Reproductive Skew in Vertebrates Reproductive skew is the study of how reproduction is partitioned in animal societies...
21 downloads
871 Views
10MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Reproductive Skew in Vertebrates Reproductive skew is the study of how reproduction is partitioned in animal societies. In many social animals reproduction is shared unequally, leading to a reproductive skew among group members. Skew theory investigates the genetic and ecological factors causal to the partitioning of reproduction in animal groups and may yield fundamental insights into the evolution of animal sociality. This book brings together new theory and empirical work, mostly in vertebrates, to test assumptions and predictions of skew models. It also gives an updated critical review of skew theory. The team of leading contributors cover a wide range of species, from insects to humans, and discuss both ultimate (evolutionary) and proximate (immediate) factors influencing reproductive skew. Academic researchers and graduate students alike with an interest in evolution and sociality will find this material stimulating and exciting. r e i n m a r h a g e r is an NERC Research Fellow in evolutionary biology at the University of Manchester. He has been awarded University Fellowship by the German Science Foundation and was elected Senior Rouse Ball Scholar at Trinity College, Cambridge. d r . c l a r a b . j o n e s , p h . d . , p h . d . is a scientist and consultant trained by several universities, research agencies, and corporations. She has published more than 100 texts (both technical, theoretical, empirical, and popular), including several books. Currently, Dr. Clara resides in The Americas, continuing both academic, theoretical, scientific, applied, and corporate projects.
Reproductive Skew in Vertebrates Proximate and Ultimate Causes Edited by Reinmar Hager University of Manchester
Clara B. Jones Fayetteville State University, North Carolina
cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sa˜o Paulo, Delhi Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521864091 ª Cambridge University Press 2009 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2009 Printed in the United Kingdom at the University Press, Cambridge A catalog record for this publication is available from the British Library Library of Congress Cataloging in Publication data Hager, Reinmar, 1974– Reproductive skew in vertebrates : proximate and ultimate causes / Reinmar Hager, Clara B. Jones. p. cm. Includes bibliographical references and index. ISBN 978-0-521-86409-1 (hardback) 1. Vertebrates–Reproduction. 2. Sexual selection in animals. I. Jones, Clara B. II. Title. QP251.H13 2009 591.560 2–dc22 2009007300 ISBN 978-0-521-86409-1 hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
Contents
List of contributors page vii Foreword Sandra L. Vehrencamp xi Preface xix
Part I 1
Reproductive skew theory
1
Models of reproductive skew: outside options and the resolution of reproductive conflict 3 Rufus A. Johnstone and Michael A. Cant
2
Reproductive conflict and the evolution of menopause
24
Michael A. Cant, Rufus A. Johnstone, and Andrew F. Russell
Part II 3
Testing assumptions and predictions of skew models
51
Reproductive skew in female-dominated mammalian societies 53 Kay E. Holekamp and Anne L. Engh
4
The effects of heterogeneous regimes on reproductive skew in eutherian mammals 84 Clara B. Jones
5
Social skew as a measure of the costs and benefits of group living in marmots 114 Thea B. Wang, Peter Nonacs, and Daniel T. Blumstein
6
Explaining variation in reproductive skew among male langurs: effects of future mating prospects and ecological factors 134 Reinmar Hager
v
vi Contents
7
The causes and consequences of reproductive skew in male primates 165 Nobuyuki Kutsukake and Charles L. Nunn
8
Sociality and reproductive skew in horses and zebras
196
Daniel I. Rubenstein and Cassandra M. Nun˜ez
9
Reproductive skew in avian societies
227
Walter D. Koenig, Sheng-Feng Shen, Alan H. Krakauer, and Joseph Haydock
10
Reproductive skew in cooperative fish groups: virtue and limitations of alternative modeling approaches 265 Michael Taborsky
11. Reproductive skew in primitively eusocial wasps: how useful are current models? 305 Jeremy Field and Michael A. Cant
Part III 12
Resolving reproductive conflicts: behavioral and physiological mechanisms 335
Reproductive skew in female common marmosets: contributions of infanticide and subordinate self-restraint 337 David H. Abbott, Leslie Digby, and Wendy Saltzman
13
Reproductive skew in African mole-rats: behavioral and physiological mechanisms to maintain high skew
369
Chris G. Faulkes and Nigel C. Bennett
14
The causes of physiological suppression in vertebrate societies: a synthesis 397 Andrew J. Young
Part IV 15
Future directions
437
Understanding variation in reproductive skew: directions for future empirical research 439 Sarah J. Hodge
16
On the evolution of reproductive skew: a genetical view 467 W. Edwin Harris and Reinmar Hager
17
Social conflict resolution, life history, and the reconstruction of skew 480 Bernard J. Crespi
Taxonomic index Subject index
508 511
Contributors
David H. Abbott Wisconsin National Primate Research Center, University of Wisconsin, Madison, WI, USA
Nigel C. Bennett Mammal Research Institute, University of Pretoria, South Africa
Daniel T. Blumstein Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA, USA
Michael A. Cant Centre for Ecology and Conservation, University of Exeter, UK
Bernard J. Crespi Evolutionary Biology, Simon Fraser University, Burnaby, BC, Canada
Chris G. Faulkes School of Biological and Chemical Sciences, Queen Mary, University of London, UK
Jeremy Field Department of Biology and Environmental Science, University of Sussex, UK
Leslie Digby Department of Biological Anthropology and Anatomy, Duke University, Durham, NC, USA
Anne L. Engh Department of Zoology, University of Pennsylvania, Philadelphia, PA, USA
Reinmar Hager Faculty of Life Sciences, University of Manchester, UK
vii
viii List of Contributors
W. Edwin Harris Faculty of Life Sciences, University of Manchester, UK
Joseph Haydock Department of Biology, Gonzaga University, Spokane, WA, USA
Sarah J. Hodge Department of Zoology, University of Cambridge, UK
Kay E. Holekamp Department of Zoology, Michigan State University, East Lansing, MI, USA
Rufus A. Johnstone Department of Zoology, University of Cambridge, UK
Clara B. Jones Department of Psychology, Fayetteville State University, Fayetteville, NC, USA; and National Evolutionary Synthesis Center, Duke University, Durham, NC, USA
Walter D. Koenig Hastings Reservation and Museum of Vertebrate Zoology, University of California, Berkeley, CA, USA
Alan H. Krakauer Section of Evolution and Ecology, University of California, Davis, CA, USA
Nobuyuki Kutsukake Department of Evolutionary Studies of Biosystems, Graduate University for Advanced Studies Hayama, Miura-gun, Zushi, Kanagawa, Japan
Peter Nonacs Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA, USA
Cassandra M. Nun˜ez Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ, USA
Charles L. Nunn Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany; and Department of Integrative Biology, University of California, Berkeley, CA, USA
Daniel I. Rubenstein Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ, USA
Andrew F. Russell Department of Animal and Plant Sciences, University of Sheffield, UK
List of Contributors
Wendy Saltzmann Department of Biology, University of California, Riverside, CA, USA
Sheng-Feng Shen Department of Neurobiology and Behavior, Cornell University, Ithaca, NY, USA
Michael Taborsky Department of Behavioural Ecology, University of Berne, Switzerland
Sandra L. Vehrencamp Department of Neurobiology and Behavior, Cornell University, Ithaca, NY, USA
Thea B. Wang Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA, USA
Andrew J. Young Centre for Ecology and Conservation, University of Exeter, UK
ix
Foreword
A brief history of skew theory New ideas in science don’t spring out of nowhere; they combine and build from earlier ones. The concept of reproductive skew is no different, and it is nice to have this opportunity to look back, over 30 years ago now, and identify the various sources and give credit where it’s due. The story begins in 1974 at the 16th International Ornithological Congress in Canberra, Australia. During that long flight to Australia from Johannesburg, South Africa, the pilot came on the loudspeaker and announced that the United States President Richard Nixon had just resigned in disgrace over the Watergate scandal. Having been in the field in Africa all summer, this was a shocking return to civilization. Ian Rowley had convened the first symposium on cooperative breeding in birds, with himself, Lew Grimes, Glen Woolfenden, and Amotz Zahavi presenting surveys of the cooperatively breeding species in their respective continents (Australia, Africa, the Americas, and Europe) (Grimes 1976, Rowley 1976, Woolfenden 1976, Zahavi 1976). All four speakers noted that the helper-at-the-nest form of cooperative breeding, where offspring remain on the parental territory, delay breeding, and assist with the care of subsequent broods of their parents, was by far the most common one. Cooperative breeding was associated with a variety of habitats and climates, but most were characterized by sedentary residence on territories or fixed home ranges. Habitat saturation was identified by Woolfenden and Zahavi as the primary force favoring prolonged retention of offspring on the parental territory. During the contributed talk sessions, two of us, John Craig and I, spoke about our work on two species that differed significantly from the usual pattern in having multiple males and females breeding in a communal nest: groove-billed anis and pukeko. A lively workshop discussion followed, in which we argued about whether kin altruism or individually selected selfish
xi
xii Foreword behavior was the more important driving force for the evolution of cooperation. I tried to focus the discussion on the question of what might be driving the two different forms of cooperative breeding, the helper system versus the communal system, and suggested that lack of reproduction by helpers was the more pressing question to answer. Woolfenden glared at me and said no, the burden was on me, Craig, and others to explain when communal nesting would be favored, since helping was the norm and communal nesting the rarer form. The gauntlet had been thrown down, and I left the meeting determined to answer this question. In 1976, with my thesis finished and my first academic job beginning, a second motivator arose: teaching a new course (jointly with spouse Jack Bradbury). E. O. Wilson’s Sociobiology (Wilson 1975) had just appeared, and became the textbook for our course of the same name. Wilson’s review of the two alternative routes to sociality, the familial and the communal routes, was inspiring, and it quickly became clear that many animal groups, including social insects, spiders, birds, and mammalian carnivores, all showed these two forms of cooperative breeding (Eisenberg 1966, Wilson 1971, Lin & Michener 1972, Brown 1974). I felt compelled to come up with an explanation for the students. Hamilton’s rule (Hamilton 1964) could explain the altruism of helpers towards close kin, but it couldn’t explain the aggregation and mutual cooperation of unrelated individuals in a communal system. Richard Alexander’s insightful review of animal social behavior (Alexander 1974) held the next major key: he clearly articulated that an individual’s gain from remaining in a social group had to be compared to the alternatives of living alone or taking the risk of finding and joining another group. He also noted that different competitive dynamics were likely to occur in different types of groups, i.e. family groups versus groups of unrelated individuals. Finally, he asserted that group-living animals invariably form dominance hierarchies in which dominants often gain at the expense of subordinates, and coined the phrase “parental manipulation of progeny” to describe selection on parents to reduce the reproduction of certain offspring in order to increase their inclusive fitness via other offspring. A new perspective started to emerge: a focus on the power of the dominant and its greater ability to shape group composition and bias reproductive shares to its advantage. But the leverage of the dominant had limits if subordinates could opt to leave. Under conditions of habitat saturation, offspring would have poor outside options and therefore would be forced to remain with their parents, where they could be prevented from breeding by parental domination but able to increase their inclusive fitness by helping. By contrast, in unsaturated conditions groups would only form if there were some
Foreword type of mutualistic advantage, and dominants would be limited in their leverage to demand the cooperation of subordinates unless subordinates obtained enough direct fitness benefits to make staying in the group worthwhile. I first worked up the graphical analysis of the limits of dominant manipulation in unrelated groups, which then allowed me to incorporate Hamilton’s rule to compute the inclusive fitness break-even point within and outside the group for the case of related group members. Only three variables, combined in a very simple model, were required to explain the full range of high- and low-skew breeding systems: the benefit of group versus solitary breeding, the availability of options for breeding outside the group, and the coefficient of relatedness. I presented the bare bones of the model at the next Ornithological Congress in Berlin, 1978 (Vehrencamp 1980). Woolfenden was there, but he did not offer any comments. Stephen Emlen, my former thesis advisor, who also attended this congress, clearly did grasp the significance of the idea and subsequently came out with his own version of it, without formulating a fully quantitative model (Emlen 1982). It took a few more years to completely vet the mathematics of the model and explore the parameter space with different values of relatedness, group benefits, and ecological constraints. I appreciate the feedback from my mathematically competent colleagues at UC San Diego, Mike Gilpin, Ted Case, and Kurt Fristrup. I was pleased when John Maynard Smith took an interest in the model during our 1980 sabbatical leave in Sussex, UK; he not only incorporated it as an example of a two-stage game into the game theory book he was writing (Maynard Smith 1982), but he “did the sums” in a slightly different way by assuming that in larger groups subordinates made individual decisions about whether to stay or leave. To explain the occurrence of equitably breeding groups of relatives within the model’s framework, some factor that prevented enforcement by dominants such as constraints on control, reproductive inefficiency caused by within-group conflict, or coalitions of subordinates had to be invoked. If these factors were widespread, such that observed skew was often less than the maximum predicted by the model, I surmised that the original model would not be very useful. Occurrence of skews greater than predicted by the model would certainly falsify it. In the final revision of the 1983 paper (Vehrencamp 1983), a reviewer and editor were critical of the use of the word “skew” as a verb and insisted I use the term “bias.” Once the paper was published, I figured the idea would have to cook. Since it seemed very difficult to conduct experimental tests and manipulations on birds, and the molecular means for determining paternity and maternity were not fully developed, I switched to research on animal communication.
xiii
xiv Foreword Kern Reeve, along with colleagues Francis Ratnieks and Laurent Keller, gave the theoretical model a shot in the arm in the early 1990s by simplifying the math, limiting group size to two, greatly clarifying the specific predictions and tests of the model, and extending the model to make predictions about when groups should be stable versus unstable (Reeve 1991, Reeve & Ranieks 1993, Keller & Reeve 1994). In addition, they added the possibility that the degree of skew might be limited by subordinate fighting for complete control of the group’s resources, derived skew in parent–offspring associations, and generalized the skew model to N-person groups with diminishing group benefits as group size increases (Reeve & Keller 1995, 1997, Reeve & Emlen 2000). With their background in social insects, they shifted the terminology and mechanisms of group formation to a perspective different from my vertebrate approach. Whereas I conceived of groups of potential breeders initially forming for reasons of birth location or ecological conditions, followed by the emergence of a dominance hierarchy and the subsequent suppression of reproduction by subordinates just short of the point where they should leave, Reeve envisioned a dominant overwintering queen returning first to the birth nest site and then enticing later arrivals to join in a group. Words like “incentive,” “payment,” “concession,” “negotiation,” and “social contract” were incorporated to describe the processes of group formation and reproductive partitioning. Many lab groups were engaged in empirical testing of the model’s predictions. To my surprise, the reproductive biasing models suddenly became “skew theory.” A symposium devoted to skew theory was held in Sheffield in 1997 to bring together vertebrate and invertebrate perspectives. Along with the notoriety and testing came a wave of skepticism articulated by Tim CluttonBrock, who expected dominants to have perfect control and both dominants and subordinates to have perfect knowledge of reproductive shares and the availability of outside options (Clutton-Brock 1998). Beginning in 1998, an explosion of new skew models by Reeve, Michael Cant, Rufus Johnstone, Hanna Kokko, and others appeared (Cant 1998, Reeve et al. 1998, Cant & Johnstone 1999, Johnstone & Cant 1999, Johnstone et al. 1999, Kokko and Johnstone 1999, Crespi & Ragsdale 2000, Reeve 2000, Cant & Field 2001, Kokko et al. 2001, to mention only a few). In response to the criticism, these models variously modified the assumptions and mechanisms of dominant control, incorporated additional fitness components such as the potential for subordinates to inherit the dominant position, and considered the costs of enforcing skew and producing more young. A major alternative model was also developed, the tug-of-war model (Reeve et al. 1998), which assumed that
Foreword reproductive shares were the outcome of a costly conflict between two individuals of different competitive abilities. Ecological constraints do not influence reproductive partitioning in this model, but both participants reduce their combative effort when more closely related so skew tends to be unaffected or even lower than when they are less closely related. Renewed efforts to test the alternative predictions of the tug-of-war and optimal skew models sometimes found better support for the tug-of-war predictions. A new synthetic model by Reeve and Shen (2006) that combines the tug-of-war process within the constraints of options outside the group, called the bordered tug-of-war model, holds great promise as a single flexible model. This model assumes that dominant and subordinate continually adjust their payments and selfish tug-of-war efforts in response to each other (K. Reeve, personal communication). In a real way, it better incorporates the biasing mechanisms and constraints that I envisioned 30 years ago. By varying parameters that set whether one, both, or no parties concede some direct fitness to the partner to keep her/him from leaving, one can cover the whole range of traditional concessions, bordered tug-of-war, and pure tug-of-war models. Each of these three processes appears to operate under different conditions of ecological constraint, benefit to grouping, and relatedness. The current state of skew theory is well represented in the chapters of this book. Chapters by Taborsky on cooperatively breeding fish and by Kutsukake and Nunn on primates highlight the extreme range of ease and difficulty, respectively, of manipulating and measuring the key skew-determining parameters. It is heartening to read in the review of avian cooperative breeders by Koenig et al. that a meta-analysis largely supports the predictions of the traditional optimal skew model for birds. Chapters by Abbott et al., Faulkes and Bennett, and Young argue that in mammals, despite the high cost of offspring production, dominant females can attain a high skew through hormonal suppression and infanticide, for example. Jones, on the other hand, argues that high variability and unpredictability of ecological conditions limits the ability of dominants to control subordinates in socially flexible mammals. A useful review of reproductive skew studies in primitively eusocial wasps by Field and Cant points out that skew is often greater than predicted by the concessions model. They discuss whether this outcome is a clear falsification of the model, or whether either ecological constraints or inheritance of the dominant position has not been properly measured and considered. Hager also reports that future reproductive potential may be driving the observation of high reproductive skew in male langurs. Holekamp and Engh demonstrate the impact of the control one sex may have over skew in the opposite sex, a problem that continually arises in cases of vertebrate cooperative breeders containing
xv
xvi Foreword multiple members of both sexes. Cant and Johnstone explore a model similar to the bordered tug-of-war, but where mutual adjustments between dominant and subordinate are not allowed. Despite the skepticism and sometimes-heated disagreements (for example, see Magrath and Heinsohn 2000, Nonacs 2006, Nonacs 2007, and chapters by Hodge and Crespi), the collective body of theoretical and empirical work on reproductive skew has been and continues to be an engaging and fruitful field of study with broad ramification to many aspects of social behavior in animals and humans. Sandra L. Vehrencamp Cornell University
References Alexander, R. D. (1974). The evolution of social behavior. Annual Review of Ecology and Systematics, 5, 325–384. Brown, J. L. (1974). Alternative routes to sociality in jays. American Zoologist, 14, 63–80. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169. Cant, M. A. & Field, J. (2001). Helping effort and future fitness in cooperative animal societies. Proceedings of the Royal Society of London B, 268, 1959–1964. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology & Evolution, 13, 288–292. Crespi, B. J. & Ragsdale, J. E. (2000). A skew model for the evolution of sociality via manipulation: why it is better to be feared than loved. Proceedings of the Royal Society of London B, 267, 821–828. Eisenberg, J. F. (1966). The social organization of mammals. Handbuch der Zoologie, 10, 1–92. Emlen, S. T. (1982). The evolution of helping. I. An ecological constraints model. American Naturalist, 119, 29–39. Grimes, L. G. (1976). Cooperative breeding in African birds. Proceedings of the International Ornithological Congress, 16, 666–673. Hamilton, W. D. (1964). The genetical evolution of social behavior, I & II. Journal of Theoretical Biology, 7, 1–52. Johnstone, R. A. & Cant, M. A. (1999). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Johnstone, R. A., Woodroffe, R., Cant, M. A., & Wright, J. (1999). Reproductive skew in multimember groups. American Naturalist, 153, 315–331.
Foreword Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology & Evolution, 9, 98–102. Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 266, 571–578. Kokko, H., Johnstone, R. A., & Clutton-Brock, T. H. (2001). The evolution of cooperative breeding through group augmentation. Proceedings of the Royal Society of London B, 268, 187–196. Lin, M. & Michener, C. (1972). Evolution of sociality in insects. Quarterly Review of Biology, 47, 131–159. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Maynard, Smith J. (1982). Evolution and the Theory of Games. Cambridge: Cambridge University Press. Nonacs, P. (2006). The rise and fall of transactional skew theory in the model genus Polistes. Annales Zoologici Fennici, 43, 443–455. Nonacs, P. (2007). Tug-of-war has no borders: it is the missing model in reproductive skew theory. Evolution, 61, 1244–1250. Reeve, H. K. (1991). The social biology of Polistes. In K. Ross & R. Matdiews, eds., The Social Biology of Wasps. Ithaca, NY: Cornell University Press, pp. 99–148. Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Keller, L. (1995). Partitioning of reproduction in mother–daughter versus sibling associations: a test of optimal skew theory. American Naturalist, 145, 119–132. Reeve, H. K. & Keller, L. (1997). Reproductive bribing and policing evolutionary mechanisms for the suppression of within-group selfishness. American Naturalist, 150, S42–S58. Reeve, H. K. & Ranieks, F. L. W. (1993). Queen–queen conflict in polygynous societies: mutual tolerance and repoductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S. F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Rowley, I. (1976). Cooperative breeding in Australian birds. Proceedings of the International Ornithological Congress, 16, 657–666. Vehrencamp, S. L. (1980). To skew or not to skew? In R. No¨hring, ed., Acta XVII Congressus Internationalis Ornithologici. Vol. I. Berlin: Verlag der Deutschen Ornithologen-Gesellschaft. pp. 869–874.
xvii
xviii Foreword Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Wilson, E. O. (1971). The Insect Societies. Cambridge, MA: Belknap Press. Wilson, E. O. (1975). Sociobiology: the New Synthesis. Cambridge, MA: Belknap Press. Woolfenden, G. E. (1976). Cooperative breeding in American birds. Proceedings of the International Ornithological Congress, 16, 674–684. Zahavi, A. (1976). Cooperative breeding in Eurasian birds. Proceedings of the International Ornithological Congress, 16, 685–694.
Preface
We have collaborated on projects since 2001, and it became evident to us by 2004 that a book about reproductive skew incorporating theoretical, empirical, and review chapters might be timely. Because of our own research specializations, and because a large body of literature exists on skew in social insects, we generated a plan to prepare a volume on reproductive skew in vertebrates, approaching Cambridge University Press with a proposal. The present text represents our attempt to provide a “state of the art” overview of reproductive skew in vertebrate societies by some of the most active and highly regarded researchers in this field. It is our intention to highlight the most fundamental questions for students of reproductive skew, to assess the strengths and weaknesses of skew models, to critically evaluate skew in insect societies and skew in social vertebrates, and to identify important directions for future theoretical and empirical work. In her foreword to our volume, Sandra Vehrencamp has provided a brief overview of the history, theory, and empirical highlights of models of reproductive skew. It is clear from her personalized account that advances in the study of reproductive skew (the within-sex partitioning of reproduction within social groups) were presaged by early work in behavioral ecology demonstrating a relationship between dispersion and quality of limiting resources, in particular, food and nesting sites, as well as variations in social behavior, social organization, and mating systems within and between populations. The relationship between these factors is analyzed in models of reproductive skew that attempt to explain the partitioning of reproduction among individuals of the same sex in animals and offer a theoretical framework for understanding the formation of social groups. In addition to the central role of kin selection, skew theory identifies other fundamental processes that are key to the evolution of complex sociality, such as suppression of reproduction or the control of group membership, and it may thus enable us to
xix
xx Preface study one of the major transitions in evolution, from primarily solitary to social lifestyles, across a range of different taxa. Our book provides the reader with theoretical, empirical, and review chapters on a variety of model vertebrate systems exemplifying high, intermediate, and low skew and the causes and consequences of biased reproduction within groups. Furthermore, since social insects have been the classic exemplars of high-skew societies, we include a chapter on these organisms and their similarities and differences to social vertebrates. We think that a particularly helpful feature of our text is its future value as a reference tool. A related utility is the book’s presentation of many ideas for future research that have not been thoroughly investigated to date: for example, the evolution of low-skew societies and underlying mechanisms of suppression. An additional feature of Reproductive Skew in Vertebrates: Proximate and Ultimate Causes is its presentation of caveats about skew models for those who may be uncritically enthusiastic about these paradigms. The investigation of all aspects of reproductive skew is in its early stages, and we hope that our volume will help this field develop into a more mature, and critical, enterprise. As editors, we would be remiss if we failed to acknowledge and to express gratitude to the many individuals who have facilitated the process of our book’s actualization – from initial proposal to published text. We thank our initial contacts at Cambridge University Press whose interest in our ideas never faltered. Our editor at Cambridge, Martin Griffiths, has provided significant input, guidance, support and encouragement for our project. Without Martin’s expert ability to diagnose and to resolve challenges, our book may not have gone to press. Subsequent to the decision by Cambridge University Press to invite our initial book proposal and to distribute it for review, we received helpful and constructive criticism and advice from several researchers investigating the topic of reproductive skew. These critiques influenced our final decisions regarding our volume’s conceptual framework and organization, and also prompted us to include additional chapters on topics we had not previously considered. Indeed, the initial feedback we received from reviewers of our proposal was critical in stimulating our thinking about target questions and heightened our enthusiasm for our project, and we gratefully acknowledge the comments of these researchers. Our most profound thanks are extended to our contributors, who have demonstrated professionalism, expertise, patience, and, most important, good humor at every stage of our project. Without the willingness of these individuals to engage in frequent communication about their submissions and to receive our suggestions with grace, even when they may have disagreed with them, our book would not have been realized. We hope that our contributors
Preface will be pleased with the final text, as we are, and that they, their colleagues, and their students will value their contributions to Reproductive Skew in Vertebrates: Proximate and Ultimate Causes for many years to come. Among the contributors to our book that we wish to acknowledge by name are Sandra Vehrencamp and Bernard Crespi, who assumed responsibility for writing the foreword and the concluding chapter, respectively. These experts met unique challenges requiring the ability to summarize the field of reproductive skew, both retroactively and for the future. We hope that both of these contributions will help to place the book’s chapters in a broad context by linking each topic with past work, with the present state of the field, and with the literature on reproductive skew and related topics yet to be published. Finally, we wish to express personal thanks to those who have been particularly influential in our careers. Reinmar Hager is especially grateful to Rufus Johnstone, his thesis advisor at Cambridge University, for introducing him to skew theory, and to Yfke Hager, for help with editing. Clara B. Jones, likewise, thanks her dissertation advisor at Cornell University, William C. Dilger, for encouraging her interests in plants and animals and for facilitating her research interest in the evolution of social behavior. In the final analysis, it may be both necessary and sufficient to stress that we hope you find our volume a “good read” that will provide information, stimulate thinking, and generate original research, both theoretical and empirical, for its readers and their extended academic families. We look forward to these outputs as well as your measured feedback.
xxi
I
Reproductive skew theory
1
Models of reproductive skew: outside options and the resolution of reproductive conflict rufus a. j ohnstone and m ichael a. ca nt
Summary The two main types of skew models, transactional and compromise models, make different assumptions about the division of reproduction. Transactional models assume that one individual has full control over reproduction within the group, but may have to refrain from claiming all reproduction in order to prevent others leaving or evicting it from the group. Compromise models, by contrast, ignore outside options such as departing to breed elsewhere, but allow for incomplete control over reproduction within the group. Attempts to synthesize these two approaches have proved controversial. Here, we show that this controversy can be resolved using a simple principle from the economic literature on bargaining – the “outside option principle.” Even if outside options are available, they will influence the outcome of reproductive conflict within a group only if they yield greater payoffs than are available within the group. We present a novel synthetic model based on this principle, in which individuals engage in a tug-of-war over reproduction within a group, but may “ease off ” in their competitive effort in response to the threat of departure or eviction. We show that over a large range of parameter space, particularly when group productivity and relatedness among group members are high, these threats are not credible, so that opportunities outside the group do not influence the stable level of skew. However, when group productivity and relatedness are low, one or other of the players will typically ease off in competition in order to maintain group stability. Under these circumstances, outside Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
3
4
R. A. Johnstone, M. A. Cant options do influence skew. Tests which examine the relationship between skew and factors such as group productivity or ecological constraints are thus expected to yield variable results. The essential question is whether or not any of the members of a group stand to gain from its dissociation. The answer will determine whether or not outside options come into play.
Introduction The term reproductive skew refers to inequality in the distribution of breeding success among members of a group (Vehrencamp 1983, Keller & Reeve 1994). In high-skew societies, such as those of honeybees (Apis mellifera), Mexican jays (Aphelocoma ultramarina), or meerkats (Suricata suricatta), the distribution of reproductive opportunities is markedly unequal (see, e.g., Chapter 13 in this volume). One or a few breeders monopolize reproduction, while others are denied the opportunity to mate or to raise offspring (and may even, in the case of eusocial insects or naked mole-rats (Heterocephalus glaber), develop as sterile workers). In low-skew societies, by contrast, all individuals have similar opportunities to breed (see, e.g., Chapter 15). Inequality in breeding success is not a precisely defined concept. There are many different plausible ways in which to conceive and measure inequality in a multi-member group, leading to many different indices of skew (Kokko & Lindstro¨m 1997, Tsuji & Kasuya 2001). Moreover, inequality in actual reproductive success may be expected to arise simply by chance, even if all individuals have similar opportunities to breed (e.g. Haydock & Koenig 2002). Nevertheless, however one chooses to measure inequality, it is clear that there are dramatic and consistent differences in skew within and among species (Keller & Reeve 1994, Reeve & Keller 2001). Even closely related species otherwise similar in their ecology and behavior may differ markedly in this respect – compare, for instance, dwarf mongooses (Helogale parvula), in which typically only one female in a group breeds, with banded mongooses (Mungos mungo), in which most females breed in each attempt (Cant 2000). These differences in reproductive skew cry out for explanation. Models of reproductive skew attempt to provide an adaptive account of variation in skew both between and within species. They assume that there exists a conflict of interest among members of a group, such that each would benefit by obtaining a greater share of reproduction than is in the best interest of the others. Each model then predicts how (at evolutionary equilibrium) this conflict of interest will be resolved, depending upon various factors such as the benefits of group membership, the opportunity for independent breeding, and the ability of each group member to compete for resources or breeding
Models of reproductive skew opportunities, as well as to evict or exclude others from the group (Reeve & Ratnieks 1993, Keller & Reeve 1994, Johnstone 2000). Models of skew may differ in their assumptions about the extent of conflict between group members (e.g. Cant & Johnstone [1999] suggested that when production of offspring entails accelerating costs, individuals might all benefit from sharing reproductive opportunities). Most, however, are distinguished by their assumptions about the relative power of dominant and subordinate individuals. The most striking contrast in this respect is between transactional and compromise models of skew.
Transactional models of skew Transactional models of skew were the first to be developed. Vehrencamp’s (1979, 1983) seminal papers, in which she introduced the concept of reproductive skew, were built around a transactional model, and it is this approach that has been followed in most later analyses (e.g. Reeve & Ratnieks 1993, Johnstone & Cant 1999, Buston et al. 2007). In fact, when biologists talk of skew theory it is usually the transactional approach that they have in mind. The basic assumption of transactional models is that animals may concede reproductive opportunities to others, despite being potentially able to claim these opportunities for themselves, in order to maintain the stability of the group. This is advantageous because cooperation is presumed to yield productivity benefits. It may therefore pay to yield some reproduction to others in order to gain (or continue to enjoy) the benefits of associating with them. There are in fact two types of transactional model. Early analyses focused on reproductive concessions offered by dominants to retain helpful subordinates in the group – in these models, dominance takes the form of complete control over reproduction, with dominant individuals yielding breeding opportunities to subordinates so as to make it worth their while remaining in the group rather than departing (Vehrencamp 1979, 1983, Reeve & Ratnieks 1993, Reeve & Emlen 2000). By contrast with this notion of “concessions,” the “restraint” model focuses on reproductive concessions offered by subordinates to prevent the dominant evicting them – in these models, dominance takes the form of control over group membership, with subordinates refraining from claiming as large a share of reproduction as they might, lest the dominant eject them from the group (Clutton-Brock 1998, Johnstone & Cant 1999). Both types of model can be formalised in a similar way (Reeve & Ratnieks 1993, Johnstone & Cant 1999, Johnstone 2000, Buston et al. 2007). Consider a pair of individuals, related by a coefficient r. In association, the combined reproductive success of the pair, relative to that of an established lone breeder,
5
6
R. A. Johnstone, M. A. Cant is given by the parameter k (typically > 1). The expected reproductive success of an individual that disperses to breed elsewhere, again relative to that of an established lone breeder, is given by the parameter d (typically < 1, since we assume that dispersal entails some risk or cost). In a stable association, each individual must obtain sufficient reproductive success within the group that it would not gain either by leaving or by evicting the other (where eviction is possible). In the concession model, one “dominant” individual is assumed to have complete control over the distribution of reproduction within the group, subject only to the threat of departure by the “subordinate.” In this case, the subordinate in a stable group is expected to receive the minimum share of reproduction, pmin, that is compatible with group stability, i.e. a share that is just sufficient to ensure that leaving is not profitable. This share is given by ðkpmin dÞ þ rðkð1 pmin Þ 1Þ ¼ 0
ð1:1Þ
where the first term on the left-hand side represents the direct fitness impact of staying (rather than leaving) on the subordinate, and the second term the indirect fitness impact on the reproductive success of the dominant. Rearranging, we obtain pmin ¼
d r ðk 1 Þ kð1 rÞ
ð1:2Þ
(if d < r (k 1), then the subordinate does best to remain even if the dominant completely monopolizes reproduction). In the restraint model, by contrast, the “subordinate” is free to claim unsanctioned reproduction, subject only to the threat of eviction by the “dominant”. In this case, the dominant is expected to receive the minimum share of reproduction, qmin, that is compatible with group stability, i.e. a share that is just sufficient to ensure that evicting the subordinate is not profitable. This share is given by ðkqmin 1Þ þ r ðkð1 qmin Þ dÞ ¼ 0
ð1:3Þ
where the first term on the left-hand side represents the direct fitness impact of tolerating the subordinate’s presence (rather than evicting it from the group) on the dominant, and the second term the indirect fitness impact of toleration on the reproductive success of the subordinate. Rearranging, we obtain qmin ¼
1 r ð k dÞ k ð1 r Þ
ð1:4Þ
Models of reproductive skew In both models, the association will prove stable provided that pmin þ qmin < 1
ð1:5Þ
i.e. provided that the pair are together sufficiently productive that both may simultaneously receive at least their minimum required share. Substituting Equations 1.2 and 1.4 into 1.5, this yields the condition 1þd
ð1:6Þ
implying that the association will prove stable provided that the total productivity of both individuals would be reduced if it were to break up. Concessions and restraint
From Equations 1.2 and 1.6, we see that in the concession model the subordinate’s share of reproduction in a stable group increases with the opportunity for independent breeding (d), but decreases with the productivity of the group (k) and the degree of relatedness between the players (r). This makes intuitive sense – when there are greater benefits to be obtained elsewhere, the subordinate must receive a greater share of group productivity to make staying worthwhile; at the same time, the greater the productivity of the group, and the greater the relatedness between the players, the smaller the share required to satisfy this requirement. From Equations 1.4 and 1.6, however, we see that in the restraint model, the dominant’s share exhibits precisely the same trends (with the exception that it decreases with the opportunity for independent breeding only for r > 0). Consequently, the two models generally yield opposite predictions regarding skew. For instance, when group productivity increases, the concession model predicts a decrease in the subordinate’s share, leading to greater skew (provided that the dominant enjoys greater reproductive success than the subordinate). By contrast, the restraint model predicts a decrease in the dominant’s share, leading to reduced skew (again provided that the dominant enjoys greater reproductive success than the subordinate). The difference between the two models lies in the roles assigned to the two individuals. In the concession model, it is the subordinate that threatens to break up the group, and the dominant that must concede reproduction it could otherwise claim in order to maintain the association; in the terminology of Buston et al. (2007) it is the dominant that allocates reproduction to its partner. By contrast, in the restraint model it is the dominant that threatens to break up the group, and the subordinate that must concede reproduction; in this case it is the subordinate that is the allocator. The terms
7
8
R. A. Johnstone, M. A. Cant “dominant” and “subordinate” thus carry different meanings in the two models. In the concession model, dominance denotes control over the division of reproduction (subject to the threat of departure by the subordinate), while in the restraint model, dominance denotes control over group membership (i.e. the power to evict).
Compromise models The concession and restraint models yield opposing predictions about skew, but both are instances of the transactional approach. In both, one individual is assumed to exercise full control over the division of reproduction, subject only to the threat of group breakup (which may be initiated by the other individual). Compromise models instead assume that each member of the group can act selfishly to claim a greater share of breeding opportunities, at a cost to the productivity of the group as a whole (Clutton-Brock 1998, Reeve et al. 1998, Johnstone 2000). The outcome of the conflict over reproduction depends upon the level of selfish effort invested by each individual, and on their relative “strength.” In this kind of model, “dominance” typically takes the form of greater competitive ability – the dominant individual may be able to invest more effort in competition than can subordinates, or it may obtain a greater share for the same level of investment, due to superior resource-holding potential (Reeve et al. 1998). The other factor that can affect the outcome of the conflict in these models is relatedness among group members, which potentially influences the level of competitive effort that each individual will invest at equilibrium. The most influential compromise model, the tug-of-war game of Reeve et al. (1998), focuses (like the simple transactional models described above) on the interaction between two individuals, one dominant and one subordinate, who are related by a coefficient r. Both players simultaneously choose how much effort to invest in selfish competition over reproductive opportunities within the group. The levels of effort will be denoted x for the dominant and y for the subordinate. Total group productivity is equal to k(1xy), where the parameter k specifies group productivity relative to that of a lone breeder in the absence of competition. Productivity thus declines linearly with total expenditure by both players on selfish competition. The fraction of reproduction claimed by the dominant is equal to x/(x þ by), where the parameter b (<1) specifies the competitive ability of the subordinate relative to that of the dominant. Dominance in this context is thus defined by superior competitive ability.
Models of reproductive skew Given the above assumptions, the direct fitness payoffs to the two players from the tug-of-war are given by x x þ by by Ws ðx; yÞ ¼ kð1 x yÞ x þ by
Wd ðx; yÞ ¼ kð1 x yÞ
ð1:7Þ
Reeve et al. (1998) derived the (unique) stable pair of effort levels x* and y* in the basic tug-of-war game, each of which simultaneously maximizes (for the player in question, given the other’s behavior) the sum of its own direct fitness payoff plus r times that of the other player. These effort levels satisfy the first-order condition @Wd ðx; yÞ @Ws ðx; yÞ @Ws ðx; yÞ @Wd ðx; yÞ þr ¼ þr ¼0 @x @x @y @y
ð1:8Þ
and are given by 2 x ¼
3
b 7 6 2 rð1 bÞ 4qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 15 2ð1 bÞ 2 r2 ð1 bÞ þ4b 3 2
ð1:9Þ
1 2b þ r ð1 bÞ 7 6 y ¼ 41 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi5 2ð1 bÞ r2 ð1 bÞ2 þ4b
The share of reproduction obtained by the subordinate at this equilibrium increases with its relative competitive ability, b, but surprisingly it is largely insensitive to the relatedness between the competitors, r. As relatedness increases, both players reduce their competitive effort, with the result that skew changes little.
Outside options Much recent discussion of reproductive skew has focused on the issue of control. Can the dominant really prevent subordinates from claiming any share of reproduction, as concession models assume? Or are subordinates able to gain some reproductive success even against the best interests of the dominant, as in compromise models? Can subordinates even claim a large enough share that the threat of eviction becomes relevant? Framed in these terms, compromise and transactional approaches start to seem less clearly distinct. Perhaps concession and restraint models simply represent extreme cases on a continuum of
9
10
R. A. Johnstone, M. A. Cant dominant control – the former deal with cases in which the dominant has full control over reproduction, the latter with cases in which the dominant has none (except that granted by the threat of eviction). There remains, however, another fundamental difference between transactional and compromise models. The transactional models incorporate outside options – the possibility of leaving to breed elsewhere, or of evicting a competitor from the group. Indeed, it is these outside options that, in transactional analyses, set limits on the level of skew possible in a stable group. The fitness prospects of individuals that leave or are forced out of a group, usually referred to in terms of ecological constraints, thus have a major influence on skew in transactional analyses. By contrast, compromise models have generally ignored outside options, and focus only on the resolution of conflict within the group. Ecological constraints thus have no influence on the outcome of these analyses. In reality, it seems likely that animals will often have the opportunity to leave a group and breed elsewhere, or to evict others from the association. But at the same time, it also seems likely that no one individual will enjoy complete control over reproduction within the group. A realistic model of skew, therefore, needs to incorporate both competition over reproduction within the group, as in compromise models, and outside options, as in transactional models. How are these two approaches to be combined? While there have been several attempts to “synthesize” skew models in order to answer this question, there is still little agreement over which approach is appropriate ( Johnstone 2000, Reeve & Shen 2006, Nonacs 2007).
Synthetic models In a previous review paper (Johnstone 2000), one of us outlined a synthetic model of skew that attempted to incorporate outside options into the tug-of-war model of Reeve et al. (1998). In this combined model, animals were assumed first to yield concessions to one another, in the form of uncontested shares of group productivity, sufficient to ensure that neither would benefit by leaving the group. Subsequently, both were assumed to engage in a tug-of-war competition over the remaining, contested fraction of reproduction. The model suffers from a problem, however, in that the concessions required to prevent departure in the first step are calculated on the basis of total group productivity, overlooking the fact that some of this productivity will be squandered in competition during the second step. Because of the costs of competition, the original concessions lose some of their value, and can no longer be relied upon to ensure group stability when the outcome of the tug-ofwar is taken into account.
Models of reproductive skew Reeve and Shen (2006) attempted to deal with this problem in a modified synthetic model. They too allow individuals to concede uncontested shares of reproduction to one another, and assume that a tug-of-war takes place over the remaining contested fraction. However, they assume that the magnitude of the concessions and the intensity of effort in the tug-of-war are decided simultaneously. Consequently, they suggest that an “evolutionarily stable” outcome must satisfy two simultaneous conditions: (1) each player must be just indifferent between remaining in the group and choosing its outside option, and (2) each player’s competitive effort in the tug-of-war must be a best response to that of the other, given the magnitude of the concessions made by each (i.e. selfish efforts are at a Nash equilibrium). Once again, however, this approach seems to us problematic. First, it rests on the questionable assumption that the players use up all the potential benefits of group membership in the process of finding a resolution (see Nonacs 2007). Second, the putative solution is also based on the assumption that each player gives away a “free” share of reproduction to its opponent while simultaneously fighting hard for the remainder. Intuitively, it would make more sense for each player to reduce the size of the free share offered to its partner, and invest less effort in wasteful competition. This intuition is borne out by formal analysis – one can show that any solution of the form suggested by Reeve and Shen (2006) can be invaded by a mutant that offers a reduced “handout” and at the same time invests less in wasteful competition over the remainder. Nonacs (2007) forcefully critiques both of the above models, highlighting some of the same problems we have pointed out here. He then presents results of a simulation analysis in which individuals may exchange shares of uncontested reproduction while simultaneously engaging in a tug-of-war over the remainder. The results are that such exchange always proves unstable, and that levels of competitive effort evolve to the solution of the basic tug-of-war game. However, since the analysis omits any possibility of departure or eviction, this is unsurprising. Removing all outside options renders concessions pointless, and the game merely reduces to the basic tug-of-war. The question thus remains, when outside options are available, when and how will this affect the outcome of competition over reproduction within the group? If the exchange of uncontested shares of reproduction (as proposed by Johnstone 2000 and Reeve & Shen 2006) is unfeasible or unstable, how should we expect competitors to respond to the threat of the group breaking up?
Bargaining theory and the outside option principle A possible solution to the problem of synthesis can be gleaned from work in economic bargaining theory, which is concerned with problems that
11
12
R. A. Johnstone, M. A. Cant are structurally very similar to those addressed by models of skew (Nash 1950, Osborne & Rubinstein 1990, Muthoo 1999). In economics, a “bargaining situation” is one in which two or more players have a common interest to cooperate, but disagree over how the profits of cooperation should be distributed. An obvious example is trade, although formally similar problems arise in a range of other contexts, from salary negotiations and litigation to military disputes between nation states (Fearon 1995, Muthoo 1999, 2000, Powell 2002). There is a large literature on bargaining theory in economics which we will not attempt to review here; readers seeking an accessible introduction are directed to Muthoo (1999, 2000). Rather, we need only refer to a simple principle of bargaining theory, the “outside option principle” (Binmore 1985, Sutton 1986), to help solve the problem of how to synthesize models of reproductive skew. The outside option principle suggests that a focal individual’s outside option will influence the resolution of within-group conflict only if this option yields a greater payoff to the focal individual than the payoff it expects to receive through negotiation (Nash 1953). In these circumstances the individual can use a credible threat to exercise its outside option to obtain a more favorable resolution, one that is sufficient to render the threat incredible. The principle chimes with our everyday intuition about bargaining situations. Imagine, for example, a bartender earning $10 per hour who demands a raise from his employer on the grounds that the bar across the street is offering an hourly rate of $9. Clearly, a rational employer will be left unmoved by this threat of departure. Only if the other bar is offering more than $10 per hour will the employer need to take the threat seriously, and decide whether it is worth raising her own offer to retain the worker. Threats to take action that both parties know would be self-defeating are not credible, and so should exert no influence on the resolution of conflict (Nash 1953, Binmore 1985, McNamara & Houston 2002). The outside option principle suggests that the models of Johnstone (2000) and Reeve & Shen (2006) encounter problems in part because they assume that players always concede shares of reproduction to one another, even under circumstances in which the threat of eviction and the threat of departure prove incredible. This conclusion was also reached independently by P. Buston and A. Zink, in an unpublished manuscript sent to us while this chapter was being written. In the next section, we apply the outside option principle to resolve these problems, and determine when and how the possibilities of departure and of eviction influence behavior in a tug-of-war over reproduction. The key difference between our approach and that of previous synthetic models is that, rather than start with the assumption that an equilibrium
Models of reproductive skew features concession of uncontested shares of reproduction, we first focus on the resolution of the tug-of-war, and then use this to determine whether concessions are necessary. The model We focus on the interaction between two individuals, one dominant and one subordinate, who are related by a coefficient r. Dominance in this context is defined by the ability to evict the subordinate from the local breeding territory, although we also allow for the possibility that the dominant may enjoy an advantage over the subordinate in competing for reproductive opportunities within the local territory. In the extreme, the dominant may enjoy complete control over reproduction. The interaction involves two stages. In the first stage, both players simultaneously choose how much effort to invest in selfish competition over reproductive opportunities within the group. This stage is identical to the “tug-of-war” game of Reeve et al. (1998). The levels of effort will be denoted x for the dominant and y for the subordinate. Following the first stage, both players simultaneously choose whether to remain in association and accept the payoffs from the tug-ofwar, or whether to terminate the interaction. If either player chooses to terminate the interaction, then we assume that the dominant retains control of the breeding territory and thus receives a direct fitness payoff of 1 (baseline productivity for a lone breeder), while the subordinate receives a direct fitness payoff of d (1), reflecting the cost or risk associated with locating alternative breeding opportunities. The “outside options” available to the players are thus departure for the subordinate, and eviction for the dominant (though when d ¼ 1 both players are indifferent as to who leaves and who stays). Evaluation of outside options
The direct fitness payoffs to the two players from the tug-of-war, if both remain in association, depend upon the effort levels of both, and are given by x x þ by by Ws ðx; yÞ ¼ kð1 x yÞ x þ by
Wd ðx; yÞ ¼ kð1 x yÞ
ð1:10Þ
In the second stage of the game, therefore, the subordinate stands to gain by departing if and only if ðd Ws ðx; yÞÞ þ rð1 Wd ðx; yÞÞ > 0
ð1:11Þ
13
14
R. A. Johnstone, M. A. Cant while the dominant stands to gain by evicting the subordinate if and only if ð1 Wd ðx; yÞÞ þ r ðd Ws ðx; yÞÞ > 0
ð1:12Þ
If neither condition is met, then both players do best to remain in association. Under these circumstances, neither the threat of departure nor the threat of eviction is credible. Stable levels of competitive effort
How do the threats of departure and eviction in the second step of the game affect the players’ choices of competitive effort during the first step? The (unique) stable pair of effort levels x* and y* in the basic tug-of-war game (derived by Reeve et al. 1998) were given above. If the resulting direct fitness payoffs to both players are great enough that neither condition (1.11) nor condition (1.12) is satisfied for effort levels x* and y*, then the stable outcome of the game is unaffected by the possibility of departure or of eviction, because neither threat is credible given the outcome of the tug-of-war. Only if one or both of the players stands to gain by departing or evicting the other will the availability of outside options influence the resolution of the conflict. The “concession” zone
If ð1 Wd ðx ; y ÞÞ þ rðd Ws ðx ; y ÞÞ < 0 < ðd Ws ðx ; y ÞÞ þ rð1 Wd ðx ; y ÞÞ
ð1:13Þ
then the dominant stands to gain from the association, while the subordinate does best (given effort levels x* and y* ) to depart. Under these circumstances, the dominant must “ease off ” in the tug-of-war if it is to retain the subordinate in the group. This potentially leads to what we shall call a “concession” equilibrium, at which the stable effort levels of the two players, denoted xc and yc, satisfy @Ws ðx; yÞ @Wd ðx; yÞ þr ¼ 0 for x ¼ xc ; y ¼ yc @y @y
ð1:14Þ
ðd Ws ðxc ; yc ÞÞ þ r ð1 Wd ðxc ; yc ÞÞ ¼ 0
ð1:15Þ
and
Equation 1.14 implies that the subordinate does not stand to gain from a change in effort level, because this would entail a net decrease in the inclusive fitness payoff from the tug-of-war. Equation 1.15 implies that the dominant
Models of reproductive skew does not stand to gain from an increase in its effort, because this would trigger departure by the subordinate, which obtains an inclusive fitness from the tugof-war that is just sufficient to make departure unprofitable. Such an outcome will, however, only prove stable provided that ð1 Wd ðxc ; yc ÞÞ þ rðd Ws ðxc ; yc ÞÞ < 0
ð1:16Þ
If this condition is not met, then the dominant will not be selected to reduce its competitive effort in the tug-of-war sufficiently to retain the subordinate, since it would do better simply to evict its competitor. The “restraint” zone
If ð1 Wd ðx ; y ÞÞ þ r ðd Ws ðx ; y ÞÞ > 0 > ðd Ws ðx ; y ÞÞ þ rð1 Wd ðx ; y ÞÞ
ð1:17Þ
then the subordinate stands to gain from the association, while the dominant does best (given effort levels x* and y* ) to evict its competitor. Under these circumstances, the subordinate must “ease off ” in the tug-of-war if it is to be allowed to remain in the group. This potentially leads to what we shall call a “restraint” equilibrium, at which the stable effort levels of the two players, denoted xr and yr, satisfy @Wd ðx; yÞ @Ws ðx; yÞ þr ¼ 0 for x ¼ xr ; y ¼ yr @x @x
ð1:18Þ
ð1 Wd ðxr ; yr ÞÞ þ r ðd Ws ðxr ; yr ÞÞ ¼ 0
ð1:19Þ
and
Equation 1.18 implies that the dominant does not stand to gain from a change in effort level, because this would entail a net decrease in the inclusive fitness payoff from the tug-of-war. Equation 1.19 implies that the subordinate does not stand to gain from an increase in its effort, because this would trigger eviction by the dominant, which obtains an inclusive fitness from the tug-of-war that is just sufficient to make eviction unprofitable. Such an outcome will, however, only prove stable provided that ðd Ws ðxr ; yr ÞÞ þ rð1 Wd ðxr ; yr ÞÞ < 0
ð1:20Þ
If this condition is not met, then the subordinate will not be selected to reduce its competitive effort in the tug-of-war sufficiently to be tolerated by the dominant, since it would do better simply to leave.
15
16
R. A. Johnstone, M. A. Cant Group breakup
If the solution of the original tug-of-war model is unstable because one or other of the players stands to gain by exercising their outside option, but at the same time neither a concession equilibrium nor a restraint equilibrium is feasible, then we expect that the association will break up. An equilibrium at which x < x* and y < y*, i.e. at which both players simultaneously “ease off ” in the tug-of-war in order to maintain group stability, is not possible. If both players stand to gain from the outcome of the basic tug-of-war, then neither the threat of departure nor the threat of eviction is credible, so that neither player needs to adjust its effort level in response. Conversely, if neither player stands to gain from the outcome of the basic tug-of-war, then neither has any incentive to “ease off ” in order to maintain an association that is unprofitable to both. Results
We have derived analytical solutions for the boundaries of the regions of parameter space in which one obtains the different outcomes described above. Since these expressions are in some cases complex, we will not give them here; instead, Figures 1.1 and 1.2 show graphically these regions of parameter space, for the case of unrelated (r ¼ 0) and related (r ¼ 0.5) competitors, respectively. The general qualitative pattern is simple: groups are more likely to break up when productivity (k) is low, the opportunity for independent breeding (d) is great, and competitors are less closely related. Breakup is also more likely for an intermediate level of asymmetry in competitive ability between dominant and subordinate. There is a substantial region in which both the threat of departure and the threat of eviction prove incredible, so that the solution of the basic tug-of-war game is stable and unaffected by outside options. This outcome is most likely when group productivity is high, there is little opportunity for independent breeding, and competitors are more closely related. Once again, it is also more likely for an intermediate level of asymmetry in competitive ability between dominant and subordinate. When the subordinate is much weaker than the dominant, and particularly when there are substantial opportunities for independent breeding and the competitors are unrelated, the threat of subordinate departure becomes credible. Under these circumstances we obtain a “concession” equilibrium at which the dominant “eases off ” in competition to retain the subordinate in the group. Conversely, when the subordinate is not too much weaker than the dominant, and again particularly when there are substantial opportunities for independent breeding and the competitors are unrelated, the subordinate may have to “ease off ” in competition for its presence to be tolerated.
Models of reproductive skew d = 0.8
d = 0.6
4
4 No threats Dom concession
No threats
Sub restraint
3.5 Group productivity (k)
Group productivity (k)
3.5
3
Group breaks up
2.5
2
1.5
0.6 0.8 0.4 Subordinate strength (b)
2
Group breaks up
0.2
0.6 0.8 0.4 Subordinate strength (b)
No threats
3.5
3 Sub restraint 2.5
2
Group breaks up
1
d = 0.2
4
Group productivity (k)
Group productivity (k)
2.5
1
d = 0.4
4
1.5
Sub restraint
1.5
0.2
3.5
Dom concession 3
No threats
3
2.5
Sub restraint
2
1.5 Group breaks up
0.2
0.6 0.8 0.4 Subordinate strength (b)
1
0.2
0.6 0.8 0.4 Subordinate strength (b)
1
Figure 1.1 The graphs show, for unrelated players (r ¼ 0), the regions of parameter space in which: (i) the solution of the basic tug-of-war game proves stable because both the threat of departure and the threat of eviction are incredible (labeled “no threats”); (ii) the dominant must reduce its competitive effort to prevent departure of the subordinate (labeled “Dom concession”); (iii) the subordinate must reduce its competitive effort to prevent eviction by the dominant (labeled “Sub restraint”); (iv) the association proves unstable. In each graph, the relative competitive ability of the subordinate (b) increases from left to right along the horizontal axis, and group productivity (k) increases from bottom to top along the vertical axis. Different graphs give results for different levels of opportunity for independent breeding (d): high opportunity in the top-left graph, and low opportunity in the bottom-right graph.
Implications for skew
What is the significance of these different solution regions? It matters in which region a population or species falls, because depending on the nature of the outcome, patterns of skew are expected to be very different. When
17
R. A. Johnstone, M. A. Cant d = 0.8
4
3.5 Group productivity (k)
Group productivity (k)
No threats 3
2.5 Dom concession 2 Group breaks up
1.5
0.2
0.4 0.6 0.8 Subordinate strength (b)
No threats
3
2.5
2
Group breaks up
0.2
1
Sub restraint
Dom concession
1.5
d = 0.4
4
0.6 0.8 0.4 Subordinate strength (b)
1
d = 0.2
4
3.5 Group productivity (k)
3.5
3
d = 0.6
4
3.5
Group productivity (k)
18
No threats
2.5
2
3 No threats 2.5
2
Sub restraint Sub restraint
1.5
1.5 Group breaks up
Group breaks up 0.2
0.6 0.8 0.4 Subordinate strength (b)
1
0.2
0.6 0.8 0.4 Subordinate strength (b)
1
Figure 1.2 Regions of parameter space as in Figure 1.1, for related players (r ¼ 0.5).
neither the threat of departure nor the threat of eviction is credible, outside options, as we have seen, do not come into play. Under these circumstances, neither ecological constraints nor the opportunity for independent breeding nor group productivity exert any influence whatsoever on the level of skew within a group. As mentioned previously, relatedness also has little influence on stable skew in the tug-of-war game, so that in this region the relative competitive ability of the subordinate will be the only significant influence on skew (with stronger subordinates gaining a greater share of reproduction). Outside options typically come into play, as described above, when the subordinate is either very weak or very strong, group productivity is low, relatedness is low, and there are opportunities for independent breeding. In
Models of reproductive skew these “concession” and “restraint” regions, group productivity, relatedness, and the opportunity for independent breeding will influence the level of skew in a group, just as in simple transactional models of skew. In the concession region, where the dominant is forced to “ease off ” in competition in order to retain the subordinate, the latter’s share of reproduction will increase with the opportunity for independent breeding and decrease with group productivity and relatedness. By contrast, in the restraint region, where it is the threat of eviction that comes into play, the subordinate’s share of reproduction will decrease with the opportunity for independent breeding and increase with group productivity. Lastly, across both concession and restraint regions, the subordinate’s share increases with its competitive ability (just as in the basic tug-of-war).
Discussion Taking account of the “outside option principle” leads to a synthetic model of skew that does not suffer from the stability problems of previous analyses. The threat of departure and the threat of eviction come into play only when they are credible; moreover, they lead to players “easing off ” in competition, rather than exchanging uncontested shares of reproduction. The question of how individuals could possibly “agree” to refrain from competing over a proportion of the available reproductive opportunities does not, therefore, arise. The results suggest a rather different view of reproductive skew from previous attempts at synthesis. While some have argued for the universal scope and power of transactional (particularly concession) models, and their potential for unifying the study of social evolution (e.g. Reeve & Keller 2001, Buston et al. 2007), the present analysis suggests that they are relevant only under certain restricted circumstances. Where group members are closely related and/or the productivity benefits of association are great, the threat of departure or of eviction is likely to prove incredible, because the direct and indirect benefits of group membership are too large to forgo. Under these circumstances, the whole basis of transactional skew models is eliminated. No concessions need be offered, because there is no opportunity for either player to gain by dissolving the association. Consequently, manipulating outside options is not expected to influence skew (see, e.g., Langer et al. 2004, Heg et al. 2006). Rather strikingly, these circumstances of high relatedness and high group productivity are precisely those in which selection is most likely to favor association in the first place. Consequently, the most obviously beneficial interactions are those least likely to be amenable to analysis in terms of skew theory.
19
20
R. A. Johnstone, M. A. Cant Should we then abandon skew models completely? We would argue not. The burgeoning literature on biological markets and partner choice (Noe¨ et al. 1991, Noe¨ & Hammerstein 1994, Bshary & Noe¨ 2003, Sachs et al. 2004, Foster & Wenseleers 2006) attests to a growing awareness among biologists that outside options are often important. Although most of the classical applications of game theory in biology focus on “forced play” among a small number of players, there are many situations in which one individual is not forced to interact with another, but can readily switch partners or perhaps forgo any interaction at all. Under these circumstances, it is not surprising that the threat of terminating an interaction and losing a partner should exert a significant influence on behavior within the association. Such “sanctions” have become the focus of much discussion in the study of mutualism and cooperation (see, e.g., Herre et al. 1999, Ferrie`re et al. 2002, Johnstone & Bshary 2002, West et al. 2002, Bshary & Noe¨ 2003, Kiers et al. 2003, Sachs et al. 2004, Foster & Wenseleers 2006), and an elegant experimental demonstration is provided by studies of interaction between cleaner-fish and clients, in which the threat of departure by the client induces cleaners to refrain from biting live tissue (Bshary & Grutter 2002, 2005). Although this example may seem rather remote from reproductive skew, it illustrates the same principle that lies at the heart of skew theory – that individuals’ prospects outside a given association can influence the resolution of conflicts inside it. The key feature of biological “markets” that renders the threat of departure (or rejection) credible is the ready availability of alternative partners. We suggest, accordingly, that transactional skew theory is likely to prove most relevant not to highly profitable associations involving small numbers of closely related individuals, who may have little opportunity to join a similar family group elsewhere, but to looser and less cooperative associations between more distantly related individuals. Individuals in such cases may have opportunities to move between alternative, unrelated groups, so that the threat of departure may become credible, just as the threat of eviction may when group members can be easily replaced or contribute little to productivity in any case. Indeed, it is surprising that, apart from some work by Reeve (1998), there has so far been little attempt to incorporate market effects explicitly into skew theory. The response of other group members to the prospect of departure or an attempt at eviction need not, as our model shows, involve the “exchange” of uncontested shares of reproduction, with all of the questions this raises about how animals might achieve such a feat, and what prevents cheating. Individuals may simply reduce their competitive efforts in response to the risk of triggering the breakup of a group. To detect such influences is likely to be
Models of reproductive skew difficult, because an effective threat is precisely one that elicits responses that make it unprofitable to carry out. Consequently, social behavior might potentially be influenced by many “invisible” threats that remain hidden until the “rules” they enforce are broken. However, threats of this kind can be exposed by experimentally staging such violations. Wong et al. (2007), for instance, have shown that the typical size hierarchy seen in groups of the coral-dwelling goby Paragobidon xanthosomus is maintained by the latent threat of eviction: when dominant fish were experimentally paired with subordinates larger than observed under natural circumstances, eviction was the result (while dominants tolerated individuals who were smaller than themselves). Conclusion To sum up, we suggest that outside options can influence the resolution of reproductive conflict within a group, but that they will do so only when the threat of departure or of eviction is credible. Consequently, tests which examine the relationship between skew and factors such as group productivity or ecological constraints are expected to yield variable results. The essential question is whether or not any of the members of a group stand to gain from its dissociation. The answer will determine whether or not outside options come into play.
References Binmore, K. (1985). Bargaining and coalitions. In A. Roth, ed., Game Theoretic Models of Bargaining. Cambridge: Cambridge University Press, pp. 269–302. Bshary, R. & Grutter, A. S. (2002). Experimental evidence that partner choice is the driving force in the payoff distribution among cooperators or mutualists: the cleaner fish case. Ecology Letters, 5, 130–136. Bshary, R. & Grutter, A. S. (2005). Punishment and partner choice cause cooperation in a cleaning mutualism. Biology Letters, 1, 396–399. Bshary, R. & Noe¨, R. (2003). Biological markets: the ubiquitous influence of partner choice on the dynamics of cleaner fish–client reef fish interactions. In P. Hammerstein, ed., Genetic and Cultural Evolution of Cooperation. Cambridge, MA: MIT Press, pp. 167–184. Buston, P. M., Reeve, H. K., Cant, M. A., Vehrencamp, S. L., & Emlen, S. T. (2007). Reproductive skew and the evolution of group dissolution tactics: a synthesis of concession and restraint models. Animal Behaviour, 74, 1643–1654. Cant, M. A. (2000). Social control of reproduction in banded mongooses. Animal Behaviour, 59, 147–158. Cant, M. A. & Johnstone, R. A. (1999). Costly young and the partitioning of reproduction in animal societies. Behavioral Ecology, 10, 178–184.
21
22
R. A. Johnstone, M. A. Cant Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology Evolution, 7, 288–292. Fearon, J. D. (1995). Rationalist explanations for war. International Organization, 49, 379–414. Ferrie`re, R., Bronstein, J. L., Rinaldi, S., Law, R., & Gauduchon, M. (2002). Cheating and the evolutionary stability of mutualisms. Proceedings of the Royal Society of London B, 269, 773–780. Foster, K. R. & Wenseleers, T. (2006). A general model for the evolution of mutualisms. Journal of Evolutionary Biology, 19, 1283–1293. Haydock, J. & Koenig, W. D. (2002). Reproductive skew in the polygynandrous acorn woodpecker. Proceedings of the National Academy of Sciences of the USA, 99, 7178–7183. Heg, D., Bergmuller, R., Bonfils, D., et al. (2006). Cichlids do not adjust reproductive skew to the availability of independent breeding options. Behavioral Ecology, 17, 419–429. Herre, E. A., Knowlton, N., Mueller, U. G., & Rehner, S. A. (1999). The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends in Ecology and Evolution, 14, 49–53. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Bshary, R. (2002). From parasitism to mutualism: partner control in asymmetric interactions. Ecology Letters, 5, 634–639. Johnstone, R. A. & Cant, M. A. (1999). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 12, 99–103. Kiers, E. T., Rousseau, R. A., West, S. A., & Denison, R. F. (2003). Host sanctions and the legume–rhizobium mutualism. Nature 425, 78–81. Kokko, H. & Lindstro¨m, J. (1997). Measuring mating skew. American Naturalist, 149, 794–799. Langer, P., Hogendoorn, K., & Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature 428, 844–847. McNamara, J. M. & Houston, A. I. (2002). Credible threats and promises. Philosophical Transactions of the Royal Society of London B, 357, 1607–1616. Muthoo, A. (1999). Bargaining Theory with Applications. Cambridge: Cambridge University Press. Muthoo, A. (2000). A non-technical introduction to bargaining theory. World Economics, 1, 145–166. Nash, J. (1950). The bargaining problem. Econometrica, 18, 155–162. Nash, J. (1953). Two-person cooperative games. Econometrica, 21, 128–140. Noe¨, R. & Hammerstein, P. (1994). Biological markets: supply and demand determine the effect of partner choice in cooperation, mutualism and mating. Behavioral Ecology and Sociobiology, 35, 1–11.
Models of reproductive skew Noe¨, R., van Schaik, C. P., & van Hooff, J. A. R. A. M. (1991). The market effect: an explanation for payoff asymmetries among collaborating animals. Ethology, 87, 97–118. Nonacs, P. (2007). Tug-of-war has no borders: it is the missing model in reproductive skew theory. Evolution, 61, 1244–1250. Osborne, M. & Rubinstein, A. (1990). Bargaining and Markets. San Diego, CA: Academic Press. Powell, R. (2002). Bargaining theory and international conflict. Annual Review of Political Science, 5, 1–30. Reeve, H. K. (1998). Game theory, reproductive skew and nepotism. In L. A. Dugatkin & H. K. Reeve, eds., Game Theory and Animal Behaviour. Oxford: Oxford University Press, pp. 118–145. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflict in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Sachs, J. L., Mueller, U. G., Wilcox, T. P., & Bull, J. J. (2004). The evolution of cooperations. Quarterly Review of Biology, 79, 135–160. Sutton, J. (1986). Non-cooperative bargaining theory: an introduction. Review of Economic Studies, 53, 709–724. Tsuji, K. & Kasuya, E. (2001). What do the indices of reproductive skew measure? American Naturalist, 158, 155–165. Vehrencamp, S. L. (1979). The roles of individual, kin and group selection in the evolution of sociality. In P. Marler & J. G.Vandenbergh, eds., Handbook of Behavioural Neurobiology 3. Social Behaviour and Communication. New York, NY: Plenum Press, pp. 351–394. Vehrencamp, S. L. (1983). Optimal degree of skew in reproductive societies. American Zoologist, 23, 327–335. West, S. A., Kiers, E. T., Pen, I., & Denison, R. F. (2002). Sanctions and mutualism stability: when should less beneficial mutualists be tolerated? Journal of Evolutionary Biology, 15, 830–837. Wong, M. Y. L., Buston, P. M., Munday, P. L., & Jones, G. P. (2007). The threat of punishment enforces peaceful cooperation and stable queues in a coral-reef fish. Proceedings of the Royal Society of London B, 274, 1093–1099.
23
2
Reproductive conflict and the evolution of menopause m i c h a e l a. c a n t , r u f u s a. j o h n s t o n e , and andrew f . r usse ll
Summary Human females (Homo sapiens) exhibit a dramatic form of reproductive skew in which half the age classes of adults contain only nonbreeders. Among other mammals, only pilot (Globicephala spp.) and killer whales (Orcinus orca) exhibit a similar pattern. The “grandmother” hypothesis suggests that selection can favor post-reproductive survival because older females help their offspring to reproduce. But the indirect fitness gains of helping appear insufficient to outweigh the potential benefits of continued direct reproduction, so this hypothesis cannot explain why women cease reproducing in the first place. Here we present some background on menopause and describe new research which helps to understand both the strange taxonomic distribution of menopause and the timing of reproductive cessation in humans. Specifically, recent models have explored the potential reproductive conflicts that may have arisen in ancestral human families, and the influence of demography on the resolution of such conflicts. These studies suggest that an integrated model which takes into account the potential costs of reproductive competition, as well as the benefits of helping, offers a fuller understanding of the evolution of menopause. Tabar ne maiet hate kana jane bakariyon, lardiyon jyoon! (How unbecoming of parents to procreate alongside their children like goats and sheep!) Saying of the Mogra, Rajasthan, India (quoted in Patel 1994) Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
24
Reproductive conflict and the evolution of menopause Reproductive skew in human societies Human societies are characterized by a dramatic and puzzling pattern of reproductive skew. In populations exposed to natural schedules of mortality and fertility (i.e. without access to modern medicine and technology), almost half the age classes of adult human females contain only nonbreeders (Figure 2.1). The mean ages at which women give birth to their last child in natural-fertility populations cluster around 39 years (Wood 1994, p. 442), but even in hunter–gatherer societies that lack modern medicine women who reach this age can expect to live well into their sixties (Pennington 2001, Blurton Jones et al. 2002). The restriction of reproduction to certain age classes is not in itself unusual for a cooperative vertebrate, but in other species it is almost always older females who breed and younger females who do not (Emlen 1991, 1995). The reverse pattern exhibited by humans is extremely rare – among vertebrates only killer whales and pilot whales are reported to exhibit a similar reversal of breeding roles with respect to age class (Marsh & Kasuya 1986, Olesiuk et al. 1990, Whitehead & Mann 2000). Early reproductive cessation represents an evolutionary puzzle because standard life-history theory suggests that there should be no selection for somatic
Figure 2.1 Survival and fecundity in a natural-fertility human population. Data from a Taiwanese population in 1906 (redrawn from Hamilton 1966).
25
26
M. A. Cant et al. maintenance after the end of reproduction. Why then do women cease reproducing so long before they die? Despite almost 50 years of research on the evolution of menopause, this important question remains open. Current models invoke the kin-selected benefits of helping as a grandmother to explain post-reproductive survival in women, but, as we describe below, quantitative analyses suggest that these models cannot explain why women stop breeding at the time they do. Part of the problem is that current models focus solely on the direct fitness consequences of reproduction, and compare this with the indirect fitness consequences of helping. Helping is assumed to affect the fitness of other group members, while breeding is not. This approach is one-sided because it ignores the potential impact of reproduction on the fitness of other group members. Where there are limited resources within a group for reproduction, the decision to reproduce will depend on whether other females in the group will also reproduce, how many young they will produce, and how one’s own young will fare in competition if they do. Reproduction in a social context, therefore, is a game-theoretic rather than an optimization problem. Reproductive skew theory was developed to study exactly this type of problem, and so is an apt framework within which to study the evolution of patterns of reproduction in humans. In this chapter we take a fresh look at the puzzle of menopause by examining the potential reproductive conflicts in ancestral human societies, and the way in which these conflicts are likely to be resolved. We first review the main adaptive explanations for menopause, the “mother” and “grandmother” hypotheses, and highlight the empirical and theoretical difficulties that these hypotheses have encountered. We then describe results from our own recent research, which focuses on the impact of demography on kin selection across individual lifespans, and how this will affect the resolution of conflicts over reproduction within human social groups. Our aim is to show that these new models offer a fuller explanation for the pattern and timing of reproductive cessation in humans, and help to explain why, of all long-lived, social mammals, it is specifically among the lineages of great apes and toothed whales that menopause has evolved.
How old is menopause? A possible non-adaptive explanation for menopause is that it is a simple artifact of the reduction in mortality that followed agriculture and improved sanitation (reviewed by Peccei, 2001a, 2001b). The idea is that the reproductive lifespan of women reflects the expected female lifespan prior to these technological developments, but there has been insufficient time for
Reproductive conflict and the evolution of menopause selection to extend the reproductive period to match the newly elongated lifespan. If this hypothesis were correct, menopause should be absent in hunter–gatherer populations without agriculture or modern medicine because women would rarely survive beyond the age of 50. On the contrary, in the three best-studied hunter–gatherers (the !Kung of the Kalahari, Ache of Paraguay, and Hadza of Tanzania) a large fraction of women survive to postmenopausal age. For example, 64% of non-nomadic !Kung, 46% of Hadza, and 42% of Ache women live until age 50 or more (Pennington 2001). Moreover, women who survive to 45 can expect an average of 20 or more years of life thereafter (Pennington 2001, Blurton Jones et al. 2002). A pattern of menopause coupled with prolonged post-reproductive life can be inferred from ancient texts: the Bible (Psalms 90: 10) refers to an expected lifespan of 70 (the familiar “threescore years and ten”), rising to 80 years “by reason of strength”, while Aristotle (c. 360 BC) and Pliny (c. AD 77) cite an age at menopause of around 50 years (Amundsen & Diers 1970). Finally, a recent analysis of fossil molar wear (Caspari & Lee 2004) suggests that the fraction of humans surviving to become grandparents increased five-fold between the Middle and Upper Paleolithic (c. 300 000–10 000 years ago), i.e. before the emergence of agriculture. The evidence suggests, therefore, that menopause has been a feature of the life history of human females for at least the last 10 000 or 20 000 years, and possibly much longer.
Reproductive senescence Menopause is best viewed as the endpoint of unusually rapid senescence of the reproductive system relative to somatic systems. Senescence, the general decline in efficiency of bodily functions with age, is an inescapable property of both somatic and reproductive systems in iteroparous organisms. This is because random mortality ensures that older individuals always make up a smaller fraction of the breeding population than younger individuals, so genes which have negative effects on reproduction or survival early in life are more strongly opposed by selection than are genes with negative effects later in life (Medawar 1952, Williams 1957, Hamilton 1966). In theory, senescent decline should strike all body functions at a similar rate, since if one system or organ (for example, the cardiovascular system, or the renal system) declined much more rapidly than others, selection to maintain other systems would weaken, accelerating their rate of decline to match that of the most rapidly senescent (Williams 1957). Consequently, the capacity for reproduction is predicted to decline in tandem with, and at a similar rate to, other somatic systems. In most organisms this expectation is borne out: fertility in old age, like other
27
28
M. A. Cant et al. functions, is much reduced but reproduction is nevertheless still possible (Rose 1991). What in humans (and some whales) could cause the rate of reproductive senescence to become decoupled from the rate of somatic senescence? To answer this question it is not sufficient to invoke the personal costs of breeding at late ages, since these costs are themselves an evolved property of the system. Increases in, say, the rate of birth defects, or stillbirth, that occur after the age of 40 in humans are a result of, rather than a cause of, the rapid senescence of the reproductive system. Menopause marks the end of a process of rapid reproductive senescence that begins a decade earlier, and one must be careful not to invoke the effects of this rapid senescence in order to explain it. Rather, we need to consider the potential benefits of early reproductive cessation in an ancestral, non-menopausal hominid species in which fertility declined at the same rate as other bodily functions.
Adaptive explanations Two closely related adaptive explanations for the evolution of menopause are known as the “mother” and “grandmother” hypotheses. These hypotheses differ primarily in whether menopause is assumed to boost the survival or the fertility of existing young. Williams (1957) suggested that early reproductive cessation is a consequence of the long period of offspring dependency in humans. According to this hypothesis, older women may at some point gain from ceasing reproduction to invest in raising their existing children to adulthood, rather than engaging in increasingly risky breeding attempts which could leave their dependent offspring motherless (Peccei 2001a, 2001b). By contrast, other authors have emphasized the benefits of menopause for enhancing the fertility of a woman’s existing children (Hamilton 1966, Alexander 1974, Hawkes et al. 1998). An older woman whose fertility is declining due to senescence may at some point do best to switch resources from her own breeding attempts to helping to rear grandchildren. In both cases, beyond the age at which helping is more profitable than breeding there is no selection to maintain fertility, and the rate of senescence of the reproductive system is expected to increase sharply. The result is a decoupling of the rates of somatic and reproductive senescence, and a pattern of reproductive cessation long before death. Empirical data offer little support for the mother hypothesis. The chance of dying in childbirth must be very large for females to prefer reproductive cessation over continued reproduction. We can illustrate this with a simple numerical example. Consider a female at age 40 with four dependent offspring, faced with a decision of whether to cease reproduction or produce one
Reproductive conflict and the evolution of menopause more child. Let d be the chance that she dies in childbirth, which we will assume leads to the certain death of all her existing young. In this example, reproductive cessation will be favored over continued breeding if 4 > 5(1 d), or d > 20%. In reality, the chance of dying in childbirth is minuscule even among hunter–gatherers (e.g. around 1/150 in the Ache: Hill & Hurtado 1996). Data from the United States in the late nineteenth and early twentieth centuries suggest that even for mothers who gave birth at age 45–50, the risk of dying in childbirth was around 5% (Loudon 1993). Quantitative analyses using data from natural-fertility populations have concluded that the risk of death in childbirth is far too small to account for the evolution of menopause (Rogers 1993, Hill & Hurtado 1996). It is tempting to invoke other factors which might devalue late-life reproduction, such as elevated rates of fetal wastage and birth defects, but again it is important to remember that these factors are themselves the outcome of selection for rapid reproductive senescence. Investigations of the grandmother hypothesis are more numerous (see Voland et al. 2005). Data from modern hunter–gatherers and historical populations provide evidence that grandmothers can indeed boost the reproductive success of their children. Significant positive effects of grandmothers on grandchild survival have been reported in six out of seven studies of grandmothering in pre-medical societies (Lahdenpera¨ et al. 2004a, 2004b, Mace & Sear 2005). However, while these studies demonstrate a potential kin-selected benefit of grandmotherhood, quantitative analyses suggest that these benefits are not sufficiently large to explain the evolutionary maintenance of menopause at the age at which it occurs (Rogers 1993, Hill & Hurtado 1996). Hill & Hurtado (1991, 1996), for example, use data from one well-studied hunter– gatherer society (the Ache of Paraguay) to calculate the inclusive fitness payoffs of grandmothering versus continued reproduction for older women. To estimate the latter payoff they assume that in the absence of menopause fertility would decline at the same rate as somatic function. Their calculations suggest that menopause cannot be favored by kin selection in this population because there are few close kin alive for an older woman to help, and because her help has too little impact on the survival or reproduction of these kin. Rogers (1993) uses a different model and dataset (the Taiwan 1906 data used in Hamilton’s [1966] classic paper on senescence) but shows a similar result, namely that the impact of an older woman’s help on her close kin must be more substantial than has been so far documented for menopause to be favored by kin selection. Recently, however, Shanley & Kirkwood (2001) presented a model which they claim can account for the evolution of menopause around the age of 50 or 55, again using the Taiwan 1906 dataset. Unfortunately, their model uses the intrinsic rate of increase of the population as the measure of fitness to be
29
30
M. A. Cant et al. maximized by natural selection. This means that offspring and grandoffspring are counted as equally “valuable” to an older female faced with the decision of whether to continue reproduction versus help as a grandmother. As Hamilton (1966) demonstrated, selection for helping versus breeding will depend not only on the number of young produced as a result of helping or breeding, but on the relatedness to these offspring. A mother’s relatedness to her own offspring is twice that to her grandoffspring, so the fitness payoff of helping to raise an extra grandchild is, other things being equal, half that of producing another child herself. Shanley & Kirkwood’s (2001) model does not take into account relatedness, and so overestimates the fitness benefits of grandmothering by a factor of around two compared to the analyses of Hill & Hurtado (1996) and Rogers (1993). This may explain why Shanley & Kirkwood’s analysis can account for the evolution of menopause at around 50 whereas the other analyses cannot.
Physiological constraints and phylogenetic inertia While women can clearly gain fitness benefits by grandmothering, these fitness benefits are insufficient to account for the timing of reproductive cessation in human women or the evolution of menopause. The problems raised by these quantitative studies can be circumvented, however, if we assume that timing of menopause is a phylogenetic artifact or reflects some form of physiological constraint. Hawkes and co-workers (Hawkes et al. 1998, Hawkes 2003) note that the endpoint of reproductive senescence in human females occurs at the same age as in chimpanzees (Pan troglodytes), i.e. in the fifth decade of life. Consequently, they argue, it is the extended post-reproductive life of human females, not the timing of menopause, that is the derived trait to be explained (Hawkes et al. 1998). Grandmothering effects are invoked to explain the extension of the female lifespan long past the end of the phylogenetically conserved age at reproductive cessation (why males have similarly extended lifespans is not explained by this hypothesis, but see Marlowe [2000] for one perspective). This argument is unsatisfactory on its own because it assumes that stasis in the face of evolutionary change requires no special explanation. The conservation of patterns of reproductive senescence in the human lineage, despite lengthening lifespan, implies: (1) some physical or physiological constraint prohibiting evolutionary change; (2) an absence of genetic variation upon which selection can act; or (3) some form of stabilizing selection. Comparative evidence lends no support to the first possibility, i.e. that the reproductive lifespan of human females cannot be extended much past the age of 50 due to physiological constraints. Other long-lived mammals continue to breed until
Reproductive conflict and the evolution of menopause the end of life: African elephants (Loxodonta africana) reproduce in their sixties (Moss 2001) and blue whales (Balaenoptera musculus) into their nineties (Mizroch 1981). Across species, oocyte stocks are evolutionarily labile and are adjusted to lifespan and body weight (Gosden & Telfer 1987). In addition, the initial oocyte stock and rate of follicular attrition in human females is commensurate with a longer reproductive lifespan, but at around the age of 40 there is a marked increase in the follicular hazard rate so that by age 50 follicle stocks have dropped below a minimum required to sustain menstrual activity (Faddy et al. 1992, Faddy & Gosden 1996, Figure 2.2). By contrast, in laboratory rodents and rhesus macaques (Macaca mulatta) (the only other species for which similar data are available), there is no indication that the rate of follicular attrition increases later in life (Jones & Krohn 1961, Nichols et al. 2005).
Figure 2.2 The bi-phasic model of declining follicle numbers in pairs of human ovaries from neonatal age to 51 years old (modified from Faddy et al. 1992). Data are from four different autopsy studies. Note the logarithmic scale on the y-axis. A bi-exponential regression model offers a significantly better fit to the data than a single exponential regression. Follicle numbers decline at a constant exponential rate from birth until reaching a critical figure of around 25 000 at age 37.5 years, after which the exponential rate parameter increases in magnitude. Menopause occurs on average when a threshold of around 1000 follicles remain (Faddy et al. 1992).
31
32
M. A. Cant et al. Turning to the second possibility, age at menopause varies widely among individuals (with ages between 40 and 59 considered normal in both modern and natural-fertility populations), and estimates of the heritability of age at menopause range from 40% to 63% (Snieder et al. 1998, Peccei 1999). There would thus seem to be sufficient genetic variation on which natural selection could act, if prolonged fertility were advantageous. Moreover, recent evidence from a preindustrial Finnish population (Figures 2.3, 2.4) suggests that prolonged reproduction can have a substantial positive influence on a woman’s fitness (Helle et al. 2005), and that age at last reproduction is also highly heritable (Pettay et al. 2005). Taken together, this evidence suggests that any selection to extend the female’s reproductive span has been held in check by some form of opposing selection. Indeed, Hawkes (2003, p. 389) reaches a similar conclusion: Overall then, the available data on ages at last birth and menopause in chimpanzees show age-specific fertility declines in that species not substantially different from our own. Mammalian fertility, however, can extend to much older ages than it does in humans. This evidence is consistent with the argument that ancestral age-specific fertility declines have been maintained in our lineage, perhaps conserved by stabilizing selection.
Figure 2.3 Extended Lummaa family from nineteenth-century Finland showing several generations. Courtesy of Virpi Lummaa.
Reproductive conflict and the evolution of menopause
Figure 2.4 Photo of nineteenth-century nuclear Lummaa family from Finland with mother, father and all their children. Courtesy of Virpi Lummaa.
This stabilizing selection means that mutations for a later age at menopause must have been selected against, despite selection for a longer lifespan. We are thus back to our initial question: what is the nature of this selection? Why has the reproductive lifespan of human females not increased in line with their longevity, as it has in other long-lived mammals?
Reproductive competition: a new perspective on menopause We believe that previous models offer an incomplete account of the evolution of menopause because they focus solely on the kin-selected benefits of parenting and helping, and ignore the potential kin-selected costs of cobreeding. Conflict over reproduction in animal societies is expected wherever communal resources are used for the production of offspring, particularly where helpers increase the success if group members and breeders compete
33
34
M. A. Cant et al. for monopolization of those helpers. If two females produce young in the same group at the same time, each offspring will necessarily receive less food (unless twice as much food, or help, is available). Per capita success of young in a communal brood will therefore decrease with the number of young produced, as assumed in standard clutch-size theory (Lack 1947, Cant 1998, Cant & Johnstone 1999). Reproductive conflict is ubiquitous in other cooperative breeders in which there is more than one potential breeder per social group (Keller & Reeve 1994, Clutton-Brock 1998, Beekman et al. 2003, Ratnieks et al. 2006). Unlike other cooperative breeders, however, the possibility of reproductive competition in ancestral human families has been ignored. Recently, we have argued that the pattern and timing of reproductive cessation in humans is best understood as an adaptation to minimize the degree of reproductive competition between generations (Cant & Johnstone 2008). Certainly one of the consequences of the mean age at reproductive cessation in humans is that it leads to very low reproductive overlap between generations. Many primate species exhibit a post-reproductive lifespan, but there is nevertheless considerable overlap in the period for which females of older and younger generations are reproductively active. Figure 2.5A shows, for those primate species classified as exhibiting a post-reproductive lifespan (Cohen 2004), the relationship between generational overlap (calculated as the proportional overlap between the maximum lifespans of mother and daughter) and reproductive overlap (calculated as the proportional overlap between the mean reproductive spans of mother and daughter). Humans are not unusual in respect of their degree of generational overlap (77%, compared to 71% for chimpanzees, Pan troglodytes, and 73% for gorillas, Gorilla gorilla). However, they show an extraordinarily low degree of maximum reproductive overlap compared to other primates (30%, compared to 50% or more for all other species in the sample), far lower than would be expected on the basis of their generational overlap. On average, females from one generation stop breeding at just the time that the females in the next generation start to reproduce (Cant & Johnstone 2008; Figure 2.5B). If we used the regression line shown in Figure 2.5A to “reverse-engineer” the reproductive lifespan of human females, we would predict a mean age at last birth of 62, followed by menopause around 10 years later – an intriguing match to the predicted age at menopause obtained by extrapolating the constant rate of follicular attrition before the age of 40 (Figure 2.2). If the pattern of rapid reproductive senescence leading to low reproductive overlap in humans is a response to reproductive competition, why do none of the other primate species in Figure 2.5 show similar adaptations to the costs of co-breeding? Reproductive competition is expected where breeding resources are communal or the need for helpers is important and helpers are limiting.
A
1 Mouse lemur
Maximum reproductive overlap
0.9
Hanuman langur Rhesus macaque
0.8
Tamarin
Pig-tailed macaque Barbary macaque Japanese macaque
0.7
Olive baboon
0.6
Orangutan
0.5
Gorilla Chimpanzee
0.4 Human
0.3 0.5
0.6
0.7
0.8
0.9
1
Maximum generational overlap B
Orangutan
Japanese macaque 5
22
23
Chimpanzee 15
39 47
12
38 44
Human 19 38
82
Figure 2.5 Patterns of reproductive overlap in 12 primate species recently classed as exhibiting a post-reproductive lifespan (Cant & Johnstone 2008). Panel A shows maximum reproductive overlap versus maximum generational overlap. Maximum generational overlap is defined as (MLS AFB)/MLS, where AFB is average age at first birth and MLS is maximum recorded lifespan. Maximum reproductive overlap is defined as (MRS AFB)/MRS, where MRS is the maximum reproductive span, calculated as maximum age at last birth (MALB) minus AFB. For four species (chimpanzees, orangutans, Japanese macaques, and humans), published data are sufficiently detailed to calculate average reproductive overlap, defined as (ARS AFB)/ARS, where ARS is the average reproductive span (i.e. mean ALB minus AFB). Panel B summarizes the pattern of overlap for these four species. For each, horizontal bars represent the maximum lifespans of three successive generations, scaled to a standard length and offset in accordance with the value of AFB relative to MLS, with average reproductive spans shaded. The average reproductive overlap values for Japanese macaques, orangutans and chimpanzees were 0.71, 0.52, and 0.39 respectively, compared to an average reproductive overlap for humans of 0.00. For reference sources and values used to plot the figure see Cant & Johnstone (2008).
36
M. A. Cant et al. Compared to most other primates, such competition will be particularly intense in humans because they exhibit a unique degree of food sharing, and are cooperative breeders, in the sense that adults other than parents make important contributions to raising offspring (Emlen 1991, Mace & Sear 2005). In most other primates mothers look after their own young in the early years of life, provisioning them with milk during infancy, and helping them to gain experience in finding their own food – usually fruits and other plant matter. Mothers and their offspring are not reliant on shared food resources obtained by adult helpers, so co-breeders are likely to have little direct impact on each other’s reproductive success. Humans, by contrast, provision their children into adulthood and parents rely on the other group members to gather food. Food acquisition is divided between family members on the basis of sex and age class, and food collected or hunted by different group members is combined into a shared resource (Kaplan & Hill 1985, Gurven et al. 2004, Gurven 2005). Additional offspring within the same family will therefore draw on the same communal resource pool. Unless the presence of an extra female breeder somehow generates twice as much food for the group, co-breeding females will have fewer resources with which to raise their children than a female who is able to monopolize reproduction within the group. Unfortunately, the reproductive separation of generations in humans is so pronounced that it is difficult to obtain direct data on the costs of co-breeding. For example, in the natural-fertility Gambian dataset (for which data were collected between 1950 and 1974), only 5.6% of children (89/1588) had a reproductively active maternal grandmother when they were born, and there were no children born who had a reproductively active paternal grandmother (Sear et al. 2000); similar patterns also occur in pre-modern Western populations (V. Lummaa, personal communication). In many societies the reproductive separation of generations is further reinforced by cultural taboos. Among the Nyakyusa of Tanzania and some West African populations, women are required to stop breeding as soon as their first grandchild is born (Wilson 1957 p. 137, Cavalli-Sforza 1983), and similar proscriptions are found in some Asian societies (Patel 1994 p. 162 [quote at head of this chapter], Islam & Yadava 1997, Skinner, 2004). Though there are almost no data on the costs of co-breeding between generations, we can get some insight into the potential costs of co-breeding within a single family by examining offspring success under polygyny. Women in polygynous marriages (i.e. marriages in which two or more women are married to a single man) are on average less fertile than their monogamous counterparts (Isaac & Feinberg 1982, Pebley & Mbuga 1989). A large dataset from six West African countries suggests that, after controlling for socioeconomic and demographic factors, polygyny is associated with a 50%
Reproductive conflict and the evolution of menopause increase in neonatal mortality, and a doubling of post-neonatal child mortality (Amey 2002). These costs of polygyny to children have been attributed to crowding and disease transmission (Isaac & Feinberg 1982, Roth & Kurup 1988) and to reduced access to resources (Strassmann 1997) – two factors that would also apply to co-breeding between generations. While co-breeding is likely to involve costs, it remains puzzling why older women should choose to cease reproduction in the face of competition from younger women. In other animals, when two or more generations of females are present in a social group it is almost always older females that “win” the conflict over reproduction and retain breeding status, while younger females remain in the group as reproductively suppressed helpers. Why should humans be different? A critical factor is the unusual demography of humans (Cant & Johnstone 2008, Johnstone & Cant 2008). A “kinship dynamics” model suggests that the unusual dispersal and mating patterns of great apes, and some cetaceans, predispose these species to the evolution of early reproductive cessation and late-life helping behavior, in contrast to the majority of other mammalian species (Johnstone & Cant 2008). Moreover, relatedness asymmetries that arise as a result of demography are predicted to give younger females a decisive advantage in reproductive conflict with older females (Cant & Johnstone 2008). Analyses that incorporate demography can therefore help to explain both the unusual taxonomic distribution of menopause and the timing of reproductive cessation in humans, as we describe below.
Demography and kin selection across the lifespan Demography is important to consider in a model of menopause because the relatedness between a female and other breeders in her group can change as she ages, thereby affecting the strength of kin selection for acts such as helping or early reproductive cessation. Building on the infinite-island modeling framework (Wright 1931, Taylor 1992), Johnstone & Cant (2008) construct a general model of “kinship dynamics” to explore the impact of sex-biased dispersal and intra- versus extra-group mating on the strength of kin selection across the lifespan. They derive general formulae which track the changes in genetic similarity between males, females, and offspring in a local group (or “island”) as some individuals disperse, and others die and are replaced. These patterns of relatedness are then used to determine how the strength of kin selection across the lifespan varies for eight representative patterns of demography, which differ as to whether dispersal is male- or female-biased, and whether mating occurs locally or outside the group.
37
38
M. A. Cant et al. The results indicate that where dispersal is male-biased and mating occurs locally, the average relatedness of a female to the offspring produced in her local group decreases as she ages (Figure 2.6A). In these circumstances acts which benefit relatives, such as helping or reproductive restraint, are favored more strongly earlier rather than later in life. By contrast, female-biased dispersal and local mating leads to an increase with age in the average relatedness of a female to the offspring produced in her local group (Figure 2.6B). This effect arises under female-biased dispersal because a female’s relatedness to local male breeders, initially low, increases as her sons mature and remain in the group (relatedness among local female breeders, by contrast, starts low and remains low because daughters disperse). In these circumstances acts which benefit other group members are favored later, rather than earlier in life. Female-biased dispersal, therefore, predisposes females to early reproductive cessation and late-life helping. Most social mammals exhibit male-biased dispersal and female philopatry (Greenwood 1980, Pusey & Packer 1987, Clutton-Brock 1998, Lawson Handley & Perrin 2007). By contrast, three lines of evidence suggest that the evolutionary history of Homo has been characterized by female-biased dispersal and male philopatry. First, our closest primate relatives, chimpanzees, bonobos (Pan paniscus), and gorillas, are unusual among primates because they exhibit female-biased dispersal, and male dispersal is rare (Pusey et al. 1997, Boesch & Boesch-Ackermann 2000, Nishida et al. 2003, Stokes et al. 2003, Yamigawa & Kahekwa 2004, Eriksson et al. 2006). Second, patterns of variation in mitochondrial DNA and the Y-chromosome are consistent with greater rates of female than male dispersal (Seilstad et al. 1998, Oota et al. 2001), at least on the relevant, local scale (Wilder et al. 2004). Finally, female-biased transfer is common in modern human hunter–gathers (Ember 1978). For example, an influential analysis by Ember (1978) concluded that only 16.2% of 179 hunter– gatherer societies show a matrilocal pattern of residence. More recent analyses (Alvarez 2004, Marlowe 2004) have contested Ember’s classification, mainly because in the majority of human societies dispersal is merely biased towards, rather than restricted to, one sex or the other (and so, it is argued, should be classed as “bilocal”). Leaving this controversy over “strict” patrilocality aside, it remains the case that female-biased dispersal is considerably more common than the reverse pattern (Marlow 2004). Taken together, these three independent lines of evidence suggest that mutations affecting female reproductive lifespan are likely to have arisen in an ancestral social environment in which dispersal was female-biased. Johnstone and Cant’s (2008) model suggests, therefore, that humans were predisposed to evolve early reproductive cessation and late-life helping because the
Figure 2.6 Effect of demography on patterns of age-specific relatedness and selection for social acts across the lifespan (modified from Johnstone & Cant 2008). Graphs on the left show age-specific relatedness to a breeding female of other females (solid lines) and of males (dotted lines) in her group, as a function of her age. The dashed curves show mean relatedness to a female of other breeders, averaging across both sexes. Age is scaled relative to mean generation time. Results are plotted for three different demographic systems. (A) A high rate of male dispersal and a low rate of female dispersal, with mating occurring within the local group. (B) High female dispersal and low male dispersal, with mating occurring within the local group. (C) Low male and low female dispersal, with mating occurring outside the local group. Graphs on the right show the patterns of kin selection across the lifespan associated with these three systems. A female can perform social acts which result in an immediate gain of b offspring for other breeders in the group, at an immediate cost of c offspring to herself. The graphs plot the absolute magnitude of the c/b ratio below which a social action may be favored in females of different ages. Positive values indicate that selection will favor helping behavior (i.e. acts for which b > 0) when c/b falls above zero but below the value shown (in the lightly shaded area), while negative values indicate that selection will favor harming behavior (i.e. b < 0) when c/b falls below zero but above the value shown (in the heavily shaded area). For more details see Johnstone & Cant (2008).
40
M. A. Cant et al. relatedness of a female to the offspring produced in her local group increased as she aged. By contrast, most other social mammals exhibit male-biased dispersal, and so are predisposed to the evolution of early-life, rather than late-life, helping. The same model predicts that another unusual pattern of dispersal and mating can give rise to an age-specific increase in local relatedness: one in which both sexes are philopatric but mating occurs outside the local group (Figure 2.6C). Interestingly, this is precisely the demographic system exhibited by menopausal cetaceans. Both male and female resident killer whales are philopatric, but mate outside of the local group (Baird 2000, Whitehead & Mann 2000); it is not known whether transient killer whales, which do disperse, also exhibit menopause). Short-finned pilot whales (Globicephala macrorhynchus) are thought to exhibit a similar pattern – there is clear evidence for their sister species the long-finned pilot whale (G. melas) (Amos 1993), and the available genetic data suggest that the short-finned and long-finned species are comparable (Amos 1998). Again, the predicted increase in relatedness with age in this case is driven by relatedness through males. Consequently in these species mothers should direct their help towards sons, with the aim of improving their extra-group mating success. Observations of resident killer whales fit well with this prediction: mothers maintain closer associations with their adult sons than with their adult daughters, and may aid their son’s foraging efforts, or form effective alliance partners for them in agonistic encounters with other males (Baird 2000).
Relatedness asymmetries and conflict resolution What would be the consequences of female-biased dispersal for the resolution of reproductive conflict in ancestral hominids? This question can be explored using a simple model of the human social unit (Figure 2.7A; Cant & Johnstone 2008). We assume that males and females are socially monogamous, but allow for a proportion p of offspring to be fathered by unrelated males. For simplicity we assume in the basic model that only females disperse, but the qualitative results of our analysis hold where dispersal is merely biased toward, rather than restricted to, females. Females leave their natal groups at maturity, pair with a male of similar age, and join his natal social group. Consequently, when a young female first arrives in the group, she has no other genetic relatives present. This female can choose to breed herself and produce offspring to whom she is related by 1/2, or to refrain from breeding and assist the breeding attempts of the older female (i.e. the mother of her mate). This older female produces offspring to whom the younger female is
Reproductive conflict and the evolution of menopause
Figure 2.7 Schematic representation of relatedness asymmetries between generations under sex-biased dispersal (from Cant & Johnstone 2008). Male and female symbols represent parents. (A) Where females disperse and immigrate into a patrilocal group, a mother is related to the offspring of a daughter-in-law by (1 p)/4, where p is the probability of extra-pair paternity. The daughter-in-law, by contrast, is completely unrelated to the mother’s offspring. Thus the difference in relatedness to own versus other offspring is greater for younger than for older females. (B) Where males disperse and immigrate into matrilocal groups, by contrast, the difference in relatedness of a female to her own versus the other breeder’s offspring is greater for the mother than for the daughter.
unrelated. The difference in relatedness to offspring produced directly versus offspring produced by helping is therefore 1/2. The older female, by contrast, can choose to breed and produce offspring of relatedness 1/2, or refrain from breeding and help to rear grandoffspring, to whom she is related by (1 p)/4. The relatedness differential between breeding and helping for the older female is therefore 1/2 (1 p)/4, or (1 þ p)/4. This means that so long as there is any chance that her son fathered her putative grandchildren (i.e. p < 1), the difference in relatedness to offspring produced by breeding rather than helping is lower for the older female than for the younger female. As a result, a younger female will have an advantage in reproductive competition with older females because she is insensitive to the costs she inflicts on an older female by breeding. This contrasts with the situation where dispersal is male-biased (as in most social mammals). Here relatedness asymmetry favors older females over younger females, so older females are expected to have an advantage in reproductive conflict with the younger generation (Figure 2.7B).
41
42
M. A. Cant et al. A formal game-theoretic treatment of this model is provided by Cant and Johnstone (2008). Their analysis is based on the “tug-of-war” model of Reeve et al. (1998), which is the standard model for the analysis of reproductive conflict when both parties exhibit partial control over the outcome (Johnstone 2000, Cant & Shen 2006, Reeve & Ho¨lldobler 2007). Both females engage in competition over reproductive shares at the expense of total group resources. Increased selfish investment, therefore, results in a larger slice of a smaller reproductive “pie.” Regardless of the relative strength of the two females, the evolutionarily stable solution is for the older female to commit to zero reproduction and allow the younger female to claim all of the available reproduction. This is an example of an “endogenous” or “natural” Stackelberg solution in which both players prefer to act in sequence (rather than submit simultaneous “sealed bids”), and both agree on who should move first (Albaek 1990, van Damme & Hurkens 1999). Endogenous Stackelberg equilibria are interesting from a biological perspective because they can explain the evolution of commitment strategies which are profitable precisely because they cannot credibly be changed (Nesse 2001, Cant & Shen 2006). Thus, in the conflict between older and younger females the older female’s first move of zero investment is advantageous only if it is perceived to be irreversible by the younger female. Permanent sterility as a consequence of rapid reproductive senescence would be one highly effective way to commit credibly to a first move of zero investment in reproduction.
Conclusion To summarize the above arguments, a social environment in which dispersal is female-biased generates relatedness asymmetries between older and younger females and the offspring they produce. Conflict over reproduction in these circumstances is predicted to favor older females who commit to zero reproduction when females of a younger generation start to breed. Given a pattern of female-biased dispersal during the period of lengthening human lifespan, there would be little selection to extend or maintain female reproductive capacity beyond the age at which a woman might expect to become a paternal grandmother. This can account for the high and accelerating rate of oocyte loss in humans (Figure 2.2) leading to sterility in mid-life, and the exceptionally low reproductive overlap between generations in humans (Figure 2.5). The intensity of reproductive competition and the magnitude of the benefits that can be conferred by helping must also be important, however, because chimpanzees and bonobos exhibit strongly female-biased dispersal but are not unusual in their degree of reproductive overlap (Figure 2.5A).
Reproductive conflict and the evolution of menopause A number of studies have reported that the presence of a paternal grandmother has relatively little effect on offspring survival compared to that of a maternal grandmother (Bereczkei & Dunbar 1997, Sear et al. 2000, 2002), or even a negative effect on offspring survival (Beise & Voland 2002). For example, data from the Krummho¨rn region of Germany in the eighteenth and nineteenth centuries show that the chance of stillbirth for a daughter-in-law was increased by 35% if the paternal grandmother was present in the household, and these mortality costs are particularly high at the start of the daughter-inlaw’s marriage (Beise & Voland 2002, Voland & Beise 2005). While these data offer evidence of reproductive conflicts between generations in human families, the positive impact of maternal grandmothers seems at odds with our assumption of a female-biased dispersal system. It is important to distinguish, however, between the evolutionary origins of a lifehistory trait and the behavioral strategies that are employed once that trait has evolved. The universality of menopause in modern humans, despite vast differences in social systems and access to resources, illustrates the flexibility of behavior compared to the physiological processes underlying rapid reproductive senescence. The reproductive conflict model does not imply that older females should not help daughters if the social system subsequently changes to become less female-biased, or mothers are able to maintain kin ties with their daughters. Indeed, given a choice between helping daughters versus sons, mothers should direct their help preferentially towards daughters, since grandchildren through sons may have been fathered by extra-pair males. From a woman’s perspective, therefore, a flexible or “bilocal” system that allowed her to direct care toward daughters late in life would be preferable to strict patrilocality throughout her life. Most modern forager societies exhibit a degree of flexibility of this kind (Alvarez 2004, Marlowe 2004). The model is of course simplistic in many respects, but it remains a useful tool with which to make testable predictions. For example, the assumptions of the model could be tested by examining whether relatedness asymmetries exist within family units of natural-fertility populations using genealogical and/or genetic data. Given a larger sample size, from either current or historical populations, one should be able to detect a cost to females of breeding alongside reproductive grandmothers, similar to the demonstrated costs of co-breeding within generations in polygynous marriages. The fact that the ancestral social system may have changed in more recent times offers an opportunity to test our model. Attempts to test the predictions of the model could utilize the wide variety of patterns of dispersal and marital residence exhibited by modern humans (Ember 1978, Marlowe 2004). If a system of male-biased dispersal and matrilocality were to persist for many generations, our model predicts that
43
44
M. A. Cant et al. selection to minimize reproductive overlap would weaken. Consequently, we would predict mean age at last birth to be higher in historically matrilocal compared to patrilocal societies, and cultural restrictions on reproduction by grandmothers to be less prevalent in the former than the latter. The stability of the dispersal system over time could be inferred from genetic data, since genetic studies show that matrilocal societies show less variation in mitochondrial DNA, and greater variation in Y-chromosome DNA, than patrilocal societies (Oota et al. 2001). More speculatively, if the rate of reproductive senescence is adjusted on an individual level we might predict that women who have only daughters (or have daughters first) should cease reproduction (and perhaps undergo menopause) later than women who have only sons (or have sons first). The models described in this chapter raise their own questions about the evolutionary origins and taxonomic distribution of menopause. Why is this particular life-history pattern so unusual among mammals, and absent from other vertebrates? Male philopatry and female-biased dispersal is characteristic of other primates which do not exhibit menopause, such as the hamadryas baboon (Papio hamadryas: Pusey & Packer 1987, Hammond et al. 2006) and red colobus (Piliocolobus rufomitratus: Marsh 1979). It is also the pattern exhibited by large cooperatively breeding canids (Moehlman & Hofer 1997), and the majority of cooperatively breeding birds (Greenwood 1980, Brown 1987). Clearly, neither male-biased philopatry nor a benefit of cooperative breeding are sufficient to account for the evolution of menopause. The magnitude and nature of the potential fitness benefits that can be conferred by helping will be of key importance. Cetacean biologists have suggested that the main benefit that can be conferred by older female killer whales and pilot whales is information and experience (McAuliffe & Whitehead 2005). The potential benefits of information transfer are probably even greater in humans, and other aspects of human social biology (e.g. communal food-gathering, short inter-birth intervals, tool use) may also contribute to making helping a cost-effective strategy later in life. It is important to remember, however, that the indirect fitness benefits of helping represent just one side of the equation. The reproductive life history of highly social animals such as humans and toothed whales will also be shaped, in substantial part, by conflict over reproduction. An integrated approach which considers all the potential inclusive fitness consequences of social acts within ancestral families promises to yield a much-improved understanding of menopause. Acknowledgments We thank Bill Amos, Michael Briga, Rebecca Chicot, Tim CluttonBrock, Nick Davies, Jeremy Field, Øistein Holen, Reinmar Hager, Sarah Hrdy,
Reproductive conflict and the evolution of menopause Clara B. Jones, Rebecca Kilner, Nobu Kutsukake, Laurent Lehmann, Virpi Lummaa, Ruth Mace, Katie McAuliffe, Francis Ratnieks, and Joan Silk, and two anonymous reviewers, for discussion and helpful comments on earlier manuscripts. MAC and AFR are funded by Royal Society University Research Fellowships. References Albaek, S. (1990). Stackelberg leadership as a natural solution under cost uncertainty. Journal of Industrial Economics, 38, 335–347. Alexander, R. D. (1974). The evolution of social behaviour. Annual Review of Ecology and Systematics, 5, 325–383. Alvarez, H. P. (2004). Residence groups among hunter–gatherers: a view of the claims and evidence for patrilocal bands. In B. Chapais & C. M. Berman, eds., Kinship and Behavior in Primates. Oxford: Oxford University Press, pp. 420–442. Amey, F. K. (2002). Polygyny and child survival in sub-Saharan Africa. Social Biology, 49, 74–89. Amos, W. (1993). Use of molecular probes to analyse pilot whale pod structure: two novel analytical approaches. Symposia of the Zoological Society of London, 66, 33–48. Amos, W. (1998). Culture and genetic evolution in whales. Science, 284, 2055a. Amundsen, D. W. & Diers, C. J. (1970). Age of menopause in classical Greece and Rome. Human Biology, 42, 79–86. Baird, R. W. (2000). The killer whale: foraging specializations and group hunting. In J. Mann, R. Connor, P. L. Tyack, & H. Whitehead, eds., Cetacean Societies. Chicago, IL: University of Chicago Press, pp. 127–153. Beekman, M., Komdeur, J., & Ratnieks, F. L. W. (2003). Reproductive conflicts in animal societies: who has power? Trends in Ecology and Evolution, 18, 277–282. Beise, J. & Voland, E. (2002). A multilevel event history analysis of the effects of grandmothers on child mortality in a historical German population (Krummho¨rn, Ostfriesland, 1720–1874). Demographic Research, 7, 469–497. Bereczkei, T. & Dunbar, R. I. M. (1997). Female-biased reproductive strategies in a Hungarian Gypsy population. Proceedings of the Royal Society of London B, 264, 17–22. Blurton Jones, N. G., Hawkes, K., & O’Connell, J. F. (2002). The antiquity of postreproductive life: are there modern impacts on hunter–gatherer postreproductive lifespans? American Journal of Human Biology, 14, 184–205. Boesch, C. & Boesch-Achermann, H. (2000). The Chimpanzees of the Tai Forest. Oxford: Oxford University Press. Brown, J. L. (1987). Helping and Communal Breeding in Birds. Princeton, NJ: Princeton University Press. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184.
45
46
M. A. Cant et al. Cant, M. A. & Johnstone, R. A. (2008). Reproductive conflict and the separation of reproductive generations in humans. Proceedings of the National Academy of Sciences of the USA, 105, 5332–5336. Cant, M. A. & Shen, S. -F. (2006). Endogenous timing in competitive interactions among relatives. Proceedings of the Royal Society of London B, 273, 171–178. Caspari, R. & Lee, S. H. (2004). Older age becomes common late in human evolution. Proceedings of the National Academy of Sciences of the USA, 101, 10895–10900. Cavalli-Sforza, L. L. (1983). The transition to agriculture and some of its consequences. In D. J. Ortner, ed., How Humans Adapt: a Biocultural Odyssey. Washington, DC: Smithsonian Institute Press, pp. 103–126. Clutton-Brock, T. H. (1998). Female transfer and inbreeding avoidance in social mammals. Nature, 337, 70–72. Cohen, A. A. (2004). Female post-reproductive lifespan: a general mammalian trait. Biological Reviews, 79, 733–750. Ember, C. (1978). Myths about hunter gatherers. Ethnology, 17, 439–448. Emlen, S. T. (1991). Evolution of cooperative breeding in birds and mammals. In J. R. Krebs and N. B. Davies, eds., Behavioural Ecology: an Evolutionary Approach, 3rd edn. Oxford: Blackwell, pp. 301–337. Emlen, S. T. (1995). An evolutionary theory of the family. Proceedings of the National Academy of Sciences of the USA, 92, 8092–8099. Eriksson, J., Siedel, H., Lukas, D., et al. (2006). Y-chromosome analysis confirms highly sex-biased dispersal and suggests a low male effective population size in bonobos (Pan paniscus). Molecular Ecology, 15, 939–949. Faddy, M. J. & Gosden, R. G. (1996). Ovary and ovulation: a model conforming the decline in follicle numbers to the age of menopause in women. Human Reproduction, 11, 1484–1486. Faddy, M. J., Gosden, R. G., Gougeon, A., Richardson, S. J., & Nelson, J. F. (1992). Accelerated disappearance of ovarian follicles in mid-life: implications for forecasting menopause. Human Reproduction, 7, 1342–1346. Gosden, R. G. & Telfer, E. (1987). Numbers of follicles and oocytes in mammalian ovaries and their allometric relationships. Journal of Zoology, 211, 169–175. Greenwood, P. J. (1980). Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour, 28, 1140–1162. Gurven, M. (2005). To give or not to give: the behavioural ecology of human food transfers. Behavioral and Brain Sciences, 27, 543–559. Gurven, M., Hill, K., & Jakugi, F. (2004). Why do foragers share and sharers forage? Explorations of social dimensions of foraging. Research in Economic Anthropology, 23, 19–43. Hamilton, W. D. (1966). The moulding of senescence by natural selection. Journal of Theoretical Biology, 12, 12–45. Hammond, R. L., Lawson Handley, L. J., Winney, B. J., Bruford, M. W., & Perrin, N. (2006). Genetic evidence for female-biased dispersal and gene flow in a polygynous primate. Proceedings of the Royal Society of London B, 273, 479–484.
Reproductive conflict and the evolution of menopause Hawkes, K. (2003). Grandmothers and the evolution of human longevity. American Journal of Human Biology, 15, 380–400. Hawkes, K., O’Connell, J. F., Blurton Jones, N. G., Alvarez, H., & Charnov, E. L. (1998). Grandmothering, menopause, and the evolution of human life histories. Proceedings of the National Academy of Sciences of the USA, 95, 1336–1339. Helle, S., Lummaa V., & Jokela, J. (2005). Are reproductive and somatic senescence coupled in humans? Late, but not early reproduction correlated with longevity in historical Sami women. Proceedings of the Royal Society of London B, 272, 29–37. Hill, K. & Hurtado, A. M. (1991). The evolution of reproductive senescence and menopause in human females. Human Nature, 2, 315–350. Hill, K. & Hurtado, A. M. (1996). Ache Life History: the Ecology and Demography of a Foraging People. New York, NY: De Gruyter. Isaac, B. L. & Feinberg, W. E. (1982). Marital form and infant survival among the Mende of rural upper Bambara Chiefdom, Sierra-Leone. Human Biology, 54, 627–634. Islam, M. M. & Yadava, R. C. (1997). Estimation of fecundability: levels and trends in Bangladesh. Journal of Family Welfare, 43, 13–20. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Cant, M. A. (2008). Sex differences in dispersal and the evolution of helping and harming. American Naturalist, 172, 318–330. Jones, E. C. & Krohn, P. L. (1961). The effects of hypophysectomy on age changes in the ovaries of mice. Journal of Endocrinology, 21, 469–496. Kaplan, H. & Hill, K. (1985). Food sharing among Ache foragers: tests of explanatory hypotheses. Current Anthropology, 26, 223–245. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Lack, D. (1947). The significance of clutch size. Ibis, 89, 309–352. Lahdenpera¨, M., Lummaa, V., & Russell, A. F. (2004a). Menopause: why does fertility end before life? Climacteric, 7, 327–331. Lahdenpera¨, M., Lummaa, V., Helle, S., Tremblay, M., & Russell, A. F. (2004b). Fitness benefits of prolonged post-reproductive lifespan in women. Nature, 428, 178–181. Lawson Handley, L. J. & Perrin, N. (2007). Advances in our understanding of mammalian sex-biased dispersal. Molecular Ecology, 16, 1559–1578. Loudon, I. (1993). Death in Childbirth: an International Study of Maternal Care and Maternal Mortality 1800–1950. Oxford: Oxford University Press. Mace, R. & Sear, R. (2005). Are humans cooperative breeders? In E. Voland, A. Chasiotis, & W. Shiefenhovel, eds., Grandmotherhood: the Evolutionary Significance of the Second Half of Female Life. New Brunswick, NJ: Rutgers University Press, pp. 143–159. Marlowe, F. W. (2000). The patriarch hypothesis: an alternative explanation of menopause. Human Nature, 11, 27–42. Marlowe, F. W. (2004). Marital residence among foragers. Current Anthropology, 45, 277–284.
