal Behavior DIETRICH VON HOLST Stress and Immune Response VICTOR APANIUS

Volume 26 Sexual Selection in Seawood Flies THOMAS H. DAY AND ´ S. GILBURN ANDRE Vocal Learning in Mammals VINCENT M. JANIK AND PETER J. B. SLATER Behavioral Ecology and Conservation Biology of Primates and Other Animals KAREN B. STRIER How to Avoid Seven Deadly Sins in the Study of Behavior MANFRED MILINSKI Sexually Dimorphic Dispersal in Mammals: Patterns, Causes, and Consequences LAURA SMALE, SCOTT NUNES, AND KAY E. HOLEKAMP Infantile Amnesia: Using Animal Models to Understand Forgetting MOORE H. ARNOLD AND NORMAN E. SPEAR Regulation of Age Polyethism in Bees and Wasps by Juvenile Hormone SUSAN E. FAHRBACH Acoustic Signals and Speciation: The Roles of Natural and Sexual Selection in the Evolution of Cryptic Species GARETH JONES

Behavioral Variability and Limits to Evolutionary Adaptation P. A. PARSONS Developmental Instability as a General Measure of Stress ANDERS PAPE MOLLER Stress and Decision-Making under the Risk of Predation: Recent Developments from Behavioral, Reproductive, and Ecological Perspectives STEVEN L. LIMA Parasitic Stress and Self-Medication in Wild Animals G. A. LOZANO Stress and Human Behavior: Attractiveness, Women’s Sexual Development, Postpartum Depression, and Baby’s Cry RANDY THORNHILL AND F. BRYANT FURLOW Welfare, Stress, and the Evolution of Feelings DONALD M. BROOM Biological Conservation and Stress HERIBERT HOFER AND MARION L. EAST

Volume 28

Understanding the Complex Song of the European Starling: An Integrated Ethiological Approach MARCEL EENS

Sexual Imprinting and Evolutionary Processes in Birds: A Reassessment CAREL TEN CATE AND DAVE R. VOS

Representation of Quantities by Apes SARAH T. BOYSEN

Techniques for Analyzing Vertebrate Social Structure Using Identified

CONTENTS OF PREVIOUS VOLUMES

Individuals: Review and Recommendations HAL WHITEHEAD AND SUSAN DUFAULT Socially Induced Infertility, Incest Avoidance, and the Monopoly of Reproduction in Cooperatively Breeding African Mole-Rats, Family Bathyergidae NIGEL C. BENNETT, CHRIS G. FAULKES, AND JENNIFER U. M. JARVIS Memory in Avian Food Caching and Song Learning: A General Mechanism or Different Processes? NICOLA S. CLAYTON AND JILL A. SOHA Long-Term Memory in Human Infants: Lessons in Psychobiology CAROLYN ROVEE-COLLIER AND KRISTIN HARTSHORN Olfaction in Birds TIMOTHY J. ROPER Intraspecific Variation in Ungulate Mating Strategies: The Case of the Flexible Fallow Deer SIMON THIRGOOD, JOCHEN LANGBEIN, AND RORY J. PUTMAN

Volume 29 The Hungry Locust STEPHEN J. SIMPSON AND DAVID RAUBENHEIMER Sexual Selection and the Evolution of Song and Brain Structure in Acrocephalus Warblers CLIVE K. CATCHPOLE Primate Socialization Revisited: Theoretical and Practical Issues in Social Ontogeny BERTRAND L. DEPUTTE

501

Ultraviolet Vision in Birds INNES C. CUTHILL, JULIAN C. PARTRIDGE, ANDREW T. D. BENNETT, STUART C. CHURCH, NATHAN S. HART, AND SARAH HUNT What Is the Significance of Imitation in Animals? CECILIA M. HEYES AND ELIZABETH D. RAY Vocal Interactions in Birds: The Use of Song as a Model in Communication DIETMAR TODT AND MARC NAGUIB

Volume 30 The Evolution of Alternative Strategies and Tactics H. JANE BROCKMANN Information Gathering and Communication during Agonistic Encounters: A Case Study of Hermit Crabs ROBERT W. ELWOOD AND MARK BRIFFA Acoustic Communication in Two Groups of Closely Related Treefrogs H. CARL GERHARDT Scent-Marking by Male Mammals: Cheat-Proof Signals to Competitors and Mates L. M. GOSLING AND S. C. ROBERTS Male Facial Attractiveness: Perceived Personality and Shifting Female Preferences for Male Traits across the Menstrual Cycle IAN S. PENTON-VOAK AND DAVID I. PERRETT The Control and Function of Agonism in Avian Broodmates HUGH DRUMMOND