47
48
M. A. Cant et al. Marsh, C. W. (1979). Female transference and mate choice among Tana River red colobus. Nature, 281, 568–569. Marsh, H. & Kasuya, T. (1986). Evidence for reproductive senescence in female cetaceans. Reports of the International Whaling Commission (Special Issue), 8, 57–74. McAuliffe, K. & Whitehead, H. (2005). Eusociality, menopause and information in matrilineal whales. Trends in Ecology and Evolution, 20, 650. Medawar, P. B. (1952). An Unsolved Problem in Biology. London: HK Lewis. Mizroch, S. A. (1981). Analyses of some biological parameters in the Antarctic fin whale. Reports of the International Whaling Commission, 31, 425–434. Moehlman, P. D. & Hofer, H. (1997). Cooperative breeding, reproductive suppression, and body mass in canids. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 76–129. Moss, C. J. (2001). The demography of an African elephant (Loxodonta africana) population in Amboseli, Kenya. Journal of Zoology, 255, 145–156. Nesse, R. M., ed. (2001). Evolution and the Capacity for Commitment. New York, NY: Sage Foundation Press. Nichols, S. M., Bavister, B. D., Brenner, C. A., et al. (2005). Ovarian senescence in the rhesus monkey (Macaca mulatta). Human Reproduction, 2, 79–83. Nishida, T., Corp, N., Hamai, M., et al. (2003). Demography, female life history and reproductive profiles among the chimpanzees of Mahale. American Journal of Primatology, 59, 99–121. Olesiuk, P., Bigg, M. A., & Ellis, G. M. (1990). Life history and population dynamics of resident killer whales (Orcinus orca) in the coastal waters of British Columbia and Washington State. Reports of the International Whaling Commission (Special Issue), 12, 209–243. Oota, H., Settheetham-Ishida, W., Tiwawech, D., Ishida, T., & Stoneking, M. (2001). Human mtDNA and Y-chromosome variation is correlated with matrilocal versus patrilocal residence. Nature Genetics, 29, 20–21. Patel, T. (1994). Fertility Behaviour: Population and Society in a Rajasthan Village. Oxford; Delhi: Oxford University Press. Pebley, A. R. & Mbugua, W. (1989). Polygyny and fertility in sub-Saharan Africa. In R. J. Lesthaeghe, ed., Reproduction and Social Organization in Sub-Saharan Africa. Berkeley, CA: University of California Press, pp. 339–365. Peccei, J. S. (1999). First estimates of heritability in the age of menopause. Current Anthropology, 40, 553–558. Peccei, J. S. (2001a). Menopause: adaptation or epiphenomenon? Evolutionary Anthropology, 10, 43–57. Peccei, J. S. (2001b). A critique of the grandmother hypothesis: old and new. American Journal of Human Biology, 13, 434–452. Pennington, R. (2001). Hunter–gatherer demography. In C. Panter-Brick, R. H. Layton, & P. Rowley-Conwy, eds., Hunter–Gatherers: an Interdisciplinary Perspective. Cambridge: Cambridge University Press, pp. 170–204.
Reproductive conflict and the evolution of menopause Pettay, J. E., Kruuk, L. E. B., Jokela, J., & Lummaa, V. (2005). Heritability and genetic constraints of life-history trait evolution in preindustrial humans. Proceedings of National Academy of Sciences of the USA, 102, 2838–2843. Pusey, A. E. & Packer, C. (1987). Dispersal and philopatry. In B. B. Smuts, D. L. Cheney, R. M. Seyfarth, & R. W. Wrangham, eds., Primate Societies. Chicago, IL: University of Chicago Press, pp. 250–266. Pusey, A. E., Williams, J., & Goodall, J. (1997). The influence of dominance rank on the reproductive success of female chimpanzees. Science, 277, 828–831. Ratnieks, F. L. W., Foster, K. R., & Wenseleers, T. (2006). Conflict resolution in insect societies. Annual Review of Entomology, 51, 581–608. Reeve, H. K. & Ho¨lldobler, B. (2007). The emergence of a superorganism through intergroup competition. Proceedings of the National Academy of Sciences of the USA, 104, 9736–9740. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Rogers, A. R. (1993). Why menopause? Evolutionary Ecology, 7, 406–420. Rose, M. R. (1991). Evolutionary Biology of Aging. New York, NY: Oxford University Press. Roth, E. A. & Kurup, K. B. (1988). Demography and polygyny in a southern Sudanese agro-pastoralist society. Culture, 8, 67–73. Sear, R., Mace, R., & McGregor, I. A. (2000). Maternal grandmothers improve nutritional status and survival of children in rural Gambia. Proceedings of the Royal Society of London B, 267, 1641–1647. Sear, R., Steele, F., McGregor, I. A., & Mace, R. (2002). The effects of kin on child mortality in rural Gambia. Demography, 39, 43–63. Seilstad, M. T., Minch, E., & Cavalli-Sforza, L. L. (1998). Genetic evidence for a higher female migration rate in humans. Nature Genetics, 20, 278–280. Shanley, D. P. & Kirkwood, T. B. L. (2001). Evolution of the human menopause. Bioessays, 23, 282–287. Skinner, G. W. (2004). Grandparental effects on reproductive strategizing Nobi villagers in Early Modern Japan. Demographic Research, 11, 111–148. Snieder, H., MacGregor, A. J., & Spector, T. D. (1998). Genes control the cessation of a woman’s reproductive life: a twin study of hysterectomy and age at menopause. Journal of Clinical Endocrinology and Metabolism, 83, 1875–1880. Stokes, E. J., Parnell, R. J., & Olejniczak, C. (2003). Female dispersal and reproductive success in wild western lowland gorillas (Gorilla gorilla gorilla). Behavioral Ecology and Sociobiology, 54, 329–339. Strassmann, B. I. (1997). Polygyny as a risk factor for child mortality among the Dogon. Current Anthropology, 38, 688–695. Taylor, P. D. (1992). Altruism in viscous populations: an inclusive fitness approach. Evolutionary Ecology, 6, 352–356. Van Damme, E. & Hurkens, S. (1999). Endogenous Stackelberg leadership. Games and Economic Behavior, 28, 105–129.
49
50
M. A. Cant et al. Voland, E. & Beise, J. (2005). “The husband’s mother is the devil in the house”: data on the impact of the mother-in-law on stillbirth mortality in historical Krummho¨rn (18th–19th century Germany) and some thoughts on the evolution of postgenerative female life. In E. Voland, A. Chasiotis, & W. Shiefenhovel, eds., Grandmotherhood: the Evolutionary Significance of the Second Half of Female life. New Brunswick, NJ: Rutgers University Press, pp. 239–255. Voland, E., Chasiotis, A., & Shiefenhovel, W., eds. (2005). Grandmotherhood: the Evolutionary Significance of the Second Half of Female Life. New Brunswick, NJ: Rutgers University Press. Whitehead, H. & Mann, J. (2000). Female reproductive strategies of cetaceans. In J. Mann, R. Connor, P. L. Tyack, & H. Whitehead, eds., Cetacean Societies. Chicago, IL: University of Chicago Press, pp. 219–246. Wilder, J. A., Kingan, S. B., Mobasher, Z., Pilkington, M. M., & Hammer, M. F. (2004). Global patterns of human mitochondrial DNA and Y-chromosome structure are not influenced by higher migration rates of females versus males. Nature Genetics, 36, 1122–1125. Williams, G. C. (1957). Pleiotropy, natural selection, and the evolution of senescence. Evolution, 11, 398–411. Wilson, M. (1957). Rituals of Kinship among the Nyakyusa. London: Oxford University Press. Wood, J. W. (1994). Dynamics of Human Reproduction. New York, NY: De Gruyter. Wright, S. (1931). Evolution in Mendelian populations. Genetics, 16, 97–159. Yamagiwa, J. & Kahekwa, J. (2004). Dispersal patterns, group structure, and reproductive parameters of eastern lowland gorillas at Kahuzi in the absence of infanticide. In M. M. Robbins, P. Sicotte, & K. J. Stewart, eds., Mountain Gorillas: Three Decades of Research at Karisoke. Cambridge: Cambridge University Press, pp. 89–122.
II
Testing assumptions and predictions of skew models
3
Reproductive skew in femaledominated mammalian societies k a y e . ho l e k a m p a n d a n n e l . e ng h
Summary We review available data documenting reproductive skew in the small group of mammals characterized by female dominance over males, focusing mainly on lemurs and spotted hyenas (Crocuta crocuta). Although most females in all lemur species examined here appear to bear young at each opportunity, we know very little about variation in longer-term reproductive success or rates of reproduction among female lemurs. Therefore we cannot draw firm conclusions in regard to reproductive skew among female lemurs except that at present this appears to be slight. However, current data show that female lemurs typically mate with multiple males, and that a substantial fraction of litters containing multiple offspring is sired by more than one male. The extent of reproductive skew in male lemurs varies among species, but there is a slight trend, among the lemur species for which genetic data exist, for male skew to decrease as the intensity of female dominance increases. Variance in reproductive success among female spotted hyenas appears to be substantially greater than it is in male-dominated species in which plural breeding occurs. In this species, female dominance, combined with virilization of the external genitalia, may increase female control over mating to its extreme limit, such that we find very little reproductive skew among males relative to that found in other polygynous mammals. The most dominant male hyenas often achieve very little reproductive success. Overall, reproductive skew among females in female-dominated mammals appears to be the same as or slightly greater than that in male-dominated species, whereas skew among males in femaledominated species generally tends to be relatively low. Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
53
54
K. E. Holekamp, A. L. Engh Introduction Early models of reproductive skew (e.g. Vehrencamp 1983a, 1983b) assumed that the distribution of reproduction within vertebrate social groups is determined by interactions occurring within each sex. However, as the study of sexual selection revealed new mechanisms of multisexual mate choice and intrasexual competition for mates operating both before and after copulation in many species (e.g. Birkhead & Moller 1993; Eberhard 1996), it became apparent that there is in fact much potential for females to affect reproductive success in males and vice versa. Although reproductive success is a property of individuals, usually measured as the number of surviving offspring an individual produces in its lifetime, reproductive skew describes the extent to which reproductive success varies among members of a specific population of animals. Recent studies of reproductive skew have emphasized that many gregarious vertebrates occur in stable mixed-sex groups, and that under these circumstances in particular, control over reproduction exerted by members of one sex might affect reproductive skew in the other (e.g. Whittingham & Dunn 1998, Cant 2000, Van Hooff 2000). Cant and Reeve (2002) addressed the question of how female control influences the distribution of paternity among males in cooperatively breeding vertebrates. However, a related question that, to our knowledge, has not previously been addressed asks more broadly whether, and how, reproductive skew within animal societies is affected by which sex is socially dominant. In most mammals, sexual dimorphisms in body size, weaponry, and aggressive behavior occur in association with contest competition among males for access to females or for control of the resources needed by females for successful reproduction (Darwin 1871, Short & Balaban 1994). As males are usually larger, better armed, and more aggressive than females, they seldom have difficulty achieving social dominance over females, and males easily win contests for resources needed by both sexes. Examples abound among ungulates (e.g. red deer, Cervus elaphus), cercopithecine primates (e.g. baboons, Papio spp.), and both pinniped and fissiped carnivores (e.g. elephant seals, Mirounga angustirostris, and lions, Panthera leo, respectively). Indeed, males are larger than females, and socially dominant to them, in all but a handful of mammalian species. Our focus in this chapter is on the small group of “role-reversed” species in which females are socially dominant to males. We define female dominance as the ability of adult females to win against adult males in contests over resources, and to evoke submissive behavior from males in dyadic contexts. Although females in some species dominate males seasonally (e.g. reindeer, Rangifer tarandus; Espmark 1964), here we consider
Reproductive skew in female-dominated mammalian societies only species in which female dominance occurs year-round. We focus our attention on those species for which data exist documenting individual variation in reproductive success. The species that satisfy these criteria include two species of mole-rats (naked mole-rats, Heterocephalus glaber, and common molerats, Cryptomys hottentotus), spotted hyenas (Crocuta crocuta), and several species of lemurs. Although females dominate males in some populations of bonobos (Pan paniscus: Stanford 1998, Vervaecke et al. 2000; but see Paoli et al. 2006) and Garnett’s greater bushbabies (Otolemur garnettii: Hager & Welker 2001), data documenting reproductive skew are rare in bonobos (Gerloff et al. 1999) and unavailable in O. garnettii, so we discuss these species only briefly here. The species satisfying our inclusion criteria range in size from roughly 30 g (naked mole-rats) to 60 kg (spotted hyenas), and they occupy a variety of ecological niches. Some of these female-dominated mammals are seasonal breeders (common mole-rats and lemurs) whereas others breed year-round (naked mole-rats and spotted hyenas). Mean group size in these species ranges from one (e.g. gray mouse lemurs, Microcebus murinus: Eberle & Kappeler 2004b) to approximately 80 (naked mole-rats: Bennett & Faulkes 2000). Thus female dominance represents one of a very small number of traits all these species have in common. Here we will examine effects of female dominance in these species on the partitioning of reproduction within each sex (Table 3.1). Because the mole-rats are the subject of their own chapter in this volume (Chapter 13), we focus here on lemurs and spotted hyenas.
Reproductive skew among lemurs Lemurs are arboreal forest-dwelling primates that consume a wide array of plant material and insects. All but two of the roughly 50 extant lemur species are endemic to Madagascar (Richard & Dewar 1991, Goodman & Benstead 2005), where, due to long dry seasons and highly seasonal rainfall, reproduction among lemurs is also highly seasonal. Most lemurs are characterized by low basal metabolic rate, lack of sexual dimorphism in body size, roughly even adult sex ratios, relatively small group sizes compared to haplorhine primates, targeted female–female aggression, high infant mortality, and female dominance (Richard & Dewar 1991, Wright 1999). To date, no lemur species are known to be dominated by males (Wright 1999). However, the degree of female dominance among lemur species is variable. In one species, Eulemur fulvus, there is no obvious sex-related dominance structure, whereas in others females have feeding priority (Propithecus verreauxi, Hapalemur griseus, Eulemur mongoz, Phaner furcifer), show partial dominance (Eulemur coronatus, Daubentonia madagascarensis, Mirza coquereli), or are clearly dominant to
55
56
K. E. Holekamp, A. L. Engh Table 3.1 Social system and partitioning of reproduction in female-dominated mammals Degree
Degree of skew among
of female Species
dominance
Social system
Males
Females
Indri (Indri indri)
High
One-male, one-female
Unknown
Unknown
Multi-male, multi-female Moderate
Moderate
groups with offspring Ring-tailed lemur
High
(Lemur catta) Ruffed lemur (Varecia
groups High
Loose multi-male,
High
Multi-male, multi-
variegata) Milne-Edwards sifaka
Unknown
Unknown
Moderate?
Unknown
Unknown
Unknown
Unknown
Unknown
Moderate
Moderate
Moderate
Moderate
Moderate
Very high
Moderate
Very high
Multi-male, multi-female Unknown
Unknown
multi-female groups
(Propithecus diadema
female groups
edwardsi) Golden-crowned sifaka
High
(Propithecus tattersalli) Black lemur (Eulemur
female groups High
macaco) Grey mouse lemur
Multi-male, multiMulti-male, multifemale groups
High
(Microcebus murinus)
Solitary with small female sleeping groups
Spotted hyena (Crocuta
High
crocuta) Naked mole-rat
multi-female groups High
Large, multi-male,
High
Large, multi-male,
(Heterocephalus glaber) Common mole-rat
multi-female groups Moderate
coronatus) Alaotran gentle lemur
to high
multi-female groups
(Cryptomys hottentotus) Crowned lemur (Eulemur
Large, multi-male,
groups Moderate
(Hapalemur griseus
One-male, multi-female groups
Moderate to
Very low
high
alaotrensis) Aye-aye (Daubentonia
Moderate
Solitary
Moderate?
Unknown
Moderate
Small multi-male,
Moderate to
Low
madagascariensis) Verreaux’s sifaka (Propithecus verreauxi)
multi-female groups
Coquerel’s mouse lemur Moderate (Mirza coquereli) Fat-tailed dwarf lemur
Low
Low?
Moderate?
Low?
Unknown
Unknown
Unknown
Unknown
kin clusters Moderate
(Cheirogaleus medius) Garnett’s greater
Solitary in matrilineal
high
One-male, one-female groups with offspring
Moderate
Solitary in matrilineal kin clusters
bushbaby (Otolemur garnettii) Mongoose lemur (Eulemur mungoz)
Moderate
One-male, one-female groups with offspring
Reproductive skew in female-dominated mammalian societies Table 3.1 (cont.) Degree
Degree of skew among
of female Species
dominance
Social system
Males
Females
Fork-marked lemur
Moderate
One-male, one-female
Unknown
Unknown
Moderate
Multi-male, multi-
Low
Moderate?
High
Low?
High
Low?
Unknown
Unknown
(Phaner furcifer) Bonobo (Pan paniscus)
groups with offspring female groups
Red-fronted brown lemur None (Eulemur fulvus rufus) Mayotte brown lemur
Small multi-male, multi-female groups
None
(Eulemur fulvus
Small multi-male, multi-female groups
mayottensis) Pygmy mouse lemur (Microcebus berthae)
Unknown
Solitary, possibly with female kin clusters
males in all contexts (Microcebus murinus, Lemur catta, Indri indri, Varecia variegata, Eulemur macaco: reviewed in Radespiel & Zimmerman 2001). Here we will examine skew in lemur species with well-characterized mating systems, presented in broad order of increasing clarity with which these species exhibit female dominance over males. Verreaux’s sifaka (Propithecus verreauxi)
P. verreauxi (Figure 3.1) live in multi-male multi-female groups that range in size from 2 to 14 members (Lawler et al. 2003). Most groups contain one or two adult females, several adult males, and immatures (Jolly 1998). Both males and females have a clear dominance hierarchy (Brockman 1999, Kraus et al. 1999). Females are usually philopatric, but if a female matures in a group that already contains several breeding females, these adults may force her to disperse (Lawler et al. 2003). All males disperse at puberty (Richard 1992), but this appears to be voluntary. Females often attack older males and force them to leave the group, possibly as a means of avoiding mating with their fathers (Richard 1992, Lawler et al. 2003). Females give birth to only one infant each year (Richard et al. 2002), and only females care for young. Prime-aged females are more fertile and have higher infant survival than do very young and very old females (Richard et al. 2002). Heavier females are more likely to give birth than lighter females, and their young are more likely to survive to weaning (Richard et al. 2000, Lewis & Kappeler 2005). Richard et al. (2002) allege that reproductive success varies
57
58
K. E. Holekamp, A. L. Engh
Figure 3.1 Vereaux’s sifaka (Propithecus verreauxi) is a lemur species characterized by female dominance. The photo shows a young individual. Courtesy of Claudia Fichtel.
substantially among females; however, besides body weight and longevity, they do not identify characteristics associated with increased success. Given that only 48% of adult females give birth each year, infant survival is poor (52%), and female reproduction is not contingent on the previous year’s reproduction, there could potentially be moderate reproductive skew among female P. verreauxi. However, using data from 14 female P. verreauxi that were observed for at least 10 years (Richard et al. 2002), we calculated that there is little variance in female reproductive success (mean ¼ 0.309 offspring surviving to age 1 per year, variance ¼ 0.018, n ¼ 14). Although P. verreauxi are highly seasonal breeders (Lawler et al. 2003), the estrous periods of individual females are asynchronous (Brockman & Whitten 1996). Females often mate with more than one male, though never with all group males (Brockman 1999). Multiple mating may be a strategy to confuse paternity and thereby prevent infanticide, which has been observed in this species (Richard et al. 2002). During the breeding season, males often make forays into neighboring groups (Brockman 1999, Lawler et al. 2003). Females, however, seem to prefer residents (i.e. established immigrants). They mate more often with older, dominant males than with young subordinates and non-residents (Brockman et al. 1998, Brockman 1999), and resident males sire
Reproductive skew in female-dominated mammalian societies 68% of the offspring each year (Lawler et al. 2003). Dominant males try to guard receptive females and harass them if they mate with other males (Richard 1992, Lawler et al. 2003). However, females may refuse to mate with dominant males, and may mate surreptitiously at the group’s periphery or during intergroup conflicts (Brockman 1999). Males engage in numerous, drawn-out fights during the breeding season, leaving ample opportunity for females to mate with whomever they choose (Lawler et al. 2005). In groups without a stable male hierarchy, females may actually provoke fights between males, and usually mate with the winner (Richard 1992). Male reproductive success is highly skewed in this species. Of 70 males sampled in one population, 36 did not father any young, while the remaining 34 fathered 90 infants; on average, each male fathered 1.3 offspring, but the high variance (3.3) indicated strong skew (Lawler et al. 2005). Alaotran gentle lemur (Hapalemur griseus alaotrensis)
This highly endangered subspecies of the gray gentle lemur is found only in the marshes of Madagascar’s Lac Alaotra. They are small, cathemeral folivores (Mutschler 2002) that live in groups of 2–9 individuals, including one or two breeding females, their offspring, and one breeding male (Mutschler et al. 2000, Nievergelt et al. 2002). Females appear to be largely philopatric, whereas males disperse after puberty (Nievergelt et al. 2002). Because both male and female offspring may remain in the group even after they have reached adult size, and male offspring never reproduce until after dispersal, groups contain more breeding females than males despite the fact that adult sex ratios are even (Nievergelt et al. 2002). Reproductive females may mitigate competition from offspring by evicting maturing daughters (Mutschler 1999). Emigrant females then attempt to establish new groups with males, but their mortality rates are assumed to be high (Nievergelt et al. 2002). Both males and females participate in territory defense (Nievergelt et al. 1998), but only females care for young. On average, females produce 1.03 offspring each year (Nievergelt et al. 2002), with no difference in yearly reproductive rate between females in one-female groups and females in two-female groups (Nievergelt et al. 2002; but see Mutschler et al. 2000). Most females give birth over several consecutive years, suggesting that reproductive skew among females is extremely low (Nievergelt et al. 2002). Unlike many other lemur species, H. griseus have a long breeding season, lasting six months (Mutschler et al. 2000). In light of the relatively low levels of reproductive synchrony among females, it should be possible for males to monopolize within-group reproduction. Genetic data suggest that this is exactly what happens. Ninety-two percent of 59 infants were or could have been
59
60
K. E. Holekamp, A. L. Engh fathered by the group male (Nievergelt et al. 2002). A minimum of 8% of all offspring were sired by extra-group males; however, in 60% of these cases, the group male was closely related to the mother of the extra-group infants, and in the other 40%, the female mated with a solitary male after two or more years of unsuccessful reproduction with the group male (Nievergelt et al. 2002). Male reproductive success is twice as high in two-female groups as in one-female groups (Nievergelt et al. 2002), so there may be substantial reproductive skew among males in this species. However, turnover of group males is quite high, presumably because new immigrants expel established males (Nievergelt et al. 2002). If males compete more aggressively for two-female groups, the extent of skew might be tempered by shorter male tenure in these groups. Coquerel’s mouse lemur (Mirza coquereli)
M. coquereli are solitary lemurs that forage alone and rarely share sleeping nests with other adults (Kappeler 1997a). Females’ ranges overlap with those of several other females. Males’ home ranges, in contrast, do not overlap with those of other males, except during the breeding season, when their ranges quadruple in size. Although genetic analyses suggest that dispersal is facultative in both sexes, females tend to settle close to their natal range, creating clusters of matrilineal kin (Kappeler et al. 2002). Reproduction in this species is highly seasonal (Kappeler 1997a). Females give birth to 1–2 infants each year, and most appear to bear young (Stanger 1995, Kappeler et al. 2002), but the degree of reproductive skew among females is unknown. Paternity among M. coquereli males is relatively evenly divided. Because females emit estrus advertisement calls, as do female Daubentonia and Microcebus (Stanger 1995), it has been inferred that they seek to attract multiple mates. Males appear to compete for access to receptive females through scramble competition. Kappeler et al. (2002) found that half of twin litters were fathered by multiple males, and that there was no spatial clustering of infants sired by particular males. In addition, almost one-third of infants could not have been fathered by resident males, suggesting that their sires made brief visits from afar during the mating season. Thus, skew among males in this species appears to be quite low. Fat-tailed dwarf lemur (Cheirogaleus medius)
Fat-tailed dwarf lemurs enter a 6- to 8-month-long torpor during the dry season, so they must mate, wean their young, and fatten for the next dry season during a brief active period each year (Mu¨ller & Thalmann 2002). C. medius live in small family groups that usually consist of a mating pair, infants, and sometimes their young from the previous year (Fietz 1999, Mu¨ller & Thalmann 2002). This
Reproductive skew in female-dominated mammalian societies species is socially monogamous. Adult pairs stay together on the same home range for several years and cooperate in rearing their offspring (Fietz 1999; Mu¨ller & Thalmann 2002). Paternal care is obligate; females who lose their partners are unable to raise young (Fietz 1999). Both males and females disperse, though males usually do so earlier than females (Mu¨ller & Thalmann 2002). Mating is seasonal, starting soon after females emerge from torpor, but not highly synchronous (Fietz 1999). Females usually twin, but can give birth to as many as five infants in a single litter (Fietz 1999, Mu¨ller & Thalmann 2002). Although there are no published data on female skew in this species, it is likely to be low, given the necessity of parental care by males. Given their relatively small testes, obligate paternal care, and social monogamy, C. medius might be expected to be genetically monogamous and show low skew as well, particularly since paired males should be able to monitor their partners’ reproductive status better than should neighboring or floating males. Male C. medius produce sperm plugs (Fietz et al. 2000), suggesting that, if paired males tend to mate with their partners before other males do, they might also have a competitive advantage over extra-pair males. However, of 16 juveniles whose paternity could be determined by Fietz et al. (2000), only 56% were sired by their social fathers. The remaining 44% were fathered either by neighboring (13%) or unsampled (31%) males. No non-territorial male floaters fathered any young. Given these data, there is potential for moderate skew among male C. medius. Some males may father several litters, whereas others clearly do not father any young. However, the exact extent of skew among male C. medius is unknown. Grey mouse lemur (Microcebus murinus)
M. murinus forage alone at night, but they spend the day sleeping in tree holes that they often share with conspecifics (Radespiel 2000). Sleeping groups are stable associations, typically consisting of 2–4 closely related females (Radespiel 2000, Radespiel et al. 2001, Eberle & Kappeler 2003), though larger groups of up to 15 have been observed (Martin 1972). Co-sleeping females regularly groom and sometimes nurse each other’s offspring (Eberle & Kappeler 2003) and have extensive home range overlap (Radespiel et al. 2001). Males usually sleep alone (Radespiel 2000). Molecular evidence suggests that most females remain in their natal areas whereas most males disperse (Radespiel et al. 2001, Wimmer et al. 2002, Fredsted et al. 2004). As the mating season approaches, males expand their home ranges so that, on average, they overlap the ranges of 11 females (Radespiel 2000, Eberle & Kappeler 2004a). Although reproduction is highly seasonal in this species, most female M. murinus are receptive on different nights than nearby females (Eberle &
61
62
K. E. Holekamp, A. L. Engh Kappeler 2002, 2004b). Females entering estrus advertise their status by increasing their frequency of scent marking (Eberle & Kappeler 2004b) and emitting ultrasonic vocalizations (Buesching et al. 1998). Females give birth to 1–3 young, and at least 70% of females conceive litters each year (Radespiel 2000, Eberle & Kappeler 2004a). Heavier females produce heavier offspring, and heavier offspring are more likely to survive to one year (Eberle & Kappeler 2004b). It is unclear whether maternal weight is correlated with other characteristics, such as age or sleeping-group size, or whether there is much variance in female reproductive success. The overall sex ratio in M. murinus populations is male-biased (Radespiel 2000), and this pattern is even more pronounced in the operational sex ratio. Because females enter estrus slightly asynchronously, and because each female is receptive for only a few hours, up to 18 or more males might compete for access to a single receptive female each night (Eberle & Kappeler 2002, 2004b). Behavioral observations confirm that as many as six males approach a receptive female at once (Eberle & Kappeler 2004b). Males gather outside the nest holes of receptive females at dusk, waiting for them to emerge (Radespiel 2000, Eberle & Kappeler 2004a). Once females emerge, males attempt to mate and to guard females from the advances of other males. Heavier males are most successful at repelling competitors (Eberle & Kappeler 2004a), but even heavy males must contend with uncooperative females. Though females show little evidence of direct mate choice, they regularly counteract males’ attempts to monopolize mating, escaping from guarding males in 70% of observed matings (Eberle & Kappeler 2004b). Females are typically promiscuous, mating with as many as seven males in one evening (Eberle & Kappeler 2002, 2004b). Data from captive animals suggest that dominant males father the vast majority of all young (Andres et al. 2001; but see Radespiel et al. 2002) and inhibit testicular development and reproductive behavior in subordinate males (Perret 1992), but there is no evidence of reproductive suppression in the wild (Schmid & Kappeler 1998). Seventeen of 26 wild litters had multiple sires (Eberle & Kappeler 2004b), but reproductive success was biased towards older and heavier males (Eberle & Kappeler 2004a). The large testes of male M. murinus suggest that sperm competition may be very important in this species (Kappeler 1997b, Schmid & Kappeler 1998), and there does appear to be a reproductive advantage to males who mate with females early in their receptive periods (Eberle & Kappeler 2004a). Schmelting (2001) suggests that male reproductive success is constrained by poor male survival. He proposes that newly immigrated males maintain small territories in order to mitigate the risks of predation. In their second breeding season, when they are more familiar with their territories, they expand their ranges and thus the number of females they encounter.
Reproductive skew in female-dominated mammalian societies Accordingly, he found that 21% of recent immigrants sired offspring in their first season, whereas 44–50% of males reproduced in their second to fourth seasons. Scramble competition, sperm competition, contest competition, and female preferences for multiple partners all appear to play roles in determining how reproduction is partitioned among grey mouse lemurs. Although reproduction is skewed towards older, heavier males who hold larger territories, females limit males’ efforts to monopolize mating, so the resulting skew among males is probably moderate. Ring-tailed lemur (Lemur catta)
On average, each troop of ring-tailed lemurs (Figure 3.2) contains 12–15 individuals comprising one or more matrilines of females, their offspring, and one or more immigrant males (Sussman 1991, Gould et al. 2003). As in many Old World monkeys, closely related female L. catta spend more time together and groom each other more often than do unrelated females (Nakamichi & Koyama 1997). Females never outrank their mothers, and kin tend to have similar ranks, but female L. catta do not inherit their mothers’ ranks, as do cercopithecine monkeys and spotted hyenas (Holekamp & Smale 1993,
Figure 3.2 Female dominance in ring-tailed lemurs (Lemur catta), with the female on the left and the male on the right. Courtesy of Peter Kappeler.
63
64
K. E. Holekamp, A. L. Engh Nakamichi & Koyama 1997, Nunn & Pereira 2000). Ring-tailed lemurs also differ from monkeys and hyenas in that they rarely form third-party alliances, and their ranks are sometimes circular and may change rapidly (Nakamichi & Koyama 1997). Female L. catta cooperate to defend their core home range against incursions by other troops (Gould et al. 2003). Large troops may be more successful at defending their home ranges, but as troop size increases, annual birth rate decreases (Jolly et al. 2002, Takahata et al. 2005). Troops that contain numerous adult females often fission when some of the females force others out ( Jolly et al. 2002, Gould et al. 2003). Breeding among ring-tails is highly seasonal, but each female is usually receptive on a different day (Periera 1991, Sauther 1991). Females give birth to 1–3 young at the start of the dry season (Parga & Lessnau 2005). Seventy-five to ninety-five percent of females reproduce each year, suggesting that there is little skew in reproductive output in this species (Sussman 1991, Sauther et al. 1999, Jolly et al. 2002, Gould et al. 2003, Parga & Lessnau 2005). Infant mortality, however, is very high (52%), so there may be great variation in the number of surviving young produced by females in a troop (Sussman 1991). High-ranking females are reported to have higher reproductive success than lower-ranking females in some populations (Taylor 1986, Takahata et al. 2005). However, female hierarchies are often unstable (Takahata et al. 2005), making it difficult to draw firm conclusions about variation in lifetime reproductive success among female L. catta, and therefore also about reproductive skew. Male L. catta disperse shortly before the breeding season (Sussman 1992). Most males emigrate for the first time at around 3–4 years of age, disperse again a few years later, then disperse every three to four years thereafter (Sussman 1992). They generally attempt to join a new troop along with one or two partners, but usually only one male successfully transfers at a time (Sussman 1992). Though males rarely join females in resource defense, they frequently cooperate to keep dispersing males from joining the troop (Nakamichi & Koyama 1997). If males are able to withstand prolonged challenges from troop members, they typically enter the troop with a low rank and remain peripheral for several months (Gould 1997). Both male and female L. catta have been observed attacking infants, but infanticide appears to be rare in this species (reviewed in Sauther et al. 1999, Ichino 2005). Males do not participate in care of young. In the week before estrus, cycling females attract a great deal of male attention. Females advertise their reproductive status by displaying their swollen vulvas and scent-marking (reviewed in Jolly 1998). Males begin to follow females several days before they are receptive, and if unreceptive females do not aggressively resist amorous males, the males constantly harass
Reproductive skew in female-dominated mammalian societies them (Sauther 1991). L. catta’s large testes and sperm plugs suggest strong sperm competition in this species (Sauther et al. 1999), and evidence from captive ring-tails suggests an advantage to being the first to mate (Pereira & Weiss 1991). In the wild, males fight viciously for access to estrous females, and some data suggest high-ranking males are the first to mate with receptive females (Jolly 1966, Sauther et al. 1999, Cavigelli & Pereira 2000). After mating, males often try to guard females against mating attempts by lower-ranking males, but most females nonetheless mate with multiple males (Sauther et al. 1999). Despite constant harassment from troop males, females frequently present to and mate with low-ranking males, transferring males, and males from other troops (Sauther 1991, Sauther et al. 1999, Parga 2002). The only males with which females consistently refuse to mate are natal males (Sauther 1991). Though they are promiscuous, the order in which females accept males suggests that they do have preferences (Sauther 1991). Genetic analysis of paternity among semi-free-ranging L. catta indicates that reproductive skew among males may be considerable (Pereira & Weiss 1991). Surprisingly, however, there is not yet a molecular analysis of paternity in wild L. catta. Behavioral data from the wild suggest that males engage largely in contest and sperm competition, but that female choice may curtail male–male competition. If females do in fact tend to mate with the highest-ranking males first, then there may be strong skew among males. In contrast, if female preferences are less predictable, or if females prefer to mate with different males in consecutive estrous periods, there may be very little male skew.
Reproductive skew among bonobos (Pan paniscus) Bonobos inhabit the forests of central Africa, living in mixed-sex groups that contain 7–8 males and 8–15 females (Kano 1996, Gerloff et al. 1999). Intensive aggression is relatively rare in this species, and social rank plays a less important role in determining priority of resource access in bonobo society than it does in societies of most other gregarious primates (de Waal 1995). Controversy exists in the literature regarding the extent of female dominance in this species, but it seems to be generally agreed that female bonobos at least have feeding priority over males (Stanford 1998, Vervaecke et al. 2000, Paoli et al. 2006). In a recent review, Kutsukake & Nunn (2006) used two different indices to estimate reproductive skew among male primates, the binomial skew index of Nonacs (2000, 2003) and the lambda index of Kokko and Lindstro¨m (1997). Although the binomial index suggested that the distribution of copulations among male bonobos differs significantly from expectations based on random
65
66
K. E. Holekamp, A. L. Engh mating, the lambda index was far lower than that found in most populations of chimpanzees (Pan troglodytes), a species in which males inevitably dominate females (Kutsukake & Nunn 2006). However, the lambda values calculated for male bonobos did not differ from those calculated for many other primate species in which males dominate females (Kutsukake & Nunn 2006). It is important to note that the analysis of Kutsukake & Nunn (2006) was based on mating frequency rather than on molecular analyses of paternity in each species. Similarly, although observations of captive bonobos suggest that highranking females may interfere with other females’ copulations (Vervaecke & van Elsacker 2000), the lack of molecular paternity analyses leaves it unclear whether this behavior results in any measurable skew among females in wild populations.
Reproductive skew among spotted hyenas (Crocuta crocuta) Spotted hyenas are large terrestrial predators occurring throughout sub-Saharan Africa. They live in complex multi-generational social groups, called clans, ranging in size from 6 to 90 individuals. Each Crocuta clan contains one to several matrilines of natal females and their offspring, as well as one to several adult immigrant males. Most habitats in which spotted hyenas occur appear to be saturated such that clan territories form a mosaic covering the entire landscape (Kruuk 1972, Boydston et al. 2001). Relatedness is high within matrilines but, on average, clan members are only very distantly related due to high levels of male-mediated gene flow among clans, and mean relatedness declines only slightly across clan borders (Van Horn et al. 2004). Clans are fission–fusion societies in which all members recognize each other, defend a common territory against neighboring clans, and rear their cubs at a single communal den (Kruuk 1972, Henschel & Skinner 1991). Individual hyenas spend much of their time alone or in small groups, particularly when foraging, but they also join together during territorial defense, interactions with competitors, and at ungulate kills. Members of multiple hyena matrilines frequently join forces to defend their kills against lions or hyenas from other clans, and by doing so risk serious injury or death (Kruuk 1972, Mills 1990, Henschel & Skinner 1991, Hofer & East 1993, Boydston et al. 2001). Thus unrelated clan-mates serve as essential allies in competition for resources. The carcasses of large ungulates represent extremely rich, rare, and ephemeral food patches that occur unpredictably in space and time, so clan members compete intensively over them. Every Crocuta clan is structured by a rigid linear dominance hierarchy, and an individual’s position in this hierarchy determines its priority of access to food (Kruuk 1972, Tilson & Hamilton 1984, Frank 1986,
Reproductive skew in female-dominated mammalian societies Mills 1990). Within a clan, all adult females are socially dominant to adult males not born in the clan (Smale et al. 1993, 1997). Before cubs reach puberty, they attain ranks in the clan’s dominance hierarchy immediately below those of their mothers (Holekamp & Smale 1993, Smale et al. 1993). Virtually all males disperse between 24 and 62 months of age, and they do this voluntarily, but females typically spend their entire lives in the natal clan (Henschel & Skinner 1987, Smale et al. 1997, Boydston et al. 2005). Excluding the occasional adult male who never disperses, adult natal males always dominate adult females ranked lower than their own mothers in the clan’s hierarchy for as long as they remain in the natal clan. However, when males disperse, they behave submissively to all new hyenas encountered outside the natal area, and this is the point at which females come to dominate males (Smale et al. 1993, 1997). By joining a new clan, each immigrant assumes the lowest rank in that clan’s dominance hierarchy (Smale et al. 1997). Males disperse alone, and relatedness among immigrant males is extremely low (Van Horn et al. 2004). Dispersing females never join existing clans elsewhere, but instead leave the natal clan alone or with some clan-mates to form an entirely new clan if an opening occurs in the mosaic of neighboring territories (Mills 1990, Holekamp et al. 1993). Because females who disperse alone appear to experience very low reproductive success, female dispersal effectively occurs only by clan fission. Although subordinate females monitor reproductive opportunities outside their natal territory (Holekamp et al. 1993), vacancies rarely occur in the local mosaic of territories, so clan fission events are similarly rare. Overall it appears that opportunities for male dispersal arise frequently, but opportunities for female dispersal seldom occur. Whereas male spotted hyenas are quite mobile, their immigration to new clans appears constrained by the severe aggression directed at dispersers by resident immigrant males (Boydston et al. 2001, Szykman et al. 2003). Males appear to compete intensively for membership in new clans, so reproductive skew among male hyenas might be expected to vary with the intensity of this competition (Reeve & Emlen 2000). Intrasexual social rank among immigrant males is highly correlated with immigrants’ tenure in the new clan, such that those arriving first dominate those arriving later (Smale et al. 1997). Social status among males is not linked with variation in body size or weaponry. Low-ranking males who have recently joined a clan rise in rank as high-ranking, longer-tenured males die or leave the clan (East & Hofer 2002, Engh et al. 2002). Thus, rather than fighting with other males for social status, males acquire status by queuing (Smale et al. 1997, East & Hofer 2001, Alberts et al. 2003). The queuing convention observed among immigrant males is relatively strict (Kokko & Johnstone 1999); males who
67
68
K. E. Holekamp, A. L. Engh manage to join a new clan rarely use aggression to enhance their status, but instead simply wait their turn to rise in rank (East & Hofer 2001). Female Crocuta are slightly larger than males, and they are heavily masculinized with respect to many aspects of their morphology and behavior (Lindeque & Skinner 1982, Glickman et al. 1992). In addition to being socially dominant to males, females are also more aggressive (Smale et al. 1993, Szykman et al. 2003), and their external genitalia are heavily virilized, such that the female urinates, copulates, and gives birth through a fully erectile pseudopenis (Matthews 1939). The female hyena’s masculinized genitalia, combined with her social dominance over breeding males, allow her an extraordinarily high degree of control over mating relative to that experienced by female mammals living in male-dominated societies (Engh et al. 2002, East et al. 2003). In most Crocuta populations studied to date, births occur throughout the year, so reproductive synchrony among females is generally very low (Kruuk 1972, Lindeque & Skinner 1982, Mills 1990, Holekamp et al. 1999). Litters born to wild hyenas usually contain one or two cubs, although triplets occur occasionally. Males do not participate in care of young, nor do females participate in the care of offspring other than their own. Approximately 8% of all hyena deaths with known causes in East African Crocuta clans are due to infanticide (Hofer & East 1995). Most successful attempts at infant-killing are perpetrated by adult females (Frank 1996), although adult males have also been observed to make failed attempts (Kruuk 1972, East & Hofer 1995). Dominant females sometimes kill cubs of subordinates, but subordinate females also kill dominant cubs (Frank 1996, East et al. 2003). Infanticide by subordinates can potentially decrease reproductive skew (Young & CluttonBrock 2006). Interestingly, reproductive skew among female spotted hyenas appears to be greater than that documented among females of male-dominated species characterized by plural breeding (Holekamp & Smale 2000). Although all female Crocuta produce young, they do so at rates that vary strongly with social rank. For example, the highest-ranking female hyenas produce an average of approximately 2.5 offspring per year, while the lowest-ranking females produce only 0.5 offspring per year, a five-fold difference (Holekamp et al. 1996). This stands in marked contrast to the far more egalitarian distribution of reproduction among females in other plural breeders. For example, in female lions the most extreme difference in annual cub production between any two pride-mates is only 0.35 cubs (Packer et al. 2001). Clans of spotted hyenas are similar in size, structure, and composition to the societies of many cercopithecine primates. Like spotted hyenas, these primates are plural breeders. Although high-ranking female primates garner
Reproductive skew in female-dominated mammalian societies greater reproduction than do subordinates, the rank-related variation in female reproductive success in all these primate species is small compared to that in spotted hyenas (Holekamp & Smale 2000). In addition to producing offspring at higher rates, high-ranking female hyenas bear daughters that generally grow faster, are more likely to survive to adulthood, start breeding earlier, and enjoy longer reproductive lifespans than do daughters of mid- or low-ranking females (Holekamp et al. 1996, Hofer & East 1996). Direct aggressive harassment does not appear to function importantly in the reproductive suppression of subordinate females in this species (Holekamp et al. 1996). Furthermore, the occurrence of infanticide is also often inconsistent with suppression of reproduction in subordinate hyenas. Instead, the key determinant of reproductive success among female Crocuta is rank-related priority of access to food resources. Those female hyenas enjoying the greatest reproductive success are those most effectively able to use aggression to displace conspecifics from carcasses during competitive feeding. Reproductive skew among male Crocuta
Crocuta’s mating system is polygynous, but matings are not monopolized by high-ranking males (Engh et al. 2002, East et al. 2003). Dispersal status and length of residence as immigrants in new clans after dispersal are the strongest determinants of reproductive success among male hyenas. Adult natal male hyenas are socially dominant to immigrant males, and often show strong sexual interest in clan females (Holekamp & Smale 1998), yet they sire only 3% of cubs in their natal clans, whereas immigrants sire 97% (Engh et al. 2002), indicating that females prefer to mate with immigrants. Among resident immigrant males, social rank is correlated with male reproductive success, but tenure predicts this far better than does male rank. Immigrants do not typically begin to sire offspring until they have resided in their new clan for one or two years, during which time they occupy low rank positions in the male queue (Engh et al. 2002, East et al. 2003). At least 40% of female Crocuta mate with multiple males during any given estrous period, and 25–40% of twin litters are multiply sired (Engh et al. 2002, East et al. 2003). Work by Cant & Reeve (2002) underscores the importance of evaluating whether males or females control paternity in multi-male vertebrate societies before attempting to examine predictions of skew models. The authors suggest experimental manipulations in which alpha males are removed briefly when females are fertile to assess the extent to which females control the distribution of paternity. If females control paternity, removal of alpha males should have little consistent effects on their share of paternity. By contrast, if alpha males normally control paternity, their removal should lead to a reduction in
69
70
K. E. Holekamp, A. L. Engh their share of paternity. Crocuta offer a natural experiment of this sort. Patterns of paternity in spotted hyenas do not conform, as they do in many other gregarious mammals, to a dominance-based “priority of access” model (Altmann et al. 1996) predicting that the number of offspring sired by males should directly reflect both male social rank and the number of females simultaneously in estrus. Hyena immigrants in the top half of the male hierarchy father more cubs than those in the bottom half, but males of all ranks sire offspring, and the alpha male generally sires fewer cubs than do males in lower rank positions (Engh et al. 2002); again, this suggests an important role for female choice. Overall reproductive skew among male Crocuta is considerably less than that documented in many other group-living mammals in which males dominate females (ungulates: Clutton-Brock et al. 1988, Pemberton et al. 1992, Hirotani, 1994; rodents: Sherman et al. 1991; primates: Smith 1993, de Ruiter et al. 1994, Altmann et al. 1996, Bercovitch & Nurnberg 1997, Setchell et al. 2005; carnivores: Keane et al. 1994, Creel et al. 1997, Girman et al. 1997). In contrast to what occurs in most polygynous mammals, male–male contest competition appears to have little influence over male reproductive success in the spotted hyena. Instead, female choice of mates appears to be the key determinant of patterns of paternity. Females clearly prefer immigrants over adult natal males, and they frequently choose immigrant males other than the alpha male as mates (Engh et al. 2002). High-ranking male hyenas cannot monopolize reproduction if females prefer not to mate with them. Female control over mating has thus reduced selection for male fighting ability, and has led to low levels of male–male combat and the evolution of a male social queue (East et al. 2003).
How well do female-dominated mammals satisfy assumptions of skew models? In accordance with transactional models, dominant lemurs benefit from the presence of subordinates in the group; subordinate lemurs are often needed by dominants for cooperative defense of territories. Ecological constraints clearly limit dispersal in lemurs, group stability appears to be important to all group members, and subordinates in some lemur species are constrained by the threat of expulsion from the group by dominants. Similarly, mole-rats (see Chapter 13) also appear to satisfy the assumptions of transactional skew models (e.g. Vehrencamp 1983a, 1983b, Reeve & Emlen 2000) quite well. Spotted hyenas also appear to satisfy some of the assumptions underlying transactional models. Specifically, dominant spotted hyenas need subordinate allies, without which high-ranking animals risk loss of individual ungulate
Reproductive skew in female-dominated mammalian societies carcasses and the group territory. Furthermore, group stability is important to both high- and low-ranking hyenas; rates of female–female aggression and wounding are greatly elevated during periods of social upheaval (Van Meter & Holekamp, unpublished data). Finally, ecological constraints limit dispersal in hyenas, particularly among females. However, lemurs and spotted hyenas fail to satisfy various other assumptions underlying transactional models. Most importantly, transactional models assume that dominant individuals have complete control over allocation of reproduction within the group, a condition that is clearly not met by either lemurs or hyenas. Furthermore, transactional models assume that dominants can evict subordinates when this is in the best interests of the dominants, but this is not true in hyenas. The ability of dominant hyenas to control subordinate reproduction is severely constrained by the fission–fusion sociality characteristic of this species, and eviction potential is low because subordinates are very well armed and often larger than dominants. In accordance with compromise skew models (e.g. Clutton-Brock 1998, Reeve et al. 1998), both dominants and subordinates reproduce in the societies of lemurs and hyenas because neither has complete control over reproduction. Moreover, individuals clearly struggle over the distribution of reproduction rather than conceding each other a share of it to their mutual benefit. However, contrary to assumptions underlying compromise models, ecological constraints are extremely important in limiting dispersal in both lemurs and hyenas, and group stability does enhance reproductive success of both dominant and subordinate group-mates. Furthermore, in violation of another assumption of compromise models, dominant hyenas may be constrained by the threat of subordinate departure, as clans regularly fission whenever vacant habitat becomes available (e.g. Holekamp et al. 1993). Overall it appears that the female-dominated mammals considered in this chapter satisfy the assumptions of transactional skew models slightly better than those of compromise models. That is, ecological constraints and benefits of group membership appear to be important determinants of skew in all these species, as they are in many male-dominated mammals. However, lemurs and hyenas fail to meet some of the critical assumptions of both transactional and compromise models, so we would expect predictions derived from those models to be of similarly limited value in regard to these species. We suggest that, although skew models are useful in predicting partitioning of reproduction in mole-rats (see Chapter 13), perhaps current models invoke too many simplistic assumptions to account for patterns of reproduction and aggression in societies like those of lemurs and spotted hyenas. For example, most skew models ignore individual variation within dominant or subordinate
71
72
K. E. Holekamp, A. L. Engh classes, and incorrectly assume that all members of each class are identical (Smith et al. 2007). One possible exception to this generalization is the recruiter–joiner model, in which Hamilton (2000) uses optimal skew theory to predict group size and how resources will be divided within groups of social foragers. Assuming that reproductive output by female spotted hyenas is directly related to their food intake, an assumption strongly supported by empirical field data (Holekamp et al. 1996, Hofer & East 2003), the “limited control” transactional model presented by Hamilton (2000) appears to predict partitioning of reproduction among female spotted hyenas better than do other skew models.
Conclusion Although reproductive skew in both sexes has been well documented in mole-rats and spotted hyenas, the paucity of data from other species with female dominance (e.g. lemurs and bonobos) does not presently allow us to arrive at any firm general conclusions about skew in female-dominated mammals. Variation in reproductive skew among these species is enormous: it is extremely high among female mole-rats, moderate among female spotted hyenas, and apparently quite low in female lemurs. Thus, overall variation in skew among these species appears to be no different than that documented in male-dominated mammals. Dominant female mole-rats can control reproduction in subordinates (Faulkes et al. 1991, Lacey & Sherman 1997) far more effectively than can dominant lemurs or hyenas, suggesting that female molerats satisfy the assumptions of original transactional skew models (e.g. Vehrencamp 1983a, 1983b) much better than do these other animals. With the exception of naked mole-rats, reproductive skew among males in female-dominated mammals is greater than that among females, even in socially monogamous species, and is similar to that found in male-dominated species. Females in female-dominated societies must be somewhat freer from constraints upon reproduction than their male-dominated counterparts, because they have higher priority of access to food and other resources critical for reproduction than males. Therefore, in female-dominated species, we might expect the average productivity of females to be higher than in male-dominated species. However, since reproductive skew largely reflects intrasexual competition, there is no reason to expect a priori that female skew should necessarily differ between female-dominated and male-dominated societies. Indeed, in some cases the question of which sex is socially dominant appears to be moot in regard to skew. For example, female skew among male-dominated Damaraland molerats (Cryptomys damarensis) does not differ from that in female-dominated naked
Reproductive skew in female-dominated mammalian societies and common mole-rats (Bennett & Faulkes 2000). Similarly, in the one lemur species (Eulemur fulvus) that is not characterized by female dominance, older, heavier females have higher reproductive success than younger, lighter females (Overdorff et al. 1999), a pattern of skew virtually identical to that in Propithecus verreauxi, in which females dominate males. Although this implies that skew is independent of female dominance patterns, many gaps remain in our knowledge about the species discussed in this review. Furthermore, the analysis presented here may be confounded by our attempts to compare reproductive skew among species with different types of social organization. For example, we have attempted here to evaluate patterns of reproductive skew in gregarious, monogamous, and solitary lemurs. However, although we see no reason why skew cannot be calculated within a population of solitary or monogamous animals as well as within a particular social group of gregarious animals, “reproductive skew” traditionally refers only to the latter. In any case, until more studies examine variance in lifetime reproductive success, it will be difficult to come to any firm conclusions about the degree or determinants of skew among female lemurs in particular. More generally, conclusions about the extent of reproductive skew among female mammals, regardless of which sex is socially dominant, should ideally be based on data documenting offspring survivorship rather than simply the number of females breeding within a particular social group, because early mortality among offspring can profoundly affect long-term reproductive success. There is a suggestion in the work reviewed in this chapter that skew among females in some female-dominated mammals may be slightly greater than that in male-dominated species with comparable group compositions. The degree of skew among females is likely to be affected by the extent to which priority of access to resources influences reproductive output. In hyenas and mole-rats this appears to be considerably greater than in lemurs, and the reproductive skew among female lemurs appears to be correspondingly low. If in fact reproductive skew among females tends to be greater in female-dominated than in male-dominated species, then males in these species should be selective with respect to their own choice of mates. Indeed, we already know that male spotted hyenas strongly prefer to mate with high-ranking females (Szykman et al. 2001). Female dominance, and the associated high level of female control over reproduction, tends to decrease skew among males. Although few studies have used molecular tools to assess partitioning of reproduction among male lemurs, virtually all female lemurs mate with multiple males at each estrus, and they often choose to mate with low-ranking resident males or males from other groups. Most male lemurs invest little in reproduction and therefore
73
74
K. E. Holekamp, A. L. Engh might be expected to exhibit a higher degree of skew than females. However, compared to male-dominated species, reproductive success among male lemurs is probably influenced relatively weakly by male–male competition and relatively strongly by female preferences. As a result, we might expect that skew among male lemurs, on average, should be less than that documented in male-dominated mammals living in groups of comparable size. On the other hand, if females prefer males who excel in intraspecific competition, this prediction would be nullified. We find that skew among male lemurs spans a wide spectrum. Male P. verreauxi show strong skew, whereas skew among male M. coquereli is essentially absent. Female dominance is expressed at a low level in P. verreauxi, but is much more strongly pronounced in M. coquereli. In several lemur species, females counteract males’ attempts to monopolize reproduction, so it is likely, on the whole, that there is less skew among male lemurs than among maledominated mammals. However, the significance of this difference is unclear, since even in male-dominated species, females often successfully resist males’ attempts to control mating (e.g. Eberhard 1996). Among spotted hyenas, the female’s masculinized genitalia may increase female control over mating to its extreme limit, and in this species we see a much lower degree of reproductive skew among males relative to that found in other polygynous mammals (e.g., compare Engh et al. 2002 with Altmann et al. 1996). The most dominant males in this species often achieve very little reproductive success. In neither lemurs nor hyenas do dominant males determine whether females are permitted some promiscuity, as females in both groups clearly make their own decisions regarding whether or not to mate with more than one male. Competition among male mammals for access to females is often intense, and can result in high variance in reproductive success among males, especially in polygynous species. In most mammals, sexual selection favors traits in males that are attractive to females and that enhance male competitive ability. The range of reproductive skew among males in female-dominated species is generally less than it is in male-dominated mammals. Among female dominated species, we seldom find cases in which one male monopolizes all reproduction. Even in naked mole-rats, multiple males usually get to mate, and females bear litters sired by multiple males (Bennett & Faulkes 2000). Given the apparent strength of the effects of female choice in female-dominated mammalian species, it appears that males have been obliged to develop strategies to maximize their reproductive success that supplement or replace male–male combat. We see much more reliance in female-dominated species than in the majority of mammals on alternative modes of sexually selected interactions including endurance rivalry (e.g. queuing), sperm competition, use of copulatory plugs,
Reproductive skew in female-dominated mammalian societies and mate-guarding. Coercive sex is not possible in female-dominated species, and female dominance makes it difficult or impossible for males to monopolize females. Each female determines whether or not she will be monopolized by a single male and, if so, which male this will be. Dominant females can tolerate or refuse male mating attempts according to their own reproductive interests, and this unusual degree of female control appears to reduce the strength of the relationship between social status and reproductive success in males. Acknowledgments This chapter was produced with partial support of funds from the US National Science Foundation in grants IBN0343381 and IOB0618022. References Alberts, S. C., Watts, H. E., & Altmann, J. (2003). Queuing and queue-jumping: longterm patterns of reproductive skew among male savannah baboons, Papio cynocephalus. Animal Behaviour, 65, 821–840. Altmann, J., Alberts, S. C., Haines, S. A., et al. (1996). Behavior predicts genetic structure in a wild primate group. Proceedings of the National Academy of Sciences of the USA, 93, 5797–5801. Andres, M., Gachot-Neveu, H., & Perret, M. (2001). Genetic determination of paternity in captive grey mouse lemurs: pre-copulatory sexual competition rather than sperm competition in a nocturnal prosimian? Behaviour, 138, 1047–1063. Bennett, N. C. & Faulkes, C. G. (2000). African Mole-Rats: Ecology and Eusociality. Cambridge: Cambridge University Press. Bercovitch, F. B. & Nurnberg, P. (1997). Genetic determination of paternity and variation in male reproductive success in two populations of rhesus macaques. Electrophoresis, 18, 1701–1705. Birkhead, T. & Moller, A. (1993). Female control of paternity. Trends in Ecology and Evolution, 8, 100–104. Boydston, E. E., Morelli, T. L., & Holekamp, K. E. (2001). Sex differences in territorial behavior exhibited by the spotted hyena (Crocuta crocuta). Ethology, 107, 369–385. Boydston, E. E., Kapheim, K. M., Van Horn, R. C., Smale, L., & Holekamp, K. E. (2005). Sexually dimorphic patterns of space use throughout ontogeny in the spotted hyaena (Crocuta crocuta). Journal of Zoology, London, 267, 271–281. Brockman, D. K. (1999). Reproductive behavior of female Propithecus verreauxi at Beza Mahafaly, Madagascar. International Journal of Primatology, 20, 375–398. Brockman, D. K. & Whitten, P. L. (1996). Reproduction in free-ranging Propithecus verreauxi: estrus and the relationship between multiple partner matings and fertilization. American Journal of Physical Anthropology, 100, 57–69.
75
76
K. E. Holekamp, A. L. Engh Brockman, D. K., Whitten, P. L., Richard, A. F., & Schneider, A. (1998). Reproduction in free-ranging male Propithecus verreauxi: the hormonal correlates of mating and aggression. American Journal of Physical Anthropology, 105, 137–151. Buesching, C. D., Heistermann, M., Hodges, J. K., & Zimmermann, E. (1998). Multimodal oestrus advertisement in a small nocturnal prosimian, Microcebus murinus. Folia Primatologica, 69, 295–308. Cant, M. A. (2000). Social control of reproduction in banded mongooses. Animal Behaviour, 59, 147–158. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Cavigelli, S. A. & Pereira, M. E. (2000). Mating season aggression and fecal testosterone levels in male ring-tailed lemurs (Lemur catta). Hormones and Behavior, 37, 246–255. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H., Albon, S. D. & Guinness, F. E. (1988). Reproductive success in male and female red deer. In T. H. Clutton-Brock, ed., Reproductive Success. Chicago, IL: University of Chicago Press, pp. 325–343. Creel, S. (1997). Rank and reproduction in cooperatively breeding African wild dogs: behavioral and endocrine correlates. Behavioral Ecology, 8, 298–306. Darwin, C. (1871). The Descent of Man and Selection in Relation to Sex. London: John Murray. de Ruiter, J. R., van Hooff, J. A. R. A. M., & Scheffrahn, W. (1994). Social and genetic aspects of paternity in wild long-tailed macaques (Macaca fascicularis). Behaviour, 129, 203–224. de Waal, F. B. M. (1995). Bonobo sex and society. Scientific American, 272, 82–88. East, M. L. & Hofer, H. (2001). Male spotted hyenas (Crocuta crocuta) queue for status in social groups dominated by females. Behavioral Ecology, 12, 558–568. East, M. L. & Hofer, H. (2002). Conflict and cooperation in a female-dominated society: a reassessment of the “hyperaggressive” image of spotted hyenas. Advances in the Study of Behavior, 31, 1–30. East, M. L., Burke, T., Wilhelm, K., Greig, C., & Hofer, H. (2003). Sexual conflicts in spotted hyenas: male and female mating tactics and their reproductive outcome with respect to age, social status and tenure. Proceedings of the Royal Society of London B, 270, 1247–1254. Eberhard, W. G. (1996). Female Control: Sexual Selection by Cryptic Female Choice. Princeton, NJ: Princeton University Press. Eberle, M. & Kappeler, P. M. (2002). Mouse lemurs in space and time: a test of the sociobiological model. Behavioral Ecology and Sociobiology, 51, 131–139. Eberle, M. & Kappeler, P. M. (2003). Cooperative breeding in grey mouse lemurs (Microcebus murinus). Folia Primatologica, 74, 367–368. Eberle, M. & Kappeler, P. M. (2004a). Sex in the dark: determinants and consequences of mixed male mating tactics in Microcebus murinus, a small solitary nocturnal primate. Behavioral Ecology and Sociobiology, 57, 77–90.
Reproductive skew in female-dominated mammalian societies Eberle, M. & Kappeler, P. M. (2004b). Selected polyandry: female choice and intersexual conflict in a small nocturnal solitary primate (Microcebus murinus). Behavioral Ecology and Sociobiology, 57, 91–100. Engh, A. L., Funk, S. M., Van Horn, R. C., et al. (2002). Reproductive skew among males in a female-dominated society. Behavioral Ecology, 13, 193–200. Espmark, Y. (1964). Studies in dominance–subordination relationship in a group of semi-domestic reindeer (Rangifer tarandus). Animal Behaviour, 12, 420–426. Faulkes, C. G., Abbott, D. H., Liddell, C. E., George, L. M., & Jarvis, J. U. M. (1991). Hormonal and behavioral aspects of reproductive suppression in female naked mole rats. In P. W. Sherman, J. U. M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole Rat. Princeton, NJ: Princeton University Press, pp. 426–445. Fietz, J. (1999). Monogamy as a rule rather than exception in nocturnal lemurs: the case of the fat-tailed dwarf lemur, Cheirogaleus medius. Ethology, 105, 259–272. Fietz, J., Zischler, H., Schwiegk, C., et al. (2000). High rates of extra-pair young in the pair-living fat-tailed dwarf lemur, Cheirogaleus medius. Behavioral Ecology and Sociobiology, 49, 8–17. Frank, L. G. (1986). Social organization of the spotted hyaena (Crocuta crocuta). II. Dominance and reproduction. Animal Behaviour, 35, 1510–1527. Frank, L. G. (1996). Female masculinization in the spotted hyena: endocrinology, behavioral ecology and evolution. In J. L. Gittleman, ed., Carnivore Behavior, Ecology, and Evolution. Ithaca, NY: Cornell University Press, vol. 2, pp. 78–131. Fredsted, T., Pertoldi, C., Olesen, J. M., Eberle, M. & Kappeler, P. M. (2004). Microgeographic heterogeneity in spatial distribution and mtDNA variability of gray mouse lemurs (Microcebus murinus, Primates: Cheirogaleidae). Behavioral Ecology and Sociobiology, 56, 393–403. Gerloff, U., Hartung, B., Fruth, B., Hohmann, G., & Tautz, D. (1999). Intracommunity relationships, dispersal patterns and paternity success in a wild-living community of bonobos (Pan paniscus) determined from DNA analysis of faecal samples. Proceedings of the Royal Society of London B, 266, 1189–1195. Girman, D. J., Mills, M. G. L., Geffen, E., & Wayne, R. K. (1997). A molecular genetic analysis of social structure, dispersal, and interpack relationships of the African wild dog (Lycaon pictus). Behavioral Ecology and Sociobiology, 40, 187–198. Glickman, S. E., Frank, L. G., Pavgi, S., & Licht, P. (1992). Hormonal correlates of “masculinization” in female spotted hyenas (Crocuta crocuta). I. Infancy to sexual maturity. Journal of Reproduction and Fertility, 95, 451–462. Goodman, S. M. & Benstead, J. P. (2005). Updated estimates of biotic diversity and endemism for Madagascar. Oryx, 39, 73–77. Gould, L. (1997). Intermale affiliative behavior in ringtailed lemurs (Lemur catta) at the Beza-Mahafaly Reserve, Madagascar. Primates, 38, 15–30. Gould, L., Sussman, R. W., & Sauther, M. L. (2003). Demographic and life-history patterns in a population of ring-tailed lemurs (Lemur catta) at Beza Mahafaly Reserve, Madagascar: a 15-year perspective. American Journal of Physical Anthropology, 120, 182–194.
77
78
K. E. Holekamp, A. L. Engh Hager, R. & Welker, C. (2001). Female dominance in African lorises (Otolemur garnettii). Folia Primatologica, 72, 48–50. Hamilton, I. M. (2000). Recruiters and joiners: using optimal skew theory to predict group size and the division of resources within groups of social foragers. American Naturalist, 155, 684–695. Henschel, J. R. & Skinner, J. D. (1987). Social relationships and dispersal patterns in a clan of spotted hyaenas, Crocuta crocuta, in the Kruger National Park. South African Journal of Zoology, 22, 18–23. Henschel, J. R. & Skinner, J. D. (1991). Territorial behavior by a clan of spotted hyaenas, Crocuta crocuta. Ethology, 88, 223–235. Hirotani, A. (1994). Dominance rank, copulatory behavior and estimated reproductive success in male reindeer. Animal Behaviour, 48, 929–936. Hofer, H. & East, M. L. (1993). The commuting system of Serengeti spotted hyaenas: how a predator copes with migratory prey. II. Intrusion pressure and commuter’s space use. Animal Behaviour, 46, 559–574. Hofer, H. & East, M. L. (1995). Population dynamics, population size, and the commuting system of Serengeti spotted hyaenas. In A. R. E. Sinclair and P. Arcese, eds., Serengeti II. Chicago, IL: University of Chicago Press, pp. 332–363. Hofer, H. & East, M. L. (1996). The components of parental care and their fitness consequences: a life-history perspective. Verhandlungen der Deutschen Zoologischen Gesellschaft, 89, 149–164. Hofer, H. & East, M. L. (2003). Behavioral processes and costs of co-existence in female spotted hyenas: a life-history perspective. Evolutionary Ecology, 17, 315–331. Holekamp, K. E. & Smale, L. (1993). Ontogeny of dominance in free-living spotted hyaenas: juvenile rank relations with other immature individuals. Animal Behaviour, 46, 451–466. Holekamp, K. E. & Smale, L. (1998). Dispersal status influences hormones and behavior in the male spotted hyena. Hormones and Behavior, 33, 205–216. Holekamp, K. E. & Smale, L. (2000). Feisty females and meek males: reproductive strategies in the spotted hyena. In K. Wallen & J. Schneider, eds., Reproduction in Context. Cambridge. MA: MIT Press, pp. 257–285. Holekamp, K. E., Ogutu, J. O., Dublin, H. T., Frank, L. G., & Smale, L. (1993). Fission of a spotted hyaena clan: consequences of prolonged female absenteeism and causes of female emigration. Ethology, 93, 285–299. Holekamp, K. E., Smale, L., & Szykman, M. (1996). Rank and reproduction in female spotted hyaena. Journal of Reproduction and Fertility, 108, 229–237. Holekamp, K. E., Smale, L., & Cooper, S. (1997). Hunting rates and hunting success in the spotted hyaena. Journal of Zoology, London, 242, 1–15. Holekamp, K. E., Szykman, M., Boydston, E. E., & Smale, L. (1999). Association of seasonal reproductive patterns with changing food availability in an equatorial carnivore. Journal of Reproduction and Fertility, 116, 87–93.
Reproductive skew in female-dominated mammalian societies Ichino, S. (2005). Attacks on a wild infant ring-tailed lemur (Lemur catta) by immigrant males at Berenty, Madagascar: interpreting infanticide by males. American Journal of Primatology, 67, 267–272. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Jolly, A. (1966). Lemur Behavior. Chicago, IL: University of Chicago Press. Jolly, A. (1998). Pair-bonding, female aggression and the evolution of lemur societies. Folia Primatologica, S1, 1–13. Jolly, A., Dobson, A., Rasamimanana, H. M., et al. (2002). Demography of Lemur catta at Berenty Reserve, Madagascar: effects of troop size, habitat and rainfall. International Journal of Primatology, 23, 327–353. Kano, T. (1996) Male rank order and copulation rate in a unit-group of bonobos at Wamba, Zaire. In W. C. McGrew, L. F. Marchant, & T. Nishida, eds., Great Ape Societies. Cambridge: Cambridge University Press, pp. 135–145. Kappeler, P. M. (1997a). Intrasexual selection in Mirza coquereli: evidence for scramble competition polygyny in a solitary primate. Behavioral Ecology and Sociobiology, 45, 115–127. Kappeler, P. M. (1997b). Intrasexual selection and testis size in strepsirhine primates. Behavioral Ecology, 8, 10–19. Kappeler, P. M., Wimmer, B., Zinner, D., & Tautz, D. (2002). The hidden matrilineal structure of a solitary lemur: implications for primate social evolution. Proceedings of the Royal Society of London B, 269, 1755–1763. Keane, B., Waser, P. M., Creel S. R., et al. (1994). Subordinate reproduction in dwarf mongooses. Animal Behaviour, 47, 65–75. Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 266, 571–578. Kokko, H. & Lindstro¨m, J. (1997). Measuring the mating skew. American Naturalist 149, 794–799. Kraus, C., Heistermann, M., & Kappeler, P. M. (1999). Physiological suppression of sexual function of subordinate males: a subtle form of intrasexual competition among male sifakas (Propithecus verreauxi)? Physiology and Behavior, 65, 855–861. Kruuk, H. (1972). The Spotted Hyena: a Study of Predation and Social Behavior. Chicago, IL: University of Chicago Press. Kutsukake, N. & Nunn, C. L. (2006). Comparative tests of reproductive skew in male primates: the role of demographic factors and incomplete control. Behavioral Ecology and Sociobiology, 60, 695–706. Lacey, E. A. & Sherman, P. W. (1997). Cooperative breeding in naked mole rats: implications for vertebrate and invertebrate sociality. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 267–301. Lawler, R. R., Richard, A. F., & Riley, M. A. (2003). Genetic population structure of the white sifaka (Propithecus verreauxi verreauxi) at Beza Mahafaly Special Reserve, southwest Madagascar (1992–2001). Molecular Ecology, 12, 2307–2317.
79
80
K. E. Holekamp, A. L. Engh Lawler, R. R., Richard, A. F., & Riley, M. A. (2005). Intrasexual selection in Verreaux’s sifaka (Propithecus verreauxi verreauxi). Journal of Human Evolution, 48, 259–277. Lewis, R. J. & Kappeler, P. M. (2005). Seasonality, body condition, and timing of reproduction in Propithecus verreauxi verreauxi in the Kirindy Forest. American Journal of Primatology, 67, 347–364. Lindeque, M. & Skinner, J. D. (1982). Fetal androgens and sexual mimicry in spotted hyaenas (Crocuta crocuta). Journal of Reproduction and Fertility, 65, 405–410. Martin, R. D. (1972). A preliminary field study of the lesser mouse lemur (Microcebus murinus J. F. Miller 1777), Tierpsychologie Zeitschrift fu¨r, 9, 43–89. Matthews, L. H. (1939). Reproduction in the spotted hyaena, Crocuta crocuta (Erxleben). Philosophical Transactions of the Royal Society B, 230, 1–78. Mills, M. G. L. (1990). Kalahari Hyaenas: The Behavioural Ecology of Two Species. London: Unwin Hyman. Mu¨ller, A. E. & Thalmann, U. (2002). Biology of the fat-tailed dwarf lemur (Cheirogaleus medius E. Geoffroy 1812): new results from the field. Evolutionary Anthropology, S1, 79–82. Mutschler, T. (1999). The Alaotran gentle lemur (Hapalemur griseus alaotrensis): a study in behavioural ecology. Unpublished Ph.D. thesis, University of Zurich, Switzerland. Mutschler, T. (2002). Alaotran gentle lemur: some aspects of its behavioral ecology. Evolutionary Anthropology, S1, 101–104. Mutschler, T., Nievergelt, C. M. & Feistner, A. T. C. (2000). Social organization of the Alaotran gentle lemur (Hapalemur griseus alaotrensis). American Journal of Primatology, 50, 9–24. Nakamichi, M. & Koyama, N. (1997). Social relationships among ring-tailed lemurs (Lemur catta) in two free-ranging troops at Berenty Reserve, Madagascar. International Journal of Primatology, 18, 73–93. Nievergelt, C., Mutschler, T., & Feistner, A. T. C. (1998). Group encounters and territoriality in wild Alaotran gentle lemurs (Hapalemur griseus alaotrensis). American Journal of Primatology, 46, 251–258. Nievergelt, C. M., Mutschler, T., Feistner, A. T. C., & Woodruff, D. S. (2002). Social system of the Alaotran gentle lemur (Hapalemur griseus alaotrensis): genetic characterization of group composition and mating system. American Journal of Primatology, 57, 157–176. Nonacs, P. (2000). Measuring and using skew in the study of social behavior and evolution. American Naturalist, 156, 577–589. Nonacs, P. (2003). Measuring the reliability of skew indices: is there one best index? Animal Behaviour, 65, 615–627. Nunn, C. L. & Pereira, M. E. (2000). Group histories and offspring sex ratios in ringtailed lemurs (Lemur catta). Behavioral Ecology and Sociobiology, 48, 18–28. Overdorff, D. J., Merenlender, A. M., Talata, P., Telo, A., & Forward, Z. A. (1999). Life history of Eulemur fulvus rufus from 1988–1998 in southeastern Madagascar. American Journal of Physical Anthropology, 108, 295–310.
Reproductive skew in female-dominated mammalian societies Packer, C., Pusey, A. E., & Eberly, L. E. (2001). Egalitarianism in female African lions. Science, 293, 690–693. Paoli, T., Palagi, E., & Borgognini Tarli, S. M. (2006) Reevaluation of dominance hierarchies in bonobos (Pan paniscus). American Journal of Physical Anthropology, 130, 116–122. Parga, J. A. (2002). Male dominance rank reversal during the breeding season in ringtailed lemurs (Lemur catta): changes resulting from female mate choice. American Journal of Physical Anthropology, S34, 123. Parga, J. A. & Lessnau, R. G. (2005). Female age-specific reproductive rates, birth seasonality, and infant mortality of ring-tailed lemurs on St. Catherine Island: 17-year reproductive history of a free-ranging colony. Zoo Biology, 24, 295–309. Pemberton, J. M., Albon, S. D., Guinness, F. E., Clutton-Brock, T. H., & Dover, G. A. (1992). Behavioral estimates of male mating success tested by DNA fingerprinting in a polygynous mammal. Behavioral Ecology, 3, 66–75. Pereira, M. E. (1991). Asynchrony within estrous synchrony among ringtailed lemurs (Primates: Lemuridae). Physiology and Behavior, 49, 47–52. Pereira, M. E. & Weiss, M. L. (1991). Female mate choice, male migration, and the threat of infanticide in ringtailed lemurs. Behavioral Ecology and Sociobiology, 28, 141–152. Perret, M. (1992). Environmental and social determinants of sexual function in the male lesser mouse lemur (Microcebus murinus). Folia Primatologica, 59, 1–25. Radespiel, U. (2000). Sociality in the gray mouse lemur (Microcebus murinus) in northwestern Madagascar. American Journal of Primatology, 51, 21–40. Radespiel, U. & Zimmermann, E. (2001). Female dominance in captive gray mouse lemurs (Microcebus murinus). American Journal of Primatology, 54, 181–192. Radespiel, U., Sarikaya, Z., Zimmermann, E., & Bruford, M. W. (2001). Sociogenetic structure in a free-living nocturnal primate population: sex-specific differences in the grey mouse lemur (Microcebus murinus). Behavioral Ecology and Sociobiology, 50, 493–502. Radespiel, U., Dal Secco, V., Drogemuller, C., et al. (2002). Sexual selection, multiple mating and paternity in grey mouse lemurs, Microcebus murinus. Animal Behaviour, 63, 259–268. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model . Behavioral Ecology, 11, 640–647. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Richard, A. F. (1992). Aggressive competition between males, female-controlled polygyny and sexual monomorphism in a Malagasy primate, Propithecus verreauxi. Journal of Human Evolution, 22, 393–406. Richard, A. F. & Dewar, R. E. (1991). Lemur ecology. Annual Review of Ecology and Systematics, 22, 145–175.
81
82
K. E. Holekamp, A. L. Engh Richard, A. F., Dewar, R. E., Schwartz, M., & Ratsirarson, J. (2000). Mass change, environmental variability and female fertility in wild Propithecus verreauxi. Journal of Human Evolution, 39, 381–391. Richard, A. F., Dewar, R. E., & Ratsirarson, J. (2002). Life in the slow lane? Demography and life histories of male and female sifaka (Propithecus verreauxi verreauxi). Journal of Zoology, London, 256, 421–436. Sauther, M. L. (1991). Reproductive behavior of free-ranging Lemur catta at Beza Mahafaly Special Reserve, Madagascar. American Journal of Physical Anthropology, 84, 463–477. Sauther, M. L., Sussman, R. W., & Gould, L. (1999). The socioecology of the ringtailed lemur: thirty-five years of research. Evolutionary Anthropology, 8, 120–132. Schmelting, B. (2001). Reproductive tactics in male grey mouse lemurs (Microcebus murinus, J. F. Miller 1777) in Northwestern Madagascar. Unpublished Ph.D. thesis, School of Veterinary Medicine, Hannover, Germany. Schmid, J. & Kappeler, P. M. (1998). Fluctuating sexual dimorphism and differential hibernation by sex in a primate, the gray mouse lemur (Microcebus murinus). Behavioral Ecology and Sociobiology, 43, 125–132. Setchell, J. M., Charpentier, M., & Wickings, E. J. (2005). Mate guarding and paternity in mandrills: factors influencing alpha male monopoly. Animal Behaviour, 70, 1105–1120. Sherman, P. W., Jarvis, J. U. M., & Alexander, R. D. (1991). The Biology of the Naked MoleRat. Princeton, NJ: Princeton University Press. Short, R. V. & Balaban, E., eds. (1994). The Differences Between the Sexes. Cambridge: Cambridge University Press. Smale, L., Frank, L. G., & Holekamp, K. E. (1993). Ontogeny of dominance in free-living spotted hyaenas: juvenile rank relations with adults. Animal Behaviour, 46, 467–477. Smale, L., Nunes, S., & Holekamp, K. E. (1997). Sexually dimorphic dispersal in mammals: patterns, causes, and consequences. Advances in the Study of Behavior, 26, 181–250. Smith, D. G. (1993). A 15-year study of the association between dominance rank and reproductive success of male rhesus macaques. Primates, 34, 471–480. Smith, J. E., Memenis, S. K., & Holekamp, K. E. (2007). Rank-related partner choice in the fission–fusion society of the spotted hyena (Crocuta crocuta). Behavioral Ecology and Sociobiology, 61, 753–765. Stanford, C. B. (1998). The social behavior of chimpanzees and bonobos. Current Anthropology, 39, 399–420. Stanger, K. (1995). Vocalizations of some cheirogaleid prosimians evaluated in a phylogenetic context. In L. Alterman, G. Doyle, & M. Izard, eds., Creatures of the Dark. New York, NY: Plenum, pp. 353–376. Sussman, R. W. (1991). Demography and social organization of free-ranging Lemur catta in the Beza Mahafaly Reserve, Madagascar. American Journal of Physical Anthropology, 84, 43–58. Sussman, R. W. (1992). Male life history and intergroup mobility among ringtailed lemurs (Lemur catta). International Journal of Primatology, 13, 395–413.
Reproductive skew in female-dominated mammalian societies Szykman, M., Engh, A. L., Van Horn, R. C., et al. (2001). Association patterns between male and female spotted hyenas reflect male mate choice. Behavioral Ecology and Sociobiology, 50, 231–238. Szykman, M., Engh, A. L., Van Horn, R. C., et al. (2003). Rare male aggression directed toward females in a female-dominated society: baiting behavior in the spotted hyena. Aggressive Behavior, 29, 457–474. Takahata, Y., Koyama, N., Ichino, S., & Miyamoto, N. (2005). Inter- and within-troop competition of female ring-tailed lemurs: a preliminary field report. African Study Monographs, 26, 1–14. Taylor, L. (1986). Kinship, dominance, and social organization in a semi-free-ranging group of ring-tailed lemurs (Lemur catta). Unpublished Ph.D. thesis, Washington University, St. Louis, MO, USA. Taylor, L. & Sussman, R. W. (1985). A preliminary study of kinship and social organization in a free-ranging group of Lemur catta. International Journal of Primatology, 6, 601–614. Tilson, R. T. & Hamilton, W. J. (1984). Social dominance and feeding patterns of spotted hyaenas. Animal Behaviour, 32, 715–724. Van Hooff, J. A. R. A. M. (2000). Relationships among non-human primate males: a deductive framework. In P. M. Kappeler, ed., Primate Males: Causes and Consequences of Variation in Group Composition. Cambridge: Cambridge University Press, pp. 183–191. Van Horn, R. C., Engh, A. L., Scribner, K. T., Funk, S. M., & Holekamp, K. E. (2004). Behavioral structuring of relatedness in the spotted hyena (Crocuta crocuta) suggests direct fitness benefits of clan-level cooperation. Molecular Ecology, 13, 449–458. Vehrencamp, S. L. (1983a). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Vehrencamp, S. L. (1983b). Optimal skew in cooperative societies. American Zoologist, 23, 327–355. Vervaecke, H. & van Elsacker, L. (2000). Sexual competition in a group of captive bonobos (Pan paniscus). Primates, 41, 109–115. Vervaecke, H., de Vries, H., & van Elsacker, L. (2000). Dominance and its behavioral measures in a captive group of bonobos (Pan paniscus). International Journal of Primatology, 21, 47–68. Whittingham, L. A., & Dunn, P. O. (1998). Male parental effort and paternity in a variable mating system. Animal Behaviour, 55, 629–640. Wimmer, B., Tautz, D., & Kappeler, P. M. (2002). The genetic population structure of the gray mouse lemur (Microcebus murinus), a basal primate from Madagascar. Behavioral Ecology and Sociobiology, 52, 166–175. Wright, P. C. (1999). Lemur traits and Madagascar ecology: coping with an island environment. Yearbook of Physical Anthropology, 42, 31–72. Young, A. J. & Clutton-Brock, T. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387.
83
4
The effects of heterogeneous regimes on reproductive skew in eutherian mammals c l a r a b. j on e s
Summary Although several models of reproductive skew have been proposed, “transactional” models, on the one hand, and “tug-of-war” or “incomplete control” models, on the other hand, are most commonly debated. The former hold that dominants control group size by yielding “incentives” (a share of total group productivity) to subordinates, while the latter advance the idea that the dominant’s control over one or more subordinates is incomplete. While high skew has been identified in several mammalian societies, most research on this topic shows that social mammals are likely to display intermediate, low, or variable reproductive skew. In an attempt to explain this pattern of results, the present chapter shows that mammals have evolved to cope with heterogeneous environmental regimes (abiotic and biotic), yielding a eutherian adaptive complex comprised of endothermy, relative brain enlargement, and behavioral flexibility. It is argued that these adaptations to environmental unpredictability favor the evolution of alternative phenotypes as well as situation- and condition-dependent responses decreasing the likelihood that dominants will be able to monopolize subordinates, including females, and that high skew will be observed. Additional research is required to highlight similarities and, most important, differences in the routes to sociality between insects, birds, and mammals (see Vehrencamp 1979). A simple mathematical model is presented linking reproductive suppression by a dominant to his/her influence on a subordinate and consequent Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
84
The effects of heterogeneous regimes on reproductive skew ability to monopolize the subordinate. This treatment shows that, depending upon conditions, both types of skew models are realistic. Reproductive skew is discussed in relation to allocation decisions and the different reproductive tactics of female and male mammals. All other things being equal, reproductive skew in heterogeneous regimes may be limited for taxa with relatively generalized phenotypes, because exposure to unpredictable conditions increases the likelihood of error, uncertainty, and risk of response outcomes, decreasing the ability to control a conspecific.
Introduction Individuals living in reproductive units (e.g. colonies, groups, troops) may share reproduction more or less equally. Alternatively, one or a few individuals may monopolize a group’s reproductive output. The apportionment of reproduction within groups by sex (“reproductive skew”: Vehrencamp 1983) may predict other features of animal sociality, such as agonistic (Hager 2003a) and foraging (Jones 2004) behaviors or responses to predation (Bian et al. 2005). Several recent treatments have argued that theories of reproductive skew may yield general formulations for the evolution of social behavior (Reeve & Keller 1996, Heinze & Keller 2000, Reeve & Emlen 2000, Reeve 2001). There is some disagreement, however, about the relative utility of “transactional” models, on the one hand, and “tug-of-war” or “incomplete control” models, on the other (Johnstone 2000, Hager 2003b), and alternate models of skew besides the most influential ones have also been proposed (e.g. Crespi & Ragsdale 2000). Each of these models analyzes social behavior as rule-governed, evolutionarily stable outcomes of variables (e.g. “ecological constraints”, coefficients of relatedness) predicting the apportionment of reproduction within groups. In concession models of skew, a subtype of transactional skew models, the presence of one or more same-sex subordinates is beneficial to the dominant that may extend “incentives” (i.e. one or more shares of total group reproductive productivity) to the subordinate in order to decrease the likelihood that (s)he will emigrate or escalate an interaction. Incentives are theorized to induce a subordinate to remain in the group and, since it is the dominant who “decides” to yield or not to yield incentives to a subordinate, the dominant ultimately determines group size in these models. In “tug-of-war” models (Clutton-Brock 1998), a class of compromise models of reproductive skew, dominants have “incomplete control” over the reproduction of same-sex group members. In this condition, subordinates may have options other than leaving the reproductive unit (dispersal) if the costs of remaining outweigh the
85
86
C. B. Jones associated benefits. Reproductive skew in groups with “incomplete control” by dominants may often be lower, on average, than reproductive skew in groups described by the classic transactional models if subordinates in the former condition have more behavioral options to escape a dominant’s control (e.g. alternative reproductive behaviors). Although competing skew models can be found in the literature, there is broad agreement that the degree of reproductive skew within groups is a function of three primary factors: (1) dispersal costs, (2) costs of reproduction, and (3) the coefficient of within-group relatedness (r) (Vehrencamp 2000). A primary focus of the present chapter concerns the consequences of environmental heterogeneity for variability in these three factors among social mammals. While no general model of reproductive skew has received broad consensus, researchers have shown that Hamilton’s (1964) seminal treatments require expansion beyond their exclusive emphasis upon coefficients of relatedness (r) and inclusive fitness maximizing (Johnstone 2000, West et al. 2001, 2002, Reeve 2001, Reeve & Shen 2006). In addition, several investigators have pointed out that factors determining reproductive skew may vary within and between taxa (e.g. Emlen 1984, Solomon & French 1997, Lacey & Sherman 1997, Vehrencamp 2000, Crespi 2005). The present chapter treats reproductive repression in heterogeneous regimes as one measure of “repression of competition” (Frank 1995, 2003, 2006) influencing patterns of reproductive skew in social mammals. For this report, environmental heterogeneity is defined as spatial and temporal variation in a stimulus or stimulus array endogenous or exogenous to the organism and representing selection pressures inducing alternative behavioral phenotypes (Piersma & Drent 2003, Jones 2005a). I propose in this chapter that, in general, social mammals, including humans, are adapted to cope with, to manage, and to respond to changing conditions with responses likely to yield intermediate, low, or variable – rather than high – reproductive skew, all other things being equal. A-priori evaluation of social evolution in heterogeneous regimes suggests that environmental unpredictability is likely to decrease the value, accuracy, and thus utility of a dominant’s “staying incentives” or “peace incentives” (see Vehrencamp 1983), leading to conditions favoring “incomplete control.” This hypothetical scenario supports the conclusions of Johnstone (2000) that “concession” models and “tug-of-war” models are not mutually exclusive.
Eutherian adaptations to environmental heterogeneity Extensive reviews of the evolutionary history of mammals may be found elsewhere (Eisenberg 1981, Vaughan et al. 2000, Feldhamer et al. 2004,
The effects of heterogeneous regimes on reproductive skew Rose & Archibald 2005). In Eisenberg’s (1981, p. 7) words, “There have been many natural experiments resulting from the early separation of the various phyletic lines that we recognize today as the class Mammalia”. While this statement highlights diversity among mammals, the present treatment focuses upon particular selection pressures and conserved features of eutherians which may help to explain variable configurations of reproductive skew among social taxa. Following Eisenberg’s (1981; see also Rose & Archibald 2005) report, the morphological grade characterizing mammals was attained by descendants of therapsids during the Triassic, more than 250 Ma. The emphasis in this chapter concerns changes in the mammalian line beginning in the early Cretaceous period (125 Ma), when early placentals first appeared. These groups and related mammalian assemblages (e.g. marsupials) demonstrated dentition characteristic of an insectivore–frugivore trophic level, limiting food resources with an unpredictable dispersion in time and space. The radiations of mammals during the Cretaceous occurred in consort with the radiation of angiosperms, events inducing plant–mammal coevolution. Marsupial and eutherian radiations were distinct and well advanced by the late Cretaceous, and differentiation of modern orders was recognizable by the Eocene (Tertiary period), 54.8–33.7 Ma. Mammalian adaptations to environmental heterogeneity are assumed to have been driven by plant–mammal interactions, factors thought to explain phenotypic plasticity and relative brain enlargement in these animals (Eisenberg 1981, Lillegraven et al. 1987, Jones 1995, Feldhamer et al. 2004, Gingerich 2006, Wilbur & Rudolf 2006).
The eutherian adaptive complex: endothermy, relative brain enlargement, and behavioral flexibility The eutherian adaptive complex was driven by environmental heterogeneity, in particular, variability in food abundance and dispersion (Eisenberg 1981, Jerison 1983, Vaughan et al. 2000, Rose & Archibald 2005, Gingerich 2006, McNab 2006). In brief, first principles of ecology indicate that the size and composition of groups change in response to temporal environmental heterogeneity and may have important consequences for the survival and fecundity of organisms (Roughgarden 1979, Pulliam & Caraco 1984, Wang et al. 2006). Population abundance and structure (including group size) through time is an attribute of resource predictability (Roughgarden 1979). High resource predictability and high resource quality, relatively homogeneous spatial dispersion of resources and resource tracking by the animal population is expected to favor resource defense (e.g. contest competition or territoriality)
87
88
C. B. Jones by individuals or small groups, on average, whereas low resource predictability and large distance or high variation in distance between patches may make resources indefensible (unmonopolizable), yielding large average group size (Schoener 1971, Emlen & Oring 1977, Roughgarden 1979, Pulliam & Caraco 1984). Since temporal unpredictability of resources may be positively correlated with spatial uncertainty (“patchiness”), foraging in groups may reduce average searching time per individual group member. Thus, environmental predictability will be inversely correlated with group size (Wittenberger 1980, Pulliam & Caraco 1984). The resultant population structure has significant consequences for genes and the individuals that carry them (Hewitt & Butlin 1997). Population structure may be evident as subdivision into demographic subunits or groups which represent an evolutionary compromise among those parameters yielding optimal inclusive fitness to individuals (Wilson 1975, Wittenberger 1980, Pulliam & Caraco 1984, Dunbar 1996). As Wilson (1975) pointed out, the frequency distribution of group sizes will be a function of those phenomena leading individuals to join and to leave groups combined with the selection pressures on individual responses to these forces. The parameters determining modal group size in a population, thus, are ultimately expressed as adaptations of individuals to local environments (Wilson 1975, Wittenberger 1980, Dunbar 1996). In the same local conditions, males and females may adopt different adaptive tactics and strategies for optimizing lifetime reproductive success due to the constraints of anisogamy (differential investment in gametes by each sex: Trivers 1972), intragenomic conflict (Burt & Trivers 2006), and/or intersexual conflict (costs imposed by one sex upon the other: Rice 2000). In mammals, these factors have consequences for the reproductive strategies of females, who are expected to adopt those behavioral programs conferring the greatest benefits from the conversion of resources, especially food, into offspring. In the same conditions, the fitness of mammalian males will depend upon their ability to monopolize females (or resources required by females: Bradbury & Vehrencamp 1977, Emlen & Oring 1977, Wittenberger 1980, Wrangham 1980, Nunn 1999) (see Figure 4.1). Eisenberg’s review (1966; see also Eisenberg 1981) demonstrated that most mammals exhibit a spatially dispersed (“solitary”) dispersion pattern in time and space excluding periods of mating and maternal care of young. Apparently, then, the benefits of group life have not outweighed sociality’s costs for most mammalian taxa. Nonetheless, complex social behavior in which adults of one or both sexes exhibit tolerance and consequent association in time and space may be found in several mammalian groups, in particular in
The effects of heterogeneous regimes on reproductive skew
Figure 4.1 An adult female mantled howler monkey (Alouatta palliata palliata) displaying genital hypertrophy which may have evolved with a suite of traits (e.g. female choice) characterizing female emancipation (Emlen & Oring 1977, Jones 2005a: 61–78; 91–92) in this species. Female emancipation reflects the unreliability of female monopolization by males and probably arises as a result of the unpredictability of resources, in particular, food, in time and space. Female emancipation and related characteristics of this species (e.g. multiple mating) are likely to constrain the potential for high skew among males (see Jones 1985, Jones & Corte´s-Ortiz 1998).
carnivores, cetaceans, and primates (Feldhamer et al. 2004; see also Eisenberg 1981, pp. 424–425, Gingerich 2006). Some members of these taxa, in addition to some rodents (e.g. the naked mole-rats, Heterocephalus glaber: Lacey & Sherman 1997; see Solomon & French 1997), exhibit noteworthy mechanisms
89
90
C. B. Jones of reproductive control, cooperation, or division of labor. While several authors have advanced classification systems for social taxa (see review in Jones 2005a, pp. 115–121; also Jones 2005b, p. 20), Crespi’s (2005) recent schema is employed in this chapter. In brief, assessing social states in terms of trade-offs (differential optima) of costs and benefits to reproductive success in different environmental regimes, Crespi (2005, p. 570) envisions three classes of social system: (1) “eusocial,” in which “the trade-off involves two permanently distinct types of individual, and this permanence results in two independently evolving phenotypic systems;” (2) “cooperatively breeding,” in which “the difference between ‘breeders’ and ‘helpers’ is not permanent – individuals are ‘totipotent’ and can switch roles;” and (3) “communal,” in which individuals are also totipotent but “there is only one type of individual, who both breeds and engages in helping.” The adaptive significance of endothermy
Feldhamer et al. (2004, pp. 113–115) provide a detailed summary of the significance of endothermy (maintaining a relatively constant body temperature by means of heat produced inside the body), pointing out that this adaptation explains the ability of mammals to survive “inhospitable environments.” Endothermy becomes, then, the primary component of the eutherian adaptive axis, permitting mammals to remain active over a broad range of environmental conditions (Vaughan 1978). Endothermy is most likely a necessary condition for relative brain enlargement, possibly favored initially in changing environments by the role of the hypothalamus and thyroid gland in regulating body temperature and by benefits associated with increasing endothermy’s basic functions of centralized coordination and control (see Jerison 1983). The adaptive significance of relative brain enlargement
The brain, in particular the neocortex, affords a central location for organismal coordination and control (Eisenberg 1981, Jerison 1983) in response to environmental stimuli. For mammals and many other vertebrates, the neocortex is associated with behavioral flexibility (Matzel et al. 2003, Hsu et al. 2005) and mechanisms of learning that may vary in complexity (e.g. habituation, imprinting, associative processes, social learning, cognition). The mammalian neocortex is also capable of assessing likelihoods of reward and selfish gain in conditions of uncertainty (Rilling et al. 2002, de Quervain et al. 2004, King-Casas et al. 2005, Schultz 2006), features that may enhance fitness in heterogeneous regimes (e.g. Bian et al. 2005, Jansen & Stumpf 2005). Ecological and social factors are thought to be the primary selective factors promoting
The effects of heterogeneous regimes on reproductive skew relative brain enlargement in mammals (Shultz & Dunbar 2005). These factors are not mutually exclusive, since the social decisions that individuals make will reflect resource abundance and dispersion in local conditions, a perspective fundamental to behavioral ecology (Emlen & Oring 1977). Since females are expected to be “energy maximizers” (Schoener 1971) and more sensitive than males (“time minimizers”: Schoener 1971) to ecological regimes, all other things being equal, future investigations of the relative import of ecological and social variables for relative brain enlargement require that males and females be analyzed separately. These fundamental differences between the sexes are expected to have broad implications for the behavioral decisions made by eutherians, including tactics and strategies of reproductive allocation (Feldhamer et al. 2004, Jones 2005a, Kussell & Leibler 2005). Related to the evolution of endothermy and large brains is the observation that both adaptations are energetically costly (Eisenberg 1981, Jones 2005a, Isler & van Schaik 2006), with consequent impacts upon metabolic processes and behavioral, including sociosexual, decisions made by females and males. Relative brain enlargement will be associated with a suite of traits including longevity, late sexual maturity, and iteroparity (Eisenberg 1981, Kappeler & Pereira 2003), characteristics associated with adult over juvenile survival (see, for example, Jones 1997). Temporal and spatial allocation patterns dependent upon “sensing,” learning, and memory rather than stochastic (“bethedging”) tactics and strategies (Jansen & Stumpf 2005, Kussell & Leibler 2005) are diagnostic of many eutherians, and, combined with other mammalian traits (e.g. modes of locomotion and reproduction), probably distinguish these taxa from other families (insects, birds) noteworthy for high grades of sociality. Many eutherians, then, exploited an open niche favoring the capacity for “responsive switching” (Jansen & Stumpf 2005, Kussell & Leibler 2005) in heterogeneous regimes. Responsive switching entails facultative physiological and behavioral changes in the face of environmental fluctuations, with consequent benefits (e.g. adjustment to changing rates of abiotic and biotic environmental change, generation of novel responses) and costs (e.g. the maintenance of requisite molecular and physiological machinery, energetic “neglect” if the environment rarely changes, production of maladapted individuals). Investigating the trade-offs of responsive switching in social mammals may increase our understanding of patterns of reproductive skew, including sociosexual patterns of energy and time allocation. It will also be important to study when and under what conditions stochastic phenotype switching and responsive switching are employed within and between eutherians
91
92
C. B. Jones and other taxa, possibly yielding differential patterns of reproductive skew. Taxa exhibiting responsive switching may be more likely to display low and intermediate reproductive skew, since individuals of these groups are expected to have more control over their responses to environmental change, possibly yielding a broader range of reproductive tactics and strategies to achieve reproduction directly or to avoid control or influence by conspecifics. The adaptive significance of behavioral flexibility
The third component of the mammalian axis, behavioral flexibility, reflects selection in heterogeneous regimes, optimizing genotypic and phenotypic success in conditions of uncertainty or risk (Meyers & Bull 2002, Piersma & Drent 2003, Reader & MacDonald 2003, Jones 2005a; see also Lewontin 1957). Behavioral flexibility may be promoted by the coordination and control functions accompanying relative brain enlargement, facilitating an organism’s adjustment to changing conditions, abiotic (e.g. temperature, humidity) or biotic (e.g. feeding or social regimes), and endogenous (e.g. immunological, hormonal) or exogenous (e.g. mate quality, interaction rates) stimuli (see Mayr 1963, Jerison 1983, Rutherford 2000). Recent evidence demonstrates that mechanisms switching responses from one behavioral pattern to another are genetically induced in mammals (Choi et al. 2005; see also WestEberhard 1979, Jones & Agoramoorthy 2003). Behavioral flexibility, a reversible component of the phenotype (Piersma & Drent 2003), yields alternative phenotypes across time and space for the same individual, whose patterns of response are more or less “totipotent” (the ability of an individual to perform most or all of a society’s roles). Because most social mammals, including humans (Homo sapiens), are communal rather than eusocial or cooperatively breeding, the study of alternative behavioral phenotypes in mammals is primarily a study of polyphenisms (environmentally switched alternative phenotypes) rather than polymorphisms (genotypically regulated alternative phenotypes), characteristic of social insects (Wilson 1971) and naked mole-rats (O’Riain et al. 2000). Totipotency renders relatively generalized phenotypes or the ability to change one’s patterns of response rapidly, all other things being equal, in response to abiotic or biotic (including social) regimes (see West-Eberhard 1979, 2003). This trait also allows an individual to exhibit relatively specialized responses for varying periods of time (e.g. temporarily displaying a helper role: Nicolson 1987, adjusting lactation to food shortages: Dall & Boyd 2004; modifying intensities of offspring competition in response to growth rates: Stockley & Parker 2002, exploiting seasonal opportunities: Gockel & Ruf 2001).
The effects of heterogeneous regimes on reproductive skew The eutherian adaptive complex and reproductive skew Exposure to environmental heterogeneity favored adaptations predisposing eutherians to intermediate, low, or variable levels of reproductive skew, since responses are likely to be condition-dependent and/or facultative (Vehrencamp 2000; see also Waynforth et al. 1998). Condition-dependent responses will be enhanced by large brains and behavioral flexibility favoring alternative behaviors, and some of these responses will be a function of learning and higher-order processes (e.g. imitation, problem solving). Outcomes of sociosexual interactions are likely to be highly variable in many, if not most, eutherian societies, since exposure to environmental heterogeneity will decrease the accuracy of behavioral decisions (e.g. attempts by dominants to repress subordinates’ competition, including reproduction) by increasing response variability and, consequently, error. Further, alternative response patterns may arise where phenotypes are exposed to unpredictable regimes, often creating sociosexual opportunities (e.g. parasitic or cooperative phenotypes: Taborsky 1994). These alternative phenotypes may obtain shares of a group’s reproductive output over the short or long terms, decreasing within-group skew. Large brains and behavioral flexibility may also facilitate decisions to play a “waiting game” (e.g. queueing: Voigt & Streich 2003, Alberts et al. 2003; see Kokko & Johnstone 1999) or to adopt a temporarily unproductive or maladaptive phenotype, possibly imposed by an influential group member (see Jones 2005a). Phenotypic hitchhiking, whereby one individual, usually a subordinate, “hitchhikes” on the phenotype of another, usually a dominant, may permit individuals to “buy time” by adopting an alternative strategy until the abiotic or biotic (including social) environment becomes more auspicious for the expression of optimal responses. Subordinate males, for example, may form a temporary coalition or alliance with one or more dominant males while ascending a group’s dominance hierarchy and/or while waiting for opportunities to displace dominants. For example, Jones (1980) reported a coalition between a young male pretender and a dominant male. Over time, the young male rose to dominant rank in his group, and the former dominant lost this position. A coalition between these two males expelled another high-ranking male from the group. In another case, a high-ranking male was observed to assume temporary subordinate rank within his group while a wound healed (C. B. Jones, personal observation). These and other examples (e.g. where individuals become “helpers” to other group members) of phenotypic hitchhiking have the potential to influence reproductive skew within a group.
93
94
C. B. Jones Research on sex and reproduction in mammals has been dominated by studies of male tactics and strategies until relatively recently (Andersson 1994, Ellis 1995, Jennions & Petrie 1997, Dixson 1998, Shahnoor & Jones 2003). Research is needed to investigate causes and consequences of within-group relative contributions of direct (selfish) reproduction (“reproductive skew”) by mammals of both sexes in an attempt to describe and test patterns and processes of despotic (high skew) or egalitarian (intermediate or low skew) relationships or variations in reproductive skew across conditions (e.g. season, habitat) or groups of the same species. Low, intermediate, or variable reproductive skew among males in polygynous and/or multi-male groups may also be promoted by the virtually ubiquitous tendency of eutherian females to mate multiply (Wolff & Macdonald 2004, Jones 2005a, pp. 100–102). This female adaptation may impose costs of time and of ejaculate production upon males (Preston et al. 2001, Eady & Hardy 2001), a condition likely to constrain high skew in this sex. Investigations are also required to identify those features of mammals, mammalian phylogeny, and mammalian regimes differentiating their patterns of reproductive skew from those found in birds and insects, taxa more predisposed, on average, to high-skew societies than documented for mammals. One possibility is that individualistic strategies are more highly elaborated in social mammals, a condition that would dampen the evolution of high grades of sociality (Jones 2005a) and thus high-skew societies. Most, if not all, groups of social mammals are distinguished by the presence of helpers, usually subordinate females assisting the reproductive output of dominant females, to whom they are often related genetically. High skew has been reported for some social mammals (naked mole-rats: Sherman et al. 1991; male lions, Panthera leo: Packer et al. 1991; some marmoset, Callithrix spp., and tamarin, Saguinus spp., monkeys: Abbott 1993; black rhinoceros, Diceros bicornis: Garnier et al. 2001; male Antarctic fur seal, Arctocephalus gazella: Hoffman et al. 2003; female house mice, Mus musculus: Rusu & Krackow 2004; male whitefaced capuchin monkeys, Cebus capucinus: Jack & Fedigan 2006). Nonetheless, as summarized in Table 4.1, most research on social mammals supports the view that reproductive skew within groups is low, intermediate, or highly variable and, as this chapter suggests, differentially responsive to local conditions. This observation supports the argument that, all other things being equal, mammals are selected to endure and to manage direct exposure to environmental uncertainty, an evolutionary scenario leading not only to the suite of characteristics presented as the eutherian adaptive complex but also to a predisposition to make the “best of a bad job” (see Austad 1984, Brockmann 2001, Ebensperger & Blumstein 2006).
The effects of heterogeneous regimes on reproductive skew Table 4.1 Selected overview of eutherian societies with low, intermediate, or variable reproductive skew. In these studies, skew was generally not calculated quantitatively. Thus, classification is typically based on inferences from researchers’ descriptions. Reproductive Skew Order (species)
(sex)
Notes/references
Low (females)
Coefficient of relatedness poor predictor of
Chiroptera Bechstein’s bat (Myotis bechsteinii)
group structure and reproductive skew (Kerth et al. 2002)
Brown long-eared bat
Low (males)
(Plecotus auritus)
Kin selection did not account for patterns of natal philopatry (Burland et al. 2001)
Primates Brown lemur (Eulemur
Variable (males)
Reproductive skew strongly influenced by
Low (females)
Suppression of competition frequently
fulvus mayottensis) Ring-tailed lemur
female choice (Gachot-Neveu et al. 1999)
(Lemur catta)
initiated by dominants to subordinates (Cavigelli et al. 2003)
Moustached tamarin (Saguinus mystax)
Variable (moderate to high (males)
Woolly spider monkey Low (females and or muriqui
males)
(Brachyteles
Evidence of multiple mating by females (Goldizen 1988, Huck et al. 2005) No evidence of dominance hierarchies or reproductive suppression (Strier 1986, 1992)
arachnoides) Vervet monkey
Variable (males)
(Cercopithecus aethiops)
Results supported “limited control” model of reproductive skew (Whitten & Turner 2004)
Rhesus macaque
Variable (males)
(Macaca mulatta)
Coefficients of relatedness did not predict patterns of reproductive skew, and results supported “limited control” model (Widdig et al. 2004)
Barbary macaque (Macaca sylvanus)
Low (females and males)
Correlation between social rank and reproductive success appears to be absent; low skew a function of weak environmental constraints (Ku¨mmerli & Martin 2005)
Mandrill (Mandrillus sphinx)
Variable (females and males)
Success of dominant males generally high but decreased as number of reproductive males in group increased; skew influenced by heterozygocity (Setchell et al. 2005, Charpentier et al. 2005)
Yellow baboon (Papio cynocephalus)
Variable (males)
Reproductive skew a function of male tenure and group size (Alberts et al. 2003)
95
96
C. B. Jones Table 4.1 (cont.) Reproductive Skew Order (species)
(sex)
Notes/references
Mountain gorilla
Low to intermediate
Results support “tug of war” model of
(Gorilla beringei
(males)
beringei)
reproductive skew since subordinate male can escape monopolization by silverback (Bradley et al. 2005)
Chimpanzee (Pan
Variable (females)
troglodytes)
Linear dominance hierarchies not evident, and females have dispersed configuration of foraging (Pusey et al. 1997)
Human (Homo sapiens)
Variable (females)
Menopause increases reproductive skew among females (Foster & Ratnieks 2005; see also Austad 1997, Packer et al. 1998, Fedigan & Pavelka 2001)
Human (Homo sapiens)
Variable (males)
Promiscuity and slavery increase reproductive skew among males (Dickemann 1997, Betzig, 1997a, 1997b)
Wealthy modern human (Homo sapiens)
Variable (females and males)
“Snowballing” resources increases reproductive skew (Hill & Reeve 2005)
Carnivora Meerkat (Suricata suricatta)
Variable but usually high (females)
Results support “limited control” models of reproductive skew (Clutton-Brock et al. 2001, Young & Clutton-Brock 2006)
Banded mongoose
Low (females)
(Mungos mungo)
No evidence of reproductive suppression of subordinates by dominants or of inbreeding avoidance; egalitarianism among females may be induced by benefits of cooperative breeding (de Luca & Ginsburg 2001, Gilchrist et al. 2004)
Spotted hyena (Crocuta Intermediate to low crocuta)
(males)
Results support “limited control” models of reproductive skew, of which female choice and male tenure are significant determinants (Engh et al. 2002)
African lion (Panthera
Low (females)
leo)
Egalitarianism among females may be induced by benefits of cooperative breeding and high costs of female–female aggression (Packer et al. 2001)
Gray seal (Halichoerus grypus)
Variable (males)
Reproductive skew influenced by tendency of females to mate multiply, in particular, with males at sea (Worthington Wilmer et al. 1999, Ambs et al. 1999)
The effects of heterogeneous regimes on reproductive skew Table 4.1 (cont.) Reproductive Skew Order (species)
(sex)
Notes/references
Intermediate to low
Levels of reproductive skew influenced by
Cetacea Humpback whale (Megaptera
(males)
novaeangliae)
operational sex ratio and consequently low competition among males for mates (Cerchio et al. 2005)
Artiodactyla White-tailed deer
Variable (males)
Some evidence of assortative mating by age
Variable (males)
Evidence of multiple mating by females
(Odocoileus virginianus) Pronghorn antelope
(Sorin 2004)
(Antilocapra americana)
including multiple paternity of litters (Carling et al. 2003)
Soay sheep (Ovis aries)
Intermediate to low
Camargue stallion
Low (females);
(males) (Equus caballus)
variable (males)
Depletion of ejaculate with subsequent matings limits skew (Preston et al. 2001) Male reproductive skew a function of maternal rank, female choice, and success of coalitions (Feh 1990)
Rodentia Wood mouse
Variable (females)
(Apodemus sylvaticus)
High reproductive skew associated with mother–daughter breeding groups and with large differences in maternal investment in litters (Gerlach & Bartmann 2002)
African striped mouse (Rhabdomys pumilio)
Variable (females and males)
Patterns of philopatry and consequent reproductive skew a function of habitat (“social flexibility”) (Schradin & Pillay 2005)
Alpine marmot
Variable (females)
(Marmota marmota)
Reproductive suppression costly for dominant females, and skew decreased with increasing group size (King & Allaine´ 2002, Hackla¨nder et al. 2003)
Yellow-bellied marmot Variable (females) (Marmota flaviventris)
Reproductive skew a function of female group size, and lower skew in larger groups was associated with increased female–female aggression (Armitage & Schwartz 2000)
97
98
C. B. Jones Reproductive skew and reproductive allocation patterns in social mammals All other things being equal, mammalian females are more likely than males to exhibit offspring care (Eisenberg 1966, 1981, Trivers 1972) and cooperation (Jones 2005b, Box 4.1), although both responses may be induced in
Box 4.1. Why are females more likely than males to be social? The simple answer to this question is, Because the benefits of sociality are greater to females than to males. Why? It has been suggested that sociality is most likely to evolve as an energy-saving strategy (Heinze & Keller 2000; see also Jones & Agoramoorthy 2003, Jones 2005b). Females (“energymaximizers”: Schoener 1971) have more to gain, all other things being equal, from energy-saving opportunities, and are expected to be more
The effects of heterogeneous regimes on reproductive skew Box 4.1. (cont.) sensitive than males to energetic costs (Jones 2005b). These conditions will predispose females to sociality (e.g. altruism, philopatry) where these responses deliver energetic gains benefiting inclusive fitness (Queller 1997). The graphical model describes the costs (C ) or benefits (B) to female inclusive fitness (expected lifetime reproductive success) of the relative degree of sociality as a function of differential energy savings, from low (–) to high (þþ). Benefits will increase and then level off as the costs increase linearly (because resources, in particular food, are limiting), and the maximum net benefit (benefit minus cost) to females should occur at threshold x. The location of x will depend upon the positions and shapes of the benefit and cost curves, a function of environmental unpredictability over the short and long terms. Energy limitation may explain why the highest grades of sociality (eusociality and cooperative breeding) are relatively uncommon among mammals (Jones & Agoramoorthy 2003), for which endothermy and large brain size are energetically costly (Eisenberg 1981, Feldhamer et al. 2004; see also Jansen & Stumpf 2005, Kussell & Leibler 2005), and broad patterns of reproductive skew in mammals may correspond to regional differences in resource dispersion in the tropics (Fleming et al. 1987).
conspecifics by force (taking away behavioral control), coercion (imposing costs on non-cooperators), manipulation (exerting influence for selfish advantage), or exploitation (use of another’s resources for selfish ends: Crespi & Ragsdale 2000; see also Frank 1995, 2003, 2006, Helms Cahan 2001, Jones 2005a, 2005b; Figure 4.2). Since most eutherian taxa are characterized by sexual dimorphism, with females the smaller sex (Crook 1972, Eisenberg 1981), and since energy limitation will usually be more important to a female than to a male mammal (Schoener 1971, Jones 2005a, pp. 61–78, 2005b; Box 4.1), female mammals are expected to be more vulnerable to environmental perturbations, particularly in food resources, than a male in similar conditions, all other things being equal. As a result, allocation decisions determining reproductive skew among group-living mammalian females, such as the differential benefits and costs of cooperating with or suppressing another female’s reproduction, or of dispersing away from a natal group rather than remaining, will generally have greater consequences for a female’s lifetime reproductive success. This condition obtains because of a female’s greater expected parental investment compared with males (Trivers 1972), because
99
100
C. B. Jones
Figure 4.2
such decisions are likely to increase female inter-birth intervals, because males are likely to parasitize the reproductive allocation patterns of females (Alexander et al. 1997, Jones 2007), and because sexual conflict will usually benefit males (Rice 2000, Jones 2005a, pp. 61–78). Male mammals, on the other hand, with an inherently higher reproductive capacity than females in the same conditions (Trivers 1972), should benefit more from selfish tactics and strategies (e.g. avoidance of other males) likely to maximize their reproductive potential, and should be capable and willing to bear greater costs (e.g. exposure to heterogeneous conditions or higher tolerance for behavioral inaccuracy and risk) in their pursuit of fertilizable females. Simply, males can afford to make more reproductive errors than females, all other things being equal. This asymmetry between the sexes will dampen the potential benefits that mammalian males, including those with lower resource-holding potential, might gain from associating with their mate(s) or other males, since these gains are likely to be outweighed by the benefits of individualized tactics and strategies, including risk taking, to maximize future reproduction.
The effects of heterogeneous regimes on reproductive skew The literature on mammals supports this preliminary scenario (see, for example, Wilson 1975, Wittenberger 1980, Feldhamer et al. 2004). Mammalian females living in groups, for example, are more likely to be philopatric than are males of the same species in the same conditions (Greenwood 1980), and male mammals are unlikely to be social (Eisenberg 1966, 1981, Wilson 1975). This pattern suggests that dispersal costs, costs of reproduction, and/or the benefits of remaining within matrilines are high for female mammals, and that the benefits of sociality are low for mammalian males. The inherent differences in adaptations to group living between the sexes of this class are likely to explain the outcomes summarized in Table 4.1. As Le Galliard et al. (2005, p. 206) point out, “Sociality typically requires, first, some form of altruistic behavior through which individuals sacrifice their own fitness for the benefit of others . . . and, second, some reduction in individual mobility, allowing for sustained interaction, which exacerbates competition for local resources.” All other things being equal, female and male mammals will differentially weight the costs and benefits of the two requirements relative to their abiotic and biotic environments, including the temporal and spatial distribution of conspecifics (kin and non-kin). The resulting allocation decisions will determine patterns of within-sex reproductive skew within groups. Reproductive skew and reproductive suppression in heterogeneous regimes Crespi & Ragsdale (2000) argue that the apportionment of reproduction within groups results from cooperation (mutualistic benefits), force, persuasion (providing benefits to cooperators), or coercion. These four factors describe competitive relations between group members of the same sex yielding asymmetries in their contribution to total group productivity, and may be determined by repression of reproductive behaviors, usually repression of subordinate reproduction by dominants. More effective reproductive repression (e.g. punishment, policing) will generally lead to higher reproductive skew (but see Young & Clutton-Brock 2006). Factors such as environmental heterogeneity, however, may dampen the potential for high-skew societies by increasing the uncertainty, and thereby error, of any behavioral decision by a potential repressor. Heterogeneous regimes will also favor alternative phenotypes and behavioral flexibility, increasing the likelihood that potential targets of reproductive repression, usually subordinates, will partially or completely escape repression attempts (e.g. by policing, such as infanticide: see Hager & Johnstone 2004, Jones 2005c, 2007, by coalitions and alliances with other group members against the
101
102
C. B. Jones repressor: see Flack et al. 2006; by social parasitism, including phenotypic manipulation: see Jones 2005a, 2005b). Such post-repression events have the potential to significantly modify power relations within mammalian groups (Feh 1990, Jones 2000, 2005d, 2006, Gros-Louis et al. 2003, Kutsukake & Hasegawa 2005, Saj & Sicotte 2005). A simple model by May & Anderson (1990) can be modified to describe the potential benefits and costs of reproductive repression. This model measures the fitness of a parasite as a measure of the parasite’s influence on its host. Assume that the fitness of a potential repressor (a social parasite) can be measured as reproductive rate (R0), a density-dependent value. The effectiveness of reproductive repression may be linked to a repressor’s influence on the recipient (host) of its act (e.g. increased inter-birth intervals, increased mortality rates, or decreased litter size experienced by the recipient will decrease his/her fitness). May & Anderson’s formulation can be written for reproductive repression such that R0 ¼ yðNÞ=ða þ b þ vÞ
where y is effectiveness of reproductive repression, N is population density of potential recipients of the act, a is rate of recipient cost (e.g. mortality rate: see Le Galliard et al. 2005) from intensity of reproductive repression, b is rate of recipient’s cost from all but intensity of reproductive repression, and v is recovery rate (the recipient’s ability to completely or partially escape the deleterious effects of reproductive repression). The report by Bradley et al. (2005) on male mountain gorillas (Gorilla beringei) with low to intermediate skew provides a preliminary example of May & Anderson’s (1990) model where y is incomplete, a is not high because of the subordinate’s ability to adopt alternative behavioral tactics and strategies, b is lower than the costs of dispersal, and v is intermediate to high because of the subordinate’s ability to avoid or escape monopolization by the dominant (“negative reinforcement”: Jones 2002). Following May & Anderson (1990), R0 increases as a decreases when y, b, and/ or v are independent. Under these conditions, reproductive repression would not be favored by selection, since the costs of attempted repression would outweigh its benefits. This situation is likely to arise in heterogeneous regimes where the correlation of abiotic and biotic events is decreased. In such “noisy,” stochastic circumstances, the potential for high reproductive skew should be minimized, and “limited control” models of reproductive skew are more likely to apply. Where y, b, and/or v are correlated, however, reproductive repression should be favored, and the intensity of reproductive repression should be determined by the relative degree of benefit to the actor (the repressor), all
The effects of heterogeneous regimes on reproductive skew other things being equal, consistent with “concession” models of reproductive skew. Correlation of y, b, and/or v determines environmental predictability and may be employed as cues for sociosexual decision making.
Conclusions and prospects The arguments advanced in the present chapter hold that social mammals exposed to unpredictable environments are pre-adapted to heterogeneous regimes, increasing the likelihood that these taxa will exhibit low, intermediate, or variable patterns of reproductive skew compared to social insects and cooperative birds. In future, it will be important to measure environmental heterogeneity (see, for example, Jones 1997) in relation to dispersal costs, costs of reproduction, and within-group relatedness for a broad range of social animals, including humans, in order to explain variations in reproductive skew within and between populations and species. Similar treatments have argued that the benefits of assisting the reproduction of kin and their competitors (West et al. 2002) and the benefits of dispersal (Jones 2005a, pp. 18–25) are a function of local competition, and variations in the intensity of local competition are likely, as well, to strongly influence patterns of altruism and philopatry among all group members whose interests conflict. Since the intensity of local competition will vary with environmental predictability (e.g. Emlen & Oring 1977), reproductive skew should reflect these conditions. Frank (2003, p. 693) asserted, “Repression of competition within groups joins kin selection as the second major force in the history of life shaping the evolution of cooperation.” Future studies need to evaluate the differential significance of dispersal costs, costs of reproduction, and the coefficient of within-group relatedness (r) within and between the orders (insects, birds, and mammals) characterized by expansion of sociality (Table 4.1; see also Vehrencamp 1979, 2000, Kokko & Johnstone 1999, Reeve & Shen 2006). It is expected that, for social mammals, decisions to respond or not to respond are more dependent upon an act’s consequences for the competitive regimes of the actor’s direct (progeny) or indirect (other relatives) kin (Jones 2005c; see also West et al. 2002), rather than r, per se, possibly due to the fluctuating effects of heterogeneous regimes. It is possible that, for eutherians, the benefits of altruism and restricted mobility are more often outweighed by the benefits of dispersal, compared to social insects and birds. Partially supporting these ideas, unpredictable conditions induce increased rates of bisexual dispersal in some social mammals (Jones 1999). Repression of reproduction may be the primary determinant of altruism and limited mobility, and thus reproductive skew, in mammals. Studies are
103
104
C. B. Jones required to identify the range of response thresholds to heterogeneous conditions characteristic of the sexes in each mammal species. Finally, because anthropogenic perturbations continue to increase in frequency, duration, and intensity in many parts of the world, increasing the likelihood of mammal extinctions, patterns of reproductive skew may be linked to the persistence of populations in some areas, elevating the importance of studies of the relationship between reproductive skew and reproductive rate within groups. A recent search for the phrase “reproductive skew” in the Journal of Mammalogy yielded only 13 citations. Hopefully, students of mammals will embrace this topic for a more complete understanding of conserved and derived social traits within the class. Acknowledgments Jack Bradbury, Bob Johnston, Fred Stollnitz, “Griff” Ewer, Don Wilson, Norm Scott, Richard W. “Thor” Thorington, and John Eisenberg were my first models in mammalogy. I am fortunate that these specialists shared their knowledge with me during my earliest years as a student and researcher. I am indebted to my postdoctoral advisor, Dick Lewontin, who, at every turn, addressed stochasticity’s impact on animal, including human, populations. I appreciate the constructive criticism provided by Lee Drickamer and Reinmar Hager on an early version of this chapter. This chapter was discussed by the NESCent Mammal Reading Group (Kathleen Smith, Samantha Hopkins, Samantha Price, Louise Roth, and Clara Jones), whose comments proved useful in revising the manuscript. Chris Maher shared an anecdote about woodchucks highlighting the behavioral flexibility of “solitary” mammals. This chapter is dedicated with gratitude to the memory of R. F. “Griff ” Ewer, from my 1973 field season at Barro Colorado Island, Panama. Supported by the National Evolutionary Synthesis Center (NESCent), NSF #EF-0423641. References Abbott, D. H. (1993). Social conflict and reproductive suppression in marmoset and tamarin monkeys. In W. A. Mason & S. P. Mendoza, eds., Primate Social Conflict. Albany, NY: State University of New York Press, pp. 331–372. Alberts, S. C., Watts, H. E., & Altmann, J. (2003). Queuing and queue-jumping: longterm patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Animal Behaviour, 65, 821–840. Alexander, R. D., Marshall, D. C., & Cooley, J. R. (1997). Evolutionary perspectives on insect mating. In J. C. Choe & B. J. Crespi, eds., The Evolution of Mating Systems in Insects and Arachnids. Cambridge: Cambridge University Press, pp. 4–31.
The effects of heterogeneous regimes on reproductive skew Ambs, S. M., Boness, D. J., Bowen, W. D., Perry, E. A., & Fleischer, R. C. (1999). Proximate factors associated with high levels of extraconsort fertilization in polygynous grey seals. Animal Behaviour, 58, 527–535. Andersson, M. (1994). Sexual Selection. Princeton, NJ: Princeton University Press. Armitage, K. B. & Schwartz, O. A. (2000). Social enhancement of fitness in yellowbellied marmots. Proceedings of the National Academy of Sciences of the USA, 97, 12149–12152. Austad, S. N. (1984). A classification of alternative reproductive behaviors and methods for field-testing ESS models. American Zoologist, 24, 309–319. Austad, S. N. (1997). Postreproductive survival. In S. A. Austad, ed., Between Zeus and the Salmon: the Biodemography of Longevity. Washington, DC: National Academy of Sciences, pp. 161–174. Betzig, L. (1997a). Roman polygyny. In L. Betzig ed., Human Nature: a Critical Reader. Oxford: Oxford University Press, pp. 375–398. Betzig, L. (1997b). Why a despot? In L. Betzig, ed., Human Nature: a Critical Reader. Oxford: Oxford University Press, pp. 399–401. Bian, J., Wu, Y., & Liu, J. (2005). Breeding behavior under temporal risk of predation in male root voles (Microtus oeconomus). Journal of Mammalogy, 86, 953–960. Bradbury, J. W. & Vehrencamp, S. L. (1977). Social organization and foraging in emballonurid bats III: mating systems. Behavioral Ecology and Sociobiology, 2, 1–17. Bradley, B. J., Robbins, M. M., Williamson, E. A., et al. (2005). Mountain gorilla tug-of-war: silverbacks have limited control over reproduction in multimale groups. Proceedings of the National Academy of Sciences of the USA, 102, 9418–9423. Brockmann, H. J. (2001). The evolution of alternative strategies and tactics. Advances in the Study of Behavior, 30, 1–51. Burland, T. M., Barratt, E. M., Nichols, R. A., & Racey, P. A. (2001). Mating patterns, relatedness and the basis of natal philopatry in the brown long-eared bat, Plecotus auritus. Molecular Ecology, 10, 1309–1321. Burt, A. & Trivers, R. (2006). Genes in Conflict. Cambridge, MA: Harvard University Press. Carling, M. D., Wiseman, P. A., & Byers, J. A. (2003). Microsatellite analysis reveals multiple paternity in a population of wild pronghorn antelopes (Antilocapra americana). Journal of Mammalogy, 84, 1237–1243. Cavigelli, S. A., Dubovick, T., Levash, W., Jolly, A., & Pitts, A. (2003). Female dominance status and fecal corticoids in a cooperative breeder with low reproductive skew: ring-tailed lemurs (Lemur catta). Hormones and Behavior, 43, 166–179. Cerchio, S., Jacobsen, J. K., Cholewiak, D. M., Falcone, E. A., & Merriwether, D. A. (2005). Paternity in humpback whales, Megaptera novaeangliae: assessing polygyny and skew in male reproductive success. Animal Behaviour, 70, 267–277. Charpentier, M., Setchell, J. M., Prugnolle, F., et al. (2005). Genetic diversity and reproductive success in mandrills (Mandrillus sphinx). Proceedings of the National Academy of Sciences of the USA, 102, 16723–16728.
105
106
C. B. Jones Choi, G. B., Dong, H., Murphy, A. J., et al. (2005). Lhx6 delineates a pathway mediating innate reproductive behaviors from the amygdale to the hypothalamus. Neuron, 46, 647–660. Clutton-Brock, T. H. (1998). Reproductive concessions and skew in vertebrates. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H., Brotherton, P. N., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Crespi, B. J. (2005). Social sophistry: logos and mythos in the forms of cooperation. Annales Zoologici Fennici, 42, 569–571. Crespi, B. J. & Ragsdale, J. E. (2000). A skew model for the evolution of sociality via manipulation: why it is better to be feared than loved. Proceedings of the Royal Society of London B, 267, 821–828. Crook, J. H. (1972). Sexual selection, dimorphism, and social organization in the primates. In B. Campbell, ed., Sexual Selection and the Descent of Man, 1871–1971. Chicago: Aldine, pp. 231–281. Dall, S. R. X. & Boyd, I. L. (2004). Evolution of mammals: lactation helps mothers to cope with unreliable food supplies. Proceedings of the Royal Society of London B, 271, 2049–2057. de Luca, D. W. & Ginsberg, J. R. (2000). Dominance, reproduction, and survival in banded mongooses: towards an egalitarian social system? Animal Behaviour, 61, 17–30. de Quervain, D. J., Fischbacher, U., Treyer, V., et al. (2004). The neural basis of altruistic punishment. Science, 305, 1246–1247. Dickemann, M. (1997). Paternal confidence and dowry competition: a biocultural analysis of purdah. In L. Betzig, ed., Human Nature: a Critical Reader. Oxford: Oxford University Press, pp. 311–328. Dixson, A. F. (1998). Primate Sexuality: Comparative Studies of the Prosimians, Monkeys, Apes, and Human Beings. Oxford: Oxford University Press. Dunbar, R. I. M. (1996). Determinants of group size in primates: a general model. In W. G. Runciman, J. Maynard Smith, & R. I. M. Dunbar, eds., Evolution of Social Behaviour Patterns in Primates and Man. Oxford: Oxford University Press, pp. 33–58. Eady, P. E. & Hardy, I. C. W. (2001). Overt versus covert competition in Soay sheep. Trends in Ecology and Evolution, 16, 279–280. Ebensperger, L. A. & Blumstein, D. T. (2006). Sociality in New World hystricognath rodents is linked to predators and burrow digging. Behavioral Ecology, 17, 410–418. Eisenberg, J. F. (1966). The social organizations of mammals. Handbuch der Zoologie, 10, 1–92. Eisenberg, J. F. (1981). The Mammalian Radiations: an Analysis of Trends in Evolution, Adaptation, and Behavior. Chicago, IL: University of Chicago Press. Ellis, L. (1995). Dominance and reproductive success among nonhuman animals: a cross-species comparison. Ethology and Sociobiology, 16, 257–333. Emlen, S. T. (1984). Cooperative breeding in birds and mammals. In J. R. Krebs & N. B. Davies, eds., Behavioural Ecology: An Evolutionary Approach, 2nd edn. Oxford: Blackwell, pp. 305–339.
The effects of heterogeneous regimes on reproductive skew Emlen, S. T. & Oring, L. W. (1977). Ecology, sexual selection, and the evolution of mating systems. Science, 197, 215–223. Engh, A. L., Funk, S. M., Van Horn, R. C., et al. (2002). Reproductive skew among males in a female-dominated mammalian society. Behavioral Ecology, 13, 193–200. Fedigan, L. M. & Pavelka, M. S. M. (2001). Is there adaptive value to reproductive termination in Japanese macaques? A test of the maternal investment hypothesis. International Journal of Primatology, 22, 109–125. Feh, C. (1990). Long-term paternity data in relation to different aspects of rank for Camargue stallions, Equus caballus. Animal Behaviour, 40, 995–996. Feldhamer, G. A., Drickamer, L. C., Vessey, S. H., & Merritt, J. F. (2004). Mammalogy: Adaptation, Diversity, Ecology, 2nd edn. Boston, MA: McGraw-Hill. Flack, J. C., Girvan, M., de Waal, F. B., & Krakauer, D. C. (2006). Policing stabilizes construction of social niches in primates. Nature, 439, 426–429. Fleming, T. H., Breitwisch, R., & Whitesides, G. H. (1987). Patterns of tropical vertebrate frugivore diversity. Annual Review of Ecology and Systematics, 18, 91–109. Foster, K. R. & Ratnieks, F. L. W. (2005). A new eusocial vertebrate? Trends in Ecology and Evolution, 20, 363–364. Frank, S. A. (1995). Mutual policing and repression of competition in the evolution of cooperative groups. Nature, 377, 520–522. Frank, S. A. (2003). Perspective: repression of competition and the evolution of competition. Evolution, 57, 693–705. Frank, S. A. (2006). Social selection. In C. W. Fox & J. B. Wolf, eds., Evolutionary Genetics: Concepts and Case Studies. Oxford: Oxford University Press, pp. 350–363. Gachot-Neveu, H., Petit, M., & Roeder, J. J. (1999). Paternity determination in two groups of Eulemur fulvus mayottensis: implications for understanding mating strategies. International Journal of Primatology, 20, 107–119. Garnier, J. N., Bruford, M. W., & Goossens, B. (2001). Mating system and reproductive skew in the black rhinoceros. Molecular Ecology, 10, 2031–2041. Gerlach, G. & Bartmann, S. (2002). Reproductive skew, costs, and benefits of cooperative breeding in female wood mice (Apodemus sylvaticus). Behavioral Ecology, 13, 408–418. Gilchrist, J. S., Otali, E., & Mwanguhya, F. (2004). Why breed communally? Factors affecting fecundity in a communal breeding mammal: the banded mongoose (Mungos mungo). Behavioral Ecology and Sociobiology, 57, 119–131. Gingerich, P. D. (2006). Environment and evolution through the Paleocene–Eocene thermal maximum. Trends in Ecology and Evolution, 21, 246–253. Gockel, J. & Ruf, T. (2001). Alternative seasonal reproductive strategies in wild rodent populations. Journal of Mammalogy, 82, 1034–1046. Goldizen, A. W. (1988). Tamarin and marmoset mating systems: unusual flexibility. Trends in Ecology and Evolution, 3, 36–40. Greenwood, P. J. (1980). Mating systems, philopatry, and dispersal in birds and mammals. Animal Behaviour, 28, 1140–1162. Gros-Louis, J., Perry, S., & Manson, J. H. (2003). Violent coalitionary attacks and intraspecific killing in wild white-faced capuchin monkeys (Cebus capucinus). Primates, 44, 341–346.
107
108
C. B. Jones Hackla¨nder, K., Mo¨stl, E., & Arnold, W. (2003). Reproductive suppression in female Alpine marmots, Marmota marmota. Animal Behaviour, 65, 1133–1140. Hager, R. (2003a). Models of reproductive skew applied to primates. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 65–101. Hager, R. (2003b). The effects of dispersal costs on reproductive skew and withingroup aggression in primate groups. Primate Report, 67, 85–98. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Hamilton, W. D. (1964). The evolution of social behavior. Journal of Theoretical Biology, 7, 1–52. Heinze, J. & Keller, L. (2000). Alternative reproductive strategies: a queen perspective in ants. Trends in Ecology and Evolution, 15, 508–512. Helms Cahan, S. (2001). Co-operation and conflict in ant foundress associations: insights from geographical variation. Animal Behaviour, 61, 819–825. Hewitt, G. M. & Butlin, R. K. (1997). Causes and consequences of population structure. In J. R. Krebs & N. B. Davies, eds., Behavioural Ecology: an Evolutionary Approach. Oxford: Blackwell, pp. 350–372. Hill, S. E. & Reeve, H. K. (2005). Low fertility in humans as the evolutionary outcome of snowballing resource games. Behavioral Ecology, 16, 398–402. Hoffman, J. I., Boyd, I. L., & Amos, W. (2003). Male reproductive strategy and the importance of maternal status in the Antarctic fur seal Arctocephalus gazelle. Evolution, 57, 1917–1930. Hsu, M., Bhatt, M., Adolphs, R., Tranel, D., & Camerer, C. F. (2005). Neural systems responding to degrees of uncertainty in human decision-making. Science, 310, 1680–1683. Huck, M., Lottker, P., Bohle, U. R., & Heymann, E. W. (2005). Paternity and kinship patterns in polyandrous moustached tamarins (Saguinus mystax). American Journal of Physical Anthropology, 127, 449–464. Isler, K. & van Schaik, C. P. (2006). Metabolic costs of brain size evolution. Biology Letters, 2, 557–60. Jack, K. M. & Fedigan, L. M. (2006). Why be alpha male? Dominance and reproductive success in wild white-faced capuchins (Cebus capucinus). In A. Estrada, P. A. Garber, M. Pavelka, & L. Luecke, eds., New Perspectives in the Study of Mesoamerican Primates: Distribution, Ecology, Behavior, and Conservation. New York, NY: Springer, pp. 367–386. Jansen, V. A. A. & Stumpf, P. H. (2005). Making sense of evolution in an uncertain world. Science, 309, 2005–2007. Jennions, M. D. & Petrie, M. (1997). Variation in mate choice and mating preferences: a review of causes and consequences. Biological Reviews, 72, 283–327. Jerison, H. J. (1983). The evolution of the mammalian brain as an informationprocessing system. In J. F. Eisenberg & D. G. Kleiman, eds., Advances in the Study of Mammalian Behavior. Shippensburg, PA: American Society of Mammalogists, pp. 113–146.
The effects of heterogeneous regimes on reproductive skew Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Jones, C. B. (1980). The functions of status in the mantled howler monkey (Alouatta palliata Gray): intraspecific competition for group membership in a folivorous Neotropical primate. Primates, 21, 389–405. Jones, C. B. (1985). Reproductive patterns in mantled howler monkeys: estrus, mate choice, and copulation. Primates, 26, 130–142. Jones, C. B. (1995). Howler monkeys appear to be preadapted to cope with habitat fragmentation. Endangered Species UPDATE, 12, 9–10. Jones, C. B. (1997). Life history patterns of howler monkeys in a time-varying environment. Boletin Primatologico Latinoamericano, 6, 1–8. Jones, C. B. (1999). Why both sexes leave: effects of habitat fragmentation on dispersal behavior. Endangered Species UPDATE, 16, 70–73. Jones, C. B. (2000). Alouatta palliata politics: empirical and theoretical aspects of power. Primate Report, 56, 3–21. Jones, C. B. (2002). Negative reinforcement in primate societies related to aggressive restraint. Folia Primatologica, 73, 140–143. Jones, C. B. (2004). The number of adult females in groups of polygynous howling monkeys (Alouatta spp.): theoretical inferences. Primate Report, 68, 7–25. Jones, C. B. (2005a). Behavioral Flexibility in Primates: Causes and Consequences. New York, NY: Springer. Jones, C. B. (2005b). Social parasitism in mammals with particular reference to Neotropical primates. Mastozoologı´a Neotropical, 12, 19–35. Jones, C. B. (2005c). A preliminary test of the Van Schaik model of male coalitions for Costa Rican mantled howler monkeys (Alouatta palliata). Laboratory Primate Newsletter, 44, July, 3–6. Jones, C. B. (2005d). A preliminary test of the Van Schaik model of male coalitions for Costa Rican mantled howler monkeys (Alouatta palliata). Laboratory Primate Newsletter, 44, www.brown.edu/primate/current.html. Jones, C. B. (2006). Exploring the utility of Van Schaik’s model of male coalitions for female mantled howler monkeys (Alouatta palliata). Laboratory Primate Newsletter, 45, www.brown.edu/primate/current.html. Jones, C. B. (2007). The evolution of exploitation in humans: “Surrounded by strangers I thought were my friends”. Ethology, 113, 499–510. Jones, C. B., & Agoramoorthy, G. (2003). Alternative reproductive behaviors in primates: towards general principles. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 103–139. Jones, C. B. & Corte´s-Ortiz, L. (1998). Facultative polyandry in the howling monkey (Alouatta palliate): Carpenter was correct. Boletin Primatologico Latinoamericano, 7, 1–7. Kappeler, P. M. & Pereira, M. E., eds. (2003). Primate Life Histories and Socioecology. Chicago, IL: University of Chicago Press. Kerth, G., Safi, K., & Ko¨nig, B. (2002). Mean colony relatedness is a poor predictor of colony structure and female philopatry in the communally breeding Bechstein’s bat (Myotis bechsteinii). Behavioral Ecology and Sociobiology, 52, 203–210.
109
110
C. B. Jones King, W. J. & Allaine´, D. (2002). Social, maternal, and environmental influences on reproductive success in female Alpine marmots (Marmota marmota). Canadian Journal of Zoology, 80, 2137–2143. King-Casas, B., Tomlin, D., Anen, C., et al. (2005). Getting to know you: reputation and trust in a two-person economic exchange. Science, 308, 78–83. Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 266, 571–578. Ku¨mmerli, R. & Martin, R. D. (2005). Male and female reproductive success in Macaca sylvanus in Gibraltar: no evidence for rank dependence. International Journal of Primatology, 26, 1229–1249. Kussell, E. & Leibler, S. (2005). Phenotypic diversity, population growth, and information in fluctuating environments. Science, 309, 2075–2078. Kutsukake, N. & Hasegawa, T. (2005). Dominance turnover between an alpha and a beta male and dynamics of social relationships in Japanese macaques. International Journal of Primatology, 26, 775–800. Lacey, E. A. & Sherman, P. W. (1997). Cooperative breeding in naked mole-rats: implications for vertebrate and invertebrate sociality. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 267–301. Le Galliard, J. F., Ferrie`re, R., & Dieckmann, U. (2005). Adaptive evolution of social traits: origin, trajectories, and correlations of altruism and mobility. American Naturalist, 165, 206–224. Lewontin, R. C. (1957). The adaptations of populations to varying environments. Cold Spring Harbor Symposium of Quantitative Biology, 22, 395–408. Lillegraven, J. A., Thompson, S. D., McNab, B. K., & Patton, J. L. (1987). The origin of eutherian mammals. Biological Journal of the Linnean Society, 32, 281–336. Matzel, L. D., Han, Y. R., Grossman, H., et al. (2003). Individual differences in the expression of a “general” learning ability in mice. Journal of Neuroscience, 23, 6423–6433. May, R. M. & Anderson, R. M. (1990). Parasite–host coevolution. Parasitology, 100, S89–101. Mayr, E. (1963). Animal Species and Evolution. Cambridge, MA: Belknap Press. McNab, B. K. (2006). The energetics of reproduction in endotherms and its implication for their conservation. Integrative and Comparative Biology, 46, 1159–1168. Meyers, L. A. & Bull, J. J. (2002). Fighting change with change: adaptive variation in an uncertain world. Trends in Ecology and Evolution, 17, 551–557. Nicolson, N. A. (1987). Infants, mothers, and other females. In B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham, & T. T. Struhsaker, ed., Primate Societies. Chicago, IL: University of Chicago Press, pp. 330–342. Nunn, C. L. (1999). The number of males in primate social groups: a comparative test of the socioecological model. Behavioral Ecology and Sociobiology, 46, 1–13.
The effects of heterogeneous regimes on reproductive skew O’Riain, M. J., Jarvis, J. U. M., Alexander, R., Buffenstein, R., & Peeters, C. (2000). Morphological castes in a vertebrate. Proceedings of the National Academy of Sciences of the USA, 97, 13194–13197. Packer, C., Gilbert, D., Pusey, A. E., & O’Brien, S. J. (1991). A molecular genetic analysis of kinship and cooperation in African lions. Nature, 351, 562–565. Packer, C., Tatar, M., & Collins, A. (1998). Reproductive cessation in female mammals. Nature, 392, 759–761. Packer, C., Pusey, A. E., & Eberly, L. E. (2001). Egalitarianism in female lions. Science, 293, 690–693. Piersma, T. & Drent, J. (2003). Phenotypic flexibility and the evolution of organismal design. Trends in Ecology and Evolution, 18, 228–233. Preston, B. T., Stevenson, I. R., Pemberton, J. M., & Wilson, K. (2001). Dominant rams lose out by sperm depletion. Nature, 409, 681–682. Pulliam, R. & Caraco, T. (1984). Living in groups: Is there an optimal group size? In J. R. Krebs & N. B. Davies, eds., Behavioral Ecology: an Evolutionary Approach. Oxford: Blackwell, pp. 122–147. Pusey, A., Williams, J., & Goodall, J. (1997). The influence of dominance rank on the reproductive success of female chimpanzees. Science, 277, 828–831. Queller, D. C. (1997). Why do females care more than males? Proceedings of the Royal Society of London B, 264, 1555–1557. Reader, S. M. & MacDonald, K. (2003). Environmental variability and primate behavioural flexibility. In S. M. Reader & K. M. Laland, eds., Animal Innovation. Oxford: Oxford University Press, pp. 83–116. Reeve, H. K. (2001). In search of unified theories in sociobiology: help from social wasps. In L. A. Dugatkin, ed., Model Systems in Behavioral Ecology: Integrating Conceptual, Theoretical, and Empirical Approaches. Princeton, NJ: Princeton University Press, pp. 57–71. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Keller, L. (1996). Relatedness asymmetry and reproductive sharing in animal societies. American Naturalist, 148, 764–769. Reeve, H. K., & Shen, S. -F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Rice, W. R. (2000). Dangerous liaisons. Proceedings of the National Academy of Sciences of the USA, 97, 12953–12955. Rilling, J., Gutman, D., Zeh, T., et al. (2002). A neural basis for social cooperation. Neuron, 35, 395–405. Rose, K. D. & Archibald, J. D., eds. (2005). The Rise of Placental Mammals. Baltimore, MD: Johns Hopkins University Press. Roughgarden, J. (1979). Theory of Population Genetics and Evolutionary Ecology: An Introduction. New York, NY: Macmillan.
111
112
C. B. Jones Rusu, A. S. & Krackow, S. (2004). Kin-preferential cooperation, dominance-dependent reproductive skew, and competition for mates in communally nesting female house mice. Behavioral Ecology and Sociobiology, 56, 298–305. Rutherford, S. L. (2000). From genotype to phenotype: buffering mechanisms and the storage of genetic information. Bioessays, 22, 1095–1105. Saj, T. L. & Sicotte, P. (2005). Male takeover in Colobus vellerosus at Boabeng-Fiema Monkey Sanctuary, central Ghana. Primates, 46, 211–214. Schoener, T. W. (1971). Theory of feeding strategies. Annual Review of Ecology and Systematics, 2, 369–404. Schradin, C. & Pillay, N. (2005). Intraspecific variation in the spatial and social organization of the African striped mouse. Journal of Mammalogy, 86, 99–107. Schultz, W. (2006). Behavioral theories and the neurophysiology of reward. Annual Review of Psychology, 57, 87–115. Setchell, J. M., Charpentier, M., & Wickings, E. J. (2005). Mate guarding and paternity in mandrills: factors influencing alpha male monopoly. Animal Behaviour, 70, 1105–1120. Shahnoor, N. & Jones, C. B. (2003). A brief history of the study of sexual selection and reproductive competition in primatology. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 1–43. Sherman, P. W., Jarvis, J. U., & Alexander, R. D., eds. (1991). The Biology of the Naked Mole-Rat. Princeton, NJ: Princeton University Press. Shultz, S. & Dunbar, R. I. M. (2005). Both social and ecological factors predict ungulate brain size. Proceedings of the Royal Society of London B, 273, 207–215. Solomon, M. G. & French, J. A. (1997). Cooperative Breeding in Mammals. Cambridge: Cambridge University Press. Sorin, A. B. (2004). Paternity assignment for white-tailed deer (Odocoileus virginianus): mating across age classes and multiple paternity. Journal of Mammalogy, 85, 356–362. Stockley, P. & Parker, G. A. (2002). Life history consequences of mammal sibling rivalry. Proceedings of the National Academy of Sciences of the USA, 99, 12932–12937. Strier, K. B. (1986). The behavior and ecology of the woolly spider monkey, or muriqui (Brachyteles arachnoids E. Geoffroy 1806). Unpublished Ph. D. dissertation, Harvard University. Strier, K. B. (1992). Atelinae adaptations: behavioral strategies and ecological constraints. American Journal of Physical Anthropology, 88, 515–524. Taborsky, M. (1994). Sneakers, satellites, and helpers: parasitic and cooperative behavior in fish reproduction. Advances in the Study of Behavior, 23, 1–100. Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell, ed., Sexual Selection and the Descent of Man, 1871–1971. Chicago, IL: Aldine, pp. 136–179. Vaughan, T. A. (1978). Mammalogy. Philadelphia, PA: Saunders. Vaughan, T. A., Ryan, J. M., & Czaplewski, N. J. (2000). Mammalogy, 2nd edn. Fort Worth, TX: Harcourt College Publishers. Vehrencamp, S. L. (1979). The roles of individual, kin, and group selection in the evolution of sociality. In P. Marler & J. Vandenbergh, eds., Handbook of Behavioral
The effects of heterogeneous regimes on reproductive skew Neurobiology: Social Behavior and Communication (Volume 3). New York, NY: Plenum, pp. 351–394. Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Vehrencamp, S. L. (2000). Evolutionary routes to joint-female nesting in birds. Behavioral Ecology, 11, 334–344. Voigt, C. C. & Streich, W. J. (2003). Queuing for harem access in colonies of the greater sac-winged bat. Animal Behaviour, 65, 149–156. Wang, G., Hobbs, N. T., Boone, R. B., et al. (2006). Spatial and temporal variability modify density dependence in populations of large herbivores. Ecology, 87, 95–102. Waynforth, D., Hurtado, A. M., & Hill, K. (1998). Environmentally contingent reproductive strategies in Mayan and Ache males. Evolution and Human Behavior, 19, 369–385. West, S. A., Murray, M. G., Machado, C. A., Griffin, A. S., & Herre, E. A. (2001). Testing Hamilton’s rule with competition between relatives. Nature, 409, 510–513. West, S. A., Pen, I., & Griffin, A. S. (2002). Cooperation and competition between relatives. Science, 296, 72–75. West-Eberhard, M. J. (1979). Sexual selection, social competition, and evolution. Proceedings of the American Philosophical Society, 123, 222–234. West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. Oxford, UK: Oxford University Press. Whitten, P. L., & Turner, T. R. (2004). Male residence and the patterning of serum testosterone in vervet monkeys (Cercopithecus aethiops). Behavioral Ecology and Sociobiology, 56, 565–578. Widdig, A., Bercovitch, F. B., Streich, W. J., et al. (2004). A longitudinal analysis of reproductive skew in male rhesus macaques. Proceedings of the Royal Society of London B, 271, 819–826. Wilbur, H. M. & Rudolf, V. H. W. (2006). Life-history evolution in uncertain environments: bet hedging in time. American Naturalist, 168, 398–411. Wilson, E. O. (1971). The Insect Societies. Cambridge, MA: Belknap Press. Wilson, E. O. (1975). Sociobiology: the New Synthesis. Cambridge, MA: Belknap Press. Wittenberger, J. F. (1980). Group size and polygamy in social mammals. American Naturalist, 115, 197–222. Wolff, J. O. & Macdonald, D. W. (2004). Promiscuous females protect their offspring. Trends in Ecology and Evolution, 19, 127–134. Worthington Wilmer, J., Allen, P. J., Pomeroy, P. P., Twiss, S. D., & Amos, W. (1999). Where have all the fathers gone? An extensive microsatellite analysis of paternity in the grey seal (Halichoerus grypus). Molecular Ecology, 8, 1417–1429. Wrangham, R. W. (1980). An ecological model of female-bonded primate groups. Behaviour, 75, 262–299. Young, A. J. & Clutton-Brock, T. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387.
113
5
Social skew as a measure of the costs and benefits of group living in marmots t h e a b . w a n g , pe t e r n o n a c s , a n d d a n i e l t . b l u m s t e i n
Summary In group-living animals reproduction is often skewed such that some group members reproduce more than others. In addition to reproductive skew, group members may also exhibit social skew, where some individuals show particular behaviors more often than others. Significant social skew in behaviors such as anti-predator defense or social interactions may influence survival and reproduction. Therefore, social skew has the potential to translate into reproductive skew and affect group productivity. We measured social skew across groups in a population of yellow-bellied marmots (Marmota flaviventris). Several behaviors such as agonistic interactions, affiliative interactions, and first emergence were significantly skewed in most groups. Alarm calling, however, was infrequently skewed more than would be expected by random chance. Thus, marmot groups do not appear to have behavioral roles in terms of individuals acting like sentinels. Although significant social skew was present, it did not obviously affect fitness as measured by female reproductive success for each group. However, skew in individual-directed behavior (e.g. agonistic and affiliative interactions) did significantly correlate with the level of reproductive skew. Finally, the results were independent of the scale at which groups were defined. Behavioral variability appears to occur similarly across the entire marmot population. The results of this study illustrate that the quantification of social skew has potential to be a powerful tool for understanding the evolution of sociality. Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
114
Social skew and group living in marmots Introduction Animal groups are characterized by variation in the degree to which individuals exhibit specific behaviors and variance across group members in the distribution of such variables has come to be called “skew” (Reeve & Ratnieks 1993, Keller & Reeve 1994, Ruzzante et al. 1995, Kokko & Lindstro¨m 1997, Kokko et al. 1999, Nonacs 2000). Within cooperative groups, research has largely focused on reproductive skew (Johnstone 2000, Magrath & Heinsohn 2000, Nonacs 2001, Reeve & Keller 2001). Indeed, the degree of reproductive skew has been proposed as a valuable evolutionary metric for comparing species (Sherman et al. 1995, Lacey & Sherman 2005). Such a eusociality index would differentiate between species by how strictly breeder and non-breeder roles are defined within groups. Similar levels of reproductive skew could therefore imply similar evolutionary pressures across taxonomically very different groups (e.g. ants similar to naked mole-rats and paper wasps similar to cooperatively breeding birds). This eusociality index for comparing species has been criticized by Costa & Fitzgerald (1996, 2005) on the basis that many species show complex cooperative behavior, but do not skew reproduction. However, these authors offer no mathematical solution to this problem, and it may be impossible as a practical matter to have a single skew index that measures variance simultaneously across various social behaviors and reproduction. Instead, Nonacs (2000, 2001) suggested that skew indices could be applied separately across reproductive and non-reproductive behaviors. The degree of skew in nonreproductive behaviors could correlate with outcomes for the group in survival, foraging, or reproduction. Thus, the degree to which individual group members have defined roles may have positive or negative consequences for group success in terms of survival and productivity. Here, we will extend the use of skew to study roles in yellow-bellied marmots (Marmota flaviventris; Figure 5.1), a moderately social ground-dwelling rodent (Frase & Hoffmann 1980). A “role” within a group is defined in this chapter by the presence of significant skew. For example, if the frequency of alarm calling in a group of marmots shows significant skew, this could imply that the more alert individuals are acting as sentinels. Conversely, a lack of skew would imply that there is no sentinel role within groups and all individuals are equally likely to watch for predators. Roles could exist for any activity with variance in events or productivity between individuals. The combination of all the roles will define the structure and patterns of relationships within the group.
115
116
T. B. Wang et al.
Figure 5.1 Yellow-bellied marmot pup, sex unknown. Courtesy of Daniel T. Blumstein.
Skew indices
Skew can be described in over 20 different mathematical ways (Kokko et al. 1999, Nonacs 2000), and, for purposes of comparing social groups, not all methods are equally valid (Nonacs 2003). Suppose we want to measure whether alarm-calling behavior in marmots is skewed. We record K alarm calls over some time period, and if there are N marmots in our defined group, the mean would be K/N alarm calls per marmot. The first problem in quantifying skew is that our null hypothesis of no skew is not an expectation that each marmot gives exactly K/N alarms. This would imply that marmots are somehow dividing up sentinel duty exactly equally, and this would require a behavioral explanation as much as would a finding of significant skew. Instead, our null expectation would be a random distribution around the mean, with some animals calling more and others calling less due to random chance. Therefore, a skew index must have a set value for a random distribution. A second problem arises if all marmots are not present for equal periods of time. If for instance, a marmot does not emerge from its burrow, it would not be expected to give an alarm call. Therefore, apparent skew in alarm calling could be an artifact of different times spent above ground. Any skew index that
Social skew and group living in marmots cannot adjust for differing times spent in the group is not reliably measuring behavioral differences and roles (Crespi & Yanega 1995). Nonacs (2000, 2003) compared a variety of skew indices and recommended the B index as the most useful and reliable under a wide range of assumptions. The B index was found to be sensitive to robust differences in skew and can compare groups with different productivities, sizes, and differential residence times. This allows us to account for random processes and differential survival or duration within a group. Therefore, the B index is resistant to bias due to group numerical characteristics. It was also determined to be more powerful and to have better-defined statistical properties than other skew indices (Nonacs 2000). The B index as we use it here is not a direct replacement for the eusociality index, but rather a way of measuring skew in the performance of any social behavior. The B index calculates variance across i individuals in a group, R(pi – ni/N)2, where pi is the proportion of total events performed or benefits received by the i-th individual and ni is the time the i-th individual spent in the group. Differential time spent in the group is handled by changing N from the number of individuals to the total time spent in the group across all individuals. Thus, variance is the observed proportion of all acts across all individuals, minus their proportional contribution to the total group time. A random distribution has a positive variance, the magnitude of which varies with K and N. Fortunately, we can estimate a random distribution of K events across N individuals as following a binomial distribution (Sokal & Rohlf 1995). The B index, therefore, is the observed variance minus the expected binomial variance, and would equal zero if the behavioral activity was distributed randomly. Social skew
To study behavioral roles in social groups, the group itself must first be defined. Groups can be determined according to geographic boundaries, but these boundaries and therefore group composition may depend on the scale at which interactions occur (e.g. alarm calls can potentially connect individuals that never physically interact). Thus, one individual can be part of many different spatially nested groups. For example, marmots that live in a meadow can be considered as a group. However, within the meadow, members may share different burrow systems and at a finer scale some individuals may be associated with specific burrows. Which geographic level to use for study may depend on the question being asked and its scale of biological relevance. Returning to our example of alarm calling, we could use a definition of a group that contains the active space of a call (i.e. all individuals within earshot), or we could focus on the set of subjects that could both hear and see a potential
117
118
T. B. Wang et al. caller. This distinction is important for the specific benefits that subjects can obtain. Potentially more information might be transmitted by a multimodal signal (i.e. seeing and hearing the caller: Partan & Marler 2005). Biologically meaningful group boundaries, however, may become apparent only after experimental observation when statistical patterns emerge. These patterns could identify group membership from similar individual responses to detected signals or resulting fitness consequences. For any given defined group, it is possible to calculate skew in behavior. There are no a-priori predictions that marmot groups with well-defined behavioral roles would be more or less efficient, in terms of productivity and survival, than groups without defined behavioral roles. Certainly there will be trade-offs in time allocation. If individuals spend less time in social interactions they could allocate more time to foraging, grooming, or other forms of self-maintenance (e.g. K. A. Pollard & D. T. Blumstein, unpublished data). Energy gain and injury avoidance could come from reduced territory defense, when only a few individuals fight off transient individuals. In addition, animals that engage in the majority of the activity could draw predators away from the rest of the group. The rest of the group may be able to forage and rest unnoticed by predators. On the other hand, the existence of behavioral roles could cause group productivity to decline. One individual performing more aggressive interactions could disrupt the other members. An individual dominating all behavioral activity could suppress other group members and prevent them from engaging in personally beneficial activities. In this way, behavioral roles in non-reproductive behaviors could have a strong influence on mating and reproductive skew. A few aggressive individuals could prevent others from gaining enough nutrition to be able to reproduce, or even prevent contact with members of the opposite sex. Do behavioral roles have fitness consequences, and are they evolutionarily important? This might depend on whom they affect and the level of the skew. For example, if juveniles, but not adult females, have structured behavioral roles, these roles might have less of an impact on offspring production. On the other hand, if group success is limited by juvenile survival, then behavioral roles may have more fitness consequences. Roles could also vary in the degree of the consequences of skew. Behavioral roles within groups where dominants completely suppress subordinates from grooming and foraging activities would have a large fitness effect. Groups could also compensate for the effects of behavioral roles. If dominant individuals are able to prevent subordinates from reproducing but they end up having more offspring themselves, total group productivity would stay the same.
Social skew and group living in marmots In this chapter we will examine yellow-bellied marmot reproductive and non-reproductive behavior for evidence of skewed social roles. We will use the B index as a metric for suggesting roles within groups and to identify potential fitness consequences of behavioral skew.
Methods Study animals and study site
Yellow-bellied marmots, a semi-fossorial ground-dwelling sciurid rodent, were studied in the upper East River valley near the Rocky Mountain Biological Laboratory, Colorado, USA (Figure 5.2). This population has been observed continuously since 1962 (Armitage 1991, Ozgul et al. 2006) and social groups subdivide most colony sites (Blumstein et al. 2006). Each year of the study, all subjects were live-trapped and marked (with fur dye to facilitate observations from afar, and with ear tags for permanent identification). Subjects were observed from a distance known not to influence their behavior (this varied by site and ranges from 50 to 200 m), through 8 · 40 binoculars and 15–45 · spotting scopes, during morning (06:30–10:00) and afternoon (16:00– 19:00) activity bouts. All-event recording (Martin & Bateson 1993) quantified social interactions (occurring about once every 20 minutes), and alarm calling
Figure 5.2 Example of yellow-bellied marmot habitat at the Rocky Mountain Biological Laboratory, Colorado. The shown site is called “Picnic” where groups P18, P20, and P21 can be found. Courtesy of Lucretia Olson.
119
120
T. B. Wang et al. (which is much less frequent). For the following analysis of social skew, we focus on data collected in the 2004 active season (April through August). Group composition
Social groups were defined by calculating a simple ratio association index (Cairns & Schwager 1987). The ratio calculates the proportion of observations in which individuals are seen together (at the same burrow entrance) as a fraction of all times they were seen both together and alone. We used the entire set of observations and trap locations from 2004 to calculate ratios. We measured skew across a range of association levels (0.1, 0.3, 0.5, and 0.7). The low 0.1 association level compared groups that included individuals observed in the same general geographic colony (which consists of several burrows) and the higher levels are increasingly smaller subdivisions within the colony. Different association levels allow us to examine behavioral roles at different scales in the population. We used a 0.5 association level to compare skew across different age–sex classes. This level is appropriate for individuals known to share burrows, and has been used in previous studies (Baird & Whitehead 2000, Nanayakkara & Blumstein 2003). The age–sex classes that we considered included: all marmots, all marmots excluding juveniles, adult females and yearling females, adult males and yearling males, only adult females, and only adult males. We used SOCPROG 2.2 (Whitehead 2004), a program for analyzing social structure, to calculate association indices and plot dendrograms that helped us identify social groups. Group names consist of a one- or two-letter prefix that indicates a geographic colony and a specific number. Non-reproductive behaviors
We analyzed skew in individual-directed and group-directed behaviors. Individual-directed behaviors are defined as those that have a clear recipient, such as allogrooming or biting. We grouped individual-directed behaviors into two categories: (1) affiliative or positive, or (2) agonistic or negative (Table 5.1). In contrast, group-directed behaviors may have no targeted individual and potentially a large number of recipients. Examined group-directed behaviors were alarm calling and first emergence. Alarm calls have a demonstrable conspecific warning function (Blumstein et al. 1997). Each time a marmot initiated a bout of alarm calling, we recorded the number of individual calls within a bout, the duration of the bout, and the likely cause for the alarm. For analysis, we used bouts of alarm calls as a measure rather than the number of separate calls within a bout, because these are likely to be directed at the same stimulus. Bouts were considered separate if at least 1 minute elapsed between calls.
Social skew and group living in marmots Table 5.1 Individual-directed behaviors Affiliative (positive)
Agonistic (negative)
Follow another marmot
Posture aggressively
Forage with another marmot
Vocalize aggressively
Lie down with another marmot
Snarl
Greet
Snap
Sniff
Hiss
Play
Displace
Groom
Fight
Nurse Mate
The first marmot to emerge from the burrow each day can be thought of as a “scout.” This marmot incurs an extra degree of predation risk because it may be the first to be detected by a lurking predator. All subsequent marmots potentially gain information from what happens to the first marmot through group eavesdropping. Because the visibility and total activity levels during the end of the summer are reduced (thus making it difficult in our subalpine system for eavesdroppers to benefit), we only looked at emergences before July 1, 2004. First emergence was recorded during observations begun before 07:30 hours, and only when the observer could clearly identify which marmot emerged first. Skew calculation
Skew was calculated using the B index as described in Nonacs (2000). It determines whether K events are spread randomly over i number of individuals in a group. For affiliative and agonistic social behaviors and alarm calls, we transformed for each marmot its observed number of social behaviors (ki) into its proportional contribution to the total number of social interactions (ki/K ¼ pi). Because individuals were observed for different lengths of time, group size was defined relative to the total number of minutes that all individuals were observed (Nt), such that weighted mean group size equals Nt/nmax, where nmax is the maximum time any individual could be present (often equal to the length of time the group was observed). Therefore, the expected proportion of K events performed by the i-th individual is the number of minutes it was observed, divided by the total time (¼ ni/Nt). Observed variance is the sum across all marmots of (pi ni/Nt)2.
121
122
T. B. Wang et al. To calculate the B index value for a group we subtracted the expected variance from the observed variance. The expected variance follows the binomial distribution such that it equals (1 – 1/N)/K. For first emergence, Nt was total number of animals known to be present in a group (independent of whether they did or did not emerge on a particular day), and all ni equalled one. The significance of skew for each group was determined by simulation. We assigned a probability level that the observed B value was due to random chance for each group. All B index values and their significance levels were calculated using Skew Calculator (available at: www.eeb.ucla.edu/Faculty/ Nonacs). Statistically significant skew as measured by the B index values implies that marmots within a group have different behavioral propensities that might be considered behavioral roles. Fitness consequences
The next question is whether strongly defined roles within groups have fitness consequences. Therefore, we correlated observed social skew with reproductive skew (measured again with the B index) and overall reproductive success of groups. Individual reproductive success was measured by the number of pups born to each female. For mixed litters, we calculated the average number of pups per associated adult female (Armitage 2004). Group reproductive success was measured by the total number of pups divided by the total number of adult females in the group. The range of the B index increases with larger group size (N) and both N and K affect the absolute values that represent complete skew and completely equal sharing (see Nonacs 2000 for details on calculating the potential minimum and maximum B index values). Therefore, we cannot use them for across-group comparisons. Instead, we converted the B index values to standardized values by dividing positive and negative B values by the absolute values of the maximum and minimum possible B value (Nonacs 2000). This creates a parameter range of 1 to 1. The adjusted B index should be used with caution because it tends to overweight distributions with less than random variance (Nonacs 2003). For our purposes, however, this does not represent a problem as we do not use the adjusted B index to determine deviations from random expectations. Results
When groups were defined using association indices ranging from 0.1 to 0.7, we found no substantial differences in the proportion of groups that indicated significant skew. This was found for both group-directed and individual-directed behaviors (Figure 5.3). Higher association levels subdivided the
Social skew and group living in marmots
Proportion of significant results
(a)
1
5
Alarm Calling
5
First Emergence
0.8
0.6 7 6 0.4
0.2
9
10
6
11
0 0.1
(b)
0.3 0.5 Association level value
1
Affiliative 7
Proportion of significant results
6 0.8
0.7
Agonistic 8
6 8
9
9 11
0.6
0.4
0.2
0 0.1
0.3 0.5 Association level value
0.7
Figure 5.3 Proportion of significantly skewed groups at each of the four levels of association. The total number of groups for each behavior at that association level is indicated at the top of each bar. (a) Group-directed behavior; (b) Individual-directed behavior.
population into more groups but similar proportions of them were significantly skewed. Not all groups for the four behaviors were used because we could not calculate skew for groups with only 1 member or only 1 event. We
123
124
T. B. Wang et al. found consistently significant skew in first emergence, affiliative, and agonistic social behaviors. At least 60% of the groups at each association level were significantly skewed in these three behaviors. Alarm calling was much less often significantly skewed. A detailed analysis at the 0.5 level of association shows that skew for all behaviors was consistent across different demographic and age–sex classes (Table 5.2). In general, groups that were significantly skewed across all marmots were also skewed when juveniles were excluded from the analysis, and when only one sex was included. In addition, neither age nor sex was a significant predictor for which marmots were the most likely to emerge first in the morning (Table 5.3). Skew in individual or group-directed behaviors did not significantly correlate with a group’s reproductive success. Reproduction itself, however, was significantly skewed across females in four of the seven groups, and marginally skewed in another (Table 5.4). Two groups, however, shared reproduction more equally than would have been predicted by chance, and across all seven groups reproduction was more skewed than expected by chance. There was no significant relationship between a group’s reproductive skew and the mean number of offspring produced per adult female in the group (Figure 5.4), nor did reproductive skew predict the success of individual females. Skew in both individual-directed social interactions was significantly associated with reproductive skew (Figure 5.5). Group P20, however, is an outlier in these comparisons, and if excluded, neither relationship remains significant. Although both individual-directed behaviors exhibited significant levels of skew, there was no significant correlation between the skew of affiliative and the skew of agonistic social behaviors (Figure 5.6). Skew in neither of the group-directed behaviors significantly correlated with reproductive skew.
Discussion Within a social group or neighborhood, functional roles can result from individuals that are recognizably different in their personalities and propensities towards certain actions (Bouchard & Loehlin 2001, Sih et al. 2004). While the majority of the observed yellow-bellied marmot groups exhibited significant skew across several behaviors, the functional implications of this skew were not dramatic. Sentinels and scouts
We found that in almost all groups no sentinel role could be detected through a differential likelihood to issue alarm calls. This result parallels the
125
126
Table 5.2 Skew in behaviors at the 0.5 level of association. For some groups, there were not enough observed behaviors to calculate B values Alarm calling Group All marmots
N
BR10
11
B 0.0077
Scouting P-level
B
0.735
0.0396
P-level 0.034
Social
Social
affiliative
agonistic
B
P-level
B
P-level
0.0106
<0.001
0.0752
<0.001
BR12
4
0.2536
0.283
BR14
48
0.0005
0.486
0.0717
<0.001
0.0274
<0.001
0.0903
<0.001
M5
30
0.0296
0.076
0.0844
0.001
0.0218
<0.001
0.0582
<0.001
P18
22
0.0155
0.333
0.1236
0.057
0.0964
<0.001
0.0723
0.012
P20
15
0.2484
0.030
0.0257
0.841
0.3849
<0.001
P21
22
0.0352
0.769
0.0772
<0.001
0.0169
0.107
T2
5
0.0517
0.483
0.5178
0.033
0.4724
0.005
0.0228
0.014
0.0277
0.094
T8
10
0.4080
0.003
T9
13
0.0107
0.699
0.0373
0.054
BR10
11
0.0077
0.714
0.0396
0.028
BR12
3
0.1903
0.385
0.0897
0.104
0.0303
<0.001
All adults and yearlings 0.0106
<0.001
0.0752
<0.001
0.0342
<0.001
0.0844
<0.001
BR14
20
0.0041
0.545
0.0436
0.004
M5
20
0.0318
0.113
0.0689
0.003
0.0265
<0.001
0.0516
<0.001
P18
13
0.0234
0.202
0.0985
0.109
0.1089
<0.001
0.0679
0.008
P20
6
0.2518
0.033
0.0289
0.851
0.3837
0.003
P21
12
0.0309
0.658
0.0792
<0.001
0.0161
0.117
0.9479
0.024
T2
2
0.0160
0.607
T8
2
0.0409
0.382
T9
13
0.0107
All adult females and
N
0.0043
0.298
0.061
0.0228
P-level
B
0.0373
P-level
B 0.1353
0.015
0.0162
0.010
0.0385
0.036
0.0771
0.003
0.0570
<0.001
0.0525
<0.001
P-level
0.0303
<0.001
0.697
B
B
0.006 P-level
yearling females BR10
7
BR14
15
0.0054
0.419
M5
7
0.0082
0.421
P18
9
P20
3
0.0714
0.101
0.0227
0.004
0.2159
<0.001
0.0711
0.25
0.0570
0.006
0.0566
0.046
0.0025
0.559
P21
8
0.0501
0.772
0.0186
0.203
0.0302
0.007
0.0129
0.227
T9
5
0.0328
0.801
0.0556
0.269
0.0617
0.003
0.0288
0.237
BR10
4
0.0200
0.662
0.0247
0.583
0.0034
0.793
0.0640
0.001
BR14
5
0.1530
0.760
0.0327
0.291
0.0175
0.004
0.3705
<0.001
0.0294
0.565
0.1785
0.048
0.0321
<0.001
0.0291
0.011
0.0738
<0.001
All adult males and yearling males
M5
13
P18
4
P20
3
P21
4
T9
8
0.3259 0.0134
0.0414
0.713
0.5179
0.006
0.1667
1.000
0.1364
<0.001
0.1637
0.073
0.0875
0.045
0.0318
<0.001
0.0077
0.224
0.130 0.362
Social skew and group living in marmots Table 5.3 The distribution of the number of observed first-emergence events across marmots within groups, by age–sex class (no. first / no. in class) Group
Adult females
Adult males
Yearling females
Yearling males
BR10
1/3
2/1
11/4
7/2
BR14
14/8
2/3
7/6
5/2
M5
4/3
3/3
6/4
2/10
P18
5/8
0/1
0/1
0/3
P21
10/7
2/2
1/1
1/2
T9
2/3
0/2
4/2
10/6
Total
36/32
9/12
29/18
25/25
Table 5.4 Reproductive skew among adult females Group
Females
Pups
B value
P
BR10
3
12
0.1111
0.064
BR14
8
31
0.0018
0.447
M5
3
9
0.1481
0.040
P18
8
13
0.0514
0.015
P20
3
12
0.0417
0.927
P21
7
12
0.2163
<0.001
T9
3
14
0.1293
0.022
Means
5.0
14.7
0.0875
<0.001
5.0 Mean pups per female
4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 –1
–0.8
–0.6
–0.4
–0.2
0
0.2
0.4
Reproduction B value (adjusted) Figure 5.4 Mean number of pups per female for each group relative to the adjusted B index values (P ¼ 0.526).
127
T. B. Wang et al.
Reproduction B value (adjusted)
0.2 0.1 0 –0.1 –0.2 –0.3 –0.4 –0.5
P20
–0.6 –1
–0.8
–0.6
–0.4
–0.2
0
0.2
0.4
Affiliative B value (adjusted) 0.4 Reproduction B value (adjusted)
128
0.2 0.0 –0.2 –0.4 –0.6 –0.8 –1.0 –0.1
P20 0.0
0.1 0.2 0.3 0.4 Agonistic B value (adjusted)
0.5
0.6
Figure 5.5 Skew in behavior versus skew in reproduction: (a) P ¼ 0.0023 and (b) P ¼ 0.0003 with all the data points. If the outlier point (group P20 in both panels) is removed, there is no significant relationship.
finding of Blumstein et al. (1997) that most age–sex classes call at similarly low rates. In contrast, however, the majority of groups had individuals that were significantly likely to act as scouts (defined as the first marmot to emerge from its burrow in the morning). Emerging first is a risky behavior as this individual would be the first to encounter a waiting predator (predators were observed outside marmot burrows early in the morning), and other marmots would gain information from the experience of the first one out. Overall, which marmots had a propensity to emerge first was not significantly predicted by age or sex. Thus, a tendency to act as a scout may reflect consistent behavioral differences in the personalities of marmots relative to their risk
Social skew and group living in marmots 0.50 0.45 Agonistic B value
0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 –0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
Affiliative B value Figure 5.6 The relationship between index values for agonistic and affiliative behaviors.
taking (Blumstein et al. 2004). In contrast, Armitage et al. (1996) found that adult males spent more time above ground than other age–sex classes. This would suggest that adult males act more often as scouts and have increased opportunities to use other marmots as scouts. Whether scouting truly has a social role requires further study. Individual-directed versus group-directed behaviors
Behavioral skew was more pronounced across groups in individualdirected behaviors than in group-directed behaviors (Figure 5.3). This also suggests stable personality differences across individuals in willingness to positively or negatively interact. These differences are not created by different patterns of interaction across the sexes, as both females and males show the same skew patterns (Table 5.2). We cannot rule out, however, that relatedness differences rather than personality differences play a role in creating skew in individual-directed behavior. We currently do not have precise coefficients of relatedness calculated for the observed animals. Therefore, particularly affiliative pairs may be close relatives, and agonistic pairs may be genetically distant or unrelated. A mixture of such pairs within a group could create an overall skewed pattern of behavior. A pattern of interaction that is based primarily on relatedness would, however, also predict similar and correlated levels of skew across groups in affiliative and agonistic acts. Such a correlation does not exist between these two measures (Figure 5.6), which suggests that aggressive and cooperative behavior may vary separately across individual
129
130
T. B. Wang et al. yellow-bellied marmots. Therefore, at least one of the traits must be independent of genetic relatedness. The results described in this study are not highly sensitive to the size of the defined group as categorized by association index. This implies that the underlying factors affecting social skew are scale-independent. In other words, similar skew at low and high association levels implies that behavioral variability in personality appears to be the same within groups as it is across groups. This was unexpected, because some behaviors are limited to individuals that physically interact (e.g. affiliative behaviors), while others (e.g. alarm calling) can affect individuals with whom the caller has no physical relationship. Lower association indices imply a reduced opportunity for physical interaction. Given variation in the nature of who could be influenced by a behavior, we expected that social skew would be influenced by our definition of group. We also found that within colonies there was no evidence of individuals associating with others into groups that have similar behavioral patterns. It is possible that behavioral roles are constrained in some way, and that groups cannot be that different from each other. Further studies are required to elucidate why social behaviors appear to be scale-independent. Functional implications
We found no relationship between either behavioral or reproductive skew and per capita fitness of females in terms of reproductive success. The degree to which a group had defined behavioral roles did not predict either more or less reproduction. Therefore, within the population we studied there is no evidence for strong selection operating to favor more advanced levels of sociality with pronounced social and reproductive skew. Yellow-bellied marmots appear stably ensconced at an intermediate level of social complexity (Nonacs 2001, Helms Cahan et al. 2002), a finding that parallels other classifications of this species (Armitage 1981, Michener 1983, Blumstein & Armitage 1998). We are unaware of other studies with vertebrates or social insects that have correlated social skew to group success, but the results from other species and populations would be interesting for comparisons. We found two interesting relationships between behavioral skew and reproductive skew. Reproductive skew was positively correlated with affiliative skew and negatively with agonistic skew (Figure 5.5). This suggests that when reproductive skew occurs, it is not created through agonistic actions of a very dominant individual. It seems possible that because the existence of skew in different behaviors is not due to one particular individual, patterns of social skew in groups would be more resistant to changes in group composition. However, it should be noted that the results for reproductive, affiliative, and
Social skew and group living in marmots agonistic skew are strongly influenced by the presence of one group (P20). Our dataset is currently not large enough to determine whether this group is a true outlier, or indicative of a more robust relationship. However, the analysis has identified this one group as behaving in a very different manner from others. Therefore, it will be interesting to pursue why this group appears to have different dynamics within its interactions. Following the entire RMBL population for several more years will allow us to answer this question. Overall, we did not find a large effect of behavioral roles on group productivity. It is possible that our productivity measurement was insensitive to effects of behavioral roles. This could be taken into account by measuring weight gain or some other group factor. In this population of yellow-bellied marmots, productivity could be influenced more by past behavioral interactions or environmental factors, such as snowfall the year before, than by current behavioral interactions. We were also limited by a small number of groups and low event numbers for some behaviors. Increased size in both of these factors could possibly elucidate the effects of behavioral roles.
Conclusion In summary, the study of social skew and the identification of behavioral roles can lead to insight and clarification regarding social evolution and social complexity. Using social skew for comparisons avoids taxa-based terminology, which is an important step in identifying patterns of sociality in social insects and vertebrates (Brockmann 1997). We can look at groups with similar social skew and behavioral roles to identify common constraints between social species as well as look at the effect of common constraints. In this chapter, we have outlined the first steps of an approach to studying group behavior that can potentially be applied across numerous other taxa, and that may become a framework for comparison across groups that do not have clear reproductive skew (see Costa & Fitzgerald 2005). In regard to yellowbellied marmots, there are several future areas of research. For example, additional group- and individual-directed behaviors can be examined to see if the latter have consistently higher levels of skew. Also, the B index identifies the presence of behavioral roles within a group, but not the specific individuals responsible. More detailed analysis may identify predictive characters for particular behavioral types. Finally, it is of interest to be able to follow the same groups across years and investigate whether social skew changes over time and whether there are any fitness consequences. The application of skew metrics to behavioral phenomena may have a very informative future.
131
132
T. B. Wang et al. References Armitage, K. B. (1981). Sociality as a life-history tactic of ground squirrels. Oecologia, 48, 36–49. Armitage, K. B. (1991). Social and population dynamics of yellow-bellied marmots: results from long-term research. Annual Review of Ecology and Systematics, 22, 379– 407. Armitage, K. B. (2004). Lifetime reproductive success of territorial male yellowbellied marmots. Oecologia Montana, 13, 28–34. Armitage, K. B., Salsbury, C. M., Barthelmess, E. L., Gray, R. C., & Kovach, A. (1996). Population time budget for the yellow-bellied marmot. Ethology, Ecology and Evolution, 8, 67–95. Baird, R. W. & Whitehead, H. (2000). Social organization of mammal-eating killer whales: group stability and dispersal patterns. Canadian Journal of Zoology, 78, 2096–2105. Blumstein, D. T. & Armitage, K. B. (1998). Life history consequences of social complexity: a comparative study of ground-dwelling sciurids. Behavioral Ecology, 9, 8–19. Blumstein, D. T., Steinmetz, J., Armitage, K. B., & Daniel, J. C. (1997). Alarm calling in yellow-bellied marmots: II. Kin selection or parental care? Animal Behaviour, 53, 173–184. Blumstein, D. T., Runyan, A., Seymour, M., et al. (2004). Locomotor ability and wariness in yellow-bellied marmots. Ethology, 110, 615–634. Blumstein, D. T., Ozgul, A., Yovovich, V., Van Vuren, D. H., & Armitage, K. B. (2006). Predation risk predicts presence and persistence of yellow-bellied marmot colonies. Journal of Zoology London, 270, 132–138. Bouchard, T. J. Jr. & Loehlin, J. C. (2001). Genes, evolution, and personality. Behavior Genetics, 31, 243–273. Brockmann, H. J. (1997). Cooperative breeding in wasps and vertebrates: the role of ecological constraints. In J. C. Choe & B. J. Crespi, eds., Evolution of Social Behavior in Insects. Ithaca, NY: Cornell University Press: pp. 348–371. Cairns, S. J. & Schwager, S. J. (1987). A comparison of association indices. Animal Behaviour, 35, 1454–1469. Costa, J. T. & Fitzgerald, T. D. (1996). Developments in social terminology: Semantic battles in a conceptual war. Trends in Ecology and Evolution, 11, 285–289. Costa, J. T. & Fitzgerald, T. D. (2005). Social terminology revisited: Where are we ten years later? Annales Zoologici Fennici, 42, 559–564. Crespi, B. J. & Yanega, D. (1995). The definition of eusociality. Behavioral Ecology, 6, 109–115. Helms Cahan, S., Blumstein, D. T., Sundstrom, L., Liebig, J., & Griffin, A. (2002). Social trajectories and the evolution of social behavior. Oikos, 96, 206–216. Frase, B. A. & Hoffmann, R. S. (1980). Marmota flaviventris. Mammalian Species, 135, 1–8. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26.
Social skew and group living in marmots Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Kokko, H. & Lindstrom, J. (1997). Measuring the mating skew. American Naturalist, 149, 794–799. Kokko, H., Mackenzie, A., Reynolds, J. D., Lindstrom, J. & Sutherland, W. J. (1999). Measures of inequality are not equal. American Naturalist, 154, 358–382. Lacey, E. A. & Sherman, P. W. (2005). Redefining eusociality: concepts, goals and levels of analysis. Annales Zoologici Fennici, 42, 573–577. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Martin, P. & Bateson, P. (1993). Measuring Behaviour, 2nd edn. Cambridge: Cambridge University Press. Michener, G. R. (1983). Kin identification, matriarchies, and the evolution of sociality in ground-dwelling sciurids. In J. F. Eisenberg & D. G. Kleiman, eds., Advances in the Study of Mammalian Behavior: Stillwater, OK: American Society of Mammalogists, pp. 528–572. Nanayakkara, D. D. & Blumstein, D. T. (2003). Defining yellow-bellied marmot social groups using association indices. Oecologia Montana, 12, 7–11. Nonacs, P. (2000). Measuring and using skew in the study of social behavior and evolution. American Naturalist, 156, 577–589. Nonacs, P. (2001). A life history approach to group living and social contracts between individuals. Annales Zoologi Fennici, 38, 239–254. Nonacs, P. (2003). Measuring the reliability of skew indices: is there one best index? Animal Behaviour, 65, 615–627. Ozgul, A., Armitage, K. B., Blumstein, D. T., Van Vuren, D. H. & Oli, M. K. (2006). Effects of patch quality and network structure on patch occupancy dynamics of a yellow-bellied marmot metapopulation. Journal of Animal Ecology, 75, 191–202. Partan, S. R. & Marler, P. (2005). Issues in the classification of multimodal communication signals. American Naturalist, 166, 231–245. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflict in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Ruzzante, D. E., Hamilton, D. C., Kramer, D. L., and Grant, J. W. A. (1995). Scaling of the variance and the quantification of resource monopolization. Behavioral Ecology, 7, 199–207. Sherman, P. W., Lacey, E. A., Reevem, H. K., and Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108. Sih, A., Bell, A. M., Johnson, J. C., & Ziemba, R. E. (2004). Behavioral syndromes: An integrative overview. Quarterly Review of Biology, 79, 241–277. Sokal, R. R. & Rohlf, F. J. (1995). Biometry, 3rd edn. New York, NY: W. H. Freeman. Whitehead, H. (2004). SOCPROG: programs for analyzing social structure. Halifax, Nova Scotia: Department of Biology, Dalhousie University.
133
6
Explaining variation in reproductive skew among male langurs: effects of future mating prospects and ecological factors re inmar hager
Summary The partitioning of reproduction in a group, or reproductive skew, is affected by multiple ecological, genetic, and population parameters. The ways in which these factors interact to determine reproductive skew have been formally modeled, but few attempts have been made to assess the utility of skew theory in advancing our understanding of primate social systems and observed levels of skew. In the study presented here, I demonstrate how reproductive skew among male Hanuman langurs (Semnopithecus entellus) can be calculated using paternity data. In contrast to expectations, skew was found to increase with the number of males in a group. In a further analysis, I estimated skew for 21 populations of Asian langurs and investigated the ecological and population factors influencing variation in this parameter. The results lend support to model predictions that future mating prospects, as measured in chances of individual survival and number of potential mates, can be a significant predictor of skew. Furthermore, increased population density positively affected skew, which may reflect ecological constraints on dispersal for subordinates. These results suggest that skew theory may offer a useful framework for understanding the effects of multiple factors determining variation in reproductive skew.
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
134
Male langurs: future mating prospects and ecological factors Introduction One of the key aims of studies in evolutionary biology, behavioral ecology, and primatology is to analyze the factors influencing variation in individual fitness and associated behavioral traits (e.g. Clutton-Brock & Harvey 1979, Krebs & Davies 1993). The interactions between key genetic and ecological parameters and features of animal societies were formally modeled in the late 1970s in the pioneering work of Sandra Vehrencamp, who developed a model of the degree to which fitness can be biased among group members (Vehrencamp 1979, 1983a, 1983b). This skew in individual fitness was assumed to reflect an equilibrium between competition and cooperation in groups, and might be applied to categorize animal societies. In the following 20 years, the initial models were further developed, focusing on the bias in the distribution of reproduction among group members (Reeve & Ratnieks 1993, Emlen 1994, 1995, Keller & Reeve 1994), or reproductive skew. Skew theory investigates associated genetic and ecological correlates of reproductive skew and could thus be used to quantify features of sociality and compare levels of sociality between and within species (Shermann et al. 1995). This has prompted some initial enthusiasm that skew theory might be a good foundation upon which to build a framework to understand sociality, in particular where skew models focused on conditions for stable groups (Reeve & Ratnieks 1993, Keller & Reeve 1994, Reeve & Keller 2004, Reeve & Shen 2006). For example, a feature of group-living animals that can be directly predicted is the degree to which dominants monopolize reproduction (high skew) and under what conditions a more egalitarian distribution of reproduction is to be expected (low skew). In many skew models (to be precise, in transactional models: see below), one can make explicit predictions about the expected stable group size. Its implications for group living address the fundamental question of why individuals should remain in a group rather than disperse. Furthermore, group foraging has been analyzed using the skew framework (Hamilton 2000) as well as other characteristics of sociality such as within-group aggression (e.g. Reeve & Nonacs 1997, Reeve & Keller 1997, Cant & Johnstone 2000) and levels of infanticide (Johnstone & Cant 1999a, Hager & Johnstone 2004). Assuming that levels of reproductive skew reflect a way of quantifying sociality, we can then proceed to analyze the multiple factors that account for variation in this parameter. While key parameters have been identified in the early skew models (i.e. relatedness, group productivity, costs of solitary breeding), the original set has been extended to incorporate individual fighting abilities, costs of offspring production for individual group members, and the
135
136
R. Hager ability to commit discriminate or indiscriminate infanticide (Reeve & Nonacs 1997, Cant & Johnstone 1999, Johnstone & Cant 1999a, Hager & Johnstone 2004), allowing, on the one hand, a much more detailed investigation of skew tailored to different social systems. On the other hand, an ever-increasing number of more detailed models with sometimes disparate predictions (see e.g. Johnstone 2000, Reeve & Keller 2001) may render it more difficult to assess the utility of skew theory. Testing predictions of skew models in primates
Despite the relative popularity of skew theory among theoretical scientists, the majority of empirical studies with tests of model predictions were limited to social insects or birds (Vehrencamp 2000, Reeve & Keller 2001, Magrath et al. 2004), with only a few studies in mammalian systems (e.g. meerkats, Suricata suricatta: Clutton-Brock et al. 2001; naked mole-rats, Heterocephalus glaber: Faulkes & Bennett 2001; marmots, Marmota marmota: Allaine´ 2000). While the initial treatments on skew theory may be of a more technical nature, the paucity of data on paternity (which would allow determining levels of skew), and the focus on social insects have to some degree impeded potential applications of skew theory in the field of primatology. The question at issue is to determine whether our understanding of the ultimate and proximate causes of primate sociality can be advanced by skew theory. More specifically, we can ask if specific features of primate social behavior are better explained using skew theory than by other existing hypotheses (e.g. the priority of access model: Altmann 1962) or if any additional insight might be gained. One of the first attempts to make primatologists aware of the potential scope of skew theory for the study of primate social systems (Hager 2003a) was followed by a small number of studies investigating paternity data in primates and compared model predictions with empirical results to infer evolutionary causes of the observed distribution of reproduction in these species (e.g. common marmoset, Callithrix jacchus: Dietz 2004; mountain gorilla, Gorilla beringei: Bradley et al. 2005; rhesus macaques, Macaca mulatta: Widdig et al. 2004). In the following section, I will briefly review the main concepts of skew theory, summarize the key model predictions, and examine their application to primate systems. Models of reproductive skew
Reproductive-skew models may be classified into two basic categories, transactional and compromise models, while so-called synthetic models attempt to accommodate features of both transactional and compromise models (Johnstone 2000, Reeve 2000, Reeve & Keller 2001, Hager 2003a).
Male langurs: future mating prospects and ecological factors In transactional skew models it is assumed that a dominant, who is in full control of group reproduction, concedes a share of reproduction to subordinates because the presence of the latter increases the group’s (and thus the dominant’s) fitness (see e.g. Buston 2004) – in essence reflecting a “transaction” between dominant and subordinates. Subordinates are limited in dispersing by ecological constraints on breeding outside the group, and may accept a share of the group’s reproduction conceded by the dominant, who will offer just enough fitness benefits to make staying worthwhile for the subordinate. In theory, the share of reproduction offered by the dominant is the minimum share needed to retain the subordinate. Implicit in this argument is that subordinate reproduction occurs at the expense of the dominant’s share, weighted by the effects of relatedness between the two. This basic framework is referred to as a concession model. A variation on this treatment has been developed (the restraint model), in which it is assumed that dominants only control group membership but that reproduction is controlled by subordinates (Johnstone & Cant 1999b). By contrast, compromise models ignore constraints imposed by group stability and assume that control over group reproduction is incomplete. Rather than assuming a transaction between group members, each individual struggles in a kind of tug-of-war for its share, which is determined mainly by individual fighting abilities and relatedness between group members (Reeve et al. 1998, Johnstone 2000). In compromise models, ecological factors do not play a direct role in affecting skew. Finally, the subordinate is assumed to be less efficient in converting group resources into individual reproduction than the dominant. Skew models thus use a small set of key genetic and ecological variables (e.g. relatedness, cost of dispersal and offspring production, fitness of subordinates and dominants: see Reeve & Keller 2001, Hager 2003a) to explain, in a simplified manner, variation in levels of reproductive skew. For instance, concession models predict that skew increases with dispersal costs because the dominant concedes a lower share of reproduction due to the limited breeding options for subordinates outside the group. A comprehensive summary of the main model assumptions and predictions can be found in Hager (2003a), Johnstone (2000), and Reeve & Keller (2001). To date, two of three studies in primates that have used skew theory to analyze the bias in reproduction (rather than just investigating differences in individual reproductive success outside this framework) found some support for compromise models (Widdig et al. 2004, Bradley et al. 2005). By contrast, one study reported that neither transactional nor compromise models can adequately elucidate the diversity of social and mating systems in
137
138
R. Hager cooperatively breeding callitrichids (Dietz 2004). Work in a semi-free-ranging population of rhesus macaques on Cayo Santiago found no evidence that the pattern of reproductive skew followed the predictions of concession models (Widdig et al. 2004). For instance, distantly related males had a lower reproductive success than close relatives of the dominant male, which is in contrast to concession-model predictions. Further lack of support for concession models comes from a study in wild mountain gorillas, which showed that neither the dominant nor the subordinate male had full control over reproduction, suggesting a tug-of-war situation described by compromise models (Bradley et al. 2005). It should be noted, however, that none of the above studies set out to explicitly test predictions of compromise versus transactional models. Conclusions were based mainly on lack of support for some predictions of one class of models rather than a test between the two (see also Chapter 10 on testing predictions of skew models). Future prospects and skew
Skew theory has advanced over the years, and a number of new models have been developed, introducing additional parameters that are thought to affect reproductive skew, including the benefits of inheriting breeding status (Kokko & Johnstone 1999, Cant & Johnstone 1999, 2000, Hager & Johnstone 2004). In primatology, it has long been acknowledged that rank as well as chances of inheriting a dominant position affect individual reproductive success in both sexes (Cowlishaw & Dunbar 1991, Ellis 1995, Packer et al. 1995, Alberts et al. 2003). Kokko & Johnstone (1999) formally modeled the effects of this on the distribution of reproduction in a group. Skew is said to be influenced by the subordinate’s chances of inheriting breeding status at a later stage because in weighing the costs and benefits of leaving (and attempting to breed elsewhere) subordinates may remain in the group and “queue” for their turn. Under certain conditions, even if subordinates never obtain any concessions from the dominant, staying is still better because of the chances to breed later; skew will thus be lower than if the future prospects of mating were low. In other words, a subordinate’s decision to leave or remain in the group is likely to be influenced by its future prospects to reproduce (Stacey & Ligon 1987). Moreover, results of the model have shown that the delayed benefits of queuing may have a far greater effect on skew and group composition than other key parameters such as relatedness. In primates, this line of reasoning is supported by the observation of social queuing in wild savannah baboons (Papio cynocephalus). Investigating 32 group years, Alberts et al. (2003) showed that individuals may attain dominant status or higher ranks, often by waiting, and that this significantly affects the level of (mating) skew.
Male langurs: future mating prospects and ecological factors Given these results of empirical and theoretical studies, it seems crucial to incorporate measures of future mating prospects in analyses that seek to explain variation in reproductive skew in primates. From this inference follows the question: which factors are most likely to influence an individual’s future prospects? An individual’s prospects of gaining matings in the future will be determined first by its chances of survival, and second by the number of potential mates and competitors. Both factors, in turn, mutually depend on multiple ecological and genetic factors such as resource availability, and it is one key (and difficult) task in studies of skew to quantify these measures in a meaningful way. The study system: Hanuman langurs
Langurs appear to be a good system in which to explore whether skew theory can further our understanding of primate social systems and the underlying causes for variation in the distribution of reproduction. First, social organization is extremely diverse within and between species, male dispersal is high, and individuals are likely to face the trade-offs as described in skew models. Second, it is possible to obtain data on paternity (for one population at present: Ramnagar) as well as on other ecological and population variables, thus allowing an analysis of the factors influencing skew. Finally, many studies suggest that key assumptions of skew models are met, such as control over reproduction and the potential for group benefits due to subordinates’ presence, the potential for queuing, or the propensity to disperse. Hanuman langurs (Presbytis entellus) live in one-male or multi-male groups consisting of up to 120 individuals throughout the Indian subcontinent in habitats ranging from desert to mountains (Hrdy 1977, Winkler et al. 1984, Koenig & Borries 2001). Male dispersal patterns are characterized by dispersal from the natal group or secondary dispersal. Further, males may form all-male bands that attempt to take over other groups (Rajpurohit & Sommer 1991, Newton & Dunbar 1994, Borries 2000). While it is unverified whether the dominant male fully controls reproduction in multi-male groups, he certainly dominates the reproductive output of the group and is also most likely to interrupt subordinate matings (Sommer 1989, Newton & Dunbar 1994). The presence of subordinates in the group may confer several advantages to the dominant, for example in the context of predator detection and communal foraging. In addition, it has been observed that subordinate resident males may act as infant protectors (Borries et al. 1999) and that subordinate males were tolerated by the dominant to prevent a rival from entering the group (Hrdy 1977). Further details on langur life history may be found for example in Koenig & Borries (2001) and Hrdy (1977).
139
140
R. Hager
Objective of the study
The aim of this study is first to show how reproductive skew among primate males can be calculated and how ecological and population factors influencing variation in levels of skew can be investigated, with a specific focus on the effects of future mating prospects. Second, I ask the question whether the explanations given by skew theory present a realistic way of interpreting the effects of multiple variables on skew in this system and whether we gain additional insights from applying skew theory to understand sociality in langurs, in particular in the light of existing theories such as the priority-of-access model (Altmann 1962). Third, I highlight a number of assumptions in skew models that need to be revisited in order to give a more realistic description of the factors affecting male skew. The study presented here suggests directions for future analyses of skew in primates that include both ecological and population variables as predictors of skew. I do not attempt, however, to explicitly test whether skew in langurs is best described by a tug-of-war or a transactional mechanism. Methods Measuring reproductive skew
While the concept of reproductive skew seems an intuitively straightforward way of measuring how reproduction is shared among group members, the quantification of skew has been more contentious (Tsuji & Kasuya 2001, Nonacs 2003). At one end of a continuum, high skew occurs if one or a few individuals monopolize reproduction in a group, whereas there is no skew when reproduction is egalitarian. Thus, reproductive skew refers to the bias in reproduction among individuals of the same sex (Reeve & Keller 1995, Clutton-Brock 1998, Johnstone 2000). Essentially, a group of animals is divided into non-breeders and breeders of one sex, and skew is calculated among the breeders, although the definition of who should be included in this category may prove more difficult in some species (see e.g. Chapter 9). In most skew indices, the level of skew ranges from 0 (even distribution of reproduction) to 1 (monopolized reproduction), although other measures have been devised (Kokko et al. 1999, Tsuji & Kasuya 2001). On statistical and mathematical grounds, a number of potential problems when measuring skew have been pointed out over the years (Nonacs 2003). First, some level of skew will arise purely due to random processes. Thus, it is problematic that a level of skew is returned in the analysis which in fact does not exist. Second, when mean reproductive output is very low among group members, the level of
Male langurs: future mating prospects and ecological factors skew may appear high simply due to sampling error (Tsuji & Kasuya 2001). To address these problems, initial skew indices have been refined, leading to a large number of indices at hand (Kokko et al. 1999, Tsuji & Kasuya 2001). Which index should be chosen when studying reproductive skew and its underlying causes in primates? To date, only two indices have been commonly applied in empirical studies: Nonacs’ B index (Nonacs 2000) and Pamilo & Crozier’s S3 index (Pamilo & Crozier 1996), mainly because of their reliable statistical properties. In addition to the need to accommodate statistical accuracy, practical considerations are likely to play a role in the choice of index, such as the availability of specific data. In studies on skew in primates, Nonacs’ B index has been applied to measure reproductive skew in groups of mountain gorillas (Bradley et al. 2005) and in male rhesus macaques (Widdig et al. 2004) because it allows comparing groups of different sizes and productivity. The calculation of this index requires data on the tenure of individuals, although this will be simplified if all individuals were present throughout the period over which skew is measured. However, if no information on individual tenure is available, or if it is incomplete, a different index may be chosen. In the dataset used in this study (Launhardt et al. 2001), a given subordinate’s contribution to the group’s reproductive output was unknown and could thus not be related to measures of tenure. Therefore, and due to uncertainties in the tenure of subordinates, I shall use the effective-number-ofbreeding-individuals index (Pamilo & Crozier 1996), which has been used in other empirical work on skew (Vehrencamp 2000). Calculating levels of skew in Hanuman langurs
I calculated reproductive skew in Hanuman langurs from populations in Ramnagar, India, using the effective-number-of-breeding-individuals skew index S3 (Pamilo & Crozier 1996) and published paternity data (Launhardt et al. 2001). This index measures skew as a function of the total number of reproductive individuals and their individual reproductive contribution, and is expressed as X Skew ¼ N T 1= p2i =ðNT 1Þ
ð6:1Þ
where NT is the total number of actually reproductive individuals and pi is the reproductive contribution of the i-th individual. Skew was calculated for each group and year for which paternity data were available, rather than per group across years, because mean tenure of alpha males in multi-male groups was only 1.47 mating seasons and subordinates as well as females were joining and leaving the groups across years in both one-male and multi-male groups
141
142
R. Hager (Launhardt et al. 2001). This pattern is caused by multiple social and ecological factors, which in turn may affect an individual’s trade-off of staying versus leaving (subordinates) or conceding a share of reproduction to subordinates or not (dominant). In another study in macaques, reproductive skew has been calculated for each year (Widdig et al. 2004), while Bradley et al. (2005) used longterm data and calculated skew per group (see also below in the Discussion). The original study on paternity assigned the same share of the overall proportion of reproduction to non-alpha resident males while only the alpha male’s contribution was determined (Launhardt et al. 2001). Thus, I assumed an equal share of reproduction among the adult subordinates in my analysis. However, a rank order may exist among subordinate males (see e.g. Borries 2000), which, in turn, may result in an unequal share of reproduction among them. In order to estimate how this would affect calculated levels of skew, I have conducted an additional analysis assuming that only the highest-ranking subordinate obtained any successful matings. Furthermore, it should be noted that skew in one-male groups will not invariably be complete because unidentified non-resident males are able to sire young (Launhardt et al. 2001), possibly leading to lower skew in one-male groups with multiple offspring (e.g. Ohsawa et al. 1993). With a few exceptions (Jodhpur), all Asian langur populations are characterized by the existence of one-male groups (Srivastava & Dunbar 1996), which have thus been included in the calculation of skew. Groups in which no young were born were excluded both from the calculation of skew and from the analysis of factors that may affect skew. Note that there is one dominant male per group. Estimating and analyzing reproductive skew in langur populations
In order to explore which ecological and population factors may affect variation in levels of skew in populations for which no paternity data are available, I first analyzed data on the Ramnagar population to find variables that significantly predict skew and are available for most Asian langur populations. Using the Ramnagar dataset, I conducted a quadratic regression analysis that predicts the level of skew using the number of males as the independent variable. Then I collated data from the literature on 11 original or derived variables from 21 Asian langur populations (Srivastava & Dunbar 1996, Treves & Chapman 1996). Table 6.1 lists the parameters that are used in the skew analysis. Ecological parameters
As a measure of ecological constraint for a given population, I followed Emlen’s model (1982) and used the amount of annual rainfall in addition to the
Male langurs: future mating prospects and ecological factors Table 6.1 Ecological and life-history parameters used in the GLM analysis of potential predictors of estimated levels of skew in 21 Asian langur populations. Details of individual parameters are given in the text. Ecological Annual precipitation (mm) Number of months with less than 50 mm of rain Predation risk Dispersal costs Life-history/population Number of potential mates per male Population density (n individuals per km2) Proportion of one-male groups in the population Number of non-group males per bisexual group Birth rate Home-range size (km2) Birth index
number of months per year with less than 50 mm of rain. High levels of rainfall may reflect overall better conditions (Emlen 1982, Clutton-Brock et al. 2001). Predation risk was assigned following Treves & Chapman (1996) as an index of disturbance of the predator community whereby 1 refers to an intact and 0 to a disturbed community. In the latter case, the risk of predation is assumed to be lower than in the former (Treves & Chapman 1996). Furthermore, in transactional skew theory, dispersal costs play a crucial role in determining whether an individual will leave and attempt breeding outside the focal group or stay and accept a share of reproduction (e.g. Clutton-Brock 1998, Reeve & Emlen 2000, Hager 2003b). While it is widely accepted that dispersal costs are affected by multiple variables such as predation risk or uncertainties of resource availability (e.g. Moore & Ali 1984), quantifying costs of dispersal has been consistently difficult (e.g. Greenwood 1980, Pusey & Packer 1987). Given the data available, here I have approximated dispersal costs as a composite measure of predation risk and population density, for which published data from langurs are available. First, predation risk is generally accepted as a key cost of dispersal such that costs increase with higher predation risk due to unfamiliar habitat, loss of group protection, etc. (e.g. Isbell 1994, Alberts & Altmann 1995). Assuming that population density at moderate to high densities reflects more saturated habitats (Sterck 1998), the resulting increase in competition for limited resources such as food and mates is thought to affect the costs of dispersal (Tanner 1966, Emlen 1982, Morris 1992, 2002). In the
143
144
R. Hager analyses presented here, I have assumed that costs of dispersal increase with both predation risk and population density. In order to quantify this measure, population density was divided into three categories: high, intermediate, and low, the boundaries defined by the upper and lower limits of the first, second, and third quartiles. Predation risk was used as a categorical variable as described above. Both factors were then used in an additive model such that, for example, dispersal costs were highest when both population density and predation risk were high. Life-history/population parameters
In order to approximate future male mating prospects, I calculated the number of females per male in a given group or population, assuming that, everything else being equal, more females per male reflect better chances of mating per male. Furthermore, I added the number of extra-group males per bisexual group, which may serve as an indicator of potential intrasexual competition to monopolize groups. The birth rate in a population refers to the number of births per female per year, and the birth index, ranging from 1 to 12, denotes seasonality. Small numbers in this index indicate that births occur during a few months per year (1 means that all births occurred during one month) while the maximum value, 12, refers to an even distribution of births over the year (Srivastava & Dunbar 1996). Finally, home-range size has been included. Reproductive skew in all populations was estimated using a quadratic regression. Data on within-group relatedness were not available. As a final step in the analysis, I employed general linear models to assess the effects of the above parameters on levels of skew using a stepwise approach (Sokal & Rohlf 1995, Grafen & Hails 2002) with predation risk, birth index, and dispersal costs as categorical variables and the remainder as covariates. Error structures were checked for normality. Results Reproductive skew in the Ramnagar population
Using the effective-number-of-breeding-individuals skew index (Equation 6.1; Pamilo & Crozier 1996) and published paternity data (Launhardt et al. 2001), I first calculated levels of reproductive skew in the Ramnagar populations following the assumption that all subordinates obtained an equally large share of reproduction (Launhardt et al. 2001). However, if a hierarchy exists among subordinates it is reasonable to assume that reproduction may be unequally distributed, with higher-ranking subordinates gaining a greater share of the offspring not sired by the dominant in the group. This, in turn,
Male langurs: future mating prospects and ecological factors Table 6.2 Skew in the Ramnagar population. Adapted and modified from Launhardt et al. (2001). It should be noted that the number of infants sired by the dominant male represents a statistical measure of the probability of paternity (Launhardt et al. 2001). Skew 2 assumes that all offspring not sired by the alpha male were fathered by the highest-ranking subordinate. Group O-troop
P-troop
A-troop
X-troop B-troop
Year
Adult males
Infants
Alpha sired
Skew
Skew 2
1992
5
2
1.57
0.852
0.873
1993
5
9
6.21
0.750
0.813
1994
4
3
2.57
0.883
0.891
1995
7
5
4.57
0.967
0.967
1996
3
7
2.14
0.005
0.631
1992
3
7
1.14
0.173
0.813
1994
3
3
2.57
0.828
0.837
1995
3
3
1.57
0.210
0.502
1992
2
2
1.57
0.490
0.490
1993
1
1
1
1
1
1994
1
2
2
1
1
1995
1
2
2
1
1
1993
1
5
5
1
1
1994
1
1
1
1
1
1995
1
5
5
1
1
may affect the level of overall skew. Therefore I performed an additional analysis assuming that all young not sired by the dominant male were fathered by the highest-ranking subordinate. Table 6.2 shows that skew is complete in all one-male groups and ranges from very high levels to an almost egalitarian distribution in multi-male groups. Reproductive skew can vary strongly over the years, even within groups. For example, in the O-troop skew reaches a peak in 1995 with a level of 0.967 while in the following year reproduction was shared almost equally among males. The variation of skew across years and groups is less if we assume that the highest-ranking subordinate sired all young that were not fathered by the dominant. Comparing levels of skew assuming an equal subordinate share to those when assuming a monopolized subordinate share for the same year and group we find only small differences, with the exception of the P-troop in 1992 and the O-troop in 1996. Here, skew is much higher when one subordinate sires all non-dominant offspring. Because the dominant male in P-troop sired only 1.14 young out of 7, the subordinate male is assigned all remaining young such that he dominated reproduction in this year, which in turn results in high skew. It should be
145
R. Hager
1.0
Skew
146
0.5
0.0 1
2
3
4 5 Number of males
6
7
Figure 6.1 Reproductive skew in the Hanuman langur population from Ramnagar, India. The graph follows Equation 6.2.
noted that the number of infants sired by the dominant male represents a statistical measure of the probability of paternity (Launhard et al. 2001). In order to estimate levels of skew in populations for which no paternity data are available, we need to find a significant correlation between predictor variables and skew in the Ramnagar population. Assuming an equal share for all subordinates, reproductive skew is given by a quadratic regression, with the number of males as the independent variable: Skew ¼ 1:36292 0:481220 Nm þ 0:0648801ðNm Þ2
ð6:2Þ
where Nm is the number of males in the group in a given year. Figure 6.1 shows that skew in Ramnagar across all years and groups follows a U-shaped curve of group size accounting for 46.8% of the variation (F2,14 ¼ 5.27, p ¼ 0.023, n ¼ 16). Skew is complete in all one-male groups and falls to a low when groups contain 3–5 males (including the dominant), increasing again thereafter. If we assume that reproduction in the group was divided only between the dominant and the highest-ranking subordinate, we find a similarly significant relationship (F2,14 ¼ 4.97, p ¼ 0.027, n ¼ 16). In other words, whether or not subordinate reproduction follows a rank order or is monopolized does not affect the overall relationship between the number of males in the group and skew. Estimating skew in Asian langur populations
While several life-history and ecological parameters are known for many populations of langurs, genetic data on paternity and relatedness of
Male langurs: future mating prospects and ecological factors group members remain scarce. Therefore I estimated levels of skew in these populations using Equation 6.2 (Table 6.3). This shows that across all populations the distribution of reproduction is predicted to be skewed, although within each population the level of skew may vary between groups and years, as suggested by the data from Ramnagar. The highest levels of skew are predicted for populations in Deotala, Jaipur, and Mahdav, while skew takes intermediate values in Orcha Forest and Rajaji. Figure 6.2 illustrates the differences in estimated skew among Asian langur populations. If we were able to investigate whether or not the same set of factors influences skew in populations with similar skew, we could then attempt to assess whether skew could be used to classify and categorize populations. Predictor variables of estimated skew levels
Having estimated levels of skew in langur populations, we can now explore which of the life-history variables listed in Table 6.3 may explain variation in skew. All main effects and meaningful two-way interactions were analyzed using general linear models and the results of this analysis are summarized in Table 6.4. The minimal model shows that three predictor variables have significant main effects: the number of potential mates per male, the level of average annual rainfall in the focal population, and population density. First, the measure of future male mating prospects has a significant positive effect on skew (GLM; F1,12 ¼ 48.26, p < 0.001). In addition, the amount of average annual rainfall also positively impacts on skew (F1,12 ¼ 27.62, p ¼ 0.001), as does population density (GLM; F1,12 ¼ 8.84, p ¼ 0.021). In other words, skew is predicted to increase with increasing future mating prospects, annual rainfall, and population density. These effects are additive in that they explain part of the variation in levels of skew, over and above the effects of the other variables. Skew is affected to roughly the same extent by annual rainfall as it is by population density and is given by the following model equation: Estimated skew ¼ 0:145 þ 0:04½future prospects þ 2 · 103 ½population density þ 1:75 · 104 ½annual rain
ð6:3Þ
The positive effect of rainfall on skew is displayed in Figure 6.3, which shows the residual variation explained by rainfall after controlling for number of mates and population density. No significant second-order interactions between the predictor variables can be reported. Further, I detected no significant effects of birth rate, home range size, the average group size in the population, or the birth index characteristic of a population detected.
147
148
149
Table 6.3 Predicted levels of skew in 21 Asian langur populations in order of latitude. Levels of skew were estimated using Equation 6.2. Adapted from Treves & Chapman (1996) and Srivastava & Dunbar (1996). Empty cells refer to unavailable data, and populations with no multi-male groups have been excluded. Days/ Population
Birth rate
density
(births per
Home
month with
Estimated
Potential
Precipitation
% one-male
(individuals/
female per
Range
Predation
50 mm
Birth
Location
mates/male
per year (mm)
groups
km2)
year)
size (km2)
index
rain
index
skew
Kanha
7.67
1600
87
46.0
0.43
0.75
0
8
2.5
0.879
Dharwar
4.71
1492
77
91.5
0.20
0.20
1
5
8
0.732
Mundanthurai
2.63
-
-
-
-
-
-
-
0.811
Karnataka 1
7.17
-
-
-
-
-
1
-
0.707
level of
Karnataka 2
5.07
-
-
-
-
-
1
-
0.816
Andhra
5.25
-
-
-
-
-
1
-
0.759
Pradesh Junbesi
3.27
2500
33
1.5
0.11
6.60
1
5
2.5
0.787
Orcha forest
1.97
2030
0
4.4
0.72
5.30
0
5
7.0
0.471
Jaipur
17.80
648
84
12.5
0.51
-
1
8
0.947
Ambagarh
33.50
-
-
66.0
-
-
0
-
0.879
Rajaji
3.97
1200
25
90.0
-
1.50
0
6
2.5
Abu forest
6.74
1800
88
31.6.0
0.65
0.40
1
8
9
Abu town
5.56
1800
88
71.8
0.65
0.40
1
8
0.477 0.683 0.707
Ranthambhore
8.48
900
67
14.6
0.32
-
0
8
4.0
0.537
Kanha meadow Deotalao
4.79 10.50
-
-
46.1 -
-
-
0 -
-
2.5
0.582 0.947
1.0
0.477
Gir forest
5.53
674
40
121.5
0.31
0.40
0
8
Mahdav
13.30
-
-
-
-
-
-
-
0.947
Bangladesh
3.80
-
-
-
-
-
1
-
0.660
Raipur
5.61
388
57
-
0.35
-
1
7
1.0
0.524
Simla
3.33
1800
42
24.6
0.51
2.80
-
5
1.5
0.476
R. Hager
1 0.9
Estimated level of reproductive skew
150
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21
Figure 6.2 Estimated levels of reproductive skew for 21 Asian langur populations. The populations are numbered 1–21 in the same order as in Table 6.3.
Moreover, neither dispersal costs, predation risk, the mean number of infants, nor the proportion of one-male groups and their number per bisexual group were found to exert a significant effect on levels of skew. It should be noted that the number of females as a separate predictor does not significantly influence skew either (GLM; F1,12 ¼ 0.01, p ¼ 0.952).
Discussion The objective of this study was to illustrate how variation in reproductive skew among male primates may be explored, and to assess the potential of skew theory in helping to understand why specific factors influence skew and ultimately primate social systems. Using the S3 index, I calculated reproductive skew for five groups of Hanuman langurs in Ramnagar across several years. The analysis revealed a U-shaped curve of skew when plotted against the number of male group members. Skew was complete in all one-male groups, fell to a minimum with 2–3 subordinate males in the group, and increased again thereafter. Finally, several predictor variables of estimated skew levels in Asian langur population were identified using multiple regression. Chances of gaining matings in the future, as reflected in number of mates per male and survival chances (approximated by annual rainfall), were found
Male langurs: future mating prospects and ecological factors Table 6.4 GLM model of factors influencing predicted levels of reproductive skew in 21 populations of Asian langurs, listed in order of p value. Full model
DF
F
p
Number of potential mates per male
1
48.26
0.001
Annual precipitation (mm)
1
27.62
0.001
Population density (n individuals per km2)
1
8.84
0.021
Birth rate
1
5.32
0.069
Dispersal costs
2
14.98
0.180
Home-range size (km2)
1
1.70
0.262
Mean group size
1
0.88
0.386
Birth index
1
0.28
0.622
Predation risk
1
0.25
0.666
Mean number of infants
1
0.08
0.793
Proportion of one-male groups
1
0.05
0.833
Number of non-group males per bisexual group
1
0.04
0.860
Minimal model
Coefficient
SE coefficient
Constant
0.145
0.122
Number of potential mates per male
0.040
0.009
Population density (n individuals per km2)
0.002
7.9 · 104
Annual precipitation (mm)
0.002
5.7 · 105
Percentage variance accounted for: 88.10
to exert a significant positive effect on skew. In addition, environmental constraints, as measured by population density, positively affected skew. Reproductive skew is predicted to increase additively with all three parameters. In contrast to the other two studies on skew in primates (Widdig et al. 2004, Bradley et al. 2005), which found no support for predictions of concession models, here I demonstrate that variation in male reproductive skew may be explained by considering ecological and population variables, and that their effects can be explained in the framework of transactional skew models. However, it should be noted that no genetic data were available on withingroup relatedness, so the effect of r on skew could not be ascertained. Reproductive skew in Ramnagar populations: a potential role for infanticide?
Reproductive skew among male Langurs in the Ramnagar population plotted as a function of the number of males in the group followed a U-shaped curve with a minimum at just over three males in the group (Figure 6.1). It is surprising to see that skew increases when more than 3–4 males are present in
151
R. Hager Model residuals (number mates and population density)
152
0.2
0.1
0.0
–0.1
–0.2 500
1500 Annual rain [mm]
2500
Figure 6.3 GLM residuals of the two predictors potential mates and population density as a function of annual rainfall.
the group, since one would have expected it to become more difficult for the dominant to control subordinate reproduction when more males are present. Several factors need to be considered when attempting to explain this pattern. First, it should be noted that the observed pattern is mainly due to skew being complete in one-male groups, although skew declines also from two to three males in the group and increases thereafter. Second, even if one-male groups were excluded, skew would be low with around 3–4 males and higher with more males in the group. This suggests that the analysis reflects a realistic pattern of the distribution of skew, so we can now explore adaptive explanations. The role of infanticide in the control of reproduction has been modeled by Hager & Johnstone (2004), who found that the ability to discriminate own young from those of other group members, coupled with the option of infanticide, can have strong effects on reproductive skew because the threat of infanticide is effective in preventing subordinate reproduction. Hanuman males are certainly capable of committing infanticide, with most attacks being reported in the context of group takeovers both in one-male groups (e.g. Vogel & Loch 1984) and in multi-male groups (Boggess 1984, Borries 1997). The Hager & Johnstone (2004) model shows in particular that in order for the threat of infanticide to affect subordinate reproduction, little or no actual infanticide needs to occur, because subordinates refrain from siring young that otherwise would be killed by the dominant. Applied to the pattern of skew found in the Ramnagar population, one could speculate that with only a few extra males (one or two), the dominant may be able to directly control subordinate reproduction, for example by interrupting mating attempts or controlling access to females in the group. Infanticide is not invoked in this situation
Male langurs: future mating prospects and ecological factors because of the potential costs of accidentally killing one’s own young relative to possible fitness gains. However, the more males there are in the group, the more difficult it will become to physically control subordinate reproduction, and the dominant may threaten to kill any young that are being sired by other males, even at the risk of incurring fitness costs by killing his own young. In other words, the dominant may employ alternative strategies of controlling reproduction depending on the number of males in the group which affect his trade-off of the costs and benefits of physical control versus the threat of infanticide. The transition between these two strategies may occur when there are 2–3 extra males in the group, at which point it becomes difficult to directly control subordinate reproduction, and too costly to attempt to do so with more males. Testing this hypothesis will remain difficult, however, because of the constraints on experimental manipulation of male numbers in primate groups that could potentially lead to elevated levels of infanticide. As an alternative to the infanticide hypothesis, skew may be low in groups with 2–3 subordinate males because all males (dominant and subordinate) may be able to monopolize a given female to a greater extent than with more males in the group. Within-group aggression may increase with the number of males in the group, and the costs of attempting to monopolize a female may become too high such that queuing, yielding less costly but also less immediate chances of breeding, may become the better alternative for subordinates. In turn, this would result in higher skew, as suggested by the Ramnagar data. Factors influencing reproductive skew in langur populations: future mating prospects
Estimated levels of skew in Asian langur populations were found to be affected by chances of obtaining future matings and population density. Concession theory predicts that skew should be higher when future prospects are good, because benefits from queuing increase. Under such conditions, subordinates are expected to accept a smaller share of current reproduction when remaining in the group (Kokko & Johnstone 1999). Results of the analysis presented here lend support to this hypothesis: skew among males was positively influenced by their future mating prospects, which has been approximated by chances of survival and potential mating partners per male, whereby a higher ratio reflects better chances. One caveat here is that as the number of females per male increases the variance in reproductive success in the limited sex is predicted to increase (Emlen & Oring 1977). This means that mating chances may increase asymmetrically between males, with higher-ranking males gaining most of the benefits. However, it is reasonable to assume that monopolization becomes generally more difficult as the number of females increases.
153
154
R. Hager In langurs, staying in the group and queuing for a breeding position may become an option for subordinate males due to costs associated with dispersal (e.g. Rajpurohit & Sommer 1991). While to date no study has explicitly investigated the possibility of queuing in langurs, several observations provide evidence for this hypothesis. In both multi-male groups and all-male bands, it has been reported that individuals may rise in the hierarchy to attain the breeding position, with the result that most rank changes occur during male immigration, but also among resident males (Rajpurohit et al. 1995, Borries 1997). Other studies in primates have shown that an individual’s rank position is a good predictor of mating and reproductive success (Cowlishaw & Dunbar 1991, Ellis 1995). Altmann (1962) suggested that a ranking system may indeed function as a queue to attain dominance rank, for example in savannah baboons (Alberts et al. 2003) or Japanese macaques (Macaca fuscata: Sprague et al. 1998). Furthermore, in langurs it has been observed that during a group takeover subordinate males provided coalitional support, and Sommer (1988) suggested that such males gain delayed benefits, for instance through increased mating chances – a situation equivalent to that described in transactional skew models. Thus, queuing as a result of weighing costs and benefits of dispersal may be a viable strategy for subordinate males to increase their reproductive success. It should be noted that queuing in general may not apply to female langurs, due to their inverse dominance system whereby younger females are at the top of the hierarchy and then decline thereafter (Borries et al. 1991). Implicit in the argument that queuing for future matings can be decisive in affecting skew is the assumption that both individual survival and number of potential mates at a given point in time indeed reflect the conditions projected into the future. If the number of females per male is highly variable over the years (e.g. in populations with a high emigration/immigration rate), or if rainfall, and thus the chance of survival, is unpredictable, such measures of future prospects are more likely to estimate short-term mating prospects. Therefore, it seems important to ask how reliable the measures of future mating prospects for a given population are. In addition, it may be helpful to investigate in a separate study whether any of the population or ecological parameters influence the dominant’s ability to control subordinate reproduction. For instance, the structure of the habitat in which the group lives could make it more or less easy to interrupt subordinate mating attempts. Ecological constraints
Ecological constraints determine whether the trade-off of costs and benefits associated with dispersal and attempting to breed independently is
Male langurs: future mating prospects and ecological factors positive or negative. In concession models, for example, this will affect the share of reproduction the dominant needs to offer in order to retain the subordinate. Analyzing the factors influencing variation in skew provides support for the assumptions made in skew models about the significance of ecological constraints. First, I found that the amount of annual rainfall positively affects skew. Higher levels of rainfall have previously been assumed to reflect good environmental conditions (e.g. Emlen 1982), which, everything else being equal, may mean better chances of survival. Assuming that queuing for mating chances is an option for subordinate males, this would result in higher skew. Second, population density was found to affect skew such that skew is predicted to increase with population density. Following Sterck (1998), higher population densities occur in saturated habitats, in which we may expect higher competition for limiting resources (Tanner 1966, Morris 1992). Such conditions place constraints on dispersal and successful joining of other groups (Jones et al. 1988), which, in turn, make staying in a given group more attractive, so that skew should be higher. In addition, when resources are very limited in a habitat, retaining subordinates in the group may effectively reduce group productivity due to increased within-group competition over resources. In the concession framework, the staying incentive may then be reduced by the dominant, resulting in higher skew. It is important to note again that the analysis was based on estimated levels of skew derived from the significant relationship between number of males and skew in the Ramnagar population, which may deviate from that in other populations. Costs of dispersal
Many studies in primates have demonstrated that dispersing incurs fitness costs in terms of higher mortality, predation risk, and fighting with other males (Rajpurohit & Sommer 1991, van Schaik & Ho¨rstermann 1994). Concession models of skew predict, everything else being equal, high costs of dispersal to result in high skew (Johnstone 2000). Results of my analysis, however, did not detect any effect of dispersal costs on estimated skew in Asian langurs. Several factors could account for this finding. The most parsimonious of these would be that my proxy measure of dispersal costs (taking into account the additive effects of predation risk and population density) does not adequately reflect costs of dispersal in the langur system, implying the possibility that skew may still be affected by this variable had it been measured in a different way. Given the paucity of existing data and the non-existence of a generally agreed parameter to quantify dispersal costs, we are left in such an analysis with making assumptions about which recorded parameters are most likely to represent costs of dispersal (see Methods). For example, it could be
155
156
R. Hager that dispersal costs for a given individual are not at their highest when both predation risk and population density are high, but rather that the chance of individual predation is a negative function of population density because the attack rate per individual may be lower. At present, it is difficult to ascertain whether dispersal costs really do not impact on skew, or whether any effect could not be detected with the measures used. Assuming that dispersal costs can, indeed, be approximated as conducted in this analysis, we can ask why costs of dispersal did not affect levels of skew as posited by skew models. It may be that in langurs predation risk plays only a minor role in determining costs of dispersal (as opposed to fighting with other males, unknown habitat, etc.). Indeed, only one predator attack has been reported in the literature (Borries 1997). It should be noted that animals are usually not observed for the whole duration of their lifetime and may simply disappear for unknown reasons. This illustrates, again, the immense difficulty in quantifying the true effects of predation mortality and hence costs of dispersal. Future directions
Future studies of reproductive skew and the factors that influence variation in this measure will benefit from a number of improvements both in data collection and in conceptual aspects, which I will only briefly discuss here. Foremost is the collection of data that allow calculation of skew across years and groups, i.e. paternity data with complete records of individual tenure and, ideally, values for within-group relatedness. Second, primatologists should endeavor to collect and quantify the key variables identified by skew theory (see above), the more problematic being dispersal costs, future mating prospects, and group productivity. Third, as yet, there is no consensus regarding the period over which one should measure skew (skew per group year or over several years; see Bradley et al. 2005, Widdig et al. 2004), nor does it appear to be conceptually clear why one should prefer one option over the other. In primates, a starting point would be to consider over what time period individuals in a given species are likely to trade off, for example, staying versus leaving (from the subordinate’s perspective), or what the potential gains of subordinate presence are compared to the loss of direct reproduction if a share is conceded to subordinates (from the dominant’s perspective). The length of the breeding season and average tenure are variables that could be considered here (Figure 6.4). Of course, it would be highly interesting to compare how skew varies between years against the level of skew over longer periods of time (see Altmann et al. 1996). Can we identify groups by a reasonably constant level of skew, and can we correlate this to specific population and ecological parameters that are characteristic of the focal population?
Male langurs: future mating prospects and ecological factors
Figure 6.4 Among the Hanuman langurs at Mount Abu, Rajasthan, India. The average tenure length for a resident male was 27 months. This picture shows a former alpha male (grinding his teeth in the foreground) together with young males, all ousted when a rival male took over their troop. Two mothers with infants, attempting to avoid attacks on their infants by the new male, are also temporarily traveling with them (details can be found in Hrdy’s 1977 book The Langurs of Abu). Courtesy of Sarah B. Hrdy/Anthro-Photo.
Within-group aggression
One key feature of group-living animals is the level of within-group aggression, which is thought to be linked to reproductive skew in a group (Reeve & Nonacs 1997, Reeve 2000, Reeve & Keller 2001, Hager 2003b). A particular role in determining levels of aggression between members of one sex is assigned to dispersal costs. For example, in a concession model by Cant & Johnstone (2000), high costs of dispersing result in an increase of aggression because the subordinate (who is constrained in dispersing) stands to gain a potentially large share of reproduction by being highly aggressive, the extent of which depends on how aggression affects group productivity. On the other hand, in a tug-of-war situation, skew will be determined largely by the subordinate’s fighting ability, and will be high when the latter is poor because the
157
158
R. Hager subordinate can claim only a small fraction of reproduction (Johnstone 2000). In the context of queuing, it is reasonable to assume that a consequence of better future mating prospects could be to reduce costly activities such as fighting to gain matings within a group. Therefore, one testable prediction is that within-group aggression should be lower in groups in which subordinates have good future prospects than in groups with poor prospects, all other things being equal. The effects of female choice on male reproductive skew
All models of skew analyze how reproduction is shared among samesex individuals in a group and how population, ecological, and individual factors influence the dominant’s and subordinate’s trade-offs. However, in many systems, including primates, females are likely to exert an influence on male reproductive skew, for example by sneaky matings with subordinates, by extra-group copulations, or by totally disregarding any male except the alpha. Among primates are numerous examples of female choice in some form (e.g. Stumpf & Boesch 2005). Without attempting any formal description, it seems reasonable that skew among males could be negatively affected by females choosing to occasionally mate with a male different from the dominant, perhaps to increase the genetic diversity of their young or as a security against infanticide should the subordinate become the dominant at a later stage (Hausfater & Hrdy 1984). While it seems clear from many empirical studies that female choice impacts male reproductive skew, to date this has not been given due attention in skew theory (but see Cant & Reeve 2002). For example, if subordinates can expect to obtain a certain level of matings through female choice, they may accept a lower “direct” share conceded by the dominant, and skew might be higher than without female choice.
Conclusion The study presented here shows that skew theory can help explain causes of variation in reproductive skew in langurs. The factors found to influence levels of estimated skew among males can be explained by a concession mechanism if queuing for mating opportunities is a realistic option in langurs. The question at issue, then, is: do we advance our understanding of primate sociality by using skew theory to explain the distribution of reproduction, compared to alternative explanations? A central tenet in primatology is the priority-of-access model (Altmann 1962), which asserts that access to estrous females is largely determined by a male’s position in the dominance hierarchy. A general prediction of this model is that male
Male langurs: future mating prospects and ecological factors reproductive success is a function of his tenure and the number of estrous females at a given point in time, which has been confirmed by several studies (Altmann et al. 1996, Pope 1990). We therefore need to ask whether we gain any additional insight into primate social systems by applying skew theory? The answer, for the present author, is “Yes.” While the priority-of-access model makes predictions about individual reproductive success, the model does not offer any further explanations as to why individuals show variation in reproductive success. Which factors influence a dominant’s standing in the hierarchy? What is the role of ecological factors in determining skew? Skew theory holds the possibility of linking ecological, genetic, and population characteristics in a conceptual approach (described as transactional, compromise, or synthetic approaches: Johnstone 2000, Reeve & Shen 2006). Clearly, the factors underlying variation in skew can be analyzed regardless of skew theory, but these models have formally demonstrated the way in which such factors can be expected to interact under conditions of natural selection, and they offer testable hypotheses. Moreover, a more comprehensive analysis of reproductive skew allows us then to compare and categorize groups of animals, and perhaps species, according to their level of skew. Future studies should compare populations that feature similar skew. It will also be important to explore whether or not these comparisons are affected by similar sets of parameters and have similar consequences, for instance, for group sizes.
Acknowledgments I would like to thank Carola Borries, Volker Sommer, and two anonymous reviewers for help with the literature, and Clara B. Jones for constructive comments on the manuscript. References Allaine´, D. (2000). Sociality, mating system and reproductive skew in marmots: evidence and hypotheses. Behavioural Processes, 51, 21–34. Alberts, S. C. & Altmann, J. (1995). Balancing costs and opportunities: dispersal in male baboons. American Naturalist, 145, 279–306. Alberts, S. C., Watts, H. E. & Altmann, J. (2003). Queuing and queue-jumping: longterm patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Animal Behaviour, 65, 821–840. Altmann, J., Alberts, S. C., Haines, S. A., et al. (1996). Behavior predicts genes structure in a wild primate group. Proceedings of the National Academy of Sciences of the USA, 93, 5797–5801.