502

CONTENTS OF PREVIOUS VOLUMES

Volume 31 Conflict and Cooperation in a Female-Dominated Society: A Reassessment of the ‘‘Hyperaggressive’’ Image of Spotted Hyenas MARION L. EAST AND HERIBERT HOFER Birdsong and Male–Male Competition: Causes and Consequences of Vocal Variability in the Collared Dove (Streptopelia decaocto) CAREL TEN CATE, HANS SLABBEKOORN, AND MECHTELD R. BALLINTIJN Imitation of Novel Complex Actions: What Does the Evidence from Animals Mean? RICHARD W. BYRNE Lateralization in Vertebrates: Its Early Evolution, General Pattern, and Development LESLEY J. ROGERS Auditory Scene Analysis in Animal Communication STEWART H. HULSE Electric Signals: Predation, Sex, and Environmental Constraints PHILIP K. STODDARD How to Vocally Identify Kin in a Crowd: The Penguin Model THIERRY AUBIN AND PIERRE JOUVENTIN

Volume 32 Self-Organization and Collective Behavior in Vertebrates IAIN D. COUZIN AND JENS KRAUSE Odor-Genes Covariance and Genetic Relatedness Assessments: Rethinking

Odor-Based Recognition Mechanisms in Rodents JOSEPHINE TODRANK AND GIORA HETH Sex Role Reversal in Pipefish ANDERS BERGLUND AND GUNILLA ROSENQVIST Fluctuating Asymmetry, Animal Behavior, and Evolution JOHN P. SWADDLE From Dwarf Hamster to Daddy: The Intersection of Ecology, Evolution, and Physiology That Produces Paternal Behavior KATHERINE E. WYNNE-EDWARDS Paternal Behavior and Aggression: Endocrine Mechanisms and Nongenomic Transmission of Behavior CATHERINE A. MARLER, JANET K. BESTER-MEREDITH, AND BRIAN C. TRAINOR Cognitive Ecology: Foraging in Hummingbirds as a Model System SUSAN D. HEALY AND T. ANDREW HURLY

Volume 33 Teamwork in Animals, Robots, and Humans CARL ANDERSON AND NIGEL R. FRANKS The ‘‘Mute’’ Sex Revisited: Vocal Production and Perception Learning in Female Songbirds KATHARINA RIEBEL Selection in Relation to Sex in Primates JOANNA M. SETCHELL AND PETER M. KAPPELER

CONTENTS OF PREVIOUS VOLUMES

Genetic Basis and Evolutionary Aspects of Bird Migration PETER BERTHOLD

Precursors and Limitations for the Evolution of Language ¨ HLER KLAUS ZUBERBU

Vocal Communication and Reproduction in Deer DAVID REBY AND KAREN MCCOMB

Vocal Self-stimulation: From the Ring Dove Story to Emotion-Based Vocal Communication MEI-FANG CHENG

Referential Signaling in Non-Human Primates: Cognitive

503

Harald Lachnit ET AL., Fig. 4. Activity pattern of glomeruli in the antennal lobe of the bee when stimulated with 37 different odors. High activity as measured optophysiologically is indicated by red/dark colors. Top left: Schematic representation of the antennal lobe with the numbers of the glomeruli measured with the optophysiological technique (see text). The 13 pictures (top right) show the activity patterns of different olfactory stimuli ranging from single odors, mixtures of two or three compounds, or mixtures composed of many compounds such as in the case of floral odors (Cit, citral; Ger, geraniol; Iso, isoamylacetate; Pfm, peppermint oil; Org, orange; Car, carnation; Lnd, lime blossom; Lim, limonene; Cin, cineol; Eug, eugenol; Lio, linalool; Mnt, menthol; Cio, dl-citronellol. Bottom left: Twenty-four activity patterns showing the activity patterns of alkanes (first row), primary alcohols (second row), aldehydes (third row), and secondary ketones (fourth row). The number of carbon atoms in each of these four chemical classes is indicated above each column (C-5 to C-10). Bottom right: Images as they appear in the microscope for the four respective odors in the C-10 column. Adapted from data published by Sachse et al. (1999) and Galizia et al. (1999).