159
160
R. Hager Altmann, S. A. (1962). A field study of the sociobiology of rhesus monkeys, Macaca mulatta. Annals of the New York Academy of Sciences, 102, 338–435. Boggess, J. (1984). Infant killing and male reproductive strategies in langurs (Presbytis entellus). In G. Hausfater & S. B. Hrdy, eds., Infanticide: Comparative and Evolutionary Perspectives. New York, NY: Aldine, pp. 283–310. Borries, C. (1997). Infanticide in seasonally breeding multimale groups of Hanuman langurs (Presbytis entellus) in Ramnagar (South Nepal). Behavioral Ecology and Sociobiology, 41, 139–150. Borries, C. (2000). Male dispersal and mating season influxes in Hanuman langurs living in multi-male groups. In P. M. Kappeler, ed., Primate Males. Cambridge: Cambridge University Press, pp. 146–158. Borries, C., Sommer, V. & Srivastava, A. (1991). Dominance, age, and reproductive success in free-ranging female hanuman langurs (Presbytis entellus). American Journal of Primatology, 12, 231–257. Borries, C., Launhardt, K., Epplen, C., Epplen, J. T. & Winkler, P. (1999). Males as infant protectors in Hanuman langurs (Presbytis entellus) living in multimale groups: defence pattern, paternity and sexual behaviour. Behavioral Ecology and Sociobiology, 46, 350–356. Bradley, B. J., Robbins, M. M., Williamson, E. A., et al. (2005). Mountain gorilla tug-of-war: silverbacks have limited control over reproduction in multi-male groups. Proceedings of the National Academy of Sciences of the USA, 102, 9418–9423. Buston, P. M. (2004). Does the presence of non-breeders enhance the fitness of breeders? An experimental analysis in the clown anemonefish Amphirion percula. Behavioral Ecology and Sociobiology, 57, 23–31. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Cant, M. A. & Johnstone, R. A. (2000). Power struggles, dominance testing, and reproductive skew. American Naturalist, 155, 406–417. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H. & Harvey, P. H. (1979). Comparison and adaptation. Proceedings of the Royal Society of London B, 205, 547–565. Clutton-Brock, T. H., Brotherton, P. N., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Cowlishaw, G. & Dunbar, R. I. M. (1991). Dominance rank and mating success in male primates. Animal Behaviour, 41, 1045–1056. Dietz, J. M. (2004). Kinship structure and reproductive skew in cooperatively breeding primates. In B. Chapais & C. M. Berman, eds., Kinship and Behavior in Primates. Oxford: Oxford University Press, pp. 223–241. Ellis, L. (1995). Dominance and reproductive success among nonhuman animals: a cross-species comparison. Ethology and Sociobiology, 16, 257–333.
Male langurs: future mating prospects and ecological factors Emlen, S. (1982). The evolution of helping. I. An ecological constraints model. American Naturalist, 119, 29–39. Emlen, S. (1994). Benefits, constraints and the evolution of family. Trends in Ecology and Evolution, 9, 282–285. Emlen, S. (1995). An evolutionary theory of the family. Proceedings of the National Academy of Sciences of the USA, 92, 8092–8099. Emlen, S. & Oring, L. (1977). Ecology, sexual selection, and the evolution of mating systems. Science, 197, 215–223. Faulkes, C. G. & Bennet, N. (2001). Family values: group dynamics and social control of reproduction in African mole-rats. Trends in Ecology and Evolution, 16, 184–190. Grafen, A. & Hails, R. (2002). Modern Statistics for the Life Sciences. Oxford: Oxford University Press. Greenwood, P. J. (1980). Mating systems, philopatry, and dispersal in birds and mammals. Animal Behaviour, 28, 1140–1162. Hager, R. (2003a). Models of reproductive skew applied to primates. In Jones, C. B. ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Direction. Norman, OK: American Society of Primatologists, pp. 65–101. Hager, R. (2003b). The effects of dispersal costs on reproductive skew and withingroup aggression in primate groups. Primate Report, 67, 941–949. Hager, R. (2004). Behavioural ecology. Encyclopedia of Life Sciences. Wiley: Chichester. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Hamilton, I. M. (2000). Recruiters and joiners: using optimal skew theory to predict group size and the division of resources within groups of social foragers. American Naturalist, 155, 684–695. Hausfater, G. & Hrdy, S. B. (1984). Comparative and evolutionary perspectives on infanticide: introduction and overview. In G. Hausfater & S. B. Hrdy, eds., Infanticide: Comparative and Evolutionary Perspectives. New York, NY: Aldine, pp. xiii-xxxv. Hrdy, S. B. (1977). The Langurs of Abu. Cambridge, MA: Harvard University Press. Isbell, L. A. (1994). Predation on primates: ecological patterns and evolutionary consequences. Evolutionary Anthropology, 3, 61–71. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 105, 5–26. Johnstone, R. A. & Cant, M. A. (1999a). Reproductive skew and indiscriminate infanticide. Animal Behaviour, 57, 243–249. Johnstone, R. A. & Cant, M. A. (1999b). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Jones, W. T., Waser, P. M., Elliott, L. F., Link, N. E. & Bush, B. B. (1988). Philopatry, dispersal, and habitat saturation in the banner-tailed kangaroo rat, Dipodomys spectabilis. Ecology, 69, 1466–1473. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Koenig, A. & Borries, C. (2001). Socioecology of Hanuman langurs: the story of their success. Evolutionary Anthropolgy, 10, 122–137.
161
162
R. Hager Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 266, 571–578. Kokko, H., Mackenzie, A., Reynolds, J. D., Lindstro¨m, J. & Sutherland, W. J. (1999). Measures of inequality are not equal. American Naturalist, 72, 358–382. Krebs, J. R. & Davies, N. B. (1993). An Introduction to Behavioural Ecology. 3rd edn. Oxford: Blackwell Science. Launhardt, K., Borries, C., Hardt, C., Epplen, J. T. & Winkler, P. (2001). Paternity analysis of alternative male reproductive routes among the langurs (Semnopithecus entellus) of Ramnagar. Animal Behaviour, 61, 53–64. Magrath, R. D., Heinsohn, R. G., & Johnstone, R. A. (2004). Reproductive skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. Moore, J. & Ali, R. (1984). Are dispersal and inbreeding avoidance related? Animal Behaviour, 32, 94–112. Morris, D. W. (1992). Scales and costs of habitat selection in heterogeneous landscapes. Evolutionary Ecology, 6, 412–432. Morris, D. W. (2002). Measuring the allele effect: positive density dependence in small mammals. Ecology, 83, 14–20. Newton, P. N. & Dunbar, R. I. M. (1994). Colobine monkey society. In A. G. Davies & J. F. Oates, eds., Colobine Monkeys: Their Ecology, Behaviour and Evolution. Cambridge: Cambridge University Press, pp. 311–346. Nonacs, P. (2000). Measuring and using skew in the study of social behavior and evolution. American Naturalist, 156, 577–589. Nonacs, P. (2003). Measuring the reliability of skew indices: is there one best index? Animal Behaviour, 65, 615–627. Ohsawa, H., Inoue, M., & Takenaka, O. (1993). Mating strategy and reproductive success of male patas monkeys. Primates, 34, 533–544. Packer, C., Collins, D. A., Sindimwo, A., & Goodall, J. (1995). Reproductive constraints on aggressive competition in female baboons. Nature, 373, 60–63. Pamilo, P. & Crozier, R. H. (1996). Reproductive skew simplified. Oikos, 75, 533–535. Pope, T. R. (1990). The reproductive consequences of male cooperation in the red howler monkey: paternity exclusion in multi-male and single-male troops using genetic markers. Behavioral Ecology and Sociobiology, 27, 439–446. Pusey, A. E. & Packer, C. (1987). Dispersal and philopatry. In B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham, & T. T. Struhsacker, eds., Primate Societies. Chicago, IL: Chicago University Press, pp. 250–281. Rajpurohit, L. S. & Sommer, V. (1991). Differences in mortality among langurs (Presbytis entellus) of Jodhpur, Rajasthan. Folia Primatologica, 56, 17–27. Rajpurohit, L. S., Sommer, V., & Mohnot, S. M. (1995). Wanderers between harems and bachelor bands: male Hanuman langurs (Presbytis entellus) at Jodhpur in Rajasthan. Behaviour, 132, 255–299. Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382.
Male langurs: future mating prospects and ecological factors Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Keller, L. (1995). Partitioning of reproduction in mother–daughter versus sibling associations: a test of optimal skew theory. American Naturalist, 145, 119–132. Reeve, H. K. & Keller, L. (1997). Reproductive bribing and policing as evolutionary mechanisms for the suppression of within-group selfishness. American Naturalist, 150, 42–58. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Nonacs, P. (1997). Within-group aggression and the value of group members: theory and a field test with social wasps. Behavioral Ecology, 8, 75–82. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflict in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Sherman, P. W., Lacey, E. A., Reeve, H. K., & Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108. Sokal, R. R. & Rohlf, F. J. (1995). Biometry. 3rd edn. New York, NY: Freeman. Sommer, V. (1988). Male competition and coalitions in langurs (Presbytis entellus) at Jodhpur, Rajasthan, India. Human Evolution, 3, 261–278. Sommer, V. (1989). Sexual harassment in langur monkeys (Presbytis entellus): competition for ova, sperm, and nurture? Ethology, 80, 205–217. Sprague, D. S., Suzuki, S. S., Takahashi, H., & Sato, S. (1998). Male life history in natural populations of Japanese macaques: migration, dominance rank, and troop participation of males in two habitats. Primates, 39, 351–363. Srivastava, A. & Dunbar, R. I. M. (1996). The mating system of Hanuman langurs: a problem in optimal foraging. Behavioral Ecology and Sociobiology, 39, 219–226. Stacey, P. B. & Ligon, J. D. (1987). Territory quality and dispersal options in the acorn woodpecker, and a challenge to the habitat saturation model of cooperative breeding. American Naturalist, 130, 654–676. Sterck, E. H. M. (1998). Female dispersal, social organization, and infanticide in langurs: are they linked to human disturbance? American Journal of Primatology, 44, 235–254. Stumpf, R. M. & Boesch, C. (2005). Does promiscuous mating preclude female choice? Female sexual strategies in chimpanzees (Pan troglodytes verus) of the Tai National Park, Cote d’Ivoire. Behavioral Ecology and Sociobiology, 57, 511–524. Tanner, J. T. (1966). Effects of population density on growth rates of animal populations. Ecology, 47, 733–745.
163
164
R. Hager Treves, A. & Chapman, C. A. (1996). Conspecific threat, predation avoidance, and resource defense: implications for grouping in langurs. Behavioral Ecology and Sociobiology, 39, 43–53. Tsuji, K. & Kasuya, E. (2001). What do the indices of reproductive skew measure? American Naturalist, 158, 155–165. van Schaik, C. P. & Ho¨rstermann, M. (1994). Predation risk and the number of adult males in a primate group: a comparative test. Behavioral Ecology and Sociobiology, 35, 261–272. Vehrencamp, S. (1979). The roles of individual, kin, and group selection in the evolution of sociality. In P. Marler & J. Vandenbergh, eds., Handbook of Behavioral Neurobiology: Social Behavior and Communication. New York, NY: Plenum Press, pp. 351–394. Vehrencamp, S. (1983a). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Vehrencamp, S. (1983b). Optimal degree of skew in cooperative societies. American Zoologist, 23, 327–335. Vehrencamp, S. (2000). Evolutionary routes to joint-female nesting in birds. Behavioral Ecology, 11, 334–344. Vogel, C. & Loch, H. (1984). Reproductive parameters, adult-male replacements, and infanticide among free-ranging langurs (Presbytis entellus) at Jodphur (Rajasthan), India. In G. Hausfater & S. B. Hrdy, eds., Infanticide: Comparative and Evolutionary Perspectives. New York, NY: Aldine, pp. 237–255. Widdig, A., Bercovitch, F., Streich, W. J., et al. (2004). A longitudinal analysis of reproductive skew in male rhesus macaques. Proceedings of the Royal Society of London, B, 271, 819–826. Winkler, P., Loch, H. & Vogel, C. (1984). Life history of Hanuman langurs (Presbytis entellus). Folia Primatologica, 43, 1–23.
7
The causes and consequences of reproductive skew in male primates n o b u y u k i k u t s u k a k e an d c h a r l e s l . n u n n
Summary This chapter discusses the underlying causes and consequences of reproductive skew in male primates. Although our understanding of the causes of skew is still in its infancy, empirical studies thus far support the compromise framework (e.g. tug-of-war model) rather than the concession model. Our assessment of the different models also suggests that the priority-of-access (POA) model makes predictions that are very similar to the compromise framework, but that skew models expand significantly on the POA model by including additional factors that relate to patterns of reproduction within groups. Our phylogenetic comparative analyses on mating skew in male primates also provide supporting evidence for the tug-of-war model, as mating skew decreased as the number of males increased, suggesting that monopolization of females becomes more difficult when there are more rivals (Emlen & Oring 1977). However, there have been no studies that represent strong tests of skew models, possibly because of difficulties in estimating parameters that are necessary for quantitative analyses. Future research could help to clarify the causes of skew, including development of mathematical models that are more suitable to primate societies, empirical studies based on paternity tests, and comparative approaches to investigate interspecific patterns of skew in other biological systems. Previous studies commonly investigated the causes of skew, but fewer have considered the consequences of skew on other physiological and social parameters such as within-group relatedness and sexually transmitted diseases. Of
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
165
166
N. Kutsukake, C. L. Nunn these, it appears that effects on within-group relatedness could have the largest effects on patterns of primate sociality. The introduction of reproductive-skew models into primate research is likely to provide new insights into primate social and reproductive behavior in the future, while a primate perspective is likely to stimulate new theoretical work on reproductive skew.
Introduction Reproductive-skew theory attempts to explain the uneven distribution of reproductive success among same-sexed group members by multiple social, ecological, and genetic factors (reviewed in Johnstone 2000). Reproductiveskew theory has often been divided into two broad categories known as transactional and compromise frameworks. These frameworks differ according to the assumptions that each of them makes. In a version of the transactional framework known as the concession model, the dominant individual controls the reproduction of subordinates and allows them to reproduce in return for the subordinate staying in the group (i.e. the dominant offers a “staying incentive”: Vehrencamp 1983a, 1983b, Keller & Reeve 1994, Clutton-Brock 1998, Johnstone 2000). Retaining the subordinate is assumed to increase group productivity (i.e. total reproductive output of a group) and fitness benefits of a dominant, relative to the alternative of the subordinate leaving the group. In contrast, the tug-of-war model, which is part of the compromise framework, suggests that the dominant individual is unable to control the reproduction of subordinates completely (Reeve et al. 1998, Cant 1998, Clutton-Brock 1998); the division of reproduction is therefore determined by competition between a dominant and subordinate (Reeve et al. 1998, Cant 1998, Clutton-Brock 1998), which is assumed to decrease group productivity. These models can be expanded into systems with more than two individuals competing for reproduction (Johnstone et al. 1999, Reeve & Emlen 2000), including queuing systems (i.e. a subordinate acquiring a higher dominance position in the future: Kokko & Johnstone 1999, Ragsdale 1999, Mesterton-Gibbons et al. 2006). In this chapter, we consider the causes and the consequences of skew in male primates. Although recent research has synthesized the transactional and compromise frameworks into single conceptual models (Johnstone 2000, Reeve & Shen 2006), the classic dichotomy of the transactional (e.g. concession model) and compromise frameworks (e.g. tug-of-war model) provides a useful starting point for investigating reproductive skew in primates and will therefore be used here. Social primates live in relatively stable social groups. In these groups, males can exhibit considerable variation in the degree to which reproduction or matings are skewed (Cowlishaw & Dunbar 1991, Bulger 1993,
Causes and consequences of skew in male primates Kutsukake & Nunn 2006). Although inter-individual variation in male reproductive success has been a central topic in primate research (e.g. Cowlishaw & Dunbar 1991, Bulger 1993, Alberts et al. 2003, van Noordwijk & van Schaik 2004), only recently have researchers begun to investigate patterns of mating and reproduction in male primates using the theoretical framework of reproductive skew (Hager 2003, Widdig et al. 2004, Bradley et al. 2005, Kutsukake & Nunn 2006). Figure 7.1 provides an overview of the topics covered in this chapter. First we focus on the causes of skew, starting with an explanation of the priority-ofaccess model (POA) model and how this model corresponds to newer theoretical frameworks for understanding reproductive skew. In this first section, we also discuss key assumptions and predictions of the tug-of-war and concession models. Further, we review four case studies that have explicitly introduced and used paternity data to investigate predictions of skew models in primates, and we discuss a new research direction to examine predictions from skew theory using phylogenetic comparative methods (Kutsukake & Nunn 2006). In the second part of this chapter, we investigate the consequences of reproductive skew on other biological traits (Figure 7.1), focusing on two examples. The first involves the effects of skew on relatedness within groups, and the second considers how patterns of skew might influence the spread of sexually transmitted diseases.
The causes of reproductive skew The priority-of-access (POA) model
The POA model (Altmann 1962) has been the most influential framework used to explain variation in reproduction among male primates (Altmann 1962, Altmann et al. 1996, Boesch et al. 2006). The model predicts that the dominant male monopolizes reproduction within a group. However, the degree to which the dominant male succeeds in this goal is affected by the number of estrous females in the group. When two or more females are in estrus at the same time, the dominant male is unable to mate-guard all of them effectively, thus providing an opportunity for subordinate males to mate (Figure 7.2). The model therefore makes predictions for the distribution of matings within groups, with the dominant male obtaining the largest share and subordinates obtaining lesser amounts in proportion to their ranks. Empirical studies provide evidence for the dominant male’s advantages in both mating (Cowlishaw & Dunbar 1991, Bulger 1993, Ellis 1995, Alberts et al.
167
168
N. Kutsukake, C. L. Nunn
Causes Demographic Variables Male number Female number
Female influences
Ecological constraints
Estrous overlap Female choice Incest avoidance
Probability of solitary reproduction
Compromise framework
Genetic factors Relatedness among males Transactional framework
Skew among males
Within-group kin structure
Sexually transmitted disease
Consequences Figure 7.1 Scheme of the causes and consequences of reproductive skew discussed in this chapter. Black arrows indicate the effects predicted from the compromise framework (i.e. tug-of-war model), and the dashed arrows indicate the effects predicted from the transactional framework (i.e. concession model).
2003, Kutsukake & Nunn 2006) and paternity success (van Noordwijk & van Schaik 2004). In addition, some studies have investigated the effect of estrous synchrony on the distribution of matings, reproductive success, and the number of males in a group (e.g. Bulger 1993, Paul 1997, Nunn 1999a, Soltis et al. 2001, Takahashi 2004, van Noordwijk & van Schaik 2004, Alberts et al. 2006, Boesch et al. 2006). In general, these studies have shown that the more females are in estrus, the more difficult it becomes for the dominant male to control access to females. This effect of estrous synchrony has also been demonstrated in other species (e.g. domestic cats, Felis catus: Say et al. 2001). The POA model has contributed greatly to primate research, but studies in primates have produced variable results (Dunbar 1988, Cowlishaw & Dunbar 1991, Bulger 1993, van Noordwijk & van Schaik 2004, Kutsukake & Nunn 2006). In some cases, researchers have uncovered the biological reasons for departures from the POA model. For example, mate choice by females may also affect the distribution of reproduction in ways that differ from predictions of the POA model (Dunbar 1988, Soltis 2004). Females may confuse paternity by mating promiscuously and concealing ovulation – both of which should
Causes and consequences of skew in male primates
Figure 7.2 Mating in wild Japanese macaques (Macaca fuscata) at Shiga Heights, Nagano, Japan. A male is mounting an estrous female. Photo by N. Kutsukake.
decrease skew – or females may increase skew by copulating with the dominant male during periods in which the probability of fertilization is high (van Schaik et al. 1999, 2000, Nunn 1999b, van Noordwijk & van Schaik 2004). Some researchers have incorporated the effect of the number of males in evaluating the POA model (e.g. Alberts et al. 2003, 2006, Boesch et al. 2006), based on the reasoning that it should be more difficult for a dominant male to monopolize females when there are more males in the group that are competing for females (Cowlishaw & Dunbar 1991). Our comparative work – discussed below – provides evidence for this effect in analyses that control for phylogeny, suggesting that male number is the primary factor affecting skew in social primates. We refer to this framework as the extended POA model, in order to separate it from the original POA model, in which the number of males was not explicitly included (Altmann 1962). The POA and skew models are not fundamentally different in their goals of explaining the distribution of reproduction within groups. Compared to the POA model, however, the reproductive-skew framework uses a greater number of genetic and environmental variables, allowing a more detailed investigation
169
170
N. Kutsukake, C. L. Nunn of variation in male mating success. For example, skew models take into account the possibilities for males to leave the group and either attempt to breed on their own or join another group where their fitness would be greater, and thus also the need for dominant males to provide staying incentives. The concession model from the transactional framework also makes explicit assumptions about the degree of control that dominant males have over reproduction, with the tug-of-war models explicitly challenging the assumption of the dominant’s complete control of subordinate reproduction. Skew models make use of data on relationships among males, with greater skew predicted under the concession model when males are more closely related. In addition to male reproductive success, these models can be applied to investigate female reproductive success. Testing the reproductive-skew frameworks
Evaluating whether a particular skew model applies to a species requires information on multiple genetic and ecological parameters. Quantification of these parameters is difficult in any species, including primates. Moreover, experimental studies on skew conducted in other systems (e.g. Hymenoptera) are difficult or unethical to attempt in primates, in part because most primates have long lifespans and many are highly threatened. In this chapter, we first investigate the assumptions of different skew models (Johnstone 2000, Magrath et al. 2004), and second test specific predictions in observational and comparative studies. Testing assumptions of different skew models
The first assumption of the transactional framework is that the presence of subordinates increases group productivity and the dominant’s fitness. Positive relationships between male number and group productivity (e.g. efficiency of defense against extra-group males) have been reported in male primates (Wrangham 1999, Treves 2001). In wild chimpanzees (Pan troglodytes), for example, inter-group aggression is mainly conducted by males (Wrangham 1999), and the number of offspring and probability of infant survival increases with the number of males (Boesch et al. 2006). This pattern could occur through the combined effects of attracting females to the group and better defense of the territory or offspring. In another population, at Mahale, Tanzania (Figure 7.3), researchers documented that a decrease in the number of males in a small group, possibly caused by inter-group killing by the larger neighboring group, resulted in the transfer of females to the larger group (Nishida et al. 1985). Thus, this assumption that subordinates provide fitness benefits and higher group productivity could be met in species where males defend a territory or a
Causes and consequences of skew in male primates
Figure 7.3 Four male chimpanzees at Mahale Mountains National Park, Tanzania, patrolling the border to a neighboring group. Photo by N. Kutsukake.
group of females, and in other situations in which dominant males benefit from membership in multi-male groups. A key assumption of the concession model within the transactional framework is that the dominant individual has complete control over reproduction by subordinates. Field studies provide weak support for this assumption. In most species of social primates, for example, the presence of a dominant individual does not suppress the reproductive states of subordinates, and complete control must be difficult to achieve if there are too many rivals in a group. A dominant male can often interrupt subordinate matings, but in many cases he is ineffective in completely preventing copulations by subordinate males (Soltis 2004). Various studies have further shown that the degree to which the alpha male succeeds in reproduction decreases as the number of rivals increases (van Noordwijk & van Schaik 2004). Finally, complete control should be especially difficult in species living in fission–fusion societies (Dunbar 1988), where subdivision of the group into foraging parties should make it more difficult for males to monitor mating attempts by other males.
171
172
N. Kutsukake, C. L. Nunn These species include chimpanzees, bonobos (Pan paniscus), and spider monkeys (Ateles spp.). An important assumption of the compromise framework is that group productivity decreases as a result of competition between the dominant and subordinate. Infanticide by males is widely observed in primates (van Schaik & Janson 2001) and reduces group productivity. Correlational studies showed that groups with multiple males are less productive than single-male groups (black-and-white colobus, Colobus guereza: Dunbar 1987; Hanuman langurs, Semnopithecus entellus: Srivastava & Dunbar 1996), although the behavioral mechanism for how the presence of multiple males affects male–male competition – and ultimately group productivity – is largely unknown. Moreover, some studies reported positive effects of subordinate males on group productivity (red howler monkeys, Alouatta seniculus: Crockett & Janson 2000; mountain gorillas, Gorilla beringei: Watts 2000). These results suggest that the links between the number of subordinate males in a group and competition among males or group productivity is not universal. Future studies should test this assumption more broadly across species, including species living in multimale groups. This brief summary suggests that males are unlikely to have complete control over reproduction (as assumed in the concession model), and that group productivity can either increase or decrease with the number of males as well as the intensity of competition among dominant and subordinate males (as predicted by transactional and compromise frameworks, respectively). Thus, a particular skew model could be appropriate for some species but not for others, and quantitative testing of the assumptions could help to disentangle which models should be investigated in different species. Other models (and extensions of these models, such as social queuing or models that incorporate multiple individuals) make additional assumptions (Kokko & Johnstone 1999, Johnstone et al. 1999, Reeve & Emlen 2000) that would be worth investigating as skew frameworks are applied to primate mating systems. Testing specific predictions of skew models
Reproductive skew models also make different predictions regarding the effects of demographic variables (number of males and females), female reproductive traits (estrous synchrony), and relatedness among males on patterns of reproductive skew (Table 7.1). The tug-of-war model and the extended POA model predict that skew decreases as the number of males in a group increases, based on the reasoning that it will be difficult for a dominant male to control or monitor reproductive attempts by other males when more rivals are
Causes and consequences of skew in male primates Table 7.1 Predicted relationships between reproductive skew and the number of males and females, estrous overlap, and relatedness among males from three models Models “Extended” Effects on
priority-of-
reproductive skew
Priority-of-access access
Tug-of-war
Concession
Number of males
No prediction
No prediction
No prediction
No prediction
No prediction No prediction
in group Number of females in group Estrous overlap
Relatedness among
No prediction
No prediction
No prediction
males
þ
or
þ, positive relationship between variable and degree of skew; , negative relationship with skew
present (Cowlishaw & Dunbar 1991, van Noordwijk & van Schaik 2004). Similarly, increases in the number of females in a group should decrease skew if this provides more mating opportunities for subordinate males (Altmann 1962, Cowlishaw & Dunbar 1991, Bulger 1993, van Noordwijk & van Schaik 2004). Another prediction from the tug-of-war model and the POA model involves female estrous overlap (Table 7.1). Increased estrous overlap, which results from a long mating season, a long estrus duration, or socially mediated synchrony, should make it more difficult for a dominant male to monopolize a receptive female, thus decreasing skew among males (Ridley 1986, Cowlishaw & Dunbar 1991, Paul 1997, Shuster & Wade 2003). Similar effects can arise if the costs of guarding are high, causing dominant males to guard females for only part of their cycles (Packer 1979, Bercovitch 1983, Alberts et al. 1996). Few mathematical or empirical models of reproductive skew among males have considered the influence of female reproduction and behavior; exceptions include the female control model developed by Cant and Reeve (2002), and studies of acorn woodpeckers (Melanerpes formicivorus: Haydock & Koenig 2002), brown jays (Cyanocorax morio: Williams 2004), and some studies of primates (e.g. Soltis et al. 2001, Charpentier et al. 2005a, Boesch et al. 2006). Many of these exceptions involve female effects on male monopolization, and therefore address assumptions of the tug-of-war model. Furthermore, the socioecological model focuses explicitly on how female reproductive strategies influence male
173
174
N. Kutsukake, C. L. Nunn behavior (Wrangham 1980, Sterck et al. 1997, Dunbar 1988, Nunn 1999a, van Schaik et al. 1999), an area where primatology has the potential to contribute to further development of skew models. Finally, the tug-of-war model predicts no relationship between male relatedness and skew (Table 7.1). This relationship could even be negative in circumstances in which males exert weaker control over close relatives (Reeve et al. 1998). In some circumstances, for example, dominants could increase their inclusive fitness by exerting fewer restrictions on mating by related subordinates, thus generating a negative association between relatedness and skew. The concession model predicts no association between demographic factors or estrous synchrony and skew (Table 7.1). Instead, this model predicts that relatedness will affect patterns of skew, with high relatedness associated with high skew, assuming related subordinate males can receive their “staying incentive” in the form of inclusive fitness benefits (Keller & Reeve 1994, Johnstone 2000). Case studies of the causes of reproductive skew
Data on patterns of reproductive success in primates are becoming increasingly available (van Noordwijk & van Schaik 2004), offering potential for investigating whether compromise or transactional frameworks are more appropriate for studying skew in male primates (Widdig et al. 2004, Setchell et al. 2005, Charpentier et al. 2005a, Bradley et al. 2005, Boesch et al. 2006). As reviewed below, the compromise framework appears to offer a better fit for primate males, and the extended POA model may be equally powerful in explaining patterns of reproductive skew among male primates. Even so, transactional frameworks may account for additional variation in male skew, particularly when males defend territories, as this is one way by which group productivity can increase with the number of males (see above). In what follows, we review four case studies of male skew in multi-male multi-female primate groups (Table 7.2). These examples are not meant to be exhaustive; rather we use selected examples to reveal how skew theory provides new insights to variation in male reproductive success in primates. Rhesus macaque (Macaca mulatta)
In rhesus macaques, males disperse from their natal groups, while females remain in the group in which they were born. Widdig et al. (2004) investigated reproductive skew in a population on Cayo Santiago, Puerto Rico, and found that the top sire fathered between 19% and 30% of the offspring per year over a 6-year period, while 69–79% of males sired no infants at all. In terms
Table 7.2 Summary of empirical studies investigating reproductive skew in male primates Model “Extended” Study site, Species
group
Rhesus
Cayo Santiago,
macaque
a
# male
# female
Other
Relatedness
Estrous
Priority of
priority of
Tug-of-
among males
overlap
access
access
war
Concession
factors
ns (but two
ns
Yes
No
Hetero-
Puerto Rico
most
(Macaca
(free-ranging
successful
mulatta)
provisioned)
males were
important
zygosity
related in two of five years) Mandrill
b, c
CIRMF mandrill
(Mandrillus
colony, Gabon
sphinx)
(semi-freeranging provisioned) Setchell
b
Yes
Yes
-—
—
Yes
Partially yes
+ in one
Not
—
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
et al. 2005
Mountain gorilla
c
Charpentier
+
et al. 2005a Karisoke,
ns
f
avoidance ns
Virunga,
(of four)
investigated,
(Gorilla
Volcanoes
group
but overlap is
beringei)
National Park,
unlikely
Rwanda (wild) Chimpanzee
d
Taı¨ National
(Pan
Park, Cote
troglodytes)
d’Ivoire
—
—
Yes
Yes
(wild) Comparative analyses
Incest
—
ns
ns
ns
Partially yes
f
e
(31 species) þ, positive relationship; , negative relationship; ns, uncorrelated a Widdig et al. 2004 b controlled for the effect of estrous overlap by calculating the deviance of observed data from expected value from the priority-of-access model c Bradley et al. 2005 d Boesch et al. 2006 e Kutsukake and Nunn 2006 f “partial” because the effect of estrous overlap was not confirmed
Causes and consequences of skew in male primates of specific tests, the authors showed that: (1) males exhibited significant variation in skew, with a measure of skew (the B index: Nonacs 2000) significantly different from zero in most tests; (2) the B index was not significantly associated with either average pairwise relatedness among males or female synchrony (estimated indirectly from births); (3) heterozygosity of MHC genes predicted male reproductive success, highlighting the potential role of female choice. The authors concluded that their results support the compromise framework, as the concession model would predict a lower level of relatedness among breeders and stronger control of group reproduction by resident (dominant) males. Mandrill (Mandrillus sphinx)
In the wild, mandrills live in groups of up to several hundred individuals (Abernethy et al. 2002). Behavior in these groups has not been investigated, largely due to the difficulties of habituating and observing behavior of mandrills in their natural habitat. Important information on this species has been provided by research from a semi-free-ranging captive colony (CIRMF Mandrill Colony, Gabon). In this population, only the alpha male exhibits the distinctive secondary sexual traits (e.g. bright facial color) characteristic of this extremely sexually dimorphic species. Although multiple males are present in the colony, paternity analyses have shown that the alpha male fathers 69% of offspring, indicating extreme reproductive skew among males (Setchell et al. 2005). Two studies have investigated different aspects of reproductive skew in this colony. Although the authors studied the same groups, some results differed between the studies, in part due to differences in the specific aims of each study, in samples collected and variables analyzed, and in statistical approaches. In one of these studies, Setchell et al. (2005) investigated deviations in the alpha male’s reproductive success from the expected value based on the POA model. The authors showed that departures from the POA model increased as the number of males in a group increased (Table 7.2), which fits predictions from the extended POA and the tug-of-war models. By comparison, Charpentier et al. (2005a) studied factors affecting the failure of alpha males to sire offspring. These authors reported that relatedness among males, female estrous synchrony, relatedness between the dominant male and females, and the number of males affected paternity of the dominant male. Specifically, (1) the dominant male’s reproduction decreased as relatedness among males increased; (2) estrous overlap decreased reproduction by the dominant male; (3) relatedness between the dominant male and females negatively affected reproduction by the alpha male. Although the behavioral mechanism for this
177
178
N. Kutsukake, C. L. Nunn result is unknown, incest avoidance may have played a role, because the degree of heterozygosity correlated positively with individual reproductive success (Charpentier et al. 2005b). (4) In contrast to predictions of the tug-of-war model, the number of males correlated positively with the proportion of offspring sired by the alpha male (Charpentier et al. 2005a). To explain this result, Charpentier et al. (2005a) suggested that competition among subordinates increased with the number of males, deflecting competition away from the dominant male. Although the effect of male number differed between the studies, both Setchell et al. (2005) and Charpentier et al. (2005a) concluded that the limited control model best characterized this species; predictions of the concession model were never supported. Setchell et al. (2005) also noted that conditions in wild mandrills might produce weaker patterns of control than those found in the colony studied by these authors. Mountain gorilla (Gorilla beringei)
There is variation in the number of males in groups of mountain gorillas in the Virunga Mountains, Rwanda, with multi-male groups representing 40% of the groups in the population (Robbins et al. 2001). Female reproductive cycles are short, and it is rare that the receptive periods of two or more females overlap. The lack of overlap should tend to enable the dominant male to monopolize reproduction within a group. However, paternity analyses have shown that subordinates also reproduce to some extent (about 15%), suggesting that reproductive skew (estimated by the B index) is high but not complete (Bradley et al. 2005). In another study, Robbins & Robbins (2005) used an individual-based simulation model with demographic parameters from the same population studied by Bradley et al. (2005) to investigate the expected reproductive success of subordinates that remain in their group. The model revealed that remaining in a group benefits a subordinate more than dispersing. However, the model revealed that the dominant does not benefit from retention of subordinates, suggesting that dominant males do not concede reproduction. Thus, both Bradley et al. (2005) and Robbins & Robbins (2005) concluded that reproductive skew in this population corresponds better to predictions of the tug-of-war model than those of concession models. Chimpanzee (Pan troglodytes)
In chimpanzees, males remain in their natal group and exhibit a high degree of fission–fusion sociality. Females develop sexual swellings when they are in estrus, with synchronous estrous relatively common. The dominant
Causes and consequences of skew in male primates male has higher reproductive success, but subordinate males also reproduce (Constable et al. 2001). Boesch et al. (2006) investigated paternity in chimpanzees from Taı¨ National Park, Cote d’Ivoire, using long-term records. They found that the alpha male’s proportion of reproduction decreased as the number of males increased and when female estrous overlap increased. These results therefore agree with predictions from both the extended POA model and the tug-of-war model. Summary of case studies
Overall, these studies suggest that limited control is a characteristic of male behavior in primates (Table 7.2), and that the tug-of-war model or the extended POA model can explain variation in skew among male primates. However, these studies do not completely reject the concession model, for the following reasons. First, empirical studies mainly tested predictions from mathematical models that were designed for systems other than primates, often assuming that the group contains only two individuals – a dominant and a subordinate. In contrast, mathematical models that incorporate more realistic parameters, such as three or more group members or the possibility of social queuing by subordinates, predict a reduced necessity of offering incentives by a dominant individual to a subordinate (in particular to unrelated subordinates) relative to the two-player models (Kokko & Johnstone 1999, Ragsdale 1999, Johnstone et al. 1999, Reeve & Emlen 2000, Reeve & Shen 2006). This makes it difficult to draw firm predictions for how relatedness should correlate with patterns of skew. Second, no studies in primates have succeeded in accurately quantifying parameters that are necessary to test the skew model, in large part because it is difficult to conduct experimental manipulations in primates and to quantify model parameters. Finally, the concession and tug-of-war models are not mutually exclusive, and can in fact coexist within a single framework (Johnstone 2000, Reeve & Shen 2006). Phylogenetic comparative analyses
Skew models have been regarded as a unifying framework for understanding the diversity of social systems seen in animals (Keller & Reeve 1994, Sherman et al. 1995), but surprisingly few studies have examined broad evolutionary patterns of skew within one clade of either vertebrates or invertebrates (Faulkes et al. 1997, Boomsma & Sundstro¨m 1998, Duffy et al. 2000). Such comparative perspectives are important in reproductive skew research for at least four reasons. First, comparative studies provide a means to understand the factors generating broad evolutionary patterns of skew (Nonacs 2000) and therefore can assess the generality of a pattern across species, leading to
179
180
N. Kutsukake, C. L. Nunn
Figure 7.4 Phylogenetic comparative analyses on the relationship between male number in a group and the maximum mating proportion (the proportion of mating by the most successful male). For further details on these analyses, see Kutsukake & Nunn (2006).
greater unification of models of social evolution. Second, comparative approaches offer an opportunity to test assumptions and predictions of skew models from an evolutionary perspective. Third, by identifying differences among species, comparative results can point to new variables to investigate in future field or laboratory research. Finally, comparative research can be used to generate new hypotheses, which can then be tested in the field or laboratory, or refined through theoretical models. Prior cross-species comparisons in primates examined the effects of seasonality on variance in mating or reproductive success (Cowlishaw & Dunbar 1991, Paul 1997). We conducted a phylogenetic comparative analysis on the determinants of “mating” skew in male primates, based on a database of species in multi-male primate groups (in total from 84 studies, representing 31 species in 17 genera: Kutsukake & Nunn 2006). Since few studies have investigated the distribution of paternity for a sufficient number of primate species, we investigated mating distribution. While many studies have shown that mating frequency predicts reproductive success (e.g. Smith 1981, Pope 1990, Ohsawa et al. 1993, de Ruiter et al. 1994, Paul & Kuester 1996, Soltis et al. 1997, Alberts et al. 2006), other studies failed to find such links (e.g. Curie-Cohen et al.
Causes and consequences of skew in male primates 1983, Shively & Smith 1985, Inoue et al. 1991, 1993), possibly because many matings in primates are likely to be non-reproductive (Soltis 2004). To deal with this problem, we used data that are most tightly linked to male reproductive success whenever possible; specifically, we preferred data on ejaculation frequency more than copulation frequencies, and copulation data at times when conception was most likely to take place (Kutsukake & Nunn 2006). Genetic information on actual reproduction in groups would clarify these issues and allow skew to be examined more directly, but such data are not yet sufficiently available to test the predictions in a comparative context. In quantifying the magnitude of mating skew, we focus here on results using the “maximum mating proportion” (Bulger 1993), which is the proportion of matings by the most successful male. We also examined other skew indices, including the B index (Nonacs 2000) and lambda (Kokko & Lindstro¨m 1997). We investigated the effects of three variables: demographic factors (the number of males or females in a group), female reproductive factors associated with the difficulties of monopolizing estrous females (i.e. duration of the breeding season, duration of estrus, and measures of estrous overlap), and male dispersal pattern (categorized as male philopatry or male dispersal). Regarding male dispersal pattern, the concession model predicts high skew in male-philopatric species relative to species in which males disperse because there is (1) a high probability that a dominant male has a brother within a group and (2) a lower probability that subordinates will disperse. Taken together, these factors reduce the need for the dominant male to provide a staying incentive. The main results of our study (Kutsukake & Nunn 2006) can be summarized as follows. First, based on Nonac’s B index, mating was significantly skewed among males in 75.4% of cases (43/57 cases), and the alpha male or resident male tended to mate more frequently. Second, using the independent contrasts method (Felsenstein 1985) and stepwise multiple regression, we found that only male number correlated with mating skew (P < 0.001), with the proportion of mating by the most successful male falling as the number of males in a group increases (Figure 7.4). Finally, neither female reproductive proxies nor male dispersal pattern affected mating skew. Overall, these results are most consistent with the tug-of-war model and partially consistent with the extended POA model (in the sense that the number of males negatively affected skew). This result raises the possibility that the effects of estrous synchrony are not universal to all primate species, that its effects are weak, or that synchrony is difficult to quantify, all of which would limit our ability to detect a significant association in comparative analyses given existing data. Although the intensity
181
182
N. Kutsukake, C. L. Nunn of the correlation between dominance rank and reproductive success was affected by seasonality (Paul 1997), so far, few studies have investigated paternity among males in relation to estrous synchrony (Setchell et al. 2005, Charpentier et al. 2005a, Boesch et al. 2006). One could argue that the concession model also predicts that mating skew should decrease as the number of males increases, specifically if the dominant male needs to pay staying incentives to each subordinate male. However, we also found a similar negative relationship in an intraspecific analysis of wild chimpanzees (Kutsukake & Nunn 2006). The negative relationship is not expected in a male-philopatric species, such as the chimpanzee, because subordinate males have few opportunities for reproduction outside of their natal communities, and therefore do not need an incentive to stay. Even with this intraspecific analysis, however, we cannot firmly reject the concession model. For example, a negative relationship between the number of males and mating skew can also be explained by the concession model because reproductive skew may decrease when the power difference between a dominant and subordinate is small (e.g. in a group with many males: Cowlishaw & Dunbar 1991); therefore, the dominant may concede the reproduction as a “peaceful” incentive to avoid a risky fight with powerful rival males (Reeve & Ratnieks 1993). Indeed, the power differences may be smaller in a group with a large number of males because one would expect that males are, on average, more similar in age (and therefore competitive ability). This idea needs further testing, but highlights the difficulty of testing between different skew models, even in well-studied mammalian species. In addition, our comparative study does not reject the possibility that the concession model applies to a particular primate species, even if it is not a general explanation for patterns of skew across primates. As is shown by a recent synthetic model, the transactional framework and compromise framework are not mutually exclusive (Johnstone 2000, Reeve & Shen 2006). So one model may fit one species but not others, or in certain demographic or ecological situations but not in others within a species. For example, even within a species, the dominant male may be able to exert complete control in a small group in which there is only one subordinate, but not in a large group with multiple subordinates. This possibility can be tested by investigating how the effects of relatedness on reproductive skew vary according to the number of subordinate males in groups. Although we focused on short time intervals, such as a single breeding season, our approach can be used to examine complex life-history trajectories (patterns of lifetime reproductive success). In addition, it could be applied to both sexes. For example, reproductive success among female primates can be
Causes and consequences of skew in male primates estimated using long-term data. Finally, it would be interesting to apply this approach to other clades in which data on reproduction and phylogeny are widely available, such as birds and social insects, and in other well-studied mammalian groups, such as rodents, ungulates, and carnivores. Applying comparative approaches to other biological systems
Comparative tests can focus on either the predictions or the assumptions of skew models, and testing is possible if researchers have quantitative data for the distribution of reproduction or mating among group males. Here, in an attempt to stimulate further comparative research in other clades of animals, we list several methodological practices for conducting comparative tests of predictions related to reproductive skew models. (1)
(2)
(3)
(4)
Carefully choose the hypotheses, predictions, or assumptions to be tested. Within the framework of the models and the biology of the organisms, the researcher needs to consider alternative explanations and how different parameters might influence the predictions of a skew model. It is also important to incorporate the characteristics that are specific to the study animals, because some parameters are difficult to quantify in some clades. Collect data on mating or reproductive skew and other important variables such as group composition (e.g. number of males and females), relatedness, female behaviors, and reproductive biology. Data on reproduction are available in many non-primates (e.g. Ellis 1995), which could be used for comparative analyses. It is also important to obtain a phylogeny for the group of species being studied. “Supertrees” and other large-scale, dated phylogenies are now available for many species (Bininda-Emonds 2004), making this process easier than in the past. Quantify the distribution of reproduction using several skew proxies (Nonacs 2003). Many studies will not provide these measures directly, and may not even provide information for the comparative biologist to calculate the measures. Thus it might be necessary to use a simple index that maximizes sample size (in terms of the number of species). Test the hypotheses using phylogenetic comparative methods, such as independent contrasts (Felsenstein 1985, Harvey & Pagel 1991, Nunn & Barton 2001). It is important to check whether the data show phylogenetic signal (Blomberg & Garland 2002), to test the statistical and evolutionary assumptions, and to determine whether the results are robust to alternative assumptions.
183
184
N. Kutsukake, C. L. Nunn Consequences of reproductive skew Previous studies have mainly investigated the causes of skew, and have tested specific models. An important new direction in skew research is to consider the consequences of reproductive skew on other biological traits, including social structure and individual social strategies (Figure 7.1; Heinze 1995, Widdig et al. 2001, Cant & English 2006). For example, in some systems, the number of breeders and characteristics of the breeding queue could influence optimal group size (Cant & English 2006). With the goal of developing new questions for future studies, we briefly discuss two consequences of reproductive skew in male primates: (1) effects on within-group relatedness and (2) the spread of disease. Reproductive skew and within-group relatedness
In species characterized by high skew, infants born within a short period are more likely to be paternally related. For example, Widdig et al. (2004) found that in a high-skew rhesus macaque troop at Cayo Santiago, 74% of the infants had at least one paternal sibling in the group, and individuals had almost four times as many paternal as maternal siblings. In contrast, infants in low-skew societies are more likely to be fathered by different males, thus tending to reduce the level of relatedness at the group level. The paternal relatedness among group members should have a major impact on a wide range of social behaviors, including affiliation, cooperation, competition, and mate choice (Hamilton 1964, Chapais & Berman 2004). Several studies have suggested that individuals recognize paternal relationships and adjust their behavior accordingly. For example, skew is high in male western gorillas, and the silverbacks of different groups are closely related (Bradley et al. 2004). This result may explain the occurrence of nonagonistic encounters between groups observed in this species – unexpected in such a sexually dimorphic species, in which male–male competition is likely to be intense. Paternal half-siblings are more affiliative with one another than unrelated individuals in rhesus macaques (Widdig et al. 2001) and in savannah baboons (Papio cynocephalus: Smith et al. 2003, see also Silk et al. 2006). Furthermore, in baboons, paternal half-siblings showed less affiliative and sexual behavior during consortships than did unrelated pairs (Alberts 1999). As a final example, infants were supported by a biological father (Buchan et al. 2003) or were not the target of infanticide by the biological father in species living in multi-male groups (Borris et al. 1999a, 1999b, Soltis et al. 2000).
Causes and consequences of skew in male primates When reproductive skew is high and the dominant male’s tenure is long enough for his female offspring to become sexually mature, it may be adaptive for the dominant male to discriminate the paternity of the offspring and avoid mating with his daughters. In wild white-faced capuchin monkeys (Cebus capucinus), for example, the probability of reproduction by the alpha male varied with whether or not a female was a daughter of the alpha male, with a lower probability of reproduction between the alpha male and his daughter (Muniz et al. 2006). It would be interesting to investigate whether such incestavoidance mechanisms are more common in primates, especially since previous studies yielded inconclusive results (Table 7.2; Constable et al. 2001). If incest avoidance is an important selective force, strong skew combined with long male tenure could reduce future opportunities for the alpha male to reproduce within a group, thus creating an incentive for secondary dispersal. In addition, it will be important to uncover the proximate mechanisms responsible for identifying paternal kin (Rendall 2004). It is possible, for example, that dominant males make use of their information on the monopolization of receptive females as a proximate cue to assess the probability of their paternity. Similarly, for human observers, it may be possible to estimate the magnitude of reproductive skew a posteriori from the genetic relatedness among infants and juveniles in a group. Reproductive skew and the spread of infectious disease
Reproductive skew can also have consequences for patterns of social contact within social units, thus affecting the spread of disease within primate groups (Nunn & Altizer 2006). In a high-skew primate group under the tug-ofwar model, for example, one or a few males will gain access to receptive females. Thus there is likely to be intense competition among males as they fight to improve or maintain their dominance ranks. This fighting causes wounds through biting and scratching and may result in the spread of disease, as demonstrated in the case of retroviruses (SIV and STLV) in a semi-freeranging colony of mandrills (Nerrienet et al. 1998). In addition to being involved in male intrasexual competition, a high-ranking male in a high-skew society also has better access to mates, resulting in higher rates of sexual contact. Thus, such a male can act as a contact point for sexually transmitted diseases (STDs: Graves & Duvall 1995), potentially even selecting for reduced skew (Thrall et al. 2000, Kokko et al. 2002). If a female is already infected with an STD at the time a new male rises in rank, this male is likely to become infected shortly after he attains high rank; he can thus serve as the source of infection for the many females that he mates with during his tenure as the alpha male.
185
186
N. Kutsukake, C. L. Nunn A number of models have investigated the epidemiology of STDs in both human (Anderson & May 1991) and non-human systems (Thrall et al. 1997, Boots & Knell 2002, Kokko et al. 2002). In the context of variance in male mating skew, for example, Thrall et al. (2000) developed an individual-based model to explore how variance in mating success, patterns of female dispersal, and mortality rates of both sexes influence the spread of STDs. Given that the simulated population had an equal number of males and females, every male would have one female if there was no skew (equivalent to monogamy); each additional female assigned to a male means one less female for another male, resulting in increased reproductive skew. The simulations revealed that the prevalence of STDs is higher as the degree of polygyny (reproductive skew) increases. A challenge in applying these concepts to generate testable predictions is that low skew in multi-male multi-female primate groups can also favor the spread of an STD. Thus, if males have relatively equal access to females, this could result in a higher rate of mating with more males throughout the female’s cycle, possibly as a strategy to reduce the risk of infanticide (Hrdy & Whitten 1987, van Schaik et al. 1999). And of course, increased promiscuity should increase the spread of an STD (Anderson & May 1991). This promiscuity is likely to increase the prevalence to even higher levels than revealed by models of STD spread under skew (e.g. Thrall et al. 2000), especially if most subordinates have some mating success. The STD model of Thrall et al. (2000) provides a way out of this conundrum, however, because it also predicts a higher prevalence of infection in females than in males as reproductive skew increases, i.e. a sex difference is predicted. Kokko et al. (2002), in a different modeling approach, confirmed that female choice for a particular (presumably high-ranking) male can also lead to higher prevalence of infection in females. Thus, a critical prediction is that higher skew will produce not only high prevalence (relative to, say, monogamy); increasing skew should also produce a sex difference in the prevalence of an STD, with higher prevalence in females than in males. This prediction has been tested and supported using data on STDs in primates (Nunn & Altizer 2004). A next step is to examine whether sex differences also correlate with skew and other variables influencing the establishment of an STD, including mortality rates, dispersal, and differences in transmission probabilities between the sexes (e.g. with females potentially being more susceptible to an STD). In addition, it will be important to bring queuing or age-dependency into the STD models, because if most individual males have some mating opportunities over their lifetimes, the difference in STD exposure between the sexes may become more narrow.
Causes and consequences of skew in male primates Conclusion This chapter has discussed the causes and consequences of reproductive skew in male primates. Empirical studies have shown that the tug-ofwar model may better explain the pattern of skew among males than the concession model. Our comparative study revealed a negative association between the number of males in a group and skew, which agrees with previous findings in primates (Setchell et al. 2005, Boesch et al. 2006, reviewed in van Noordwijk & van Schaik 2004) and also agrees with predictions from the tug-ofwar model. Therefore, we tentatively conclude that incomplete control is a general characteristic of male primates, but more studies are needed to test the assumptions or predictions of the concession model. The priority-of-access (POA) model (Altmann 1962) highlighted the effect of estrous overlap on the distribution of reproduction among males in multi-male multi-female groups, including non-primates. A major conclusion of our chapter is that the POA model – especially an extended version that incorporates the number of males – is almost indistinguishable from the compromise framework. This is particularly true with regard to the model predictions (Table 7.1). It might therefore seem that the skew framework represents “new wine in an old bottle.” This would be misleading, however, as the skew framework is actually much broader than the previous POA model. For example, similar to models of reproductive skew, it builds significantly on POA by encapsulating factors involving relatedness, breeding opportunities, and costs of dispersal. Several challenges remain for the future. First, the present mathematical models of reproductive skew were not designed to investigate primate social systems. In particular, it would be worthwhile to develop skew models that incorporate three or more players (Johnstone et al. 1999, Reeve & Emlen 2000), social queuing (Kokko & Johnstone 1999, Ragsdale 1999, Mesterton-Gibbons et al. 2006), female influences such as incest avoidance (Cant & Reeve 2002, Johnstone 2000), and female choice for males with particular biological traits (“good genes” or high dominance rank). Recent mathematical models in which one individual adjusts behavior in response to the behavior of the other individual (negotiation game: McNamara et al. 1999, Cant & Shen 2006) may be more appropriate in primates, because social interactions in primates change temporally according to the strategy of opponents. Also, an individual-based model based on empirical demographic parameters would be a useful tool for generating more refined predictions for patterns of skew in primates (e.g. Robbins & Robbins 2005).
187
188
N. Kutsukake, C. L. Nunn Second, no empirical studies of primates have successfully estimated the parameters that are needed to distinguish among the different skew models. These parameters include the links between competition within groups and group productivity and ecological constraint that determine the probability of solitary reproduction. This may represent a limitation of skew theory, with very few predictions distinguishing the different models. Nonetheless, experimental studies, including manipulating group composition, would help to more formally test skew theory in primates. Such tests could be conducted in semi-free-ranging groups. Third, most of the studies in primates estimate skew in a relatively short time period. Thus, it is unknown how short-term skew is associated with longterm (i.e. lifespan) reproductive success (Altmann et al. 1996). The consequences of reproductive skew have been largely unexplored, yet these topics offer great opportunities for future research in primates. Independent of the causes of skew, how a given level of skew affects social structure, individual decision making, and other biological traits that relate to reproduction is a promising area for both empirical and theoretical research. For example, investigating the relationship between skew and the prevalence of STDs could have important implications for conservation biology, given that STDs often cause sterility (Canfield et al. 1991, Lockhart et al. 1996). In conclusion, bringing the skew paradigm to primatology may yield new perspectives for understanding primate behavior, specifically by integrating more diverse factors that are relevant to male and female decisions on group formation, interactions within groups, and reproductive strategies. Thus, skew models could play a major role in developing an integrative model of primate socioecology. Key future directions will involve developing skew models that are more appropriate for primates, collecting data to test the assumptions and predictions of these models, and investigating the consequences of reproductive skew for primate behavior. Moreover, a primate perspective on reproductive skew should help to ground models of skew more firmly, specifically in the context of multiple competitors and queuing within groups.
Acknowledgments We thank Reinmar Hager and Clara B. Jones for their invitation to contribute this chapter, and Mike Cant, Sarah Hodge, Kavita Isvaran, Jo Setchell, and two anonymous reviewers for helpful comments and discussion. This study was supported by JSPS Research Fellowships, RIKEN Special PostDoctoral Researchers Program, financed by JSPS core-to-core program HOPE (to NK) and the Max Planck Society (to CN).
Causes and consequences of skew in male primates References Abernethy, K. A., White, L. J. T., & Wickings, E. J. (2002). Hordes of mandrills (Mandrillus sphinx): extreme group size and seasonal male presence. Journal of Zoology, London, 258, 131–137. Alberts, S. C. (1999). Paternal kin discrimination in wild baboons. Proceedings of the Royal Society of London B, 266, 1501–1506. Alberts, S. C., Altmann, J., & Wilson, M. L. (1996). Mate guarding constrains foraging activity of male baboons. Animal Behaviour, 51, 1269–1277. Alberts, S. C., Watts, H. E., & Altmann, J. (2003). Queuing and queue-jumping: longterm patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Animal Behaviour, 65, 821–840. Alberts, S. C., Buchan, J. C., & Altmann, J. (2006). Sexual selection in wild baboons: from mating opportunities to paternity success. Animal Behaviour, 72, 1177– 1196. Altmann, S. A. (1962). A field study of the sociobiology of the rhesus monkey, Macaca mulatta. Annals of the New York Academy of Sciences, 102, 338–435. Altmann, J., Alberts, S. C., Haines, S. A., et al. (1996). Behavior predicts genetic structure in a wild primate group. Proceedings of the National Academy of Sciences of the USA, 93, 5797–5801. Anderson, R. M. & May, R. M. (1991). Infectious Diseases of Humans: Dynamics and Control. Oxford: Oxford University Press. Bercovitch, F. B. (1983). Time budgets and consortships in olive baboons (Papio anubis). Folia Primatologica, 41, 180–190. Bininda-Emonds, O. R. P., ed. (2004). Phylogenetic Supertrees: Combining Information to Reveal the Tree of Life. Dordrecht: Kluwer. Blomberg, S. P. & Garland, T., Jr. (2002). Tempo and mode in evolution: phylogenetic inertia, adaptation and comparative methods. Journal of Evolutionary Biology, 15, 899–910. Boesch, C., Kohou, G., Nene, H., & Vigilant, L. (2006). Male competition and paternity in wild chimpanzees of the Tai forest. American Journal of Physical Anthropology, 130, 103–115. Boomsma, J. J. & Sundstro¨m, L. (1998). Patterns of paternity skew in Formica ants. Behavioral Ecology and Sociobiology, 42, 85–92. Boots, M. & Knell, R. J. (2002). The evolution of risky behaviour in the presence of a sexually transmitted disease. Proceedings of the Royal Society of London B, 269, 585–589. Borries, C. K. L., Epplen, C., Epplen J. T., & Winkler, P. (1999a). Males as infant protectors in Hanuman langurs (Presbytis entellus) living in multimale groups: defense pattern, paternity and sexual behavior. Behavioral Ecology and Sociobiology, 46, 350–356. Borries, C., Launhardt, K., Epplen, C., Epplen, J. T., & Winkler, P. (1999b). DNA analyses support the hypothesis that infanticide is adaptive in langur monkeys. Proceedings of the Royal Society of London B, 266, 901–904.
189
190
N. Kutsukake, C. L. Nunn Bradley, B. J., Doran-Sheehy, D. M., Lukas, D., Boesch, C., & Vigilant, L. (2004). Dispersed male networks in western gorillas. Current Biology, 14, 510–513. Bradley, B. J., Robbins, M. M., Williamson, E. A., et al. (2005). Mountain gorilla tug-ofwar: silverbacks have limited control over reproduction in multimale groups. Proceedings of the National Academy of Sciences of the USA, 102, 9418–9423. Buchan, J. C., Alberts, S. C., Silk, J. B., & Altmann, J. (2003). True paternal care in a multi-male primate society. Nature, 425, 179–181. Bulger, J. B. (1993). Dominance rank and access to estrous females in male savanna baboons. Behaviour, 127, 67–103. Canfield, P. J., Love, D. N., Mearns, G., & Farram, E. (1991). Chlamydial infection in a colony of captive koalas. Australian Veterinary Journal, 68, 167–169. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169. Cant, M. A. & English, S. (2006). Stable group size in cooperative breeders: the role of inheritance and reproductive skew. Behavioral Ecology, 17, 560–568. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Cant, M. A. & Shen, S.-F. (2006). Endogenous timing in competitive interactions among relatives. Proceedings of the Royal Society of London B, 273, 171–178. Chapais, B. & Berman, C. M. (2004). Kinship and Behavior in Primates. Oxford: Oxford University Press. Charpentier, M., Peignot, P., Hossaert-McKey, M., et al. (2005a). Constraints on control: factors influencing reproductive success in male mandrills (Mandrillus sphinx). Behavioral Ecology, 16, 614–623. Charpentier, M., Setchell, J. M., Prugnolle, F., et al. (2005b). Genetic diversity and reproductive success in mandrills (Mandrillus sphinx). Proceedings of the National Academy of Sciences of the USA, 102, 16723–16728. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Constable, J., Ashley, M., Goodall, J., & Pusey, A. (2001). Noninvasive paternity assignment in Gombe chimpanzees. Molecular Ecology 10, 1279–1300. Cowlishaw, G. & Dunbar, R. I. M. (1991). Dominance rank and mating success in male primates. Animal Behaviour, 41, 1045–1056. Crockett, C. M. & Janson, C. H. (2000). Infanticide in red howlers: female group size, male membership, and a possible link to folivory. In C. P. van Schaik & C. H. Janson, eds., Infanticide by Males and its Implications. Cambridge: Cambridge University Press, pp. 75–98. Curie-Cohen, M., Yoshihara, D., Luttrell, L., et al. (1983). The effects of dominance on mating behavior and paternity in a captive troop of rhesus monkeys (Macaca mulatta). American Journal of Primatology, 5, 127–138. de Ruiter, J. R., van Hooff, J. A. R. A. M., & Scheffrahn, W. (1994). Social and genetic aspects of paternity in wild long-tailed macaques (Macaca fascicularis). Behaviour, 129, 203–224. Duffy, J. E., Morrison, C. L., & Rios, R. (2000). Multiple origins of eusociality among sponge-dwelling shrimps (Synalpheus). Evolution, 54, 503–516.
Causes and consequences of skew in male primates Dunbar, R. I. M. (1987). Habitat quality, population dynamics, and group composition in colobus monkeys (Colobus guereza). International Journal of Primatology, 8, 299–329. Dunbar, R. I. M. (1988). Primate Social Systems. Ithaca, NY: Cornell University Press. Ellis, L. (1995). Dominance and reproductive success among nonhuman animals: a cross-species comparison. Ethology and Sociobiology, 16, 257–333. Emlen, S. T. & Oring, L. W. (1977). Ecology, sexual selection, and the evolution of mating systems. Science, 197, 215–223. Faulkes, C. G., Bennett, N. C., Bruford, M. W., et al. (1997). Ecological constraints drive social evolution in the African mole-rats. Proceedings of the Royal Society of London B, 264, 1619–1627. Felsenstein, J. (1985). Phylogenies and the comparative method. American Naturalist, 125, 1–15. Graves, B. M. & Duvall, D. (1995). Effects of sexually transmitted diseases on heritable variation in sexually selected systems. Animal Behaviour, 50, 1129–1131. Hager, R. (2003). Models of reproductive skew applied to primates. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 65–101. Hamilton, W. D. (1964). The genetical evolution of social behaviour. Journal of Theoretical Biology, 7, 1–16, 17–52. Harvey, P. H. & Pagel, M. D. (1991). The Comparative Method in Evolutionary Biology. Oxford: Oxford University Press. Haydock, J. & Koenig, W. D. (2002). Reproductive skew in the polygynandrous acorn woodpecker. Proceedings of the National Academy of Sciences of the USA, 99, 7178–7183. Heinze, J. (1995). Reproductive skew and genetic relatedness in Leptothorax ants. Proceedings of the Royal Society of London B, 261, 375–379. Hrdy, S. B. & Whitten, P. L. (1987). Patterning of sexual activity. In B. B. Smuts, D. L. Cheney, R. W. Seyfarth, R. W. Wrangham, & T. T. Struhsaker, eds., Primate Societies. Chicago, IL: University of Chicago Press, pp. 370–384. Inoue, M., Mitsunaga, F., Ohsawa, H., et al. (1991). Male mating behaviour and paternity discrimination by DNA fingerprinting in a Japanese macaque group. Folia Primatologica, 56, 202–210. Inoue, M., Mitsunaga, F., Nozaki, M., et al. (1993). Male dominance rank and reproductive success in an enclosed group of Japanese macaques: with special reference to post-conception mating. Primates, 34, 503–511. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A., Woodroffe, R., Cant, M. A., & Wright, J. (1999). Reproductive skew in multimember groups. American Naturalist, 153, 315–331. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 265, 571–578.
191
192
N. Kutsukake, C. L. Nunn Kokko, H. & Lindstro¨m, J. (1997). Measuring mating skew. American Naturalist, 149, 794–799. Kokko, H., Ranta, E., Ruxton, G., & Lundberg, P. (2002). Sexually transmitted disease and the evolution of mating systems. Evolution, 56, 1091–1100. Kutsukake, N. & Nunn, C. L. (2006). Comparative tests of reproductive skew in male primates: the roles of demographic factors and incomplete control. Behavioral Ecology and Sociobiology, 60, 695–706. Lockhart, A. B., Thrall, P. H., & Antonovics, J. (1996). Sexually transmitted diseases in animals: ecological and evolutionary implications. Biological Reviews of the Cambridge Philosophical Society, 71, 415–471. Magrath, R. D., Heinsohn, R. G., & Johnstone, R. A. (2004). Reproductive skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. McNamara, J. M., Gasson, C. E., & Houston, A. I. (1999). Incorporating rules for responding into evolutionary games. Nature, 401, 368–371. Mesterton-Gibbons, M., Hardy, I. C. W., & Field, J. (2006). The effect of differential survivorship on the stability of reproductive queueing. Journal of Theoretical Biology, 242, 699–712. Muniz, L., Perry, S., Manson, J. M., et al. (2006). Father–daughter inbreeding avoidance in a wild primate population. Current Biology, 16, R156–157. Nerrienet, E., Amouretti, X., Muller-Trutwin, M. C., et al. (1998). Phylogenetic analysis of SIV and STLV type I in mandrills (Mandrillus sphinx): indications that intracolony transmissions are predominantly the result of male-to-male aggressive contacts. Aids Research and Human Retroviruses, 14, 785–796. Nishida, T., Hiraiwa-Hasegawa, M., Hasegawa, T., & Takahata, Y. (1985). Group extinction and female transfer in wild chimpanzees of the Mahale Mountains. Zeitschrift fu¨r Tierpsychologie, 67, 284–301. Nonacs, P. (2000). Measuring and using skew in the study of social behavior and evolution. American Naturalist, 156, 577–589. Nonacs, P. (2003). Measuring the reliability of skew indices: is there one best index? Animal Behaviour, 65, 615–627. Nunn, C. L. (1999a). The number of males in primate social groups: a comparative test of the socioecological model. Behavioral Ecology and Sociobiology, 46, 1–13. Nunn, C. L. (1999b). The evolution of exaggerated sexual swellings in primates and the graded signal hypothesis. Animal Behaviour, 58, 229–246. Nunn, C. L. & Altizer, S. M. (2004). Sexual selection, behavior and sexually transmitted diseases. In P. M. Kappeler & C. P. van Schaik, eds., Sexual Selection in Primates: a Comparative Perspective. Cambridge: Cambridge University Press, pp. 117–130. Nunn, C. L. & Altizer, S. M. (2006). Infectious Diseases in Primates: Behavior, Ecology and Evolution. Oxford: Oxford University Press. Nunn, C. L. & Barton, R. A. (2001). Comparative methods for studying primate adaptation and allometry. Evolutionary Anthropology, 10, 81–98.
Causes and consequences of skew in male primates Ohsawa, H., Inoue, M., & Takenaka, O. (1993). Mating strategy and reproductive success of male patas monkeys (Erythrocebus patas). Primates, 34, 533–544. Packer, C. (1979). Male dominance and reproductive activity in Papio anubis. Animal Behaviour, 37, 37–45. Paul, A. (1997). Breeding seasonality affects the association between dominance and reproductive success in non-human male primates. Folia Primatologica, 68, 344–349. Paul, A. & Kuester, J. (1996). Differential reproduction in male and female Barbary macaques. In J. E. Fa & D. G. Lindburg, eds., Evolution and Ecology of Macaque Societies. Cambridge: Cambridge University Press, pp. 293–317. Pope, T. R. (1990). The reproductive consequences of male cooperation in the red howler monkey: paternity exclusion in multi-male and single-male troops using genetic markers. Behavioral Ecology and Sociobiology, 27, 439–446. Ragsdale, J. E. (1999). Reproductive skew theory extended: the effect of resource inheritance on social organization. Evolutionary Ecological Research, 1, 859–874. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Ratnieks, F. L.W. (1993). Queen–queen conflicts in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. New York, NY: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S. -F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Rendall, D. (2004). “Recognizing” kin: mechanisms, media, minds, modules and muddles. In B. Chapais & C. Berman, eds., Kinship and Behavior in Primates. Oxford: Oxford University Press, pp. 295–316. Ridley, M. (1986). The number of males in a primate troop. Animal Behaviour, 34, 1848–1858. Robbins, A. M. & Robbins, M. M. (2005). Fitness consequences of dispersal decisions for male mountain gorillas (Gorilla beringei beringei). Behavioral Ecology and Sociobiology, 58, 295–309. Robbins, M. M., Sicotte, P., & Stewart, K. J. (2001). Mountain Gorillas: Three Decades of Research at Karisoke. Cambridge: Cambridge University Press. Say, L., Pontier, D., & Natoli, E. (2001). Influence of oestrus synchronization on male reproductive success in the domestic cat (Felis catus L.). Proceedings of the Royal Society of London B, 268, 1049–1053. Setchell, J. M., Charpentier, M., & Wickings, E. J. (2005). Mate-guarding and paternity in mandrills (Mandrillus sphinx): factors influencing monopolisation of females by the alpha male. Animal Behaviour, 70, 1105–1120. Sherman, P. W., Lacey, E. A., Reeve, H. K., & Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108.
193
194
N. Kutsukake, C. L. Nunn Shively, C. & Smith, D. G. (1985). Social status and reproductive success of male Macaca fascicularis. American Journal of Primatology, 9, 129–135. Shuster, S. M. & Wade, M. J. (2003). Mating Systems and Mating Strategies. Princeton, NJ: Princeton University Press. Silk, J. B., Altmann, J., & Alberts, S. C. (2006). Social relationships among adult female baboons (Papio cynocephalus) I. Variation in the strength of social bonds. Behavioral Ecology and Sociobiology, 61, 183–195. Smith, D. G. (1981). The association between rank and reproductive success of male rhesus monkeys. American Journal of Primatology, 1, 83–90. Smith, K., Alberts, S. C., & Altmann, J. (2003). Wild female baboons bias their social behaviour towards paternal half-sisters. Proceedings of the Royal Society of London B, 270, 503–510. Soltis, J., Thomsen, R., & Takenaka, O. (2001). The interaction of male and female reproductive strategies and paternity in wild Japanese macaques, Macaca fuscata. Animal Behaviour, 62, 485–494. Soltis, J. (2004). Mating systems. In B. Thierry, M. Singh, & W. Kaumanns, eds., Macaque Societies: a Model for the Study of Social Organization. Cambridge: Cambridge University Press, pp. 135–151. Soltis, J., Mitsunaga, F., Shimizu, K., et al. (1997). Sexual selection in Japanese macaques, II: female mate choice and male–male competition. Animal Behaviour, 54, 737–746. Soltis, J., Thomsen, R., Matsubayashi, K., & Takenaka, O. (2000). Infanticide by resident males and female counter-strategies in wild Japanese macaques (Macaca fuscata). Behavioral Ecology and Sociobiology, 48, 195–202. Srivastava, A. & Dunbar, R. I. M. (1996). The mating system of Hanuman langurs: a problem in optimal foraging. Behavioral Ecology and Sociobiology, 39, 219–226. Sterck, E. H.M., Watts, D. P., & van Schaik, C. P. (1997). The evolution of female social relationships in nonhuman primates. Behavioral Ecology and Sociobiology, 4, 291–309. Takahashi, H. (2004). Do males have a better chance of mating when the number of estrous females is equal to or greater than the males’ ordinal rank? Testing the hypothesis in Japanese macaques. American Journal of Primatology, 63, 95–102. Thrall, P. H., Antonovics, J., & Bever, J. D. (1997). Sexual transmission of disease and host mating systems: within-season reproductive success. American Naturalist, 149, 485–506. Thrall, P. H., Antonovics, J., & Dobson, A. P. (2000). Sexually transmitted diseases in polygynous mating systems: prevalence and impact on reproductive success. Proceedings of the Royal Society of London B, 267, 1555–1563. Treves, A. (2001). Reproductive consequences of variation in the composition of howler monkey (Alouatta spp.) groups. Behavioral Ecology and Sociobiology, 50, 61–71. van Noordwijk, M. A. & van Schaik, C. P. (2004). Sexual selection and the careers of primate males: paternity concentration, dominance acquisition tactics and
Causes and consequences of skew in male primates transfer decisions. In P. M. Kappeler & C. P. van Schaik, eds., Sexual Selection in Primates: a Comparative Perspective. Cambridge: Cambridge University Press, pp. 208–229. van Schaik, C. P. & Janson, C. H. (2001). Infanticide by Males and its Implications. Cambridge: Cambridge University Press. van Schaik, C. P., van Noordwijk, M. A., & Nunn, C. L. (1999). Sex and social evolution in primates. In P. Lee, ed., Comparative Socioecology of Primates. Cambridge: Cambridge University Press, pp. 204–240. van Schaik, C. P., Hodges, K., & Nunn, C. L. (2000). Paternity confusion and the ovarian cycles of female primates. In C. P. van Schaik & C. H. Janson, eds., Infanticide by Males and Its Implications. Cambridge: Cambridge University Press, pp. 361–387. Vehrencamp, S. L. (1983a). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Vehrencamp, S. L. (1983b). Optimal skew in cooperative societies. American Zoologist, 23, 327–355. Watts, D. P. (2000). Causes and consequences of variation in male mountain gorilla life: histories and group membership. In P. M. Kappeler, ed., Primate Males: Causes and Consequences of Variation in Group Composition. Cambridge: Cambridge University Press, pp. 169–180. Widdig, A., Nu¨rnberg, P., Krawczak, M., Streich, W. J., & Bercovitch, F. B. (2001). Paternal relatedness and age proximity regulate social relationships among adult female rhesus macaques. Proceedings of the National Academy of Sciences of the USA, 98, 13769–13773. Widdig, A., Bercovitch, F. B., Streich, W. J., Sauermann, U., Nuernberg, P., & Krawczak, M. (2004). A longitudinal analysis of reproductive skew in male macaques. Proceedings of the Royal Society of London B, 271, 819–826. Williams, D. A. (2004). Female control of reproductive skew in cooperatively breeding brown jays (Cyanocorax morio). Behavioral Ecology and Sociobiology, 55, 370–380. Wrangham, R. W. (1980). An ecological model of female-bonded primates. Behaviour, 75, 262–299. Wrangham, R. W. (1999). The evolution of coalitionary killing. Yearbook of Physical Anthropology, 42, 1–30.
195
8
Sociality and reproductive skew in horses and zebras d a n i e l i . r u b e n s t e i n a n d ca s sa nd r a m . n u n˜ e z
Summary The outcome of competition for resources or mates often leads to individual differences in reproductive success. In populations of equids, such as those of horses and zebras, skewed distributions of reproduction emerge because a limited number of individuals achieve disproportionate gains. For both sexes, skew results from differences in rank, age, and degree of social stability, although skew is generally greater for males than for females. Adult male horses and zebras typically establish “harem” groups by bonding with a number of mature females. Although the number of females that dominants bond with can be quite variable, potentially high levels of skew are rarely reached because subordinate males adopt alternative mating tactics that exact concessions from partners, whether they are dominant stallions or other subordinates. Successful breeding females also rely on support from subordinates to minimize feeding competition by keeping group size small, and this, too, reduces skew among females. The conflict of interest between the sexes arising over differences in optimal group size, along with the tendency for females to leave groups when sexually harassed, induces, but limits, the aggression that males direct towards females. Thus female concessions can shape both female and male levels of skew, but they also can be modulated by male behavior. Thus it appears that for equids the level of skew that emerges depends on key phenotypic traits and how their distribution among individuals constrains reproduction, as well as on how social relationships within and between the sexes affect the ability of a small number of individuals to monopolize resources involved in reproduction.
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
196
Sociality and reproductive skew in horses and zebras Inequality is a pervasive feature of virtually all societies, as is the attempt by the disadvantaged to attenuate it. Humans (Homo sapiens) have gone so far as to construct political systems to reduce, if not eliminate, differences in payoffs or outcomes. The Marxist ideal “From each according to his abilities, to each according to his needs” (Marx 1875) represents an egalitarian extreme. While animals may not codify behavior into philosophical norms, actions within many animal societies often redistribute payoffs away from the successful few, thus moderating heavily skewed resource allocations and reproductive opportunities. In mammals alone, skewed patterns of reproductive success are seen in male macaques (Macaca mulatta: Widdig et al. 2004), rhinoceroses (Diceros bicornis: Garnier et al. 2001), and even sex-role-reversed species such as hyenas (Crocuta crocuta: Engh et al. 2002). And in gorillas (Gorilla beringei: Robbins et al. 2006), meerkats (Suricata suricatta: Clutton-Brock et al. 2001), marmots (Marmota spp.: Allaine´ 2000), and mountain zebras (Equus zebra: Lloyd & Rasa 1989), skewed patterns of reproduction occur among females as well. Reductions in inequality often emerge when the less favored change the rules of the game by adopting strategies that deviate from the typical strategy (Rubenstein 1980). Alternatively, when the haves and the have-nots need each other to prosper, the status quo is maintained by the actions of those receiving the least. Either those having gained the most provide incentives to the less welloff, or the well-off are unable to completely impose their will when confronted by the actions of others (Rubenstein 1978, Vehrencamp 1983, Reeve et al. 1998, Johnstone 2000, Reeve & Shen 2006). In order to understand why particular strategies and relationships evolve and produce skewed payoffs and outcomes, knowing the underlying causes of inequality is critical. In many animal societies inequality results from differences in rank or fighting ability. Dominant females often monopolize resources needed to produce and raise offspring, while dominant males often monopolize reproductive access to females (Rubenstein 1994). For males, status and the bodily condition necessary for achieving and maintaining dominance is driven by sexual selection. But less well understood are the factors affecting status in non-sex-role-reversed females, especially in species such as equids, where competition for essential ecological resources is minimal (Rubenstein 1994). Dominance itself is influenced by many other factors such as age, bodily condition, reproductive state, and environmental harshness, thus pushing an understanding of the causes of reproductive skew one step back. Factors other than dominance and its determinants are also likely to play a role in shaping patterns of reproductive skew. These might include kinship as well as the nature and strength of relationships with individuals of different ages and sex. In this chapter we will explore these issues in order to understand the causes
197
198
D. I. Rubenstein, C. M. Nun˜ez and consequences of unequal and skewed reproduction in populations of feral horses (Equus caballus) and plains zebras (E. burchelli). We will illustrate the degree to which reproduction is skewed in each sex and we will identify causes. We will then examine strategies that develop to attenuate these differences among both females and males. Lastly, we will explore management and conservation consequences of these attenuation strategies.
Patterns of sociality and methods Equid societies come in two forms (Klingel 1977, Moehlman 1979, Rubenstein 1986). In one, typified by feral and wild horses (Equus caballus) as well as plains (E. burchelli) and mountain zebras (E. zebra), females and their immature offspring associate for long periods with one male. This is the socalled “harem,” and as offspring of both sexes mature they disperse. Thus the breeding members of these family groups are typically non-relatives, and kinship is unlikely to play a strong role in shaping social or reproductive decisions. In the other, typified by Grevy’s zebras (E. greryi), Asiatic (E. hemionus) and African wild asses (E. africanus), and the kiang (E. kiang), females form loose and transitory associations with many other females and males. This is a fission–fusion society and, typically, females of different reproductive states segregate and use different habitats. Since both classes of females contain reproductively active individuals, males forgo bonding with any class, instead establishing territories around water. In this way these territories contain the home ranges of lactating females with young foals as well as the routes that non-lactating females use to come to water (Rubenstein 1986, 1994, Sundaresan et al. 2007). Variants on these themes do emerge, and they provide insights into the ecological forces that shape social relationships as well as reproductive strategies and payoffs. For example, some horse populations have harems that defend territories (Rubenstein 1980, 1981) and some have “harems” with two stallions rather than one (Rubenstein 1982, 1986, Feh 1999, Linklater & Cameron 2000a, 2000b). Moreover, in plains zebra societies harems often coalesce into herds. Thus “harem species” can form multi-male as well as unimale groups, providing polygynous opportunities for males and polyandrous opportunities for females. Harem males can also defend territorial space and, at least in plains zebras, males can bring their harems together to form herds. A second tier of sociality, with fission–fusion elements, emerges as herds form and dissolve. The social and reproductive patterns described in this chapter come from long-term studies (1983 to the present) of free-ranging horses on Shackleford
Sociality and reproductive skew in horses and zebras
Figure 8.1 A highly successful Shackleford female and her young on the Shackleford Banks, North Carolina, USA. Photo courtesy of Cheryl Burke, Carteret County News Times.
Banks, North Carolina, USA (Figure 8.1), and plains zebras on the Ol Pejeta Conservancy and unfenced commercial ranches around the Mpala Research Center, Kenya, from 2000 to 2004 (Figure 8.2). Standard behavioral observations on time budgets (Rubenstein 1986, Rubenstein & Hack 2004), associations (Rubenstein et al. 2007, Sundaresan et al. 2007), and movements (Rubenstein 1986, 1994) were made on individuals identifiable by unique color patterns and freeze brands (horses) or stripes (plains zebras). Patterns of reproductive success (means, variances, and skew) are represented by standard statistical moments. Paternity analyses were performed by the US National Park Service using allozymes and DNA (E. G. Cothran, personal communication).
Background features and framework Equids illustrate some special social, ecological, and demographic features. First, equids, as large-bodied, hindgut fermenters with wide tooth rows, can subsist on low-quality forage. As a result, competition for such abundant food should be low, especially in mesic habitats. Instead of contesting for individual food items, females move to nearby items of equivalent
199
200
D. I. Rubenstein, C. M. Nun˜ez
Figure 8.2 Two male plains zebras fighting. This species is characterized by harem structures where one, but occasionally a second, male defends a group of females. Fights occur both between intruder and resident males and between two resident males. Photo by Dan Rubenstein.
value. Rubenstein (1994) shows that contests among female horses are infrequent, occurring at a rate of about one every 10 hours. Similarly, foraging success of plains zebras is not impeded as harem size increases (Rubenstein 1994). This suggests that vegetation eaten by horses and plains zebras facilitates, rather than hinders, group living. As a result, contests over foraging location or individual food items should not be a major factor exacerbating reproductive differences among females. Second, as Figure 8.3 illustrates, sex-specific survivorship schedules for Shackleford horses are somewhat atypical for polygynous, sexually selected mammals. In this population male survivorship is not always lower than that of females. Instead the curves cross shortly after maturity. As Berger (1986) notes, competition is intense among juvenile equid males as they attempt to establish dominance relationships early in life. Not surprisingly, female survival is much higher during this period, when females are building up bodily condition. Once a male’s place in society is established, however, the emergence of stable linear dominance hierarchies among harem males (Rubenstein 1994) lowers the costs of maintaining dominance, thus increasing male
Sociality and reproductive skew in horses and zebras Sex Differences in Survival
1000
Number of Horses
800
Male Female
600
400
200
0
0
2
4
6 Age
8
10
12
Figure 8.3 Age-specific survivorship of male and female horses on Shackleford Banks.
survival. For females the opposite is true: once foaling commences, female mortality increases and remains high and constant from year to year. When taken together, both features suggest that augmenting feeding success to maintain good body condition will matter to females, but strategies for increasing intake should not involve contest competition. As hindgut fermenters and throughput feeders, foraging success is likely to be mediated more by time available for feeding, than by excluding neighbors from particular resource patches. Group-living females could increase foraging time by coordinating foraging and vigilance activities among themselves, or they could rely on males to assume most of the vigilance costs while they devote more time to foraging. Equid females choose the latter by selecting males that limit driving and herding behavior and whose heightened vigilance reduces sexual approaches and harassment by other males (Rubenstein 1986). Since male reproductive success is determined by the number of females in the group, reproductive patterns of both males and females will be affected more by social than by ecological factors. Features such as age, rank, and social stability are likely to have important consequences on differences in reproduction and what individuals can do to ameliorate them. Since reproductive
201
202
D. I. Rubenstein, C. M. Nun˜ez success depends both on how many young are born over a lifetime and how many survive to the age of independence – the age when direct parental investment ceases – we will use total number born over an extended period of time and total number that survive until 1 year of age as measures of success. Our analyses will necessarily examine windows of time during which individuals of different ages will span different portions of these times. Consequently, reproductive rate as measured by number of foals born per unit time and interbirth interval will also be measured. Since these measures of reproductive success are likely to be influenced by age, residuals about best-fit regression lines of each of the key reproductive variables on age will be used to control for age effects. Patterns of reproductive inequality and skew Reproduction among females and males in both horses and zebras is highly variable, but not always skewed (Table 8.1). For female horses having left their natal group for at least 3 years, only a few have failed to give birth during their lifetimes. The distribution of total births is bell-shaped, with 50% of the population giving birth to between two and six offspring during this period. Yearly foaling rate is similarly bell-shaped, with the average female giving birth approximately once every other year (Table 8.1). For plains zebras, moments of variance and skew are greater than those of horses for both the total number of young born and yearly foaling rate (Table 8.1). Reproduction in males for both species is more variable and more skewed than for females (Table 8.1). Figure 8.4 shows the distribution of the total number of foals sired by male horses; similar distributions and moments are seen for male plains zebras (Table 8.1). In general, measures of male fertility are more variable and show more skew than measures of female fertility in both harem-dwelling equids (Table 8.1).
Factors affecting female reproductive success Age
Many features of reproduction in horses and plains zebras are affected by age. On Shackleford Banks the fraction of females giving birth in any given year increases with age (Figure 8.5). Whereas only 27% of females less than 6 years of age have given birth, 55% of those older than 10 have done so. Moreover, the total number of foals produced during the 17-year period increases the longer a female lives (Figure 8.6) as does the number of offspring surviving to independence (Figure 8.7). In addition, as females age, their yearly
Sociality and reproductive skew in horses and zebras Table 8.1 Summary statistics of distributions of various measures of reproductive successes for male and female horses and plains zebras. Means, variances, and skews are calculated from the first three moments of the frequency distributions. For horses, statistics before and after management are presented. Males Mean
Females
Variance
Skew
Mean
Variance
Skew
Horses (natural conditions) Total young born
4.92
24.83
1.96
3.76
6.14
0.59
No. surviving to
–
–
–
3.04
3.08
0.69
1.11
0.44
0.30
0.42
0.04
0.48
independence Yearly foaling rate
Horses (after management) Total young born
2.08
5.51
1.16
2.05
0.79
0.74
No. surviving to
–
–
–
1.46
0.67
2.03
0.52
0.26
0.70
0.20
0.04
0.54
independence Yearly foaling rate Plains zebras Total young born
2.34
2.07
1.00
0.84
0.66
0.54
No. surviving to
–
–
–
0.16
0.15
2.32
0.66
0.21
1.78
0.29
0.07
0.81
independence Yearly foaling rate
25 20 15 10 5
io
ny si s Ad am Pa c St hin eg o as W us in st on Sp oc Li k za rd M er li D n ig ge Ph r oe ni x Sl as h Za ne C ae sa r Ev an To ob a Ax l C as e Ed g H e om e Lu r c Si ifer sy ph Sp us hy nx
0
D
Number of offspring sired
30
Stallion
Figure 8.4 Distribution of the number of offspring sired by individual male horses over their lifetimes. Moments of the distribution found in Table 8.1.
203
D. I. Rubenstein, C. M. Nun˜ez 0.6
Fraction giving birth/yr
0.5 0.4 0.3 0.2 0.1 0 Young
Middle Age Female Age
Old (> 10 yrs)
Figure 8.5 Fraction of female horses giving birth per year as a function of age. v2 ¼ 24.7, p < 0.0001; df ¼ 2; n ¼ 610.
12
B
10
No. of foals born
204
B BB
8 6 4 2
BBBBBB BBB BBB B BB BB B BB B BBBBBB BBBBBBB BBBBBBB
0
B
B
BBBBBBB 0 2 4 6 8 10 12 14 16 18 Female Age
Figure 8.6 Total number of foals produced by female horses since leaving their natal harems. F [1, 155] ¼ 679.48; p < 0.0001; r ¼ 0.91.
birth rate also increases (Figure 8.8). For plains zebra females, age-dependent birth and survival patterns are similar (Table 8.2). Rank
Although a greater proportion of high-ranking females give birth each year than middle- or low-ranking females (Table 8.3), rank is age-dependent (v2 ¼ 18.9; df ¼ 2; p < 0.001). When residuals of total number of foals born, or
No. foals surviving to independence
Sociality and reproductive skew in horses and zebras 7
B
6
B
5
BBB
4
B
3
BB
2
BBBB BBB BB
BBB BB
B
B
BBBBBBBBBB
1
BBBBBBBBB
B
0 0
2
4
6
8 10 12 14 16 18 Female Age
Figure 8.7 Number of foals surviving to independence at 1 year of age as a function of female age. F [1,97] ¼ 175.96; p < 0.0001; r ¼ 0.83.
Yearly foaling rate (no. born/yr.)
1
B
0.9 0.8
B
0.7
B
B
0.6
B
B B
B
B BB B B B B B B B B B B B B B B B B B B B B B B B B B B B B BB B B
0.5 0.4 0.3 0.2 0.1 0 0
B
B
B
BBBBBBB 2 4 6 8 10 12 14 16 18 Female Age
Figure 8.8 Yearly foaling rate by female horses of different ages. F [1, 155] ¼ 44.12; p < 0.0001; r ¼ 0.47.
total number surviving to age of independence, are compared for high-, middle-, and low-ranking females, high-ranking females outperform females of lower rank (Table 8.3). Male rank also has a strong effect on both the total number of young produced and the total reared to the age of independence by females after accounting for age (Table 8.3). Females bonded to males of higher rank bear
205
206
D. I. Rubenstein, C. M. Nun˜ez Table 8.2 Age-dependent birth and survival patterns in plains zebras x-variable
y-variable
Female age
Birth rate (mean)
(SE)
Old
0.30
0.02
Middle
0.26
0.04
Young
0.10
0.12
Female age
Total foals born (mean)
(SE)
Old
0.92
0.05
Middle
0.68
0.08
Young
0.21
0.13
F
df
p value
r
3.32
2,298
< 0.05
0.21
F
df
p value
r
6.58
2,352
< 0.002
0.25
more young, and more of them survive to independence, than females associating with males of lower rank (Table 8.3). Social stability
Since males vary in the degree to which they provide females with material rewards (Rubenstein 1986), females faring less well should switch groups in search of better conditions. Since older females have had more opportunities to change groups than younger females, residuals from the best-fit line of number of moves regressed against age were compared with residuals of the best-fit lines regressing total number of young produced, or surviving to 1 year of age, against age. In both cases, after controlling for age, females’ reproductive success is inversely related to the number of different harems a female joins. Females moving frequently among males bear fewer young (F [1, 152] ¼ 11.99; p < 0.001) and have fewer of them survive to independence (F [1, 107] ¼ 13.72; p < 0.0005) than females bonded more tightly to males. How long-term residency leads to enhanced reproductive success is related to reduced harassment by males. In general, females associating with higherranking males, or males rising in rank quickly, are herded and harassed less than females associating with lower-ranking males (Figure 8.9; Rubenstein 1994). And females that are harassed more tend to be females that are most likely to switch harems (Figure 8.10). How females determine with which male to associate after leaving a harassing male could be based either on cues learned from witnessing male–male contests (Rubenstein & Hack 1992) or from direct feedback of male herding tendencies after joining a new harem. If indirect cues were the deciding factor, then females should generally move from high- to low-harassment males on their initial foray. If such information
Sociality and reproductive skew in horses and zebras Table 8.3 Factors affecting female reproductive success in horses x-variable
y-variable
x-variable
y-variable
Female rank
Proportion giving birth
High
0.61
Middle
0.38
Low
0.31
x2 10.68
df
p value < 0.005
2
Residuals of total foals Female rank
versus age
F
(mean)
(SE)
High
0.28
.25
Middle
0.20
.31
Low
0.61
.27
3.33
df 2.54
p value
r
< 0.05
0.32
Residuals of number Female rank
surviving to independence (mean)
(SE)
High
0.34
0.18
Middle
0.21
0.28
Low
0.67
0.27
Female rank
Residuals of foaling rate (mean)
(SE)
High
0.08
0.04
Middle
0.05
0.05
Low
0.02
0.04
Male rank
Residuals of total foals (mean)
(SE)
High
0.016
0.15
Low
0.61
0.20
F 4.38
F 1.03
F 5.87
df 2.37
df 2.54
df 1.45
p value < 0.02
p value < 0.5
r 0.44
r 0.19
p value
r
< 0.02
0.33
Residuals of number Male rank
surviving to independence (mean)
(SE)
High
0.01
0.24
Low
0.63
0.27
F 3.27
df 1.26
p value
r
< 0.08
0.32
is not perceivable or is unreliable, then the initial male chosen should not on average lead to reduced harassment. As Figure 8.11 shows, females often find themselves in a worse situation after their initial choice, and only after many moves do they find males that harass them less than their original partner.
207
D. I. Rubenstein, C. M. Nun˜ez
Herds/Hr/Female
0.3
0.2
0.1
0.0 High
Low Male Rank
Figure 8.9 Rates of herding received by female horses in harems led by high- and low-ranking males (t ¼ 2.16; p < 0.05; df ¼ 14).
3 Male Harassment Rate
208
2
1
0
Mover
Stayer Type of Female
Figure 8.10 Rates of herding received by harem female horses that remain in the harem or ultimately leave (t ¼ 2.23; p < 0.05; df ¼ 22).
Factors affecting male reproductive success Age
Both the number of total young sired by a male and his yearly siring rate increase with age. As males mature they accumulate more offspring over time (F [1,60] ¼ 173.9; p < 0.0001; r ¼ 0.86), yet they also tend to sire more offspring per year (F [1,60] ¼ 5.84; p < 0.02; r ¼ 0.28). Both measures of
12 11 10 9 8 7 6 5 4 3 2 1 -0.5
Individual Females
Individual Females
Sociality and reproductive skew in horses and zebras
0.0 0.5 1.0 1.5 Initial Change in Harassment Rate
2.0
12 11 10 9 8 7 6 5 4 3 2 1 –2
–1
0
1
Final Change in Harassment Rate
Figure 8.11 Paired comparisons of changes in harassment rates received by 12 female horses leaving their harems. Initial changes refer to differences in the rates they received from their original male and the male of the first group they joined (paired t ¼ 6.2; p < 0.005; df ¼ 11). Final changes refer to differences in the rates received from their original male and the male whose group they finally move to (paired t ¼ 4.24; p < 0.005; df ¼ 11).
reproductive success result from the fact that males accumulate females as they age. On becoming a harem leader, the number of females in the harem is smaller than in the year before they lose their harem (Table 8.4). Rank
Dyadic interactions among Shackleford males yield clear-cut dominance rankings (Rubenstein 1994). These manifest themselves in two ways. First, for harems headed by single males, rankings are highly linear and second, harems headed by single males are always dominant to those led by two males. When the population of males whose rank remained the same throughout the study period is divided into thirds, males in the top third of the hierarchy sire more total offspring and have higher yearly reproductive rates than males in the middle or lower third of the hierarchy (Table 8.4). As before, these differences are the result of differences in the number of mares in a male’s harem (Table 8.4). When rank is assessed by type of harem, males that are sole harem leaders sire more total offspring and have higher yearly siring rates than males in dualmale harems (Table 8.4). Most multi-male harems (80%), however, are led by younger males that become single harem leaders later in life. When the effect of age is controlled for by plotting the residuals of total number of offspring sired or yearly siring rate against age, single-male harem leaders fare significantly better with respect to siring rate and slightly better in terms of total foals sired than multi-male leaders (Table 8.4). Even more striking is the fact
209
210
D. I. Rubenstein, C. M. Nun˜ez Table 8.4 Factors affecting male reproductive success in horses x-variable
y-variable
Harem longevity
Number of females (mean)
First year
1.73
0.29
Penultimate year
3.42
0.28
Harem longevity
Number of foals sired (mean)
df
p value
r
4.17
78
< 0.0001
0.82
t
df
p value
r
4.99
78
< 0.0001
0.20
F
df
p value
r
3.60
2,28
< .05
.61
F
df
p value
r
8.02
2,107
< 0.0006
0.73
F
df
p value
r
6.98
2,107
< 0.002
0.58
F
df
p value
r
2.22
1,45
0.14
0.20
F
df
p value
r
4.00
1,47
< 0.05
0.51
t
df
p value
r
2.15
11
< 0.05
0.54
t
df
p value
r
1.51
4
< 0.20
060
(SE)
First year
1.9
0.16
Penultimate year
0.75
0.14
Harem rank
Total foals sired (mean)
(SE)
High
8.3
1.62
Middle
3.8
1.61
Low
2.5
1.54
Harem rank
Yearly siring rate (mean)
(SE)
High
1.58
0.15
Middle
1.29
0.14
Low
0.60
0.19
Harem rank
Number of females (mean)
t
(SE)
(SE)
High
3.88
0.33
Middle
2.78
0.31
Low
1.96
0.23
Harem type
Total young sired (mean)
(SE)
Single male
4.62
0.75
Multi-male
1.20
2.16
Harem type
Yearly siring rate (mean)
(SE)
Single male
1.29
0.09
Multi-male
0.86
0.22
Harem history:
Residuals of number of
males leading both
foals sired
harem types Single male
0.21
0.24
Multi-male
0.45
0.19
Herding intensity
Residuals of number of
(number / hour)
foals sired
Low (< 1/hour)
1.78
0.41
High (> 1/hour)
0.91
0.25
Sociality and reproductive skew in horses and zebras that when the siring rates of males experiencing both roles over a lifetime are compared after correcting for age, siring rate is significantly higher for males when they become sole harem leader than when they share leadership with a partner (Table 8.4). Social stability
Just as harassment rate affects female movements and fertility, it affects the stability of harems and male reproductive performance as well. Males that direct the movements of their females by overzealous herding experience higher female movement rates (F [1,3] ¼ 44.22; p < 0.007; r ¼ 0.96). And, although only a strong trend, when the population of single harem leaders is divided into high-intensity and low-intensity herders, the yearly siring rate of low-intensity males is almost twice as high as that of highintensity males (Table 8.4). Adopting alternative reproductive strategies or balancing selfish effort and the granting of concessions for access to critical resources are often thought of as different mechanisms that individuals use to reduce inequality and skew. For male equids this need not be the case. Male horses display a variety of routes to adulthood (Figure 8.12). The typical route involves leaving the natal group between 2 and 3 years of age and joining a bachelor group consisting of other non-breeding males. Bachelorhood is a period where males wander and socialize with other bachelors and harem males. Consequently dominance relationships begin to be developed with a significant portion of the male population. When a bachelor rises to the top of a group’s hierarchy he begins to challenge harem leaders. In the typical route to adulthood about half of these high-ranking bachelors take over a harem intact (harem route); the other half form a bond with a young female dispersing from her natal group (monogamous route). Two other less common alternative routes to adulthood are possible (Figure 8.12). In one, a young dispersing male forgoes joining a bachelor group, instead joining a harem group first as a peripheral male and finally as a secondary male (satellite route). This type of coalition illustrates how dominants sometimes offer concessions to retain the services of subordinates, as assumed by concession models of skew (Johnstone 2000). In the other, the dispersing male joins a bachelor group, but rather than spending time socializing and ascending the dominance queue, he forms a partnership with another recently arrived young male and together they leave the bachelor group and form a bond with a young dispersing female (polyandrous route). The probabilities associated with each of these routes to adulthood are shown in Figure 8.12, and Table 8.5 combines these likelihoods with the
211
212
D. I. Rubenstein, C. M. Nun˜ez Male Mating Tactics Harem Group
j
j
subadult .30
Bachelors Satellite
subadult .70 Harem
.10 .90 .50
Multi Multi
.10 .40
Polyandrous
Monogamous
Figure 8.12 Pathways to adulthood of Shackleford horse males. All juveniles leave their natal groups. Females either join an existing harem group or bond to a highranking bachelor male. Males have four options available. They can follow the typical harem route by joining a bachelor group, rising in rank and taking over an existing harem. Or they can leave the bachelor group as an experienced young adult and bond with an emigrating female (monogamous route). Alternatively, they can join a bachelor group for a short period, leaving with a similar-aged low-ranking subadult and bond with an emigrating female (polyandrous route). Some males, however, forgo joining a bachelor group altogether and take up residence as a secondary male in an existing harem (satellite route). Probabilities of the various pathways are shown adjacent to pathway arrows.
reproductive values presented above to generate estimated lifetime reproductive success of males pursuing these different ontogenetic pathways. Not too surprisingly the typical harem route is expected to lead to the highest lifetime output. But what is striking is that two of the other trajectories are not far behind. And given that the patterns presented above show that there are large differences in both male and female reproductive success associated with
Sociality and reproductive skew in horses and zebras Table 8.5. Male reproductive payoffs Male mating strategy Reproductive component
Satellite
Polyandrous
Monogamous
Harem 6
Potential breeding years
10
8
8
Number of females
3 – >5
1 – >5
1 – >5
3 – >5
Inter-birth interval (years)
3 – >2
3 – >2
3 – >2
2
Number of receptive females
1 – >2.5
0.33 – >2.5
0.33 – >2.5
1.5 – >2.5
Percentage paternity
10 – >100
50
100
100
Yearly reproductive success
0.1 – >2.5
0.2 – >2.5
0.7 – >2.5
1.2 – >2.5
Lifetime reproductive success
10.3
7.3
10.4
12.8
rank of harem males, males adopting the satellite and monogamous routes are likely to perform as well as subordinates adopting the harem route. The typical harem route derives its reproductive advantage because males take over an entire harem, usually containing a full complement of breeding females (4–6). As described above, females in single-male harems, especially if they are stable and the male is of high rank, have high yearly foaling rates and low inter-birth intervals. If takeover males are able to prevent other males from stealing these females and their harassment rate does not drive females away, then the expected reproductive success of takeover males will be great. But this route comes with an opportunity cost associated with time lost while rising in rank within bachelor society. It is precisely this cost that allows males adopting alternative routes to adulthood to flourish. Males adopting the monogamous route trade off time spent in acquiring dominance within bachelor groups with time spent increasing the size of their harems from one female to many. And while males adopting the satellite tactic reduce this opportunity cost to a minimum by associating with many reproductive females with almost no delay, their fertility is likely to be low initially since single harem stallions typically sire about 90% of a harem’s offspring. While pairs of males adopting the polyandrous route also reduce the time to adulthood, they must share paternity, which will range from 10% to 50%. Behavioral alternatives, inequality, and skew Males
While this series of trade-offs makes it possible for alternatives to coexist, their existence reduces inequality in male reproductive success. If males did not avail themselves of these alternatives the reproductive disparity
213
214
D. I. Rubenstein, C. M. Nun˜ez among dominant and subordinate single harem stallions is likely be even greater. Moreover, the behaviors that contribute to reducing reproductive inequality and stabilize these alternatives are the same behaviors that reduce reproductive skew. When males are challenged by other males for access to females, males must balance driving the challenger away and insuring that their females are not poached or do not stray. If only the typical route to adulthood existed, and males were only challenged by one stallion at a time, then balancing these competing demands would be straightforward: after a short period of assessment the subordinates would submit and herd their females away; if not, then dominants would escalate (Rubenstein & Hack 1992). Although herding females away from trouble solves a subordinate’s immediate problem, the likelihood of female departures in the long run is high and thus provides the impetus for males to adopt alternative ontogenetic pathways. Since two-thirds of the alternatives entail young males attempting to get a head start on breeding by forming alliances either with similar age-mates (polyandrous route) or mature stallions on the decline (satellite route), the evolutionary stability of the alternatives depends on the behavioral dynamics that sustain these coalitions. In both types of dual-stallion harems, the subordinate stallion is almost always first in approaching and engaging an intruder male. If his approach causes the intruder to retreat then little is risked and little energy is expended. Given his subordinate status, however, these initial contests usually escalate. Sniffing, squealing, and then biting, chasing, or rearing are commonplace (Rubenstein & Hack 1992). Since the intruder is far from his home area, the resident often has an ownership advantage and many (44%) of the contests end with the resident subordinate winning. But in slightly more than half of such contests (56%) the subordinate retreats, inducing the dominant male to join the fray. But this is where the two types of coalitions differ. In coalitions emerging via the satellite route, in virtually all contests (98%) the resident dominant wins. What is important to note, however, is that before the partner joined the harem, the solo stallion was losing encounters and was dropping in rank. By associating with a younger partner the situation becomes reversed. In the three instances of satellite coalitions forming, each original harem holder had been losing females. With the creation of these partnerships the harems began to grow again. In one instance the group grew from two to ten females. In coalitions begun via the polyandrous route, both males are usually in the bottom half of the breeding hierarchy. As a result, even with the dominant engaging the intruder first, resident dominants win less than half of encounters (38%) they assume. In no case, however, was a female ever abducted or seen to stray during a contest.
Sociality and reproductive skew in horses and zebras Clearly, coalitions of males protect against losing females while increasing the likelihood of winning, and perhaps even create long-term benefits by deterring future intrusions from neighboring males. For their services, subordinate males receive mating opportunities. In polyandrous-originating coalitions, subordinate males are rarely prevented (7%) from mating with estrous females. They are allowed to flehmen, inspect and lick female genital areas and, if the female doesn’t kick while a subordinate is mounting or walk out from under him once he is mounted, then thrusting and ejaculation occurs without interference by the dominant stallion. In 96% of these cases the dominant male subsequently mates with the female. If the dominant mates first, however, the subordinate male usually attempts to mate but succeeds only 88% of the time. Thus sperm mixing is common and paternity sharing is likely (Ginsberg & Rubenstein 1990). Although the situation is similar for subordinate males opting for satellite-route partnerships, the percentage of mating attempts by subordinates is lower. Given that group size can be up to five times larger for satellite partnerships than polyandrous ones, absolute paternity benefits for subordinates in these two coalition types are not likely to be that different. In over 17 years of studying the Shackleford horse population before population control began, only 12 coalitions formed and only three lasted until one partner died (Rubenstein 1982). As the above analysis shows, the expected lifetime reproductive success of these alternative strategies is predicted to be lower than that of the typical route to adulthood. Therefore it is not surprising that most coalitions disband after a few years. But while they last, what mechanism maintains them? Females generally do not stray far from a group even when in estrus. And given that the initiation and culmination of a successful mating takes between 2 and 5 minutes, it is unlikely that matings by a coalition partner will go unnoticed. That subordinates mate so often and without interference suggests that dominants are allowing them to do so. While these behavioral events tend to support the idea that concession giving by dominants is a mechanism for maintaining the stability of groups, dominants continuously reinforce their dominance by initiating fights before, during, and after the breeding season. Maintaining rank differential within coalitions appears to be important, but whether it plays a role in conditioning inferiority in subordinates so that they initiate fewer matings, thus increasing inequality within the coalition, or whether it is part of the overall maturation process that leads to breaking up the partnership, is hard to determine. In one of the three cases of satellite-originating partnerships, the subordinate spent most of his time associating with all new females joining the harem as it grew. On a regular basis the dominant would disrupt these associations by chasing
215
216
D. I. Rubenstein, C. M. Nun˜ez the subordinate away. Eventually the subordinate left the coalition to take over a neighboring harem, and within a few days those females from the old harem with whom the subordinate had formed strong associations moved to his new harem. Thus it appears that dominants continuously reinforce their dominance over subordinates for good reason. Frequent low levels of aggression by dominants appear to limit freedom of movement by subordinates. But making subordinates “think twice” before attempting to mate also reveals that dominants are unable to enforce exclusive control of mating. Consequently, granting concessions and enforcing only limited control over subordinates – a form of selfish effort – appear to be operating jointly as Johnstone (2000) and Reeve et al. (1998) suggest and as Reeve & Shen (2006) explicitly capture in the “bordered tug-of-war” model. Females
As we have seen above, reproductive inequality and skew also exists among females, even if the magnitude of these differences is smaller than among males. Since female rank affects reproductive success, dominant females have some control over the reproduction of subordinates. Yet subordinates stay. Why do they do so? In part they stay because they are tolerated by the dominants. Subordinates help limit the number of immigrating females joining harem groups (personal observation). Even in species where scramble competition is the norm, as it is in equids, competition will increase if groups get too large. Therefore, allowing subordinates to stay as an aid in preventing even more females from joining represents a cost to dominants. As predicted by concession models of skew (Johnstone 2000), granting residency appears to be a price that dominants pay to derive even greater gains. But do dominants actually have the ability to determine the optimal size of their harems? The answer is mostly likely no – and understanding why provides a more compelling reason why subordinates stay. Stallions typically stand near immigrating females, protecting them from approaches by other females and thus limiting the ability of resident females to force immigrants from harems. Because male rank and the social stability it provides affects female reproduction more than does female rank (Rubenstein 1986), when male and female reproductive interests come into conflict, females are caught in a bind. Consequently, they rarely confront males, and male interests tend to prevail. The limited ability of females to control their own reproductive success is underscored by two unfortunate “experiments” that occurred on Shackleford Banks when two harem stallions inadvertently died during immobilization by another team of researchers. Since this immobilization was being carefully monitored, we were able to get detailed, fine-grained behavioral data on the
Sociality and reproductive skew in horses and zebras social behavior of the unguarded females. Almost immediately these groups were besieged by harem leaders and high-ranking bachelors. For up to four days these females fended off advances, but the intrusion rate was over 10 times higher than when their harem males were present. The percentage of time females foraged decreased from 72% of every hour to less than 40%, not only because of direct approaches by males, but because all female groups retreated to less frequented areas where forage was less abundant and of lower quality. After four days each group disintegrated. This inadvertent perturbation to the normal social structure highlights the central role that males play in the lives of females. Their increased vigilance reduces their grazing time to 80% of that of their females, but it also buffers their females from extensive intrusions. As a result, females are provided with increased time for grazing and a means of enhancing their reproductive success (Rubenstein 1986). Our observations also suggest that frequent reciprocal and mutualistic actions that would be required of females to hold groups together in the absence of males are costly. Thus female–female transactions in which dominants would limit their aggression towards subordinates could in theory encourage group cohesion as predicted by concession models of skew (Johnstone 2000). But even if females attempted to grant such concessions to foster residency, their actions would be superseded by female–male transactions. But what form are these transactions likely to take? Males have much to gain by keeping as many females as possible in harems. Subordinate males do this by adopting a “best of a bad job” strategy of herding females away from competitors rather than driving competitor males away from females (Rubenstein 1994). In the long run such behavior reduces subordinate harem size and ultimately reproductive success. Dominants, however, provide more stable and less stressful environments for females, where their reproductive success is enhanced. Dominants also attract more females than subordinates. For males, both factors increase reproductive success. Skew dynamics
Understanding how males and females balance the tension between cooperation and competition provides insights into the processes determining levels of reproductive inequality and skew. For group-living species, reproductive skew emerges when three conditions pertain: dominants must be able to limit the reproduction of subordinates; subordinates must do better by staying than by leaving; and dominants must do worse if subordinates leave rather than stay. For horses all three conditions are met, but because of novel social dynamics. First, although competition among females is limited and involves “scramble” rather than the “contest” behavior, dominant females
217
218
D. I. Rubenstein, C. M. Nun˜ez control the reproduction of subordinates to a limited degree (Table 8.3). Second, subordinate females do help dominants by preventing additional females from entering the group and intensifying competition. And third, subordinates faring poorly sometimes do leave harems in search of better opportunities, but the process of integrating into new ones is difficult and not without costs (Figures 8.9–8.11). Thus for some subordinate females that are balancing the marginal gains of staying with leaving, concessions could tip the balance enticing them to stay and join in mutualistically beneficial behavior. But dominants do not appear to offer such concessions because males intervene to keep these subordinates – as well as many future subordinates – in their groups. Since the size of harem groups that benefit polygynous males will usually be larger than those that benefit females (Armitage 1981, 1986), reproductive tension is created among males and females. While agonism by dominant females does induce some subordinates to leave when groups get large, aggressive herding by males usually overrides the actions of their dominant females. But if males are too aggressive, subordinate females receiving herding actions by stallions will leave anyway. Thus, to maintain large harems, dominant males must curtail excessive aggressive behavior. Ultimately, the ability of males to balance agonistic and affiliative tendencies towards females will determine how many, and which, females stay. Just as dominant males in multi-male harems maintain their coalitions by conceding reproduction while simultaneously engaging in occasional bouts of aggression to condition subordinates to limit reproductive competition, harem males will strive to maintain harems larger than preferred by their females by using aggression against females in limited and discretionary ways. Rarely in animal societies are reproductive interests among group-living individuals of one sex mediated by interactions and transactions with members of the other sex. But it may be widespread in the equids. In plains zebras, herds form as harems coalesce (Rubenstein & Hack 2004). Harems come together because cuckolding pressure imposed by bachelors is high, higher than levels reached in horses. Presumably high levels of predation favor bachelor males forming long-term associations with a large number of males facing similar risks. When confronted with coordinated actions of many bachelors, harem stallions reduce their individual cuckolding risk by banding together and keeping bachelors on the edge of the herd (Rubenstein & Hack 2004). Clearly males benefit by forming herds, but they can only do so because female foraging success is similar whether harems are in or out of herds. To insure that females remain indifferent, males in herds incur reductions in time spent foraging above those incurred when their harems are alone on a
Sociality and reproductive skew in horses and zebras landscape (Rubenstein & Hack 2004). This represents a significant concession to maintaining herd stability.
Impact of management on reproductive inequality and skew In 2000 the US National Park Service began managing the Shackleford horse population with PZP immunocontraception of females. Along with yearly removal of females, this kept horse numbers well below carrying capacity. One immediate consequence of these interventions was an increased birth rate and a release from density-dependence. What little competition for resources existed in the more natural state, the current state offers even less. Other consequences were seen with respect to changes in behavior. Young females, especially those coming into estrus for the first or second time, showed reductions in time spent feeding and increases in time spent standing (Rogers 2001, Constantino 2002). Males also showed greater tendencies to herd females, especially those having received contraception (Stroeh 2001). After six years of intervention, patterns of reproductive inequality and skew of males and females have changed. As shown in Table 8.1, the variance in the distribution of total births, yearly foaling and siring rate, and the number of young surviving to the age of independence have all decreased. In many respects, the underlying factors affecting these measures of reproductive success remain unchanged after management. Age still has a major impact on the number of foals born during the 6 years (F [1,42] ¼ 8.15; p < 0.007; r ¼ 0.40), on the foaling rate (F [1,77] ¼ 56.79; p < 0.0001; r ¼ 0.64), and on the number of harem changes made by females (F [1,70] ¼ 8.95; p < 0.005; r ¼ 0.33). After computing appropriate residuals to account for the influence of age, female rank still affects the total number of foals born throughout the period as well as the annual foaling rate (Table 8.6). And while females associating with males of higher rank perform better reproductively than females associating with lower-ranking males, the effect is of reduced significance (Table 8.6). Since females receiving contraception continue to cycle throughout the entire reproductive season, genital inspection and harassment remains high for all males (Stroeh 2001). Consequently, differences in the way males of different ranks interact with females have been diminished. Reducing both the variance and skew in the number of foals born or sired during this six-year period should increase effective population size, since more females will contribute similar numbers of genes to future generations and a greater proportion of males will sire offspring (Falconer 1960). Given that the US Park Service is holding population numbers between 120 and 130, supporting
219
220
D. I. Rubenstein, C. M. Nun˜ez Table 8.6 Factors affecting reproductive success in horses after management commenced x -variable
y-variable
Female rank
Total foals sired (mean)
F
df
p value
r
37.17
2,31
< 0.003
.57
3.72
2,48
< 0.03
0.36
1.59
1,70
< 0.25
0.24
2.92
2,19
<0.07
0.28
(SE)
High
0.54
0.19
Middle
0.24
0.36
Low
0.43
0.18
Female rank
Yearly foaling rate (mean)
(SE)
High
0.075
0.04
Middle
0.003
0.07
Low
0.60
0.03
Female rank
Residual numbers of mares (mean)
(SE)
High
0.25
0.26
Middle
-0.27
0.47
Low
0.32
0.23
Male rank
Foaling rate (mean)
(SE)
High
0.27
0.05
Middle
0.20
0.08
Low
0.01
0.07
inherent features of a species’ social biology that reduce the loss of genetic diversity is important.
Discussion In our populations of horses and plains zebras reproductive inequalities and skew occur for males and females. At least for male horses, variation and skew would have been greater had it not been for the opportunity for lowperforming males to join groups of different types. Such variants are maintained because of the many ways in which adjustments in concessional and selfish effort can be made. In all multi-male harems, complete control by dominants is absent either because aging males are dropping in rank (satellite route) or because young partners begin their reproductive lives as inexperienced equals (polyandrous route). Both situations meet the assumptions of the
Sociality and reproductive skew in horses and zebras bordered tug-of-war model, and the relatively peaceful coexistence that emerges fits one of the model’s predictions (Reeve & Shen 2006). If there are no strong ecological or social reasons for maintaining groups, the model predicts that the degree of strife within groups will be low. Given that joining bachelor groups is a ready alternative that only incurs a moderate cost of delaying reproduction for a few years, it is not surprising that fights among coalition partners are infrequent and of low intensity. Coalitions among males are not limited to our horse and plains zebra populations. While coalitions are relatively rare in our study (< 10% of harem groups in any given year), they are common in other horse populations. On a nearby North Carolina island 50% of the harem groups were led by two males (Stevens 1990); multi-male groups comprise 12% of harem groups in the Great Basin (Berger 1986); 33% of groups in the Kaimanawa Range are led by multiple males; and 38% of the reproductive units in the Camargue horses contain two or more males (Feh 1999). Recently there has been much debate as to whether these multi-male units are coalitions formed by mutualistic or reciprocal relationships (Feh 1999, Linklater & Cameron 2000a). Linklater (2000) and Linklater & Cameron (2000b) claim that multi-male associations are not true coalitions in part because the authors assume that the reproductive success of males and females inhabiting multi-male groups must equal that of males and females in uni-male harems. Otherwise, they could not be maintained by selection. Game theoretic models that account for differences in phenotype (Parker 1984) offer many alternative predictions concerning reproductive payoffs that would permit multi-male social units to be maintained. And if males and females residing in them do so only temporarily as part of a complex life cycle, then the existence of such social variants is easier to sustain; as payoffs produced during one period are incorporated into measures of overall lifetime reproductive success their contribution is diluted. Feh’s field study (1999) and critique (2001) provide ample evidence that multi-male groups are functional coalitions. As a result, understanding the behavioral dynamics that hold them together can provide insights into how evolutionary forces generate skewed reproductive patterns. In accord with concession models of skew (Johnstone 2000), we have shown the degree to which both dominants and subordinates act selfishly and provide concessions that maintain these groups until balance among the payoffs changes. And change they will as phenotypes and environmental conditions change. Feh (1999) describes trade-offs associated with the different roles that are similar to those we find: subordinate coalition members take more risks and dominants rarely interfere with matings of subordinates, although they are aware that they are occurring. Feh also goes one step further and calculates the degree to
221
222
D. I. Rubenstein, C. M. Nun˜ez which dominants allow subordinates to mate. In the Camargue population, subordinates sire about 25% of multi-male groups’ offspring, a value that is consistent with our expectations. What makes our study unique is that we also explore the more complicated dynamics underlying unequal and skewed reproduction of females. While not as great as in males, female reproductive skew is also significant. Despite abundant vegetation and physiologies that make subsistence on low-quality food economical, agonism persists among females and dominance hierarchies emerge. That dominance affects reproduction demonstrates harem females invest in selfish behavior. Mutual support among females in keeping additional females from joining harems benefits females of all ranks, and shows that dominants are prepared to provide concessions to retain the services of subordinates. But group cohesion can only be maintained if subordinates also offer concessions to dominants at critical times. Such concessions are seen occasionally at watering points where access is extremely limited (Rubenstein 1994). Yet even though these conditions meet the assumptions of the bordered tug-of-war model, they are not the conditions that matter most in shaping skewed reproduction of females. Male preferences for even larger harems generate pressures to override the preferences of females. But in meeting their selfish needs, males could often drive away precisely those females they want to retain. Thus males are left balancing agonistic and affiliative tendencies with those of their females. As the model suggests, the optimal mix is likely to change with changes in group size as well as demography. Thus we expect that when there are many harems available rates of herding and harassment should decrease. The bordered tug-of-war model predicts that when ecological conditions are harsh, group breakup will be constrained and destructive conflict within groups should increase. What makes equids special is that they are among the few grazers that experience little competition. Marmots (Marmota spp.) are also grazers, but they are less free to wander widely in search of forage since they are tied to burrows for safety from predators (Armitage 1981, 1986). Thus they should experience greater competition than horses and plains zebras and they should exhibit more strife among dominants and subordinates. This appears to be true for most of the socially monogamous species. The need to maintain warmth during hibernation apparently selects for small numbers of young and thus complete reproductive suppression of subordinates by dominants (Allaine´ 2000). Reproductive skew is less pronounced in the yellow-bellied marmots (Marmota flaviventris), even though all females compete for limited resources. Apparently females cooperate in keeping other groups of females from usurping parts of the home range (Allaine´ 2000), much as female equids cooperate to keep additional immigrant females from joining harems. In both species emigration
Sociality and reproductive skew in horses and zebras by females is possible, thus reducing skew, but in marmots, males appear not to be involved in adjusting concessional and selfish investments. In our study we showed that the dynamics shaping sociality in horses operate in plains zebras as well. Dominance differences set the stage for inequitable and skewed reproduction in males and females of both haremliving species. The importance of male rank in balancing concessional and selfish effort is a thread in both species’ social cloth. Yet in plains zebras this thread appears stronger where ecological conditions make it more difficult for males to maintain control of females and their reproduction. Incomplete control of reproduction necessitates transactions between dominants and subordinates in order to stabilize social orders. When these transactions involve individuals within and between the sexes it is not surprising that a variety of mating and social structures can coexist. This is especially true for horses, and should be searched for in other species where conflicts of interest among many phenotypes are common.
Conclusions Inequality in reproductive success in horses and zebras exists in both sexes, although both variance and skew is greater for males than for females. Differences in age and status shape the distributions of reproduction for both males and females. Social stability also determines degree of skew. After controlling for age, females that change groups often breed less often than those that remain attached to particular males for long periods of time. And for males, greater reproductive success accrues to those that curtail harassment of females. Perhaps what is most striking about equids that form long-term breeding associations among males and females, as is the case in horses and zebras, is that the magnitude of skew is lower than could be expected. Such attenuation emerges because of two factors. One is that males can adopt a number of alternative reproductive tactics. Each alternative relies on concessions to subordinates that boost reproduction above what would be achieved at that age by adopting the typical tactic. Moreover, each alternative tactic enhances the reproduction of the concession giver over what he would have gained without joining the partnership. Similarly for females, concessions enable non-harem females to gain access to a dominant male while limiting the loss of reproduction that females already bonded to such an “attractive” male would incur if group size were to grow even larger. The other factor is that intersexual conflict reduces reproductive differences within the sexes because neither males nor females can completely enforce their self-interest. While males would prefer larger harems, which would enhance male skew,
223
224
D. I. Rubenstein, C. M. Nun˜ez the aggression necessary to effectively maintain large groups would most likely drive females away. And for females the ability to keep group size small by peripheralizing additional females, thus reducing female skew by allowing wandering females to be spread more equitably among males, is modulated by male herding. Thus, while phenotypic differences such as age and size and their distribution within populations play important roles in shaping patterns of skew, a variety of intra- and intersexual relationships that characterize harem-living equids also influence sex-specific patterns of skew. The extent to which the interplay between these forces shapes skew in other harem-living species, or even species exhibiting other mating systems, needs to be explored. Acknowledgments We thank the US National Park Service, people and the Government of Kenya (Research permit MOST 13/001/29C 80 vol. II) for allowing us to carry out this study, and Cape Lookout National Seashore (North Carolina), the Mpala Research Centre (Kenya) and the Ol Pejeta Conservancy (Kenya) for allowing us to work on their lands. Support for these studies came from the National Science Foundation awards CNS-0205214, IOB-9874523, and BSR-8352137. Thanks also go to Mace Hack, Justine Cordingley, Jessica Rogers, Oliver Stroeh, and Virginia Costantino for their assistance in gathering field data.
References Allaine´, D. (2000). Sociality, mating system and reproductive skew in marmots: evidence and hypotheses. Behavioural Processes, 51, 21–34. Armitage, K. B. (1981). Sociality as a life-history tactic of ground squirrels. Oecologia, 48, 36–49. Armitage, K. B. (1986). Marmot polygyny revisited: determinants of male and female reproductive strategies. In D. I. Rubenstein & R. W. Wrangham, eds., Ecological Aspects of Social Evolution: Birds and Mammals. Princeton, NJ: Princeton University Press, pp. 303–331. Berger, J. (1986). Wild Horses of the Great Basin. Chicago, IL: Chicago University Press. Clutton-Brock, T. H., Brotherton, P. N. M., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Costantino, V. S. (2002). Immunocontraception on feral mares: intended and unintended effects on hormone levels and behavior. Unpublished B. A. senior thesis, Princeton University. Engh, A. L., Funk, S. M., Horn, R. C. V., et al. (2002). Reproductive skew among males in a female-dominated mammalian society. Behavioral Ecology, 13, 193–200.
Sociality and reproductive skew in horses and zebras Falconer, D. S. (1960). Introduction to Quantitative Genetics. New York, NY: Ronald Press. Feh, C. (1999). Alliances and reproductive success in Camargue stallions. Animal Behavior, 57, 705–713. Feh, C. (2001). Alliances between stallions are more than just multimale groups: reply to Linklater and Cameron (2000). Animal Behaviour, 61, F27–F30. Garnier, J. N., Bruford, M. W., & Goossens, B. (2001). Mating systems and reproductive skew in the black rhinoceros. Molecular Ecology, 10, 2031–2041. Ginsberg, J. R. & Rubenstein, D. I. (1990). Sperm competition and zebra mating behaviour. Behavioral Ecology and Sociobiology, 26, 427–434. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Klingel, H. (1977). Observations on social organization and behaviour of African and Asiatic wild asses (Equus africanus and Equus hemionus). Zeitschrift fur Tierpsychologie (Ethology), 44, 323–331. Linklater, W. L. (2000). Adaptive explanation in socio-ecology: lessons from the Equidae. Biological Reviews of the Cambridge Philosophical Society, 75, 1–20. Linklater, W. L. & Cameron, E. Z. (2000a). Distinguishing cooperation from cohabitation: the feral horse case study. Animal Behaviour, 59, F17–F21. Linklater, W. L. & Cameron, E. Z. (2000b). Tests for cooperative behaviour between stallions. Animal Behaviour, 60, 731–743. Lloyd, P. H. & Rasa, O. A. E. (1989). Status, reproductive success and fitness in Cape mountain zebra (Equus zebra zebra). Behavioral Ecology and Sociobiology, 25, 411–420. Marx, K. (1875). Marx/Engels Selected Works, Vol. 3. Moscow: Progress Publishers, 1970. Moehlman, P. D. (1979). Behavior and ecology of feral asses (Equus asinus). National Geographic Society Research Reports, 405, 405–411. Parker, G. A. (1984). Evolutionary stable strategies. In J. R. Krebs & N. B. Davies, eds., Behavioural Ecology: an Evolutionary Approach, 2nd edn. Oxford: Blackwell, pp. 30–61. Reeve, H. K. & Shen, S. -F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders. Behavioral Ecology, 9, 267–278. Robbins, A. M., Robbins, M. M., Gerald-Steklis, N. & Steklis, H. D. (2006). Age-related patterns of reproductive success among female mountain gorillas. American Journal of Physical Anthropology, 131, 511–521. Rogers, J. E. (2001). The effects of an immunocontraceptive on the social behavior of female feral horses (Equus caballus). Unpublished B.A. senior thesis, Princeton University. Rubenstein, D. I. (1978). On predation, competition, and the advantages of group living. Perspectives in Ethology, 3, 205–232. Rubenstein, D. I. (1980). On the evolution of alternative mating strategies. In J. E. R. Stadden, ed., Limits to Action: the Allocation of Individual Behaviour. New York, NY: Academic Press, pp. 65–100.
225
226
D. I. Rubenstein, C. M. Nun˜ez Rubenstein, D. I. (1981). Behavioural ecology of island feral horses. Equine Veterinarian Journal, 13, 27–34. Rubenstein, D. I. (1982). Reproductive value and behavioural strategies: coming of age in monkeys and horses. Perspectives in Ethology, 5, 469–487. Rubenstein, D. I. (1986). Ecology and sociality in horses and zebras. In D. I. Rubenstein & R. W. Wrangham, eds., Ecological Aspects of Social Evolution: Birds and Mammals. Princeton, NJ: Princeton University Press, pp. 282–302. Rubenstein, D. I. (1994). The ecology of female social behavior in horses, zebras and asses. In P. Jarman and A. Rossiter, eds., Animal Societies: Individuals, Interactions and Organisations. Kyoto: Kyoto University Press, pp. 13–28. Rubenstein, D. I. & Hack, M. (1992). Horse signals: the sounds and scents of fury. Evolutionary Ecology, 6, 254–260. Rubenstein, D. I. & Hack, M. (2004). Natural and sexual selection and the evolution of multi-level societies: insights from zebras with comparisons to primates. In P. Kappeler and C. P. van Schaik, eds., Sexual Selection in Primates: New and Comparative Perspectives. Cambridge: Cambridge University Press. Rubenstein, D. I., Sundaresan, S., Fischhoff, I., & Saltz, D. (2007). Social networks in wild asses: comparing patterns and processes among populations. In A. Stubbe, P. Kaczensky, K. Wesche, R. Samjaa, & M. Stubbe, eds., Exploration into the Biological Resources of Mongolia, Vol. 10. Halle: Martin-Luther-University HalleWittenberg, pp. 159–176. Stevens, E. (1990). Instability of harems of feral horses in relation to season and presence of subordinate stallions. Behaviour, 112, 149–161. Stroeh, O. M. (2001). The effects of management strategies on the behavioral ecology of the Shackleford Banks male horses. Unpublished B.A. thesis, Princeton University. Sundaresan, S. R., Fischhoff, I. R., & Rubenstein, D. I. (2007). Male harassment influences female movements and associations in Grevy’s zebra (Equus grevyi). Behavioral Ecology, 18, 860–865. Vehrencamp, S. L. (1983). Optimal degree of skew in cooperative societies. American Zoologist, 223, 327–335. Widdig, A., Bercovitch, F. B., Streich, W. J., et al. (2004). A longitudinal analysis of reproductive skew in male rhesus macaques. Proceedings of the Royal Society of London, B, 271, 819–826.
9
Reproductive skew in avian societies w a l t e r d . k o e n i g , sh e n g - f e n g s h e n , al a n h . kr a k a u e r , and joseph haydock
Summary Reproductive skew theory provides a predictive theory of the extent of reproductive sharing that is expected to occur in societies consisting, at least potentially, of multiple co-breeders. Here, we discuss some of the challenges that skew theory faces as it attempts to form the basis of a unified theory of social evolution in birds. These include the problem of distinguishing potential versus actual reproductive roles, encompassing extra-group parentage and sexual conflict, predicting the distribution of group size, and determining the appropriate null model against which to test empirical results. Despite these and other problems with skew theory as currently developed, a compilation of prior studies indicates some degree of consistency with the predictions of concession or optimal skew theory. More surprisingly, a meta-analysis indicates that interspecific patterns of sociality offer reasonably good matches to the predictions of the concession model of reproductive skew. Strong support for skew theory remains lacking, and experimental tests sufficient to reject alternative skew models have yet to be performed in birds. Nonetheless, these results offer encouragement that additional theoretical work in this field may eventually yield a useful framework for understanding the remarkable diversity of avian sociality.
Introduction An irony of kin-selection and inclusive-fitness theory (Hamilton 1964) is that these concepts, conceived primarily as a solution to the paradox of Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
227
228
W. D. Koenig et al. altruism, should turn out to be so useful in providing a framework for interpreting conflict among close relatives. Indeed, as more social species have been studied in detail, it has become clear that cooperation and conflict go hand in hand in most, if not all, societies. Nowhere is the uneasy coexistence of cooperation and conflict more evident than in cooperative breeders, in which more than a pair of individuals assist in the production of offspring. Cooperatively breeding birds have proved a particularly fertile taxon for studying social behavior, both because they show a considerable range in mating systems and because they are generally diurnal, observable, and easily manipulated. Equally important, the close associations of individuals from multiple generations found in such societies result in opportunities for conflict and competition rarely found in non-group-living taxa. Simultaneously, the fact that associated individuals are often close genetic relatives opens the door for the evolution of altruistic and cooperative behaviors that are less likely to evolve among non-relatives that may only interact one or a few times during their lives. The standoff that appears to exist between the highly developed cooperation and equally sophisticated competition observed in cooperative breeders is to a considerable extent what makes such species so exciting and interesting to study. Most commonly, cooperatively breeding groups consist of nuclear families in which only a pair of individuals produce offspring that then delay dispersal, remaining on their natal territories as “helpers-at-the-nest” for a variable length of time, often several years, rather than disperse and attempt to breed on their own. Such helpers are typically non-breeders, presumably because of incest avoidance, despite being reproductively mature and fully capable of breeding. A classic example is the Florida scrub-jay (Aphelocoma coerulescens), which is both socially and genetically monogamous but lives in territorial groups containing up to several non-breeding helpers (Woolfenden & Fitzpatrick 1984, Quinn et al. 1999). Helpers in this situation can be considered subordinates in a high-skew society, although, as discussed below, whether this approach yields new insight into the situation or not is debatable. Beyond such cases, there is now an increasing number (although still a minority) of cases in which three or more individuals share reproduction, at least potentially, to a greater or lesser extent. These systems are quite diverse. A few species, most notably several species of manakins in the family Pipridae and the wild turkey (Meleagris gallapavo), exhibit what is known as “cooperative courtship” in which males form coalitions for the purpose of attracting and mating with females, but do not otherwise contribute parental care (Foster 1977, MacDonald 1989, Krakauer 2005a). More commonly birds form bisexual groups within which the mating system is quite variable, including
Reproductive skew in avian societies “cooperative polygyny” (if groups consist of one male and multiple females), “cooperative polyandry” (one female and multiple males), or “cooperative polygynandry” (multiple breeders of both sexes). Two examples of the latter are the acorn woodpecker (Melanerpes formicivorus: Koenig & Mumme 1987) and the dunnock (Prunella modularis: Davies 1992), both of which exhibit cooperative polygynandry in which more than one male may compete for matings with multiple females. The focus of this book, reproductive skew, provides a predictive and potentially quantitative theory as to the degree of reproductive sharing that should take place in such societies. Skew theory currently involves at least two distinct approaches. The first approach encompasses what are known as “transactional” skew models, in which participants are hypothesized to engage in a kind of social contract by which they agree to cooperate in the group in exchange for the assurance of obtaining sufficient fitness benefits to outweigh the alternative of leaving the group and breeding solitarily. Transactional models generally assume complete control of reproduction by one individual and a high degree of mutualism among the participants, focus on different assumptions (such as the possibility of reproduction being increasingly costly as more young are produced: Cant & Johnstone 1999), and explicitly consider group stability as well as the degree of reproductive sharing. Thus far, most empirical work has focused on testing transactional models, and thus we will primarily focus on these models here as well. A second framework, known as “compromise” models, ignores group stability, instead postulating that skew is determined solely by the relative competitiveness of the participating individuals (Reeve et al. 1998, Johnstone 2000). Thus, the most common form of compromise models, “tug-of-war models,” do not assume any social contract among the individuals involved. On the other hand, by eliminating the often unrealistic assumption of complete control, tug-of-war models accommodate the likelihood of overt competition among co-breeders, a frequently observed phenomenon that is not expected under transactional models (K. Reeve, personal communication). Attempts to combine the cogent aspects of these two approaches have been made by both Reeve (2000) and Johnstone (2000), and more recently by Reeve & Shen (2006) by means of a “bordered” tug-of-war model.
Goals As discussed in detail elsewhere in this volume, skew theory provides a framework integrating ecologic, genetic, and social factors to predict the patterns of reproductive partitioning in animal societies. The goal of skew theory
229
230
W. D. Koenig et al. is ambitious: to provide a unified theory of social evolution ultimately leading to “a comprehensive theory that relates all of the major features of animal societies to their ecological contexts, genetic structures, and internal power asymmetries” (Reeve & Keller 2001, p. 348). Although no one is likely to be surprised if this goal is not easily attained, it is clearly desirable to know whether skew theory as we currently understand it is heading in the right direction and appears capable of providing insight into the social behavior of different kinds of societies. One of our primary aims here is to see whether these goals are being met with the data available from avian societies. First we discuss several key issues related to avian social behavior that skew theory has largely ignored, or that are difficulties in applying skew theory as currently developed. Although there is overlap, we do not attempt to duplicate all the issues raised by the several excellent prior reviews of reproductive skew in birds (Johnstone 2000, Magrath & Heinsohn 2000, Magrath et al. 2004). As these reviews make clear, the available empirical data in birds fall far short of what is needed to unambiguously test the assumptions of any particular skew model, much less test quantitative predictions in ways that allow discriminating among the different models. Although we fully agree with this rather damning criticism, our goal here, as in any good review, is to stimulate rather than stifle further research. Toward this end, we devote the latter half of the chapter to examining the empirical evidence, both intra- and interspecific, relevant to reproductive skew and avian social behavior, concluding with a prospectus of how we envision skew theory evolving in the future and a discussion of the extent to which it is ever likely to meet Reeve & Keller’s (2001) ambitious goal.
Skew theory and birds: issues Before summarizing the available reproductive skew data for birds, we discuss a series of problems having to do generally with applying skew theory to avian social organization. The issues we focus on are not necessarily limited to birds and are by no means comprehensive. For additional critiques of skew theory as it applies to birds, see Johnstone (2000), Magrath & Heinsohn (2000), and Magrath et al. (2004). Incest avoidance and potential versus actual reproductive roles
Skew quantifies the extent to which reproduction is shared among potentially reproductive individuals. Thus, in order to measure skew, individuals that are not potentially reproductive should be excluded. Often this is neither controversial nor ambiguous: for example, it is generally clear that
Reproductive skew in avian societies fledglings or other immature, non-reproductive individuals should be excluded, since such individuals are incapable of breeding. Unfortunately, this distinction quickly becomes problematical when considering adult helpers that are physiologically capable of breeding but constrained socially from doing so. If the constraint involves reproductive suppression on the part of a dominant, such individuals should clearly be considered as potential breeders, as this is exactly the kind of situation for which skew theory was developed. But what if the constraint is incest avoidance? Incest avoidance is a genetic mechanism that functions as a constraint on breeding by reducing the potential fitness benefits of reproduction between closely related individuals. Helpers in many cooperative breeders live in groups where their only breeding options are incestuous, and consequently are non-breeders not because they are physiologically incapable of breeding but because the fitness benefits of doing so are presumably too low to make the effort worthwhile. Although this contrasts sharply with co-breeders that do not receive a share of reproduction, distinguishing the two can be problematical. Consider the acorn woodpecker, whose social system involves both matesharing by relatives and offspring that delay dispersal to become non-breeding helpers (Figure 9.1). We know that incest is not only avoided in this population, but that incest avoidance, rather than reproductive competition, is the primary mechanism restricting reproduction by offspring (Koenig et al. 1998, Koenig & Haydock 2004). Consequently, we are able to define non-breeding helpers operationally as birds fledged within the group whose parent of the opposite sex (or an opposite-sex co-breeder of their parent) is still present within the group. Meanwhile, co-breeders are defined based on their potential, rather than their actual, success at parenting offspring. For joint-nesting females, reproductive skew is absent and in most cases all females classified as breeders lay eggs (Haydock & Koenig 2002). For co-breeding males, however, things are not so straightforward. Bias in reproductive success between co-breeder males appears to be quite strong, on the order of 6 : 1 or more for individual nests, such that a majority of broods are sired by a single male and many co-breeding males do not sire offspring in any individual nesting attempt. Even if they successfully copulated with the female and their sperm had good opportunity to fertilize eggs, many males will by chance alone fail to sire offspring in a particular nest, since the number of young fledged is usually small, on the order of 2–5. Such “failed breeders” cannot be distinguished from nonbreeding helpers based on patterns of parentage, since neither sire any offspring. Instead, the distinction between the two classes of birds rests solely on
231
232
W. D. Koenig et al.
Figure 9.1 Two adult male acorn woodpeckers, a group-breeding species common in oak woodlands of western North America. Acorn woodpeckers exhibit both helping at the nest by non-breeders and cooperative polygynandry in which multiple co-breeder males compete for matings with 1–3 joint-nesting females, all of which lay their eggs in the same nest cavity. Co-breeders of the same sex (males and females) are almost always close relatives, usually siblings but sometimes parents and offspring or a combination of the two. Skew is low among joint-nesting females, who regularly destroy each other’s eggs but ultimately share maternity of eggs equally. In contrast, one male often sires all young in a nesting attempt, and skew among co-breeding males is thus apparently quite high. This results in two distinct kinds of non-reproductive males: helpers that do not attempt to mate due to incest avoidance, and potential cobreeders that fail to sire young in a particular breeding attempt. Photo ª 2000 by Ron Mumme.
the difference between their perceived reproductive roles. Is it necessary or desirable to distinguish these two sets of individuals? Dismissing the potential reproductive status of failed breeders and classifying them as non-breeding helpers along with the offspring that are constrained
Reproductive skew in avian societies from breeding due to incest avoidance clearly obscures an important functional difference between them and can be misleading. Reeve & Keller (1995), for example, proposed that the lack of reproduction by helpers observed in many cooperative breeders in which helpers are offspring, compared to systems in which helpers are siblings, supports a prediction of optimal skew theory stemming from the difference in symmetry of genetic relatedness between the potential breeders in the two situations and their offspring. As pointed out by Emlen (1996), however, the observation that helpers that are offspring do not usually breed is more parsimoniously explained by incest avoidance rather than optimal skew theory. That is, because of incest avoidance, such helpers are not part of the breeding activities of the group and therefore their lack of reproduction can neither be explained by skew theory nor provide support for it. On the other hand, there are at least two arguments for including nonbreeding helpers that are offspring of the “true” breeders in a group when considering or trying to measure reproductive skew. First, although such individuals have been empirically found not to successfully parent offspring in many species, excluding them rests on the assumption that reproductive suppression is primarily, if not exclusively, due to incest avoidance rather than reproductive competition or some other mechanism that is accommodated by skew theory. This is a reasonable assumption only if one agrees that incest avoidance is a strong and universal factor affecting mating patterns in vertebrate societies, a hypothesis that has certainly not been generally accepted historically (Bengtsson 1978, Shields 1982, Pusey & Wolf 1996), and remains controversial today for both theoretical and empirical reasons (Pusey & Wolf 1996, Koenig & Haydock 2004). The second argument for including non-breeding helpers when measuring skew is that even if incest avoidance is important in a society, “non-breeding helpers” are not necessarily excluded from reproduction. In rare cases they may breed incestuously; although incest avoidance appears to be strong in most vertebrate societies, it is not absolute (e.g. Haydock et al. 2001, Koenig & Haydock 2004). More problematically, as discussed in the next section, such helpers in at least some cases attempt to breed outside of their home group, contributing to high levels of extra-group matings. To the extent that either of these occur, excluding such helpers from consideration of reproductive skew becomes all that much more subjective. Equally disturbingly, by excluding the occurrence of incest and extra-group matings, the range of behaviors that reproductive-skew theory is left to address narrows, leaving only a small fraction of the diversity of avian mating systems within its purvey. How much of a difference does this make? Jamieson (1997) calculated skew using the Pamilo & Crozier (1996) index for multi-male and multi-female
233
234
W. D. Koenig et al. Table 9.1 Reproductive skew (mean ± SE [n broods]) of 128 broods (n offspring > 1) within multi-male and multi-female groups of acorn woodpeckers at Hastings Reservation for which parentage data were determined (Dickinson et al. 1995, Haydock et al. 2001) using the B index of Nonacs (2000, 2003). Category of individuals
Males
Females
Excluding non-breeding helpers
0.215 ± 0.025 [85]
0.034 ± 0.036 [32]
Including non-breeding helpers
0.298 ± 0.022 [106]
0.221 ± 0.030 [69]
groups of pukeko (Porphyrio porphyrio) at one of his study sites two ways and found a difference of a factor of 2.5 between the value calculated for females including helpers (0.42) and that excluding helpers (0.17). As a second example, we have calculated skew for acorn woodpeckers using the B index (Nonacs 2000, 2003) both including and excluding non-breeding helpers (Table 9.1). For males, skew increases a modest 39% if non-breeding helpers are included in the calculations. However, for females, skew increases over six-fold from near 0 to 0.221, concurrently increasing from 16% to 74% of the comparable value for males. Thus, both the magnitude and to some extent the pattern of reproductive skew are dependent on the assumptions being made about who is and who is not a potential breeder in the population. Extra-group parentage
There is considerable variation within cooperatively breeding birds in the frequency of extra-group reproduction, including both extra-group matings and conspecific brood parasitism (see Tables 9.3 and 9.4, below). Among 15 studies of female co-breeding species, for example, conspecific brood parasitism is known to occur in at least six, while at least four (brown jay, Cyanocorax morio; white-browed scrubwren, Sericornis frontalis; Taiwan yuhina, Yuhina brunneiceps [Figure 9.2], and Seychelles warbler, Acrocephalus sechellensis) are known to exhibit a significant (10%) incidence of extra-group matings. In the extreme, females may mate preferentially or exclusively with extra-group males, in which case skew theory is useless in predicting partitioning of reproduction among males within groups. This approaches what has in fact been found in the cooperatively breeding superb fairy-wren (Malurus cyaneus: Mulder et al. 1994, Dunn et al. 1995). Extra-group parentage clearly affects reproductive partitioning within groups, and thus reproductive skew. However, skew theory, designed thus far to address reproductive sharing among co-breeders within groups, has yet to be extended to integrate the potential for reproductive “sharing” by extra-pair
Reproductive skew in avian societies
Figure 9.2 A group of four yuhinas, a small species of joint-nesting babbler endemic to Taiwan, where it is found in a wide range of forest habitats. Yuhina social groups consist of 1–3 monogamous pairs that are generally non-relatives and share the same nest along with an occasional unpaired individual. Skew is low among both males and females, and egg destruction is rarely observed. How yuhinas resolve breeding conflict and achieve such seemingly peaceful cooperation with non-relatives remains to be answered. Photo ª 2007 by Jinn-Juang Hsu.
or extra-group males. Empirically, no current index of skew takes into account the loss of parentage breeders suffer within a group due to extra-group reproduction, and there is no obvious way to apply skew indices to situations in which a resident male, of which there may be only one, shares paternity with one or more extra-group males with which he may or may not interact. Theoretically, this situation clearly illustrates that monopolization of reproduction by a resident male may require control over more individuals than one’s co-breeders. Rather, such males must exert control over their female mate, or alternatively over all potential sires in the population attempting to obtain – by force in some cases – extra-pair matings with that female. Within socially monogamous species, variation in extra-pair matings appears to be related to a variety of factors, including breeding density, genetic variation, and sexual conflict (Petrie & Kempenaers 1998), and it is likely that similar factors are also important in influencing the frequency of extra-group matings in cooperative breeders. Skew models have yet to be extended to
235
236
W. D. Koenig et al. encompass this intriguing phenomenon, although work by Shellman-Reeve & Reeve (2000) applying a transactional model to the problem of extra-pair matings in socially monogamous species suggests that skew models could readily be developed to address this issue in the more complex cooperative breeding arena as well. Sexual conflict
Extra-group matings are a specific example of the more general omission of sexual conflict, defined as individuals of one sex imposing reproductive costs on the other for the benefit of the imposing sex, in most skew models. Specifically, it is evident that behavior on the part of one sex can often subvert the ability of the dominant member of the opposite sex to control reproduction of subordinates, thus violating this key assumption of skew theory. For example, females may actively choose to mate with more than one male to ensure that their eggs are fertilized, to increase genetic diversity in their offspring, or to obscure paternity and thereby increase the likelihood of obtaining parental care from more than one male (e.g. Stacey 1982). If females successfully seek matings from more than one male, or different males, within the social group, then male reproduction may end up being more egalitarian than predicted by skew theory alone. Alternatively, even when reproductive skew among males within the group is high due to female choice, if females seek matings from males outside the social group and prefer different extragroup mates than those preferred as within-group mates, variance in reproductive success among males in the population may be low (Webster et al. 1995). Among cooperatively breeding birds, the classic example of sexual conflict is the dunnock, in which Davies (1990, 1992) has shown that the fitness consequences of mate sharing are positive for females that are able to monopolize the parental effort of multiple males, but not for the males themselves. Conversely, having access to plural-nesting females allows male dunnocks to achieve higher fitness but reduces the reproductive success of the females themselves. Thus, the mating system, including how reproduction is shared within groups, appears to be mainly driven by sexual conflict rather than factors considered by skew models. Even more generally, skew models, to the extent that they focus exclusively on within-sex competition, ignore any means by which the opposite sex may affect the pattern of reproductive sharing among potential co-breeders within a group. Given the importance of such behaviors in so many systems (Andersson 1994), this is clearly a serious omission. Thus far, the only attempt at extending a transactional skew framework to include sexual conflict and the possibility of female control over the
Reproductive skew in avian societies distribution of paternity is that of Cant & Reeve (2002). Their results demonstrate that incorporating this basic aspect of many vertebrate societies yields contrasting predictions about patterns of reproductive skew depending on the assumptions made about the relationship between paternity, paternal care, and the options available to subordinate males. It is likely that some level of female control is at work in many, if not all, avian cooperative breeders, and thus that skew models dealing with such societies will need to consider the behavior of both sexes in order to successfully predict patterns of reproductive partitioning. Group size and sociality
Several papers have extended transactional models to groups of more than two co-breeding individuals (Johnstone et al. 1999, Reeve & Emlen 2000, Reeve & Jeanne 2003). What has yet to be addressed, however, is variation in group size and in the proportion of the population living in groups. This omission is paradoxical, given the explicit focus of skew theory on whether subordinates should remain in or leave groups. In acorn woodpeckers at Hastings Reservation, California, USA, for example, only half of groups consist of two or more co-breeder males and only 23% consist of joint-nesting females (usually two, but rarely three). Indeed, the most common breeding group composition is a breeding pair of individuals (plus a variable number of nonbreeding helpers), which make up 42% of all breeding groups. Among social birds in general, variation in the frequency of breeding coalitions varies considerably, from <10% among female oystercatchers (Haematopus ostralegus) and Arabian babblers (Turdoides squamiceps) to 90% in Taiwan yuhina (see Table 9.3, below). Can skew theory provide guidelines to assist in explaining the frequency of co-breeding groups in the population? Or, even more ambitiously, can it help explain differences in group size across populations within and among species? Thus far, the answer is no. Transactional skew theories are predicated on the assumption that individuals are able to choose to leave and breed independently if they do not receive sufficient fitness benefits from remaining in a group. In acorn woodpeckers, adult breeders do frequently leave reproductive positions for other opportunities (Haydock & Koenig, unpublished data), and it is likely that such fluidity exists in other systems also. However, current skew models do not predict the frequency with which individuals should share mates or variation in group size within or among species. This is clearly another area in which future development of skew theory might be particularly useful. As an example involving the concordance between group formation and relatedness, consider male coalitions of wild turkeys. Considered a textbook example of kin selection (Watts & Stokes 1971), recent molecular evidence has
237
238
W. D. Koenig et al. confirmed that males are closely related, and only the dominant male reproduces (Krakauer 2005a). Coalitions fathered offspring at approximately seven times the level of solitary males, and the indirect fitness benefits for subordinate members more than made up for the loss of independent breeding opportunities. Such high skew among co-breeding relatives is generally consistent with predictions of optimal skew theory. Conversely, skew theory predicts greater reproductive sharing among male coalitions of non-relatives, for which there should be strong selection given the immense reproductive advantage of coalitions compared with solitary males. Such coalitions of unrelated males have not been observed. Optimal skew theory, which assumes that conflict is avoided because of complete dominance, offers little to resolve this discrepancy. Instead, it may be explainable if coalitions of non-relatives resolve conflict in a completely different way than do related males, perhaps by means of a costly “tug-of-war” that ultimately renders cooperation unprofitable. Unfortunately, in lieu of unrelated coalitions to observe or the possibility of appropriate field experiments, this hypothesis, entailing the intriguing possibility that different skew models apply in different situations within the same population, remains speculative. Null models and relatedness asymmetries
Finally, there are several technical issues related to measuring reproductive skew in birds that deserve comment. It is important to keep in mind that although the basis of most avian societies is a set of closely related individuals, such societies still can differ in ways that are critical in terms of how they affect the predictions of the various skew models. For example, when cobreeders are siblings, relatedness is symmetrical: that is, each individual is equally closely related to the other’s offspring. In contrast, when co-breeders are parents and offspring, relatedness is potentially asymmetrical, with the offspring being more closely related to the parent’s subsequent young than the parent is to his or her grandchildren. Such differences have important consequences to predictions of reproductive skew in all models that are sensitive to relatedness (Reeve & Keller 1996). Specifically, when relatedness of cobreeders is asymmetrical, offspring produced by the dominant are typically equivalent, from the subordinate’s perspective, to the subordinate’s own offspring, and thus skew is generally expected to be greater than if relatedness is symmetrical. Unfortunately, asymmetrical relatedness only occurs if the parent mates with a relative (most often, presumably, the other parent) of the co-breeding offspring while the co-breeding offspring mates with a different, unrelated individual. It does not arise in the situation, such as occurs in acorn woodpeckers, in which
Reproductive skew in avian societies offspring only co-breed after all breeders of the opposite sex have disappeared and been replaced by non-relatives. Thus, asymmetrical relatedness is only likely to be an issue in very complex societies in which related and unrelated individuals of the same sex potentially co-breed, such as white-winged choughs (Corcorax melanorhamphos: Heinsohn et al. 2000). A second important issue, given the small sample sizes generally available from vertebrate societies, is the need to quantify skew using an index that is not sensitive to factors such as group size or differences in productivity that may differ between the groups being compared in a way that will obscure the pattern of interest (Nonacs 2000, 2003). Related to this problem is the need to test values statistically against an appropriate random model (Nonacs 2003). Such models can be modified to deal with quite complicated situations, such as when both productivity varies among groups and there is variability among group size. This latter difficulty is particularly problematical, since parentage cannot necessarily be divided exactly the same way among different numbers of co-breeders (Haydock & Koenig 2003). Unfortunately, using an appropriate skew index and testing it against an apparently valid null model does not guarantee that the results will be meaningful. There are almost always going to be multiple ways to calculate skew, even using the same index. In many cases it will be possible to calculate skew using different categories of group members (Jamieson 1997). Often it will be possible to calculate skew on a variety of time scales, including a broodby-brood basis, a seasonal basis, or (at least theoretically) a lifetime basis, differences that can lead to very different conclusions. Moreover, skew can be calculated for differing subsets of the population or across the population as a whole, although an index encompassing all groups, whether they involve cobreeders or not, clearly measures something other than reproductive sharing as usually envisioned in skew theory. A related issue that is becoming more important as molecular techniques continue to advance is to keep in mind that knowledge of outcomes – specifically of parentage – does not necessarily provide insight as to mechanisms and thus may complement but usually not replace behavioral observations. Indeed, by failing to provide any hints concerning potential mechanisms, a purely molecular approach can be distinctly misleading. As an example, consider again co-breeding male acorn woodpeckers. When measured within broods, such coalitions exhibit significant reproductive skew, with a single male frequently gaining paternity of all young. However, across broods there is often switching of paternity, leading to considerably lower skew when multiple nesting attempts are combined (Haydock & Koenig 2002, 2003). Such switching, which has also been observed in brown jays (Williams
239
240
W. D. Koenig et al. 2004) and to some extent in Gala´pagos hawks (Buteo galapagoensis: Faaborg et al. 1995), raises the possibility that, contrary to what is usually assumed to be the case, paternity may not be independently determined for each egg within broods. Indeed, for acorn woodpeckers, an excellent fit to the observed level of reproductive skew can be obtained when nesting attempts for the same set of males are combined and it is assumed that paternity is a winner-take-all affair 72% of the time, the observed probability of single-sired broods (Haydock & Koenig 2002). Thus the observed skew level can simply be attributed to nonindependence of paternity. Further behavioral observations and physiological studies on sperm competition may help to resolve how molecular results should be interpreted. To the extent that such non-independence of paternity is the case, the observed outcome of high skew within broods is a red herring, suggesting that there is a strongly unequal probability of paternity among co-breeder males (that is, high skew) when in fact the chances of co-breeders fathering any particular brood (and thus any particular offspring) may actually be equal. The take-home message is that in order to determine the appropriate null model and thus to interpret a particular pattern of skew, it may often be necessary to understand the mechanisms by which the pattern is obtained.
Tests of skew theory in birds We now turn to empirical studies of skew in birds in order to examine the extent to which current theory successfully predicts patterns of skew within and among populations of birds. In general, we focus on the predictions of optimal skew or concessions theory, since this is the only framework that has thus far been tested in birds to any significant extent. However, it is important to keep in mind that the predictions of optimal skew theory are in some cases opposite to those of other skew models (e.g. the restraint or the tugof-war model), and thus that support or non-support for this particular model does not constitute a test of skew theory in general. With this in mind, we first consider intraspecific tests and then present a meta-analysis across species using published data on reproductive skew among birds in which co-breeding is known to occur. Intraspecific tests
We are aware of 12 intraspecific tests of concession skew theory involving seven different species (Table 9.2). Tests involved a wide range of comparisons, from different populations to differences among years or the sexes. All three major ecological and social factors encompassed by skew
Reproductive skew in avian societies theory have been tested, although there appears to be no study that claims to compare differences in the benefits of group living independent of ecological constraints. Two of the tests focused on the frequency and apparent strength of competitive interactions rather than differences in reproductive skew. All studies entail one or more caveats limiting the extent to which the comparison is a strong test of skew theory (Table 9.2). We discuss each briefly. Pukeko (Porphyrio porphyrio)
Jamieson (1997) contrasted skew in two populations of this polygynandrous species, one of which involved closely related co-breeders apparently living under conditions of high constraints and the other of which apparently involved unrelated co-breeders living under less ecologically restricted conditions. Comparisons were made in the degree of reproductive skew and the amount of dominance-related aggression, both of which were predicted to be higher in social groups composed of close kin in which ecological constraints were high. As predicted, skew was higher in the population living under higher ecological constraints in which co-breeders were genetic relatives. There was also some indication that dominance interactions were more prevalent and intense in the high-constraint, high-relatedness population. There are several caveats to these results. With respect to the dominance interactions, the differences were not entirely clear-cut and, as pointed out by Jamieson (1997), the assertion of dominance could by itself lead to high skew rather than being a consequence of high skew. As for the calculated skew values, differences between the populations were not significantly different for males, values for the high-skew population were estimated from a nearby study, and the index used (Pamilo & Crozier 1996) is influenced by group size and productivity (Nonacs 2000, 2003). Consequently, the observed differences in skew between the populations may to some extent be artifactual rather than due to differences in ecological constraints or relatedness of cobreeders. Common moorhen (Gallinula chloropus)
McRae (1996a) found that the proportion of eggs laid by dominant females in joint nests of this species was significantly higher when co-breeders were first-order relatives compared to when they were not known to be relatives, in accordance with optimal skew theory. However, as she points out, the sample size for apparently unrelated females was small (n ¼ 3, one of which was an adopted daughter) and the result is potentially confounded by age effects, which were highly significant in the population.
241
242
Table 9.2 Summary of intraspecific tests of concessions theory in social birds Results consistent with Species
Comparison
Difference
Prediction
prediction?
Comments
References
Pukeko (Porphyrio
Between
Constraints;
Greater skew in
Yes
Difference not
Jamieson 1997
porphyrio)
populations
relatedness
population with higher
significant for males;
constraints and related
skew not tested against
breeders
random models; data from one population inferred from prior nearby study.
Pukeko
Dominance
Constraints;
Greater dominance in
interactions
relatedness
high-skew societies
quantitative data;
with high constraints
possible circularity
Yes
No unambiguous
Jamieson 1997
and close relatedness Common moorhen
Female coalitions
Relatedness
(Gallinula chloropus)
Greater skew among
Yes
related cobreeders
Small sample size;
McRae 1996b
possibly confounded by age effects
Acorn woodpecker (Melanerpes formicivorus)
Between sexes
Constraints;
Greater skew in sex
Yes
No skew among
Haydock &
benefits of group
with higher constraints
females; male skew
Koenig 2002,
living
and greater group
depends on
2003
benefits
assumption of paternity independence
243
Acorn woodpecker
Male coalitions
Relatedness
Greater skew among
No
related cobreeders
Difference in
Haydock &
relatedness probably
Koenig 2003
small; small sample size Acorn woodpecker
Male coalitions
Constraints
Greater constraints and No
Male skew depends on Haydock &
(coalition size)
greater skew in smaller
assumption of
coalitions
paternity
Koenig 2003
independence Acorn woodpecker
Among years
Constraints
Greater skew in years
Partially
No significant
Haydock &
(reproductive
with fewer vacancies
relationship when
Koenig 2003
vacancies)
and higher constraints
relative, rather than absolute, number of vacancies used
Brown jay (Cyanocorax
Between sexes
morio)
Constraints;
Greater skew in sex
relatedness
with higher constraints
Relatedness
Greater competition
Yes
Female competitive interactions
Williams 2004
nests/years may change
and higher relatedness Brown jay
Switching between observed pattern
No
among closely related
Relatedness based on
Williams 2004
band sharing
cobreeders White-winged chough
Coalition structure
(Corcorax
Constraints;
Greater skew during
Complex social
Heinsohn et al.
relatedness
period of higher
structure consisting of
2000
constraints and higher
mix of related and
melanorhamphos)
Yes
relatedness Arabian babbler (Turdoides squamiceps)
Male coalitions
Relatedness
Greater skew among related cobreeders
unrelated coalitions Yes
Relatedness based on
Lundy et al.
band sharing; possibly
1998
confounded by group stability
244 Table 9.2 (cont.) Results consistent with Species
Comparison
Difference
Prediction
prediction?
Comments
White-browed
Male coalitions
Relatedness
Greater skew among
Yes
Possible confound with Whittingham
scrubwren (Sericornis frontalis)
related cobreeders
extra-group paternity
References
et al. 1997
Reproductive skew in avian societies Acorn woodpecker (Melanerpes formicivorus)
Polygynandry occurs regularly in this species in conjunction with helping at the nest by non-breeders. Based on comparative demography, Haydock & Koenig (2002, 2003) found that males appeared to be living under conditions of greater ecological constraints and were able to gain greater fitness benefits by living in groups than females. Optimal skew theory predicts that these differences should result in greater skew among co-breeder males, and this was indeed found to be the case. Although the difference was in accordance with skew theory, other findings were not. Females, for example, exhibited no skew whatsoever, sharing reproduction more equally than expected by chance. This pattern is enforced by egg destruction (Mumme et al. 1983, Koenig et al. 1995), a phenomenon eliminating any opportunity for females to control reproduction. Meanwhile, male skew is high within broods but is unrelated to male age or physical characteristics and often switches between broods. This suggests the possibility that paternity of eggs may not be independently determined, as is normally assumed, and that co-breeder males may actually have equal chances of siring eggs despite the high skew found in individual nests. Haydock & Koenig (2003) tested several other predictions of optimal skew theory as well. Among males there was no difference in skew in coalitions that were close relatives (r ¼ 0.5) versus coalitions where males were less closely related, contrary to expectations. There also tended to be an increase, rather than the expected decrease, in skew with increasing coalition size, which was assumed to correlate with lower ecological constraints since larger coalitions are known to be more competitive for reproductive vacancies and thus should provide more opportunities for leaving the group. The last test involved comparing reproductive skew in different years, indexing the degree of ecological constraints by counting the number of male reproductive vacancies occurring within the population during the three months prior to each nesting attempt. Skew was higher when there were fewer vacancies, in accord with skew theory. However, when dividing the mean number of vacancies by the number of individuals potentially competing for those vacancies, the results were no longer significant, thus providing at best only partial support for skew theory. Brown jay (Cyanocorax morio)
Females in this tropical jay can apparently be either joint or plural nesting, although the latter was not observed in the parentage study by Williams (2004). Williams compared skew in the two sexes, predicting that it should be greater among females, for which dispersal is apparently more
245
246
W. D. Koenig et al. constrained and relatedness was higher based on band-sharing. As expected from concession theory, skew was significantly higher among females. Williams (2004) also compared whether he observed aggressive interactions between joint-nesting females and found that females that have higher band-sharing scores, and thus were presumably more closely related, were less, rather than more, likely to engage in nesting aggression. White-winged chough (Corcorax melanorhamphos)
Heinsohn et al. (2000) documented the amalgamation of coalitions of relatives that occurred in this species following a severe drought and reported that among such groups males were more likely to share reproduction than females and that skew was higher prior to the drought, when groups were composed exclusively of relatives, than afterwards, when they were composed of related and unrelated birds. This latter finding supports the expected pattern of higher skew when co-breeders are close relatives. Arabian babbler (Turdoides squamiceps)
In a genetic analysis of this species, Lundy et al. (1998) found significantly higher band-sharing between males that did not share paternity compared to males that did, supporting the prediction of skew theory that there should be higher skew among more closely related individuals. A possible confound in this study was group stability, since shared paternity was more likely to occur following major changes in group structure, which were also more likely to involve non-relatives. White-browed scrubwren (Sericornis frontalis)
Similar to the Arabian babbler, Whittingham et al. (1997) found that males in this species were more likely to share paternity among non-relatives, a result they carefully disentangled from incest avoidance. However, this population also exhibited a fairly high incidence of extra-group paternity, which was more frequent in groups where the subordinate male was unrelated to the dominant. As a result, although subordinate males gained more parentage when unrelated to the dominant, the overall share of reproductive monopolization by the dominant did not differ appreciably between groups in which males were related (17 of 26 young; 66% paternity) compared to groups in which males were unrelated (12 of 24 young; 50% paternity; Fisher exact test p ¼ 0.21). To conclude, of the 12 intraspecific tests summarized in Table 9.2, the results of nine (75%) are fully or partially consistent with optimal skew theory. Of the seven involving differences in ecological constraints, six (86%) are consistent with skew theory at least in part, while only three of five (60%) tests
247
Table 9.3 Summary of the life-history and behavioral correlates of skew in female social birds. “Joint” species are those in which females nest communally; “plural” species are those in which females produce separate nests within a larger social group. CBP, conspecific brood parasitism. Mean
Extra-
Frequency
coalition
Relatedness
Ecological
Benefits of
Dominance
Competition group
Species
of coalitions
size
of coalitions
constraints
coalitions
obvious?
obvious?
matings?
Magpie goose
“Common”
2.0 (joint)
High
?
Low
Yes
Skew?
References
Yes (egg
Low (?) (but Low
Frith and
(Anseranas
tossing
CBP
Davies 1961,
semipalmata)
likely)
occurs?)
Horn et al.
Egg tossing
Low (but
possible
CBP
1986, McRae
common)
1996a, 1996b
1996 Common
21%
2.1 (joint)
High
High
Moderate
No
moorhen (Gallinula
Low
Gibbons
chloropus) Pukeko
26%
2.0 (joint)
Low
Low
Low
Yes
No
(Porphyrio
No (but
Low
CBP occurs)
Jamieson et al. 1994,
porphyrio)
Lambert
(Otokia)
et al. 1994, Jamieson 1997
Oystercatcher
1%
2.0 (joint)
Low
High
Low
No
No
?
Low
(Haematopus
Heg & van Treuren 1998
ostralegus) Acorn woodpecker (Melanerpes formicivorus)
23%
2.1 (joint)
High
High
Low
No
Yes (egg tossing)
No
Low/none
Haydock &
(switching) Koenig 2002, 2003
248 Table 9.3 (cont.) Mean
Extra-
Frequency
coalition
Relatedness
Species
of coalitions
size
of coalitions
Guira cuckoo
94%
3.1 (joint)
Low
Ecological
Benefits of
Dominance
Competition group
constraints
coalitions
obvious?
obvious?
matings?
Skew?
References
Low
Moderate (?) No
Yes (egg
No (but
Low/none
Macedo 1992,
tossing;
CBP occurs) (switching) Macedo et al.
(Guira guira)
infanticide) Groove-billed
30%
2.7 (joint)
Low
High
Moderate
Yes
ani (Crotophaga
Yes
2004, 2005 No (but
Low
(egg-tossing) CBP occurs)
Vehrencamp 1977, Koford
sulcirostris)
et al. 1986, 1990, Vehrencamp & Quinn 2004
Dunnock
38%
2.2 (plural) Low
Low
Low
No
No
Low
Low
(Prunella
Davies 1990, 1992
modularis) Brown jay
2.1 (plural; Moderate
Moderate to Low
(Cyanocorax
29%
sometimes (62% related)
low
morio)
joint)
Yes
Yes (some
17%
Low
egg tossing)
Lawton & Lawton 1985, Williams et al. 1994, Williams 2004
Taiwan yuhina 90% (Yuhina brunneiceps)
2.3 (joint)
Low
Low
High
Yes
Yes (egg tossing)
Low
Low
Yuan et al. 2004, S.-F. Shen unpubl. data
249
Pukeko
79%
2.0 (joint)
High
High
Low
Yes
No
(Shakespear)
No (but
Moderate
CBP occurs)
Jamieson et al. 1994, Lambert et al. 1994, Jamieson 1997
Tasmanian
42%
2.0 (joint)
High
High
native hen
Moderate to No
No
No
low
Moderate
Goldizen et al.
(> males)
1998, 2000
(Gallinula mortierii) Seychelles
30%
2.0 (joint)
High
High
Present on
warbler
high, but
(Acrocephalus
not low
sechellensis)
quality
Yes
No
40%
Moderate
Komdeur 2005
territories White-winged
32%
2.5 (joint)
chough
Moderate
High
(50% related)
(foraging
destruction;
Heinsohn et
constraints)
egg tossing)
al. 2000
(Corcorax
High
Yes
Yes (nest
Low (?)
High
Rowley 1978,
melanorhamphos) Arabian babbler (Jurdoides squamiceps)
9%
2.0 (joint)
High
High
Moderate to low
Yes
No (?)
No
High
Zahavi 1990, Lundy et al. 1998
250
Table 9.4 Summary of the life-history and behavioral correlates of skew in male social birds. The last two species exhibit cooperative courtship; all other species are either cooperatively polyandrous, in which males share one or more females (“unpaired”), or species in which male–female pairs form breeding coalitions (“paired”). Freq. of
Coalition
Relatedness
Ecological
Benefits of
Dominance
Species
coalitions
size
of coalitions
constraints
coalitions
obvious?
matings
Skew?
References
Gala´pagos hawk
68%
2.5
Low
High (?)
Low
No
Low
Low
Faaborg &
(switching)
Bednarz 1990,
(Buteo galapagoensis)
Extra-group
(unpaired)
Faaborg et al. 1995 Groove-billed ani
30%
2.7 (paired)
Low
High
Moderate
Yes
Low
Low
Vehrencamp
(Crotophaga
1977, Koford
sulcirostris)
et al. 1986, 1990, Vehrencamp & Quinn 2004
Taiwan yuhina
90%
2.5 (paired)
Low
Low
High
Yes
10%
Low
(Yuhina brunneiceps)
Yuan et al. 2004, S.-F. Shen unpubl. data
Dunnock (Prunella
64%
2.0
25%
2.4
modularis) Tasmanian native hen (Gallinula mortierii)
Low
Low
Low
No
Low
Low
High
High
Moderate
No
No
Moderate
Goldizen et al.
to low (<
1998,
females)
Goldizen et al.
(unpaired) (unpaired)
Davies 1990, 1992
to low
2000
251 White-winged
79%
chough (Corcorax
3.5 (partly
Moderate to
High
unpaired)
low (23% of
(foraging
pairs related)
constraints)
melanorhamphos) White-browed
31%
scrubwren
High
2.2
Moderate (50% Moderate (?) Low
(unpaired)
of coalitions
(Sericornis frontalis)
Yes
Yes
Low (?)
12%
Moderate to
Heinsohn
low
et al. 2000
Moderate to
Whittingham
low
et al. 1997,
related)
R. Magrath pers. comm.
Pukeko (Porphyrio
46%
porphyrio) (Otokia)
2.3
Low
Low
Low
(unpaired)
Yes (but
No
Moderate
Jamieson et al.
unrelated to
1994, Lambert
paternity)
et al. 1994, Jamieson 1997
Acorn woodpecker
49%
(Melanerpes
2.6
High
High
Moderate
No
No
(unpaired)
Moderate (?) Haydock & (switching)
Koenig 2002,
Moderate
Lawton &
(switching;
Lawton 1985,
> females)
Williams et al.
formicivorus) Brown jay
2003 High
? (unpaired)
Low
Low
Yes
No
17%
(Cyanocorax morio)
1994, Williams 2004 Brown skua
33%
(Catharacta
2.2
Low
Moderate (?) Low
No
No
(unpaired)
Moderate
Millar et al.
(switching)
1994
Moderate
Jamieson et al.
to high
1994, Lambert
antarctica) Pukeko (Shakespear)
100%
3.2 (unpaired)
High
High
Low
Yes
No
et al. 1994, Jamieson 1997
252 Table 9.4 (cont.) Freq. of
Coalition
Relatedness
Ecological
Benefits of
Dominance
Species
coalitions
size
of coalitions
constraints
coalitions
obvious?
matings
Skew?
References
Arabian babbler
34%
2.0
High
High
Moderate to
Yes
No
High (but
Zahavi 1990,
< females)
Lundy et al.
High
MacDonald
(Turdoides
(unpaired)
Extra-group
low
squamiceps) Long-tailed
1998 100%
manakin
2.0 (þ 7.1
Low
?
High
Yes
No
auxilliaries)
1989,
(Chiroxiphia linearis)
MacDonald & Potts 1994
Wild turkey (Meleagris gallapavo)
33%
2.4
High
?
High
Yes
No
High
Krakauer 2005a, 2005b
Reproductive skew in avian societies involving differences in relatedness were in the direction predicted. There are several reasons why these results provide at best only weak support of optimal skew theory, including that the predictions are qualitative, in some cases relatively superficial, may be biased toward positive findings, and generally fail to consider alternative models. Nonetheless, they offer at least some encouragement that optimal skew theory may be useful in understanding, if not predicting, patterns of reproductive sharing within some avian populations. Clearly additional work remains to be done. Recent advances in quantifying and statistically testing skew, and in molecular techniques allowing estimation of relatedness, should help future workers obtain more accurate and less biased estimates of these variables. Better knowledge of mechanisms will help resolve questions concerning the significance of paternity switching between nests and whether it is always appropriate to assume that paternity of eggs is determined independently of other eggs within the same brood. These advances, combined with increased sample sizes and additional studies, should ultimately provide the opportunity for considerably stronger intraspecific tests of optimal skew theory than have been conducted thus far. Interspecific tests: a meta-analysis
Based on the literature, we summarize the cogent information currently available on reproductive skew for different populations of birds, dividing the data by sex (Tables 9.3 and 9.4). Each species is listed once, except for the two populations of Pukeko studied by Jamieson (1997), which were considered separately due to the contrasting conditions under which they appeared to be living. In each case, we estimated, based on the data provided, the frequency of co-breeding coalitions and the mean size of coalitions (i.e. the mean number of individuals in coalitions excluding singletons). We then went on to categorize (usually as being “high,” “moderate,” or “low”) the overall level of reproductive skew exhibited by the population along with relatedness of coalitions, the benefits of reproductive coalitions, and the relative strength of ecological constraints to independent reproduction within the system. These latter three variables are widely acknowledged to be key ecological factors in the evolution of cooperative breeding, although their precise role and relative importance in particular systems is the subject of considerable debate (e.g. Dickinson & Hatchwell 2004, Ekman et al. 2004). We also noted whether dominance was obvious or not, whether (for females) competition among co-breeders was evident, whether extra-group matings in the form of conspecific brood parasitism and/or extra-group paternity occurs, whether males and females are generally paired or unpaired, and (for males) whether male parental care occurs or not. Male parental care is found in the majority of
253
254
W. D. Koenig et al. Table 9.5 Relationships between life-history parameters and coalition frequency, coalition size, and reproductive skew across 15 populations of co-breeding female birds (Table 9.3) and 15 populations of co-breeding male birds (Table 9.4). Not all parameters were available for all species. Tests are Spearman rank correlations (rs listed) or Mann–Whitney U tests (z value listed; ‘ þ ’ larger coalitions or greater skew with increases in or greater presence of the variable; ‘ ’ smaller coalitions or less skew with increases in or greater presence of the variable). þ ¼ p 0.10; * ¼ p 0.05; ** ¼ p 0.01). Coalition frequency Males Coalition frequency
–
Females –
Mean coalition size
Reproductive skew
Males
Females
Males
Females
0.09
0.44
0.01
0.06
0.06
0.31
(rs) Coalition size (rs) Genetic relatedness
0.09
0.44
0.20
0.21
0.32
–
0.53*
–
0.03
0.43
0.73**
0.36
0.16
0.34
0.35
0.67
0.26
0.61
0.51
–
0.44þ
0.51*
(rs) Ecological
0.11
0.53*
0.23
0.44þ
1.03
1.13
constraints (rs) Benefits of group
0.44þ
living (rs) Dominance
0.50
(z value) Competition
–
2.57**
–
1.52
(z value) Extra-group mating
0.00
0.32
0.74
0.12
0.89
1.04
0.39
1.09
1.58
2.29*
1.92þ
0.26
0.55
–
0.28
–
2.26*
–
(z value) Paired vs. unpaired (z value) Paternal care (z value)
species (which are either cooperatively polyandrous or in which pairs of birds form cobreeding coalitions), but not in the last two species (long-tailed manakin, Chiroxiphia linearis, and wild turkey), in which males exhibit cooperative courtship. The results confirm the existence of considerable variability across each of these parameters. They also confirm some interesting differences between male and female mate-sharing. In particular, both coalition frequency and mean coalition size are significantly larger among males than females (Wilcoxon signed-ranks tests; n ¼ 9; both p < 0.05), supporting the hypothesis that
Reproductive skew in avian societies females have a harder time sharing breeding opportunities (although not necessarily sharing reproduction per se) than males. This difference is most likely related to the fact that female competition, often taking the form of egg destruction and infanticide, is often much more destructive than male competition, and thus co-breeding is often likely to be more beneficial for males than for females. To what extent do other interspecific patterns correspond to those predicted from optimal skew theory? We used Spearman rank correlations and Mann–Whitney U tests to look for associations between coalition frequency, mean coalition size, and reproductive skew and the various life-history variables summarized in Tables 9.3 and 9.4. Results (Table 9.5) yielded no significant correlations with coalition frequency. However, several significant (p 0.05) and near-significant (p 0.10) correlations emerged with the other two variables. For males, mean coalition size was larger in populations living under greater ecological constraints, whereas for females, mean coalition size was larger in populations where female co-breeders were less closely related, the benefits of group living were greater, competition between females was more evident in the form of egg tossing and infanticide, and birds tended to be paired rather than unpaired. Further, the skew analysis also yielded several promising results. Skew in both sexes tended to be higher when co-breeders were more closely related. In addition, for females, skew was higher in populations living under conditions of greater ecological constraints and higher benefits of group living. Skew tended to be greater among males when they were unpaired rather than paired to females, and among the two species that do not exhibit male parental care. Thus, all three predicted relationships between ecological and genetic factors and reproductive skew were upheld to some degree in females. Only one of the three predicted relationships was upheld in males, but this was the one between skew and genetic relatedness, which is the most counterintuitive relationship predicted by optimal skew theory (Keller & Reeve 1994). As discussed earlier, skew theory does not in general predict any explicit relationships about coalition size. Encouragingly, however, the results involving this variable are intuitively reasonable. That is, we expect mean coalition size to increase with greater ecological constraints, as found in males, and with greater benefits of group living, as found in females. It is also satisfying that overt competition among co-breeding females appears to be greater in populations with larger coalitions, and interesting that coalitions tend to be larger among species in which male–female pairs form, since this may offer at least the start of a hypothesis explaining perplexing variation in the pair-bonding behavior (or lack thereof) observed in different cooperative breeders.
255
256
W. D. Koenig et al. The finding that coalition size is inversely correlated with genetic relatedness is also intriguing, and may have to do with factors other than individual strategic decisions. For example, if potential coalition members are limited to only close relatives, such as apparently is the case in wild turkeys (Krakauer 2005a, 2005b), then the frequency of those relatives will depend upon clutch size and patterns of mortality, and coalitions in such species are likely to be smaller for purely demographic reasons rather than as any reflection of differences in optimal group size. These results offer support for the intriguing possibility that not only particular ecological and genetic factors may predict reproductive skew, but that skew is associated with specific behavioral traits across species. Greater female skew is found in species in which ecological constraints and the benefits of group living are relatively large and genetic relatedness relatively high. Such species tend to exhibit more obvious dominance but less overt competition than low-skew species. Among males, greater skew is associated with closer genetic relatedness, while high-skew species tend to be species in which birds are unpaired and males do not participate in raising young. The correlation between pair-bonding and lower skew suggests that pairing behavior within complex social groups containing multiple breeders counters monopolization by a single individual that might otherwise result in high skew, while the relationship between high skew and lack of male parental care is interesting due to its possible relationship to multiple mating by females (Krakauer 2005b). In the two species in which males do not provide paternal care, there is no potential phenotypic benefit to females mating with more males than necessary to fertilize their eggs, and skew tends to be high. In contrast, in the species where males provide paternal care, females could potentially benefit by mating with multiple males and decreasing male skew because of the relationship between paternal care and opportunity of parentage (Stacey 1982, Hartley & Davies 1994, Cant & Reeve 2002). The lower skew in these species is consistent with this possibility. We conclude that the predictions of optimal skew theory are, at least to some extent, upheld across populations with the currently available data. This level of support is more than we had originally expected. Again, however, these analyses are at best a preliminary attempt at examining skew theory at an interspecific level. More data, phylogenetically controlled analyses, and tests in which alternative models are explicitly pitted against each other will be necessary before it will be possible to conclude that there is strong support for optimal skew theory among birds. Nonetheless, they yield hope that despite its shortcomings, skew theory may eventually provide a valuable tool for understanding many aspects of avian social behavior.
Reproductive skew in avian societies Conclusion Given the logistic difficulties of applying skew theory, combined with the failure of skew theory as usually envisioned to address several key variables in avian social behavior, including the frequency of extra-group matings, variation in group size and sociality, and sexual conflict, it would seem a priori unlikely that skew theory can fulfill the goal of providing a framework for unifying the major features of sociality. Given this starting point, the summary of empirical results reviewed here provides a relatively optimistic outlook for the application of skew theory to an understanding of avian social behavior. Within social birds, numerous key questions remain and experimental studies have yet to be performed, yet a number of studies provide at least some comparative evidence supporting the possibility that reproductive skew can be predicted based on knowledge of ecological constraints, the relative benefits of group living, and genetic relatedness (Table 9.2). More impressively, across species there is evidence not only that these factors can predict reproductive sharing but that such sharing correlates, at least to some extent, with several of the major features of avian societies including the degree of competition between co-breeders, whether birds are paired or unpaired, whether males provide paternal care or not, and to a lesser extent even the extent and obviousness of dominance and reproductive competition. The fact that many of these same parameters correlate with mean coalition size is particularly exciting, since this suggests the possibility that the factors needed to predict reproductive skew may also provide insight into the degree of sociality expressed by different species. Unfortunately, the weakness of many of these relationships, combined with our inability to reject alternative hypotheses, confirms that we have a long way to go before a unifying theory of social evolution, even just for birds, can be achieved. The fundamental spirit of all skew models is to capture the means by which animals resolve their breeding conflicts. Thus far there are mainly two nonmutually exclusive approaches that have been proposed: transactional and compromise models. Transactional models describe the situation in which a controlling dominant may yield certain benefits to a subordinate in order to change the latter’s behavior to the former’s benefit. In contrast, in tug-of-war models breeding conflicts are resolved by costly competition between group members (see also chapter 1). Adding to this central idea, we envision several directions for future advances, both theoretically and empirically. First, it would be useful to broaden the scope of skew theory and relax some of its assumptions in order to apply it to a greater number of systems. One logical frontier for skew theory to encompass, for example, would be to
257
258
W. D. Koenig et al. explain reproductive partitioning in aggregations such as leks (Kokko & Lindstro¨m 1996). Given the large variation in lek size within species (Ho¨glund & Alatalo 1995), as well as recent findings of divergent patterns of male relatedness (Ho¨glund et al. 1999, Petrie et al. 1999, Shorey et al. 2000, Gibson et al. 2005), leks may eventually provide excellent systems for testing models of reproductive skew. Theoretically, a promising step in this direction is potentially the “bordered” tug-of-war model (Reeve & Shen 2006), which combines the option of independent reproduction and reproductive incentive with the potential for within-group conflict, which, as the authors point out, is in theory rendered pointless in optimal skew models as a consequence of the assumption of complete dominant control. Their synthesis model demonstrates that, despite their proliferation, the diversity of skew models can be understood within a general theoretical framework by allowing researchers to ask not only how conflict is resolved in a particular society, but also why a particular mechanism is used in a specific case, thereby moving significantly closer to the goal of understanding reproductive partitioning in a given situation. Most theoretical developments thus far have concentrated on extending transactional models, albeit in various realistic and thus potentially useful ways. We suggest that theoreticians should focus more on modeling alternative conflict-resolution mechanisms and clarifying the relationships between different models. For empiricists, at least two approaches are available. First is attempting to distinguish between different models by testing suites of predictions (e.g. Langer et al. 2004). Second is direct experimental testing of conflict-resolution mechanisms, such as have been performed in social wasps (Reeve & Nonacs 1992) and cooperatively breeding fishes (Heg et al. 2006). Such experiments clearly provide methods of testing the predictions of alternative reproductive-skew models. Although progress is being made, experimental investigation of reproductive skew remains lacking in avian systems. In addition, other key aspects of cooperative-breeding societies, including extra-group matings, sexual conflict influencing reproductive partitioning, and variability in group size and composition, are still largely left behind by current skew models. A truly unified theory of avian social behavior, let alone social evolution in general, will clearly require much further investigation.
Acknowledgments We thank Kern Reeve, Michael Taborsky, an anonymous reviewer, and the editors for their comments, and the Department of Neurobiology and
Reproductive skew in avian societies Behavior, Cornell University, for accommodating the senior author during the preparation of the manuscript. Our work on avian social behavior has been supported by the National Science Foundation.
References Andersson, M. (1994). Sexual Selection. Princeton, NJ: Princeton University Press. Bengtsson, B. O. (1978). Avoiding inbreeding: at what cost? Journal Of Theoretical Biology, 73, 439–444. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Davies, N. B. (1990). Dunnocks: cooperation and conflict among males and females in a variable mating system. In P. B. Stacey & W. D. Koenig, eds., Cooperative Breeding in Birds: Long-term Studies of Ecology and Behavior. Cambridge: Cambridge University Press, pp. 455–485. Davies, N. B. (1992). Dunnock Behaviour and Social Evolution. Oxford: Oxford University Press. Dickinson, J. L. & Hatchwell, B. J. (2004). Fitness consequences of helping. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge, Cambridge University Press, pp. 48–66. Dickinson, J. L., Haydock, J., Koenig, W. D., Stanback, M. T., & Pitelka, F. A. (1995). Genetic monogamy in single-male groups of acorn woodpeckers, Melanerpes formicivorus. Molecular Ecology, 4, 765–769. Dunn, P. O., Cockburn, A., & Mulder, R. A. (1995). Fairy-wren helpers often care for young to which they are unrelated. Proceedings of the Royal Society of London B, 259, 339–343. Ekman, J., Dickinson, J. L., Hatchwell, B. J., & Griesser, M. (2004). Delayed dispersal. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 35–47. Emlen, S. T. (1996). Reproductive sharing in different types of kin associations. American Naturalist, 148, 756–763. Faaborg, J. & Bednarz, J. C. (1990). Gala´pagos and Harris’ hawks: divergent causes of sociality in two raptors. In P. B. Stacey & W. D. Koenig, eds., Cooperative Breeding in Birds: Long-term Studies of Ecology and Behavior. Cambridge: Cambridge University Press, pp. 359–383. Faaborg, J., Parker, P. G., Delay, L. et al. (1995). Confirmation of cooperative polyandry in the Gala´pagos hawk (Buteo galapagoensis). Behavioral Ecology and Sociobiology, 36, 83–90. Foster, M. S. (1977). Odd couples in manakins: a study of social organization and cooperative breeding in Chiroxiphia linearis. American Naturalist, 111, 845–853. Frith, H. J. & Davies, S. J. J. F. (1961). Ecology of the magpie goose Anseranas semipalmata. CSIRO Wildlife Research, 113, 555–557.
259
260
W. D. Koenig et al. Gibbons, D. W. (1986). Brood parasitism and cooperative breeding in the moorhen, Gallinula chloropus. Behavioral Ecology and Sociobiology, 19, 221–232. Gibson, R. M., Pires, D., Delaney, K. S., & Wayne, R. K. (2005). Microsatellite DNA analysis shows that greater sage grouse leks are not kin groups. Molecular Ecology, 14, 4453–4459. Goldizen, A. W., Putland, D. A., & Goldizen, A. R. (1998). Variable mating patterns in Tasmanian hens (Gallinula mortierii): correlates of reproductive success. Journal of Animal Ecology, 67, 307–317. Goldizen, A. W., Buchan, J. C., Putland, D. A., Goldizen, A. R., & Krebs, E. A. (2000). Patterns of mate-sharing in a population of Tasmanian native hens Gallinula mortierii. Ibis, 142, 40–47. Hamilton, W. D. (1964). The genetical evolution of social behaviour. I, II. Journal of Theoretical Biology, 7, 1–52. Hartley, I. R. & Davies, N. B. (1994). Limits to cooperative polyandry in birds. Proceedings of the Royal Society of London B, 257, 67–73. Haydock, J. & Koenig, W. D. (2002). Reproductive skew in the polygynandrous acorn woodpecker. Proceedings of the National Academy of Sciences of the USA, 99, 7178– 7183. Haydock, J. and Koenig, W. D. (2003). Patterns of reproductive skew in the polygynandrous acorn woodpecker. American Naturalist 162, 277–289. Haydock, J., Koenig, W. D., & Stanback, M. T. (2001). Shared parentage and incest avoidance in the cooperatively breeding acorn woodpecker. Molecular Ecology, 10, 1515–1525. Heg, D. & van Treuren, R. (1998). Female–female cooperation in polygynous oystercatchers. Nature, 391, 687–691. Heg, D., Bergmu¨ller, R., Bonfils, D., et al. (2006). Cichlids do not adjust reproductive skew to the availability of independent breeding options. Behavioral Ecology, 17, 419–429. Heinsohn, R., Dunne, P., Legge, S., & Double, M. (2000). Coalitions of relatives and reproductive skew in cooperatively breeding white-winged choughs. Proceedings of the Royal Society of London B, 267, 243–249. Ho¨glund, J. & Alatalo, R. V. (1995). Leks. Princeton, NJ: Princeton University Press. Ho¨glund, J., Alatalo, R. V., Lundberg, A., Rintamaki, P. K., & Lindell, J. (1999). Microsatellite markers reveal the potential for kin selection on black grouse leks. Proceedings of the Royal Society of London B, 266, 813–816. Horn, P. L., Rafalski, J. A., & Whitehead, P. J. (1996). Molecular genetic (RAPD) analysis of breeding magpie geese. Auk, 113, 552–557. Jamieson, I. G. (1997). Testing reproductive skew models in a communally breeding bird, the pukeko, Porphyrio porphyrio. Proceedings of the Royal Society of London B, 264, 335–340. Jamieson, I. G., Quinn, J. S., Rose, P. A., & White, B. N. (1994). Shared paternity among non-relatives is a result of an egalitarian mating system in a communally breeding bird, the pukeko. Proceedings of the Royal Society of London B, 257, 271–277.
Reproductive skew in avian societies Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A., Woodroffe, R., Cant, M. A., & Wright, J. (1999). Reproductive skew in multimember groups. American Naturalist, 153, 315–331. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Koenig, W. D. & Haydock, J. (2004). Incest and incest avoidance. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 142–156. Koenig, W. D. & Mumme, R. L. (1987). Population Ecology of the Cooperatively Breeding Acorn Woodpecker. Princeton, NJ: Princeton University Press. Koenig, W. D., Mumme, R. L., Stanback, M. T., & Pitelka, F. A. (1995). Patterns and consequences of egg destruction among joint-nesting acorn woodpeckers. Animal Behaviour, 50, 607–621. Koenig, W. D., Haydock, J., & Stanback, M. T. (1998). Reproductive roles in the cooperatively breeding acorn woodpecker: incest avoidance versus reproductive competition. American Naturalist 151, 243–255. Koford, R. L., Bowen, B. S., & Vehrencamp, S. L. (1986). Habitat saturation in groove– billed anis (Crotophaga sulcirostris). American Naturalist, 127, 317–337. Koford, R. R., Bowen, B. S., & Vehrencamp, S. L. (1990). Groove-billed anis: joint nesting in a tropical cuckoo. In P. B. Stacey & W. D. Koenig, eds., Cooperative Breeding in Birds: Long-term Studies of Ecology and Behavior. Cambridge: Cambridge University Press, pp. 333–356. Kokko, H. & Lindstro¨m, J. (1996). Kin selection and the evolution of leks: whose success do young males maximize? Proceedings of the Royal Society of London B, 263, 919–923. Komdeur, J. (2005). No evidence for adaptive suppression of joint laying by dominant female Seychelles warblers: an experimental study. Behaviour, 142, 1669–1684. Krakauer, A. H. (2005a). Kin selection and cooperative courtship in wild turkeys. Nature, 434, 69–72. Krakauer, A. H. (2005b). The evolution of cooperative male courtship: kin selection and the mating system of the wild turkey Meleagris gallopavo. Unpublished Ph.D. thesis, University of California, Berkeley. Lambert, D. M., Millar, C. D., Jack, K., Anderson, S., & Craig, J. L. (1994). Single- and multilocus DNA fingerprinting of communally breeding pukeko: do copulations or dominance ensure reproductive success? Proceedings of the National Academy of Sciences of the USA, 91, 9641–9645. Langer, P., Hogendoorn, K., & Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature, 428, 844–847. Lawton, M. F., & Lawton, R. O. (1985). The breeding biology of the brown jay in Monteverde, Costa Rica. Condor, 87, 192–204. Lundy, K. J., Parker, P. G., & Zahavi, A. (1998). Reproduction by subordinates in cooperatively breeding Arabian babblers is uncommon but predictable. Behavioral Ecology and Sociobiology, 43, 173–180.
261
262
W. D. Koenig et al. MacDonald, D. W. (1989). Correlates of male mating success in a lekking bird with male–male cooperation. Animal Behaviour, 37, 1007–1022. MacDonald, D. W. & Potts, W. K. (1994). Cooperative display and relatedness among males in a lek-mating bird. Science, 266, 1030–1032. Macedo, R. H. (1992). Reproductive patterns and social organization of the communal guira cuckoo (Guira guira) in central Brazil. Auk, 109, 786–799. Macedo, R. H., Cariello, M. O., Graves, J., & Schwabl, H. (2004). Reproductive partitioning in communally breeding guira cuckoos, Guira guira. Behavioral Ecology and Sociobiology, 55, 213–222. Macedo, R. H. F., Quinn, J. S., & Lima, M. R. (2005). Reproductive skew and individual strategies: infanticide or cooperation? Acta Ethologica, 8, 92–102. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Magrath, R. D., Johnstone, R. A., & Heinsohn, R. G. (2004). Reproductive skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. McRae, S. B. (1996a). Family values: costs and benefits of communal nesting in the moorhen. Animal Behaviour, 52, 225–245. McRae, S. B. (1996b). Brood parasitism in the moorhen: brief encounters between parasites and hosts and the significance of an evening laying hour. Journal of Avian Biology, 27, 311–320. Millar, C. D., Anthony, I., Lambert, D. M., et al. (1994). Patterns of reproductive success determined by DNA fingerprinting in a communally breeding oceanic bird. Biological Journal of the Linnean Society, 52, 31–48. Mulder, R. A., Dunn, P. O., Cockburn, A., Lazenby-Cohen, K. A., & Howell, M. J. (1994). Helpers liberate female fairy-wrens from constraints on extra-pair mate choice. Proceedings of the Royal Society of London, Series B, 255, 223–229. Mumme, R. L., Koenig, W. D., & Pitelka, F. A. (1983). Reproductive competition in the communal acorn woodpecker: sisters destroy each other’s eggs. Nature, 305, 583–584. Nonacs, P. (2000). Measuring and using skew in the study of social behavior and evolution. American Naturalist, 156, 577–589. Nonacs, P. (2003). Measuring the reliability of skew indices: is there one best index? Animal Behaviour, 65, 615–627. Pamilo, P. & Crozier, R. H. (1996). Reproductive skew simplified. Oikos, 75, 533–535. Petrie, M. & Kempenaers, B. (1998). Extra-pair paternity in birds: explaining variation between species and populations. Trends in Ecology and Evolution, 13, 52–58. Petrie, M., Krupa, A., & Burke, T. (1999). Peacocks lek with relatives even in the absence of social and environmental cues. Nature, 401, 155–157. Pusey, A. & Wolf, M. (1996). Inbreeding avoidance in animals. Trends in Ecology and Evolution, 11, 201–206. Quinn, J. S., Woolfenden, G. E., Fitzpatrick, J. W., & White, B. N. (1999). Multi-locus DNA fingerprinting supports genetic monogamy in Florida scrub-jays. Behavioral Ecology and Sociobiology, 45, 1–10.
Reproductive skew in avian societies Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Jeanne, R. L. (2003). From individual control to majority rule: extending transactional models of reproductive skew in animal societies. Proceedings of the Royal Society of London B, 270, 1041–1045. Reeve, H. K. & Keller, L. (1995). Partitioning of reproduction in mother–daughter versus sibling associations: a test of optimal skew theory. American Naturalist, 145, 119–132. Reeve, H. K. & Keller, L. (1996). Relatedness asymmetry and reproductive sharing in animal societies. American Naturalist, 148, 764–769. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Nonacs, P. (1992). Social contrasts in wasp societies. Nature, 359, 823–825. Reeve, H. K. & Shen, S.-F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Rowley, I. (1978). Communal activities among white-winged choughs Corcorax melanorhamphos. Ibis, 120, 178–197. Shellman-Reeve, J. S. & Reeve, H. K. (2000). Extra-pair paternity as the result of reproductive transactions between paired mates. Proceedings of the Royal Society of London B, 267, 2543–2546. Shields, W. M. (1982). Philopatry, Inbreeding Avoidance, and the Evolution of Sex. Albany, NY: State University of New York Press. Shorey, L., Piertney, S., Stone, J., & Ho¨glund, J. (2000). Fine-scale genetic structure on Manacus manacus leks. Nature, 408, 352–353. Stacey, P. B. (1982). Female promiscuity and male reproductive success in social birds and mammals. American Naturalist, 120, 51–64. Vehrencamp, S. L. (1977). Relative fecundity and parental effort in communally nesting anis, Crotophaga sulcirostris. Science, 197, 403–405. Vehrencamp, S. L. & Quinn, J. S. (2004). Joint laying systems. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge, UK: Cambridge University Press, pp. 177–196. Watts, C. R. & Stokes, A. W. (1971). The social order of turkeys. Scientific American, 224 (6), 112–118. Webster, M. S., Pruett-Jones, S., Westneat, D. F., & Arnold, S. J. (1995). Measuring the effects of pairing success, extra-pair copulations and mate quality on the opportunity for sexual selection. Evolution, 49, 1147–1157.
263
264
W. D. Koenig et al. Whittingham, L. A., Dunne, P. O., & Magrath, R. D. (1997). Relatedness, polyandry and extra-group paternity in the cooperatively-breeding white-browed scrubwren. Behavioral Ecology and Sociobiology, 40, 261–270. Williams, D. A. (2004). Female control of reproductive skew in cooperatively breeding brown jays (Cyanocorax morio). Behavioral Ecology and Sociobiology, 55, 370–380. Williams, D. A., Lawton, M. F., & Lawton, R. O. (1994). Population growth, range expansion, and competition in the cooperatively breeding brown jay, Cyanocorax morio. Animal Behaviour, 48, 309–322. Woolfenden, G. E. & Fitzpatrick, J. W. (1984). The Florida Scrub Jay: Demography of a Cooperative-Breeding Bird. Princeton, NJ: Princeton University Press. Yuan, H.-W., Liu, M., & Shen, S.-F. (2004). Joint nesting in Taiwan yuhinas: a rare passerine case. Candor, 106, 862–872. Zahavi, A. (1990). Arabian babblers: the quest for social status in a cooperative breeder. In P. B. Stacey & W. D. Koenig, eds., Cooperative Breeding in Birds: Long-Term Studies of Ecology and Behavior. Cambridge: Cambridge University Press, pp. 105–130.
10
Reproductive skew in cooperative fish groups: virtue and limitations of alternative modeling approaches michael taborsky
Summary Four types of cooperation between conspecific competitors can be differentiated in fish reproduction: joint defense of a spawning site or territory, joint preparations for spawning, cooperative spawning, and cooperative brood care. Long-lasting associations allowing for different reproductive shares of partners are mainly found in species that cooperate in territory defense and brood care. Here I use skew theory to scrutinize patterns of reproductive participation among related and unrelated group members in cooperatively reproducing cichlids and wrasses. A comparison of five species from which sufficient data are available does not reveal an obvious relationship between average relatedness, or group size, and reproductive skew levels, as predicted from respective skew models. It is remarkable that superficially similar cooperative systems in fish may be based on distinctly different parameter combinations, even in closely related species. I discuss five alternative schemes to understand the patterns of reproductive participation in cooperatively reproducing fish, including kin-selection theory, reciprocity models, manipulation or coercive strategies, models of alternative reproductive tactics, and a dynamic modeling approach. A comparison of approaches suggests that conventional skew models do not account for the complexity of evolutionary mechanisms involved in reproductive skew among members of fish groups. Alternative approaches, such as reciprocity theory or models to explain the coexistence of alternative reproductive tactics, may have greater explanatory Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
265
266
M. Taborsky potential, at least in some cases. However, they have generally not been developed sufficiently to derive predictions allowing for a conclusive test. To understand how decisions of distinctly different types of group members evolve, an approach is needed that takes account of the state dynamics involved, as in fish, conditions constantly change due to their continued growth after maturation. I conclude by outlining widespread fallacies in the application of theory to interpret data, caused by the use of general models that are based on purposely simple assumptions to explain case studies while neglecting the mechanisms involved.
Introduction Reproductive-skew theory has been developed to explain the extent to which breeding is monopolized by dominant individuals in cooperatively breeding animals (Johnstone 2000). Transactional models focus on group stability and the constraints this places on the division of reproduction (Vehrencamp 1979, 1983, Stacey 1982, Keller & Reeve 1994, Clutton-Brock 1998, Reeve et al. 1998, Johnstone & Cant 1999). Dominant individuals with full control over the reproductive share of subordinates trade reproductive concessions against the help they receive (concession model), or, when their control is incomplete, subordinates may trade the threat of eviction against their reproductive participation (restraint model). In contrast, compromise models view the division of reproduction as the outcome of a conflict between group members, where each individual has some ability to enforce its own optimum, ignoring issues of group stability (Cant 1998, Reeve et al. 1998). In a “tug-of-war” dominants and subordinates face a trade-off between maximizing overall group productivity or their own share of reproduction (Reeve et al. 1998). Synthetic models of reproductive skew view these different approaches as special cases of a general underlying theory rather than as alternative paradigms and aim at combining them into a more general model that incorporates the possibility of both transactions and intra-group conflict (Johnstone 2000, Reeve & Shen 2006). Reproductive-skew models have primarily been used to understand the extremely uneven reproductive shares of females in social insects (Reeve & Ratnieks 1993, Langer et al. 2004; see Reeve & Keller 2001 for review), but they have also been applied to the study of reproductive share in male and female vertebrates showing cooperative breeding (Cant 2000, Faulkes & Bennett 2001, Haydock & Koenig 2002, 2003, Rusu & Krackow 2004, Heg et al. 2006; see also Chapter 9). In this chapter I aim to review the suitability of reproductive-skew theory to explain the sharing of reproduction between members of groups in
Cooperative fish groups: alternative modeling approaches fish. I shall first outline how reproductive competitors in fish cooperate in one way or another to illustrate the scope of social patterns that might be better understood by the use of reproductive-skew theory. I shall then relate the available information about these systems to the predictions of various skew models and discuss alternative concepts that might be applied in order to understand the evolutionary mechanisms underlying reproductive sharing in fish societies. Finally, I shall examine the explanatory power of reproductiveskew models for the behavior of fish species where group members cooperate during reproduction.
Cooperative reproductive behavior in fish With more than 25 000 described species, fish are by far the most speciose and diverse class of vertebrates (there are more fish species than all other vertebrate classes combined: Reid & Hall 2003). This is reflected by their unparalleled variation in reproductive patterns (Breder & Rosen 1966, Keenleyside 1979, Blumer 1982, Taborsky 1987, 1994, 1998, 1999, 2001, 2008a, Turner 1993, Berglund 1997, Petersen & Warner 1998, Wisenden 1999, Avise et al. 2002). Taborsky (1994) listed 95 fish species involved in cooperative reproductive relationships or alloparental care, and more examples have been described since (Taborsky 2001, Heg & Bachar 2006). Some of these cases involve interspecific alloparental care, which is most often parasitic rather than cooperative in nature and in any case is not relevant for the question of reproductive skew among group members. If we only include cases that involve cooperation between male reproductive competitors, we can differentiate between four different forms of cooperation: (1) joint defense of spawning site or breeding territory, (2) joint preparations for spawning (nest building, courtship), (3) cooperative spawning, and (4) cooperative brood care (see Table 10.1, listing 38 species of 10 families). Regarding the latter, there are various conditions leading to brood care of foreign offspring that are not of interest for the question of reproductive skew among group members, including egg stealing, adoption of young, nest takeovers, and joint offspring defense by neighboring pairs (see Taborsky 1994 for review). In category (4) I consider only cases of alloparental care between members of long-lasting social groups, a situation resembling “helpers-at-the-nest” in birds or alloparental care in mammals. For our purpose here I only include cases with reproductively mature helpers. Below I shall describe five case examples in detail belonging to the categories (1) joint defense and (4) alloparental care outlined above. The discussion will be confined to cases from which sufficient information is available on cooperative
267
268
M. Taborsky effort of potential competitors and their reproductive participation and success, to be able to investigate the explanatory power of skew theory. Cooperative reproduction of bourgeois and satellite males (“joint defense”)
At least eight species of fish have been studied in which bourgeois males1 defending a spawning site, courting females, and providing brood care are assisted by satellite males that are tolerated in their territories (see Taborsky 1994 for review). Usually these satellites join in defense against reproductive competitors, and in turn they participate in attempts to fertilize eggs deposited by females at the defended nests or spawning sites. In the Mediterranean ocellated wrasse (Symphodus ocellatus), between one and three satellites cooperate with nest owners in defense against small reproductive parasites or sneakers (Figure 10.1), and these male partners stay together for most of the spawning period at a particular nest (Taborsky et al. 1987). Males of this species belong to one of three different life histories, which relates to their growth pattern in the first year of life (Soljan 1930, Alonzo et al. 2000). Satellites do not join in brood care or in defense against egg predators, and they leave the nest after spawning has ceased. Nest males with satellites appear to be much more successful than those without, because females visit such nests more frequently and apparently prefer to spawn there (Taborsky 2001). Reproductive cooperation in polygynandrous breeding groups (“alloparental care”)
In the West African cichlid Pelvicachromis pulcher male helpers are permanently accepted in harem territories, where they help in defense against reproductive competitors and predators of eggs and fry. In turn they participate in fertilizing eggs produced in these groups, but only from a minimum harem size of three females (Martin & Taborsky 1997). This breeding system is characterized by a high degree of reciprocity among males. Harem owners tolerate up to three male helpers, at the expense of the latters’ participation in offspring production, but the dominant males gain by their enormous defense effort. Helpers pay a high price by risky defense that often results in injuries, but dominant helpers obtain a similar reproductive success to monogamous males (Martin & Taborsky 1997). Helpers in the second and third rank positions expend more effort per success than dominant ones, but their competitive ability is apparently lower.
1
The term “bourgeois males” refers to individuals investing in privileged access to mates, by behavioral (e.g. defense, courtship), physiological, or morphological means (e.g. pheromones, secondary sexual characters: Taborsky 1997).
Cooperative fish groups: alternative modeling approaches
Figure 10.1 Spawning in Symphodus ocellatus seen from above. Three fish are in the nest built mainly of brown and red algae by the nest male, which is on the bottom of the nest (below two other fish) just in the movement of spawning (curved body). The female is in the foreground adjacent to the nest male, showing her extended belly and white genital papilla, just about to release eggs. The satellite male is above the spawning nest male and female in an S-bent position with spread opercula, attempting to herd the female into the nest. In the upper left corner of the picture, a small sneaker male is just about to rush to the nest to release sperm (blurred due to its fast approach). (Photo by Michael Taborsky taken at Pointe de la Revellata, Corsica)
Brood-care helpers (“alloparental care”)
Cooperative brood care has been described in at least 17 species of fish, most of which belong to the substrate-breeding cichlids of Lake Tanganyika (Table 10.1; Taborsky 1994; Heg & Bachar 2006 for review). In all species studied in sufficient detail to date, reproductive competition is found among group members, especially among males (Taborsky 1985, Kohler 1998, Dierkes et al. 1999, 2008, Awata et al. 2005, 2006, Heg et al. 2006). Mature brood-care helpers either were born in the territory where they help and share in reproduction, or they have joined the group at a later stage (Taborsky & Limberger 1981, Stiver et al. 2004, Awata et al. 2005, Bergmu¨ller et al. 2005a, Dierkes et al. 2005, Heg & Bachar 2006). Relatedness is variable between dominant breeders and their mature helpers, ranging from zero or very low degrees of relatedness (Awata et al. 2005, Dierkes et al. 2005, Stiver et al. 2005) to relatedness levels considerably exceeding the population mean (Taborsky & Limberger 1981,
269
270
M. Taborsky Kohler 1998). In the species from which we know most, Neolamprologus pulcher (Figure 10.2), helpers of both sexes share in direct and indirect brood care by cleaning and fanning eggs, cleaning larvae, digging out shelters, removing snails, and defending the brood and territory against predators and competitors (Taborsky & Limberger 1981, Taborsky 1984, Taborsky et al. 1986, BalshineEarn et al. 1998, Stiver et al. 2005). After reaching maturity, male helpers share in reproduction by surreptitious release of sperm while the breeders spawn (Taborsky 1985, Dierkes et al. 1999, 2008, Heg et al. 2006). Female helpers may breed in the dominant’s territory as well, or they may split off part of the territory to initiate a harem with the male breeder or pair up with a foreign male (Taborsky 1985, Heg et al. 2006, Heg & Hamilton 2008). Superficially the social system of N. savoryi looks similar, but the groups are usually haremic in this species and subdivided into subgroups, with each female assisted by her own set of helpers, and the male breeder defending the entire area encompassing all subgroups (Heg et al. 2005b). In the congener N. multifasciatus, young do not disperse from their natal territories and they share in territory defense and maintenance, especially digging (Kohler 1998). Usually more adult females than males are present per group (Taborsky 2001). Both mature male and female helpers participate regularly in reproduction (Kohler 1998). In the closely related cichlid Julidochromis ornatus, helpers stay with a breeding pair or harem in roughly 40–95% of breeding groups depending on the population, with male helpers outnumbering female helpers about twofold (Awata et al. 2005, Heg & Bachar 2006). Usually there are only one or two helpers present, but helper numbers may go up to six. These helpers are usually unrelated to the dominant breeders, they share in the defense of broods and in digging, and they participate in offspring production (Awata et al. 2005, 2006, Heg & Bachar 2006). This system is somewhat reminiscent of the polygynandrous groups in Pelvicachromis pulcher (Martin & Taborsky 1997), but in J. ornatus the pattern is more flexible, with monogamous pairs and male and female harems all occurring with and without helpers of both sexes. The group structure and cooperative brood care seems similar in the congener J. marlieri (Taborsky 1994, unpublished data, Sunobe 2000), but there is no information available for this species on reproductive participation of group members.
Use of skew theory to understand patterns of reproductive participation in cooperatively reproducing fish In this section I shall examine how reproductive-skew models can help us understand the reproductive shares of group members in the examples
271
Cooperative fish groups: alternative modeling approaches
M. Taborsky
Table 10.1 Cooperative reproduction between adult male conspecifics in fish Catostomidae Catostomus commersonii
joint spawning
Reighard 1920
Hypentelium nigricans
joint spawning
Raney & Lachner 1946
Moxostoma aureolum
joint spawning
Reighard 1920
M. duquesnei
joint spawning
Bowman 1970, Kwak & Skelly 1992
M. carinatum
joint courtship & spawning
Hackney et al. 1968
M. macrolepidotum
joint spawning
Jenkins 1970, Burr & Morris 1977
M. erythrurum
joint spawning
Jenkins 1970, Kwak & Skelly 1992
M. valenciennesi
joint spawning
Jenkins 1970, Jenkins & Jenkins
Erimyzon oblongus
joint spawning
Page & Johnston 1990
1980 Cyprinidae Nocomis micropogon
joint nest building
Reighard 1943
Notropis leptocephalus
joint nest building
Wallin 1989
joint defense
Barlow 1961
Cyprinodontidae Cyprinodon macularis Cottidae Hemilepidotus hemilepidotus joint brood defense
DeMartini & Patten 1979
Osphronemidae a
Betta brownorum
joint defense
B. persephone
joint defensea
Witte & Schmidt 1992 Witte & Schmidt 1992
Blenniidae Parablennius sanguinolentus joint defense
Santos 1985, Santos & Almada 1988, Oliveira et al. 2002
Cichlidae Pelvicachromis pulcher
joint brood defense
Martin & Taborsky 1997
Neolamprologus pulcher b
cooperative breeding
Taborsky & Limberger 1981, Taborsky 1984, 1985, Taborsky & Grantner 1998, Balshine et al. 2001, Brouwer et al. 2005, Dierkes et al. 2005, Heg et al. 2005a, Stiver et al. 2005
N. gracilis
cooperative breeding
Woodland 2002, H. H. Bu¨scher
N. helianthus
cooperative breeding
H. H. Bu¨scher pers. comm. in Heg
N. marunguensis
cooperative breeding
H. H. Bu¨scher pers. comm. in Heg
N. olivaceous
cooperative breeding
H. H. Bu¨scher pers. comm. in Heg
N. falcicula
cooperative breeding
& Bachar 2006 own obs. at Kigoma, Tanzania
N. splendens
cooperative breeding
H. H. Bu¨scher pers. comm. in Heg
N. savoryi
cooperative breeding
Taborsky & Limberger 1981;
N. multifasciatus
cooperative breeding
Kohler 1998
N. similis
cooperative breeding
Bu¨scher 1992 and pers. comm. in
Julidochromis marlieri
cooperative breeding
Taborsky & Limberger 1981,
J. regani
cooperative breedinga
Taborsky & Limberger 1981
J. ornatus
cooperative breeding
Taborsky & Limberger 1981, Awata
pers. comm. in Heg & Bachar 2006 & Bachar 2006 & Bachar 2006
& Bachar 2006 Kondo 1986; Heg et al. 2005b
Heg & Bachar 2006 Taborsky 1994, Yamagishi & Kohda 1996, Sunobe 2000
et al. 2005, 2006, Heg & Bachar 2006 J. transcriptus Chalinochromis brichardi
cooperative breeding
Kuwamura 1997
cooperative breeding
Awata et al. 2005
joint defense
Fricke 1979
Symphodus ocellatus
joint defense
Fiedler 1964, Lejeune 1985,
S. roissali
joint defense
Lejeune 1985
Pomacentridae Amphiprion akallopisos Labridae Taborsky et al. 1987 S. tinca
joint defense
Lejeune 1985
Halichoeres maculipinna
joint defense
Thresher 1979
joint courtship
Fahy 1954
Percidae Etheostoma blennioides
Examples are included only if adult males cooperate in some reproductive activity. “Joint defense” refers to defense of a spawning site or breeding territory; if broods are jointly defended, this is called “joint brood defense.” Cases where neighboring breeding pairs defend their young jointly (see Taborsky 1994 for review and examples) are not included, because it is unlikely that these associations persist during the spawning phase. In cooperatively breeding species, both adult male and female brood-care helpers may be tolerated by dominant breeders. a
Only observations from captivity are available.
b
Neolamprologus pulcher and N. brichardi have been regarded as two species (Poll 1974), but
recent molecular, morphological, and ecological evidence suggests that this is not justified (Duftner et al. 2007, E. Skubic et al. unpublished).
272
Cooperative fish groups: alternative modeling approaches
Figure 10.2 In front of a breeding shelter, a dominant breeder of Neolamprologus pulcher (left) approaches a subordinate helper, which responds with a submissive display (tail vibration, with folded fins). Below these two adult fish there is a small young (only front half visible). (Photo taken by Michael Taborsky at Kasakalawe Point, Zambia)
described above. Table 10.2 provides basic information about the reproductive biology and behavior of these example species. Reproductive skew between bourgeois and satellite males in Symphodus ocellatus
Nest males and satellite males of this species defend the nest jointly against opportunistic reproductive competitors (sneakers). Nest males are always bigger than satellite males and always dominant. They seem to be in control of group membership of their satellites, which is mediated by an elaborate exchange of display behaviors between the two parties (Taborsky et al. 1987). The cooperators in S. ocellatus are not related to each other above the population mean (Basieux 2007), because they have been subject to a planktonic dispersal phase at their larval stage and, in addition, they often belong to different age classes (Alonzo et al. 2000). The reproductive skew is high between the cooperators: as revealed by DNA microsatellite analyses, nest males sired 47% of offspring produced in their nests (median; range 2–72%; n ¼ 15 nests) when satellites were present, and satellites sired 0.63% (median; range 0–22%; n ¼ 15 nests), which is only 1.34% of the nest male’s share (Figure 10.3); all other young were sired by non-cooperative, parasitic males (Basieux 2007), which visit these nests in great numbers (Taborsky et al. 1987,
273
274
M. Taborsky Table 10.2 Basic information on the reproductive biology and behavior of the five species presented as examples in the text Symphodus
Pelvicachro-
Neolamprologus N.
Trait
ocellatus
mis pulcher
pulcher
multifasciatus ornatus
Julidochromis
Brood care of
male
biparental
biparental
biparental
male
male, female & male, female male, female
biparental
breeders Sex of subordinates male
immature
&
&
immature
immature
Brood care of helpers: defence of brood no
yes
yes
yes
yes
territory
no
yes
yes
yes
no
maintenance direct brood carea no Subordinates raise
no
yes
no
yes
no
yesb
yes
?
yesb
low
high
medium
medium
high
breeders’ success Reprod. share of male subordinates Control of group
male breeder male breeder pair members pair
membership
(þ/-)
Classification of
fixed and
ARTsc
members sequential
sequential
sequential
plasticd
(reversible
(fixed
(fixed
(fixed
sequence)
sequence)
sequence)
pair members sequentiale
sequence) Information compiled from Taborsky & Limberger 1981, Taborsky 1984, 1985, 1994, Martin & Taborsky 1997, Kohler 1998, Dierkes et al. 1999, Alonzo et al. 2000, Schradin & Lamprecht 2000, Awata et al. 2005, Brouwer et al. 2005, Heg & Bachar 2006, Basieux 2007. a
In fish this includes fanning of eggs, removing infected eggs, and cleaning of eggs, larvae,
and free-swimming fry with the mouth. b
Only correlative evidence available.
c
Alternative reproductive tactics can be fixed for life or plastic; plastic tactics may be either
performed simultaneously depending on conditions, or sequentially with a fixed or reversible sequence (see Figure 1 in Taborsky et al. 2008). d
In S. ocellatus, males do not switch between bourgeois and subordinate tactics within a
season. Males that are already large in their first season may reproduce as satellites and turn into bourgeois nest males in their second season. Males that are small in their first season reproduce as sneakers and turn into satellites in their second season. They never reproduce as bourgeois nest males (Alonzo et al. 2000). e
A sequential tactic with fixed sequence seems most likely in J. ornatus, but some individuals
might not pass through a helper phase at all (as derived from size distributions: Awata et al. 2005).
Cooperative fish groups: alternative modeling approaches
x adult subordinates/group
2.5
2.29
females
2.16
2.0 1.60 1.5
1.31
1.27 1.09
1.0
0.86 0.63
0.5
Species: N helpers:
0.4 Degree of relatedness (r)
males
S.o. 104
P.p. 50
N.p. 31
N.m. 32
J.o. 35
0.375
to male breeder to female breeder
0.3 0.205 0.2
0.158
0.1
0.043 0.029
0.033 (0) (0)
(0) (0)
S.o. a
P.p. b
0 Species: N helpers:
N.p. 99
77.1
80
N.m. 12+22c
78.8
Paternity of young in groups [%]
67.5 60
J.o. 35
sired by dom. sired by subdom.
50.4
47 40
34.7
32.5 21.2
20 10.3 0.63
Species: N helpers:
S.o. 15
P.p. 10
N.p. 12+7d
N.m. 10
J.o. 22
Figure 10.3 Number of adult subordinates per group (upper panel), relatedness between subordinate and dominant cooperators (middle panel), and levels of reproductive skew among male group members (lower panel), as determined in five species of fish with cooperative reproduction (Symphodus ocellatus, Pelvicachromis pulcher, Neolamprologus pulcher, N. multifasciatus, Julidochromis ornatus). See text for details and references. Exact values are given above each column or point, and sample sizes are provided below the abscissa. Paternity estimates were obtained
275
276
M. Taborsky Alonzo & Warner 1999). The paternity of individual sneakers is unknown, but when comparing the reproductive behavior of satellites with sneakers, by being tolerated the former stay much closer to the nest than the latter (median distance 35 cm vs. 75 cm), which provides a better position for fertilization attempts, and satellites also ejaculate more often than individual sneakers (Taborsky et al. 1987). In this cooperative system, satellite males apparently benefit from being tolerated near the nest of a dominant male, and they pay for this concession by helping to defend the nest against sneakers. Tolerance of satellites by nest males may have one of three ultimate causes: (1) defense by satellite males serves only their own reproductive benefit in the competition with sneakers: nest males have limited control over the presence of satellites and are unable Figure 10.3 (Continued) from genetic color markers (P. pulcher; here only harems with three satellite males were included in the paternity analysis) or with the help of microsatellite DNA analyses (all other species). Fractions of paternity do not necessarily add up to 100% if other males (usually non-group members) sired offspring as well. In N. multifasciatus, I included only those offspring that could be assigned to at least one group male, as the resolution of the microsatellite analysis was limited because only three loci were investigated. It seems unlikely that in this species non-group males would have many opportunities to sire offspring as well (see Kohler 1998, p. 70). a
In S. ocellatus, satellite males are unrelated to bourgeois nest males because of
planktonic dispersal of larvae; also, satellite males often belong to a different age cohort than nest males, and they switch frequently between different nest males during a reproductive season (Taborsky et al. 1987). b
Harem males and satellite males in P. pulcher belong to different color morphs; in
the field, satellites are recruited from aggregations, and under seminatural conditions they first try to establish an independent territory, and join a harem only if unsuccessful with breeding independently. Often satellites switch between harems in consecutive seasons (Martin & Taborsky 1997). For these reasons it seems safe to conclude that satellites are not related to harem males above the population mean. c
In N. multifasciatus no data are available on the relatedness between helpers and the
female breeder. Instead, separate estimates exist for the relatedness between male breeder and male helpers (upper value, r ¼ 0.375, n ¼ 12), and between male breeder and female helpers (lower value, r ¼ 0.205, N ¼ 22; Kohler 1998). d
The number of offspring sired by male breeders in N. pulcher refers to a field sample
(n ¼ 12 groups; Dierkes et al. 2008), whereas the number of offspring sired by male helpers has been derived from controlled lab experiments (n ¼ 7 groups with one or two male helpers each, Dierkes et al. 1999), as the “extra-pair young” in the field sample could not be unequivocally assigned to a certain type of male (see Dierkes et al. 2008).
Cooperative fish groups: alternative modeling approaches to keep them at bay (Yanagisawa 1987); (2) nest males benefit by the satellites’ defense effort, either directly through a reduced sneaking rate or indirectly, e.g. by making the nest more attractive to females (Taborsky 2001, Oliveira et al. 2002); (3) expulsion of satellite males is more costly to nest males than tolerating them: despite some net costs, nest males do better by not striving for complete expulsion but instead conceding some reproductive share to them to avoid even higher costs (Kodric-Brown 1977). For the first possibility, skew theory would not be relevant, because no behavioral decisions of the dominant partner are involved concerning the reproductive share of satellites and their presence at the nest. In a functional sense, the situation would be purely competitive (or parasitic) and not cooperative. The second possibility corresponds to a situation described either by concession or by compromise models, whereas the third one is adequately reflected either by restraint or by compromise models (e.g. a “tug-of-war”). To decide between these possibilities, we need to consider the degree of control of the dominant male, the benefits of cooperation, and the level of skew. The first mentioned possibility (nest males have no control over the presence of satellites) seems highly unlikely from the behavioral information we have. Nest males engage in elaborate display sequences with satellite males and make a clear difference between satellite males and sneakers regarding tolerance at the nest, which should not be the case if satellites were pure parasites like sneakers (Taborsky et al. 1987). Application of a concession model to the interaction between nest males and satellites would imply full control by the former of the latter’s reproduction, and a net benefit for nest males from the presence of satellite males. Full control over reproductive participation of competitors is hardly possible when fertilization is external (Taborsky 1994, 1998, 2008a). This clearly holds also for the relationship between nest males and satellite males in S. ocellatus (Taborsky et al. 1987, Alonzo & Warner 2000). The question of whether nest males benefit from the presence of satellites, which would allow us to decide between possibilities (2) and (3), has been investigated experimentally by removing satellites for a limited period of time to test for immediate effects of satellites on parasitic spawning (Taborsky 2001; M. Taborsky, P. Wirtz, & B. Taborsky, unpublished data), or for the entire spawning period at a nest to test for indirect long-term effects of the presence of satellites (Basieux 2007). After satellite-male removal nest males immediately increased overt attacks against intruders (i.e. mainly defense against sneakers), and they increased courtship behavior. In the long run, nest males were exposed to more parasitic fertilization attempts by other males (satellites and sneakers combined) when satellites had been removed. All this suggests a beneficial effect of satellite males to the nest males with which they cooperate.
277
278
M. Taborsky However, contrary to expectation, satellite removal had no negative long-term effect on the spawning rates of nest males, and the proportion of offspring sired by the bourgeois male was not significantly different between the experimental periods with and without satellites. This contrasts with the unmanipulated situation, where nests with satellites received more female visits and more spawnings and were more likely to be tended until larvae hatched than nests without satellites (see Taborsky 2001). This female preference for nests with satelllites may be due to a positive feedback loop between female spawning rate and the likelihood that broods will be cared for by the nest owner until hatching. If bourgeois males receive insufficient numbers of clutches, satellite males will not turn up and the nest males abandon their nest prematurely and start anew (M. Taborsky, P. Wirtz, & B. Taborsky, unpublished data). Hence females may use satellites as an indicator of a high probability that their clutches will not be wasted because they will be tended until hatching (Taborsky 2001). A similar pattern has been described in other Symphodus species (Soljan 1931, Fiedler 1964, Lejeune 1985) and in the Azorean rock-pool blenny Parablennius sanguinolentus (Santos & Almada 1988, Oliveira et al. 2002). In conclusion, the interaction between nest males and satellites does seem to be cooperative in the functional sense (West et al. 2007a) as described in possibility (2) above, because both parties benefit, but the benefit to nest males seems to be subtle, and at our present state of knowledge hypothesis (3) may still hold. The reproductive share of satellites is probably not due to unsolicited concessions of nest males, because the latter are not able to fully control the reproductive participation of the former. Hence, the interaction between nest males and satellites is probably best reflected by a tug-of-war model, and the very high skew between these two types of males despite their lack of relatedness is compatible with its theoretical expectations; it probably reflects the great differences in reproductive competitiveness between bourgeois and satellite males. Reproductive skew between males of polygynandrous Pelvicachromis pulcher
In this polymorphic species, males have three different options to reproduce, depending on their genotype and relative size. There is a genetically determined color polymorphism in the male sex. Males of the red color morph may either form pairs or harems, whereas males of the yellow morph may decide between forming a pair or joining a harem to help its owner in brood care and defense (Martin & Taborsky 1997). In cooperative groups, harem owners are always bigger than their satellites, and a dominance hierarchy strictly following the relative sizes of males is established between
Cooperative fish groups: alternative modeling approaches harem owners and satellites (up to three per territory), at least a week before spawning starts (Martin & Taborsky 1997). To my knowledge, neither ecological constraints nor group size or within-group relatedness levels have yet been experimentally manipulated in this species. However, for a comparison with other species (see below) it is worth considering reproductive skew among group members in connection with their relatedness. Groups in this species are formed from independent individuals, i.e. not from descendants of the territory where they breed and raise young. Therefore, even though relatedness between satellites and breeders has not been tested by genetic analysis, it seems safe to assume that relatedness among these cooperating males is close to the population mean, similar to relatedness in S. ocellatus. The level of reproductive skew among group members is intermediate, with more than two-thirds of young produced on average by the dominant harem owner: depending on group size, harem males sired between 67% and 77% of offspring (Figure 10.3). When three satellites were present and their offspring production was separated by rank, the dominant satellite males sired roughly 20% of the young produced in the harem, the second satellite nearly 10% and the third satellite slightly more than 3%. During a season, dominant satellites produced almost as many offspring as pair males did, but these young had only about half the survival probability compared with offspring of pair males or harem males (Martin & Taborsky 1997). For this case, where group members are not related and group membership is apparently largely controlled by the dominant breeder, the reproductive skew predicted by different models depends primarily on (1) the competitive ability of the subordinate relative to the dominant, which is probably rather low in P. pulcher (0.5); (2) the change in productivity of the group due to the subordinate’s participation in brood care, which is hard to determine without experimental manipulation in this species because harems without satellites hardly occurred (Martin & Taborsky 1997); and (3) the alternative possibilities for subordinates to produce offspring if they disperse (these clearly differ for satellites in the first, second, and third rank positions in the hierarchy because of size differences, but for the dominant satellites they can be assumed to be good if they manage to find a partner (see above); this probability is unknown for the natural situation, however). If we assume a 50–100% increase of reproductive output due to the presence of a satellite, which seems plausible, and if we derive from our data that the alternative possibility of a satellite producing offspring is about 0.5 if they disperse, compromise and tug-of-war models as well as the synthetic approach proposed by Johnstone (2000) would all predict levels of a subordinate’s reproductive share in the range of 0.25–0.5; this compares quite well to the share of dominant satellite males in our study,
279
280
M. Taborsky but the shares of second- and third-rank males were clearly lower (Martin & Taborsky 1997). However, the latter also have lower competitive abilities and worse alternative options, and their effect on productivity of the group may also be lower than the estimates given above, which were based on dominant satellites. Reproductive skew among group members in cooperative breeders Neolamprologus pulcher
In the only experimental study of reproductive skew to date in a cooperatively breeding vertebrate, Heg et al. (2006) tested for the combined effects of ecological constraints, helper size, and helper sex on reproductive skew among group members of N. pulcher. We created 32 breeding groups with one dominant pair and one large and one small unrelated helper of opposite sex each, which had access to a dispersal area with (low constraint) or without (high constraint) suitable breeding substrate (see Bergmu¨ller et al. 2005b for details). Dominant breeders in this species (and probably in any other fish species with size-structured groups of cooperators) have full control over group membership of subordinate group members, but only limited control over their share in reproduction, particularly in the male sex (Taborsky 1985, Dierkes et al. 1999, 2008). In the most basic versions of transactional and compromise reproductive-skew models, predictions differ about the relative skew in conjunction with high or low ecological constraint levels. While “restraint” and “tug-of-war” models do not predict a difference in skew between different levels of ecological constraints for independent breeding when dominant and subordinate group members are unrelated, “concession” and “synthetic” models predict that when alternative breeding options are available (i.e. low constraints), reproductive skew should be reduced, because dominants should provide incentives for helpers to stay when they have opportunities to breed independently elsewhere. Heg et al. (2006) found no support for concession and synthetic models, as the number of young sired by male helpers (4.8% overall in 42 broods containing a total of 1185 offspring) did not differ significantly between the two experimental treatment levels (low constraint: 5.22%; high constraint: 2.28%). Among females, only one helper participated in reproduction, which occurred in the high-constraint situation. Regarding the influence of size and sex of helpers on their participation in reproduction, concession and restraint models in their general form make no clear predictions, whereas tug-of-war models predict that (1) larger individuals are more likely to participate in reproduction, because they are more competitive, and (2) male subordinates should be more likely to reproduce than female subordinates, because in fish reproductive parasitism can be controlled
Cooperative fish groups: alternative modeling approaches by dominants much more easily in the female than in the male sex (see Taborsky 1994, 2008a). Heg et al. (2006) indeed found that larger helpers were more likely to participate in reproduction than small helpers, and males were more likely to participate than females, which is in accordance with a tug-ofwar scenario. A result of this experiment that is hard to reconcile with any reproductive skew model is the very high overall skew level. In experiments with groups consisting of a pair and one (six replicates) or two (one replicate) unrelated male helpers but without dispersal opportunities, Dierkes et al. (1999) found that male helpers sired on average 10.3% of offspring produced, and in the unmanipulated field situation male breeders sired on average 77.1% of offspring produced in their territories (Dierkes et al. 2008; Figure 10.3). In this field sample, the sires of offspring not produced by the dominant male breeders could not be unequivocally determined, but it seems likely that male helpers were also involved. These two studies suggest that the reproductive skew in N. pulcher might be a bit lower under natural conditions than suggested by the experimental results of Heg et al. (2006), but still the vast majority of offspring are sired by the dominant male breeder. In females, reproductive skew is even higher: in the field, the dominant female breeder sired all young in 11 of 12 groups, and in one group neither male nor female breeder sired any of the young collected in that territory, which suggests that they had taken over the territory very recently (i.e. after the fry had been produced: Dierkes et al. 2008). In controlled laboratory conditions female helpers were found to reproduce in the dominant breeders’ territory dependent on the availability of an own “private” breeding shelter (Heg & Hamilton 2008). Interestingly, when female breeders were assisted by a large male helper, they produced larger clutches when this helper had an independent breeding opportunity than when he had none (difference 25%), whereas when they had a large female helper, they produced smaller clutches when their helper could breed independently (difference 20%: Heg et al. 2006). The first result might reflect a tendency of female breeders to provide incentives to their male helpers to stay instead of leaving to breed independently (Cant & Reeve 2002, Hamilton & Heg 2007), because when more eggs are laid the number of young they could sire at home would increase. At the same time, the larger clutch size might alleviate reproductive conflict between the male breeder and the male helper, as predicted by tug-of-war models allowing for compensatory thirdparty effects (Hamilton & Heg 2007). The second result might reflect a response of the dominant breeder female to the higher level of reproductive competition with her large female helper if the latter cannot breed independently outside the home territory. The only reproductive participation of a female helper took place in the situation without alternative breeding options, which
281
282
M. Taborsky might hint at a higher level of reproductive competition among female group members when subordinates cannot breed independently. By producing greater clutches when female helpers cannot breed elsewhere, dominant females might deter spawning between the dominant breeder males and female helpers in the home territory, as this would be less rewarding. None of the existing skew models make predictions about strategic or compensatory clutch-size variation in response to reproductive competition between females in connection with ecological constraints or alternative breeding options. Regarding relatedness, concession models predict a positive relationship between the relatedness among competitors and reproductive skew, whereas restraint models predict the opposite (but see Hamilton & Heg 2007). Tug-ofwar and synthetic models make no clear predictions about this relationship, as their outcome depends to a large extent on particular assumptions, for example about the competitive ability of group members. There is as yet no experimental study of N. pulcher varying relatedness levels within groups to check for effects on reproductive skew. However, we know average relatedness levels among group members and reproductive skew from the field (Dierkes et al. 2005, 2008, Stiver et al. 2005). Relatedness between mature helpers of either sex and male breeders is no different from the population mean (between 0 and 0.1), whereas relatedness to the female breeders is significantly higher (0.13–0.21, no difference between male and female helpers: Dierkes et al. 2005). This difference in relatedness to male and female breeders reflects the fact that the replacement rate of breeder males is twice as high as that of females; helpers usually stay when breeders are replaced (Taborsky 1984, Balshine-Earn et al. 1998). The resulting relatedness differences should generate a greater conflict potential between male than between female group members. Reproductive skew was high among males (with 77.1% of offspring sired by the male breeder), and absolute among females (with 100% of offspring sired by the female breeder: Dierkes et al. 2008). These sex differences are in accord with concession rather than with restraint models, as the lack of relatedness among males might cause male breeders to concede some reproductive share to subordinate male helpers to make them stay, whereas among related females such concession might not be required. Neolamprologus multifasciatus
In this snail-brooding cichlid, groups are long-lasting and rather small, with one or two adult male subordinates present at c. 38% of pairs and between one and four adult female subordinates present at c. 64% of pairs (Kohler 1998). This sex difference is significant, i.e. among reproductively mature helpers there are on average more females than males (Taborsky 2001). However,
Cooperative fish groups: alternative modeling approaches among adults only the ranges of male group members overlap within the family territory, whereas the female group members defend exclusive subterritories against each other (Schradin & Lamprecht 2002). In addition to adults, there are subadult and juvenile conspecifics permanently present in the majority of breeding groups. These families consist usually of highly related individuals. Nine of 12 genotyped male helpers (75%) were probably sons (or brothers) of the current dominant male breeder, whereas this applied to nine of 22 female helpers (41%; Kohler 1998). Relatedness between female breeders and adult helpers was not tested. Sex-specific relatedness differences among group members are connected to a greater tendency of females to change between groups, which apparently results from tolerance differences of the dominant breeders (Schradin & Lamprecht 2000). Reproductive skew is apparently higher in males than in females. Dominant breeder males sired 78.8% of offspring, and 21.2% were sired by the largest male helper in the group (Figure 10.3; derived from data provided in Kohler 1998, Table 3.5; n ¼10 groups with adult male helpers and 66 offspring that could be assigned to a group male; 20 offspring for which paternity of neither the breeder male nor the male helper could be excluded were assigned to both types of males in equal shares). In 10 of 18 groups of which two or more consecutive broods were analyzed, the young were produced by two (n ¼ 9) or three (n ¼1) females each. If we consider the relatedness differences between male and female helpers to the male breeder, the higher relatedness of male helpers coincides with a greater reproductive skew, which is compatible with concession models. However, comparing parentage of offspring in groups with female helpers that were related to the male breeder (i.e. probably daughters, n ¼6) with those where they were not (n ¼6), there is no apparent difference: in all six cases with related female helpers these participated in reproduction, and in four of six cases with unrelated female helpers the latter did so as well (p ¼ 0.45; Fisher exact test, two-tailed; compiled from data in Kohler 1998). Experimental evidence in the field revealed that ecological constraints (availability of suitable breeding shelters, predation pressure) are extremely high, which is probably mainly responsible for the delayed dispersal of individuals from their natal territories (Kohler 1998, Schradin & Lamprecht 2000, 2002). As only one population has been investigated in this respect and the ecological conditions were obviously rather uniform there, no analysis of a variation in reproductive skew in relation to a variation of ecological conditions is possible. Group-size variation was too low in the study population to allow for a test of a relationship with reproductive skew.
283
284
M. Taborsky Julidochromis ornatus
Mature helpers in this Lake Tanganyika cichlid are mostly unrelated to the breeders they assist (only 14% of helpers were related to either the male or the female breeder: Awata et al. 2005). Breeders are always bigger than their helpers, but attack them rarely (the frequency of aggressive behavior of male breeders towards helpers was similar to that towards their mate, occurring about once in 100 minutes: Awata et al. 2005). No expulsions of helpers have been witnessed in the field, but on the contrary helpers are sometimes attacked by their breeders when they roam too far from the breeding shelter, which causes them to return immediately (Heg & Bachar 2006). There was no reproductive skew in either male or female group members: male helpers sired on average 41% of young in their groups, apparently either by simultaneous spawning with the breeders or with the dominant female alone, and female helpers produced on average 56% of young. In pairs with a male helper the numbers of young sired by male breeders and male helpers were not significantly different (Awata et al. 2005; note that the numbers given in Figure 10.3 are slightly different, because there offspring that could not be assigned to either group male are also included in the calculation of paternity proportions). Male helpers siring offspring were no bigger than those not doing so, which suggests that the reproductive participation of helpers does not depend on their relative competitiveness. The numbers of young sired by male breeders with one male helper (the most common cooperative breeding unit), by such helpers, and by monogamous males without helpers were all very similar, which may be partly due to a tendency for more young to be produced in groups with helpers (Awata et al. 2005; but see Heg & Bachar 2006, who did not find such tendency in their population). There is evidence that larger young switch between breeding groups and helpers immigrate into breeders’ territories (Awata et al. 2005, Heg & Bachar 2006). Hence in this species helpers are either young that remained in their natal territories or immigrants from elsewhere (see Figure 10.3 for relatedness estimates). In J. ornatus no experimental manipulations of ecological constraints, relatedness levels, or group size have been performed, and natural variation in relatedness levels and group sizes was too low to draw any conclusions about their relationship with reproductive skew (see above). However, it is interesting to consider the apparent lack of reproductive skew in this species in comparison to other species, in particular to other cooperatively breeding cichlids. It does not make sense to compare levels of ecological constraints when reliable data from the natural situation are missing in most cases. Regarding relatedness, J. ornatus helpers are apparently as unrelated to male breeders as is the case in N. pulcher, but here they are similarly unrelated to the
Cooperative fish groups: alternative modeling approaches female breeders as well, which is different from the situation in N. pulcher (Awata et al. 2005, Dierkes et al. 2005). In contrast, N. multifasciatus groups are composed of highly related individuals. Reproduction is more egalitarian in J. ornatus than in the two cooperative breeding systems known that are also largely combined of unrelated fish (P. pulcher and N. pulcher), but also more egalitarian than in the cichlid that shows the highest relatedness levels among cooperatively breeding fish (N. multifasciatus). When comparing the five cooperatively breeding species described here in detail, from the limited set of data available there is no obvious relationship between average relatedness and reproductive-skew levels. This might partly result from “third-party” influences on reproductive conflict, e.g. compensatory clutch size adjustments of females to alleviate male–male conflict within the group (Hamilton & Heg 2007), which in turn might go along with corresponding egg-size adjustments (Taborsky et al. 2007). Regarding group size, there is no obvious relationship with reproductive skew either when comparing the species with smaller (S. ocellatus, P. pulcher, J. ornatus) and larger (N. multifasciatus, N. pulcher) group sizes with each other. The highest and lowest levels of skew are both found in the species with the smallest group sizes, whereas the two species with larger groups showed rather intermediate levels of reproductive skew (see Figure 10.3; these relationships hold also when immature group members are included). Alternative approaches towards understanding patterns of reproductive participation in cooperatively reproducing fish In this section I shall discuss alternative approaches that may be used instead of skew theory to model reproductive competition in cooperative groups, including kin-selection theory, reciprocity models, game-theory models developed for alternative reproductive tactics, and varied approaches used to model particular case studies. Kin-selection theory
The theory of kin selection was developed originally to explain altruistic behavior between relatives, the most extreme of which is complete abandonment of reproduction for the sake of raising relatives (Hamilton 1964), i.e. complete reproductive skew. Hamilton’s rule states that the ratio between fitness costs (c) to the donor (e.g. helper) and fitness benefits (b) to the receiver (e.g. breeder or its offspring) should be lower than the degree of relatedness (r) between the two for altruistic behavior to be favored by natural selection: c=b < r
285
286
M. Taborsky This means that costs to donors, benefits to recipients, and relatedness between them need to be considered (West et al. 2007b). Kin-selection theory was initially used to explain reproductive investment patterns mainly in social insects, particularly where haplodiploidy creates relatedness asymmetries (Trivers & Hare 1976, Queller & Strassmann 1998, Chapuisat & Keller 1999). However, kin-selection models can be applied also to the study of sociality and reproductive skew in other animals, because systematic relatedness asymmetries such as those caused by haplodiploidy are of course no precondition (Cant & Field 2001, 2005, Griffin & West 2003). Even though seemingly altruistic traits may have direct fitness benefits as well (Griffin & West 2002) and competition between relatives may limit the importance of kin selection in generating reproductive altruism (Queller 1994, West et al. 2002), kin-selection theory has been widely viewed as the basic concept to explain helping and reproductive skew (Hamilton 1964, Michod 1982, Queller & Strassmann 1998, Griffin & West 2003, Wenseleers & Ratnieks 2006a). The effects of relatedness on expected levels of reproductive skew are explicitly modeled in some of the reproductive-skew theory (particularly transaction models; see above), but there is also much potential to apply kin-selection theory independently of conventional skew models (e.g. Cant & Field 2005). Reciprocity models
Since Trivers (1971) focused the attention of ethologists and evolutionary biologists on reciprocity as a potential mechanism causing cooperative behavior, a prolific literature has emerged on the theory of reciprocal cooperation (for review see Killingback & Doebeli 2002, Nowak 2006), and empiricists have attempted to clarify whether cooperation among animals may be based on reciprocity mechanisms (for review see Dugatkin 1997, Sachs et al. 2004). How can reciprocity theory possibly explain the relationship between cooperative behavior and reproductive competition in members of social groups? A crucial element of such a reciprocal relationship is that different commodities are traded against each other, such as “if you scratch my back, I share my treat with you.” Indeed, trading has been observed in cooperative breeders, where helpers pay with costly helping behavior for being allowed to stay in a safe territory (Taborsky 1985, Balshine-Earn et al. 1998, Bergmu¨ller & Taborsky 2005), depending on ecological constraints that make staying beneficial, such as predation risk and access to shelters (Taborsky 1984, Heg et al. 2004, Bergmu¨ller et al. 2005b; see Bergmu¨ller et al. 2007 for a discussion of payto-stay as a reciprocity mechanism). The helpers gain by increased survival probabilities (Taborsky 1984, Heg et al. 2004), and they may gain by territory inheritance (Balshine-Earn et al. 1998) and participation in reproduction
Cooperative fish groups: alternative modeling approaches (Dierkes et al. 1999, Heg et al. 2006). However, abstaining from reproduction can also be part of such trading: subordinate group members do not compete for reproduction with dominants in exchange for access to resources and protection. As the net fitness effect of a helper under pay-to-stay alone is expected to be negative to dominant breeders (Hamilton & Taborsky 2005), reciprocity theory is very suitable for modeling reproductive skew in cooperative breeders. One example that might illustrate this is the relationship between nest males and satellite males in Symphodus ocellatus, where satellites pay to stay with submissive behavior and nest defense against sneakers, but nest males do not sire more young when satellites are present, due to the reproductive participation of satellites (see above). Another example is given by the cooperatively breeding cichlid Neolamprologus pulcher (Taborsky 1984, Balshine-Earn et al. 1998, Bergmu¨ller & Taborsky 2005, Bergmu¨ller et al. 2005b), where reproductive participation of helpers affects the tolerance threshold of breeders (Taborsky 1985, Dierkes et al. 1999, Skubic et al. 2004). Reciprocity theory has explanatory potential for other helper systems in fish (and other taxa) as well, but presently suitable data and respective model applications are missing. Enforced cooperation
Rewarding cooperators or punishing non-cooperators can enforce cooperation in a group (Boyd & Richerson 1992, Clutton-Brock & Parker 1995, Tebbich et al. 1996, Frank 2003, Wenseleers & Ratnieks 2006b, Young et al. 2006, West et al. 2007b). The propensity to cooperate can be strongly affected by the suppression of competition within groups. Mutual policing and enforcement of reproductive fairness may be responsible for the evolution of increasing social complexity (Leigh 1971, Frank 1995). Repression of competition within groups is believed to often enhance group success in competition against other groups (Frank 2003), which is somewhat reminiscent of group augmentation benefits (Kokko et al. 2001). This may be relevant in particular when leaving the group is no real option, as is the case in many highly social insects, where survival and reproduction outside of the group is highly unlikely. Under these conditions, punishment has been shown to effectively suppress reproduction of workers (Ratnieks 1988, Ratnieks & Visscher 1989, Wenseleers & Ratnieks 2006b), which is apparently affected also by group size (Bourke 1999). Even though such situations have been considered in compromise models (e.g. the tug-of-war model: Reeve et al. 1998), approaches different from conventional reproductive-skew models have been developed to understand the importance of enforcement for the evolution of cooperation and reproductive abstinence (Boyd & Richerson 1992, Frank 1995, 2003,
287
288
M. Taborsky Wenseleers et al. 2004). It is as yet unclear whether applying such an approach would be beneficial also for an analysis of cooperative reproduction in certain fish species or populations, particularly where ecological constraints such as heavy predator pressure may render dispersal unrewarding (e.g. Neolamprologus pulcher: Taborsky & Limberger 1981, Taborsky 1984, Heg et al. 2004; N. multifasciatus: Kohler 1998). Models of alternative reproductive tactics (ARTs)
In a large range of taxa the members of cooperatively breeding groups employ ARTs (Taborsky 1994, Koenig & Dickinson 2008, Taborsky et al. 2008). This applies, for example, to birds in both sexes (Faaborg & Patterson 1981, Cockburn 2004, Vehrencamp & Quinn 2004), including among others dunnocks (Prunella modularis: Davies 1992), Gala´pagos hawks (Buteo galapagoensis: Faaborg et al. 1995), alpine accentors (Prunella collaris: Hartley et al. 1995), and Smith’s longspur (Calcarius pictus: Briskie et al. 1998). Also in many cooperatively reproducing fish species, groups consist of individuals employing ARTs (Taborsky 1987, 1994), with cooperative effort and reproductive success of group members being skewed to various degrees (Taborsky 1999, 2001, 2008a). In the male sex, it seems that in general the individuals that show more reproductive effort, i.e. the bourgeois competitors, are more successful in direct reproductive competition than males employing alternative tactics. This does not mean, however, that the formers’ average lifetime fitness will exceed that of satellite or helper males (Taborsky 2008a). There are two possibilities: either (1) the alternative tactic is a transient stage during life, i.e. both roles are performed by the same males at different phases of their lives (e.g. in some cooperatively breeding cichlids: Taborsky 1984, 1985, Stiver et al. 2004, Heg & Bachar 2006; see Table 10.2); or (2) different types of males specialize in different tactics throughout life (e.g. in Pelvicachromis pulcher: Martin & Taborsky 1997; Symphodus ocellatus: Alonzo et al. 2000; see Table 10.2). In the latter case, average lifetime reproductive success of different types of males should be similar and balanced by frequency-dependent selection, if the coexistence of male types is based on heritable variation (Gross 1996, Taborsky et al. 2008). If it results from conditional responses to the environment (physical, biotic, or social) or to individual quality variation (depending e.g. on conditions during ontogeny), males adopting alternative tactics do not necessarily attain a similar lifetime reproductive success (Repka & Gross 1995, Gross & Repka 1998, Taborsky 1998, Taborsky et al. 2008). Evolutionary stability of alternative phenotypes within populations that are based on heritable variation, and which involves frequency dependence, has four requirements (Brockmann & Taborsky 2008): (1) a mechanism by which
Cooperative fish groups: alternative modeling approaches distinct phenotypes can develop (e.g. a genetic polymorphism: Martin & Taborsky 1997, Taborsky 2001), (2) heritable variation in the mechanisms controlling the expression of the alternative phenotypes, (3) disruptive selection favoring multiple phenotypes rather than only one “optimum,” and (4) crossing fitness curves of the alternative tactics depending on relative frequencies. Fitness effects of cooperation between reproductive competitors influence the two latter conditions, which provides the opportunity to model the coexistence of alternative phenotypes in a cooperative-breeding framework (e.g. with the help of dynamic game theory: E. Skubic, I. M. Hamilton, & M. Taborsky, unpublished data). Modeling helper behavior as ART may be appropriate to clarify the selective forces promoting cooperation and reproductive participation in cooperative breeders in general, because the behavior of subordinates is a distinctly different alternative to the behavior of breeders, it is usually conditional, and it may often be frequency-dependent (see also Koenig & Dickinson 2008). Specific models developed to explain case studies
Among theory developed to understand animal behavior we must distinguish between heuristic models aiming to find simple general rules, such as the majority of reproductive-skew theory, and specific models aiming to depict particular examples. If one aspires to understand the evolutionary mechanisms responsible for decisions of animals in a particular example and situation, heuristic models are not the method of choice because of their overall simplicity and often inadequate assumptions. This can be illustrated by an example from cooperatively breeding fish, where helpers are capable of participating in reproduction, but at the cost of an increased expulsion risk. In fish, indeterminate growth causes a permanent change of an individual’s conditions and options (Taborsky 1999). For example, if a male helper grows larger it may produce more sperm, so its potential to succeed in sperm competition when participating in the breeder’s spawning increases (Taborsky 1998, Dierkes et al. 1999). This also means that it is a greater threat to the male breeder, especially if they are not related. Therefore, the relationship between breeders and helpers will change with time (Taborsky 1985). In addition, the alternative options helpers face outside of their own group change as well, because larger size means a greater potential to compete for territories (Balshine-Earn et al. 1998) and a reduced predation risk when leaving home (Heg et al. 2004). Therefore, if we wish to understand the criteria by which helpers should decide to participate in reproduction in their home territory, we must include size and growth to grasp the dynamics of the underlying processes. This can be incorporated by using a dynamic modeling approach,
289
290
M. Taborsky where body size is used as a state variable (Skubic et al. 2004). Applying this method to the cooperatively breeding cichlid Neolamprologus pulcher revealed that punishment by breeders if helpers parasitize their reproduction, relatedness between breeders and helpers, and the parasitism capacity of helpers are key variables for the decision of helpers to participate in reproduction, and hence will influence reproductive skew. Furthermore, future fitness expectations after a helper has left the group have important effects on optimal decisions of helpers (as for instance mortality risk differs between group members and non-group members). The virtue of this approach is that due to its comparatively high inherent realism it can be parameterized with data obtained from a particular system, so it leads to testable, quantitative predictions (Skubic et al. 2004). Similar approaches have been used to model the influence of future fitness effects, relatedness, and mortality on dispersal decisions in cooperatively breeding dwarf mongooses (Helogale parvula; with age, status, and relatedness included as state variables: Lucas et al. 1997), and subordinate female dispersal and eviction probabilities of dominant females in meerkats (Suricata suricatta; using rank and group size as state variables: Stephens et al. 2005). Interestingly, the results derived from such an approach may contradict the predictions from heuristic reproductive skew models (see Skubic et al. 2004: the predictions regarding the relationship between relatedness and reproductive skew are opposite between this model and the restraint model of Johnstone & Cant 1999 when applied to the same situation). Further progress in predicting how breeders and helpers interact and share reproduction can be achieved by modeling a dynamic game to take account of both of their individual optima and capacities (E. Skubic, I. M. Hamilton, & M. Taborsky, unpublished data).
Conclusions and perspectives: where next? The amazing variation of cooperative reproduction in fishes provides excellent opportunities to investigate general evolutionary mechanisms responsible for patterns of cooperation and conflict, altruism and parasitism among reproductive competitors at various levels of complexity. In this context, it is of special interest that superficially similar cooperative systems in fish may be based on distinctly different parameter combinations, even in closely related species. For example, the often polygynandrous group structure of Neolamprologus pulcher and N. multifasciatus is remarkably similar, as is the type and quantity of cooperation and reproductive competition shown among group members, although the relatedness pattern is completely different (Kohler 1998, Taborsky 2001, Dierkes et al. 2005). In Julidochromis ornatus large
Cooperative fish groups: alternative modeling approaches male helpers and breeders are similarly unrelated as in N. pulcher, but for very different reasons, and the reproductive skew among males differs remarkably between these two species (Dierkes et al. 1999, 2008, Stiver et al. 2004, Awata et al. 2005, Heg & Bachar 2006, Heg et al. 2006). The complex influence of crucial parameters on reproductive cooperation and skew can be illustrated by the relation between relatedness and different types of helping behaviors performed in N. pulcher. In this cichlid, work effort of helpers varies with the degree of relatedness to male and female breeders differently, dependent on their own and the breeder’s sex, suggesting that helpers react to much more complex cues than the average relatedness within groups. For example, helpers invest more in defense when they are unrelated to the male breeder, but less when they are unrelated to the female breeder (Stiver et al. 2005), and male and female helpers share differently in reproduction (Taborsky 1985, Heg et al. 2006, Dierkes et al. 2008). Helpers might invest in direct brood care of kin produced by their mother, but increase defense for being allowed to stay and breed when unrelated to the dominant male in the group. Fortunately there are sufficient data available from some cooperatively reproducing fish species to examine the explanatory power of alternative approaches to unravel evolutionary mechanisms underlying the reproductive sharing among group members. From the above discussion we can conclude that: (1) the variability of cooperative behavior shown between reproductive competitors in fish is stupendous, and arguably unmatched among vertebrates. (2) Conventional reproductive-skew models do not account for the complexity of evolutionary mechanisms involved in reproductive skew among members of fish groups. They are helpful, however, in identifying parameters that might constitute important components of such mechanisms. (3) Alternative approaches such as reciprocity theory or models to explain the coexistence of ARTs may have greater explanatory potential in particular cases, but they have generally not been developed sufficiently to fulfill this aim. (4) To understand how decisions evolve of different types of group members such as breeders and helpers, or bourgeois and satellite individuals, a more elaborate approach is needed that takes account of the state dynamics involved. Apart from the general limitations of skew models to explain reproductive sharing among members of cooperative groups, which are mainly due to their inherent simplicity and unrealistic assumptions (as outlined e.g. by Magrath & Heinsohn
291
292
M. Taborsky 2000, Kokko 2003, Nonacs 2006, 2007, Hamilton & Heg 2007, Chapter 9 in this volume), in fish the situation is additionally complicated by the fact that conditions constantly change as these animals continue to grow after reaching maturity. Also, reproductive-skew models have been developed mainly to explain eusocial insects, where groups are stable; usually this is not so in fish species exhibiting cooperative reproduction. More generally, reproductive-skew models often consider a single decision by a dominant and a single decision by a subordinate, which is a far cry from grasping the natural situation – and hence they leave us stuck at a stage of vague guesses about potential evolutionary mechanisms. Empirical research in behavioral ecology tends to take coincidence between predictions of a model and actual data as proof that the respective model explains the causal mechanisms underlying a trait. This is a plausible but dangerously misguided approach, especially if it builds on heuristic models that do not grasp more than a puny minimum of factors involved in the evolution of animal decisions. Naturally, this is no critique of the models, because by unveiling potential relationships between important factors they guide our thinking and thereby serve an important purpose. It is a critique, however, of the naive use of models to “explain,” for example, cooperative behavior and reproductive skew in any particular case. It seems relevant here to point out three frequent fallacies. These are not confined to the issue of reproductive skew, but are rather a general problem in behavioral ecology (Taborsky 2008b); however, the reproductive skew literature provides ample illustration. (1)
Predictions about a relationship between crucial factors, e.g. relatedness and skew, are derived from reproductive-skew models and tested with a dataset from a natural population providing the required information. The aim of the approach is to find out whether an existing pattern can be explained by applying one or the other model. This very frequent approach is flawed in two important ways. First, the intention of modelers is usually to investigate general principles, not to “explain” particular systems found in nature. Accordingly, the assumptions underlying reproductive-skew models are generally simple, to keep the model practical and intelligible. Without thorough evaluation of these assumptions any application of reproductive-skew theory to a given dataset will be meaningless, because any observed pattern is compatible with some variant of such general models (Johnstone 2000; e.g. Hamilton & Heg 2007 demonstrate how a small extension of original assumptions may reverse the predictions).
Cooperative fish groups: alternative modeling approaches
(2)
(3)
Second, a correlative approach cannot reliably unravel functional or causal relationships (Milinski 1997). Many mechanisms may cause a given pattern, and a model may show that one particular mechanism can create that pattern. This, however, does not prove that it is the mechanism responsible for the pattern in any specific case. Predictions derived from heuristic skew models are tested by an experiment. Obviously, this hitherto amazingly rare approach is much less flawed than the first one, because it is not as strongly affected by the problem that coincidence between predictions and data does not allow conclusions to be made about about causal relationships. However, it is constrained by the difficulty of taking account of all important parameters involved and their interactions. Notably, reproductive-skew models rest usually on extremely simple assumptions, by intention excluding a large number of potentially important variables and their interactions. Therefore an experiment must ensure that the assumptions of the model are fully met and that no significant factor or relationship between variables has gone unnoticed. It is important to bear in mind that even when an experimental result meets a prediction, this does not mean that the underlying hypothesis provides an adequate explanation for the result. For example, if I use a skew model predicting a direction of influence of relatedness among group members on reproductive skew, and my experimental manipulation reveals an effect as predicted, the underlying mechanisms may still not be reflected by the model; it may well be that the change of the relatedness structure incidentally affected dominance among group members, which in turn affected dispersal propensity, which in turn affected cooperation propensity (e.g. in a pay-to-stay scenario), which then also affected reproductive skew. If the experiment had been done differently (e.g. no dominance differences among group members between treatments), the result might have differed substantially. The only way out of this dilemma is to unravel underlying mechanisms. As long as this is not fully accomplished, even a fit between a model’s predictions and results of an experiment aiming to test them is no proof that the model explains the system. Empirical data are used to “test a skew model.” The logic of this approach is that if a relationship predicted by a model is not confirmed by empirical data, the model is probably wrong. This approach suffers from a fundamental misunderstanding of important elements of the scientific method: if a theoretical model is not flawed by logical or formal errors, it is “true,” full stop. Empirical data can neither prove
293
294
M. Taborsky nor disprove it. The conclusion that a mismatch between data and predictions allows is either that the assumptions of the model were not met in the empirical study, or that the model does not reflect the issue it aims to abstract. In either case, the model may turn out to be irrelevant for our understanding of reproductive skew in a particular case or – after applying sufficient scrutiny – also in general. This, and no more, can we learn about the quality of a model from a mismatch between its predictions and empirical data.
For obvious reasons I have not cited examples for the flaws described above, but the reader might take delight in an attempt to categorize publications dealing with reproductive skew in cooperative breeders into one of the three categories. Is there a way out? Yes there is. The important points are that (1) an experimental approach is used and (2) the underlying mechanisms (ultimate and proximate) are unraveled. This is vastly demanding and does not provide quick answers, but it is indispensable (Taborsky 2008b). Heuristic skew theory will help to identify important factors and critically assess one’s assumptions, thereby guiding research. It should not be misinterpreted as providing predictions that are easily testable with a particular case study. If empiricists and theoreticians are to benefit from each other, modelers should strive to use realistic assumptions (i.e. as close to reality as possible), and ethologists should use theoretical models as guidance to ask the right questions and to generate falsifiable hypotheses (Cant & Field 2001). An alternative is to produce models aimed at explaining specific relationships in a particular system, which can generate testable, quantitative predictions when parameterized with the respective data (Skubic et al. 2004). Both approaches are suited to improving our understanding of mechanisms causing reproductive skew, and should lead to a constructive interaction between theory and empiricism. Acknowledgments I am grateful to Reinmar Hager, Dik Heg, Daniel Rankin, Eva Skubic, and Barbara Taborsky for constructive comments on the manuscript, to Barbara Taborsky for help with Figure 10.3, and to the Swiss National Science Foundation for financial support (SNF grant no. 3100A0–105626). References Alonzo, S. H. & Warner, R. R. (1999). A trade-off generated by sexual conflict: Mediterranean wrasse males refuse present mates to increase future success. Behavioral Ecology, 10, 105–111.
Cooperative fish groups: alternative modeling approaches Alonzo, S. H. & Warner, R. R. (2000). Allocation to mate guarding or increased sperm production in a Mediterranean wrasse. American Naturalist, 156, 266–275. Alonzo, S. H., Taborsky, M., & Wirtz, P. (2000). Male alternative reproductive behaviors in a Mediterranean wrasse, Symphodus ocellatus: evidence from otoliths for multiple life-history pathways. Evolutionary Ecology Research, 2, 997–1007. Avise, J. C., Jones, A. G., Walker, D., & DeWoody, J. A. (2002). Genetic mating systems and reproductive natural histories of fishes: lessons for ecology and evolution. Annual Review of Genetics, 36, 19–45. Awata, S., Munehara, H., & Kohda, M. (2005). Social system and reproduction of helpers in a cooperatively breeding cichlid fish (Julidochromis ornatus) in Lake Tanganyika: field observations and parentage analyses. Behavioral Ecology and Sociobiology, 58, 506–516. Awata, S., Heg, D., Munehara, H., & Kohda, M. (2006). Testis size depends on social status and the presence of male helpers in the cooperatively breeding cichlid Julidochromis ornatus. Behavioral Ecology, 17, 372–379. Balshine, S., Leach, B., Neat, F., et al. (2001). Correlates of group size in a cooperatively breeding cichlid fish. Behavioral Ecology and Sociobiology, 50, 134–140. Balshine-Earn, S., Neat, F. C., Reid, H., & Taborsky, M. (1998). Paying to stay or paying to breed? Field evidence for direct benefits of helping behavior in a cooperatively breeding fish. Behavioral Ecology, 9, 432–438. Barlow, G. W. (1961). Social behavior of the desert pupfish, Cyprinodon macularius, in the field and in the aquarium. American Midland Naturalist, 65, 339–359. Basieux, N. (2007). Cooperation in reproduction between unrelated male competitors in a Mediterranean wrasse. Unpublished masters thesis, University of Berne, Switzerland. Berglund, A. (1997). Mating systems and sex allocation. In J.-G. J. Godin, ed., Behavioural Ecology of Teleost Fishes. Oxford: Oxford University Press, pp. 237–265. Bergmu¨ller, R. & Taborsky, M. (2005). Experimental manipulation of helping in a cooperative breeder: helpers “pay to stay” by pre-emptive appeasement. Animal Behaviour, 69, 19–28. Bergmu¨ller, R., Heg, D., Peer, K., & Taborsky, M. (2005a). Extended safe havens and between-group dispersal of helpers in a cooperatively breeding cichlid. Behaviour, 142, 1643–1667. Bergmu¨ller, R., Heg, D., & Taborsky, M. (2005b). Helpers in a cooperatively breeding cichlid stay and pay or disperse and breed, depending on ecological constraints. Proceedings of the Royal Society of London B, 272, 325–331. Bergmu¨ller, R., Johnstone, R. A., Russell, A., & Bshary, R. (2007). Integrating cooperative breeding into theoretical concepts of cooperation. Behavioural Processes, 76, 61–72. Blumer, L. S. (1982). A bibliography and categorization of bony fishes exhibiting parental care. Zoological Journal of the Linnean Society, 76, 1–22. Bourke, A. F. G. (1999). Colony size, social complexity and reproductive conflict in social insects. Journal of Evolutionary Biology, 12, 245–257.
295
296
M. Taborsky Bowman, M. L. (1970). Life history of the black redhorse, Moxostoma duquesnei (Lesueur), in Missouri. Transactions of the American Fisheries Society, 99, 546–559. Boyd, R. & Richerson, P. J. (1992). Punishment allows the evolution of cooperation (or anything else) in sizable groups. Ethology and Sociobiology, 13, 171–195. Breder, C. M. & Rosen, D. E. (1966). Modes of Reproduction in Fishes. Garden City, NY: Natural History Press. Briskie, J. V., Montgomerie, R., Poldmaa, T., & Boag, P. T. (1998). Paternity and paternal care in the polygynandrous Smith’s longspur. Behavioral Ecology and Sociobiology, 43, 181–190. Brockmann, H. J. & Taborsky, M. (2008). Alternative reproductive tactics and the evolution of alternative allocation phenotypes. In R. F. Oliveira, M. Taborsky, & H. J. Brockmann, eds., Alternative Reproductive Tactics: an Integrative Approach. Cambridge: Cambridge University Press, pp. 25–51. Brouwer, L., Heg, D., & Taborsky, M. (2005). Experimental evidence for helper effects in a cooperatively breeding cichlid. Behavioral Ecology, 16, 667–673. Burr, B. M. & Morris, M. A. (1977). Spawning behavior of the shorthead redhorse, Moxostoma macrolepidotum, in Big Rock Creek, Illinois. Transactions of the American Fisheries Society, 106, 80–82. Bu¨scher, H. H. (1992). Ein neuer Cichlide aus dem Tanganjikasee: Neolamprologus similis n. sp. Aquarien- und Terrarienzeitschrift, 45, 520–525. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169. Cant, M. A. (2000). Social control of reproduction in banded mongooses. Animal Behaviour, 59, 147–158. Cant, M. A. & Field, J. (2001). Helping effort and future fitness in cooperative animal societies. Proceedings of the Royal Society of London B, 268, 1959–1964. Cant, M. A. & Field, J. (2005). Helping effort in a dominance hierarchy. Behavioral Ecology, 16, 708–715. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Chapuisat, M. & Keller, L. (1999). Testing kin selection with sex allocation data in eusocial Hymenoptera. Heredity, 82, 473–478. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H. & Parker, G. A. (1995). Punishment in animal societies. Nature, 373, 209–216. Cockburn, A. (2004). Mating systems and sexual conflict. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 81–101. Davies, N. B. (1992). Dunnock Behaviour and Social Evolution. Oxford: Oxford University Press. Demartini, E. E. & Patten, B. G. (1979). Egg guarding and reproductive biology of the red Irish lord, Hemilepidatus hemilepidotus (Tilesius). Syesis, 12, 41–55.
Cooperative fish groups: alternative modeling approaches Dierkes, P., Taborsky, M., & Kohler, U. (1999). Reproductive parasitism of broodcare helpers in a cooperatively breeding fish. Behavioral Ecology, 10, 510–515. Dierkes, P., Heg, D., Taborsky, M., Skubic, E., & Achmann, R. (2005). Genetic relatedness in groups is sex-specific and declines with age of helpers in a cooperatively breeding cichlid. Ecology Letters, 8, 968–975. Dierkes, P., Taborsky, M., & Achmann, R. (2008). Multiple paternity in the cooperatively breeding fish Neolamprologus pulcher. Behavioral Ecology and Sociobiology, 62, 1581–1589. Duftner, N., Sefc, K. M., Koblmuller, S., et al. (2007). Parallel evolution of facial stripe patterns in the Neolamprologus brichardilpulcher species complex endemic to Lake Tanganyika. Molecular Phylogenetics and Evolution, 45, 706–715. Dugatkin, L. A. (1997). Cooperation Among Animals: an Evolutionary Perspective. Oxford: Oxford University Press. Faaborg, J. & Patterson, C. B. (1981). The characteristics and occurrence of cooperative polyandry. Ibis, 123, 477–484. Faaborg, J., Parker, P. G., Delay, L., et al. (1995). Confirmation of cooperative polyandry in the Galapagos hawk (Buteo galapagoensis). Behavioral Ecology and Sociobiology, 36, 83–90. Fahy, W. E. (1954). The life history of the northern greenside darter, Etheostoma blennioides blennioides Rafinesque. Journal of the Elisha Mitchell Scientific Society, 70, 139–205. Faulkes, C. G. & Bennett, N. C. (2001). Family values: group dynamics and social control of reproduction in African mole-rats. Trends in Ecology and Evolution, 16, 184–190. Fiedler, K. (1964). Verhaltensstudien an Lippfischen der Gattung Crenilabrus (Labridae, Perciformes). Zeitschrift fu¨r Tierpsychologie, 21, 521–591. Frank, S. A. (1995). Mutual policing and repression of competition in the evolution of cooperative groups. Nature, 377, 520–522. Frank, S. A. (2003). Perspective: repression of competition and the evolution of cooperation. Evolution, 57, 693–705. Fricke, H. W. (1979). Mating system, resource defence and sex change in the Anemonefish Amphiprion akallopisos. Zeitschrift fu¨r Tierpsychologie, 50, 313–326. Griffin, A. S. & West, S. A. (2002). Kin selection: fact and fiction. Trends in Ecology and Evolution, 17, 15–21. Griffin, A. S. & West, S. A. (2003). Kin discrimination and the benefit of helping in cooperatively breeding vertebrates. Science, 302, 634–636. Gross, M. R. (1996). Alternative reproductive strategies and tactics: diversity within sexes. Trends in Ecology and Evolution, 11, 92–98. Gross, M. R. & Repka, J. (1998). Stability with inheritance in the conditional strategy. Journal of Theoretical Biology, 192, 445–453. Hackney, P. A., Tatum, W. M., & Spencer, S. L. (1968). Life history study of the river redhorse, Moxostoma carinatum (Cope), in the Cahaba River, Alabama, with notes on the management of the species as a sport fish. Proceedings of the Southeast Association of Game and Fish Commissioners, 21, 324–332.
297
298
M. Taborsky Hamilton, I. M. & Heg, D. (2007). Clutch-size adjustments and skew models: effects on reproductive partitioning and group stability. Behavioral Ecology, 18, 467–476. Hamilton, I. M. & Taborsky, M. (2005). Unrelated helpers will not fully compensate for costs imposed on breeders when they pay to stay. Proceedings of the Royal Society of London B, 272, 445–454. Hamilton, W. D. (1964). The genetical evolution of social behaviour. Journal of Theoretical Biology, 7, 1–16, 17–52. Hartley, I. R., Davies, N. B., Hatchwell, B. J., et al. (1995). The polygynandrous mating system of the alpine accentor, Prunella collaris. 2. Multiple paternity and parental effort. Animal Behaviour, 49, 789–803. Haydock, J. & Koenig, W. D. (2002). Reproductive skew in the polygynandrous acorn woodpecker. Proceedings of the National Academy of Sciences of the USA, 99, 7178– 7183. Haydock, J. & Koenig, W. D. (2003). Patterns of reproductive skew in the polygynandrous acorn woodpecker. American Naturalist, 162, 277–289. Heg, D. & Bachar, Z. (2006). Cooperative breeding in the Lake Tanganyika cichlid Julidochromis ornatus. Environmental Biology of Fishes, 76, 265–281. Heg, D. & Hamilton, I. M. (2008). Tug-of-war over reproduction in a cooperatively breeding cichlid. Behavioral Ecology and Sociobiology, 62, 1249–1257. Heg, D., Bachar, Z., Brouwer, L., & Taborsky, M. (2004). Predation risk is an ecological constraint for helper dispersal in a cooperatively breeding cichlid. Proceedings of the Royal Society of London B, 271, 2367–2374. Heg, D., Brouwer, L., Bachar, Z., & Taborsky, M. (2005a). Large group size yields group stability in the cooperatively breeding cichlid Neolamprologus pulcher. Behaviour, 142, 1615–1641. Heg, D., Bachar, Z., & Taborsky, M. (2005b). Cooperative breeding and group structure in the Lake Tanganyika cichlid Neolamprologus savoryi. Ethology, 111, 1017–1043. Heg, D., Bergmuller, R., Bonfils, D., et al. (2006). Cichlids do not adjust reproductive skew to the availability of independent breeding options. Behavioral Ecology, 17, 419–429. Jenkins, R. E. (1970). Systematic studies of the catostomid fish tribe Moxostomatini. Unpublished Ph.D. thesis, Cornell University, USA. Jenkins, R. E. & Jenkins, D. J. (1980). Reproductive behavior of the greater redhorse, Moxostoma valenciennesi, in the Thousand Islands region. Canadian Field-Naturalist, 94, 426–430. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Cant, M. A. (1999). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Keenleyside, M. H. A. (1979). Diversity and Adaptation in Fish Behaviour. Berlin: Springer. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–103.
Cooperative fish groups: alternative modeling approaches Killingback, T. & Doebeli, M. (2002). The continuous prisoner’s dilemma and the evolution of cooperation through reciprocal altruism with variable investment. American Naturalist, 160, 421–438. Kodric-Brown, A. (1977). Reproductive success and the evolution of breeding territories in pupfish (Cyprinodon). Evolution, 31, 750–766. Koenig, W. D. & Dickinson, J. L. (2008). Cooperative breeding as an alternative reproductive tactic. In R. F. Oliveira, M. Taborsky & H. J. Brockmann, eds., Alternative Reproductive Tactics: An Integrative Approach. Cambridge: Cambridge University Press, pp. 451–470. Kohler, U. (1998). Zur Struktur und Evolution des Sozialsystems von Neolamprologus Multifasciatus (Cichlidae, Pisces), dem kleinsten Schneckenbuntbarsch des Tanganjikasees. Aachen: Shaker Verlag. Kokko, H. (2003). Are reproductive skew models evolutionarily stable? Proceedings of the Royal Society of London B, 270, 265–270. Kokko, H., Johnstone, R. A., & Clutton-Brock, T. H. (2001). The evolution of cooperative breeding through group augmentation. Proceedings of the Royal Society of London B, 268, 187–196. Kondo, T. (1986). Feeding habits of Lamprologus savoryi (Teleostei: Cichlidae) with reference to its social behaviour. Physiology and Ecology Japan, 23, 1–15. Kuwamura, T. (1997). The evolution of parental care and mating systems among Tanganyikan cichlids. In H. Kawanabe, M. Hori, & M. Nagoshi, eds., Fish Communities in Lake Tanganyika. Kyoto: Kyoto University Press, pp. 59–86. Kwak, T. J. & Skelly, T. M. (1992). Spawning habitat, behavior and morphology as isolating mechanisms of the golden redhorse, Moxostoma erythrurum and the black redhorse, M. duquesnei, two syntopic fishes. Environmental Biology of Fishes, 34, 127–137. Langer, P., Hogendoorn, K., & Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature, 428, 844–847. Leigh, E. G., Jr. (1971). Adaptation and Diversity. San Francisco, CA: Freeman, Cooper. Lejeune, P. (1985). Etude e´coe´thologique des comportements re´producteurs et sociaux des Labridae me´diterrane´ens des genres Symphodus (Rafinesque, 1810) et Coris (Lacepede, 1802). Cahiers d’Ethologie Applique´e, 5, 1–208. Lucas, J. R., Creel, S. R., & Waser, P. M. (1997). Dynamic optimization and cooperative breeding: an evaluation of future fitness effects. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 171–198. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Martin, E. & Taborsky, M. (1997). Alternative male mating tactics in a cichlid, Pelvicachromis pulcher: a comparison of reproductive effort and success. Behavioral Ecology and Sociobiology, 41, 311–319. Michod, R. E. (1982). The theory of kin selection. Annual Review of Ecology and Systematics, 13, 23–55.
299
300
M. Taborsky Milinski, M. (1997). How to avoid seven deadly sins in the study of behavior. Advances in the Study of Behavior, 26, 159–180. Nonacs, P. (2006). The rise and fall of transactional skew theory in the model genus Polistes. Annales Zoologici Fennici, 43, 443–455. Nonacs, P. (2007). Tug-of-war has no borders: it is the missing model in reproductive skew theory. Evolution, 61, 1244–1250. Nowak, M. A. (2006). Five rules for the evolution of cooperation. Science, 314, 1560–1563. Oliveira, R. F., Carvalho, N., Miranda, J., et al. (2002). The relationship between the presence of satellite males and nest-holders’ mating success in the Azorean rock-pool blenny Parablennius sanguinolentus parvicornis. Ethology, 108, 223–235. Page, L. M. & Johnston, C. E. (1990). Spawning in the creek chubsucker, Erimyzon oblongus, with a review of spawning behavior in suckers (Catostomidae). Environmental Biology of Fishes, 27, 265–272. Petersen, C. W. & Warner, R. R. (1998). Sperm competition in fishes. In T. R. Birkhead & A. P. Mo¨ller, eds., Sperm Competition and Sexual Selection. San Diego, CA: Academic Press, pp. 435–463. Poll, M. (1974). Contribution a` la faune ichthyologique du lac Tanganika, d’apre`s les re´coltes de P. Brichard. Revue Zoologie Africaine, 88, 99–110. Queller, D. C. (1994). Genetic relatedness in viscous populations. Evolutionary Ecology, 8, 70–73. Queller, D. C. & Strassmann, J. E. (1998). Kin selection and social insects. Bioscience, 48, 165–175. Raney, E. C. & Lachner, E. A. (1946). Age, growth and habits of the hog sucker, Hypentelium nigricans (Le Sueur), in New York. American Midland Naturalist, 36, 76– 86. Ratnieks, F. L.W. (1988). Reproductive harmony via mutual policing by workers in eusocial hymenoptera. American Naturalist, 132, 217–236. Ratnieks, F. L. W. & Visscher, P. K. (1989). Worker policing in the honeybee. Nature 342, 796–797. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflict in polygynous societies: Mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S.-F. (2006). A missing model in reproductive skew theory: The bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Reid, G. M. & Hall, H. (2003). Reproduction in fishes in relation to conservation. In W. V. Holt, A. R. Pickard, J. Rodger, & D. E. Wildt, eds., Reproductive Science and Integrated Conservation. Cambridge: Cambridge University Press, pp. 374–393.
Cooperative fish groups: alternative modeling approaches Reighard, J. (1920). The breeding behaviors of the suckers and minnows. I. The suckers. Biological Bulletin, 38, 1–33. Reighard, J. (1943). The breeding habits of the river chub, Nocomis micropogon (Cope). Papers of the Michigan Academy of Science, Arts and Letters, 28, 397–423. Repka, J. & Gross, M. R. (1995). The evolutionarily stable strategy under individual condition and tactic frequency. Journal of Theoretical Biology, 176, 27–31. Rusu, A. S. & Krackow, S. (2004). Kin-preferential cooperation, dominancedependent reproductive skew, and competition for mates in communally nesting female house mice. Behavioral Ecology and Sociobiology, 56, 298–305. Sachs, J. L., Mueller, U. G., Wilcox, T. P., & Bull, J. J. (2004). The evolution of cooperation. Quarterly Review of Biology, 79, 135–160. Santos, R. S. (1985). Parentais e sate´lites: ta´cticas alternativas de acasalamento nos machos de Blennius sanguinolentus pallas (Pisces: Blenniidae). Arquipe´lago S. C. N., VI, 119–146. Santos, R. S. & Almada, V. C. (1988). Intraspecific variations in reproductive tactics in males of the rocky intertidal fish Blennius sanguinolentus in the Azores. In G. Chelazzi & M. Vannini, eds., Behavioral Adaptation to Intertidal Life. New York, NY: Plenum Press, pp. 421–447. Schradin, C. & Lamprecht, J. (2000). Female-biased immigration and male peacekeeping in groups of the shell-dwelling cichlid fish Neolamprologus multifasciatus. Behavioral Ecology and Sociobiology, 48, 236–242. Schradin, C. & Lamprecht, J. (2002). Causes of female emigration in the group-living cichlid fish Neolamprologus multifasciatus. Ethology, 108, 237–248. Skubic, E., Taborsky, M., McNamara, J. M., & Houston, A. I. (2004). When to parasitize? A dynamic optimization model of reproductive strategies in a cooperative breeder. Journal of Theoretical Biology, 227, 487–501. Soljan, T. (1930). Die Fortpflanzung und das Wachstum von Crenilabrus ocellatus Forsk., einem Lippfisch des Mittelmeeres. Zeitschrift fu¨r wissenschaftliche Zoologie, 137, 150–174. Soljan, T. (1931). Brutpflege durch Nestbau bei Crenilabrus quinquemaculatus Risso, einem adriatischen Lippfisch. Zeitschrift fu¨r Morphologie und O¨kologie der Tiere, 20, 132–135. Stacey, P. B. (1982). Female promiscuity and male reproductive success in social birds and mammals. American Naturalist, 120, 51–64. Stephens, P. A., Russell, A. F., Young, A. J., Sutherland, W. J., & Clutton-Brock, T. H. (2005). Dispersal, eviction, and conflict in meerkats (Suricata suricatta): an evolutionarily stable strategy model. American Naturalist, 165, 120–135. Stiver, K. A., Dierkes, P., Taborsky, M., & Balshine, S. (2004). Dispersal patterns and status change in a co-operatively breeding cichlid Neolamprologus pulcher: evidence from microsatellite analyses and behavioural observations. Journal of Fish Biology, 65, 91–105. Stiver, K. A., Dierkes, P., Taborsky, M., Gibbs, H. L., & Balshine, S. (2005). Relatedness and helping in fish: examining the theoretical predictions. Proceedings of the Royal Society of London B, 272, 1593–1599.
301
302
M. Taborsky Sunobe, T. (2000). Social structure, nest guarding and interspecific relationships of the cichlid fish (Julidochromis marlieri) in Lake Tanganyika. African Study Monographs, 21, 83–89. Taborsky, B., Skubic, E., & Bruintjes, R. (2007). Mothers adjust egg size to helper number in a cooperatively breeding cichlid. Behavioral Ecology, 18, 652–657. Taborsky, M. (1984). Broodcare helpers in the cichlid fish Lamprologus brichardi: their costs and benefits. Animal Behaviour, 32, 1236–1252. Taborsky, M. (1985). Breeder–helper conflict in a cichlid fish with broodcare helpers: an experimental analysis. Behaviour, 95, 45–75. Taborsky, M. (1987). Cooperative behaviour in fish: coalitions, kin groups and reciprocity. In Y. Ito, J. L. Brown, & J. Kikkawa, eds., Animal Societies: Theories and Facts. Tokyo: Japan Scientific Societies Press, pp. 229–237. Taborsky, M. (1994). Sneakers, satellites, and helpers: parasitic and cooperative behavior in fish reproduction. Advances in the Study of Behavior, 23, 1–100. Taborsky, M. (1997). Bourgeois and parasitic tactics: do we need collective, functional terms for alternative reproductive behaviours? Behavioral Ecology and Sociobiology, 41, 361–362. Taborsky, M. (1998). Sperm competition in fish: bourgeois males and parasitic spawning. Trends in Ecology and Evolution, 13, 222–227. Taborsky, M. (1999). Conflict or cooperation: what determines optimal solutions to competition in fish reproduction? In R. Oliveira, V. Almada, & E. Goncalves, eds., Behaviour and Conservation of Littoral Fishes. Lisbon: ISPA. Taborsky, M. (2001). The evolution of parasitic and cooperative reproductive behaviors in fishes. Journal of Heredity, 92, 100–110. Taborsky, M. (2008a). Alternative reproductive tactics in fish. In R. F. Oliveira, M. Taborsky & H. J. Brockmann, eds., Alternative Reproductive Tactics: an Integrative Approach. Cambridge: Cambridge University Press, pp. 251–299. Taborsky, M. (2008b). The use of theory in behavioural research. Ethology, 114, 1–6. Taborsky, M. & Grantner, A. (1998). Behavioural time–energy budgets of cooperatively breeding Neolamprologus pulcher (Pisces: Cichlidae). Animal Behaviour 56, 1375–1382. Taborsky, M. & Limberger, D. (1981). Helpers in fish. Behavioral Ecology and Sociobiology, 8, 143–145. Taborsky, M., Hert, E., Siemens, M., & Stoerig, P. (1986). Social behaviour of Lamprologus species: functions and mechanisms. Annales de Muse´e Royal de l’Afrique Centrale, Sciences Zoologiques, 251, 7–11. Taborsky, M., Hudde, B., & Wirtz, P. (1987). Reproductive behaviour and ecology of Symphodus (Crenilabrus) ocellatus, a European wrasse with four types of male behaviour. Behaviour, 102, 82–118. Taborsky, M., Oliveira, R. F., & Brockmann, H. J. (2008). The evolution of alternative reproductive tactics: concepts and questions. In R. F. Oliveira, M. Taborsky, & H. J. Brockmann, eds., Alternative Reproductive Tactics: an Integrative Approach. Cambridge: Cambridge University Press, pp. 1–21.
Cooperative fish groups: alternative modeling approaches Tebbich, S., Taborsky, M., & Winkler, H. (1996). Social manipulation causes cooperation in keas. Animal Behaviour, 52, 1–10. Thresher, R. E. (1979). Social behavior and ecology of two sympatric wrasses (Labridae: Halichoeres spp.) off the coast of Florida. Marine Biology, 53, 161–172. Trivers, R. L. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35–57. Trivers, R. L. & Hare, H. (1976). Haplodiploidy and evolution of social insects. Science, 191, 249–263. Turner, G. F. (1993). Teleost mating behaviour. In T. J. Pitcher, ed., Behaviour of Teleost Fishes. London: Chapman & Hall, pp. 307–332. Vehrencamp, S. L. (1979). The roles of individual, kin, and group selection in the evolution of sociality. In P. Marler & J. G. Vandenbergh, eds., Handbook of Behavioral Neurobiology 3, Social Behavior and Communication. New York, NY: Plenum Press, pp. 351–394. Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Vehrencamp, S. L. & Quinn, J. S. (2004). Joint laying systems. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 81–101. Wallin, J. E. (1989). Bluehead chub (Nocomis leptocephalus) nests used by yellowfin shiners (Notropis lutipinnis). Copeia, 1989, 1077–1080. Wenseleers, T. & Ratnieks, F. L.W. (2006a). Comparative analysis of worker reproduction and policing in eusocial hymenoptera supports relatedness theory. American Naturalist, 168, E163–E179. Wenseleers, T. & Ratnieks, F. L.W. (2006b). Enforced altruism in insect societies. Nature, 444, 50. Wenseleers, T., Helantera, H., Hart, A., & Ratnieks, F. L.W. (2004). Worker reproduction and policing in insect societies: an ESS analysis. Journal of Evolutionary Biology, 17, 1035–1047. West, S. A., Pen, I., & Griffin, A. S. (2002). Cooperation and competition between relatives. Science, 296, 72–75. West, S. A., Griffin, A. S., & Gardner, A. (2007a). Social semantics: altruism, cooperation, mutualism, strong reciptrocity and group selection. Journal of Evolutionary Biology, 20, 415–432. West, S. A., Griffin, A. S., & Gardner, A. (2007b). Evolutionary explanations for cooperation. Current Biology, 17, R661–R672. Wisenden, B. D. (1999). Alloparental care in fishes. Reviews in Fish Biology and Fisheries, 9, 45–70. Witte, K. E. & Schmidt, J. (1992). Betta brownorum, a new species of anabantoids (Teleostei: Belontiidae) from northwestern Borneo, with a key to the genus. Ichthyological Exploration of Freshwaters, 2, 305–330. Woodland, D. (2002). Breeding Neolamprologus gracilis. Cichlid News Magazine, 11, 30–34.
303
304
M. Taborsky Yamagishi, S. & Kohda, M. (1996). Is the cichlid fish Julidochromis marlieri polyandrous? Ichthyological Research, 43, 469–471. Yanagisawa, Y. (1987). Social organization of a polygyneous cichlid Lamprologus furcifer in Lake Tanganyika. Japanese Journal of Ichthyology, 34, 82–90. Young, A. J., Carlson, A. A., Monfort, S. L., et al. (2006). Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proceedings of the National Academy of Sciences of the USA, 103, 12005–12010.
11
Reproductive skew in primitively eusocial wasps: how useful are current models? jeremy fiel d and michael a. c ant
Summary In this chapter, we compare the predictions of reproductive-skew models with data from primitively eusocial wasps, the insect taxon in which skew has been best studied. These wasps share some key biological features with cooperatively breeding vertebrates, but represent a more experimentally tractable system. We describe a useful classification of skew models based on concepts of battleground and resolution models, and suggest how the basic biology of a taxon can help to identify which models and predictions in our classification are relevant. In primitively eusocial wasps, dominants have been assumed to control the allocation of reproductive shares at low cost. A priori, we therefore expect dominants to offer the minimum share required to retain a subordinate in the group (the staying incentive), or deter it from fighting (the peace-incentive). Optimization constraints are unlikely to apply because the cost of producing eggs is relatively low and non-accelerating. Among eight detailed genetic studies of primitively eusocial wasps, only one has found strong support for the concession model of skew. None of the other studies found clear relationships between skew and relatedness, productivity, or relative body size. Skew was typically high, often uniformly high across groups. There are several possible explanations for this apparent lack of fit between empirical studies and the concession model. First, there are shortcomings of the data, such as small sample sizes and uncertainty concerning the chance of inheritance by subordinates. Second, strong ecological Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
305
306
J. Field, M. A. Cant constraints and a good chance of inheritance reduce the need for staying incentives, in which case other factors such as the threat of fighting must be invoked to explain reproductive sharing. Across studies, the predictions of the peace-incentive model were not supported, since there was no consistent correlation between relative body size and skew. However, there is experimental evidence from one species of Polistes that skew is linked to the probability of escalated conflict, and that body size may not be a good predictor of fighting ability in wasps. A final possibility is that skew is determined by a simple convention, in which case we would not expect it to depend on variables such as relatedness. An important challenge for future empirical studies is to determine the direction of causality between skew and other behaviors, such as aggression and helping.
Introduction Vehrencamp’s (1983) theory of reproductive skew, elaborated by Reeve & Ratnieks (1993), assumed that a dominant individual has complete control over reproductive partitioning. The dominant may yield a reproductive “concession” to another individual in order to induce it to remain in the group. In assuming complete dominant control, Reeve & Ratnieks (1993) had in mind co-foundress associations of Polistes wasps. There are usually fewer than five individuals in these associations, and the dominant can monitor the entire nest, where all reproduction must occur, in a matter of seconds. If ever there was a scenario where complete control by the dominant seemed feasible, this was it. In this chapter, we review studies of reproductive skew in primitively eusocial insects, and evaluate the implications of the empirical results for reproductive-skew theory in general. We focus on primitively eusocial wasps because by far the most work has been carried out on them (but see Paxton et al. 2002, Langer et al. 2004 in bees). By primitively eusocial, we mean that, as in cooperatively breeding vertebrates, there are no obvious morphological differences between helpers (“workers”) and reproductives (“queens”), except sometimes in mean size. It is therefore reasonable to assume that any individual can potentially reproduce, as has been demonstrated through observation and sometimes experimental manipulation (e.g. Hughes et al. 1987, Field & Foster 1999, Strassmann et al. 2004). This raises the possibility that studies of tractable insect systems may provide insight into the factors affecting the evolution of skew in vertebrate societies. The chapter is divided into four sections. In the first section we describe a useful framework for classifying models of reproductive skew, based on the
Primitively eusocial wasps: current models concepts of battleground and resolution models of evolutionary conflict (Godfray 1995, Cant 2006). We discuss the implicit assumptions underlying the models and the relevance of these to biological systems. In the second section, we summarize the nesting biology of primitively eusocial wasps and review empirical studies of skew that have been conducted on them. One of the main empirical findings is that skew is typically high. In the third section we discuss possible explanations for this general pattern in the light of our theoretical framework. In the fourth section, we conclude with some remarks about the similarities and differences between insect and vertebrate systems, and the future directions for studies of reproductive conflict in both taxa.
Types of skew model Skew theory is an attempt to understand what is essentially an economic problem: how to share the profits of a cooperative association. The profits in question are the extra young that a group can raise compared to a solitary breeder. As with many economic problems, a good first step is to simplify the analysis by focusing on the interaction between two individuals only. A general feature of skew models is that they start by assuming an asymmetry between the players: one individual is labeled dominant and the other subordinate. This is reasonable in a biological model, because social animals usually do form hierarchies of some form for access to resources such as food or mates. Different skew models make different assumptions about the nature of the asymmetry that distinguishes the two individuals. Transactional models assume, albeit implicitly, a sequential structure to the game. That is, one player makes a “first move” which is observed by the other player before it decides on its response. The game then ends: in game theory these are known as one-shot sequential models or Stackelberg models (von Stackelberg 1934). The concession and peace-incentive models (Vehrencamp 1983, Reeve 1991, Reeve & Ratnieks 1993) assume that the dominant makes the first move, allocating reproductive shares to itself and the subordinate at zero cost. The subordinate then chooses whether to stay peacefully or disperse (in the concession model), or to fight for control of the nest (in the peace-incentive model). The role of first mover puts the dominant in a much stronger position than the subordinate, because the first mover can propose a division that is just acceptable to the subordinate (in other types of game, there may be a second-mover advantage: for example, where winning a conflict involves outbidding an opponent). By contrast, the restraint model (Johnstone & Cant 1999) allows subordinates to choose a division of reproduction first, after which dominants
307
308
J. Field, M. A. Cant can respond by evicting them from the group. Dominant status in this model is therefore defined as the ability to evict one’s opponent, rather than the ability to allocate reproductive shares. In contrast to these one-shot sequential games, the tug-of-war model of Reeve et al. (1998a) assumes that no player has the advantage of a first move: both players invest simultaneously in selfish acts to increase their share of reproduction, but dominant individuals are assigned an efficiency or strength advantage in their conflict with subordinates. Finally, the costly young model of Cant & Johnstone (1999) is an optimization rather than a game-theoretic model: dominants have full control over reproduction but can maximise inclusive fitness by sharing with related subordinates if the costs of offspring production rise with the number produced. The variety of model types and their assumptions about what dominance entails has hampered attempts to produce a coherent framework capable of accommodating all the various models (Johnstone 2000, Reeve 2000, Magrath et al. 2004, Buston et al. 2007). There are always models that must be left out, or treated as special cases (e.g. Cant & Johnstone 1999). This problem can be remedied, however, by borrowing some concepts from other areas of evolutionary theory that deal with the resolution of conflict between parties with shared interests. In particular the concept of battleground and resolution models in parent–offspring conflict theory is equally applicable to reproductive-skew theory, and can encompass models with different assumptions and informational structure (simultaneous, sequential, etc.). The distinction between the two types of model is explained below. Battleground models
Battleground models seek to define the zone of conflict between two parties over a limited resource. The limits of the battleground are found by solving for the best outcome of the conflict from the perspective of the first party, then the other one, assuming that each party can choose any division of the resource at zero cost. If there were no overlapping interests between the two players (i.e. the game were zero-sum), the best choice of partition would be to take all of the resource and leave the opponent with none. The insight from skew models, however, is that individuals in a cooperative association have a shared interest in group productivity. The level of this shared interest varies with genetic and ecological factors and acts to constrain the zone of conflict between the players. The defining feature of transactional skew models, for example, is that they incorporate “outside options,” that is, alternative strategies to peaceful cooperation that can be triggered if the share an individual is offered is too low. Even though a first mover can propose any division of
Primitively eusocial wasps: current models reproductive shares, it must still take into account the outside options of the other player when choosing its allocation. The response of the second mover acts as a threat to constrain the maximum share of reproduction that the first mover can take. In transactional models there are two types of constraint arising from these outside options: group-stability constraints set by the threat of departure or eviction, and peace constraints set by the threat of fighting by either party if its own share of reproduction falls too low (Figure 11.1A & B). A third set of constraints, which may overlay those of the transactional models, arises not because of outside options or threats, but because the law of diminishing returns can apply to reproduction as it does to other resources. For example, increased production of offspring will often entail accelerating costs (Cant & Johnstone 1999, Cant 2006). Under these circumstances each player has a kin-selected incentive to share reproduction with a relative, so there is nothing to gain from attempting to increase one’s own share above a threshold level (Figure 11.1C). These optimization constraints will be particularly important for vertebrates, because offspring are costly to produce and there are usually relatively stringent physical or physiological constraints on the number of young that can be produced. Resolution models
Rather than defining the zone of conflict, resolution (or compromise) models attempt to explore how conflict within it will be resolved. Resolution models assume that both players exert partial, costly control over the outcome, so that the result is a compromise between the best possible outcomes for dominants and subordinates. Unlike transactional models, in which one party gets to allocate shares to both individuals at zero cost, resolution models assume that “pulling” the division of reproduction in one’s favor involves direct fitness costs. They also assume that a player can choose its own level of effort in the competition, but not that of its opponent (Reeve et al. 1998, Cant 1998). The resolution will depend on the relative costs to each individual of a given level of competitive effort. The best-known resolution model is the tug-of-war (Reeve et al. 1998a, Johnstone 2000), in which dominants and subordinates can increase their own share of reproduction at a cost to group productivity. To put it another way, increased effort leads to a larger slice of a smaller “pie.” The tug-of-war provides a very broad framework to model the resolution of conflict over communal resources, and may be a particularly useful tool to help understand evolutionary transitions to cooperation across levels of organization (Reeve & Ho¨lldobler 2007). However, its very abstraction and generality make it rather difficult to test. Other models sacrifice some generality by
309
J. Field, M. A. Cant (A) Group-stability constraints Sub's share p
310
(B) Peace constraints
shold
re on th evicti
resh
g th
tin figh
Battleground staying ince
ntive
Relatedness
RR FF
pea
old
Battleground
ce in
cen
tive
Relatedness
(C) Optimization constraints ben ef (su icial s b co har ntro e l) Battleground hare ial s ol) efic r ben m cont (do
Relatedness
Figure 11.1 Defining the battleground of reproductive-skew theory. Three types of constraint may limit the degree to which one party can suppress the other. (A) Group stability constraints are set by the outside options available to the subordinate. Where dominants can choose any partition (at no cost), they will push the subordinate’s share p down to the staying incentive. Where subordinates can choose the partition, they will claim up to the dominant’s eviction threshold. (B) Peace constraints are set by the threat of the subordinate or dominant to fight if its share is reduced below a threshold. The solid-line and dotted-line constraints illustrate the bounds of the battleground when the outcome of fights is the subordination of the loser (role reversal, RR) or the death of the loser (fatal fight, FF), respectively. (C) Optimization constraints arise where an increasing reproductive share brings diminishing net fitness returns. In these circumstances an individual with choice over the partition can maximize its inclusive fitness by allocating a share to the other.
making specific assumptions about the behavioral mechanism through which individuals attempt to control reproduction: for example, elevated offspring production (Cant 1998), infanticide (Johnstone & Cant 1999, Hager & Johnstone 2004), or aggression (Reeve & Ratnieks 1993, Cant et al. 2006b). These models are useful because in addition to testing the predicted outcome of conflict, one can measure the behaviors that are assumed to reflect each party’s “effort” and compare these with the levels of effort predicted by the model. It is important to recognize some of the limitations of resolution models as models of behavioral conflict. The tug-of-war, like many other evolutionary models (e.g. the biparental care model of Houston & Davies 1985) solves for the evolutionarily stable combination of fixed effort levels. That is, the best effort levels given that neither player can observe and respond to the other on a behavioral time scale (McNamara et al. 1999, Cant & Shen 2006). In the conflict over reproduction, however, group members clearly do observe and respond to each others’ attempts to claim reproduction, for example by egg tossing or egg destruction (Vehrencamp 1977, Mumme et al. 1983), infanticide (Young & Clutton-Brock 2006), or acts of aggression (Reeve & Nonacs 1992,
Primitively eusocial wasps: current models Cant et al. 2006b). Because responsiveness is such a crucial determinant of the outcome of conflict, an important challenge for future theoretical work is to incorporate such responses in a biologically meaningful way. One promising approach (developed to study sexual conflict over parental care) is to solve for evolutionarily stable “rules for responding” rather than evolutionarily stable fixed efforts (McNamara et al. 1999). A similar approach to reproductive conflict may shed light on a raft of cooperative and agonistic behaviors that seem to be involved in negotiation or bargaining over reproduction. This type of model would represent an advance because it would help to identify and understand what the process of conflict resolution looks like in nature, and whether the outcome depends on the details of the bargaining process. Finally, there have been two notable attempts to produce a synthetic model which incorporates both group stability constraints and a tug-of-war within them (Johnstone 2000, Reeve & Shen 2006). Both of these models make somewhat arbitrary assumptions about how the presence of outside options influences the resolution process. The problem is that the sequence of decisions – the initial offer, the decision of whether to pursue an outside option, and how much to invest in selfish competition – is never made explicit, which makes it difficult to assess whether the models are plausible biologically, and whether the solutions obtained are sensitive to the assumed order of play. However, a very similar problem crops up in economic bargaining theory (Muthoo 2000), and theoretical work in this field suggests a simple general principle (the “outside option principle”) that may help to resolve this problem in skew theory. We have given an explicit account of the informational structure and underlying assumptions of different skew models – sequences of moves, responsiveness, etc. – because this information can help to distinguish between the models empirically. Transactional models, for example, assume that an individual will respond behaviorally if its share of reproduction falls below the threshold set by its outside option, and that this threat constrains the degree to which it can be exploited. Threats may be hard to detect, however, until the social rules are broken. For example, Wong et al. (2007) showed that the threat of eviction constrains subordinate growth in a goby size hierarchy by introducing fish that were closer in size to a dominant than is usually observed in nature: these “rule-breaking” fish were evicted. Similarly, to test whether threats of eviction, departure, or fighting constrain the level of skew it may be necessary to perturb the status quo by manipulating reproductive shares. A detailed understanding of model assumptions can help in both the design and the interpretation of future studies.
311
312
J. Field, M. A. Cant A second reason for making model assumptions explicit is as an aid to evaluating whether data are consistent with one or another model. Information on the biology and natural history of an organism can help to rule out some models as irrelevant or based on inappropriate assumptions, and suggest reasons for the fit, or lack thereof, between data and theory. With this aim in mind, we focus in the next section on empirical studies of skew in a particularly tractable taxonomic group, primitively eusocial wasps, and describe the key features of their nesting biology that can help to differentiate between the various models.
Testing reproductive skew theory in a model system: primitively eusocial wasps Primitively eusocial wasps are attractive experimental systems for the study of skew, with some key similarities to cooperatively breeding vertebrates. These typically include individual totipotency, small group sizes, and strong constraints on independent reproduction. Before applying our theoretical framework to studies of primitively eusocial wasps, we first outline the nesting biology of the wasps themselves. Nesting biology of primitively eusocial wasps
The primitively eusocial wasps that have been investigated in relation to reproductive skew are in the family Vespidae, subfamilies Polistinae (paper wasps, including Polistes) and Stenogastrinae (hover wasps, including Liostenogaster and Parischnogaster). Paper wasps and hover wasps probably represent two independent origins of eusociality (Hines et al. 2007). The genus Polistes includes more than 200 species that occur throughout most of the world (see Reeve 1991 for a review). In seasonal habitats where Polistes has been best studied, the nesting cycle begins in spring when overwintered females (foundresses) start building their characteristic paper nests attached to plants, rocks, man-made structures, etc. (Figure 11.2). Foundresses have already been inseminated, usually by a single male, soon after emerging from their natal nests the previous autumn. They store the sperm in a muscular sac, the spermatheca. Sperm can then be released throughout their lives, as required to fertilize eggs that will produce female offspring. Males are haploid and arise from unfertilized eggs. In some populations, almost all nests have only a single foundress, whereas in other populations some or almost all nests have more than one foundress, with 10 or more not infrequent in some populations of P. dominulus (Shreeves et al. 2003). On multiple-foundress nests, typically one “dominant” foundress lays most or all of the eggs, while the others (“subordinates”)
Primitively eusocial wasps: current models
Figure 11.2 Pre-emergence nest of Polistes dominulus attached to a cactus in southern Spain. The six foundresses are individually marked on the thorax with paint spots, and white silk caps are visible closing the cells that contain pupae. Photo by J. Field.
forage for insect prey, which is pulped up and fed to larvae. Where there is only a single foundress, she must carry out all tasks alone. Each additional foundress typically enables the group to rear more offspring, although per capita productivity is usually negatively correlated with group size. Larvae mature to adulthood in late spring/early summer, denoting the end of the “founding” or “pre-emergence” phase of the nesting cycle. Many of the first female offspring become workers on their natal nests, where they forage for larval provisions. From then onwards, the foundress typically ceases foraging and concentrates on egg laying. The workers help the foundress to rear further offspring, some of which are reproductives of both sexes. After mating with reproductives from other nests, the male reproductives die, and the females enter diapause to become the next year’s new foundresses. From the description so far, it may seem surprising that almost all studies to date have examined skew among offspring laid before worker emergence (e.g. Field et al. 1998a, Seppa¨ et al. 2002, Liebert & Starks 2006): many of these offspring become workers, to which skew theory may not apply. There are, however, two reasons why even the first brood of offspring may include reproductives. First, in a population of P. fuscatus, a proportion of the first brood apparently enter diapause to become foundresses the following year
313
314
J. Field, M. A. Cant (Reeve et al. 1998b; see also Starks 2001). This observation needs replicating in other species, but parallels earlier findings in the primitively eusocial bee Halictus rubicundus, which has a similar life cycle to temperate Polistes. In H. rubicundus, the proportion of maturing females that choose to enter diapause increases as the season progresses, until no further workers are produced (Yanega 1989). Early Polistes offspring have a second chance of becoming reproductive, via foundress replacement. In many populations, foundresses often die before the end of the nesting season, at which point workers take over egg laying (e.g. Strassmann 1981, Queller & Strassmann 1988). Every worker is therefore a potential replacement reproductive, which co-foundresses should compete to produce (Field et al. 1998a). The significant proportion of males among the first brood in some species presumably reflects the fact that some first-brood females will reproduce (e.g. Field et al. 1998a; but see Seppa¨ et al. 2002, Tsuchida et al. 2004, Liebert et al. 2005b). The second group of primitively eusocial wasps in which reproductive skew has been measured are the hover wasps (Stenogastrinae: see reviews in Turillazzi 1991, Field 2008). There are approximately 50 described species, all restricted to the southeast Asian–Papuan tropics. Hover wasps differ notably from Polistes in that brood rearing continues all year, with no winter diapause. Nests are usually founded by a single female, occasionally joined later by one or more others. Female offspring may remain on their natal nests as helpers, or leave to pursue other strategies such as founding new nests (e.g. Samuel 1987, Field et al. 1998b). Nests are small (< 100 brood-rearing cells: Figure 11.3) and group size is typically 1–4 females, very rarely exceeding 10: there is not the sudden increase in group size that occurs when the first brood reaches adulthood in Polistes. Although reproductive skew is usually high (see below), any female can eventually inherit the egg-laying position, so that all offspring are potential reproductives. Implications of wasp biology for reproductive conflicts
Two assumptions about wasp biology have been used to eliminate the areas of parameter space in Figure 11.1 that are likely to be irrelevant. First, optimization constraints may be relatively unimportant in wasps compared with vertebrates, so that net benefits will increase in a near-linear fashion with increasing reproductive share. This is because subordinates provide most of the costly provisioning effort, whereas the eggs laid by dominants, although not cost-free (Field et al. 2007), probably represent a smaller proportion of the total costs of reproduction than in vertebrates. In wasps, we therefore expect group-stability and peace constraints (Figure 11.1A & B) to be much more important than optimization constraints.
Primitively eusocial wasps: current models
Figure 11.3 Nest of the hairy-faced hover wasp (Liostenogaster flavolineata) attached to the underside of a bridge in Malaysia. The nest is made of mud, and the wasps are individually marked on the thorax with paint spots. Photo by A. Cronin.
A second assumption relates to the resolution of conflict within the battleground. Here it is usually assumed that dominants can exercise control over reproduction at little direct cost. This is because nests are small, and egg laying is a conspicuous activity: a female inserts her abdomen into a cell and remains more-or-less motionless for one or more minutes. Furthermore, dominants rarely leave the nest, so that they are probably the only individuals that can prevent their eggs being replaced by other members of the group. It seems reasonable, therefore, to assume that a dominant can make an “offer” to concede a share of reproduction to a subordinate, but not vice versa. Moreover, since dominants prevent unsanctioned reproduction by subordinates at little cost, we might expect a subordinate’s reproduction to be pushed down to the lower group-stability or peace constraint boundary, so that the predictions associated with this boundary apply (Figure 11.1). This has been the implicit assumption behind attempts to test skew theory in primitively eusocial wasps, leading to the familiar predictions from the concessions and peace-incentives models: skew should be positively correlated with genetic relatedness, group productivity, and the relative fighting ability of the dominant (Reeve & Ratnieks 1993). Below, we review tests of these predictions that have been conducted to date.
315
316
J. Field, M. A. Cant Studies of reproductive skew in primitively eusocial wasps
Microsatellite markers have been employed to test the predictions of reproductive-skew models in five species of Polistes and three species of hover wasp (Table 11.1). Studies have typically tested for relationships between skew and (1) helper–dominant genetic relatedness (n ¼ 8 studies), (2) helper– dominant body-size ratios, assumed to reflect differences in fighting ability (n ¼ 6), (3) group productivity (n ¼ 5), and, in Polistes, (4) season (n ¼ 3). Only one study has found strong support for the concession model: in P. fuscatus, skew was positively correlated with both relatedness and productivity, though not with body-size ratios (Reeve et al. 2000). None of the other studies found clear relationships between skew and relatedness, productivity, or body-size differences. The only consistent finding across studies is that in Polistes skew is significantly (n ¼ 2 studies) or almost significantly (n ¼ 1) greater among “late” (younger) offspring than “early” (older) offspring in pre-emergence nests (Field et al. 1998a, Reeve et al. 2000, Seppa¨ et al. 2002). The same pattern was reported by Peters et al. (1995) in P. annularis. It has been argued that this pattern supports concessions theory because ecological constraints increase during the season: there is a decrease in the time available for subordinate co-foundresses to initiate new nests and produce offspring before winter (Field et al. 1998a, Reeve et al. 2000, Seppa¨ et al. 2002). However, the tug-of-war model could make the same prediction if the relative fighting ability of subordinates decreases during the season because subordinates carry out more energy-expensive activities (foraging) than the dominant, and suffer repeated harassment from the dominant (Field et al. 1998a, Seppa¨ et al. 2002). In addition, skew could be lower among early offspring for reasons unrelated to social-contracts theory. Foundresses tend to switch nests during the first part of the pre-emergence phase, potentially leading to changes in dominance and periods of transition when dominance is in flux (Field et al. 1998a). Limitations of the data
Some features of studies to date could partially explain the lack of fit between data and models. First, sample sizes have been relatively small: 6–23 groups per study (Table 11.1). Statistical power to detect real relationships will therefore be low. Second, the validity of using body-size ratios as a surrogate for differences in fighting ability is unproven. In a recent study of P. dominulus, Cant et al. (2006b) found that relative body size did not predict when escalated fights over dominance occurred, or the duration or outcome of such fights. Third, there was little variation in genetic relatedness in some studies (e.g. Field et al. 1998a, Seppa¨ et al. 2002), but considerable variation in others (e.g. Queller et al. 2000, Fanelli et al. 2005, Liebert & Starks 2006). Even in
Primitively eusocial wasps: current models Table 11.1 Results from studies that have used microsatellites to investigate reproductive skew in primitively eusocial wasps. Gaps are where data were absent from the original source. % nests with skew <1
Dominants Species
larger than
Mean
and data source
subordinates? skew
Mean
(total
Offspring
co-foundress
sample size)
genotypeda
relatedness
Polistes P. bellicosus (Field et al. 1998a)
No
0.84S1
50% (14)
younger broodc
0.67
Yes
0.47P
87% (23)
older
0.57
b
P. fuscatus
female broodc
(Reeve et al. 2000) P. carolina
No
0.65S2
(17)
all broodc
0.64
0.56S1
50% (6)
all brood
0.13
0.88P
41% (17)
older broodc
0.25
(Seppa¨ et al. 2002) P. aurifer (Liebert et al. 2005a) P. dominulus (USA) (Liebert &
(0.21–0.43)
Starks 2006)d Hover wasps L. flavolineata
No
0.95S1
15% (13)
eggs and
No
0.87S2
21% (19)
eggs
0.33
No
0.92S1
22% (9)
eggs and
0.46
small larvae
(Sumner et al. 2002) P. mellyi
0.52
(Fanelli et al. 2005) P. alternata (Bolton et al. 2006) S1 P
corrected S index of Keller & Krieger (1997);
small larvae S2
S index of Pamilo & Crozier (1996);
proportion of offspring produced by the most productive foundress. All three skew indices
can range in value from zero (equal reproduction by all females) to 1.0 (reproduction monopolized by a single female). a
A smaller age-range of offspring genotyped provides a better estimate of skew at a particular
time, whereas a wider range is a better reflection of lifetime skew. b
Some data recalculated from Appendix B of Field et al. (1998a).
c
All offspring genotyped were laid before worker emergence.
d
Means for skew and relatedness in P. dominulus are from the 10 two-female nests in Liebert &
Starks’s (2006) study, calculated from their Figure 1. Data for 7 nests with >2 foundresses were not given. The bracketed range 0.21–0.43 is the range in mean co-foundress relatedness among two years in a separate Italian population of P. dominulus (Queller et al. 2000).
317
318
J. Field, M. A. Cant populations with significant variation, however, predicted relationships between skew and relatedness rely on foundresses being able to respond to the variation. There is little evidence for discrimination of relatedness at the individual level in social insects, except when relatedness is correlated with obvious cues such as offspring sex (Keller 1997). In Polistes, for example, foundresses may discriminate natal nest-mates as a class, which could include cousins as well as sisters, rather than responding to relatedness per se (Queller et al. 1990, Gamboa 2004; but see Queller et al. 2000 in relation to P. dominulus). Individuals that switch nests to join non-natal nest-mates might be particularly informative, because switching would represent a cue correlated with relatedness, but such joiners have not generally been identified and may usually be rare. A final difficulty with testing skew theory in primitively eusocial wasps follows from a major conclusion of the studies to date: skew at any one time is typically high, often uniformly high. This is especially true in the three studies of hover wasps, in which mean skew exceeded 0.85. Only 15–20% of nests exhibit any reproduction at all by subordinates (Table 11.1), representing only 2–4 nests in each study. With so little variation in skew, there was little opportunity to test for correlations between skew and potential explanatory variables. However, there was considerable variation in potential explanatory variables themselves, suggesting either that these variables do not determine skew in the predicted way, or that some unmeasured variable consistently takes values that lead to high skew across all groups irrespective of the values taken by measured variables (see Discussion, below, and Sumner et al. 2002, Bolton et al. 2006). Mean skew is generally lower in Polistes (Table 11.1), with 50% or more nests typically exhibiting some reproduction by subordinates. However, this difference may partly be because skew was usually measured across a wider range of offspring ages than in hover wasps (Table 11.1).
Discussion In primitively eusocial wasps, the dominant has been assumed to have complete control over reproduction at little or no cost. Yet empirical work fails to support the predictions of the concession model, with the exception of the study by Reeve et al. (2000). Shortcomings of the data could partly explain this apparent failure: small sample sizes, no direct measurement of adult fighting ability, and the fact that many offspring in pre-emergence nests are destined to become non-reproductive workers. There is also a lack of variation in genetic relatedness in some studies, although wasps may lack the ability to discriminate relatedness at the individual level. An additional feature of studies to date
Primitively eusocial wasps: current models has been that skew is typically high, often uniformly high across groups, perhaps especially in hover wasps (Table 11.1). We now discuss three possible reasons for this latter finding which fall within the framework of skew theory: (1) strong ecological constraints, (2) the possibility of subordinates inheriting the dominant position, and (3) the costs of escalated conflict. We also discuss the possibility that conventions, rather than social contracts, are the mechanism by which skew is determined. Ecological constraints
Like tests of other inclusive-fitness models in social insects, tests of skew models have tended to focus on the predicted effects of variation in relatedness. Yet two of the three parameters in the basic concessions model are ecological: the productivity of a potential subordinate if she chooses to nest independently, and the productivity of a group if the potential subordinate joins it – both measured relative to the productivity of a lone dominant. There is little evidence of strong physiological constraints in primitively eusocial wasps: almost any individual can potentially reproduce (Reeve 1991, Field & Foster 1999). Ecological constraints, however, do appear to be strong, providing a potential explanation for the high skews observed, and for the absence of correlations between skew and other variables. Constraints are strong because adults have short lives compared with the development periods of their offspring (Queller 1989). Offspring are helpless larvae, requiring progressive feeding and continuous adult protection in order to survive: the death of an independent nester typically leads to the failure of all of her part-reared offspring (Queller 1996, Field et al. 2000, Shreeves et al. 2003). Among 19 populations of polistines surveyed by Queller (1996), on average only 34% of independent nesters (range 0–62%) lived long enough to produce any surviving offspring. In the hairy-faced hover wasp (Liostenogaster flavolineata), fewer than 50% can expect to produce independent offspring (Samuel 1987, Field et al. 2000). Subordinate helpers may live no longer than independent nesters, but their investment can be preserved after their death through various forms of insurance (Queller 1996). For example, after a helper dies, the offspring that she contributed to are usually reared through to independence by her surviving nest-mates (Field et al. 2000, Shreeves et al. 2003, but see Tibbetts & Reeve 2003). Short lifespans for independent nesters compared with the offspring development period could help to explain why most subordinates are prepared to accept little or no direct reproduction. The importance of this life history for female reproductive decisions is implied by the positive correlation between independent nesting failure rates and the frequency of multiple foundress
319
320
J. Field, M. A. Cant associations across Polistes populations (Reeve 1991). Care is needed with this interpretation, however, because ecological constraints could cut both ways: if the dominant herself would have little reproductive success without helpers, she may be prepared to offer a large reproductive incentive to induce them to stay. However, this argument applies primarily to the first helper. While all helpers have the same expected payoff through independent nesting, from the dominant’s viewpoint each successive helper increases the probability of group survival by a smaller amount, and might thus receive a smaller incentive. Insurance mechanisms have been little investigated in cooperatively breeding vertebrates, but may be less important because adult lifespans are longer relative to offspring development time in vertebrates, and because vertebrate groups are unable to recycle excess offspring left after a carer dies (Queller 1996, Shreeves et al. 2003). In a penetrating review, Nonacs et al. (2006) went further by asking whether field estimates of survival and productivity in relation to group size are quantitatively consistent with observed levels of skew in 11 populations of primitively eusocial wasp. Although ecological data were consistent with high skew among close relatives (sisters, r ¼ 0.75 in haplodiploids), skews observed in Polistes were more extreme than what should be adaptive for more distant relatives such as cousins (r ¼ 0.1875). Almost all group members are sisters in some Polistes (e.g. Field et al. 1998a, Seppa¨ et al. 2002), but cousins are frequent in other populations (e.g. Reeve et al. 2000, Field et al. 2006), and unrelated cofoundresses occur commonly in P. dominulus (Queller et al. 2000, Liebert & Starks 2006). Nonacs et al. (2006) concluded that social-contracts models fail to predict patterns of skew in wasps, but we believe that this conclusion could need qualifying. In the next two sections, we discuss how two features of wasp biology might help to explain the high skews observed within the socialcontracts framework. First, however, we discuss aspects of the available data that might also explain the discrepancy. Eight of the 11 populations analyzed by Nonacs et al. were the same species, P. dominulus, the only species in which associations of unrelated co-foundresses are common (Queller et al. 2000). Of the other three populations analyzed, P. aurifer is also unusual in that multi-female groups are hardly more productive than independent nesters. Multi-female groups are indeed rare in P. aurifer, and it is perhaps no surprise if cooperation is not adaptive. Ecological data from the remaining two populations, P. fuscatus and L. flavolineata, were not inconsistent with observed skews. As noted above, co-foundresses in many Polistes species may respond to mean natal nest-mate relatedness rather than relatedness at the individual level (but see Queller et al. 2000). Mean relatedness in the study of P. fuscatus by Reeve et al. (2000) was 0.57, well above the
Primitively eusocial wasps: current models threshold required to explain even complete monopolization of reproduction by the dominant. In L. flavolineata, Nonacs et al. estimated group survival using 40 groups that were monitored for 2 months (Shreeves & Field 2002). Their calculations suggested that while subordinates could accept complete monopolization of reproduction by a dominant sister (r ¼ 0.75), they should require significantly lower skew if the dominant is a cousin (r ¼ 0.1875). In reality, dominants almost completely monopolize reproduction at any one time in L. flavolineata, yet cooperation among cousins is not infrequent (Sumner et al. 2002, Field et al. 2006). However, using a different dataset, consisting of survivorship data for individual females extrapolated over the 100-day offspring development period, Queller’s (1996) insurance-based model predicts that complete monopolization should be acceptable above a relatedness threshold of 0.21. This is close to the value expected for cousins and well below the mean observed relatedness of approximately 0.5 that is relevant if hover wasps cannot discriminate relatedness at the individual level (see Field et al. 2000, 2006). This highlights the fact that the ecological data available to Nonacs et al. (2006) were sometimes based on small samples and were originally collected for varying purposes, often from populations different from those where skew itself was measured. The synthesis by Nonacs et al. (2006) suggests that ecological constraints alone cannot explain the skews observed in some populations of P. dominulus, but this may not be true of primitively eusocial wasps in general. Inheritance
Strong ecological constraints may be one factor with the potential to explain the high skews seen in most primitively eusocial wasps. A second factor, which could act in concert, is inheritance. Subordinates that have a good enough chance of eventually inheriting an egg-laying position themselves may accept a high skew while they wait to inherit (Kokko & Johnstone 1999, Ragsdale 1999). The model of Kokko & Johnstone shows that incorporating inheritance greatly reduces the parameter space for which a subordinate requires a staying incentive to remain in the group (Figure 11.4). Inheritance has a similar effect in the peace-incentive model: subordinates have more to lose from risky fights and so are less likely to require a share of reproduction to deter them from challenging. Like cooperatively breeding vertebrates, primitively eusocial wasps typically live in small groups in which subordinates have a predictable chance of inheriting dominance by outliving the individuals ranked above them. Hard data are scarce, however: studies of Polistes have generally not been focused on inheritance and usually report its frequency within only one slice of the
321
J. Field, M. A. Cant 1 stable survivals
322
0.5 unstable
0 0
0.5 dispersal fitness, x
1
Figure 11.4 The influence of inheritance on the stability of two-player associations (redrawn from Kokko & Johnstone 1999). The graph shows regions for which twoplayer associations are stable or unstable as a function of dispersal fitness x and survival rate s in Kokko & Johnstone’s model. The example shown assumes that dominant and subordinate have equal survivorship and are related by coefficient 0.5. The diagonal contours indicate the region where the dominant must offer a staying incentive to maintain group stability. Contours show staying incentives which increase in magnitude from left to right in steps of 0.05. For comparison, the gray shaded area to the left indicates the sub-region for which groups can be stable if there is no possibility of inheritance by the subordinate (as assumed in the original concession model). Inheritance greatly increases the region of parameter space for which stable groups can form, and greatly reduces the need for staying incentives for them to do so.
nesting cycle. For example, Cant & Field (2001) found that 4/20 dominants were replaced before offspring emergence in P. dominulus, and Queller et al. (2000) report that 10% of subordinates could expect to inherit during that time (see also Nonacs et al. 2006). On 30% of nests of P. bellicosus and P. carolina, offspring genotypes indicated that most of the younger offspring were not produced by the foundress that had produced most of the older offspring (Field et al. 1998a, Seppa¨ et al. 2002, see also Peters et al. 1995). Similarly, 25% of P. fuscatus colonies had lost their original dominant foundress within 19 days of worker emergence (Reeve et al. 2000). These data do suggest that inheritance is common, but are too incomplete to provide robust estimates. Inheritance may be particularly important in the relatively aseasonal tropical environment of hover wasps, in which nests are perennial and waiting times are unconstrained by the arrival of winter (Field et al. 1999, Shreeves & Field 2002). Escalated conflict and peace constraints
In the queue to inherit an egg-laying position, it is the highest-ranking subordinates that have the best chance of surviving to inherit: those at lower
Primitively eusocial wasps: current models ranks have little chance (e.g. Field et al. 1999, Cant & Field 2001). Theoretically, this means that lower-ranked individuals would require a greater staying incentive to remain in the group than higher-ranked individuals. On the contrary, P. dominulus rank-2 subordinates exhibit much greater ovarian development than subordinates of lower rank (M. Cant & S. English, unpublished data). This suggests that low-ranked foundresses favor group membership over dispersal even though they receive little direct reproduction and have little chance of inheriting (Cant & English 2006). In these circumstances the threat of departure is an empty, non-credible threat, and cannot be used as leverage to obtain a share of reproduction. If subordinate reproduction is not explained by the threat of departure, what can account for variation in the level of subordinate reproduction in primitively eusocial wasps? One possibility that has been little studied to date is that reproductive shares reflect the threat of aggression or escalated conflict. Where fights result in the death of the loser, subordinates will remain in the group with little or no reproduction rather than risk a fight to the death with the dominant. Dominants, for their part, will allow a subordinate to claim a large share of reproduction before they are selected to fight. Wasps possess a deadly weapon in the form of a sting, so it is possible that the high skews observed in nature reflect the potentially lethal nature of fights over dominance. Where fights lead to the subordination rather than the death of the loser, however, the zone of conflict is much narrower. Subordinates will require greater peace incentives and dominants will have a lower fighting threshold. For any given level of dominant control, we would expect reproduction to be shared more evenly where fights are less risky (Figure 11.1B). Is there any evidence that dominants offer peace incentives to subordinates to avoid escalated fights? A recent study of P. dominulus lends support to the central assumption of the peace-incentive model that increased reproductive suppression should be associated with an elevated risk of escalated conflict. Cant et al. (2006b) induced conflict over dominance rank by temporarily removing dominant foundresses to allow the second-ranking female to inherit the nest. Once the replacement dominant was established, they replaced the original dominant and recorded the resulting interaction between the two wasps. Rank-2 subordinates with lower levels of ovarian development, and those that stood to inherit larger, more productive groups, were more likely to engage in escalated fights with the returning dominant (Figure 11.5). Relative body size, by contrast, had no effect on the probability of an escalated conflict. These results suggest that reproductive suppression will lead to an increased threat of escalated conflict, and hence that dominants can deter challenges by offering subordinates a share of reproduction. Interestingly, all of the escalated
323
J. Field, M. A. Cant
2.5 Subordinate ovarian score
324
2 1.5 1
0.5 0 2
4 3 6 7 5 Group size (productivity)
8
Figure 11.5 Results of the study by Cant et al. (2006b) showing that reproductive suppression is associated with an increase in the probability of escalated conflict in Polistes dominulus. Fights over dominant status were induced experimentally by removing the dominant for 3–8 days to allow the rank-2 subordinate to inherit, after which the original dominant was reintroduced. Closed circles are those rank-2 females that entered into an escalated contest with the returning dominant; open circles are those rank-2 females that immediately submitted. The solid line shows the regression for all rank-2 females. Both group size and the level of subordinate ovarian development have significant effects on the probability of an escalated contest. Also plotted as a dotted line is the non-significant regression of ovarian development in rank-1 individuals versus group size.
conflicts observed by Cant et al. (2006b) led to the subordination of the loser: foundresses apparently stopped short of employing stings in fights over dominance. This may be because most of the foundresses involved were full sisters, and so had a strong kin-selected incentive not to kill their opponents. Cant et al. (2006b) found that subordinate ovarian development increased with group size (and hence with productivity), consistent with the idea that dominants adjust the level of suppression according to the threat of escalated fighting. However, this is also the pattern expected if dominants lose reproductive control in larger groups (Clutton-Brock 1998, Field et al. 1998a). To test definitively whether dominants respond to the threat of escalated conflict by adjusting skew would require an experimental manipulation. For example, one could try to manipulate subordinate reproductive status to look for an effect on the probability of escalated conflict, or manipulate subordinate fighting ability to look for an effect on skew.
Primitively eusocial wasps: current models Aggression: negotiation or protest?
Many models of aggression assume that fights over dominance are all-or-nothing affairs leading to a specific outcome (Parker 1974, Reeve & Ratnieks 1993, Cant & Johnstone 2000, Cant et al. 2006a, 2006b). Much of the aggression observed in cooperative animal societies, however, is of a milder, non-lethal form. These low-level acts of aggression may reflect a process of negotiation or bargaining within the battleground of reproductive conflict. For example, aggressive displays may signal each party’s strength and motivation to enter into an escalated conflict over reproduction, allowing a resolution in terms of reproductive shares to be reached without escalation. An alternative to the hypothesis that skew is determined by the threat of aggression, however, is that the distribution of reproduction is determined in some way first, and levels of aggression reflect a subordinate’s response to this level of skew (e.g. Reeve & Ratnieks 1993). The first hypothesis assumes that the level of aggression acts to determine skew (as part of a “negotiation”), whereas the second assumes that skew determines the level of aggression (which takes the form of a “protest”). The issue of the direction of causality between skew and other behaviors such as helping and aggression is rarely discussed, but is extremely important for attempts to understand individual variation in helping behavior and aggression (Cant et al. 2006a), colony-level attributes such as stable group size and productivity (Cant & English 2006), and interspecific differences in social behavior. To date, a few studies have simply tested for a correlation between skew and aggression. There is some evidence that subordinates are less aggressive when skew is high (Field et al. 1998a, Seppa¨ et al. 2002), although results could be confounded by effects of activity level on aggression (Nonacs et al. 2004). Determining the direction of causality will often require more innovative experimental approaches to manipulate one factor (e.g. skew) and look for an effect on another factor (e.g. helping, aggression, or group size). Disturbing the status quo would also help to reveal whether behavior is shaped by threats, in the way that transactional models assume (Wong et al. 2007). The information gained from such tests would greatly advance our understanding of reproductive skew and social evolution in general.
Conventions
A final explanation for the lack of fit between models and data in primitively eusocial wasps, lying outside of the skew framework presented here, is that group members obey a simple convention, such as that the current dominant is the only egg-layer (Field et al. 1998a, Nonacs 2001, Seppa¨ et al. 2002). There is some evidence that the identity of the dominant is determined
325
326
J. Field, M. A. Cant conventionally in both Polistes co-foundress associations and in hover wasps (Seppa¨ et al. 2002, Bridge & Field 2007). Dominants are usually no larger on average than subordinates, and the dominant is frequently not the largest wasp on individual nests (Table 11.1). In the hairy-faced hover wasp, dominance is determined largely by relative age, which may represent an arbitrary convention (Bridge & Field 2007). Could skew itself be determined conventionally? A convention by which only the dominant reproduces would avoid competition over reproduction and the resulting costs to the group, as well as the sensory costs required for individuals to keep track of skew. Conventions might be particularly likely in situations where potential subordinates require little or no incentive to remain peacefully in the group, such as where subordinates are closely related to dominants and have little chance of successful reproduction alone, and where fights are costly (Nonacs 2001, Seppa¨ et al. 2002). Apart from the low relatedness among P. dominulus co-foundresses, these conditions may apply to the populations listed in Table 11.1. If skew is determined conventionally, we do not expect it to be correlated with variables such as relatedness (Nonacs 2001). Variation in skew might instead reflect dominant turnover and periods when dominance is unresolved after nest-switching, foundress death, etc. Arguing against skew conventions, however, are the patterns consistent with the concessions model reported by Reeve et al. (2000) in P. fuscatus, and perhaps also the frequent aggressive interactions observed in Polistes co-foundress associations, if these reflect negotiation over reproduction. Furthermore, although conventional mechanisms would explain the lack of fit between data and models that assume social-contracts mechanisms, they do not resolve the question of why subordinates sometimes appear to accept a higher skew than is adaptive. Concluding remarks: reproductive skew in insects and vertebrates Primitively eusocial insects such as paper wasps and hover wasps share a number of features with vertebrate systems, in which experimental manipulations are difficult and data are harder to collect. Perhaps most importantly, all group members retain the ability to reproduce, groups are usually small so that group membership typically offers substantial future fitness benefits, and there are usually stringent constraints on independent reproduction. The latter two features reduce the likelihood that group-stability constraints define the lower bound of the battleground over reproduction, since both inheritance and tight ecological constraints tend to reduce the required staying incentive to zero. In primitively eusocial wasps, therefore, as
Primitively eusocial wasps: current models in vertebrates, subordinates will often favor joining a social queue, even as a non-breeder, rather than attempting to breed independently. If the threat of departure is rarely credible (because subordinates prefer staying to dispersal), the outcome of reproductive conflict is unlikely to be sensitive to variation in the level of this non-credible threat. This may account for the finding that experimental manipulation of ecological constraints has little or no effect on patterns of skew in allodapine bees or cichlid fish (Langer et al. 2004, Heg et al. 2006). This result may also reflect a lack of information on the part of dominants, or subordinates, about the likely success of nesting attempts outside the group. In cooperative mammals, both dominants and subordinates will find it difficult to obtain information on the range of breeding opportunities outside the group, since territories are often contiguous and vigorously defended. Birds, by virtue of flight, will usually be in a better position to detect when vacancies arise outside their group, so that the resolution of reproductive conflict may be more sensitive to variation in outside options. Wasps can potentially obtain even better information: not only can they fly, but groups do not defend feeding territories. However, although this may apply to subordinates, dominant wasps rarely leave the nest, so that they may have no information about a subordinate’s options outside the group. A subordinate’s threat to disperse is credible only if the dominant can also detect the level of ecological constraint. A possible consequence of staying on the nest is that a dominant can commit to strategies that are insensitive to short-term changes in these constraints. While wasps may resemble vertebrates in some respects, there are also key differences in basic biology that will have important effects on the way in which conflict over reproduction is resolved. Perhaps the most important of these differences is the cost of producing young. In social insects, eggs are probably relatively cheap to produce (but see Field et al. 2007), and any overproduction of offspring can be recycled through oophagy (Mead et al. 1994, Shreeves et al. 2003). In birds and mammals, by contrast, the production of offspring represents a significant energy expenditure on the part of the parent, before any rearing costs are taken into account (Creel & Creel 1991, Monaghan & Nager 1997). Fish may represent an intermediate case (Heg et al. 2006). In birds and mammals, the marginal fitness benefits of offspring production will decline as more offspring are produced (this is the basis of Lack’s [1947] clutch-size argument). In these circumstances, additional offspring are expensive for a dominant but cheap for a subordinate: when the two are related, dominants have a kin-selected incentive to share reproduction. We have already described how the possibility of inheritance removes the need for staying incentives. In birds and mammals, diminishing returns on increasing offspring production
327
328
J. Field, M. A. Cant mean that the lower bound of the battleground over reproduction will often be defined by optimization constraints, or “beneficial shares” (Cant & Johnstone 1999; Figure 11.1C). Beneficial sharing is the simplest mechanism to account for reproductive skew, because it does not require any social contract or negotiation (Cant 2006). For this reason, in birds and mammals beneficial sharing should be the first explanation for reproductive sharing to be ruled out. Other key factors, such as the degree of reproductive control, will vary widely in both insect and vertebrate systems, and will depend strongly on the particular social biology of the species in question. For example, dominant paper wasps have been thought to exercise full control at no cost. However, the ability to suppress subordinates may be limited by the threat of a risky, escalated conflict, even though actual fights are rarely observed (Cant et al. 2006b). In birds and mammals, dominants often exercise control by killing the offspring of subordinates, but the efficacy of this threat will depend on their ability to distinguish a subordinate’s young from their own young and avoid retaliatory attacks from the victim. The ability to discriminate parentage will also vary widely between insect systems. Even subtle forms of reproductive control, such as the use of inhibitory pheromones, must ultimately be backed up by force in order to be evolutionarily stable (Keller & Nonacs 1993). The resolution of reproductive conflict by the use or threat of force will depend on the weaponry of the animals, the outcome of fights over dominance (e.g. whether the loser is killed, evicted, or subjugated), and the information each party has about the state and motivation of the other. The overlap between vertebrate and insect systems in many of these key features means that, despite radical differences in biology, tractable model insect systems can continue to play an important role in understanding the evolution of reproductive skew in other taxa. References Bolton, A., Sumner, S., Shreeves, G. E., Casiraghi, M., & Field, J. (2006). Colony genetic structure in a facultatively eusocial hover wasp. Behavioral Ecology, 17, 873–880. Bridge, C. & Field, J. (2007). Queuing for dominance: gerontocracy and queuejumping in the hover wasp Liostenogaster flavolineata. Behavioral Ecology and Sociobiology, 61, 1253–1259. Buston, P. M., Reeve, H. K., Cant, M. A., Vehrencamp, S. L., & Emlen, S. T. (2007). Conceptual synthesis of concessions and restraint models of reproductive skew. Animal Behaviour, 74, 1643–1654. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169.
Primitively eusocial wasps: current models Cant, M. A. (2006). A tale of two theories: parent–offspring conflict and reproductive skew. Animal Behaviour, 71, 255–263. Cant, M. A. & English, S. (2006). Stable group size in cooperative breeders: the role of inheritance and reproductive skew. Behavioral Ecology, 17, 560–568. Cant, M. A. & Field, J. (2001). Helping effort and future fitness in cooperative animal societies. Proceedings of the Royal Society of London B, 268, 1959–1964. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Cant, M. A. & Johnstone, R. A. (2000). Power struggles, dominance testing, and reproductive skew. American Naturalist, 155, 406–417. Cant, M. A. & Shen, S.-F. (2006). Endogenous timing in competitive interactions among relatives. Proceedings of the Royal Society of London B, 273, 171–178. Cant, M. A., Llop, J. B., & Field, J. (2006a). Individual variation in social aggression and the probability of inheritance: theory and a field test. American Naturalist, 167, 837–852. Cant, M. A., English, S., Reeve, H. K., & Field, J. (2006b). Escalated conflict in a social hierarchy. Proceedings of the Royal Society of London B, 273, 2977–2984. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Creel, S. R. & Creel, N. M. (1991). Energetics, reproductive suppression and obligate communal breeding in carnivores. Behavioral Ecology and Sociobiology, 28, 263–270. Fanelli, D., Boomsma, J. J., & Turillazzi, S. (2005). Multiple reproductive strategies in a tropical hover wasp. Behavioral Ecology and Sociobiology, 58, 190–199. Field, J. (2008). The ecology and evolution of helping in hover wasps (Hymenoptera: Stenogastrinae). In J. Korb & J. Heinze, eds., The Ecology of Social Evolution. Berlin: Springer, pp. 85–108. Field, J. & Foster, W. (1999). Helping behaviour in facultatively eusocial hover wasps: an experimental test of the subfertility hypothesis. Animal Behaviour, 57, 633– 636. Field, J., Solis, C. R., Queller, D. C., & Strassmann, J. E. (1998a). Social and genetic structure of paper wasp cofoundress associations: tests of reproductive skew models. American Naturalist, 151, 545–563. Field, J., Foster, W., Shreeves, G., & Sumner, S. (1998b). Ecological constraints on independent nesting in facultatively eusocial hover wasps. Proceedings of the Royal Society of London B, 265, 973–977. Field, J., Shreeves, G., & Sumner, S. (1999). Group size, queuing and helping decisions in facultatively eusocial hover wasps. Behavioral Ecology and Sociobiology, 45, 378–385. Field, J., Shreeves, G., Sumner, S., & Casiraghi, M. (2000). Insurance-based advantage to helpers in a tropical hover wasp. Nature, 404, 869–871. Field, J., Cronin, A., & Bridge, C. (2006). Future fitness and helping in social queues. Nature, 441, 214–217.
329
330
J. Field, M. A. Cant Field, J., Turner, E., Fayle, T., & Foster, W. A. (2007). Costs of egg production and offspring provisioning: multifaceted parental investment in a digger wasp. Proceedings of the Royal Society of London B, 274, 445–451. Gamboa, G. (2004). Kin recognition in eusocial wasps. Annales Zoologici Fennici, 41, 789–808. Godfray, H. C. J. (1995). Evolutionary theory of parent–offspring conflict. Nature, 376, 133–138. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Heg, D., Bergmuller, R., Bonfils, D., et al. (2006). Cichlids do not adjust reproductive skew to the availability of independent breeding. Behavioral Ecology, 17, 419–429. Hines, H. M., Hunt, J. H., O’Connor, T. K., Gillespie, J. J., & Cameron, S. A. (2007). Multigene phylogeny reveals eusociality evolved twice in vespid wasps. Proceedings of the National Academy of Sciences of the USA, 104, 3295–3299. Houston, A. I. & Davies, N. B. (1985). The evolution of cooperation and life history in the dunnock, Prunella modularis. In R. M. Sibly and R. H. Smith, eds., Behavioural Ecology: the Ecological Consequences of Adaptive Behaviour. Oxford: Blackwell, pp. 471–487. Hughes, C. R. & Strassmann, J. E. (1988). Foundress mortality after worker emergence in social wasps (Polistes). Ethology, 79, 265–280. Hughes, C. R., Beck, M. O., & Strassmann, J. E. (1987). Queen succession in the social wasp Polistes annularis. Ethology, 76, 124–132. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Cant, M. A. (1999). Indiscriminate infanticide and reproductive skew. Animal Behaviour, 57, 243–249. Keller, L. (1997). Indiscriminate altruism: unduly nice parents and siblings. Trends in Ecology and Evolution, 12, 99–103. Keller, L. & Krieger, M. J. B. (1996). Mating success of male birds. Nature, 308, 208– 209. Keller, L. & Nonacs, P. (1993). The role of queen pheromones in social insects: queen control or queen signal? Animal Behaviour, 45, 787–794. Kokko, H. & Johnstone, R. A. (1999). Social queuing in animal societies: a dynamic model of reproductive skew. Proceedings of the Royal Society of London B, 266, 571–578. Lack, D. (1947). The significance of clutch size. Ibis, 89, 302–352. Langer, P., Hogendoorn, K., & Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature, 428, 844–847. Liebert, A. E. & Starks, P. T. (2006). Taming of the skew: transactional models fail to predict reproductive partitioning in the paper wasp, Polistes dominulus. Animal Behaviour, 71, 913–923. Liebert, A. E., Nonacs, P., & Wayne, R. K. (2005a). Solitary nesting and reproductive success in the paper wasp Polistes aurifer. Behavioral Ecology and Sociobiology, 57, 445–456.
Primitively eusocial wasps: current models Liebert, A. E., Sumana, A., & Starks, P. T. (2005b). Diploid males and their triploid offspring in the paper wasp Polistes dominulus. Biology Letters, 1, 200–203. Magrath, R. D., Johnstone, R. A., & Heinsohn, R. G. (2004). Reproductive skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. McNamara, J. M., Gasson, C. E., & Houston, A. I. (1999). Incorporating rules for responding into evolutionary games. Nature, 401, 368–371. Mead, F., Habersetzer, C., Gabouriaut, D., & Gervet, J. (1994). Dynamics of colony development in the paper wasp Polistes dominulus Christ (Hymenoptera, Vespidae): the influence of prey availability. Journal of Ethology, 12, 43–51. Monaghan, P. & Nager, R. (1997). Why don’t birds lay more eggs? Trends in Ecology and Evolution, 12, 270–274. Mumme, R.L, Koenig, W. D., & Pitelka, F. A. (1983). Reproductive competition in the communal acorn woodpecker: sisters destroy each others’ eggs. Nature, 306, 583–584. Muthoo, A. (2000). A non-technical introduction to bargaining theory. World Economics, 1, 146–166. Nonacs, P. (2001). A life-history approach to group living and social contracts between individuals. Annales Zoologici Fennici, 38, 239–254. Nonacs, P., Reeve, H. K., & Starks, P. T. (2004). Optimal reproductive-skew models fail to predict aggression in wasps. Proceedings of the Royal Society of London B, 271, 811–817. Nonacs, P., Liebert, A. E., & Starks, P. T. (2006). Transactional skew and assured fitness returns models fail to predict patterns of cooperation in wasps. American Naturalist, 167, 467–480. Pamilo, P. & Crozier, R. H. (1996). Reproductive skew simplified. Oikos, 75, 533–535. Parker, G. A. (1974). Assessment strategy and the evolution of animal conflicts. Journal of Theoretical Biology, 47, 223–243. Paxton, R. J., Ayasse, M., Field, J., & Soro, A. (2002). Complex sociogenetic organization and reproductive skew in a primitively eusocial sweat bee, Lasioglossum malachurum, as revealed by microsatellites. Molecular Ecology, 11, 2405–2416. Peters, J. M., Queller, D. C., Strassmann, J. E., & Solı´s, C. R. (1995). Maternity assignment and queen replacement in a social wasp. Proceedings of the Royal Society of London B, 260, 7–12. Queller, D. C. (1989). The evolution of eusociality: reproductive head starts of workers. Proceedings of the National Academy of Sciences of the USA, 86, 3224–3226. Queller, D. C. (1996). The origin and maintenance of eusociality: the advantage of extended parental care. In S. Turillazzi and M. J. West-Eberhard, eds., Natural History and Evolution of Paper-Wasps. Oxford: Oxford University Press, pp. 218– 234. Queller, D. C. & Strassmann, J. E. (1988). Reproductive success and group nesting in the paper wasp Polistes annularis. In T. H. Clutton-Brock, ed., Reproductive Success: Studies of Individual Variation in Contrasting Breeding Systems. Chicago, IL: University of Chicago Press, pp. 76–96.
331
332
J. Field, M. A. Cant Queller, D. C., Hughes, C. R., & Strassmann, J. E. (1990). Wasps fail to make distinctions. Nature, 344, 388. Queller, D. C., Zacchi, F., Cervo, R., et al. (2000). Unrelated helpers in a social insect. Nature, 405, 784–787. Ragsdale, J. E. (1999). Reproductive skew theory extended: the effect of resource inheritance on social organisation. Evolutionary Ecology Research, 1, 859–874. Reeve, H. K. (1991). Polistes. In K. Ross & R. W. Matthews, eds., The Social Biology of Wasps. New York, NY: Cornell University Press, pp. 99–148. Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382. Reeve, H. K. & Ho¨lldobler, B. (2007). Emergence of a superorganism through intergroup competition. Proceedings of the National Academy of Sciences of the USA, 104, 9736–9740. Reeve, H. K. & Nonacs, P. (1992). Social contracts in wasp societies. Nature, 359, 823– 825. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflicts in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Shen, S.-F. (2006). A missing model in reproductive skew theory: the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Reeve, H. K., Emlen, S. T., & Keller, L. (1998a). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Reeve, H. K., Peters, J. M., Nonacs, P., & Starks, P. T. (1998b). Dispersal of first “workers” in social wasps: causes and implications of an alternative reproductive strategy. Proceedings of the National Academy of Sciences of the USA, 95, 13737–13742. Reeve, H. K., Starks, P. T., Peters, J. M., & Nonacs, P. (2000). Genetic support for the evolutionary theory of reproductive transactions in social wasps. Proceedings of the Royal Society of London B, 267, 75–79. Samuel, C. T. (1987). Factors affecting colony size in the stenogastrine wasp Liostenogaster flavolineata. Unpublished Ph.D. thesis, University of Malaya. Seppa¨, P., Queller, D. C., & Strassmann, J. E. (2002). Reproduction in foundress associations of the social wasp, Polistes carolina, conventions, competition, and skew. Behavioral Ecology, 13, 531–542. Shreeves, G. & Field, J. (2002). Group size and direct fitness in social queues. American Naturalist, 159, 81–95. Cant, M. A., Bolton, A., & Field, J. (2003). Insurance-based advantages for subordinate co-foundresses in a temperate paper wasp. Proceedings of the Royal Society of London Series B, 270, 1617–1622. Starks, P. T. (2001). Alternative reproductive tactics in the paper wasp Polistes dominulus with specific focus on the sit-and-wait tactic. Annales Zoologici Fennici, 38, 189–199.
Primitively eusocial wasps: current models Strassmann, J. E. (1981). Wasp reproduction and kin selection: reproductive competition and dominance hierarchies among Polistes annularis foundresses. Florida Entomologist, 64, 74–88. Strassmann, J. E., Fortunato, A., Cervo, R., et al. (2004). The cost of queen loss in the social wasp Polistes dominulus (Hymenoptera: Vespidae). Journal of the Kansas Entomological Society, 77, 343–355. Sumner, S., Casiraghi, M., Foster, W., & Field, J. (2002). High reproductive skew in tropical hover wasps. Proceedings of the Royal Society of London B, 269, 179–186. Tibbetts, E. A. & Reeve, H. K. (2003). Benefits of foundress associations in the paper wasp Polistes dominulus: increased productivity and survival, but no assurance of fitness returns. Behavioral Ecology, 14, 510–514. Tsuchida, K., Saigo, T., Tsujita, S., & Takeuchi, K. (2004). Early male production is not linked to a reproductive strategy in the Japanese paper wasp, Polistes chinensis antennalis (Hymenoptera: Vespidae). Journal of Ethology, 22, 119–121. Turillazzi, S. (1991). The Stenogastrinae. In K. Ross & R. W. Matthews, eds., The Social Biology of Wasps. New York, NY: Cornell University Press, pp. 74–98. Vehrencamp, S. L. (1977). Relative fecundity and parental effort in communally nesting anis. Science, 197, 403–405. Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian species. Animal Behaviour, 31, 667–682. von Stackelberg, H. (1934). Marktform und Gleichgewicht. Vienna: Springer. Wong, M. Y. L., Buston, P. B., Munday, P. L., & Jones, G. P. (2007). The threat of punishment enforces peaceful cooperation and stabilizes queues in coral-reef fish. Proceedings of the Royal Society of London B, 274, 1093–1099. Yanega, D. (1989). Caste determination and differential diapause within the first brood of Halictus rubicundus in New York (Hymenoptera: Halictidae). Behavioral Ecology and Sociobiology, 24, 97–107. Young, A. & Clutton-Brock, T. H. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387.
333
III
Resolving reproductive conflicts: behavioral and physiological mechanisms
12
Reproductive skew in female common marmosets: contributions of infanticide and subordinate self-restraint d a v i d h . ab b o t t , l e s l i e d i g b y , a n d we n d y s a l t z m a n
Summary The common marmoset (Callithrix jacchus) is a cooperatively breeding monkey that exhibits high reproductive skew among females. At the proximate level, this high skew is maintained, for the most part, by reproductive selfrestraint in subordinates, involving specialized behavioral and neuroendocrine responses to the presence of a dominant female. When subordinates terminate this self-restraint, however, dominant females frequently control subordinates’ reproductive attempts by killing their infants. Based on data collected over 20 years from both the field and the laboratory, we propose that such infanticide constitutes not only a proximate mechanism limiting subordinate females’ reproductive success, but also an ultimate mechanism favoring selection for reproductive self-restraint in subordinate females. Our hypothesis is consistent with both the commitment model of reproductive skew (Hamilton 2004), in terms of pre-conception restraint, and the discriminate infanticide model (Hager & Johnstone 2004), in terms of infanticide as a mechanism driving subordinate self-restraint. Parallel, long-term field and laboratory studies of common marmosets provide powerful interdisciplinary approaches enabling investigation of mechanisms regulating female reproductive skew at a proximate level, while providing novel insight into potential ultimate causation.
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
337
338
D. H. Abbott et al. Introduction Among primates, moderate female reproductive skew, manifest as high reproductive success among a limited number of adult females in a social group, is associated with social dominance in many species (Abbott et al. 2003). Extreme monopoly of reproduction by only one or two females, however, is restricted to most, but not all, members of a single primate subfamily, the Callitrichinae (the marmosets and tamarins). These species, especially the well-studied common marmoset (Callithrix jacchus), present an opportunity to integrate both proximate and ultimate explanations of reproductive strategies in order to better understand the evolution and mechanisms of reproductive skew. Two principal classes of theoretical models have been developed to explain reproductive skew in cooperatively breeding species: transactional models (e.g. concession or restraint models) and compromise models (e.g. tug-of war models; reviewed in Hager 2003, Johnstone 2000). One of the variables that distinguishes the different subclasses of these models involves the question of which individuals “control” reproduction. Do dominant females “concede” reproductive control in order to entice subordinates to delay dispersal and instead provide care for the dominant female’s offspring (e.g. Keller & Reeve 1994, Reeve 1998)? Or are dominant females simply unable to completely control subordinates’ reproductive activity, resulting in subordinate females engaging in a “tug-of-war” with dominants and breeding whenever dominant females are unable to prevent them from doing so (e.g. Clutton-Brock 1998, Reeve et al. 1998)? Notably, many skew models focus on the dominant breeders and their ability, or lack of ability, to control reproduction in subordinates. Reproductive self-restraint, or selfinhibition, however, may also play a role in determining reproductive skew (Snowdon 1996, Hager & Johnstone 2004). The commitment model, for example, proposes that subordinates “commit” to reproductive suppression in exchange for reduced harassment from dominant females (Hamilton 2004). In recent years, a number of investigators, including ourselves, have argued that not all aspects of diminished reproductive function in subordinate female callitrichines can be attributed to control by dominant individuals, but instead reflect adaptive, self-imposed restraint or commitment to self-inhibition (Snowdon 1996, Abbott et al. 1997, Saltzman 2003, Yamamoto et al. in press). In this chapter, we will describe patterns of reproductive skew in female common marmosets, and the proximate mechanisms that generate this skew, including both pre-conception mechanisms (e.g. suppression of female reproductive physiology and inhibition of sexual behavior) and post-conception mechanisms (infanticide) (Figure 12.1). Based on these findings, we suggest that while pre-conception reproductive suppression in female common marmosets is
Female common marmosets: infanticide and self-restraint
Figure 12.1 Diagrammatic illustration of proposed relationships between pre- and post-conception stages of female reproduction, and inhibitory social influences operating in common marmoset social groups. Social subordination may determine suppression of ovulation, inhibition of female sexual behavior, and, in plurally breeding groups, vulnerability to infanticide. Female sexual behavior is further constrained by inbreeding avoidance. Modified from Saltzman (2003).
mediated, mechanistically, by self-restraint or self-inhibition in subordinates, it likely evolved in response to a postpartum mechanism of dominant control of reproduction – infanticide – and may be best represented by modified commitment models of reproductive skew. Kin selection, due to the high degree of relatedness among female marmoset group-mates, may have intensified the development of such commitment to reproductive self-restraint.
Social groups of common marmosets Free-living groups
The small body size (< 0.5 kg) of common marmosets reflects an evolutionary process of specialization enabling efficient colonization of marginal and disturbed forest habitats, sometimes at high population densities, in the northern extreme of the Atlantic coastal forest and arid thorn scrub of northeastern Brazil (Ferrari 1993, Rylands & Faria 1993). With limited breeding and dispersal opportunities, it is not surprising to find highly developed forms
339
340
D. H. Abbott et al. of communal living and cooperative breeding (Keller & Reeve 1994, Lacey & Sherman 1997). In species with high reproductive skew, group size and composition clearly influence the costs and benefits of breeding. The presence of other breeding females, the chances of gaining a reproductive position within a group, and the availability of unrelated males may all play important roles in determining the reproductive strategies of a particular female. Common marmosets tend to live in groups of up to 20 individuals, including multiple adult males and females (reviewed in Digby et al. 2007). These groups, like those of most other social primates, are cohesive, with group members remaining within sight of one another over the course of the day and frequently grooming one another (Digby 1995a, Lazaro-Perea et al. 2004). Groups typically contain one or two breeding females (which may be close relatives that are unrelated to the breeding male(s)) along with their siblings and/or adult offspring (Nievergelt et al. 2000, Faulkes et al. 2003); however, population crashes and periods of instability can result in groups containing unrelated females (Lazaro-Perea et al. 2000, Faulkes et al. 2003). Though polyandry (in which two males share mating access to a breeding female) has been noted in some callitrichines, including common marmosets, the latter tend toward within-group monogamy and polygyny with frequent episodes of extragroup copulations (Digby 1999, Arruda et al. 2005; but see also Lazaro-Perea et al. 2000). Groups also include non-breeding adult individuals that are not sexually active within the group, but these non-breeders may participate in extra-group copulations (Digby 1999, Lazaro-Perea et al. 2000). Similar to many other species exhibiting high reproductive skew (Reeve & Emlen 2000), common marmosets tend to show low frequencies and intensities of intra-group aggression, although agonism can increase following changes in group composition (Digby 1995a, Lazaro-Perea et al. 2000; see also Sussman et al. 2005). Attainment of dominance mostly depends on rare instances of aggression when a breeding position has been vacated, while maintenance of dominance requires little intimidation or threat (French 1997, Lazaro-Perea et al. 2000). Intrasexual dominance relationships are, nonetheless, apparent. In the field, one breeding female is typically found to be dominant to a second breeding female (in plurally breeding groups), and all breeding individuals are dominant to non-breeders (Digby 1995a, Sousa et al. 2005). Under laboratory conditions, breeders are again dominant to non-breeders, but it is often impossible to detect a clear dominance relationship between breeding females (Saltzman et al. 1997b, 2004, 2008). There are several possible explanations for this apparent disparity between field and laboratory observations. First, dominance relationships in wild groups (e.g. Digby 1995a, Arruda et al. 2005, Sousa et al. 2005, Bezerra et al.
Female common marmosets: infanticide and self-restraint 2007, Lazaro-Perea et al. 2000) are often assessed largely on the basis of the direction and frequency of mildly aggressive behavior (e.g. threats, chases, piloerection), sexual behavior, and even affiliative behavior, in addition to submissive behavior (e.g. avoidance/withdrawal, cower), whereas those in laboratory groups have been assessed strictly on the basis of relatively subtle submissive behaviors (e.g. submissive vocalizations, grimaces; e.g. Saltzman et al. 2004), which may be difficult to detect in the field. Second, data on agonistic behavior in wild groups are often collapsed across extended time periods, whereas agonistic interactions in laboratory groups may be assessed more acutely. Because dominance relationships between breeding females may be altered transiently by the presence of infants (e.g. Alonso 1986, Roda & Mendes Pontes 1998, Saltzman 2003, Bezerra et al. 2007), laboratory studies may be better able to detect short-term changes in the existence or directionality of dominance relationships than has been possible in the field. Nonetheless, for the sake of simplicity, in this chapter we will continue to refer to “dominant” and “subordinate” breeding females, while acknowledging that these dominance relationships between female marmosets may be more complex and transitory than has often been presumed. In plurally breeding laboratory families, we refer to the older breeding female (mother) as the dominant female, in spite of the absence of distinct behavioral indices of dominant or subordinate status (Saltzman et al. 2004, 2008). Common marmoset mothers typically receive extensive help with carrying and feeding infants from all adult and subadult group members (Figure 12.2), beginning on the day of parturition (Digby 1995b, Yamamoto et al. 1996). Subordinate breeding females also receive such alloparental assistance, but cooperative care may not begin until the infants are several weeks old (Digby 1995b). Since subordinate breeding females are more protective of their infants, this delay in allo-parental care may result from potential alloparents having limited access to these infants, rather than from group members being unwilling to help. Laboratory groups
In captivity, common marmosets can be housed as (1) adult male– female pairs that usually establish families (Layne & Power 2003), (2) families in which the biological mother or father has been replaced by an unrelated adult of the same sex (Saltzman et al. 2004), (3) peer groups of adolescents or young adults (Hiddleston 1978, Abbott 1984), and (4) groups of unrelated adults (Saltzman et al. 1994). Apart from the case when the breeding male is replaced by an unrelated male (Saltzman et al. 2004), most captive groups contain a single breeding female that is usually behaviorally dominant to other female
341
342
D. H. Abbott et al.
Figure 12.2 Common marmosets usually exhibit high levels of infant care, as exemplified by the young infant carried on the back of an older sibling in this photograph. Our studies, however, show that dominant females practice infanticide on infants born to subordinate females when subordinates terminate reproductive self-restraint and produce their own infants. Courtesy of Judith Sparkles.
group members (Abbott 1984, Saltzman et al. 2004). Laboratory groups are thus well controlled in terms of size and the degree of relatedness of group members, and provide opportunities for experimental manipulation and elucidation of mechanisms that are not possible with free-living groups.
Proximate regulation of reproductive skew Reproductive failure in subordinate female common marmosets is potentially mediated at numerous stages of reproduction, occurring either before or after conception (Figure 12.1; French 1997). Possible pre-conception
Female common marmosets: infanticide and self-restraint mechanisms include inhibition of ovarian follicular development and ovulation, as well as inhibition of sexual behavior regardless of whether or not ovarian function is impaired. Post-conception reproductive failure potentially involves impairments in implantation, pregnancy maintenance, lactation, or maternal behavior (Saltzman 2003).
Inhibition of sexual behavior Inhibition of sexual behavior, in addition to suppression of ovulation (intermittent or absent ovarian cycles; see below), may contribute to reproductive skew in female common marmosets. Subordinate adult females typically engage in little or no intra-group sexual behavior, both in captivity (Rothe 1975, Abbott 1984, Saltzman et al. 1997c) and in the field (Digby 1999). This behavioral inhibition may result, at least in part, from intrasexual reproductive competition and may occur in response to cues from a dominant female group-mate. For instance, in laboratory groups of unrelated adults, dominant females may disrupt sexual interactions involving subordinate females (Abbott 1984) and thus impose a degree of direct reproductive control over unrelated females. Alternatively or additionally, inhibition of sexual behavior may be mediated by avoidance of mating with close relatives. Importantly, both of these mechanisms could contribute to inhibition of sexual behavior in free-ranging groups, since these groups may comprise closely related individuals as well as unrelated immigrants (see above; Ferrari & Digby 1996, Nievergelt et al. 2000, Faulkes et al. 2003). Several lines of evidence support a role for intrasexual inhibition of sexual behavior. First, subordinate females engage in little or no sexual behavior even in laboratory groups of unrelated adults (Rothe 1975, Abbott 1984). It can be argued, however, that in this relatively unnatural social context, subordination may activate behavioral and physiological mechanisms that evolved in the context of family groups, so that subordinate females may respond to unrelated male group-mates as though they were close relatives. Second, when subordinate females are removed from these groups of unrelated adults and briefly paired with an unfamiliar male, they solicit and accept mounts (Abbott et al. 1997). Third, females may disrupt one another’s sexual interactions in both singly (Epple 1967, Rothe 1975, Abbott 1984) and plurally breeding groups (Alencar et al. 2006). Several investigators, however, find little or no evidence of mating interference in plurally breeding groups (Kirkpatrick-Tanner et al. 1996, Lazaro-Perea et al. 2000). Finally, two genetic studies of free-ranging common marmoset groups reveal that some nonbreeding adult females are not closely related to the resident breeding male,
343
344
D. H. Abbott et al. suggesting that their reproductive activity may be inhibited by intrasexual competition alone (Nievergelt et al. 2000, Faulkes et al. 2003), as found in some laboratory groups containing unrelated marmosets (Abbott 1984, Saltzman et al. 2004). Inbreeding avoidance, in addition to rank-related inhibition, has been implicated as a critical determinant of sexual behavior in subordinate females. Common marmosets typically avoid sexual interactions with familiar, closely related individuals (reviewed by Saltzman 2003). In the field, genetic studies indicate that breeding males and females within a group are usually not closely related to one another (Nievergelt et al. 2000, Faulkes et al. 2003). In captivity, daughters housed with their natal family do not engage in sexual interactions with their father or brothers, even if they are undergoing ovulatory cycles and are not clearly subordinate to their mother or another female group-mate (Abbott 1984, Saltzman et al. 1997c, 2004). Many of these daughters, however, will readily solicit and copulate with an unrelated adult male introduced into the family (Kirkpatrick-Tanner et al. 1996, Saltzman et al. 1997b, 1997c, 2004; see also Anzenberger 1985, Hubrecht 1989). Notably, we find that when an unrelated male is incorporated into the family, daughters that do not engage in sexual behavior are behaviorally subordinate to their mother and/or a sister, whereas those that mate with the unrelated male are not (Saltzman et al. 2004). Thus, expression of sexual behavior by female common marmosets may be constrained both by the presence of a behaviorally dominant female and by lack of access to an unrelated adult male (Saltzman et al. 1997b, 1997c, 2004, Sousa et al. 2005). Subordinate females in free-ranging groups may be able to overcome both of these intra-group behavioral constraints by mating with extra-group males (Hubrecht 1984, Digby 1999, Lazaro-Perea 2001, Arruda et al. 2005). LazaroPerea (2001), for example, observed 20 extra-group copulations or attempted copulations, all of which involved non-breeding females from singly breeding groups. Digby (1999) observed 24 extra-group matings which involved both breeding and non-breeding adult females in plurally breeding groups. The functional significance of these interactions, however, is not clear. Even when subordinate females become pregnant following extra-group copulations, their infants are unlikely to survive (Arruda et al. 2005). In summary, female reproductive skew in common marmosets is mediated in part by inhibition of sexual behavior in subordinates, which in turn may result from both the absence of unrelated males in the group and the presence of a dominant female. Although subordinate females may potentially use extra-group copulations to circumvent these obstacles, the reproductive consequences of such copulations are not clear.
Female common marmosets: infanticide and self-restraint Suppression of ovulation Free-living groups
Ovulation suppression in subordinate females appears to be a key mechanism generating reproductive skew in common marmosets. Anovulation, as determined by analysis of circulating progesterone concentrations in frequent (e.g. semi-weekly) blood samples, was first described in subordinate female common marmosets housed in groups of unrelated adults by Abbott & Hearn (1978). Since then, ovulation suppression has also been documented in adult females living with their natal families in captivity (Abbott 1984, Saltzman et al. 1997a, 1997b, 2004), and, more recently, in free-ranging groups in Brazil (Albuquerque et al. 2001). The prevalence of ovulation suppression in free-living common marmosets is not yet clear. Long-term endocrine monitoring of several wild groups suggests that these groups may commonly contain two cycling or pregnant females (Albuquerque et al. 2001, Sousa et al. 2005); however, some of these reproductively active subordinates undergo transient periods of anovulation, as determined from patterns of fecal progesterone concentrations (Albuquerque et al. 2001). Moreover, because these field studies have simultaneously monitored ovarian activity in only two females per group, it is unknown whether other females (i.e. rank 3 and below) exhibit ovulatory cycles or are likely to undergo suppression of ovulation. Laboratory groups of intact families
In contrast to field studies, laboratory studies – including studies of both families and groups comprising unrelated adults (discussed below) – have consistently detected ovulatory suppression in subordinate female common marmosets. Many, but not all, adult daughters fail to ovulate while housed with their parents and siblings (Figure 12.3). Several investigators have found that in slightly more than half of the families studied, all daughters were anovulatory (Abbott 1984, Hubrecht 1989, Saltzman et al. 1997a, 2004). These daughters typically undergo a rapid onset of ovulatory cycles following their removal from the family, indicating that ovulatory function is inhibited by some aspect of the family environment (Abbott & Hearn 1978, Abbott 1984). Importantly, a sizable minority of families contains at least one daughter undergoing ovulatory cycles (Figure 12.3). These cycles are characterized by low, luteal-phase plasma progesterone concentrations and extended follicular phases, as compared to those of older females not housed within their natal families (Saltzman et al. 1997a). Typically, only the oldest daughter (or one of the oldest daughters, in the case of female–female twin pairs) in a family ovulates (Saltzman et al. 1997a); however, younger daughters may undergo
345
346
D. H. Abbott et al.
Figure 12.3 Percentage of mature female common marmosets that conceived (dark shading), ovulated but did not conceive (gray shading), or did not ovulate (white areas) in 6 intact families and 11 families in which the biological father was replaced by an unrelated adult male when the eldest daughters were post-pubertal. Modified from Saltzman et al. (2004). * P < 0.05 for percentage of pregnant oldest daughters when comparing natal families and families with an unrelated breeding male.
ovulatory cycles if they are behaviorally dominant to their older sisters (Saltzman et al. 2004). Why does a daughter ovulate in some families, but not in others? One answer appears to be the daughter’s relationship with her mother: daughters that exhibit overt submissive behavior toward their mothers are significantly more likely to be anovulatory than daughters that do not behave submissively (Saltzman et al. 1997c, 2004). Interestingly, these submissive and non-submissive females do not differ in the behaviors received from their mothers (Saltzman et al. 1997c), and aggression between mothers and daughters is typically mild and infrequent (Rothe 1975, Abbott 1984, Saltzman et al. 1997c, 2004). Thus, ovulation suppression is closely associated with a daughter’s perception of herself as subordinate to another female. Interactions with males, in contrast, do not reliably influence ovulatory activity in daughters (Figure 12.3). Whereas young adult females in several other species may be stimulated to commence ovulating by exposure to unrelated, unfamiliar males, introduction of a novel male into common
Female common marmosets: infanticide and self-restraint marmoset families does not significantly increase daughters’ likelihood of ovulating and does not alter the patterning or hormonal dynamics of their ovarian cycles (Saltzman et al. 2004; but see Saltzman et al. 1997b). Therefore, suppression of ovulation in young adult female common marmosets housed with their natal families is governed primarily by intrasexual, rather than intersexual, influences. Laboratory groups of unrelated adults
In laboratory groups of unrelated adults, suppression of ovulation is even more striking than in families: while all dominant females (rank 1) undergo ovulatory cycles, about 70% of all subordinate females (ranks 2 and 3) fail to ovulate. This means, however, that about 30% of subordinate females in groups of unrelated adults exhibit one or more ovulatory cycles. Such cycles can include deficient luteal phases (Abbott 1993), suggestive of (1) insufficient gonadotropin secretion from the anterior pituitary gland to support postovulatory corpora lutea, (2) deficient pituitary gonadotropin stimulation of preovulatory follicle development (Figure 12.4, Soules et al. 1989a,1989b, Ayabe et al. 1994), or (3) a combination of both. Circulating progesterone levels are nevertheless sufficient to support pregnancy in some subordinate females. Social contraception is most effectively maintained in lower-ranking subordinates (e.g. rank 3 and below): while over 40% of rank-2 females undergo one or more ovulatory cycles, only 14% of rank-3 females do so. Thus there appears to be a graded effect of social rank on ovulatory frequency: the lower the social rank, the greater the degree of ovulation suppression. Such socially determined anovulation in groups of unrelated adults is remarkably reliable, rapidly induced, and readily reversible. The onset of anovulation in subordinate females usually occurs within 14 days of group formation (Abbott et al. 1988, Abbott & George 1991). Subsequently, resumption of ovarian cyclicity following removal of subordinate females from their social groups, or following removal of all higher-ranking females, is equally rapid: the previously subordinate females usually ovulate within about 10 days (Abbott et al. 1997), approximately the duration of a normal follicular phase (Saltzman et al. 1994). The ovaries of anovulatory, subordinate female common marmosets have been found to contain no large, pre-ovulatory follicles, in contrast to the large follicles found in the ovaries of dominants (Harlow et al. 1986). Taken together, the ovarian and hormonal data from subordinate females suggest that follicular development is inhibited with remarkable precision, so that follicles mature only up to a stage normally achieved at the onset of the follicular phase of the ovarian cycle, but fail to progress when they reach the final stages of
347
348
D. H. Abbott et al.
Figure 12.4 Diagrammatic representation of neural and neuroendocrine mechanisms inhibiting ovulation in subordinate female common marmosets. Examples are shown of dominant females’ aggressive and scent-marking displays directed at female subordinates in laboratory groups of unrelated adults. Modified from Abbott et al. (1998).
maturation that are dependent on stimulation from pituitary gonadotropins (Figure 12.4). Any diminution in this inhibition, leading to a moderate increase in pituitary gonadotropin release, could well support the final stages of follicle growth, resulting in ovulation indistinguishable from that in dominant females. Such potentially dynamic responses of subordinate females may allow them to rapidly capitalize on acute changes in group composition or social status (Lazaro-Perea et al. 2000), typical of a marmoset’s opportunistic existence (Rylands & de Faria 1993, Digby et al. 2007). While rank-2 subordinate female common marmosets may be best placed in this regard, all adult females retain the capacity to rapidly ovulate and conceive when the social environment permits, such as following the death or disappearance of a breeding female, migration to a nearby group, or fissioning of the original group into smaller groups (Lazaro-Perea et al. 2000).
Female common marmosets: infanticide and self-restraint Neural and neuroendocrine mechanisms implicated in suppression of ovulation Elucidating the underlying physiological cause of anovulation in nonbreeding female common marmosets has been confounded by findings unique to marmosets and probably all other New World primates: common marmosets synthesize and release chorionic gonadotropin (CG) from the anterior pituitary gland (Gromoll et al. 2003, Muller et al. 2004a), rather than luteinizing hormone (LH), the functionally and structurally similar hormone expressed in the anterior pituitary of Old World primates and other mammals. Correspondingly, common marmosets express altered LH cell membrane receptors that are responsive to CG, but not to LH (Zhang et al. 1997, Muller et al. 2004b). CG has greater biopotency than LH, circulates for longer, and shows more prolonged release from the pituitary under stimulation from hypothalamic gonadotropin-releasing hormone (GnRH; Tannenbaum et al. 2007). Nonetheless, the biological role of pituitary CG remains similar to that of LH in supporting pre-ovulatory development of ovarian follicles, ovulation, and postovulatory formation and function of corpora lutea (Muller et al. 2004b). A series of laboratory experiments has suggested that ovulatory suppression in subordinate female common marmosets is mediated by the brain, rather than by primary ovarian or pituitary failure. Circulating levels of pituitary CG are reduced in anovulatory daughters (Abbott 1993) and subordinate females in groups of unrelated adults (Abbott et al. 1981), confirming inadequate gonadotropin stimulation of ovarian function (Figure 12.4). Experimentally induced elevations in circulating gonadotrophin levels in subordinate females result in increased circulating levels of estradiol and ovulation (Abbott 1993, Abbott et al. 1997), indicative of appropriate ovarian responses to gonadotropin stimulation and an apparent absence of ovarian failure. Repeated injections of subordinate females with supraphysiological doses (1–2 mg) of exogenous GnRH every 60–90 minutes stimulate pituitary release of CG, resulting in physiologically appropriate circulating levels of this hormone (Abbott 1989). Although such experimental evidence points to the neuroendocrine hypothalamus as playing an overriding role in the mechanism of anovulation in subordinate female common marmosets (Figure 12.4), pulsatile release of GnRH from the hypothalamus is unaltered in subordinate compared to dominant females (Abbott et al., 1998). Instead, pituitary CG responsiveness to GnRH is diminished, since injection of more physiological doses (50 ng) of GnRH fails to elicit CG release in subordinates, in contrast to the CG rise induced in identically treated dominants (Abbott et al. 1988). The mechanism responsible for such diminished pituitary
349
350
D. H. Abbott et al. responsiveness to GnRH remains elusive, but may involve other changes in neurosecretion from the hypothalamus. Such a mechanism, however, does not include typical physiological stress responses (Abbott et al. 1997, 1998). For example, while the stress response characteristically involves elevated secretion of glucocorticoids (e.g. cortisol) from the adrenal cortex (Sapolsky et al. 2000, Stewart 2003), no such increase is found in non-breeding daughters in laboratory families (Ziegler & Sousa 2002), subordinate females in laboratory groups of unrelated adults (Saltzman et al. 1994, 1998), or non-breeding females in free-living groups of common marmosets (Sousa et al. 2005). Other manifestations of hypothalamic-related dysfunction commonly associated with ovulation failure in other species, including elevated prolactin secretion by the anterior pituitary (Bowman et al. 1978), altered diurnal rhythms and reduced body weight (Yen 2004), are also not found in captive subordinate female common marmosets (Abbott et al. 1997). Instead, the anterior pituitary in these subordinates is so highly sensitized to negative feedback by estradiol that even the low levels of estradiol released by anovulatory ovaries suppress CG release (Figure 12.4; Abbott 1988). In addition, a mechanism independent of ovarian hormone feedback, but mediated by endogenous opioid peptides, also suppresses pituitary secretion of CG (Abbott 1993). These mechanisms operate without obvious decrease or alteration in endogenous hypothalamic release of GnRH (Abbott et al. 1998). Associative learning of olfactory, visual, and behavioral cues from the dominant female maintains neuroendocrine suppression of ovulation in subordinate female common marmosets (Abbott et al. 1997), indicating a psychological conditioning component to the ovulation-suppression mechanism. Non-breeding females utilize specific cues in their social environment to minimize breeding in the presence of specific dominant females (Figure 12.4). Such specialized mechanisms mean that only a familiar dominant female in a subordinate’s own social group carries salience for suppression of ovulation, so that the disappearance of the dominant female or emigration of the subordinate to another social group could readily extinguish this conditioned reproductive inhibition. Such a scenario permits a highly labile response of ovulatory function to changes in a female common marmoset’s social status, as illustrated by a variety of systematic manipulations of laboratory groups (Abbott 1984, 1989, Abbott & George 1991). Taken together, the highly specialized mechanisms maintaining suppression of ovulation in female common marmosets reflect reliable responses to social environments that are not conducive to successful reproduction by more than one female. Such adaptations suggest a high degree of selection for the evolution of multiple mechanisms minimizing a female’s
Female common marmosets: infanticide and self-restraint likelihood of breeding under inauspicious social conditions. Not surprisingly, intermittent or absent ovulatory cycles in subordinate females are also observed across a variety of other cooperatively breeding species ranging from primates to rodents (Solomon & French 1997), possibly due to analogous selection pressures.
Lack of evidence for post-ovulatory inhibition of female reproductive physiology Clearly, intermittent or absent ovulatory cycles and inhibition of sexual behavior in subordinate female common marmosets play a key role in generating reproductive skew. Nonetheless, in both wild and captive groups, numerous subordinate females ultimately terminate their reproductive selfrestraint and begin to breed. In the field, these subordinate breeding females produce significantly fewer infants than dominant breeding females and, most strikingly, produce few or no infants that survive beyond the first few weeks of life (Digby 1995b, Arruda et al. 2005, Sousa et al. 2005). What are the mechanisms that generate post-ovulatory reproductive failure in subordinate breeding females and maintain reproductive skew in plurally breeding groups? One answer is that young subordinate females undergoing ovulatory cycles may be somewhat less likely to conceive than older dominant females, even when they have access to a suitable mate. When resident breeding males are replaced by an unrelated adult male in laboratory families containing an adult daughter, nulliparous daughters undergoing ovulatory cycles begin to conceive, but at lower rates than their mothers, as determined by endocrine and ultrasonographic monitoring (Saltzman et al. 2004). It is unclear whether this fertility difference arises from differences in social status or in age/parity. Following conception, however, gestation, parturition, and lactation proceed unimpaired in these daughters. Primiparous daughters and their mothers produce comparable numbers of infants, at least in the laboratory, with both mothers and daughters exhibiting very low rates of spontaneous abortions and stillbirths (Saltzman et al. 2004, 2008). Furthermore, all daughters appear to lactate normally, and mothers and daughters carry their infants at similar frequencies (Saltzman et al. in press; but see Digby 1995b). Thus the presence of another breeding female might constrain other females’ likelihood of ovulating and conceiving but, following conception, does not appear to further influence pregnancy outcomes or maternal care. Female common marmosets, unlike other female primates, such as baboons (Papio cynocephalus: Wasser & Starling 1988), may not be sufficiently aggressive to one another to disrupt each other’s pregnancies.
351
352
D. H. Abbott et al. Infanticide by females The primary mechanism of post-conception control of reproduction in common marmosets appears to be the killing of other females’ infants (e.g. Digby 1995b, Saltzman 2003, Arruda et al. 2005, Digby & Saltzman in press). Eight infanticides have been directly observed in wild groups of common marmosets (Table 12.1; note that rates of infanticide are unavailable as total observation time is typically not reported), with seven of these cases occurring in groups containing two breeding females. Multiple infanticides have also been observed or, more commonly, inferred in captive groups containing two breeding females (e.g. Alonso 1986, Kirkpatrick-Tanner et al. 1996, Saltzman 2003, Saltzman et al. 2008). Observed infanticides in free-ranging and captive groups typically involve dominant breeding females killing infants born to subordinate breeding females, although at least two cases in wild groups were committed by subordinate breeding females (reviewed in Digby & Saltzman in press). In a laboratory study, however, infants of both dominant and subordinate breeding females (mothers and daughters, respectively, in families containing an unrelated adult male) were highly vulnerable to infanticide (Figure 12.5), especially if another female in the group was pregnant at the time of the infant’s birth (Figure 12.6; Saltzman et al. 2008). Of the various functional hypotheses put forth to explain infanticide (Hrdy 1979), the one that appears to be most relevant to infanticide by female marmosets is the resource competition hypothesis. This hypothesis posits that infanticide enables females to gain either immediate or future access to limited resources (Hrdy 1979; see Sherman 1981, Hoogland 1995 for examples in other mammals). The overall pattern of pregnant female common marmosets committing infanticide is consistent with this scenario: the perpetrator gains increased access to resources (e.g. helpers and food, including allonursing by females that have lost their young), decreased future competition, and even a reduced likelihood that its own infants will be killed by the other breeding female (reviewed in Digby 2000). Similar infanticidal patterns are observed across a variety of other cooperatively breeding species (e.g. meerkats, Suricata suricatta: Clutton-Brock et al. 2001; black-tailed prairie dogs, Cynomys ludovicia: Hoogland 1995; African wild dogs, Lycaon pictus: van Lawick 1974). Reproductive status appears to strongly influence female common marmosets’ likelihood of committing infanticide. In both wild and captive populations, infanticidal females are typically in the last 1–2 months of their own pregnancies (Table 12.1; Digby 2000, Saltzman 2003, Saltzman et al. 2008); only one postpartum female has been observed to kill another female’s infant
353
Table 12.1 Cases and contexts of infanticide directly observed in wild and captive groups of common marmosets Perpetrator’s relationship to Wild/Captive
Victim
Perpetrator
the victim
Timing/context of infanticide
Reference
Wild
2-week-old
Subordinate
Unknown
Subordinate female, which had been
Roda & Mendes Pontes 1998
male
breeding
challenging dominant female for several
female
months, attacked infant (on two occasions, the latter being fatal). Perpetrator gave birth one month later and became the dominant female of the group.
Wild
24-day-old female
Probably the
Not closely
Dominant female harassed subordinate female
dominant
related
after the subordinate gave birth. One twin
breeding
(R < 0.14)
disappeared. The infanticide of the second
female
Digby 1995a, Nievergelt et al. 2000
twin was directly observed, but the perpetrator could not be definitively identified, as several animals were in the immediate vicinity of the attack, including the dominant female, the male carrier of the infant, and a young adult male.
Wild
Three infants
Dominant
Grandmother
In case of male infant killed by grandmother,
(one male < 1
breeding
in one case,
perpetrator had given birth one month prior
month old;
female
no details for
to the infanticide.
details not given for others)
others
Yamamoto et al. 1996, Arruda et al. 2005, Alencar et al. 2006
354 Table 12.1 (cont.) Perpetrator’s relationship to Wild/Captive
Victim
Perpetrator
the victim
Timing/context of infanticide
Wild
Infant < 1
Dominant
Unknown
Infanticide occurred approximately one month Lazaro-Perea et al. 2000
month old
breeding
following group fission. Perpetrator gave birth
female
approximately one month after the
Reference
infanticide. Wild
Infant
Dominant
Unknown
Infanticide occurred during incursions by both
estimated to
female from
groups into an unoccupied territory. Infant
be < 1 month
neigh-
was cannibalized.
old
boring group
Melo et al. 2003
containing no infants Wild
1-month-old
Subordinate
female
breeding
Unknown
Infant was wounded when first observed. Mother tried to recover infant, but it
female
repeatedly fell. Subordinate female then
Bezerra et al. 2007
attacked and consumed part of the infant. The perpetrator gave birth two months later, by which time she had become the dominant female. Captive
Two infants
Newly
Sister
The daughter of the group’s original dominant
dominant
breeding female became pregnant and
female
subsequently became the dominant female. The newly dominant female subsequently killed two of her infant sisters.
Alonso 1986
355 Captive
1-day-old infant Dominant
Grandmother
breeding
day after birth and cannibalized the carcass.
female Captive
18-day-old male Primary
Grandmother grabbed and killed infant on first Kirkpatrick-Tanner et al. 1996 Perpetrator gave birth one week later.
Grandmother
Perpetrator behaved submissively to the infant’s Saltzman 2003, Saltzman et al.
breeding
mother the day before the infant’s birth. Two
female
days after the birth, the perpetrator carried and ate the carcass of one of the twins, although infanticide was not observed. Perpetrator attacked and severely wounded the second twin at 18 days of age, necessitating euthanasia. Perpetrator gave birth 7 days later, and no further submissive behavior between the two females was observed.
2008
356
D. H. Abbott et al.
Figure 12.5 Percentage of infant common marmosets that survived to at least 1 year of age (dark shading), died spontaneously (gray shading), or were killed (white areas) in 6 intact families and 11 families in which the biological father was replaced by an unrelated adult male when the eldest daughters were post-pubertal. Infanticide was more common in groups containing an unrelated breeding male, but this trend was not significant. Modified from Saltzman et al. (2008).
(Arruda et al. 2005). On a functional level, such precisely timed infanticide reduces competition for helpers to assist with rearing the perpetrator’s soon-to-be-born offspring (Digby 2000, Saltzman 2003, Saltzman & Abbott 2005), while eliminating the possibility that an infanticidal female will accidentally kill her own young. On a proximate level, the hormones of late pregnancy are likely to influence females’ responses to infants. In a laboratory study, multiparous female common marmosets exhibited minimal attraction to and tolerance of infants during the late stages of pregnancy, especially as compared to early pregnancy or the early postpartum period (Saltzman & Abbott 2005). In addition, early-postpartum females showed identical behavioral responses to their own infants and to unfamiliar, unrelated infants, consistent with the possibility that these females may be unable to discriminate reliably among infants and that infanticidal postpartum females would therefore risk killing their own offspring (Saltzman & Abbott 2005). As noted above, subordinate breeding females suffer significantly higher rates of infant loss than do dominants (Table 12.1, Figure 12.5), although the relative contributions of infanticide and other causes of mortality are not
Female common marmosets: infanticide and self-restraint
Figure 12.6 Percentage of infant common marmosets that survived to 1 year of age (dark shading), died spontaneously (gray shading), or were killed (white areas) associated with whether or not another female was pregnant in the group when parturition occurred. Data are combined from 6 intact families and 11 families in which the biological father was replaced by an unrelated adult male when the eldest daughters were post-pubertal. Modified from Saltzman et al. (2008). ** P < 0.002 for percentage of infants killed in families that contained a pregnant female compared to families that did not.
known (Digby 1995b, Arruda et al. 2005). As a consequence of the threat of infanticide and the low probability of successfully raising young, females living with a dominant breeding female may well benefit from delaying reproduction and therefore avoiding investment in costly reproductive attempts that are unlikely to succeed (Wasser & Barash 1983).
Conclusion Findings from two decades of field observations and three decades of laboratory studies of common marmosets are providing unique insights into the causes of female reproductive skew in this cooperatively breeding primate. Although groups may frequently contain three or more adult females, no more than two females produce offspring per group and, at least in laboratory groups, this is frequently reduced to a single female. Social suppression of ovulation in behaviorally subordinate females is more pronounced in laboratory groups than in free-living groups, while inhibition of intra-group sexual behavior is prevalent among non-breeding females in both captive and freeliving populations. Intermittent or absent ovulatory cycles in subordinate females are not mediated by either generalized stress or overt aggression from
357
358
D. H. Abbott et al. the group’s dominant female. Instead, non-breeding subordinates exhibit highly specialized physiological responses to a variety of cues from dominant females, which diminish the subordinates’ ability to ovulate (Figure 12.4). Inhibition of intra-group sexual behavior by females, on the other hand, is regulated both by intrasexual dominance relationships and by access to unrelated males (Figure 12.1). With the presence of one or more unrelated males in the group, two females can breed concurrently, resulting in these females killing one another’s infants with notable frequency. Infanticide particularly impairs the reproductive success of the subordinate breeding female and is especially costly for marmosets, given their heavy investment in each breeding attempt (e.g. long gestation and high infant-to-maternal body mass ratio: Tardif 1997). Together, suppression or disruption of ovulatory cycles, inhibition of sexual behavior, and infanticide by breeding females generate high reproductive skew among female common marmosets within each social group. It is not yet known, however, whether such skew results in diminished lifetime reproductive success for subordinate females, or whether it merely reflects temporary postponement of reproduction until a (dominant) breeding position is attained. In addition to seeking out or waiting to obtain a breeding position, subordinate females may become increasingly likely to attempt to breed, even in the presence of a dominant female, with increasing age, as has been reported in cooperatively breeding golden lion tamarins (Leontopithecus rosalia: Baker et al. 2002), dwarf mongooses (Helogale parvula: Creel & Waser 1997), and meerkats (Young et al. 2006). Such a scenario would be consistent with laboratory findings of age-related increases in likelihood of ovulation in mature daughters remaining in their natal families, as well as age-related increases in aggressiveness towards unrelated adult females (Saltzman et al. 1996, 1997a). According to Wasser and Barash’s (1983) reproductive suppression model, age-related increases in reproductive attempts, even under suboptimal conditions, should be beneficial to subordinate females because of the diminishing likelihood of obtaining a dominant breeding position in their dwindling lifespan (see also Williams 1966). The low rates and low intensity of aggression among female group-mates, the absence of stress-related physiological changes in subordinates, and subordinates’ use of subtle behavioral and sensory cues from dominant females to trigger suppression of ovulation, all suggest that pre-conception reproductive inhibition in non-breeding female common marmosets is not maintained by a high degree of harassment from dominant females or socially induced stress, and is not actively imposed on subordinates by dominants (at a proximate level). Instead, such inhibition may be maintained by specialized, presumably
Female common marmosets: infanticide and self-restraint adaptive, neuroendocrine responses of subordinate females to infrequent and mild aggression from a dominant female group-mate, resulting in self-imposed inhibition of ovulation and sexual behavior. A key question, however, involves the adaptive significance of such self-restraint in the presence of a dominant female. Why would healthy adult females benefit from curtailing their own reproduction over periods of months or years, even when food resources and potential alloparents may be abundant? The answer may lie, in large part, in infanticide committed by breeding females (e.g. Hager & Johnstone 2004). If this pattern of infanticide evolved before or simultaneously with ovulation suppression, it may well have played a critical role in selecting for the evolution of specialized neuroendocrine and behavioral mechanisms to reduce the likelihood that females would breed in the presence of a dominant, potentially infanticidal, female group-mate (Saltzman 2003). Such specialized responses limiting pre-conception events would minimize subordinates’ likelihood of investing in reproductive attempts that are unlikely to succeed, possibly increasing their prospects for survival and successful reproduction in the future (Wasser & Barash 1983, Jaquish et al. 1991, Digby 1995b, Saltzman 2003, Gilchrist 2006). We propose that dominant female common marmosets do not or cannot completely control reproductive attempts by subordinate females at the preconception or pre-parturition stages (Figure 12.1). Instead, they use infanticide as a proximate mechanism to limit subordinate females’ reproductive success. In so doing, dominant females may also secondarily engage infanticide as an agent of selection favoring commitment to reproductive self-restraint in subordinates. Thus, reproductive skew in this species is generated predominantly, in a proximate sense, by self-restraint in subordinate females but, ultimately, by dominant control over subordinates. This pattern is consistent with both the commitment model (Hamilton 2004), in terms of pre-conceptive restraint, and the discriminate infanticide model (Hager & Johnstone 2004), in terms of infanticide as a mechanism driving subordinate self-restraint. Since infanticide among common marmosets may commonly involve killing of closely related kin, the stakes would seem to be extremely high for dominant females in maintaining reproductive sovereignty.
Female reproductive skew in common marmosets: comparative aspects and future directions Reproductive skew in female common marmosets closely resembles that found in several other mammalian cooperative breeders (Abbott et al. 1998), including other callitrichines (French 1997), meerkats (Clutton-Brock
359
360
D. H. Abbott et al. et al. 2001), dwarf mongooses (Creel & Waser 1997), African wild dogs (van Lawick 1974), and Damaraland mole-rats (Cryptomys damarensis: see Chapter 13). Like marmosets, these mammals live in groups comprising a mixture of close kin and unrelated individuals, and subordinate females occasionally attempt to breed in spite of the existence of behavioral and physiological mechanisms of reproductive self-restraint. Also similar to common marmosets, dominant female meerkats and dwarf mongooses sometimes kill infants born to subordinate females (Creel & Waser 1997, Clutton-Brock et al. 2001), especially when pregnant (meerkats: Young & Clutton-Brock 2006). This convergence in mechanisms restricting subordinate female reproduction is striking, given the phylogenetic distances between the mammalian taxa involved. The convergence may therefore reflect adaptation to environmental conditions that make dispersal and independent breeding potentially costly. Comparative biological approaches across mammals, such as those taken by Faulkes & Bennett (2001) with African mole-rats, are needed to elucidate commonalities across different habitats that may select for singular cooperative breeding and high reproductive skew among females. Further insights into proximate mechanisms regulating reproductive skew in female common marmosets are likely to continue to emerge from integrated research programs focusing on (1) neural and neuroendocrine mechanisms mediating ovulation suppression, (2) neuroendocrine mechanisms underlying infanticide, and (3) organismic and environmental factors influencing individual females’ “decisions” to either restrain their own reproduction or attempt to conceive and successfully rear offspring. To clarify ultimate causation of reproductive skew, future research must identify the factors imposing strict limitations on numbers of breeding females in common marmoset groups such that breeding females are driven to kill infants born to female group-mates, regardless of whether the infants are close kin. Our current progress in these areas has come from complementary behavioral and physiological studies from both the field and the laboratory. For example, field observations of plural breeding and infanticide (e.g. Digby 1995b) were reliably reproduced in the controlled setting of the laboratory (Saltzman et al. 2004, 2008). Conversely, robust laboratory findings of inhibition of sexual behavior and ovulation in subordinate females (Abbott et al. 1997) have led to incorporation of physiological and molecular genetics components in long-term field studies (Faulkes et al. 2003, Sousa et al. 2005). Such interactive, collaborative, and long-term field and laboratory research approaches have proved highly successful in identifying proximate mechanisms regulating reproduction in female common marmosets, and will continue to refine our understanding of the risks and
Female common marmosets: infanticide and self-restraint choices female marmosets take in attempting to maximize their overall reproductive success.
Acknowledgments We thank Clara B. Jones, Reinmar Hager, and two anonymous reviewers for constructive comments on an earlier draft of the manuscript. We are grateful to the many staff members of our respective laboratories and institutions for their multiple contributions to the work reported here, and to Scott J. Muller for assistance in figure and manuscript preparation. This work was supported by NIH grants HD007678, MH011417, MH060728, MH053709, and RR000167, and NSF grants IBN92-21771 and IBN96-04321, and was partly conducted at a facility constructed with support from Research Facilities Improvement Program grant numbers RR015459-01 and RR020141-01.
References Abbott, D. H. (1984). Behavioral and physiological suppression of fertility in subordinate marmoset monkeys. American Journal of Primatology, 6, 169–186. Abbott, D. H. (1988). Natural suppression of fertility. In J. P. Hearn & G. R. Smith, eds., Reproduction and Disease in Captive and Wild Animals. Symposium of the Zoological Society of London 60. Oxford: Oxford University Press, pp. 7–28. Abbott, D. H. (1989). Social suppression of reproduction in primates. In V. Standen & R. A. Foley, eds., Comparative Socioecology: the Behavioural Ecology of Humans and Other Mammals. Oxford: Blackwell, pp. 285–304. Abbott, D. H. (1993). Social conflict and reproductive suppression in marmoset and tamarin monkeys. In W. A. Mason & S. P. Mendoza, eds., Primate Social Conflict. Albany, NY: SUNY Press, pp. 331–372. Abbott, D. H. & George, L. M. (1991). Reproductive consequences of changing social status in female common marmosets. In H. O. Box, ed., Primate Respones to Environmental Change. London: Chapman & Hall, pp. 295–309. Abbott, D. H. & Hearn, J. P. (1978). Physical, hormonal and behavioural aspects of sexual development in the marmoset monkey, Callithrix jacchus. Journal of Reproduction and Fertility, 53, 155–166. Abbott, D. H., McNeilly, A. S., Lunn, S. F., Hulme, M. J., & Burden, F. J. (1981). Inhibition of ovarian function in subordinate female marmoset monkeys (Callithrix jacchus jacchus). Journal of Reproduction and Fertility, 63, 335–345. Abbott, D. H., Hodges, J. K., & George, L. M. (1988). Social status controls LH secretion and ovulation in female marmoset monkeys (Callithrix jacchus). Journal of Endocrinology, 117, 329–339. Abbott, D. H., Saltzman, W., Schultz-Darken, N. J., & Smith, T. E. (1997). Specific neuroendocrine mechanisms not involving generalized stress mediate social
361
362
D. H. Abbott et al. regulation of female reproduction in cooperatively breeding marmoset monkeys. Annals of the New York Academy of Sciences, 807, 219–238. Abbott, D. H., Saltzman, W., Schultz-Darken, N. J., & Tannenbaum, P. L. (1998). Adaptations to subordinate status in female marmoset monkeys. Comparative Biochemistry and Physiology, 119, 261–274. Abbott, D. H., Keverne, E. B., Bercovitch, F. B., et al. (2003). Are subordinates always stressed? A comparative analysis of rank differences in cortisol levels among primates. Hormones and Behavior, 43, 67–82. Albuquerque, A. C. S. R., Sousa, M. B. C., Santos, H. M., & Ziegler, T. E. (2001). Behavioral and hormonal analysis of social relationships between oldest females in a wild monogamous group of common marmosets (Callithrix jacchus). International Journal of Primatology, 22, 631–645. Alencar, A. I., Sousa, M. B. C., Abbott, D. H., & Yamamoto, M. E. (2006). Contested dominance modifies the anovulatory consequences of social subordination in female marmosets. Brazilian Journal of Medical and Biological Research, 39, 647–658. Alonso, C. (1986). Fracasso na inibic¸a˜o da reproduc¸a˜o de uma feˆmea subordinada etroca de hierarquia em um grupo familiar de Callithrix jacchus jacchus. Primatologia no Brasil, 2, 203. Anzenberger, G. (1985). How stranger encounters of common marmosets (Callithrix jacchus jacchus) are influenced by family members. Folia Primatologica, 45, 204–224. Arruda, M. F., Araujo, A., Sousa, M. B.C., et al. (2005). Two breeding females within free-living groups may not always indicate polygyny: alternative subordinate female strategies in common marmosets (Callithrix jacchus). Folia Primatologica, 76, 10–20. Ayabe, T., Tsutsumi, O., Momoeda, M., et al. (1994). Impaired follicular growth and abnormal luteinizing hormone surge in luteal phase defect. Fertility and Sterility, 61, 652–656. Baker, A. J., Bales, K., & Dietz, J. M. (2002). Mating systems and group dynamics in lion tamarins. In D. G. Kleiman & A. B. Rylands, eds., Lion Tamarins: Biology and Conservation. Washington, DC: Smithsonian Institution Press, pp. 188–212. Bezerra, B. M., Souto, A. S., & Schiel, N. (2007). Infanticide and cannibalism in a freeranging plurally breeding group of common marmosets (Callithrix jacchus). American Journal of Primatology, 69, 1–8. Bowman, L. A., Dilley, S. R., & Keverne, E. B. (1978). Suppression of oestrogeninduced LH surges by social subordination in talapoin monkeys. Nature, 275, 56–58. Brandon-Jones, D. & Groves, C. P. (2002). Neotropical primate family-group names replaced by Groves (2001) in contravention of article 40 of the International Code of Zoological Nomenclature. Neotropical Primates, 10, 113–115. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292.
Female common marmosets: infanticide and self-restraint Clutton-Brock, T. H., Brotherton, P. N., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Creel, S. R., & Waser, P. M. (1997). Variation in reproductive suppression among dwarf mongooses: interplay between mechanisms and evolution. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 150–169. Digby, L. (1995a). Social organization in a wild population of Callithrix jacchus. Part II. Intragroup social behavior. Primates, 36, 361–375. Digby, L. (1995b). Infant care, infanticide, and female reproductive strategies in polygynous groups of common marmosets (Callithrix jacchus). Behavioral Ecology and Sociobiology, 37, 51–61. Digby, L. J. (1999). Sexual behavior and extragroup copulations in a wild population of common marmosets (Callithrix jacchus). Folia Primatologica, 70, 136–145. Digby, L. J. (2000). Infanticide by female mammals: implications for the evolution of social systems. In C. P. van Schaik & C. H. Janson, eds., Infanticide by Males and its Implications. Cambridge: Cambridge University Press, pp. 423–446. Digby, L. J. & Saltzman, W. (in press). Balancing cooperation and competition in callitrichine primates: examining the relative risk of infanticide across species. In L. E. Davis, S. M. Ford, & L. Porter, eds., The Smallest Anthropoids: the Callimico/ Marmoset Radiation. New York, NY: Springer. Digby, L. J., Ferrari, S. F., & Saltzman, W. (2007). Callitrichines: the role of competition in a cooperatively breeding species. In C. Campbell, A. Fuentes, K. C. MacKinnon, M. Panger & S. Bearder, eds., Primates in Perspective. Oxford: Oxford University Press, pp. 85–106. Epple, G. (1967). Vergleichende Untersuchungen u¨ber Sexual- und Sozialverhalten der Krallenaffen (Hapalidae). Folia Primatologica, 7, 37–65. Evans, S. & Hodges, J. K. (1984). Reproductive status of adult daughters in family groups of common marmosets (Callithrix jacchus jacchus). Folia Primatologica, 42, 127–133. Faulkes, C. G. & Bennett, N. C. (2001). Family values: group dynamics and social control of reproduction in African mole-rats. Trends in Ecology and Evolution, 16, 184–190. Faulkes, C. G., Arruda, M. F., & Monteiro da Cruz, M. A.O. (2003). Matrilineal genetic structure within and among populations of the cooperatively breeding common marmoset, Callithrix jacchus. Molecular Ecology, 12, 1101–1108. Ferrari, S. F. (1993). Ecological differentiation in the Callitrichidae. In A. B. Rylands, ed., Marmosets and Tamarins: Systematics, Behaviour, and Ecology. Oxford: Oxford University Press, pp. 314–328. Ferrari, S. F. & Digby, L. J. (1996). Wild Callithrix groups: stable extended families? American Journal of Primatology, 38, 19–27. French, J. A. (1997). Proximate regulation of singular breeding in callitrichid primates. In N. G. Solomon and J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 34–75.
363
364
D. H. Abbott et al. Gilchrist, J. S. (2006). Female eviction, abortion, and infanticide in banded mongoosees (Mungos mungo): implications for social control of reproduction and synchronized parturition. Behavioral Ecology, 17, 664–669. Gromoll, J., Wistuba, J., Terwort, N., et al. (2003). A new subclass of the luteinizing hormone/chorionic gonadotropin receptor lacking exon 10 messenger RNA in the New World monkey (Platyrrhini) lineage. Biology of Reproduction, 69, 75–80. Hager, R. (2003). Models of reproductive skew applied to primates. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 65–101. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Hamilton, I. M. (2004). A commitment model of reproductive inhibition in cooperatively breeding groups. Behavioral Ecology, 4, 585–591. Harlow, C. R., Hillier, S. G., & Hodges, J. K. (1986). Androgen modulation of folliclestimulating hormone-induced granulosa-cell steroidogenesis in the primate ovary. Endocrinology, 119, 1403–1405. Hiddleston, W. A. (1978). Large scale production of a small laboratory primate: Callithrix jacchus. In T. Antikatzides, S. Erichsen, & A. Spiegel, eds., The Laboratory Animal in the Study of Reproduction. New York, NY: Gustav Fischer Verlag, pp. 51–57. Hoogland, J. L. (1995). The Black-Tailed Prairie Dog: Social Life of a Burrowing Mammal. Chicago, IL: University of Chicago Press. Hrdy, S. B. (1979). Infanticide among animals: a review, classification, and examination of the implications for the reproductive strategies of females. Ethology and Sociobiology, 1, 13–40. Hubrecht, R. C. (1984). Field observations on group size and composition of the common marmoset (Callithrix jacchus jacchus) at Tapacura, Brazil. Primates, 25, 13–21. Hubrecht, R. C. (1989). The fertility of daughters in common marmoset (Callithrix jacchus jacchus) family groups. Primates, 30, 423–432. Jaquish, C. E., Gage, T. B., & Tardif, S. D. (1991). Reproductive factors affecting survivorship in captive Callitrichidae. American Journal of Physical Anthropology, 84, 291–305. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 998–102. Kirkpatrick-Tanner, M., Aeschlimann, C., & Anzenberger, G. (1996). Occurrence of an infanticide within a captive polygynous group of common marmosets, Callithrix jacchus. Folia Primatologica, 67, 52–58. Lacey, E. A. & Sherman, P. W. (1997). Cooperative breeding in naked mole-rats: implications for vertebrate and invertebrate sociality. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 267–301.
Female common marmosets: infanticide and self-restraint Layne, D. G. & Power, R. A. (2003). Husbandry, handling, and nutrition for marmosets. Comparative Medicine, 53, 351–359. Lazaro-Perea, C. (2001). Intergroup interactions in wild common marmosets, Callithrix jacchus: territorial defence and assessment of neighbours. Animal Behaviour, 62, 11–21. Lazaro-Perea, C., Castro, C. S.S., Harrison, R., et al. (2000). Behavioral and demographic changes following the loss of the breeding female in cooperatively breeding marmosets. Behavioral Ecology and Sociobiology, 48, 137–146. Lazaro-Perea, C., Arruda, M. F., & Snowdon, C. T. (2004). Grooming as a reward? Social function of grooming between females in cooperatively breeding marmosets. Behavioral Ecology and Sociobiology, 48, 137–146. Melo, L., Mendes Pontes, A. R., & Monteiro da Cruz, M. A.O. (2003). Infanticide and cannibalism in wild common marmosets. Folia Primatologica, 74, 48–50. Muller, T., Simoni, M., Pekel, E., et al. (2004a). Chorionic gonadotrophin beta subunit mRNA but not luteinising hormone beta subunit mRNA is expressed in the pituitary of the common marmoset (Callithrix jacchus). Journal of Molecular Endocrinology, 32, 115–128. Muller, T., Gromoll, J., Simula, A. P., et al. (2004b). The carboxyterminal peptide of chorionic gonadotropin facilitates activation of the marmoset LH receptor. Experimental and Clinical Endocrinology and Diabetes, 112, 574–579. Nievergelt, C. M., Digby, L. J., Ramakrishnan, U., & Woodruff, D. S. (2000). Genetic analysis of group composition and breeding system in a wild common marmoset (Callithrix jacchus) population. International Journal of Primatology, 21, 1–20. Reeve, H. K. (1998). Game theory, reproductive skew, and nepotism. In L. Dugatkin and H. K. Reeve, eds., Game Theory and Animal Behavior. New York, NY: Cornell University Press, pp. 118–148. Reeve, H. K. & Emlen, S. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecoology, 9, 267–278. Roda, S. A. & Mendes Pontes, A. R. (1998). Polygyny and infanticide in common marmosets in a fragment of the Atlantic Forest of Brazil. Folia Primatologica, 69, 372–376. Rothe, H. (1975). Some aspects of sexuality and reproduction in groups of captive marmosets (Callithrix jacchus). Zeitschrift fu¨r Tierpsychologie, 37, 255–273. Rylands, A. B. & de Faria, D. S. (1993). Habitats, feeding ecology, and home range size in the genus Callithrix. In A. B. Rylands, ed., Marmosets and Tamarins: Systematics, Behaviour, and Ecology. Oxford: Oxford University Press, pp. 262–272. Saltzman, W. (2003). Reproductive competition among female common marmosets (Callithrix jacchus): proximate and ultimate causes. In C. B. Jones, ed., Sexual
365
366
D. H. Abbott et al. Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 197–229. Saltzman, W. & Abbott, D. H. (2005). Diminished maternal responsiveness during pregnancy in multiparous female common marmosets. Hormones and Behavior, 47, 151–163. Saltzman, W., Schultz-Darken, N. J., Scheffler, G., Wegner, F. H., & Abbott, D. H. (1994). Social and reproductive influences on plasma cortisol in female marmoset monkeys. Physiology and Behavior, 56, 801–810. Saltzman, W., Schultz-Darken, N. J., & Abbott, D. H. (1996). Behavioural and endocrine predictors of dominance and tolerance in female common marmosets, Callithrix jacchus. Animal Behaviour, 51, 657–674. Saltzman, W., Schultz-Darken, N. J., & Abbott, D. H. (1997a). Familial influences on ovulatory function in common marmosets (Callithrix jacchus). American Journal of Primatology, 41, 159–177. Saltzman, W., Schultz-Darken, N. J., Severin, J. M., & Abbott, D. H. (1997b). Escape from social suppression of sexual behavior and of ovulation in female common marmosets. Annals of the New York Academy of Sciences, 807, 567–570. Saltzman, W., Severin, J. M., Schultz-Darken, N. J. & Abbott, D. H. (1997c). Behavioral and social correlates of escape from suppression of ovulation in female common marmosets housed with the natal family. American Journal of Primatology, 41, 1–21. Saltzman, W., Schultz-Darken, N. J., Wegner, F. H., Wittwer, D. J., & Abbott, D. H. (1998). Suppression of cortisol levels in subordinate female marmosets: reproductive and social contributions. Hormones and Behavior, 33, 58–74. Saltzman, W., Pick, R. R., Salper, O. J., Liedl, K. J., & Abbott, D. H. (2004). Onset of plural cooperative breeding in common marmoset families following replacement of the breeding male. Animal Behaviour, 68, 59–73. Saltzman, W., Liedl, K. J., Salper, O. J., Pick, R. R., & Abbott, D. H. (2008). Postconception reproductive competition in cooperatively breeding common marmosets (Callithrix jacchus). Hormones and Behavior, 53, 274–286. Sapolsky, R. M., Romero, L. M., & Munck, A. U. (2000). How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine Reviews, 21, 55–89. Sherman, P. W. (1981). Reproductive competition and infanticide in Belding’s ground squirrels and other animals. In R. D. Alexander & D. Tinkle, eds., Natural Selection and Social Behavior: Recent Research and Theory. New York, NY: Chiron Press, pp. 311–331. Snowdon, C. T. (1996). Infant care in cooperatively breeding species. Advances in the Study of Behavior, 25, 643–689. Solomon, N. G. & French, J. A. (1997). The study of mammalian cooperative breeding. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 1–10.
Female common marmosets: infanticide and self-restraint Soules, M. R., McLachlan, R. I., Ek, M., et al. (1989a). Luteal phase deficiency: characterization of reproductive hormones over the menstrual cycle. Journal of Clinical Endocrinology and Metabolism, 69, 804–812. Soules, M. R., Clifton, D. K., Cohen, N. L., Bremner, W. J., & Steiner, R. A. (1989b). Luteal phase deficiency: abnormal gonadotropin and progesterone secretion patterns. Journal of Clinical Endocrinology and Metabolism, 69, 813–820. Sousa, M. B.C., Albuquerque, A. C.S. R., Albuquerque, F. S., et al. (2005). Behavioral strategies and hormonal profiles of dominant and subordinate common marmoset (Callithrix jacchus) females in wild monogamous groups. American Journal of Primatology, 67, 37–50. Stewart, P. M. (2003). The adrenal cortex. In P. R. Larsen, H. M. Kronenberg, S. Melmed, K. S. Polonsky, eds., Williams Textbook of Endocrinology, 10th edn. Philadelphia, PA: Saunders, pp. 517–664. Sussman, R. W., Garber, P. A., & Cheverud, J. M. (2005). Importance of cooperation and affiliation in the evolution of primate sociality. American Journal of Physical Anthropology, 128, 74–83. Tannenbaum, P. L., Schultz-Darken, N. J., Woller, M. J., & Abbott, D. H. (2007). Gonadotrophin-releasing hormone (GnRH) release in marmosets II: pulsatile release of GnRH and pituitary gonadotrophin in adult females. Journal of Neuroendocrinology, 19, 354–363. Tardif, S. D. (1997). The bioenergetics of parental behavior and the evolution of alloparental care in marmosets and tamarins. In N. G. Solomon and J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 11–33. van Lawick, H. (1974). Solo: the Story of an African Wild Dog. Boston, MA: Houghton Mifflin. Wasser, S. K. & Barash, D. P. (1983). Reproductive suppression among female mammals: implications for biomedicine and sexual selection theory. Quarterly Review of Biology, 58, 513–538. Wasser, S. K. & Starling, A. K. (1988). Proximate and ultimate causes of reproductive suppression among female yellow baboons at Mikumi National Park, Tanzania. American Journal of Primatology, 16, 97–121. Williams, G. C. (1966). Adaptation and Natural Selection: a Critique of Some Current Evolutionary Thought. Princeton, NJ: Princeton University Press. Yamamoto, M. E., Arruda, M. F., Sousa M. B.C., & Alencar, A. I. (1996). Mating systems and reproductive strategies in Callithrix jacchus females. Abstracts of the XVIth Congress of the International Primatological Society and XIX Conference of the American Society of Primatologists, 56. Yamamoto, M. E., Arruda, M. F., Alencar, A. I., Sousa, M. B. C., & Arau´jo, A. (in press). Mating systems and female–female competition in the common marmosets, Callithrix jacchus. In L. C. Davis, S. M. Ford, & L.. Porter, eds., The Smallest Anthropoids: the Callimico/Marmoset Radiation. New York, NY: Springer. Yen, S. S. C. (2004). Neuroendocrinology of reproduction. In J. F. Strauss & R. L. Barbieri, eds., Yen and Jaffe’s Reproductive Endocrinology: Physiology,
367
368
D. H. Abbott et al. Pathophysiology and Clinical Management, 5th edn. Philadelphia, PA: Saunders, pp. 3–73. Young, A. J. & Clutton-Brock, T. H. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387. Young, A. J., Carlson, A. A., Monfort, S. L., et al. (2006). Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proceedings of the National Academy of Sciences of the USA, 103, 12005–12010. Zhang, F. P., Rannikko, A. S., Manna, P. R., Fraser, H. M., & Huhtaniemi, I. T. (1997). Cloning and functional expression of the luteinizing hormone receptor complementary deoxyribonucleic acid from the marmoset monkey testis: absence of sequences encoding exon 10 in other species. Endocrinology, 138, 2481–2490. Ziegler, T. E. & Sousa, M. B. (2002). Parent–daughter relationships and social controls on fertility in female common marmosets, Callithrix jacchus. Hormones and Behavior, 42, 356–367.
13
Reproductive skew in African mole-rats: behavioral and physiological mechanisms to maintain high skew c h r i s g . fa u l k e s a n d n i g e l c . b e n n e t t
Summary African mole-rats of the family Bathyergidae are a powerful model system for the study of both ultimate and proximate factors affecting reproductive skew in mammals. They exhibit a range of cooperative breeding strategies and differing degrees of skew, culminating in eusocial behavior in two or possibly more species. Mating preferences encompass both facultative inbreeding/outbreeding and obligate outbreeding, and result in common features, but also some clear differences between mole-rat species in how their social systems and skew are maintained. Recent genetic studies have revealed that, in the species studied so far, colonies may contain unrelated immigrants and therefore there is the potential for reproductive conflicts of interest, because all mole-rats are reproductively totipotent. In naked mole-rats (Heterocephalus glaber), extreme socially induced suppression of reproductive physiology is apparent in both sexes, and is best explained by a dominantcontrol model, rather than by staying or peace incentives. The Damaraland mole-rat (Cryptomys damarensis) exhibits suppression of reproductive physiology in females, but not males, and in this species skew may be maintained by a combination of dominant control and incest avoidance. Other social species lack specific physiological blocks to reproduction, and skew may be maintained by incest avoidance and other, as yet unknown, behavioral means.
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
369
370
C. G. Faulkes, N. C. Bennett Introduction The African mole-rats of the family Bathyergidae have received much attention from sociobiologists since the discovery of eusociality in the naked mole-rat (Heterocephalus glaber: Jarvis 1981). Following the announcement by Jarvis & Bennett (1993) that another mole-rat species, the Damaraland mole-rat (Cryptomys damarensis), also fitted the classical definition of eusociality (i. e. reproductive division of labor, cooperative care of offspring, and overlap of generations: Michener, 1969), it became apparent that the Bathyergidae were a model system for the study of social evolution and extreme reproductive skew. In addition to the two eusocial species, which have subsequently been found to be highly divergent within the family, other species of Cryptomys studied to date are less social. However, like the naked and Damaraland mole-rats, they all display cooperative breeding, with normally a single female and a small number of males reproducing at any one time within the colony. Other species among the remaining three genera are strictly solitary. The application of molecular phylogenetics to clarify the taxonomy and evolutionary relationships of mole-rats has now opened up the possibilities of phylogenetically controlled comparative analyses, whilst microsatellite genotyping is beginning to clarify patterns of relatedness within and among social groups, mating systems, colony turnover, and dispersal, and hence to provide an accurate insight into reproductive skew (e. g. Burland et al. 2002, 2004, Bishop et al. 2004, Hess 2004). One difficulty in undertaking comparative studies of social behavior is simply defining the degree of sociality, and this topic has received much attention both in general (Crespi & Yanega 1995, Keller & Perrin 1995, Sherman et al. 1995, Costa & Fitzgerald 1996, Crespi 2005) and specifically in African mole-rats (Bennett & Faulkes 2000, Burda et al. 2000, Faulkes & Bennett 2007, O’Riain & Faulkes 2008). Notwithstanding these semantic difficulties, current research linking studies of physiology, behavior, and estimates of skew using molecular genetic data in African mole-rats is helping us to elucidate the proximate mechanisms that control who breeds and who helps in groups of cooperative breeders, and why these mechanisms may differ among species. After a brief overview of the ecological and social diversity within the family, this chapter will review our current understanding of the evolution and maintenance of reproductive skew in African mole-rats, focusing on the three best-studied species (naked, Damaraland, and common mole-rats, C. hottontotus hottontotus). We will then examine how these empirical studies can be interpreted or explained by current theoretical models of skew. Social diversity in African mole-rats African mole-rats are endemic to sub-Saharan Africa, and one of the most speciose of the hystricomorph families of rodents. Reviewing the early
African mole-rats: behavioral and physiological mechanisms literature from the late nineteenth and early twentieth centuries, Ellerman (1940) listed a total of 67 species in five genera, as follows: Heterocephalus (4 species, eusocial) Heliophobius (8 species, solitary) Georychus (3 species, solitary) Bathyergus (3 species, solitary) Cryptomys (49 species, social/eusocial).
Although there is now general agreement that Heterocephalus can be considered a monotypic genus, new studies support the species diversity reported by Ellerman for Heliophobius and Cryptomys (Faulkes et al. 1997a, 2004, Ingram et al. 2004, Van Daele et al. 2004). There have also been suggestions that genus Cryptomys should be split into two clades, Cryptomys and Coetomys (Ingram et al. 2004) or Cryptomys and Fukomys (Kock et al. 2006). Until the nomenclature stabilises, and to retain consistency with previous literature, in this chapter we have adopted the traditional single-genus terminology (Cryptomys). Figure 13.1 presents a simplified phylogeny showing the evolutionary relationships between the five genera. Fossil evidence of Heterocephalus and molecular phylogenies place the root of the bathyergid family tree in east Africa, with Heterocephalus and Heliophobius forming the basal lineages (Figure 13.1). Molecular clock estimates suggest a comparatively ancient origin, and date the common ancestor of the family to approximately 40–48 million years ago (Huchon & Douzery 2001). Tectonic upheaval in this part of Africa during the formation of the Rift Valley appears to have had a major influence on the adaptive radiation and spread of the family across sub-Saharan Africa, through changes in geomorphology, vegetation, and climate (Faulkes et al. 2004, Faulkes & Bennett 2007), and the changing patterns of drainage of major river systems (Van Daele et al. 2004, 2007). It seems highly likely that the variation in patterns of social behavior observed across the family has arisen in response to these environmental challenges, and has led to the convergent gains and/or losses of sociality among different mole-rat clades (Figure 13.1). As mentioned above, two species, the naked mole-rat (Figure 13.2) and the Damaraland mole-rat (Figure 13.3), fit the definition of eusociality derived from social insects (Jarvis 1981, Jarvis & Bennett 1993), and with further investigation other mole-rats may also be found to fall into this category (Bennett & Faulkes 2000, Burda et al. 2000). Unlike other species studied to date, both Damaraland and naked mole-rats also possess a behavioral division of labor among the non-breeding helpers within the colony. In the former, smaller individuals form a “frequent worker” group, while in the latter the frequency of work, in the form of colony maintenance behavior (foraging for food,
371
372
C. G. Faulkes, N. C. Bennett Cryptomys damarensis 99 Cryptomys anselli 100 96 Cryptomys sp. Kalomo 92
Cryptomys darlingi
100
Cryptomys whytei Cryptomys bocagei 100 Cryptomys mechowi Social/eusocial
100 100
12-17myr
Cryptomys h. pretorise
98
Cryptomys h. nimrodi
93 100
100
Cryptomys h. matalensis
Cryptomys h. hottentotuis
22-26myr Cryptomys h. mahali
100
100 3240myr
Bathyergus suillus Bathyergus janetta
100
Solitary
16-19myr Georychus capensis
4048myr
Heliophobius argenteocinereus Heterocephalus glaber
Eusocial
Thryonomys swinderianus (outgroup)
Figure 13.1 Phylogram based on maximum-parsimony analysis of 18 bathyergid mtDNA haplotypes (combined 12S rRNA and cytochrome b gene sequences), showing representative species from each genus with their social characteristics, and outgroup species Thryonomys swinderianus (cane rat). Numbers above each branch refer to the % bootstrap values following 100 replications, after weighting sites with the rescaled consistency index. Divergence times of selected internal nodes are in million years before present (myr; data adapted from Faulkes et al. 1997a, 2004).
digging and maintaining the burrow), shows a negative trend with increasing body mass (Lacey & Sherman 1991, Faulkes et al. 1991). Furthermore, there is evidence for the existence of morphologically distinct castes in naked molerats, including a disperser morph among males (O’Riain et al. 1996, Braude
African mole-rats: behavioral and physiological mechanisms
Figure 13.2 Social group of naked mole-rats in a nest chamber. Photo by Chris Faulkes.
Figure 13.3 Group of Damaraland mole-rats. Courtesy of Tim Jackson.
2000) and differences in the vertebrae of the breeding queen (O’Riain et al. 2000), and physiological castes in Damaraland mole-rats (Scantlebury et al. 2006). Table 13.1 summarizes some of the social and reproductive characteristics of the species for which information is available.
373
374
Table 13.1 A summary of the social and reproductive characteristics of seven species of African mole-rats Suppression of reproductive No. breeders/group
physiology in
(mean/max.
Role of
Mating
non-breeders
Species
Habitat type group sizes)a
EPPs ECPs helpers
preference
Males Females
Naked mole-rat
Semi-arid
?
Outbreeding/
Yes
(Heterocephalus glaber) Damaraland mole-rat
1 female/1–3
?
males (75/295) Semi-arid
(Cryptomys damarensis)
1 female/1–2
No
Yes
males (11/41)
Foraging, defense,
facultative
pup care
inbreeding
Foraging, defense,
Obligate
Yes
(Cryptomys h. hottentotus) Natal mole-rat
Mesic and semi-arid
No
Yes
outbreeding
1 female/1
Mesic
1 female/1
Bennett & Faulkes 2000,
Foraging
Obligate
No
No
Bennett & Faulkes 2000,
?
?
Foraging
Outbreeding
No
No
Bennett & Faulkes 2000
?
?
Foraging
Outbreeding
No
No
Bennett et al. 1994,
?
?
Foraging
Outbreeding
No
No
Burda 1995
?
?
Foraging
Outbreeding
No
No
Bennett & Aguilar 1995,
outbreeding
Bishop et al. 2004
male (7/9) Mesic
1 female/1
Mesic
1 female/1
(Cryptomys anselli) Giant mole-rat
Yes
male (9/16)
(Cryptomys darlingi) Ansell’s mole-rat
Jarvis & Bennett 1993, Burland et al. 2004
Yes
male (5/14)
Mesic
(Cryptomys h. natalensis) Mashona mole-rat
1 female/1
Jarvis 1981, Lacey & Sherman 1991, Bennett & Faulkes 2000
pup care Common mole-rat
References
Bennett & Faulkes 2000
male (-/20)
(Cryptomys mechowi)b
male (5/11)
Wallace & Bennett 1998, Bennett & Faulkes 2000, Burland et al. 2004
a
The data in this column are based on the usual patterns of sociality observed, with the breeding males identified by behavioral and morphological
characteristics, and in some cases paternity analysis using genetic markers. Deviations are known to occur in most species, e.g. rare cases of two breeding queens. b
Group sizes of up to 60 have been proposed for C. mechowi, but these are based on interviews with local hunters (Jarvis & Bennett 1993, Burland et al.
2004) and have not been confirmed by systematic trapping. EPPs are extra-breeding-pair paternities (but occurring within the colony), while ECP are extra-colony paternities – both EPPs and ECPs have been determined using genetic techniques.
African mole-rats: behavioral and physiological mechanisms Ultimate factors leading to cooperative breeding in mole-rats In addition to the social diversity apparent within the family, there is also considerable diversity among species in their habitat, which ranges across sub-Saharan Africa. Solitary mole-rats in the genera Heliophobius, Bathyergus, and Georychus are generally restricted to regions of higher precipitation (greater than 400 mm per annum). The two genera that exhibit social behavior (Cryptomys and Heterocephalus) are found in both mesic (with a moderate and predictable supply of moisture) and xeric (low rainfall) regions. Heterocephalus occurs exclusively in the arid regions of east Africa (parts of Kenya, Ethiopia, and Somalia). The areas they inhabit are generally characterized by low (less than 400 mm per annum) and unpredictable rainfall, with on average only four months per year having more than 25 mm of rain (approximately the quantity required to soften the soil at the depth of foraging tunnels and thus facilitate burrowing; Jarvis et al. 1994). The genus Cryptomys is the most widely distributed of all the extant bathyergids, and as with Heterocephalus, the ranges of some of these social Cryptomys (e. g. the Damaraland mole-rat) extend into areas of very low, sporadic, and unpredictable rainfall (sometimes < 200 mm per annum). However, some species also occur in mesic habitats, like the common mole-rat, and the giant mole-rat Cryptomys mechowi (Table 13.1). All African mole-rats feed on underground roots and tubers (geophytes) that are encountered by digging foraging tunnels from the central core of the burrow system (for review see Bennett & Faulkes 2000). Digging through soil is energetically expensive, and costs of burrowing may vary with soil hardness, which may in turn vary among different habitats, or within a habitat between the seasons (e. g. before and after rains). The distribution of roots and tubers eaten by mole-rats also varies with habitat. In mesic regions these food resources are more uniformly distributed, whilst in areas of low or unpredictable rainfall the plants are arid-adapted and more widely dispersed, or occur in high-density clumps that are patchily distributed. The aridity food distribution hypothesis (AFDH) proposes that increased natal philopatry, cooperative breeding, and ultimately eusocial behavior in African mole-rats may have evolved in response to the rainfall patterns of the habitat, its effects on food distribution, and the subsequent costs and risks of foraging and dispersal (Jarvis 1978, Bennett 1988, Lovegrove & Wissell 1988, Lovegrove 1991, Jarvis et al. 1994, 1998, Faulkes et al. 1997a). In this scenario sociality would be adaptive, as cooperative foraging shares out the energetic costs of burrowing and increases the chances of finding food. The clumped nature of the geophytes and, in some cases, their large size, means that once found these resources are sufficient to sustain large groups of animals.
375
376
C. G. Faulkes, N. C. Bennett Support for the AFDH playing at least a role in the evolution of sociality in the Bathyergidae comes from both inter- and intraspecific studies. Analyses of the architecture of burrows clearly demonstrates that in arid habitats colonies containing larger numbers of individuals are able to explore their habitat more efficiently, with a greater degree of complexity (calculated as fractal dimension) of the foraging tunnels than smaller groups (Le Comber et al. 2002). Furthermore, a long-term field study of Damaraland mole-rats has shown that larger colonies are at an advantage and less likely to fail when environmental conditions are at their most extreme (e. g. during periods of drought), again emphasizing the adaptive advantage of sociality (Jarvis et al. 1998). A comparative analysis of data from across the family revealed significant relationships as predicted by the AFDH between social group size and geophyte density, the coefficient of rainfall variation, and the mean number of months per year where rainfall exceeded 25 mm (Faulkes et al. 1997a). Some intraspecific data for the most social of the bathyergids, the naked mole-rat, have now also been collected. Working on the assumption that increased aridity should reduce dispersal and increase within-group relatedness (r), Hess (2004) used microsatellite genotyping to estimate relatedness in colonies across an aridity gradient in Kenya. A weak but statistically significant correlation was found in support of aridity constraining dispersal, but more work is needed to rule out the possibility that this result was influenced by a small sample size. Similarly, comparison of intra-colony relatedness in Damaraland mole-rats revealed a lower r value at Hotazel, South Africa, compared to a more arid site at Dordabis, Namibia (mean ± 95% confidence interval 0.40 ± 0.02 vs. 0.54 ± 0.04: Burland et al. 2002). Within-species comparisons of philopatry and dispersal in arid- and mesic-dwelling common mole-rats in South Africa have also shown that immigration and emigration were lower at an arid site than at a mesic one, indicating that constraints on dispersal are higher in areas of low and unpredictable rainfall (Spinks et al. 2000a). Burrow architecture was also more complex at the mesic site, reflecting the reduction in burrowing constraints in this habitat (Le Comber et al. 2002). Despite these differences, however, estimates of intra-colony relatedness were not found to differ between sites (mean ± standard error 0.23 ± 0.02 at Somerset West, South Africa, a mesic site, vs. 0.28 ± 0.03 at Steinkopf, South Africa, an arid site). Burda et al. (2000) argue against a causal relationship between cooperative foraging for dispersed food resources and the evolution of sociality in molerats. They suggest instead that the social behavior of mole-rats is a result of an ancestral tendency for “cooperative monogamy” reinforced by a subterranean lifestyle that constrains dispersal. “Eusocial” families then develop as a result of longevity, and slow pre- and postnatal growth and development (for further
African mole-rats: behavioral and physiological mechanisms discussion and counter-argument see Faulkes & Bennett 2007, O’Riain & Faulkes 2008).
Proximate factors maintaining high skew in social mole-rats An overview of skew among African mole-rats
In colonies of all cooperatively breeding mole-rats studied to date, a single female (the queen) normally breeds with one, but perhaps on occasion two or three, resident breeding males. Thus, at any one point in time within the colony, all social mole-rat groups have a high reproductive skew among females, with lower skew among the males. New insights into male skew in wild colonies of Cryptomys have recently been gained following paternity analysis using microsatellite genotypes. These studies have not only unambiguously established the incidence of breeding males resident within colonies (previously only inferred in the wild from morphological and behavioral characteristics), but have further revealed that paternity is in some cases due to both extra-pair subordinate males and non-resident males. Common mole-rats
In a study by Bishop et al. (2004), paternity was not always assigned to the largest, most dominant male within the colony, the individual traditionally viewed as the breeding male; extra-colony males were also found to be responsible for paternities. However, in addition, subordinate males within colonies, hitherto classed as “non-breeders,” were also found to sire offspring. There were significant differences in the proportions of these assignments to different males, depending on the field site. At a mesic site where ecological constraints on dispersal were lower, out of 24 colonies studied only 19% of within-colony paternities were assigned to the male identified as the breeder based on morphology, with 81% due to a different within-colony male. Overall, 29% of paternities were assigned to extra-colony males. Conversely, at an arid site where dispersal costs were higher, 79% of within-colony paternities were assigned to the male identified as the breeder and, overall, 18% of paternities at the site were assigned to extra-colony males. These disparate results clearly indicate that alternative reproductive strategies may exist for males of this species, depending on the environmental conditions and the ease of movement of males between colonies. Plural breeding within colonies among females appears to be very uncommon. In an extensive field study of common mole-rats, 49 colonies surveyed at two geographic locations all had a single queen (this was later confirmed by maternity analysis of microsatellite genotypes). Many other small-scale studies
377
378
C. G. Faulkes, N. C. Bennett have also failed to detect more than one queen per colony, although in one case plural breeding was observed in two out of 30 colonies caught at Somerset West, South Africa, over a two-year period, and in each case both queens reared offspring (N. C. Bennett, unpublished data). Damaraland mole-rats
In Damaraland mole-rats, most colonies (16 out of 18 studied) had just one breeding male resident at any particular time (Burland et al. 2004). However, in more than half of the colonies (nine out of 17 studied) no father could be assigned to at least one of the resident offspring. This indicates that the breeding female has contact with other non-resident males, either by temporarily leaving the colony or, more probably, by males transiently passing through the colony. To date, plural breeding among females in colonies of Damaraland mole-rats has not been observed, either in captivity (in more than 60 colonies), or in the wild (in over 150 colonies caught over a 15-year period from several geographic locations: J. U. M. Jarvis and N. C. Bennett, unpublished data). Naked mole-rats
In captive colonies of naked mole-rats up to three males have been observed mating with the queen (Lacey & Sherman 1991), and minisatellite DNA fingerprinting has confirmed that up to three males may sire offspring in a litter (Faulkes et al. 1997b). Among females, plural breeding has occasionally been seen in both captive and wild colonies. Braude (1991) records two instances of plural breeding. He captured a total of 2051 naked mole-rats from 23 colonies in Meru National Park, Kenya, over a three-year period and found one colony with two breeding queens. On a return field trip a year later this colony still had two queens, and a second colony was also found to contain two breeding females. Brett (1991) captured from seven wild colonies in southern Kenya, and in four of these all the individuals were caught: each had just one breeding queen. Jarvis (1985) caught two wild colonies at another site in Kenya, and also recorded a single breeding queen. Colonies maintained by Jarvis at the University of Cape Town are the only captive naked mole-rats so far reported to have contained two queens. In 1991 there were 19 colonies that were more than three years old in the laboratory, and of these, seven had exhibited plural breeding among females (six with two breeders, one with three). In four of these the two breeders tolerated one another long enough to have one or more litters, but in no cases were the offspring successfully reared. In three of these colonies fighting resulted: in one of the colonies, the second breeding female was removed, after five colony members were killed in the conflict that broke
African mole-rats: behavioral and physiological mechanisms out between the queens (although they themselves remained unharmed). In the other two colonies, the second breeding female produced one and three litters respectively, prior to killing the original breeding queen to remain as the sole breeder (Jarvis 1991). In addition to these four non-recruiting dualqueen colonies, three others did successfully rear offspring, but their survival rate was low. Other institutions maintaining naked mole-rats in captivity had not recorded plural breeding among females when the data were reviewed for 15 colonies by Braude (1991), and over a period of 20 years plural breeding has never been observed among many captive colonies (C. G. Faulkes, unpublished data). Although rare, these cases of plural breeding, especially the more detailed observations from captive colonies, provide an insight into the potential control of reproductive skew in naked mole-rats. Jarvis (1991) reports that in some of the captive colonies where plural breeding occurred it followed on from periods of good recruitment of offspring to the colony, and suggests that in the wild this might perhaps have triggered dispersal of rival females and new colony formation (if environmental conditions had been favorable). These situations in captivity where a second queen emerges after a non-breeder “escapes” from reproductive suppression invariably involve increased aggression, and often fatalities due to fighting (see later discussion on skew models and their application to mole-rats). Longer-term skew in lifetime reproductive success
Clearly, while reproductive skew within colonies from a number of mole-rat species at a particular point in time may be high (e.g. in most cases one female and one or two males may be breeding), collecting data in a “snapshot” fashion from wild colonies can be misleading and merely provides a measure of “instantaneous” skew. This may at first glance seem to give a similar measure of skew across species. However, to understand the full nature and extent of skew in mole-rats, it is important to consider a temporal component, since skew in terms of lifetime reproductive success may differ considerably between species, depending on the chances of dispersal and independent breeding. While opportunities for individual reproduction may be constrained over relatively long time periods, because of their long life span, mole-rats may gain the opportunity to breed during periods when environmental pressures on individual reproduction relax. This may occur infrequently (as has been shown for Damaraland mole-rats at Dordabis, Namibia, during periods of drought: Jarvis et al. 1994, 1998, Molteno & Bennett 2002), or predictably/seasonally in some of the cooperatively breeding species that inhabit more mesic habitats. In the case of the latter, several species have
379
380
C. G. Faulkes, N. C. Bennett been referred to as exhibiting eusocial behavior based on snapshot observations of reproductive skew in colonies, with no longitudinal data on reproductive success (e.g. C. anselli: Burda et al. 2000; C. mechowi: Burda & Kawalika 1993). In the Damaraland mole-rat, estimates from a long-term mark–release– recapture study at Dordabis, Namibia, have revealed that over a five-year period 92% of non-breeding females never had the opportunity of breeding: from a total of 403 adult non-reproductive animals 370 never attained breeding status (Jarvis & Bennett 1993, Jarvis et al. 1994). In naked mole-rats skew in lifetime reproductive success is even more extreme, with less than 0.1% (of over 4000 animals) captured as non-breeders subsequently re-caught as breeding queens (Sherman et al. 1992). Furthermore, long-term ecological studies need to be coupled with parentage analysis based on molecular genetics to unambiguously determine who is actually breeding, especially among males. The discovery that both “floater males” and subordinates in Cryptomys species may sire offspring exemplifies this point. This could not have been revealed using standard field surveying techniques. Kin structure of mole-rat colonies, reproductive suppression, and reproductive skew
Of the social mole-rats in the genus Cryptomys studied to date, both in the wild and in captivity, all appear to have a mating system that involves obligate outbreeding. Hence a colony of Cryptomys mole-rats composed of parents and offspring will necessarily exhibit high reproductive skew, simply as a result of there being no unrelated mates for adult offspring. Until recently, however, in the absence of genetic data it has not been possible to determine the kin structure of social groups in mole-rats, and therefore quantify the extent of inter-group dispersal and potential reproductive conflicts of interest that might then arise among unrelated group members. Two molecular genetic studies in Cryptomys clearly reveal that inter-group movements may be common, giving rise to the potential for reproductive conflict between the breeding and nonbreeding members of the colony (Bishop et al. 2004, Burland et al. 2004). In turn, this may lead to the emergence of socially induced reproductive suppression. Common mole-rats
In common mole-rats the majority of colonies at two study sites of varying aridity were composed of a family group together with foreign conspecifics, with more immigrants observed at the mesic site, where constraints on dispersal were predicted to be lower (Bishop et al. 2004). It is unclear whether these immigrants were all males (as evidenced by extra-colony extrabreeding pair paternities), or also included female dispersers. Although there is
African mole-rats: behavioral and physiological mechanisms the potential for breeding opportunities, high skew is observed among females, and immigrant males appear to restrict their mating to the queen. This is despite the fact that, in the common mole-rat, there are no apparent physiological blocks to reproduction in non-breeders of either sex (Spinks et al. 2000b), and all colony members appear to be potentially capable of reproduction. The regularity of rains in these habitats enables adult non-breeding females in these colonies to regularly disperse and pair with unrelated males, and as a consequence avoid conflict with the breeding queen of their natal colonies. A similar lack of physiological correlates of suppression has been observed in other mole-rat species that inhabit mesic areas, yet show a reproductive division of labor, including the Mashona mole-rat (Cryptomys darlingi: Bennett et al. 1997), the giant mole-rat (C. mechowi: Bennett et al. 2000), and another Zambian mole-rat (probably C. anselli: Burda 1995). At present, nothing is known about colony composition and turnover in these species, but we predict that their colonies should exhibit similar lability to that recorded in the common mole-rat. Damaraland mole-rats
In Damaraland mole-rats, immigrants of both sexes were identified and opposite-sex non-breeding animals were present at the same time in some colonies (Burland et al. 2004). Despite the presence of potential mates in Damaraland mole-rat colonies, only a single queen was found to be breeding. Again this leads to the question of what prevents these unrelated non-breeding animals from mating and reproducing? In the case of Damaraland mole-rats, where constraints on dispersal may impose prolonged philopatry, non-breeding females have a clear physiological block to reproduction, leading to a failure of ovulation (Bennett et al. 1996, 1999, Molteno & Bennett 2000). Nonbreeding males are not physiologically suppressed, but they do possess increased proportions of sperm with morphological defects (Maswanganye et al. 1999). Although the significance of these sperm abnormalities for fertility is unclear, captive non-breeding males make no attempt to mate with their female colony mates, which are usually close kin, presumably as a result of an incest-avoidance mechanism. This effect is so strong that in both wild and captive colonies in which the breeding female has died, all individuals will remain reproductively quiescent (sometimes for years) until a “foreign”, unrelated, individual becomes available, or dispersal/fragmentation of the colony occurs (Jarvis & Bennett 1993, Bennett et al. 1996, Rickard & Bennett 1997, Bennett & Faulkes 2000). The social cues that lead to anovulation in Damaraland mole-rats, and that prevent breeding in non-breeding female common mole-rats, remain unknown. In captivity, these species are typically
381
382
C. G. Faulkes, N. C. Bennett kept in family groups where incest avoidance maintains skew, so behavioral interactions between unrelated non-breeding adults have, to date, not been studied. Naked mole-rats
Among cooperative breeders, naked mole-rats exhibit the extreme of socially induced infertility, with non-breeders of both sexes within colonies being physiologically suppressed by the queen. In females, gonadal development, ovarian cyclicity, and ovulation are blocked, leaving them in an apparent pre-pubertal state. In males, most non-breeders have spermatozoa within the reproductive tract, but they are both reduced in number and lack normal levels of motility. In both sexes, reduced secretion of the pituitary gonadotropin, luteinizing hormone, is evident. In turn, this is thought to arise from a disruption in gonadotropin-releasing hormone (GnRH) from the hypothalamus, although immunohistochemical studies reveal that GnRH is still produced in non-breeders. These extreme (but rapidly reversible) reproductive blocks are brought about by social contact with the dominant breeding queen (for review see Faulkes & Abbott 1997, Bennett & Faulkes 2000). While incest avoidance, together with other factors, may explain how a reproductive division of labor is maintained in Cryptomys, the proximate control of reproductive skew in naked mole-rats appears to be different. One of the emerging debates in recent years has been the extent to which inbreeding versus outbreeding in naked mole-rats accounts for their system of mating, and in turn, how this may influence the evolution and maintenance of their eusocial behavior. Much attention has been given to the notion that naked mole-rats are unusual among mole-rats and mammals in general, because they may inbreed to a high degree. These incestuous tendencies were originally proposed as an important factor in explaining their eusociality (Reeve et al. 1990), and it was argued that inbreeding produced a within-kin-group genetic structure analogous to haplodiploidy in the Hymenoptera. Reeve et al. (1990) estimated intra-colony relatedness in some groups at 0.8, a value greater than the average 0.75 relatedness in haplodiploid organisms where the queen is singly mated (Hamilton 1964). Until recently, genetic data have been derived from sampling in and around Mtito Andei, in the south of Kenya (Faulkes et al. 1990a, 1997b, Reeve et al. 1990). It has been suggested that the high estimates of relatedness may in fact be an artifact of mole-rats in this particular location having been through a past population bottleneck (Braude 2000), rather than due to consanguinous mating. Furthermore, a number of studies have shown that if they are given the choice, outbreeding is the preferred system of mating in naked mole-rats
African mole-rats: behavioral and physiological mechanisms (Clarke & Faulkes 1999, Braude 2000, Ciszek 2000). Indeed, a particular “disperser morph” has been identified among male naked mole-rats in captive colonies (O’Riain et al. 1996), and these have also been observed in wild colonies (Braude 2000). These males are non-breeding in their parent colonies, but deliberately attempt to emigrate into neighboring groups and outbreed. A recent extensive genetic survey of the long-term field study site described by Braude (1991, 2000) in and around Meru, in the north of Kenya, has utilized microsatellite genotyping for the first time in this species. This study specifically set out to investigate intra-colony relatedness within the context of historical population structuring and the natural viscosity in gene flow that occurs in subterranean animals (Hess 2004). Some estimates of relatedness in naked mole-rats were much higher than those observed and described earlier for both the common and the Damaraland mole-rat, with a mean and standard error for the total population of 0.74 ± 0.02 (range 0.50 ± 0.08 to 0.75 ± 0.05 over colonies grouped into 20 regions). However, when substructuring of populations due to patterns of drainage, elevation, and historical factors was accounted for in the analysis, average intra-colony r values were not significantly different to 0.5. Values of the statistic Fis derived from the same data were negative, regardless of the geographic scale, indicating an avoidance of inbreeding. Thus the genetics study of Hess (2004) supports the idea that outbreeding is the preferred mating system in wild colonies. However, without question, naked mole-rats differ from all other mole-rat species so far described in that they will spontaneously inbreed in captivity (facultative inbreeding). Is it possible that this is an adaptive trait enabling reproduction to continue when environmental conditions prevent dispersal and outbreeding. Examples of queen succession from within the colony, which may give rise to consanguinous matings, have been described in the previous section. Furthermore, in the wild populations studied by Braude (1991), and later by Hess (2004) examples of queen successions from within the group were recorded in three of the 23 colonies over the trapping period. In one colony the queen disappeared and was replaced by a worker from within the colony. In a second, the queen was removed and subsequently replaced by the second largest female of the colony (cf. captive studies by Clarke & Faulkes 1997). Finally, in a third colony, the daughter of the existing queen developed into a second breeding female (Braude 1991). While parentage analyses were not undertaken within these wild colonies, in captivity such replacements from within result in incestuous matings. Quantification of such events and understanding breeder succession in naked mole-rats (especially among queens) is important in understanding why naked mole-rats have evolved such extreme physiological suppression of reproduction. In a society in which a high reproductive
383
384
C. G. Faulkes, N. C. Bennett skew and a large non-breeding workforce is adaptive, but where there may be facultative inbreeding due to breeder succession from within the colony (when constraints preclude dispersal), a control mechanism appears to have evolved whereby social cues bring about suppression of reproductive physiology in both sexes. Without this, reproductive conflict could arise, with non-breeders attempting to breed, either amongst each other, or with existing breeders, or with immigrants. In species like the common, Mashona, and giant mole-rats, where no physiological suppression is observed, incest avoidance alone is apparently sufficient to maintain reproductive skew because colonies may be more labile and opportunities for individual reproduction are greater. The Damaraland mole-rat is an interesting case because it is intermediate, with suppression of reproductive physiology restricted to non-breeding females. It is possible that this trait might have evolved as a result of reproductive conflict as a potential control mechanism by the queen to prevent unrelated immigrant males from mating with non-breeding females, and therefore maintain the high skew observed among females.
Which theoretical model best explains skew in mole-rats? A brief overview of skew theory
Over recent years a large number of different models have been developed to explain variation in reproductive skew across animal societies from invertebrates to mammals (for reviews see Johnstone 2000, Hager 2003). Implicit in most skew models is the concept that the dominant breeder gains from the presence of helpers, that physical retention of subordinate nonbreeders in the group is practically impossible, and that the costs for a subordinate leaving the group are high. Skew models fall into two basic types, transactional and compromise, each having somewhat different assumptions and sometimes very different predictions, even within these main categories. First, transactional models assume group stability and control of group membership by the dominant breeders, who may then also control reproduction in other group members, perhaps conceding some reproduction to retain subordinate non-breeders. These so-called concession models include inducements to stay in the group and cooperate peacefully (“staying incentives”) or to prevent subordinates fighting for reproductive control (“peace incentives”: Reeve et al. 1998, Reeve 2000). Restraint models assume that although the dominant controls group membership, the subordinate actually controls reproduction and concedes a non-eviction incentive to the dominant by not breeding (Johnstone & Cant 1999). The second main category, compromise models, assume that
African mole-rats: behavioral and physiological mechanisms group stability is not required and that neither dominant or subordinate members control reproduction (e.g. tug-of-war models: Reeve et al. 1998). Since their inception, skew models have gradually increased in complexity. For example, a concession model now allows three, rather than just two, members in a group, a dominant and two subordinates, with different levels of relatedness (Johnstone et al. 1999). However, not surprisingly, they still remain over-simplistic with respect to any real-life examples. In the case of mole-rats, colonies usually contain many individuals (depending on the species, from 8 to 300), and patterns of relatedness, group size, and kin structure may vary markedly within and among species, and even within populations over time. Furthermore, theoretical skew models, although arguing from an evolutionary perspective, generally do not take into account the proximate mechanisms within the organism that selection has to operate on. For example, in restraint models that assume that the subordinate suppresses its own reproduction to avoid group eviction, it is unclear how this restraint may actually operate from a neuroendocrine and physiological point of view. If an animal has gone through puberty and is reproductively capable, and unrelated mates are available, to simply refrain from breeding (in the absence of behavioral intervention from the dominant) implies some kind of conscious decision making on the part of the subordinate, or some interruption in the normal process of pair-bond formation. The important role of inbreeding avoidance is also generally neglected in skew models (but see Reeve et al. 1998). Most animals have an instinctive mechanism leading them to avoid breeding with close relatives. Thus, in social groups comprising parents and one or more litters of offspring, high skew will simply result from the absence of unrelated mates for subordinates, and the dominant breeder need not offer any concessions, nor should there be any struggle over reproduction. Skew theory and empirical data from mole-rats
There is no a-priori reason to posit that a single theoretical model best explains skew either within or among mole-rat species, where convergent gains and losses of sociality have occurred. In the common mole-rat, colonies are found in both mesic and arid regions, and constraints on dispersal can vary markedly between habitats. Furthermore, within a population, conditions may also dramatically change over time as a result of rainfall and fluctuations in food abundance. For example, within a species, a colony may go from a situation of “group stability,” where there is little possibility of dispersal (or eviction) because constraints on burrowing are so high, to a situation of potential “group instability” following good rains, where animals are free to disperse and establish new breeding groups. As such, it could be argued that
385
386
C. G. Faulkes, N. C. Bennett different categories of skew theory may apply at different times within a population (see below). In addition, looking for empirical evidence to support one or another skew model in African mole-rats is complex because, among females, plural breeding within a colony is very uncommon and therefore female skew is almost always high. When multiple breeding males occur, the available evidence suggests that within their colony, they only mate with the dominant breeding female. Floating males, which move around giving rise to extra-colony paternities, add a further dimension to skew among males that has previously not been considered (Burland et al. 2004, Bishop et al. 2004). The majority of field and laboratory empirical data among mole-rats have been collected from three species (naked, Damaraland, and common), and we will discuss each of these in turn. Naked mole-rats
It is clear that the extreme reproductive suppression observed in naked mole-rats is maintained by social contact with the breeding queen. If non-breeders are removed from their parent colony and either housed singly, or paired with the opposite sex, they rapidly come into breeding condition. In females ovarian cyclicity commences (Faulkes et al. 1990b), whilst in males testosterone and luteinizing hormone levels may rise to concentrations comparable to those of breeding males (Faulkes & Abbott 1991). Likewise, if a queen is removed from her colony, one or more non-breeding females will become reproductively active and replacement will occur. This is often after intense aggression among these females (Jarvis 1991, Lacey & Sherman 1991), and usually the queen is succeeded by the next most dominant female in the colony (Clarke & Faulkes 1997). Primer pheromones released by the queen do not appear to be implicated in reproductive suppression of non-breeders (Faulkes & Abbott 1993, Smith et al. 1997). Instead, agonistic interactions directed by the queen towards the subordinate non-breeders are implicated. A candidate for this role is that of shoving, a behavior almost entirely restricted to the breeding queen (but occasionally seen in breeding males: Reeve & Sherman 1991). Shoving occurs when individuals confront one another face-toface, and the initiator pushes the recipient backwards down the tunnel, sometimes over a distance of up to 1 m, and often with the shoving animal hissing (Lacey et al. 1991). Interestingly, this behavior can be seen to develop and increase in frequency (from zero incidence) when non-breeders become reproductively active (as hormone levels rise) and they attain breeder status (Margulis et al. 1995, Clarke & Faulkes 1997, 1998). Shoving has also been suggested as a factor in inciting activity in “lazy” workers (Reeve 1992), and as an honest signal of status and dominance (Hart & Ratnieks 2005).
African mole-rats: behavioral and physiological mechanisms From the aforementioned observations, it is clear that the queen controls reproduction in naked mole-rat colonies, and therefore that within the current framework of skew models this would meet with one of the main assumptions of concession models. The rare cases of plural breeding occurring among queens might be explained in the context of peace incentives, as these cases always involve fighting. In the vast majority of cases where skew in females is 1.0 (i.e. a single females monopolizes reproduction), for a concession-based model to apply one would have to assume that constraints on dispersal are so high that reproductive concessions made to subordinates are reduced to zero. Recent work on paper wasps by Cant et al. (2006) builds on the peace-incentives model of Reeve & Ratnieks (1993) by incorporating animal-contest models. They predicted that escalated conflict between breeders and non-breeders is more likely when non-breeders are reproductively suppressed but relatively strong, relatedness is low, and group productivity is high. Two of these predictions were upheld in that paper wasps were more likely to escalate conflict if they were more suppressed (with less ovarian development and therefore current share of reproduction) and if groups were larger and more productive. In naked mole-rats conflict for reproductive status is seen only between the queen and high-ranking females (Margulis et al. 1995, Clarke & Faulkes 1997, 1998). These individuals are comparable to paper wasps that are “more suppressed” (Cant et al. 2006), in that they too have not bred. Studies investigating dominance and agonistic behavior in naked mole-rats have revealed that the balance of power within colonies is such that “dominant wins” may be the stable outcome. Aggression by the queen, principally in the form of shoving, may help to maintain her status, perhaps signaling her quality and thereby deterring escalated challenges from subordinates (as in paper wasps: Cant et al. 2006). In addition, this aggression induces a physiological suppression of reproduction in the subordinates. Non-breeders do not normally challenge the queen because the cost–benefit ratio of fighting is not profitable. Fights over dominance and plural breeding are very rare, indicating that the cost–benefit ratio of escalated conflict is such that subordinates do better to wait for queens to die rather than attempt to usurp and risk fatal fighting. Damaraland mole-rats
In the Damaraland mole-rat, agonistic behavior is uncommon in captive colonies, probably because they are kept in family groups. In this context, the strong incest avoidance that is known to occur in this species would be sufficient to maintain skew. However, this does not explain the observations of physiological suppression of reproduction in non-breeding females of this species, when the other incest-avoiding social Cryptomys species do not exhibit
387
388
C. G. Faulkes, N. C. Bennett such suppression. Recent studies attempting to tease apart this enigma are equivocal and demonstrate that control of reproduction in the Damaraland mole-rat is not clear-cut. Cooney & Bennett (2000) found that when captive colonies of Damaraland mole-rats with the breeding male removed were given access to unrelated males, aggression between the breeding queen and her daughters increased to high levels from previously almost non-existent. Subsequently, in 50% of colonies, rank reversals occurred and daughters usurped their mothers. This suggests that, at least in these captive colonies, unlike in the naked mole-rat, complete control by the dominant is often not achieved. Perhaps this is because naked mole-rat queens have often attained their dominant position by fighting with rivals within colonies, whereas in Damaraland mole-rats, breeders may often obtain their position in captivity simply by default following pairing of unrelated individuals. In the harsh and unpredictable habitats occupied by the Damaraland and naked mole-rats, cooperative behavior is more or less obligatory for survival, because small colonies and pairs tend to be much less successful than larger established groups, presumably as a result of there being more helpers foraging in the latter (Jarvis et al. 1994, Braude 2000). Thus in the wild, where constraints on dispersal are high, colony attrition is low, and dominant breeders need to maintain a high degree of reproductive control in groups that might be of mixed relatedness (Burland et al. 2002, 2004), competitive attempts at dominance and reproduction may occur. This appears to be very similar to the cooperatively breeding South American primate, the common marmoset (Callithrix jacchus). Socially subordinate female common marmosets are similar to naked and Damaraland mole-rats in that ovarian cyclicity and ovulation is blocked. This phenomenon is especially clear in captive peer groups of unrelated animals, where dominant control of subordinate females is achieved through a mixture of behavioral and pheromonal cues from the breeding female. In family groups studied in captivity, some variation in the degree of physiological suppression of subordinates can be observed. In this case, again, an interesting mix of dominant control and incest avoidance appears to operate, which is remarkably similar to the observations in the Damaraland mole-rat. In this context, 46% of daughters were shown to have ovulated at least once, although none became pregnant, because unrelated males were absent and daughters avoided incest (Saltzman et al. 1997, Abbott et al. 1998; see also Chapter 12). In summary, collectively these results suggest that, as with naked mole-rats, the stable resolution depends on the relative costs and benefits of escalated conflict. When the benefits of challenging are low because winning only provides access to a related male with the inherent costs/risk of inbreeding, females make no attempt to challenge the dominance of the breeder, and
African mole-rats: behavioral and physiological mechanisms indeed may be neurologically inhibited in some way from mating with a close relative. When the benefits of challenging are greatly increased due to the presence of an unrelated male, females may enter into escalated conflict. Common mole-rats
In common mole-rats there is no evidence for any physiological component to reproductive suppression, and, as with Damaraland mole-rats, inbreeding avoidance (or the costs associated with inbreeding) might play an exclusive role in maintaining reproductive skew, particularly as natal philopatry is usually short-term, and new colony formation by unrelated pairs is usually successful. However, recent new insights into the mating behavior and colony structure in the common mole-rat now bring this into question. A genetic study by Bishop et al. (2004) has shown not only that several males are achieving intra- and extra-colony paternities, but that non-breeding females may often be in contact with unrelated males. Despite this, plural breeding among females is extremely rare, raising the question as to what maintains this high skew in females? More behavioral studies of mixed-kin groups in captivity are now needed to answer this question.
Conclusions and future directions Although currently limited to three main species (naked, Damaraland, and common mole-rats), large-scale genetic and mark–release–recapture studies have enabled significant advances to be made in understanding the proximate mechanisms that may influence reproductive skew in the Bathyergidae. In particular, knowledge of the kin structure of groups allows the relative effects of incest avoidance in family groups and potential reproductive conflicts among unrelated colony members to be teased apart. Among molerats, skew is more clear-cut and extreme in females than in males, where alternative mating tactics may also occur. Furthermore, increasing knowledge of the underlying neuroendocrinology and physiology now enables appropriate models of skew to be applied or developed in the context of the observed mechanisms of reproductive control. It is clear that while there is some overlap between species of mole-rats in the proximate control of skew (physiological suppression and incest avoidance), differences are apparent in the extent to which each contributes in different species. It is also clear that mixed-kin groups may be common even where constraints on dispersal are high, and, because all individuals in colonies are reproductively totipotent, the potential for reproductive conflict is high. In this respect, mole-rats may be similar to other cooperatively breeding animals, including both vertebrates and
389
390
C. G. Faulkes, N. C. Bennett invertebrates (Hart & Ratnieks 2005; Chapter 11, this volume). Future work now needs to focus on behavioral studies of mixed-kin groups of Cryptomys in captivity, to fully evaluate the role of dominance behavior and other social cues (e. g. pheromones) in maintaining high skew among females. Genetic studies of parentage also need to be extended to include other as yet uninvestigated social species, so more comprehensive comparative analyses can be undertaken. Despite enormous efforts that have been put into field studies of naked mole-rats, the sample sizes for colonies that have been repeatedly and completely caught out over a series of consecutive years are still relatively small. Indeed, parentage has yet to be examined in wild colonies of naked mole-rats, and in particular more data are needed on breeder succession. Finally, cross-taxon comparisons of social evolution (e.g. Hart & Ratnieks 2005; Chapter 11, this volume) are increasingly revealing both idiosyncratic differences and, perhaps surprisingly, common features amongst even very divergent animal species. Acknowledgements We are most grateful for financial support from the National Research Foundation, The University of Pretoria and The Mellon Foundation (N. C. B.), and the Natural Environment Research Council (C. G. F.). Thanks to Steve Le Comber and Steve Rossiter for proofreading and many helpful comments, and to Reinmar Hager, Clara B. Jones, and two anonymous referees for further suggestions that greatly improved the manuscript. References Abbott, D. H., Saltzman, W., Schultz-Darken, N. J., & Tannenbaum, P. L. (1998). Adaptations to subordinate status in female marmoset monkeys. Comparative Biochemistry and Physiology Part C, Pharmacology, Toxicology & Endocrinology, 119, 261–274. Bennett, N. C. (1988). The trend towards sociality in three species of southern African mole–rats (Bathyergidae): causes and consequences. Unpublished Ph.D. Thesis, University of Cape Town, RSA. Bennett, N. C. & Aguilar, G. H. (1995). The reproductive biology of the giant Zambian mole-rat, Cryptomys mechowi (Rodentia: Bathyergidae). South African Journal of Zoology, 30, 1–4. Bennett, N. C. & Faulkes, C. G. (2000). African Mole-Rats: Ecology and Eusociality. Cambridge: Cambridge University Press. Bennett, N. C., Jarvis, J. U. M. & Cotterill, F. P. D. (1994). The colony structure and reproductive biology of the Mashona mole-rat, Cryptomys darlingi from Zimbabwe. Journal of Zoology, 234, 477–487.
African mole-rats: behavioral and physiological mechanisms Bennett, N. C., Faulkes, C. G., & Molteno, A. J. (1996). Reproductive suppression in subordinate, non–breeding female Damaraland mole–rats: two components to a lifetime of socially induced infertility. Proceedings of the Royal Society of London B, 263, 1599–1603. Bennett, N. C., Faulkes, C. G., & Spinks, A. C. (1997). LH responses to single doses of exogenous GnRH by social Mashona mole-rats: a continuum of socially-induced infertility in the family Bathyergidae. Proceedings of the Royal Society of London B, 264, 1001–1006. Bennett, N. C., Faulkes, C. G., & Jarvis, J. U. M. (1999). Socially induced infertility, incest avoidance and the monopoly of reproduction in cooperatively breeding African mole-rats (family Bathyergidae). Advances in the Study of Behaviour, 28, 75–114. Bennett, N. C., Molteno, A. J., & Spinks, A. C. (2000). Pituitary sensitivity to exogenous GnRH in giant Zambian mole-rats, Cryptomys mechowi (Rodentia: Bathyergidae): support for the socially induced infertility continuum. Journal of Zoology, 252, 447–452. Bishop, J. M., Jarvis, J. U. M., Spinks, A. C., Bennett, N. C., & O’Ryan, C. (2004). Molecular insight into patterns of colony composition and paternity in the common mole–rat Cryptomys hottentotus hottentotus. Molecular Ecology, 13, 1217–1229. Braude, S. (1991). The behavior and demographics of the naked mole-rat, Heterocephalus glaber. Unpublished Ph.D. Thesis, University of Michigan, USA. Braude, S. (2000). Dispersal and new colony formation in wild naked mole-rats: evidence against inbreeding as the system of mating. Behavioral Ecology, 11, 712. Brett, R. A. (1991). The ecology of naked mole-rat colonies: burrowing, food and limiting factors. In P. W. Sherman, J. U. M. Jarvis, and R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 137–184. Burda, H. (1995). Individual recognition and incest avoidance in eusocial common mole-rats rather than reproductive suppression by parents. Experientia, 51, 411–413. Burda, H. & Kawalika, M. (1993). Evolution of eusociality in the Bathyergidae: the case of the giant mole–rat (Cryptomys mechowi). Naturwissenschaften, 80, 235–237. Burda, H., Honeycutt, R. H., Begall, S., Lo¨cker-Grutjen, O., & Scharff, A. (2000). Are naked and common mole-rats eusocial and if so, why? Behavioral Ecology and Sociobiology, 47, 293–303. Burland, T. M., Bennett, N. C., Jarvis, J. U. M., & Faulkes, C. G. (2002). Eusociality in African mole–rats: new insights from patterns of genetic relatedness in the Damaraland mole–rat (Cryptomys damarensis). Proceedings of the Royal Society of London B, 269, 1025–1030. Burland, T. M., Bennett, N. C., Jarvis, J. U. M., & Faulkes, C. G. (2004). Colony structure and parentage in wild colonies of co-operatively breeding Damaraland Mole-rats suggests a role for reproductive suppression. Molecular Ecology, 13, 2371–2379.
391
392
C. G. Faulkes, N. C. Bennett Cant, M. A., Llop, J. B., & Field, J. (2006). Individual variation in social aggression and the probability of inheritance: theory and a field test. American Naturalist, 167, 837–852. Cizek, D. (2000). New colony formation in the “highly inbred” eusocial naked molerat: outbreeding is preferred. Behavioral Ecology, 11, 16. Clarke, F. M. & Faulkes, C. G. (1997). Hormonal and behavioural correlates of dominance and queen succession in captive colonies of the eusocial naked mole-rat, Heterocephalus glaber. Proceedings of the Royal Society of London B, 264, 993–1000. Clarke, F. M. & Faulkes, C. G. (1998). Hormonal and behavioural correlates of male dominance and reproductive status in captive colonies of the naked mole-rat, Heterocephalus glaber. Proceedings of the Royal Society of London B, 265, 1391–1399. Clarke, F. M. & Faulkes, C. G. (1999). Kin discrimination and female mate choice in the naked mole–rat, Heterocephalus glaber. Proceedings of the Royal Society of London B, 266, 1995–2002. Cooney, R. & Bennett, N. C. (2000). Inbreeding avoidance and reproductive skew in a cooperative mammal. Proceedings of the Royal Society of London B, 267, 801–806. Costa, J. T. & Fitzgerald, T. D. (1996). Developments in social terminology: semantic battles in a conceptual war. Trends in Ecology and Evolution, 11, 285–289. Crespi, B. J. (2005). Social sophistry: logos and mythos in the forms of cooperation. Annales Zoologici Fennici, 42, 569–571. Crespi, B. J. & Yanega, D. (1995). The definition of eusociality. Behavioral Ecology, 6, 109–115. Ellerman, J. R. (1940). The Families and Genera of Living Rodents. Vol 1. London: Trustees of the British Musuem (Natural History). Faulkes, C. G. & Abbott, D. H. (1991). Social control of reproduction in both breeding and non-breeding male naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility, 93, 427–435. Faulkes, C. G. & Abbott, D. H. (1993). Evidence that primer pheromones do not cause social suppression of reproduction in male and female naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility, 99, 225–230. Faulkes, C. G. & Abbott, D. H. (1997). Proximate mechanisms regulating a reproductive dictatorship: a single dominant female controls male and female reproduction in colonies of naked mole–rats. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 302–334. Faulkes, C. G. & Bennett, N. C. (2007). African mole-rats: behavioral and ecological diversity. In J. Wolff & P. W. Sherman, eds., Rodent Societies: an Ecological and Evolutionary Perspective. Chicago, IL: University of Chicago Press, pp. 427–437. Faulkes, C. G., Abbott, D. H., & Mellor, A. (1990a). Investigation of genetic diversity in wild colonies of naked mole-rats by DNA fingerprinting. Journal of Zoology, 221, 87–97. Faulkes, C. G., Abbott, D. H., & Jarvis, J. U. M. (1990b). Social suppression of ovarian cyclicity in captive and wild colonies of naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility, 88, 559–568.
African mole-rats: behavioral and physiological mechanisms Faulkes, C. G., Abbott, D. H., Liddell, C. E., George, L. M., & Jarvis, J. U. M. (1991). Hormonal and behavioral aspects of reproductive suppression in female naked mole-rats. In P. W. Sherman, J. U. M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 426–445. Faulkes, C. G., Bennett, N. C., Bruford, M. W., et al. (1997a). Ecological constraints drive social evolution in the African mole-rats. Proceedings of the Royal Society of London B, 264, 1619–1627. Faulkes, C. G., Abbott, D. H., O’Brien, H. P., et al. (1997b). Micro- and macrogeographic genetic structure of colonies of naked mole–rats, Heterocephalus glaber. Molecular Ecology, 6, 615–628. Faulkes, C. G., Verheyen, E., Verheyen, W., Jarvis, J. U. M., & Bennett, N. C. (2004). Phylogeographic patterns of speciation and genetic divergence in African molerats (family Bathyergidae). Molecular Ecology, 13, 613–629. Hager, R. (2003). Reproductive skew models applied to primates. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 65–101. Hamilton, W. D. (1964). The genetical evolution of social behaviour. Journal of Theoretical Biology, 7, 1–16, 17–52. Hart, A. G. & Ratnieks, F. L.W. (2005). Crossing the taxonomic divide: conflict and its resolution in societies of totipotent individuals. Journal of Evolutionary Biology, 13, 383–395. Hess, J. (2004). A population genetic study of the eusocial naked mole-rat (Heterocephalus glaber). Unpublished Ph.D. thesis, Washington University, USA. Huchon, D. & Douzery, E. J. P. (2001). From the Old World to the New World: a molecular chronicle of the phylogeny and biogeography of hystricognath rodents. Molecular Phylogenetics and Evolution, 20, 238–251. Ingram, C., Burda, H., & Honeycutt, R. L. (2004). Molecular phylogenetics and taxonomy of the African mole-rats, genus Cryptomys and the new genus Coetomys Gray, 1864. Molecular Phylogenetics and Evolution, 31, 997–1014. Jarvis, J. U. M. (1978). Energetics of survival in Heterocephalus glaber (Ru¨ppell), the naked mole–rat (Rodentia: Bathyergidae). Bulletin of the Carnegie Museum of Natural History, 6, 81–87. Jarvis, J. U. M. (1981). Eu-sociality in a mammal: cooperative breeding in naked mole-rat Heterocephalus glaber colonies. Science, 212, 571–573. Jarvis, J. U.M. (1985). Ecological studies of Heterocephalus glaber, the naked mole–rat, in Kenya. National Geographic Society Research Reports, 20, 429–437. Jarvis J. U. M. (1991). Reproduction of naked mole-rats. In P. W. Sherman, J. U. M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 384–425. Jarvis, J. U.M. & Bennett, N. C. (1993). Eusociality has evolved independently in two genera of bathyergid mole-rats – but occurs in no other subterranean mammal. Behavioral Ecology and Sociobiology, 33, 353–360.
393
394
C. G. Faulkes, N. C. Bennett Jarvis, J. U.M., O’Riain, M. J., Bennett, N. C., & Sherman, P. W. (1994). Mammalian eusociality: a family affair. Trends in Ecology and Evolution, 9, 47–51. Jarvis, J. U.M., Bennett, N. C., & Spinks, A. C. (1998). Food availability and foraging by wild colonies of Damaraland mole-rats (Cryptomys damarensis): implications for sociality. Oecologia, 113, 290–298. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Cant, M. A. (1999). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Johnstone, R. A., Woodroffe, R., Cant, M. A., & Wright, J. (1999). Reproductive skew in multimember groups. American Naturalist, 153, 315–331. Keller, L. & Perrin, N. (1995). Quantifying the level of eusociality. Proceedings of the Royal Society of London B, 260, 311–315. Kock, D., Ingram, C. M., Frabotta, L. J., Honeycutt, R. L., & Burda, H. (2006). On the nomenclature of Bathyergidae and Fukomys N. Gen. (Mammalia : Rodentia). Zootaxa, 1142, 51–55. Lacey, E. A. & Sherman, P. W. (1991). Social organization of naked mole-rat colonies: evidence for divisions of labor. In P. W. Sherman, J. U. M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 275–336. Lacey, E. A., Alexander, R. D., Braude, S. H., Sherman, P. W., & Jarvis, J. U.M. (1991). An ethogram for the naked mole-rat: Non-vocal behaviors. In P. W. Sherman, J. U.M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 275–336. Le Comber, S., Spinks, A. C., Bennett, N. C., Jarvis, J. U.M., & Faulkes, C. G. (2002). Fractal dimension of African mole-rat burrows. Canadian Journal of Zoology, 80, 436–441. Lovegrove, B. G. (1991). The evolution of eusociality in mole–rats (Bathyergidae): a question of risks, numbers and costs. Behavioral Ecology and Sociobiology, 28, 37– 45. Lovegrove, B. G. & Wissel, C. (1988). Sociality in mole–rats: metabolic scaling and the role of risk sensitivity. Oecologia, 74, 600–606. Margulis, S. W., Saltzman, W., & Abbott, D. H. (1996). Behavioral and hormonal changes in female naked mole-rats (Heterocephalus glaber) following removal of the breeding female from a colony. Hormones and Behavior, 29, 227–247. Maswanganye, K. A., Bennett, N. C., Brinders, J., & Cooney, R. (1999). Oligospermia and azoospermia in non-reproductive male Damaraland mole-rats (Cryptomys damarensis) (Rodentia: Bathyergidae). Journal of Zoology, 248, 411–418. Michener, C. D. (1969). Comparative social behaviour of bees. Annual Review of Entomology, 14, 299–342. Molteno, A. J. & Bennett, N. C. (2000). Anovulation in non-reproductive female Damaraland mole-rats (Cryptomys damarensis). Journal of Reproduction and Fertility, 119, 35–41.
African mole-rats: behavioral and physiological mechanisms Molteno, A. J. & Bennett, N. C. (2002). Relaxation of socially induced reproductive inhibition in colonies of the eusocial Damaraland mole-rat: the effect of aridity as an ecological constraint promoting philopatry. Journal of Zoology, 256, 445–448. O’Riain, M. J. & Faulkes, C. G. (2008). African mole-rats: eusociality, relatedness and ecological constraints. In J. Heinze & J. Korb, eds., Ecology of Social Evolution. Berlin: Springer-Verlag, pp. 205–220. O’Riain, M. J., Jarvis, J. U.M., & Faulkes, C. G. (1996). A dispersive morph in the naked mole-rat. Nature, 380, 619–621. O’Riain, M. J., Jarvis, J. U.M., Buffenstein, R., Alexander, R., & Peeters, C. (2000). Morphological castes in a vertebrate. Proceedings of the National Academy of Sciences of the USA 97, 13194–13197. Reeve, H. K. (1992). Queen activation of lazy workers in colonies of the eusocial naked mole-rat. Nature, 358, 147–149. Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382. Reeve, H. K. & Ratnieks, F. L.W. (1993). Queen–queen conflicts in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Reeve, H. K. & Sherman, P. W. (1991). Intracolonial aggression and nepotism by the breeding female naked mole-rat. In P. W. Sherman, J. U.M. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole-Rat. New York, NY: Princeton University Press, pp. 337–357. Reeve, H. K., Westneat, D. F., Noon, W. A., Sherman, P. W., & Aquadro, C. F. (1990). DNA “fingerprinting” reveals high levels of inbreeding in colonies of the eusocial naked mole-rat. Proceedings of the National Academy of Sciences of the USA, 87, 2496–2500. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Rickard, C. A. & Bennett, N. C. (1997). Recrudescense of sexual activity in a reproductively quiescent colony of the Damaraland mole-rat, by the introduction of a genetically unrelated male: a case of incest avoidance in “queenless” colonies. Journal of Zoology, 241, 185–202. Saltzman, W., Schultz-Darken, N. J., & Abbott, D. H. (1997). Familial influences on ovulatory function in common marmosets (Callithrix jacchus). American Journal of Primatology, 41, 159–177. Scantlebury, M., Speakman, J. R., Oosthuizen, M. K., Roper, T. J, & Bennett, N. C. (2006). Energetics reveals physiologically distinct castes in a eusocial mammal. Nature, 440, 795–797. Sherman, P. W., Jarvis, J. U.M., & Braude, S. H. (1992). Naked mole-rats. Scientific American, 267, 72–78. Sherman, P. W., Lacey, E. A., Reeve, H. K., & Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108.
395
396
C. G. Faulkes, N. C. Bennett Smith, T. E., Faulkes, C. G., & Abbott, D. H. (1997). Combined olfactory contact with the parent colony and direct contact with non-breeding animals does not maintain suppression of ovulation in female naked mole-rats. Hormones and Behavior, 31, 277–288. Spinks, A. C., Jarvis, J. U. M., & Bennett, N. C. (2000a). Comparative patterns of philopatry and dispersal in two common mole-rat populations: implications for the evolution of mole-rat sociality. Journal of Animal Ecology, 69, 224–234. Spinks, A. C., Bennett, N. C., Faulkes, C. G., & Jarvis, J. U. M. (2000b). Circulating LH levels and the response to exogenous GnRH in the common mole-rat: implications for reproductive regulation in this social, seasonal breeding species. Hormones and Behavior, 37, 221–228. Van Daele, P. A. A. G., Dammann, P., Meier, J. L., et al. (2004). Chromosomal diversity in mole-rats of the genus Cryptomys (Rodentia: Bathyergidae) from the Zambezian region: with descriptions of new karyotypes. Journal of Zoology, 264, 317–326. Van Daele, P. A. A. G., Faulkes, C. G., Verheyen, E., & Adrians, D. (2007). African molerats (Bathyergidae): a complex radiation in Afrotropical soils. In S. Begall, H. Burda, & C. E. Schleich, eds., Subterranean Rodents: News from Underground. Heidelberg: Springer-Verlag, pp. 357–373. Wallace, E. & Bennet, N. C. (1998). The colony structure and social organisation of the giant Zambian mole-rat, Cryptomys mechowi. Journal of Zoology, 244, 51–61.
14
The causes of physiological suppression in vertebrate societies: a synthesis andrew j. young
Summary In many vertebrate societies, skews in reproductive success are underpinned by the down-regulation of one or more components of the reproductive physiology of subordinates relative to those of their same-sex dominants, a condition termed physiological suppression. The considerable body of research into the causes of physiological suppression is therefore of key relevance to attempts to understand the causes of reproductive skew. In this chapter, I draw this work together by outlining a single adaptive framework for understanding the causes of physiological suppression across vertebrate societies in general. I suggest that two largely distinct adaptive explanations exist for physiological suppression in a given subordinate at any given time: “subordinate restraint” and “active interference” by the dominant(s). Both processes may act in tandem to maintain physiological suppression across subordinates in any given species, with dominants employing active interference to counter, or guard against, any lapses in subordinate restraint. Subordinates may be expected to exercise physiological restraint whenever their expected fitness payoff from maintaining their fertility drops below zero. Any factors that diminish a subordinate’s expected fitness payoff from maintaining its fertility could therefore favor the evolution of restraint, including those that may act regardless of the presence of the dominant and those that arise directly from the presence and/or likely actions of the dominant. Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
397
398
A. J. Young Dominants may attempt to actively and forcibly disrupt the reproductive physiology of subordinates whenever (1) they stand to benefit from disrupting subordinate reproduction, and (2) fertility disruption per se is a favored means of achieving this (as opposed to other disruptive tactics such as infanticide). Dominants may actively disrupt subordinate fertility by subjecting them to chronic physiological stress. The evidence to date suggests that physiological suppression commonly arises, at least in part, from subordinates exercising restraint. Future research should investigate the extent to which this restraint can realistically be accounted for without invoking some role for the dominant (e.g. the threat of disruptive action should the subordinate fail to exercise restraint). Active interference by dominants, by means of socially induced stress, may also play a key role in precipitating physiological suppression in a range of species, potentially acting in tandem with subordinate restraint. Future research must allow for the possibility that dominants target only a subset of their subordinates at critical times, to counter, or guard against, any lapses in restraint. Finally, our understanding of the causes of physiological suppression in vertebrate societies has broad implications for empirical and theoretical attempts to explain variation in reproductive skew. These are discussed.
Introduction In many animal societies, subordinate individuals show markedly lower reproductive success than their same-sex dominants. A considerable body of empirical and theoretical research has now sought to explain these rank-related reproductive disparities, or reproductive skews, and the variation in their extent across contexts (reviewed in Johnstone 2000, Magrath et al. 2004, and elsewhere in this volume). In parallel with this work, numerous studies have now revealed that the low reproductive rates of subordinates are often underpinned by the down-regulation of one or more components of their reproductive physiology relative to that of their same-sex dominants, a condition commonly termed physiological suppression (Table 14.1). As physiological suppression may commonly reflect compromised fertility among subordinates, seeking to understand its causes, in particular whether it arises from disruptive tactics employed by dominants or from restraint exercised by subordinates, has become a key focus of research at the interface between behavioral ecology and endocrinology (e.g. Reyer et al. 1986, Bennett et al. 1996, Snowdon 1996, Abbott et al. 1997, Creel 2001, Schoech et al. 2004, Young et al. 2006). This growing body of behavioral–endocrine research has yielded valuable insights into the causes of physiological suppression in a broad range of focal
399
400
Table 14.1 Evidence of physiological suppression, defined here as the down-regulation of one or more components of the reproductive physiology of subordinates relative to that of their same-sex dominant(s), has now been reported at every level of the vertebrate reproductive axis. For a brief overview of the hypothalamic–pituitary–gonadal axis see Schoech et al. (2004) (for more detail see Johnson & Everitt 1999, Becker et al. 2002, Nelson 2004). Importantly, physiological suppression as defined here refers merely to reproductive physiological differences between dominants and subordinates, without any intended reference to the cause of those differences (i.e. it does not mean that they necessarily arise from “suppression” imposed by the dominant; they might equally arise from restraint exercised by subordinates: Box 14.1). I continue to use the term for consistency with previous work, though were we to reconsider our terminology, “physiological differences” might be more appropriate. Level of the reproductive endocrine (hypothalamic– pituitary–gonadal) axis
Examples of species in which physiological suppression has been detected at this level
Hypothalamus (storage and
Damaraland mole-rats Cryptomys damarensis (Molteno et al. 2004)
release of GnRH) Pituitary (sensitivity to GnRH challenge)
Naked mole-rats Heterocephalus glaber (Faulkes et al. 1991), Damaraland mole-rats (Bennett et al. 1999), common marmosets Callithrix jacchus, (Abbott et al. 1988), meerkats Suricata suricatta (Young et al. 2006)
Gonadotropic hormones (mean LH, CG, or FSH levels)
Naked mole-rats (Faulkes et al. 1991), Damaraland mole-rats (Bennett et al. 1999), common marmosets (Abbott et al. 1981), cotton-top tamarins Saguinus oedipus (Ziegler et al. 1987), meerkats (O’Riain et al. 2000, Carlson et al. 2004), Harris’s hawks Parabuteo unicinctus (Mays et al. 1991)
Reproductive steroid hormones (mean androgen, estrogen, and progestagen levels)
Naked mole-rats (Faulkes & Abbott 1997), alpine marmots Marmota marmota, (Arnold & Dittami 1997), common marmosets (Saltzman et al. 1998), cotton-top tamarins (French et al. 1984, Ziegler et al. 1987), meerkats (Carlson et al. 2004, Young et al. 2008), dwarf mongooses Helogale parvula (Creel et al. 1992), African wild dogs Lycaon pictus (Creel et al. 1997), Harris’s hawks (Mays et al. 1991), white-browed sparrow weavers, Plocepasser mahali (Wingfield et al. 1991), Florida scrub jays Aphelocoma coerulescens (Schoech et al. 1991), super fairy-wrens Malurus cyaneus (Peters et al. 2001, 2002)
Ovaries (ovulation or ovarian structure)
Naked mole-rats (Faulkes & Abbott 1997), Damaraland mole-rats (Molteno & Bennett 2000), common marmosets (Abbott 1984, Saltzman et al. 1997), pygmy marmosets Cebuella pygmaea (Carlson et al. 1997), cotton-top tamarins (French et al. 1984, Ziegler et al. 1987), white-browed sparrow weavers (Wingfield et al. 1991), Florida scrub jays (Schoech et al. 1996)
Testicular parameters (testicular or ejaculate properties)
Naked mole-rats (Faulkes et al. 1991), Damaraland mole-rats (Maswanganye et al. 1999), common marmosets (Abbott 1993), white-browed sparrow weavers (Wingfield et al. 1991), bell miners Manorina melanophrys (Poiani & Fletcher 1994), Florida scrub jays (Schoech et al. 1996), cichlid fish Neolamprologus pulcher (Fitzpatrick et al. 2006)
Causes of physiological suppression in vertebrate societies taxa (e.g. callitrichid primates: Chapter 12, French 1997; rodents: Chapter 13, Carter & Roberts 1997; carnivores: Creel et al. 1996, Young et al. 2006; birds: Reyer et al. 1986, Schoech et al. 1991, 2004, Wingfield et al. 1991; and fish: Buchner et al. 2004). However, comparatively little attention has been given to the challenge of drawing these insights together into a single adaptive framework for understanding the causes of physiological suppression in vertebrate societies in general. Such a framework would offer a much-needed focal point for debate for those seeking to advance our broader understanding of the causes of physiological suppression across animal societies, as well as a useful overview for those investigating the causes of physiological suppression in new model systems. Such a synthesis could also prove of particular value to empiricists and theoreticians seeking to understand the causes of variation in reproductive skew, who may be less familiar with the relevant insights from the behavioral–endocrine literature. In this chapter, I therefore outline a general adaptive framework for understanding the causes of physiological suppression in vertebrate societies (summarized in Box 14.1). I first introduce the key concepts of subordinate restraint and active interference, which I argue can be usefully considered two broadly distinct adaptive explanations for physiological suppression. I then consider in detail the range of factors that may act in concert to select for these two processes, and briefly review the evidence available in support of a role for each. I have sought throughout to highlight those areas where our knowledge remains poorly developed, and, in the penultimate section, suggest four areas in particular where research into the causes of physiological suppression might be profitably directed in the future. I close by considering the implications that our understanding of the causes of physiological suppression within vertebrate societies has for empirical and theoretical attempts to explain variation in reproductive skew across vertebrate societies. As with any attempt to draw together a broad body of research, a range of more or less complicated approaches exist, each with its own merits and associated problems. With the goal of maximizing the generality of such a synthesis, I have chosen to outline what I see as the most straightforward adaptive framework for understanding the causes of physiological suppression, highlighting in the process some of the key caveats with this approach. While these caveats could be readily incorporated into the framework outlined here, doing so at this stage would risk losing more in complication than was gained in accuracy, given the limits of our current understanding.
401
402
A. J. Young Box 14.1 The causes of physiological suppression in vertebrate societies: a synthesis I suggest that two broadly distinct adaptive explanations exist for physiological suppression in a given subordinate at any given time: physiological restraint exercised by the subordinate and active interference by the dominant(s). Both processes may act in tandem to maintain physiological suppression across subordinates in any given species, with dominants employing active interference to counter, or guard against, any lapses in subordinate restraint.
Physiological suppression arising from subordinate restraint In this scenario, subordinates down-regulate their own reproductive physiology because, given their circumstances, they stand to gain no net fitness payoff from maintaining their fertility.
Subordinate restraint will therefore be favored by any factors that diminish a subordinate’s expected fitness payoff from maintaining its fertility. Two broad classes of factors exist: (1) Factors that may act regardless of the presence of the dominant (e.g. a lack of unrelated breeding partners, poor body condition, limited breeding experience) (2) Factors that derive from the presence and/or likely actions of the dominant (e.g. the threat of infanticide or punishment should the subordinate breed, or costs arising from competition between the subordinate’s young and those of the dominant) Additional factors may predispose particular species or sexes to evolving and exercising physiological restraint, by modifying the costs and benefits of maintaining fertility and breeding for all individuals, dominants and subordinates alike (e.g. high energetic costs of breeding). Physiological suppression arising from active interference by dominants
In this scenario, dominants actively employ tactics to forcibly down-regulate the reproductive physiology of subordinates. Active interference may be favored when (1) dominants stand to benefit from actively disrupting subordinate reproduction, and (2) forcibly downregulating the reproductive physiology of subordinates is a favored means of achieving this (as opposed to infanticide, for example). If active
Causes of physiological suppression in vertebrate societies Box 14.1 (cont.) interference becomes sufficiently effective to reduce below zero a subordinate’s expected fitness payoff from attempting to maintain its fertility, selection may favor subordinates who exercise restraint.
It is unlikely that dominants can forcibly disrupt the fertility of subordinates using pheromones. Pheromones probably act as signals, in response to which subordinates may exercise restraint. Dominants may forcibly down-regulate subordinate fertility by subjecting them to chronic stress. While social stress alone cannot readily account for physiological suppression across all vertebrate societies, it may still play a key role in a range of species, potentially acting in tandem with subordinate restraint. Future studies must allow for the possibility that dominants target only a subset of their subordinates at critical times, to counter, or guard against, any lapses in subordinate restraint.
The causes of physiological suppression in vertebrate societies Any attempt to formulate adaptive explanations for a physiological phenomenon, such as physiological suppression, requires an understanding of the functional consequences that it will have for the individuals in which it occurs. As the primary consequence of physiological suppression across vertebrate societies is likely to be a degree of compromised fertility among subordinates, I have made the simplifying assumption throughout that adaptive explanations for physiological suppression should take the form of adaptive explanations for compromised fertility. While this assumption should generally prove reasonable, physiological suppression may sometimes at least partly reflect rank-related differences in the expression of traits other than fertility that are also regulated by the reproductive endocrine axis. When this is the case, the link between physiological suppression and subordinate infertility may be weakened (Box 14.2). If we assume then that physiological suppression does lead to compromised fertility, why should subordinates exhibit physiological suppression? It has long been recognized, in one form or another, that two main parties could be responsible for physiological suppression among subordinate vertebrates: subordinates could be down-regulating their own fertility or dominants could be disrupting subordinate fertility (e.g. Wasser & Barash 1983, Reyer et al. 1986, Brown 1987, Wingfield et al. 1991, Snowdon 1996). While these are frequently
403
404
A. J. Young Box 14.2 The link between physiological suppression and subordinate infertility I make the simplifying assumption, throughout this chapter, that adaptive explanations for physiological suppression (as defined in Table 14.1) should take the form of adaptive explanations for compromised fertility. While this assumption should generally prove reasonable, physiological suppression may sometimes reflect rank-related differences in the expression of traits other than fertility, that are also regulated by components of the reproductive endocrine axis. For example, while differences in the mean testosterone levels of dominant and subordinate males could underpin or reflect differences in their underlying fertility (e.g. by modifying testicular function: Desjardins & Turek 1977, Wingfield et al. 1990), they might also reflect differences in their mean investment in other androgen-dependent behaviors or traits (e.g. territorial aggression: Wingfield et al. 1990; or sexual signals: Folstad & Karter 1992) or in activities that alter androgen levels (e.g. prospecting for extra-group matings: Young et al. 2005). Indeed, as viable sperm production can occur at very low testosterone titres (Wingfield et al. 1990), subordinate males with lower testosterone titres than dominants, while technically “physiologically suppressed,” may prove just as fertile as dominants. Similar issues may arise for other components of the reproductive endocrine axes of both males and females, but as androgens regulate the expression of such a broad range of traits (see Wingfield et al. 1990, Folstad & Karter 1992, Alonso-Alvarez et al. 2007, Hau 2007) the fertility consequences of differences in mean androgen levels should be interpreted with particular caution. Where physiological suppression does partly reflect rank-related differences in the expression of traits other than fertility, a complete understanding of the ultimate causes of physiological suppression will require a knowledge of the range of traits that are regulated by the reproductive endocrine axis and the selective pressures that act upon each.
couched as alternative explanations for reproductive disparities between dominants and subordinates, the conceptual distinction between the two is often unclear, as subordinates may down-regulate their own fertility because dominants would otherwise do so forcibly. As such there is a need for precision when considering alternative adaptive explanations for physiological
Causes of physiological suppression in vertebrate societies suppression (or indeed rank-related disparities in reproductive success) and for consistency across studies in the framework and terminology employed. I suggest that a simple and general adaptive framework for understanding the causes of physiological suppression in vertebrate societies arises from contrasting scenarios where physiological suppression arises from subordinate restraint with those where it arises from active interference by other parties. Subordinate restraint refers to scenarios where physiological suppression arises from subordinates adaptively down-regulating their own reproductive physiology as, given their circumstances, they stand to gain no net fitness payoff from maintaining their fertility. Subordinate restraint may therefore be favored by any factors that reduce a subordinate’s expected fitness payoff from maintaining its fertility. This will include factors that may act regardless of the presence of dominants (e.g. a lack of access to unrelated breeding partners or poor body condition) and factors arising directly from the presence and/or likely actions of dominants (e.g. the likelihood that the dominant would disrupt a subordinate’s fertility or kill its young should the subordinate cease to exercise restraint). Active interference refers to scenarios where physiological suppression arises from one or more individuals (most likely socially dominant individuals) actively and forcibly down-regulating the reproductive physiology of subordinates (e.g. by subjecting them to chronic stress). Active interference may therefore be favored whenever (1) dominants stand to benefit from actively disrupting subordinate reproduction, and (2) disrupting subordinate fertility per se is a favored means of achieving this (as opposed to infanticide, for example). Whenever active interference becomes sufficiently effective to eliminate any net fitness payoff that a subordinate might otherwise have gained from attempting to maintain its fertility (by reducing its chances of successful reproduction below some critical threshold), selection may generally be expected to favor subordinates who exercise restraint. If subordinate restraint is considered a dichotomous phenomenon, with subordinates exercising either no restraint or complete restraint depending on their circumstances, then subordinate restraint and active interference can be usefully considered distinct alternative explanations for physiological suppression in a given subordinate at any given time, as active interference may rarely be necessary while a subordinate exercises complete restraint. The two processes may still act in tandem to maintain physiological suppression across subordinates in any given species, however, as dominants may resort to active interference whenever a particular subordinate ceases to exercise restraint (Young et al. 2006). The erection of such a distinction between subordinate restraint and active interference, while doubtless simplified (see schematic in
405
406
A. J. Young Box 14.3 A simple adaptive framework for understanding the causes of physiological suppression With subordinate restraint considered a dichotomous phenomenon (subordinates exercise either no restraint or complete restraint), subordinate restraint and active interference can be seen as alternative explanations for physiological suppression. Although they are distinct, the two processes may act in tandem to maintain physiological suppression across subordinates in any given species. Active interference by the dominant may rarely be necessary while a subordinate exercises complete restraint, but a dominant may respond with active interference to a subordinate that ceases to exercise restraint.
Box 14.3 and caveats in Box 14.4), seems the most straightforward way to formalize general explanations for physiological suppression across vertebrate societies, and, given the limits of our knowledge, is probably also the most useful at present. In the following two sections I discuss the concepts of subordinate restraint and active interference in more detail, considering the factors that may contribute to selection for them and the evidence available in support of a role for each. The key points that I raise throughout are summarized in the overview provided in Box 14.1.
Physiological suppression arising from subordinate restraint Considerable evidence now suggests that physiological suppression may commonly arise, at least in part, from subordinates exercising restraint: subordinates down-regulating their own reproductive physiology so as to maximize their fitness (e.g. Wasser & Barash 1983, Snowdon 1996, Abbott et al. 1997, Cooney & Bennett 2000, Creel 2001, Schoech et al. 2004, Young et al. 2008, see also Johnstone & Cant 1999b, Crespi & Ragsdale 2000, Hamilton 2004, Wenseleers et al. 2004 for theoretical treatments of reproductive restraint; see also Box 14.5). While subordinate restraint is now frequently invoked, in one form or another, to explain physiological suppression, little attempt has been made to draw together the broad range of factors that may contribute to selection for restraint across vertebrate societies into a single adaptive framework.
Causes of physiological suppression in vertebrate societies Box 14.4 The distinction between active interference and subordinate restraint Active interference and subordinate restraint may be usefully considered distinct alternative explanations for physiological suppression in a given subordinate at any given time (Box 14.3). Such a distinction may be generally valid, but it becomes blurred in certain situations. It may then be more appropriate to consider pure active interference and subordinate restraint as two ends of a continuum. For example, if restraint is considered a graded phenomenon (rather than a dichotomous one), one might imagine an intermediate level of subordinate restraint being met with active interference by the dominant (whenever the subordinate’s own optimal degree of [in]fertility differs from that which is optimal for the dominant). However, one could preserve the dichotomy by arguing that the observed degree of physiological suppression is ultimately arrived at through active interference by the dominant. In a second example, while active interference may generally prove unnecessary while a subordinate exercises complete restraint, selection might favor dominants who actively interfere regardless of a subordinate’s fertility state, if they cannot reliably discriminate those subordinates exercising restraint from those who are not. Subordinates might then be subjected to active interference despite exercising restraint. Scenarios of this kind could be comparatively rare, however, as selection may strongly favor subordinates who signal their infertility to dominants, if doing so absolves them from needless persecution.
Broadly speaking, exercising physiological restraint should be adaptive whenever an individual’s expected fitness payoff from attempting to maintain its fertility falls below zero. Physiological restraint among subordinates should therefore be favored by factors that reduce their expected fitness payoffs from attempting to maintain their fertility and breed, more so than those of dominants. A range of such factors exist, which can be usefully divided into two broad classes: (1) factors that may act regardless of the presence of the dominant (e.g. a lack of access to unrelated breeding partners or poor body condition), and (2) factors arising directly from the presence and/or likely actions of the dominant (e.g. the dominant’s ability to kill the subordinate’s young should the subordinate attempt to breed). Whether subordinate restraint is
407
408
A. J. Young Box 14.5 Subordinate restraint: physiological down-regulation or an absence of up-regulation? For simplicity, I have defined subordinate restraint as referring to situations where physiological suppression arises from subordinates adaptively down-regulating their own reproductive physiology. This definition involves the implicit assumption that the reproductive physiology of dominants in some sense reflects the “default” state, and that subordinate restraint must therefore involve physiological downregulation mechanisms. This needn’t necessarily be the case, however: subordinate restraint might also arise from a lack of physiological upregulation among subordinates (e.g. when they lack key stimuli to which dominants are exposed, such as those from a suitable breeding partner: Schoech et al. 1996, 2004). When considering evolutionary explanations for physiological suppression, though, the key point is that, regardless of the proximate mechanisms through which it arises, the end result of subordinate restraint (e.g. lower reproductive hormone levels among subordinates than dominants) is adaptive for subordinates. The arguments that I outline here, regarding the range of factors that may act in concert to favor the evolution of subordinate restraint, should therefore hold whether that restraint arises through physiological down-regulation, a lack of up-regulation, or a combination of the two.
favored in a given context should depend upon whether the combined effects of all such factors eliminate any net fitness payoff that the subordinate might otherwise have gained from maintaining its fertility. I discuss these two sets of factors in turn below, explaining the logic underpinning them and highlighting the evidence available in support of a role for each. I then consider how a number of additional species- and/or sex-specific factors that alter the costs or benefits of maintaining fertility for all individuals (subordinates and dominants alike) may predispose certain species and/or sexes to evolving and exercising physiological restraint. Restraint due to factors that may act regardless of the presence of the dominant
Perhaps the simplest explanation for subordinates exercising physiological restraint is that factors other than the presence of their same-sex dominant reduce their expected fitness payoff from maintaining their fertility. The strongest evidence in support of explanations of this kind is that
Causes of physiological suppression in vertebrate societies physiological suppression sometimes persists even after subordinates are separated from their same-sex dominant (e.g. Clarke et al. 2001; see also Bennett et al. 1996), suggesting that other factors, correlated with their subordinate status, were reducing their expected payoff from maintaining their fertility. The key factor that may favor physiological restraint among subordinates regardless of the presence of their dominant is a lack of access to unrelated breeding partners, a predicament commonly faced by subordinate vertebrates, who may have delayed dispersal from their natal groups (Koenig & Haydock 2004). A subordinate’s expected fitness payoff from maintaining its fertility may be substantially reduced under such circumstances, due to the inbreeding depression commonly suffered by offspring from incestuous matings (Pusey & Wolf 1996, Koenig & Haydock 2004; inbreeding depression has now been documented in a range of social vertebrates, e.g. common moorhens, Gallinula chloropus: McRae 1996; red-cockaded woodpeckers, Picoides borealis: Daniels & Walters 2000; mashona mole-rats, Cryptomys darlingi: Greeff & Bennett 2000). Accordingly, subordinates lacking access to unrelated breeding partners within their groups typically refrain from within-group matings (e.g. female common marmosets, Callithrix jacchus: Saltzman 2003, Saltzman et al. 2004; meerkats, Suricata suricatta: O’Riain et al. 2000, Young et al. 2007; Damaraland mole-rats, Cryptomys damarensis, Figure 14.1: Cooney & Bennett 2000; but see Reeve et al. 1990 and Jarvis et al. 1994 for naked mole-rats, Heterocephalus glaber) and often exhibit down-regulated reproductive physiologies relative to those with immediate access to outbreeding opportunities (e.g. female Damaraland mole-rats: Cooney & Bennett 2000; golden lion tamarins, Leontopithecus rosalia: French et al. 2003; meerkats: Clutton-Brock et al. 2001; Carlson et al. 2004). Indeed, in some species, physiological restraint due to inbreeding avoidance alone appears sufficient to explain physiological suppression among subordinates (e.g. male pied kingfishers, Ceryle rudis: Reyer et al. 1986; male redcockaded woodpeckers: Khan et al. 2001). Given the clear influence that patterns of relatedness among group members can have on their reproductive physiology and behavior, any attempt to identify the causes of physiological suppression should carefully consider the likely role of inbreeding avoidance. Subordinates do not always exercise physiological restraint when lacking unrelated breeding partners within their groups, however (e.g. Schoech et al. 1996, Saltzman et al. 2004, Young et al. 2008). This may be the case in some species simply because incest is not strictly avoided (e.g. Jarvis et al. 1994, Keane et al. 1996, McRae 1996, Clarke & Faulkes 1999, Koenig & Haydock 2004), but perhaps a more general explanation is that sporadic opportunities to outbreed may still arise sufficiently often to make it adaptive for subordinates
409
410
A. J. Young
Figure 14.1 Physiological suppression among female Damaraland mole-rats is thought to arise from subordinates exercising restraint, due to both their lack of access to unrelated breeding partners and the presence of the dominant female. Photo by Andrew Young.
to maintain their fertility regardless (e.g. opportunities for extra-group matings: Double & Cockburn 2003, Bishop et al. 2004, Young et al. 2007). A number of other factors that have received markedly less attention may also devalue a subordinate’s expected fitness payoff from maintaining its fertility regardless of the presence of the dominant, thereby favoring the evolution of physiological restraint. For example, reproductively mature females commonly experience lower breeding success when young (Wasser & Barash 1983, Forslund & Part 1995), when in poor body condition (Wasser & Barash 1983, Russell et al. 2003), or when they have little prior breeding experience (Wasser & Barash 1983, Cichon 2003). As subordinate individuals are commonly younger, lighter, and less experienced than their same-sex dominants (e.g. Creel & Creel 1991, Clutton-Brock et al. 2001), these factors may often devalue the average subordinate’s expected fitness payoff from maintaining its fertility more so than that of dominants. Comparatively few studies have investigated whether these factors do contribute to selection for physiological restraint, but evidence that mature social vertebrates often do exhibit lower circulating reproductive hormone levels when young or in poor body condition suggests that they could well play a role (e.g. Wasser & Barash 1983,
Causes of physiological suppression in vertebrate societies Arnold & Dittami 1997, Carlson et al. 2004, Young et al. 2008). While these factors may often contribute to selection for restraint among subordinates by yielding graded variation across all group members in their expected fitness payoffs from maintaining their fertility, it seems unlikely that these factors alone could account entirely for the marked disparities in the reproductive physiologies of dominants and subordinates that are typical of many highskew species (Table 14.1). Such clear physiological disparities may more commonly reflect subordinate restraint due primarily to a lack of unrelated breeding partners (which may often fall exclusively upon subordinates, having delayed dispersal from their natal groups) or to the presence and/or likely actions of their same-sex dominant (see below). Restraint due to the presence and/or likely actions of the dominant
In many animal societies, one factor that may markedly reduce a subordinate’s expected fitness payoff from attempting to maintain its fertility (thereby acting in concert with the factors outlined above to favor the evolution of physiological restraint), is the presence and/or likely actions of the same-sex dominant. Support for this idea stems from evidence that physiological suppression in some species cannot be readily explained by social stress imposed by dominants (i.e. active interference: see below), but is nevertheless lifted when the subordinate and dominant are separated (e.g. Abbott et al. 1997, Faulkes & Abbott 1997). Moreover, physiological suppression can sometimes be prolonged in the absence of the dominant by the regular presentation of cues that would normally indicate their presence (e.g. Barrett et al. 1990; see also Barrett et al. 1993, Abbott et al. 1997, Chapter 12 in this volume). While the possibility that subordinates exercise restraint due to the presence and/or likely actions of their same-sex dominant has been widely explored by theoretical approaches to reproductive skew (see Johnstone & Cant 1999a, 1999b, Hager & Johnstone 2004, Hamilton 2004, Wenseleers et al. 2004), comparatively few studies of physiological suppression have investigated this possibility in any detail. This is unfortunate, as subordinate restraint due in part to the presence and/or likely actions of the dominant may ultimately prove a pervasive cause of physiological suppression in cooperative vertebrate societies (Abbott et al. 1997, Creel 2001, Young et al. 2008, Chapters 12 and 13 in this volume). Subordinates may exercise physiological restraint in the presence of their same-sex dominant for at least two reasons: (1) the dominant may be capable of disrupting or punishing attempts by the subordinate to maintain its fertility or breed (see Johnstone & Cant 1999a, 1999b, Hager & Johnstone 2004, Hamilton 2004; Wenseleers et al. 2004 for relevant theory); and/or (2) regardless of the dominant’s capacity to interfere, subordinate females may gain a markedly
411
412
A. J. Young reduced inclusive fitness payoff from attempting to breed alongside the dominant, if their own offspring must compete with those of the dominant. A key distinction between the two is that the former requires the dominant to enforce restraint whereas the latter does not. Both processes could act in concert to favor subordinate restraint, though costs to the subordinate arising from offspring competition may be largely irrelevant if dominants have the capacity to disrupt or effectively punish all subordinate attempts to breed. Although studies of physiological suppression have yet to investigate the relative importance of these processes, I briefly consider their possible roles below. Perhaps the most parsimonious explanation for subordinates exercising physiological restraint in the presence of their same-sex dominant is that the dominant has the capacity to disrupt and/or punish subordinate attempts to breed. Dominants are known to disrupt subordinate reproduction using a range of tactics (e.g. mate guarding: Komdeur et al. 1999; harassment: Wasser & Barash 1983, Young et al. 2006; infanticide: Digby 2000), and growing evidence suggests that dominants may often punish their subordinates by evicting them from the group (e.g. Dierkes et al. 1999, Young et al. 2006, Wong et al. 2007). While the threat of disruptive action by the dominant provides a plausible explanation for subordinates exercising physiological restraint in their presence (e.g. Johnstone & Cant 1999b, Hager & Johnstone 2004, Hamilton 2004), few physiological studies have sought evidence of a direct link between the two. One finding that is at least suggestive of such a link stems from a recent study of meerkat societies: the average estrogen levels of subordinate females vary in association with temporal variation in the likelihood that any litter they conceive will be met with infanticide by the dominant, a pattern that cannot be explained by correlated variation in either subordinate glucocorticoid (GC) levels or body condition (Young et al. 2008). One plausible explanation for this pattern is that subordinates modulate their own estrogen levels so as to avoid conceiving at times when their litters are likely to be killed by the dominant. While a number of theoretical models have explored the possibility that subordinates exercise reproductive restraint due to the threat of dominant interference, studies of physiological suppression have yet to test their predictions (regarding relationships between the extent of restraint and other key variables, such as dominant–subordinate relatedness or the strength of constraints on independent breeding: Johnstone & Cant 1999a, 1999b, Hager & Johnstone 2004, Hamilton 2004, Wenseleers et al. 2004). Attempts to do so could now prove instructive, if close attention was focused first on testing the models’ assumptions (Johnstone 2000, Magrath et al. 2004). The threat of dominant interference might also precipitate restraint among subordinates if dominants were capable of disrupting or punishing
Causes of physiological suppression in vertebrate societies subordinate attempts to maintain their fertility (as opposed to subordinate attempts to breed, as discussed above). While this possibility remains largely unexplored, it could favor the evolution of subordinate restraint even in species where dominants can neither detect nor police subordinate attempts to breed. For restraint to evolve through this route, dominants would have to be capable, at least to some extent, of both detecting those subordinates not exercising restraint and disrupting their fertility or punishing them. There is considerable evidence that the reproductive states of female mammals can be discriminated by conspecifics (particularly by males, to whom fertile females may commonly advertise, but also by other females: e.g. Ziegler et al. 1993, Smith & Abbott 1998, Washabaugh & Snowdon 1998, Snowdon et al. 2006), but whether this is the case among males or in other taxa is less clear. Dominants in some species do appear capable of punishing or disrupting the fertility of subordinates (e.g. Clutton-Brock & Parker 1995b, Dierkes et al. 1999, Young et al. 2006, Wong et al. 2007), but whether they subject subordinates to either on the basis of the subordinate’s fertility per se remains unclear. Selection on dominant females to disrupt subordinate reproduction may typically stem from costs that would otherwise arise from competition between their own young and those of subordinates (see Chapter 15). Few studies have considered the fact that subordinate females would commonly experience similar costs arising from offspring competition if they were to breed successfully alongside the dominant. These costs could be high if the offspring of dominants were superior competitors (S. J. Hodge et al., unpublished data) and, in societies of close kin, may be compounded by indirect fitness costs arising from any detrimental effects of competition on the dominant’s young (e.g. Gerlach & Bartmann 2002). Indeed, these indirect costs could be acute in monogamous family groups, where subordinates that are daughters of the dominant female may be as closely related to the dominant’s young as they are to their own (Reeve & Keller 1995, 1996). By reducing the subordinate’s expected fitness payoff from maintaining its fertility and breeding alongside the dominant, these inclusive fitness costs arising from offspring competition may contribute to selection for restraint, regardless of any threat of dominant interference. The evolution of subordinate restraint due to costs of this kind provides one explanation for the finding that, in social-insect societies with immobile queens, subordinates may exercise restraint in her presence despite her inability to attack or punish them (Keller & Nonacs 1993; see also Endler et al. 2004). Studies of physiological suppression in vertebrate societies have yet to investigate whether costs arising from offspring competition may contribute to selection for subordinate restraint. If subordinate females were exercising
413
414
A. J. Young physiological restraint due in part to costs arising from offspring competition, one might predict that their restraint should be lifted if the dominant fails to breed. A number of studies provide support for this prediction. For example, reproduction by subordinate females in Mongolian gerbil (Meriones unguiculatus) groups frequently occurs when the dominant’s inter-birth intervals are unusually long, and the likelihood of subordinate reproduction increases if the dominant’s young are removed (Payman & Swanson 1980, French 1994). Similarly, the contraceptive treatment of dominant females (and/or the resulting lack of offspring) is associated with an increased likelihood of ovulation among subordinate females in common marmoset groups (Saltzman et al. 1997), and precipitates heightened intra-group aggression in goldenheaded lion tamarin (Leontopithecus chrysomelas) groups (though there was no clear change in reproductive activity: De Vleeschouwer et al. 2000, 2003). A focused investigation of the role that the costs of offspring competition to subordinates may play in precipitating restraint, regardless of any threat of dominant interference, could therefore prove rewarding. Care would need to be taken throughout, to address the possibility that any apparent release of restraint when dominants failed to breed is due instead to an associated reduction in the threat of dominant interference (e.g. dominant female meerkats only evict subordinates and kill their young when pregnant themselves: Young et al. 2006, Young & Clutton-Brock 2006). Taxon- or sex-specific factors may facilitate the evolution of physiological restraint
The array of factors outlined above can explain why subordinates in a given species should consistently exercise physiological restraint while their same-sex dominants do not, as they may all differentially reduce subordinates’ expected fitness payoffs from maintaining their fertility and attempting to breed. For a complete understanding of the factors that favor the evolution of physiological restraint, however, we must also consider a range of taxon- or sex-specific factors that may predispose particular taxa or sexes to evolving restraint, by affecting the benefits and costs to all individuals (dominants and subordinates alike) of maintaining their fertility and breeding or exercising restraint. For example, the costs of fertility maintenance and breeding may be higher for females in species where males harass fertile females (Clutton-Brock & Parker 1995a), where the maintenance of the womb lining is particularly costly (Strassmann 1996), or where reproductive investment is heavy (Creel & Creel 1991). Subordinate females in such species may therefore be more likely to evolve and exercise physiological restraint than those in species where the costs of fertility maintenance are comparatively low, as a greater expected fitness benefit from maintaining their fertility and breeding is required before
Causes of physiological suppression in vertebrate societies doing so would be adaptive (Creel & Creel 1991). While such factors alone cannot account for the evolution of physiological restraint among subordinates (as they cannot readily explain why subordinates should exercise restraint when their dominants do not), their potential to predispose particular species or sexes to evolving restraint may account for much of the variation across taxa and between the sexes in the prevalence of physiological restraint. As more data become available, comparative studies that consider the prevalence of physiological suppression across species in relation to such factors may yield particularly valuable insights (see Creel & Creel 1991 for a similar approach to reproductive suppression).
Physiological suppression arising from active interference by dominants Early studies of the causes of physiological suppression among subordinates focused on perhaps the most intuitive adaptive explanation: that selection favors dominant individuals who actively employ tactics to forcibly down-regulate the reproductive physiology of subordinates. While numerous studies have now investigated whether dominants employ such tactics (see below), surprisingly few have tested the pervasive assumption that dominants stand to benefit from doing so (see Chapter 15). While dominants may often gain direct fitness benefits from rendering their subordinates infertile (e.g. by reducing competition for matings: Komdeur et al. 1999; or competition with the offspring of others: S. J. Hodge et al., unpublished data), this need not necessarily be the case: dominants may actually rear more, or fitter, offspring when their subordinates also breed (Lewis & Pusey 1997, Hayes 2000, Chapter 15 in this volume). Any direct benefit accrued through active interference might also be offset by direct costs arising from associated energetic expenditure or risk (Bell 2007) or, potentially, from changes in the behavior of subordinates in response to their reduced reproductive share (e.g. dispersal or challenging the dominant, as envisaged by concession models of reproductive skew: Vehrencamp 1983, Reeve & Keller 1997; but see Clutton-Brock 1998). In societies of close kin, any direct costs of active interference may also be compounded by indirect costs arising from disrupting the reproductive attempts of relatives (Gerlach & Bartmann 2002, Young et al. 2006). Quantification of the inclusive fitness costs and benefits of subordinate reproduction to dominants should therefore be prioritized, to establish whether strong selection on dominants to disrupt the reproductive physiology of subordinates is quite as pervasive as is generally assumed (see also Chapter 15).
415
416
A. J. Young Attempts to explain the evolution of active interference in the fertility of subordinates must consider not only whether selection would favor the disruption of subordinate reproduction (as discussed above), but also whether fertility disruption per se would be a favored means of achieving this. In some societies, for example, dominants may successfully monopolize reproduction simply by killing the offspring of subordinates, obviating the need for active interference in subordinate fertility. In others, however, infanticide may prove less effective, due to difficulty discriminating their own offspring from those of others (Johnstone & Cant 1999a) or the evolution of subordinate counter-tactics (Agrell et al. 1998, Ebensperger 1998). Under these circumstances, the ability to disrupt the underlying fertility of subordinates may be strongly favored, as it may circumvent these constraints on control through infanticide alone (Young et al. 2006). Closer investigation of the effectiveness and limitations of the various means by which dominants could disrupt subordinate reproduction would be of particular value, therefore, for attempts to understand when fertility disruption per se should evolve. Theoretical treatments of reproductive conflict have yet to directly investigate the circumstances under which selection should favor active interference by dominants in the fertility of subordinates. While the large body of theory concerned with explaining variation in reproductive skew has begun to capture some of the complexity outlined above (see Johnstone 2000, Magrath et al. 2004 for reviews), few skew models have explicitly considered the tactics that dominants might employ to disrupt subordinate reproduction (see Johnstone & Cant 1999a, 1999b, Hager & Johnstone 2004 for exceptions), of which active interference in subordinate fertility per se is just one. As a consequence, theoretical predictions regarding the extent of reproductive skew do not yet readily translate into predictions regarding the extent of fertility disruption expected of dominants. A large body of behavioral–endocrine research has now yielded some insight into the likely prevalence of active interference by dominants in the fertility of subordinates, by seeking evidence of the proximate mechanisms through which it might be achieved. Dominants might forcibly disrupt the fertility of their subordinates in a number of ways, but the two mechanisms that have received the most attention to date are interference by means of pheromones released by dominants and by social stress induced by dominants. I discuss the evidence in support of a role for each of these mechanisms below. Active interference through pheromones
In many social insects and some vertebrate societies, the reproductive and/or physiological suppression of subordinates can be maintained in the
Causes of physiological suppression in vertebrate societies absence of the dominant by presenting them with olfactory cues that would normally reflect the dominant’s presence (Keller & Nonacs 1993; see Epple & Katz 1984, Barrett et al. 1990, Abbott et al. 1997 for vertebrate examples). Findings of this kind have led to the suggestion that dominants might be able to disrupt the fertility of subordinates by releasing pheromones that forcibly down-regulate their reproductive physiology. However, numerous studies have now questioned the validity of this hypothesis, in part because it is difficult to imagine how any signaling mechanism that exerted unconditional and deleterious effects on the fitness of the receiver could be evolutionarily stable (Keller & Nonacs 1993). If dominants did release such a pheromone, selection would be expected to strongly favor any mutation that left subordinates at least partially immune to its disruptive effects (e.g. by altering the structure or abundance of the receptors that would otherwise convey the pheromone’s effects to the reproductive axis). If it was actually possible to produce a pheromone whose disruptive effects were evolutionarily “inescapable,” it is also difficult to imagine how dominants would benefit substantially from its production, as it would therefore be expected to disrupt their own fertility as well. A more plausible explanation for such findings is that any pheromone capable of maintaining physiological suppression actually conveys honest information to the subordinate regarding its likely payoff from maintaining its fertility (e.g. by signaling the dominant’s presence, competitive vigor, or fecundity), to which the subordinate may respond adaptively by down-regulating its own reproductive physiology (i.e. exercising restraint: see above). Until evidence to the contrary becomes available, this pheromonal signaling mechanism should probably be considered the most parsimonious explanation for any apparent effects of pheromones on subordinate fertility (Keller & Nonacs 1993). With that in mind, establishing whether subordinates’ responses to such pheromones are actually adaptive (i.e. yield a net fitness payoff given the circumstances) should be considered a high priority for future work. Active interference through socially induced stress
Early attempts to explain physiological suppression among subordinates suggested that dominants might actively down-regulate subordinate fertility by inducing chronic “stress” (often reflected as chronic elevation of glucocorticoid adrenal hormones, GCs), through frequent attacks (Keverne et al. 1982, Wasser & Barash 1983, Kaplan et al. 1986, Reyer et al. 1986). This is an attractive hypothesis, as chronic stress is known to compromise fertility in a variety of taxa (Pottinger 1999), and early studies of social vertebrates supported the prediction that subordinates, who are commonly the target of
417
418
A. J. Young aggression, should show elevated GC levels (reviewed in Abbott et al. 1997, 2003, Von Holst 1998, Creel 2001, Goymann & Wingfield 2004). A number of more recent studies, however, focusing on cooperatively breeding species in particular (where physiological suppression is at its most apparent), have revealed that subordinate group members commonly show average GC levels equal to or actually lower than those of dominants (e.g. Creel et al. 1996; reviewed in Abbott et al. 1997, Creel 2001, Goymann & Wingfield 2004), and experimental work on two cooperative species found no evidence of a role for elevated GCs in subordinate infertility (Abbott et al. 1997, Faulkes & Abbott 1997). These findings, coupled with the low frequency of overt aggression in many cooperative species, have led to the suggestion that the marked physiological suppression typical of high-skew cooperative breeders may not generally be maintained through socially induced stress (e.g. Abbott et al. 1997, Creel 2001). While this evidence does suggest that social stress alone cannot readily account for the physiological suppression of subordinates in cooperative vertebrate societies, it would be premature to reject any role for stress-related suppression in these species (Young et al. 2006). Much of the evidence against the stress-related-suppression hypothesis derives from comparisons of the average GC levels of dominants and mature subordinates (reviewed in Abbott et al. 1997, von Holst 1998, Creel 2001, Goymann & Wingfield 2004). While this approach might be expected to detect stress-related suppression if dominants harassed all mature subordinates throughout the sampling period, such indiscriminate harassment may rarely be necessary (Young et al. 2006). As subordinates in cooperatively breeding species are commonly believed to exercise a degree of physiological restraint (see above), dominants may only need to impose stress-related suppression upon a potentially small subset of their subordinates who might otherwise attempt to breed. Furthermore, dominants may only benefit from attacking these likely breeders at specific times, e.g. when subordinates are close to ovulation (e.g. Wasser & Barash 1983) or when the dominant is rearing her own young (when any young born to a subordinate might otherwise compete with her own: e.g. Young et al. 2006). If stress-related suppression was employed in this way (directed at only a subset of subordinates at critical times, to guard against lapses in restraint), it could prove a difficult phenomenon to detect with comparisons of the average GC levels of dominants and all mature subordinates. Recent work on meerkat societies suggests that dominant females do employ stress-related suppression in this way, targeting only a subset of their subordinates at critical times (Young et al. 2006). Dominant female meerkats periodically subject a subset of their subordinate females to escalating
Causes of physiological suppression in vertebrate societies
Figure 14.2 Dominant female meerkats periodically subject subordinate females to temporary evictions from the group, which result in chronic stress, downregulation of their reproductive physiology, and marked reproductive dysfunction. Dominants appear to employ this tactic to guard against lapses in subordinate restraint at critical times. Photo by Andrew Young.
aggression, culminating in temporary evictions from the group for an average of three weeks at a time (Figure 14.2). The targeted subordinates suffer chronic stress (reflected as a twofold increase in GC metabolite concentrations in their feces), loss of body condition, and down-regulation of their reproductive physiology (substantially reduced pituitary sensitivity to an exogenous gonadotropin-releasing hormone [GnRH] challenge), which is accompanied by a virtual block on conception and elevated rates of abortion (Young et al. 2006). Rather than constantly harassing all subordinate females, dominant females only become aggressive when pregnant themselves (when subordinate reproduction would otherwise conflict with their own: Young & Clutton-Brock 2006) and target those subordinates with whom reproductive conflict is likely to be most acute (older, pregnant, and more distantly related females: Young et al. 2006). As the GC levels of subordinate females not being targeted are comparable with those of dominants (Young et al. 2008), attempts to test the stress-related-suppression hypothesis with comparisons of the average GC levels of dominants and all mature subordinates may have failed to detect it. As dominants in other cooperatively breeding species might well employ stressrelated suppression in a similar targeted manner, either with comparable temporary evictions (common in banded mongoose, Mungos mungo, societies
419
420
A. J. Young for example: Hodge 2003) or focused periods of aggression (e.g. among alpine marmots, Marmota marmota, of both sexes: Arnold & Dittami 1997, Hacklander et al. 2003), active interference by means of socially imposed stress could be a more widespread phenomenon than is currently recognized (Young et al. 2006). How then should one test the stress-related-suppression hypothesis so as to minimize the risk of falsely rejecting any role it may play? Comparisons of the GC levels of dominants and subordinates may remain a useful approach, but it is essential that analyses of this kind allow for the possibility that dominants target only a subset of their subordinates at critical times. Some studies of cooperative breeders have already considered the possibility that subordinates may only be targeted during critical periods, by investigating whether aggression and/or GC levels differ between key reproductive periods and other times (e.g. Schoech et al. 1991, 1996, Creel et al. 1997, Hacklander et al. 2003, Sands & Creel 2004). Indeed, Hacklander et al. found that aggression directed at subordinate female alpine marmots peaked during the dominant female’s pregnancy and that, during this time only, the average GC levels of subordinates exceeded those of dominants, leading to the suggestion of a possible role for stress in the suppression of subordinate reproduction. Fewer studies have also allowed for the likelihood that dominants target only a subset of their subordinates during such times, by subdividing subordinates into classes according to criteria that may affect their willingness to breed, such as their age or relatedness to other group members (e.g. Wingfield et al. 1991, Arnold & Dittami 1997, Hacklander et al. 2003; see also Young et al. 2008). Indeed, Arnold & Dittami’s work suggests a possible role for stress-related suppression among the subset of subordinate male alpine marmots most likely to compete with the dominant (older subordinates that were more distantly related to the dominant suffered higher rates of injury, elevated GC levels, and reduced testosterone levels). Similarly, Hacklander et al. found that dominant female alpine marmots focused their aggression on subordinate females to whom they were less related (see Young et al. 2006 for similar evidence from meerkats). Together, these findings highlight the importance of partitioning both time and subordinate classes when conducting such comparisons, and illustrate why comparisons of the average GC levels of dominants and all mature subordinates should be interpreted with caution. For species in which episodes of dominant aggression are quite clear, perhaps the approach most likely to detect a role for stress-related suppression is to investigate how periods of dominant aggression, when they do occur, affect the adrenal and reproductive physiology of the targeted subordinates (e.g. Hacklander et al. 2003, Young et al. 2006).
Causes of physiological suppression in vertebrate societies Future directions for research on the causes of physiological suppression Research over the past 20 years has yielded significant advances in our understanding of the causes of physiological suppression in vertebrate societies, but numerous challenges remain. The evidence to date strongly suggests that physiological suppression commonly arises, at least in part, from subordinates exercising restraint (e.g. Wingfield et al. 1991, Creel et al. 1996, Abbott et al. 1997, Cooney & Bennett 2000, Clutton-Brock et al. 2001, Creel 2001, Schoech et al. 2004, Young et al. 2008). However, it is less clear how often this restraint can be accounted for without invoking some role for the same-sex dominant (e.g. the likelihood of infanticide by the dominant should the subordinate attempt to breed: Hager & Johnstone 2004, Young et al. 2008, Chapters 12 and 13 in this volume). Current evidence also suggests that physiological suppression in some species may be due in part to active interference by dominants, by means of stress-related suppression (e.g. Wasser & Barash 1983, Young et al. 2006; see also Abbott et al. 1997, Creel 2001). However, the possibility that social stress may play a more widespread role in maintaining physiological suppression, potentially acting in tandem with subordinate restraint, demands closer investigation (Young et al. 2006). Below, I discuss four areas in particular where I believe research could be most profitably directed over the next 20 years, to address these and other key shortfalls in our current understanding. Is subordinate restraint typically due in part to the presence of the same-sex dominant?
Of particular interest is the extent to which physiological restraint among subordinates can realistically be accounted for without invoking a role for the presence and/or likely actions of the dominant. While subordinates may often exercise restraint when lacking unrelated mates within their groups (see above), inbreeding avoidance alone may rarely account entirely for physiological suppression (e.g. Schoech et al. 1996, Abbott et al. 1997, Faulkes & Abbott 1997, Young et al. 2008). It seems likely that a key factor underpinning subordinate restraint in many vertebrate societies may ultimately prove to be the presence and/or actions of the dominant. Attempts to advance our understanding of the role that dominants play in precipitating subordinate restraint should therefore be prioritized. The most direct way to investigate whether the dominant’s presence plays a role in precipitating restraint is to experimentally separate the subordinate and dominant, with the prediction that, if the dominant’s presence is key, this
421
422
A. J. Young should alleviate physiological suppression in the subordinate (as per Abbott et al. 1997, Faulkes & Abbott 1997). As one would make an identical prediction if physiological suppression arose through active interference, simultaneously investigating the effect of the treatment on the stress physiology of subordinates would be essential (as per Faulkes & Abbott 1997). Removing the dominant from the subordinate’s environment (rather than vice versa) may typically allow the findings to be interpreted with greater confidence, as any alleviation of suppression on the relocation of a subordinate could be due instead to changes in any number of other factors (e.g. the subordinate’s access to unrelated breeding partners, body condition, and familiarity with the environment). This experimental approach has been used to good effect in captivity, but has yet to be applied in the wild, perhaps due to the profound disruption that dominant removals would cause to valuable long-term lifehistory data and to potential complications arising from the associated social instability. A less disruptive approach would be to investigate whether disparities in the reproductive physiology of dominants and subordinates can be accounted for by controlling statistically for each of the factors, other than the presence of the dominant, that may contribute to selection for subordinate restraint (i.e. their access to unrelated breeding partners, age, body condition, and prior breeding experience; see Young et al. 2008 for an example). While this approach alone does not directly implicate a role for the dominant per se, it does at least suggest that subordinates may be exercising restraint due in part to some other key factor, of which the presence and/or likely actions of the dominant may often be the most plausible candidate. Why do subordinates exercise restraint in the presence of the dominant?
Subordinates in some species do appear to exercise physiological restraint in the presence of their same-sex dominant (e.g. Abbott et al. 1997, Faulkes & Abbott 1997), but precisely why this is adaptive is rarely totally clear. One might assume that such restraint is due to the dominant’s ability to disrupt or punish subordinate attempts to breed (e.g. through mate guarding: Komdeur et al. 1999; pregnancy disruption: Wasser & Barash 1983, Young et al. 2006; or infanticide: Digby 2000), but this need not necessarily be the case. Subordinate restraint might also be due to the dominant’s ability to disrupt or punish subordinate attempts to maintain their fertility, or to the fact that, regardless of any risk of dominant interference, the subordinate’s expected fitness payoff from breeding may be markedly reduced while the dominant is also breeding (e.g. due to costs arising from competition between litters). Where subordinates are thought to exercise physiological restraint in the dominant’s presence, research into precisely why it is adaptive to do so would therefore be valuable.
Causes of physiological suppression in vertebrate societies Whether dominants do interfere with subordinate breeding attempts may often be clear from their responses to natural instances of subordinate reproduction (e.g. Digby 2000). However, in species where subordinates never attempt to breed, or where natural instances of subordinate reproduction might have been “sanctioned” by the dominant, the experimental induction of subordinate breeding may be necessary before the dominant’s response to “unsanctioned” reproduction becomes apparent. Clearly experiments of this kind would require imaginative designs to ensure that no harm came from any dominant aggression elicited. Establishing whether dominants disrupt or punish subordinate attempts to maintain fertility (rather than subordinate attempts to breed) might be achieved by conducting hormonal manipulations to restore suppressed components of the reproductive axes of subordinates to the levels seen in dominants. Manipulations of this kind could yield key insights both into the means by which dominants enforce restraint and into the cues that they use to do so. For example, fitting a small sample of subordinate female meerkats with estrogen implants yielded estrus behavior (seeking copulations with extra-group mates), but it did not elicit clear aggression from the dominant female, suggesting that the low mean estrogen levels of subordinates probably do not reflect subordinate restraint due to the dominant policing their estrogen levels per se (Young et al., unpublished data). Establishing whether subordinates exercise restraint due in part to costs that would arise from competition between their own young and those of the dominant might best be achieved by quantifying the inclusive fitness costs and benefits to subordinates of breeding alongside the dominant (e.g. Gerlach & Bartmann 2002), and investigating whether variation in these explains variation in the extent of restraint. Does active interference by dominants often play a role alongside subordinate restraint?
While stress-related suppression alone cannot readily explain the physiological suppression of subordinates across cooperative vertebrates (Abbott et al. 1997, Creel 2001), social stress may still play a key role in precipitating physiological suppression in some cooperative vertebrates, potentially acting in tandem with subordinate restraint (e.g. Young et al. 2006). Future research should investigate the extent to which this is the case, allowing explicitly for the possibility that dominants target only a subset of their subordinates at critical times, to counter, or guard against, any lapses in restraint (Young et al. 2006). As discussed above, comparisons of the average GC levels of dominants and all mature subordinates should be interpreted with caution. Where periods of dominant aggression are apparent, monitoring the
423
424
A. J. Young consequences of that aggression for the adrenal and reproductive physiology of the targeted subordinate(s) may prove a valuable approach. Evidence of active interference might be most profitably sought in the many species where subordinates do exhibit physiological suppression but still occasionally breed, as subordinates with zero chance of successful reproduction may generally be expected to exercise complete physiological restraint. Even where physiological suppression does appear to arise solely from subordinate restraint, it is worth bearing in mind that the dominant’s capacity to forcibly disrupt the subordinate’s fertility may be the reason for that restraint. What cues do subordinates use to make adaptive restraint decisions?
One area with considerable potential for novel research is the examination of the cues that subordinates use to make adaptive restraint decisions (i.e. whether or not, and perhaps the extent to which, they should exercise physiological restraint). Our understanding of such cues, the precise information that they convey, and the mechanisms that maintain the reliability of that information, while well advanced for social-insect societies, is comparatively poorly developed for vertebrates. The limited work in this area to date has focused largely on the potential role of pheromones (e.g. Lepri & Vandenbergh 1986, Barrett et al. 1990, 1993, Faulkes & Abbott 1993, Smith et al. 1997), but it is clear that cues transmitted through other sensory modalities also convey complex information on which restraint decisions could be based (e.g. visual cues of individual identity: Tibbetts 2002; or vigor: Tibbetts & Dale 2004; and vocal cues of kinship: Sharp et al. 2005; see Barrett et al. 1993, Abbott et al. 1997). The suggestion that subordinate restraint may be maintained by the associative learning of individual-specific cues, in any of these modalities, also warrants close attention (Abbott et al. 1997, Chapter 12 in this volume). A complete understanding of the mechanisms that underpin the expression of subordinate restraint will also demand research into the neuroendocrine pathways that link the detection of such cues to the down-regulation of the reproductive axis (e.g. Shepherd 2006, Snowdon et al. 2006). Together, these long-term goals offer considerable opportunity for novel research at the interface between behavioral ecology and endocrinology over the coming 20 years.
Conclusion: implications for our understanding of reproductive skew in vertebrate societies Research into the causes of physiological suppression has close parallels with research into the causes of reproductive skew. Marked downregulation of the reproductive physiology simply constitutes a fundamental
Causes of physiological suppression in vertebrate societies form of reproductive dysfunction among subordinates. As such, while these two fields have advanced somewhat independently over the years, key insights from each are relevant to both. Indeed, while I have outlined a general adaptive framework for understanding the causes of physiological suppression in vertebrate societies, the same key concepts could be applied more broadly to understanding the causes of low reproductive success among subordinates. In this section, I briefly discuss some of the key implications that insights into the causes of physiological suppression have for empirical and theoretical attempts to understand the causes of variation in reproductive skew. Theoretical approaches to reproductive skew typically envisage the extent of skew as a product solely of intrasexual conflict between dominants and subordinates (for reviews see Keller & Reeve 1994, Johnstone 2000, Magrath et al. 2004; see Cant & Reeve 2002 for an exception). However, research into the causes of physiological suppression highlights the need to consider a range of additional factors that may affect a subordinate’s expected fitness payoff from attempting to breed (e.g. a lack of access to unrelated breeding partners, young age, and poor body condition). A number of studies have already highlighted the need to account for a probable role for inbreeding avoidance in treatments of reproductive skew (Koenig & Haydock 2004, Magrath et al. 2004), but this is just one of a number of ways in which dominants and subordinates may differ in their expected payoffs from attempting to breed, regardless of intrasexual competition. While incorporating such factors into theoretical models of reproductive sharing may prove somewhat trivial, accounting for their influence at some stage (either in models or when testing them) is essential if we are ultimately seeking congruence between the predictions of models and observed patterns of skew. As reproductive skew will typically be determined, at least in part, by reproductive conflict between dominants and subordinates, the effectiveness and limitations of the tactics available to competing parties is expected to play a key role in the outcome (reflected in the prediction of tug-of-war or compromise models that skew may depend largely upon the balance of power between dominants and subordinates: Reeve et al. 1998, Johnstone 2000, Magrath et al. 2004). The finding that dominants in some vertebrate societies appear capable of forcibly disrupting the fertility of subordinates, by subjecting them to chronic persecution and associated social stress (e.g. Wasser & Barash 1983, Young et al. 2006), is particularly important, therefore. The ability to disrupt a subordinate’s fertility directly may shift the balance of power in reproductive competition markedly in the dominant’s favor, by circumventing constraints on their ability to control subordinate reproduction by other means, such as infanticide (e.g. due to difficulties discriminating own offspring
425
426
A. J. Young Box 14.6 The causes of reproductive skew among female meerkats Dominant female meerkats breed at markedly higher rates than subordinates and produce > 80% of the offspring that survive their first month of life (Clutton-Brock et al. 2001). This marked reproductive skew is almost certainly due in part to processes other than intrasexual competition, as subordinate females show reduced mean estrogen levels and lower breeding rates when lacking unrelated breeding partners within their groups and when in poor body condition (O’Riain et al. 2000, Clutton-Brock et al. 2001, Carlson et al. 2004, Young et al. 2008). However, the dominant female and her subordinates clearly are involved in a reproductive power struggle. Dominant females appear capable of killing the majority of subordinate-female litters that might otherwise clash with their own (Clutton-Brock et al. 2001, Young & Clutton-Brock 2006). Subordinate females also become infanticidal when pregnant, however, killing litters born to both the dominant and other subordinates (Young & Clutton-Brock 2006; see also Clutton-Brock et al. 1998). Dominant females partly resolve this problem by employing a second tactic: periodically subjecting their subordinates to temporary stressful evictions from the group, which result in marked physiological down-regulation and reproductive failure among subordinates (Young et al. 2006). By forcibly disrupting subordinate fertility, these temporary evictions not only reduce the need for dominants to rely on infanticide, but also reduce the risk of infanticide by subordinates (by reducing the likelihood that subordinates will be pregnant, and hence infanticidal, when the dominant gives birth: Young et al. 2006, Young & Clutton-Brock 2006; see also Clutton-Brock et al. 1998). There is also a strong temporal component to this reproductive power struggle, as dominant females only evict subordinate females and kill their young when pregnant themselves (Clutton-Brock et al. 1998, 2001, Young et al. 2006, Young & Clutton-Brock 2006). As dominant females are typically closely related to their subordinates, this temporal variation in the extent of dominant interference may largely reflect the inclusive fitness payoffs that dominants stand to gain from tolerating those subordinate breeding attempts that do not clash with their own. As a result of this temporal variation in the extent of dominant interference, much of the reproductive success of subordinates arises from breeding attempts that are asynchronous with those of other female group members (Clutton-Brock et al. 2001, Young & Clutton-Brock 2006), and physiological evidence suggests that they may target their conceptive periods accordingly (Young et al. 2008).
Causes of physiological suppression in vertebrate societies from those of others: Johnstone & Cant 1999a, Hager & Johnstone 2004; or the evolution of subordinate counter-tactics: Agrell et al. 1998, Ebensperger 1998). Stress-related suppression may still not afford dominants complete control, however, if energetic costs of persecution constrain their ability to employ it (e.g. Young et al. 2006). Further empirical research into the prevalence of stressrelated suppression and its interplay with subordinate restraint, coupled with the development of theoretical models that investigate how the particular tactics available to competing parties affect the outcome of reproductive conflict (e.g. discriminate or indiscriminate infanticide: Johnstone & Cant 1999a, Hager & Johnstone 2004), should therefore be prioritized. Empirical insights into the balance of power in reproductive conflict will also help to highlight which theoretical approaches to reproductive skew most accurately capture the dynamics of wild systems. For example, concessionbased frameworks assume that dominants have complete control over reproductive sharing, whereas tug-of-war or compromise frameworks allow for the possibility of power struggles between individuals with varying degrees of control (see Magrath et al. 2004 for a review). While numerous studies have sought to clarify the relative merits of these approaches by testing their contrasting predictions, the sensitivity of the predictions to slight departures from the models’ assumptions, coupled with the difficulty of adequately controlling for confounding variables, leaves it difficult to provide unequivocal support for either using this approach alone (Magrath et al. 2004). Detailed behavioral and physiological studies of the competitive tactics that individuals employ in reproductive conflict may provide a more direct means of identifying the most appropriate framework, by testing the models’ assumptions regarding the extent of dominant control (Clutton-Brock 1998, Johnstone 2000, CluttonBrock et al. 2001, Magrath et al. 2004). Research into the causes of reproductive skew among female meerkats illustrates well the importance of understanding the particular tactics that individuals employ in reproductive conflict, as well as the role of factors other than intrasexual competition, for attempts to explain variation in reproductive skew (Box 14.5). As this level of complexity is probably far from unique (see Chapter 12, for example), and the key processes involved are likely to vary across species, we must remain realistic about the likelihood that any one theoretical model will account for the considerable variation in reproductive skew within and between species. However, the close integration of empirical insights into the proximate and ultimate causes of reproductive disparities within vertebrate societies with broader insights from realistic theory should still prove the most profitable approach to tackling the challenges that lie ahead.
427
428
A. J. Young Acknowledgments I would like to thank Reinmar Hager for inviting me to contribute to this volume, Nigel Bennett, Anne Carlson, Tim Clutton-Brock, Sarah Hodge, Steve Monfort, and Andy Russell for countless discussions about reproductive conflict in animal societies over the years, and Anne Carlson, Sarah Hodge, Steve Schoech, and an anonymous referee for insightful comments on this manuscript. This work was supported by research fellowships from Magdalene College, University of Cambridge, UK, and the Natural Environment Research Council, UK.
References Abbott, D. H. (1993). Social conflict and reproductive suppression in marmoset and tamarin monkeys. In W. A. Mason & S. P. Mendoza, eds., Primate Social Conflict. Albany, NY: SUNY Press, pp. 331–372. Abbott, D. H. (1984). Behavioral and physiological suppression of fertility in subordinate marmoset monkeys. American Journal of Primatology, 6, 169–186. Abbott, D. H., McNeilly, A. S., Lunn, S. F., Hulme, M. J., & Burden, F. J. (1981). Inhibition of ovarian function in subordinate female marmoset monkeys (Callithrix jacchus jacchus). Journal of Reproduction and Fertility, 63, 335–345. Abbott, D. H., Hodges, J. K., & George, L. M. (1988). Social status controls LH secretion and ovulation in female marmoset monkeys (Callithrix jacchus). Journal of Endocrinology, 117, 329–339. Abbott, D. H., Saltzman, W., Schultz-Darken, N. J., & Smith, T. E. (1997). Specific neuroendocrine mechanisms not involving generalized stress mediate social regulation of female reproduction in cooperatively breeding marmoset monkeys. Annals of the New York Academy of Sciences, 807, 219–238. Abbott, D. H., Keverne, E. B., Bercovitch, F. B., et al. (2003). Are subordinates always stressed? A comparative analysis of rank differences in cortisol levels among primates. Hormones and Behavior, 43, 67–82. Agrell, J., Wolff, J. O., & Ylonen, H. (1998). Counter-strategies to infanticide in mammals: costs and consequences. Oikos, 83, 507–517. Alonso-Alvarez, C., Bertrand, S., Faivre, B., Chastel, O., & Sorci, G. (2007). Testosterone and oxidative stress: the oxidation handicap hypothesis. Proceedings of the Royal Society of London B, 274, 819–825. Arnold, W. & Dittami, J. (1997). Reproductive suppression in male alpine marmots. Animal Behaviour, 53, 53–66. Barrett, J., Abbott, D. H., & George, L. M. (1990). Extension of reproductive suppression by pheromonal cues in subordinate female marmoset monkeys, Callithrix jacchus. Journal of Reproduction and Fertility, 90, 411–418. Barrett, J., Abbott, D. H., & George, L. M. (1993). Sensory cues and the suppression of reproduction in subordinate female marmoset monkeys, Callithrix jacchus. Journal of Reproduction and Fertility, 97, 301–310.
Causes of physiological suppression in vertebrate societies Becker, J. B., Breedlove, S. M., & Crews, D., eds. (2002). Behavioral Endocrinology. Boston, MA: MIT Press. Bell, M. B. V. (2007). Cooperative begging in banded mongoose pups. Current Biology, 17, 717–721. Bennett, N. C., Faulkes, C. G., & Molteno, A. J. (1996). Reproductive suppression in subordinate, non-breeding female Damaraland mole-rats: two components to a lifetime of socially induced infertility. Proceedings of the Royal Society of London B, 263, 1599–1603. Bennett, N. C., Faulkes, C. G., & Jarvis, J. U. M. (1999). Socially induced infertility, incest avoidance and the monopoly of reproduction in cooperatively breeding African mole-rats: family Bathyergidae. Advances in the Study of Behavior, 28, 75–114. Bishop, J. M., Jarvis, J. U. M., Spinks, A. C., Bennett, N. C. & O’Ryan, C. (2004). Molecular insight into patterns of colony composition and paternity in the common mole-rat Cryptomys hottentotus hottentotus. Molecular Ecology, 13, 1217– 1229. Brown, J. L. (1987). Helping and Communal Breeding in Birds. Princeton, NJ: Princeton University Press. Buchner, A. S., Sloman, K. A., & Balshine, S. (2004). The physiological effects of social status in the cooperatively breeding cichlid Neolamprologus pulcher. Journal of Fish Biology, 65, 1080–1095. Cant, M. A. & Reeve, H. K. (2002). Female control of the distribution of paternity in cooperative breeders. American Naturalist, 160, 602–611. Carlson, A. A., Ziegler, T. E., & Snowdon, C. T. (1997). Ovarian function of pygmy marmoset daughters (Cebuella pygmaea) in intact and motherless families. American Journal of Primatology, 43, 347–355. Carlson, A. A., Young, A. J., Russell, A. F., et al. (2004). Hormonal correlates of dominance in meerkats (Suricata suricatta). Hormones and Behavior, 46, 141–150. Carter, C. S. & Roberts, R. L. (1997). The psychobiological basis of cooperative breeding in rodents. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 231–266. Cichon, M. (2003). Does prior breeding experience improve reproductive success in collared flycatcher females? Oecologia, 134, 78–81. Clarke, F. M. & Faulkes, C. G. (1999). Kin discrimination and female mate choice in the naked mole-rat Heterocephalus glaber. Proceedings of the Royal Society of London B, 266, 1995–2002. Clarke, F. M., Miethe, G. H., & Bennett, N. C. (2001). Reproductive suppression in female Damaraland mole-rats Cryptomys damarensis: dominant control or selfrestraint? Proceedings of the Royal Society of London B, 268, 899–909. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H. & Parker, G. A. (1995a). Sexual coercion in animal societies. Animal Behaviour, 49, 1345–1365.
429
430
A. J. Young Clutton-Brock, T. H. & Parker, G. A. (1995b). Punishment in animal societies. Nature, 373, 209–216. Clutton-Brock, T. H., Brotherton, P. N. M., Smith, R., et al. (1998). Infanticide and expulsion of females in a cooperative mammal. Proceedings of the Royal Society of London B, 265, 2291–2295. Clutton-Brock, T. H., Brotherton, P. N. M., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Cooney, R. & Bennett, N. C. (2000). Inbreeding avoidance and reproductive skew in a cooperative mammal. Proceedings of the Royal Society of London B, 267, 801–806. Creel, S., Creel, N., Wildt, D. E., & Monfort, S. L. (1992). Behavioral and endocrine mechanisms of reproductive suppression in Serengeti dwarf mongooses. Animal Behaviour, 43, 231–245. Creel, S., Creel, N. M., & Monfort, S. L. (1996). Social stress and dominance. Nature, 379, 212. Creel, S., Creel, N. M., Mills, M. G. L., & Monfort, S. L. (1997). Rank and reproduction in cooperatively breeding African wild dogs: behavioral and endocrine correlates. Behavioral Ecology, 8, 298–306. Creel, S. (2001). Social dominance and stress hormones. Trends in Ecology and Evolution, 16, 491–497. Creel, S. R. & Creel, N. M. (1991). Energetics, reproductive suppression and obligate communal breeding in carnivores. Behavioral Ecology and Sociobiology, 28, 263–270. Crespi, B. J. & Ragsdale, J. E. (2000). A skew model for the evolution of sociality via manipulation: why it is better to be feared than loved. Proceedings of the Royal Society of London B, 267, 821–828. Daniels, S. J. & Walters, J. R. (2000). Inbreeding depression and its effects on natal dispersal in red-cockaded woodpeckers. Condor, 102, 482–491. De Vleeschouwer, K., Van Elsacker, L., Heistermann, M., & Leus, K. (2000). An evaluation of the suitability of contraceptive methods in golden-headed lion tamarins (Leontopithecus chrysomelas), with emphasis on melengestrol acetate (MGA) implants: (II) endocrinological and behavioural effects. Animal Welfare, 9, 385–401. De Vleeschouwer, K., Leus, K., & Van Elsacker, L. (2003). Stability of breeding and non-breeding groups of golden-headed lion tamarins (Leontopithecus chrysomelas). Animal Welfare, 12, 251–268. Desjardins, C. & Turek, F. W. (1977). Effects of testosterone on spermatogenesis and luteinizing hormone release in Japanese quail. General and Comparative Endocrinology, 33, 293–303. Dierkes, P., Taborsky, M., & Kohler, U. (1999). Reproductive parasitism of broodcare helpers in a cooperatively breeding fish. Behavioral Ecology, 10, 510–515. Digby, L. (2000). Infanticide by female mammals: implications for the evolution of social systems. In C. P. van Schaik & C. H. Janson, eds., Infanticide by Males and its Implications. Cambridge: Cambridge University Press, pp. 423–446. Double, M. C. & Cockburn, A. (2003). Subordinate superb fairy-wrens (Malurus cyaneus) parasitize the reproductive success of attractive dominant males. Proceedings of the Royal Society of London B, 270, 379–384.
Causes of physiological suppression in vertebrate societies Ebensperger, L. A. (1998). Strategies and counterstrategies to infanticide in mammals. Biological Reviews of the Cambridge Philosophical Society, 73, 321–346. Endler, A., Liebig, J., Schmitt, T., et al. (2004). Surface hydrocarbons of queen eggs regulate worker reproduction in a social insect. Proceedings of the National Academy of Sciences of the USA, 101, 2945–2950. Epple, G. & Katz, Y. (1984). Social influences on estrogen excretion and ovarian cyclicity in saddle back tamarins (Saguinus fuscicollis). American Journal of Primatology, 6, 215–227. Faulkes, C. G. & Abbott, D. H. (1997). The physiology of a reproductive dictatorship: regulation of male and female reproduction by a single breeding female in colonies of naked mole-rats. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 302–334. Faulkes, C. G., Abbott, D. H., & Jarvis, J. U. M. (1991). Social suppression of reproduction in male naked mole-rats, Heterocephalus glaber. Journal of Reproduction and Fertility, 91, 593–604. Faulkes, C. G. & Abbott, D. H. (1993). Evidence that primer pheromones do not cause social suppression of reproduction in male and female naked mole-rats (Heterocephalus glaber). Journal of Reproduction and Fertility, 99, 225–230. Fitzpatrick, J. L., Desjardins, J. K., Stiver, K. A., Montgomerie, R., & Balshine, S. (2006). Male reproductive suppression in the cooperatively breeding fish Neolamprologus pulcher. Behavioral Ecology, 17, 25–33. Folstad, I. & Karter, A. J. (1992). Parasites, bright males, and the immunocompetence handicap. American Naturalist, 139, 603–622. Forslund, P. & Part, T. (1995). Age and reproduction in birds: hypotheses and tests. Trends in Ecology and Evolution, 10, 374–378. French, J. A. (1994). Alloparents in the Mongolian gerbil: impact on long-term reproductive performance of breeders and opportunities for independent reproduction. Behavioral Ecology, 5, 273–279. French, J. A. (1997). Proximate regulation of singular breeding in callitrichid primates. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. pp. 40–81. Cambridge: Cambridge University Press. French, J. A., Abbott, D. H., & Snowdon, C. T. (1984). The effect of social environment on estrogen excretion, scent marking, and sociosexual behavior in tamarins (Saguinus oedipus). American Journal of Primatology, 6, 155–167. French, J. A., Bales, K. L., Baker, A. J., & Dietz, J. M. (2003). Endocrine monitoring of wild dominant and subordinate female Leontopithecus rosalia. International Journal of Primatology, 24, 1281–1300. Gerlach, G. & Bartmann, S. (2002). Reproductive skew, costs, and benefits of cooperative breeding in female wood mice (Apodemus sylvaticus). Behavioral Ecology, 13, 408–418. Goymann, W. & Wingfield, J. C. (2004). Allostatic load, social status and stress hormones: the costs of social status matter. Animal Behaviour, 67, 591–602. Greeff, J. M. & Bennett, N. C. (2000). Causes and consequences of incest avoidance in the cooperatively breeding mole-rat, Cryptomys darlingi (Bathyergidae). Ecology Letters, 3, 318–328.
431
432
A. J. Young Hacklander, K., Mostl, E., & Arnold, W. (2003). Reproductive suppression in female Alpine marmots, Marmota marmota. Animal Behaviour, 65, 1133–1140. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Hamilton, I. M. (2004). A commitment model of reproductive inhibition in cooperatively breeding groups. Behavioral Ecology, 15, 585–591. Hau, M. (2007). Regulation of male traits by testosterone: implications for the evolution of vertebrate life histories. Bioessays, 29, 133–144. Hayes, L. D. (2000). To nest communally or not to nest communally: a review of rodent communal nesting and nursing. Animal Behaviour, 59, 677–688. Hodge, S. J. (2003). The evolution of cooperation in the communal breeding banded mongoose. Unpublished Ph.D. thesis, University of Cambridge. Jarvis, J. U. M., O’Riain, M. J., Bennett, N. C., & Sherman, P. W. (1994). Mammalian eusociality: a family affair. Trends in Ecology and Evolution, 9, 47–51. Johnson, M. H. & Everitt, B. J. (1999). Essential Reproduction. Oxford: Blackwell. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A. & Cant, M. A. (1999a). Reproductive skew and indiscriminate infanticide. Animal Behaviour 57, 243–249. Johnstone, R. A. & Cant, M. A. (1999b). Reproductive skew and the threat of eviction: a new perspective. Proceedings of the Royal Society of London B, 266, 275–279. Kaplan, J. R., Adams, M. R., Koritnik, D. R., Rose, J. C., & Nanuck, S. B. (1986). Adrenal responsiveness and social status in intact and ovariectomized Macaca fascicularis. American Journal of Primatology, 11, 181–193. Keane, B., Creel, S. R. & Waser, P. M. (1996). No evidence of inbreeding avoidance or inbreeding depression in a social carnivore. Behavioral Ecology, 7, 480–489. Keller, L. & Nonacs, P. (1993). The role of queen pheromones in social insects: queen control or queen signal. Animal Behaviour, 45, 787–794. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–102. Keverne, E. B., Meller, R. E., & Eberhart, A. (1982). Dominance and subordination: concepts or physiological states? In A. B. Chiarelli & R. S. Corruccini, eds., Advanced Views in Primate Biology. Berlin: Springer, pp. 81–94. Khan, M. Z., McNabb, F. M. A., Walters, J. R., & Sharp, P. J. (2001). Patterns of testosterone and prolactin concentrations and reproductive behavior of helpers and breeders in the cooperatively breeding red-cockaded woodpecker (Picoides borealis). Hormones and Behavior, 40, 1–13. Koenig, W. D. & Haydock, J. (2004). Incest and incest avoidance. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 142–156. Komdeur, J., Kraaijeveld-Smit, F., Kraaijeveld, K., & Edelaar, P. (1999). Explicit experimental evidence for the role of mate guarding in minimizing loss of paternity in the Seychelles warbler. Proceedings of the Royal Society of London B, 266, 2075–2081.
Causes of physiological suppression in vertebrate societies Lepri, J. J. & Vandenbergh, J. G. (1986). Puberty in pine voles, Microtus pinetorum, and the influence of chemosignals on female reproduction. Biology of Reproduction, 34, 370–377. Lewis, S. E. & Pusey, A. E. (1997). Factors influencing the occurrence of communal care in plural breeding mammals. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 335–363. Magrath, R. D., Johnstone, R. A., & Heinsohn, R. G. (2004). Reproductive skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. Maswanganye, K. A., Bennett, N. C., Brinders, J., & Cooney, R. (1999). Oligospermia and azoospermia in non-reproductive male Damaraland mole-rats Cryptomys damarensis (Rodentia: Bathyergidae). Journal of Zoology, 248, 411–418. Mays, N. A., Vleck, C. M., & Dawson, J. (1991). Plasma luteinizing hormone, steroid hormones, behavioral role, and nest stage in cooperatively breeding Harris hawks (Parabuteo unicinctus). Auk, 108, 619–637. McRae, S. B. (1996). Family values: costs and benefits of communal nesting in the moorhen. Animal Behaviour, 52, 225–245. Molteno, A. J. & Bennett, N. C. (2000). Anovulation in non-reproductive female Damaraland mole-rats (Cryptomys damarensis). Journal of Reproduction and Fertility, 119, 35–41. Molteno, A. J., Kallo, I., Bennett, N. C., King, J. A., & Coen, C. W. (2004). A neuroanatomical and neuroendocrinological study into the relationship between social status and the GnRH system in cooperatively breeding female Damaraland mole-rats, Cryptomys damarensis. Reproduction, 127, 13–21. Nelson, R. J. (2004). An Introduction to Behavioral Endocrinology. Sunderland, MA: Sinauer Associates. O’Riain, M. J., Bennett, N. C., Brotherton, P. N. M., McIlrath, G., & Clutton-Brock, T. H. (2000). Reproductive suppression and inbreeding avoidance in wild populations of cooperatively breeding meerkats (Suricata suricatta). Behavioral Ecology and Sociobiology, 48, 471–477. Payman, B. C. & Swanson, H. H. (1980). Social influence on sexual maturation and breeding in the female Mongolian gerbil (Meriones unguiculatus). Animal Behaviour, 28, 528–535. Peters, A., Astheimer, L. B., & Cockburn, A. (2001). The annual testosterone profile in cooperatively breeding superb fairy-wrens, Malurus cyaneus, reflects their extreme infidelity. Behavioral Ecology and Sociobiology, 50, 519–527. Peters, A., Cockburn, A., & Cunningham, R. (2002). Testosterone treatment suppresses paternal care in superb fairy-wrens, Malurus cyaneus, despite their concurrent investment in courtship. Behavioral Ecology and Sociobiology, 51, 538–547. Poiani, A. & Fletcher, T. (1994). Plasma levels of androgens and gonadal development of breeders and helpers in the bell miner (Manorina melanophrys). Behavioral Ecology and Sociobiology, 34, 31–41.
433
434
A. J. Young Pottinger, T. G. (1999). The impact of stress on animal reproductive activities. In P. H. M. Balm, ed., Stress Physiology in Animals. Sheffield: Sheffield Academic Press, pp. 130–177. Pusey, A. & Wolf, M. (1996). Inbreeding avoidance in animals. Trends in Ecology and Evolution, 11, 201–206. Reeve, H. K. & Keller, L. (1995). Partitioning of reproduction in mother–daughter versus sibling associations: a test of optimal skew theory. American Naturalist, 145, 119–132. Reeve, H. K. & Keller, L. (1996). Relatedness asymmetry and reproductive sharing in animal societies. American Naturalist, 148, 764–769. Reeve, H. K. & Keller, L. (1997). Reproductive bribing and policing: evolutionary mechanisms for the suppression of within-group selfishness. American Naturalist, 150, S42–S58. Reeve, H. K., Westneat, D. F., Noon, W. A., Sherman, P. W., & Aquadro, C. F. (1990). DNA fingerprinting reveals high levels of inbreeding in colonies of the eusocial naked mole-rat. Proceedings of the National Academy of Sciences of the USA, 87, 2496–2500. Reeve, H. K., Emlen, S. T., & Keller, L. (1998). Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology, 9, 267–278. Reyer, H. U., Dittami, J. P., & Hall, M. R. (1986). Avian helpers at the nest: are they psychologically castrated? Ethology, 71, 216–228. Russell, A. F., Brotherton, P. N. M., McIlrath, G. M., Sharpe, L. L., & Clutton-Brock, T. H. (2003). Breeding success in cooperative meerkats: effects of helper number and maternal state. Behavioral Ecology, 14, 486–492. Saltzman, W. (2003). Reproductive competition among female common marmosets (Callithrix jacchus): proximate and ultimate causes. In C. B. Jones, ed., Sexual Selection and Reproductive Competition in Primates: New Perspectives and Directions. Norman, OK: American Society of Primatologists, pp. 197–229. Saltzman, W., Schultz-Darken, N. J., & Abbott, D. H. (1997). Familial influences on ovulatory function in common marmosets (Callithrix jacchus). American Journal of Primatology, 41, 159–177. Saltzman, W., Schultz-Darken, N. J., Wegner, F. H., Wittwer, D. J., & Abbott, D. H. (1998). Suppression of cortisol levels in subordinate female marmosets: reproductive and social contributions. Hormones and Behavior, 33, 58–74. Saltzman, W., Pick, R. R., Salper, O. J., Liedl, K. J., & Abbott, D. H. (2004). Onset of plural cooperative breeding in common marmoset families following replacement of the breeding male. Animal Behaviour, 68, 59–73. Sands, J. & Creel, S. R. (2004). Social dominance, aggression and faecal glucocorticoid levels in a wild population of wolves, Canis lupus. Animal Behaviour, 67, 387–396. Schoech, S. J., Mumme, R. L., & Moore, M. C. (1991). Reproductive endocrinology and mechanisms of breeding inhibition in cooperatively breeding Florida scrub-jays (Aphelocoma coerulescens). Condor, 93, 354–364.
Causes of physiological suppression in vertebrate societies Schoech, S. J., Mumme, R. L., & Wingfield, J. C. (1996). Delayed breeding in the cooperatively breeding Florida scrub jay (Aphelocoma coerulescens): inhibition or the absence of stimulation? Behavioral Ecology and Sociobiology, 39, 77–90. Schoech, S. J., Reynolds, S. J., & Boughton, R. K. (2004). Endocrinology. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press. pp. 128–141. Sharp, S. P., McGowan, A., Wood, M. J., & Hatchwell, B. J. (2005). Learned kin recognition cues in a social bird. Nature, 434, 1127–1130. Shepherd, G. M. (2006). Behaviour: smells, brains and hormones. Nature, 439, 149. Smith, T. E. & Abbott, D. H. (1998). Behavioral discrimination between circumgenital odor from peri-ovulatory dominant and anovulatory female common marmosets (Callithrix jacchus). American Journal of Primatology, 46, 265–284. Smith, T. E., Faulkes, C. G., & Abbott, D. H. (1997). Combined olfactory contact with the parent colony and direct contact with nonbreeding animals does not maintain suppression of ovulation in female naked mole-rats (Heterocephalus glaber). Hormones and Behavior, 31, 277–288. Snowdon, C. T. (1996). Infant care in cooperatively breeding species. Advances in the Study of Behavior, 25, 643–689. Snowdon, C. T., Ziegler, T. E., Schultz-Darken, N. J., & Ferris, C. F. (2006). Social odours, sexual arousal and pairbonding in primates. Philosophical Transactions of the Royal Society of London B, 361, 2079–2089. Strassmann, B. I. (1996). The evolution of endometrial cycles and menstruation. Quarterly Review of Biology, 71, 181–220. Tibbetts, E. A. (2002). Visual signals of individual identity in the wasp Polistes fuscatus. Proceedings of the Royal Society of London B, 269, 1423–1428. Tibbetts, E. A. & Dale, J. (2004). A socially enforced signal of quality in a paper wasp. Nature, 432, 218–222. Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. von Holst, D. (1998). The concept of stress and its relevance for animal behavior. Advances in the Study of Behavior, 27, 1–131. Washabaugh, K. & Snowdon, C. T. (1998). Chemical communication of reproductive status in female cotton-top tamarins (Saguinus oedipus). American Journal of Primatology, 45, 337–349. Wasser, S. K. & Barash, D. P. (1983). Reproductive suppression among female mammals: implications for biomedicine and sexual selection theory. Quarterly Review of Biology, 58, 513–538. Wenseleers, T., Hart, A. G., & Ratnieks, F. L. W. (2004). When resistance is useless: Policing and the evolution of reproductive acquiescence in insect societies. American Naturalist, 164, E154–E167. Wingfield, J. C., Hegner, R. E., Dufty, A. M., & Ball, G. F. (1990). The challenge hypothesis: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist, 136, 829–846.
435
436
A. J. Young Wingfield, J. C., Hegner, R. E., & Lewis, D. M. (1991). Circulating levels of luteinizing hormone and steroid hormones in relation to social status in the cooperatively breeding white browed sparrow weaver, Plocepasser mahali. Journal of Zoology, 225, 43–58. Wong, M. Y. L., Buston, P. M., Munday, P. L., & Jones, G. P. (2007). The threat of punishment enforces peaceful cooperation and stabilizes queues in a coral-reef fish. Proceedings of the Royal Society of London B, 274, 1093–1099. Young, A. J. & Clutton-Brock, T. H. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387. Young, A. J., Carlson, A. A., & Clutton-Brock, T. H. (2005). Trade-offs between extraterritorial prospecting and helping in a cooperative mammal. Animal Behaviour, 70, 829–837. Young, A. J., Carlson, A. A., Monfort, S. L., et al. (2006). Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proceedings of the National Academy of Sciences of the USA, 103, 12005–12010. Young, A. J., Spong, G., & Clutton-Brock, T. H. (2007). Helper males prospect for extra-group matings: alternative reproductive tactics in a cooperative mammal. Proceedings of the Royal Society of London B, 274, 1603–1609. Young, A. J., Monfort, S. L., & Clutton-Brock, T. H. (2008). The causes of physiological suppression among female meerkats: a role for subordinate restraint due to the threat of infanticide? Hormones and Behavior, 53, 131–139. Ziegler, T. E., Savage, A., Scheffler, G., & Snowdon, C. T. (1987). The endocrinology of puberty and reproductive functioning in female cotton-top tamarins (Saguinus oedipus) under varying social conditions. Biology of Reproduction, 37, 618–627. Ziegler, T. E., Epple, G., Snowdon, C. T., et al. (1993). Detection of the chemical signals of ovulation in the cotton-top tamarin, Saguinus oedipus. Animal Behaviour, 45, 313–322.
IV
Future directions
15
Understanding variation in reproductive skew: directions for future empirical research sarah j. hodge
Summary Reproductive skew models provide a powerful theoretical tool for understanding the way reproduction is shared in animal societies. However, testing skew models empirically has proved problematic, partly due to the difficulties of ensuring that all the assumptions of the model have been met, and partly because the parameters to be tested often correlate with other factors likely to influence skew. Two broad sets of processes are likely to act in tandem to influence reproductive skew in societies where dominants and subordinates are equally capable of breeding successfully. First, reproductive skew is likely to be influenced by the extent to which subordinate reproduction is costly to dominant females, as dominants would only be expected to disrupt subordinate breeding attempts if they benefit from doing so. Second, reproductive skew is likely to be influenced by the extent to which dominants are able to disrupt subordinate breeding attempts. Before testing the predictions of competing skew models, empirical work should investigate the underlying causes of variation in reproductive skew in the species in question. This will involve a detailed investigation of (1) the factors that influence female breeding success that are not a consequence of interference from other parties (such as age, condition, access to unrelated breeding partners); (2) the costs and benefits that dominant and subordinate individuals experience when breeding together; and (3) the tactics that Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
439
440
S. J. Hodge competing parties use to maximize their own breeding success. This should yield insights into the extent to which the observed level of skew is optimal for either competing party, which would provide better knowledge of which party is in control of reproduction and which skew models are applicable. This information could be used to identify those skew models that make assumptions appropriate to particular model systems (if any), and then to test their predictions. While existing skew models do encompass many of the factors likely to influence reproductive skew, there are a number of situations in which they cannot be readily applied. Future theoretical work to encompass situations where subordinates do not breed because they are in poor physical condition or because they do not have access to unrelated mating partners, and those where dominant females suffer little or no initial cost when subordinates breed, are likely to be fruitful.
Introduction Understanding the causes of variation in the way that reproduction is distributed among members of social groups, or the degree of “reproductive skew,” has been a major focus of research in behavioral ecology (Sherman et al. 1995). A central component of this research has been the development of theoretical models that attempt to predict the way reproductive skew will vary in response to changes in key social, genetic, and ecological variables (a body of work known collectively as reproductive-skew theory: reviewed by Keller & Reeve 1994, Johnstone 2000, Magrath et al. 2004). However, while reproductiveskew theory has yielded some valuable insights into the factors that may influence the distribution of reproduction, it is now clear that testing these models empirically is far from straightforward (Clutton-Brock 1998, Magrath & Heinsohn 2000, Magrath et al. 2004, Buston et al. 2007). This is because small changes to a model’s assumptions can dramatically change the predicted outcome, which means that tests of skew models can only be usefully conducted once all of the model’s assumptions have been met, even those not explicitly stated (Johnstone 2000, Magrath & Heinsohn 2000, Kokko 2003). Furthermore, many of the parameters included in the models correlate with other factors likely to influence skew, making it difficult to draw conclusions about the applicability of different models without first being aware of, and adequately controlling for, potential confounds (Emlen 1996, Clutton-Brock 1998, Magrath & Heinsohn 2000). Due to the difficulties of testing skew models, theoretical work on reproductive skew has begun to outstrip our empirical understanding. There is consequently a strong need to identify
Understanding variation in reproductive skew profitable directions for empirical research, both to advance our understanding of the causes of variation in reproductive skew and to allow more appropriate tests of existing theory. Rather than allowing theory to guide empirical research, we may further our understanding of reproductive skew by prioritizing empirical research that investigates the key processes affecting the distribution of reproductive success within model systems. While this may seem obvious, most empirical work on reproductive skew to date has concentrated on testing the predictions of skew theory, and relatively few studies have fully investigated the underlying causes of variation in reproductive skew in their study species. This is unfortunate, as without detailed information on the underlying causes of variation in skew (not just those posited in the models) it is impossible to assess which models, if any, are appropriate to test, or to interpret the findings of any empirical tests conducted. Perhaps more importantly, it is only by understanding the key processes that influence reproductive skew in different model systems that we will come to understand how closely theory matches reality, and how best to reconcile any differences between the two. Given the array of factors that could influence reproductive skew in any given system, however, it is rarely clear where one should begin when attempting to assess the factors that influence skew in a given species. In this chapter, I attempt to facilitate this process by taking a step back from existing theoretical arguments about reproductive skew and looking afresh at the factors likely to influence the distribution of reproduction between group members. My aims are (1) to clarify the broad array of processes likely to influence the way that reproductive success is distributed in animal societies, and (2) to use this information to develop a framework with which to guide future empirical study. I believe this approach will improve our understanding of the fundamental processes that influence reproductive skew within animal societies, in turn facilitating more directed tests of existing skew models and the refinement of future theory. To achieve these aims, I begin by outlining two broad processes that could give rise to low reproductive skew in animal societies. I have chosen to discuss the causes of low, rather than high reproductive skew, as departing from the more pervasive interest in high-skew species may highlight key factors that are usually overlooked when investigating variation in reproductive skew. First, as dominants are only likely to benefit from disrupting subordinate breeding attempts if they suffer a net fitness cost when subordinates breed (particularly if suppression is in itself costly), I consider a range of factors that could reduce the costs of subordinate reproduction to dominant females, thereby lowering skew. Second, I consider situations where dominants experience constraints
441
442
S. J. Hodge on their ability to disrupt subordinate breeding attempts, and discuss the potential tactics that dominants and subordinates could use to maximize their reproductive share. Having outlined two key processes likely to influence reproductive skew, I then use this information to draw up a four-step framework with which to guide future empirical research. I conclude by discussing the extent to which existing skew models capture the likely causes of variation in reproductive skew in animal societies, and by suggesting future directions for theoretical research. For simplicity, I limit my discussion to females, but many of the arguments outlined are equally applicable to males. In common with theoretical models of reproductive skew, I refer to “dominant” and “subordinate” group members, even though dominance hierarchies are not always easily determined. Nevertheless, I believe it is useful to view the evolution of reproductive skew through the perspective of hypothetical dominants and subordinates, as dominance hierarchies are likely to exist even when that relationship is not obvious to the observer. Finally, as my aim is to step back from the restrictions of existing skew models, I do not confine my examples to species that breed cooperatively, as the question of how reproduction is partitioned within social groups is equally applicable to non-cooperative social species.
Potential causes of low reproductive skew In trying to understand the factors that influence variation in reproductive skew, most theoretical models assume, albeit implicitly, that dominants and subordinates are equally capable of breeding successfully, and that the distribution of reproductive success is determined solely by the outcome of competition between dominants and subordinates. However, the reproductive success of individuals will typically be influenced by a host of other factors such as their age, condition, experience, and access to unrelated breeding partners (Wasser & Barash 1983, Snowdon 1996, Chapter 14 in this volume). These factors alone could generate marked reproductive skew, even in the absence of competition over reproduction between dominants and subordinates, and will commonly affect the observed level of skew when individuals do compete over reproduction. Seeking to understand the influence of factors other than competition between females on reproductive skew is therefore essential for any attempt to understand the causes of variation in reproductive skew in a given model system. If a hypothetical dominant and subordinate female are equally capable of breeding, however, there are two broad, non-mutually exclusive circumstances under which we might expect reproductive skew to be low. Subordinates
Understanding variation in reproductive skew might be expected to breed (1) when dominants suffer little or no net fitness cost when subordinates reproduce, as under these circumstances dominants are unlikely to benefit from developing costly reproductive suppression tactics, and (2) when dominants have incomplete control over subordinate reproduction. Below, I consider a broad range of factors that may lead to these scenarios, thereby promoting the evolution of lower reproductive skew. These factors (some of which are incorporated into existing skew models and some of which are not) are summarized in Box 15.1.
Box 15.1 Potential causes of low reproductive skew Two broad sets of processes are likely to act in tandem to promote lower reproductive skew in animal societies. Reproductive skew is likely to be influenced by (1) the extent to which subordinate reproduction is costly to dominant females, as dominants would only be expected to disrupt subordinate breeding attempts if they benefit from doing so, and (2) the extent to which dominants are able to disrupt subordinate breeding attempts. The factors outlined here are described in detail in the text.
Factors that may reduce the costs of subordinate reproduction to dominant individuals (a)
(b)
(c)
(d)
A relative abundance of resources. Dominant females may suffer few costs when subordinates breed if resources are sufficiently abundant that subordinate reproduction does not affect the resources available to the dominant female’s young. Differential resource allocation. If resources are limiting, dominants may suffer little cost when subordinates breed if their own offspring gain priority access to the resources contested (either because they receive preferential care or because dominant females produce more competitive young). Improved breeding success when subordinates breed. Dominant females may actually show higher breeding success when their subordinates breed alongside them. This could arise because subordinates assist in rearing the dominant’s young, because offspring survival is higher in larger litters, or if breeding with subordinates provides “assured fitness returns” because the subordinate can rear the dominant’s offspring should she die. Improved group stability when subordinates breed. Dominants may show higher breeding success when subordinates breed if group
443
444
S. J. Hodge Box 15.1 (Cont.)
(e)
(f)
productivity is higher when subordinates remain in the group, and allowing subordinates a share of reproduction ensures their retention in the group. Indirect fitness benefits. Even if dominants suffer a direct fitness cost when subordinates breed, the inclusive fitness cost may be low, or absent, if any direct cost is offset by indirect benefits gained from the reproductive success of related subordinates. Long-term fitness benefits. Even if dominants suffer an inclusive fitness cost in the short term when subordinates breed, this may be offset across a female’s lifetime if co-breeding females have a longer breeding lifespan (e.g. because they can invest less in rearing their offspring or are less likely to be predated). Incomplete control of subordinate breeding attempts
If dominants do suffer a net fitness cost when subordinates breed, they are likely to benefit from employing tactics to disrupt subordinate reproduction (e.g. infanticide, eviction, stress-related physiological suppression). But dominants may be constrained in their ability to control subordinate reproduction, promoting lower reproductive skew, if subordinates employ counter-tactics to evade dominant control or develop disruptive tactics of their own. The level of reproductive skew under these circumstances will therefore depend on relative ability of both females to enforce their own optima, i.e. the “balance of power.”
Factors that may reduce the costs of subordinate reproduction to dominant individuals
When skew is viewed as the product of tension between a dominant and subordinate, a key factor governing the amount of reproduction the subordinate gains will be the extent to which the dominant suffers a net fitness cost when subordinates breed. Where dominants suffer a large fitness cost, selection should strongly favor dominants who disrupt subordinate breeding attempts. Where dominants suffer little or no net fitness cost, selection is unlikely to favor dominants who disrupt subordinate breeding attempts, particularly if suppression is in itself costly (Bell 2007). Understanding the costs and benefits of subordinate reproduction to dominants is therefore central to explaining variation in reproductive skew, but remarkably few studies have attempted to quantify these costs and benefits or investigate the extent to which they may vary. Below, I consider a broad array of factors that may act
Understanding variation in reproductive skew either alone, or in concert, to reduce the net fitness costs that dominants experience when subordinates breed, thereby promoting lower reproductive skew. Where applicable, I illustrate my argument with examples that reflect a role for the factor in question, but it should be noted that it is not my intention to imply that this factor alone determines the observed level of reproductive skew in the examples chosen. A relative abundance of resources
If subordinate females breed alongside dominants, the number of offspring to be reared will increase, which could result in increased competition for food or care if resources are limiting (Lessells 1991). This could reduce the resources available to the offspring of the dominant female, generating strong selection for dominants to disrupt subordinate reproduction (Cant 1998, Cant & Johnstone 1999). Perhaps the most common situation where dominants may suffer little or no fitness cost when subordinates breed, therefore, is a relative abundance of resources for reproduction, so that subordinate group members can breed without markedly reducing the resources available to the dominant. This may help to explain why skew is low in separate-nesting plural-breeding birds, for example, where food and safe nest sites are sufficient for more than one pair within a territory (Brown 1987), and may also explain why reproductive skew is low in many group-living mammals, such as ungulates, macropods, and non-callitrichid primates (Jarman 1974, Clutton-Brock 1989, Russell 2004). Clearly, skews in reproductive success will still arise in these societies, either through variation in the ability of females to exclude others from the most valuable resources (e.g. red deer, Cervus elaphus: Clutton-Brock et al. 1988), or from dominants employing tactics to ensure that the number of breeding females does not grow too large (Hodge 2003). The critical importance of resource availability has attracted little theoretical or empirical attention in the context of reproductive skew, perhaps because direct competition over reproduction is less obvious in such societies. Nonetheless, situations where resources are relatively abundant may account for many occasions where skew is low and should therefore not be overlooked. Differential resource allocation
If resources required for reproduction are limiting, the cost of subordinate reproduction to dominant females may still be relatively low if the dominant female can be confident that her own offspring will receive a greater share of those resources available (S. J. Hodge et al. unpublished). For example, if the offspring of the dominant female receive preferential care (either from their mother or from other group members) then this may reduce the impact that the
445
446
S. J. Hodge subordinate female’s offspring have on the growth and survival of the dominant female’s young. This may help to explain why female ostriches (Struthio camelus) tolerate other female’s eggs in their clutch, as dominant females place their own eggs in the centre of the nest, ensuring effective incubation for their own eggs despite the increased clutch size (Bertram 1979, 1992). Similarly, female common eiders (Somateria mollissima) commonly cre`che young, but as mothers are aggressive towards offspring that are not their own, larger females secure relatively safe positions for their own offspring in the centre of the group (O¨st & Back 2003). While preferential care of this kind has received little attention in the context of reproductive skew, it is likely that whenever females are able to recognize their own young, they may be able to bias care towards their own offspring. If dominants are better able to ensure that their own offspring receive preferential care than subordinates, the costs that dominants experience when subordinates breed may be lowered, reducing their need to suppress subordinate reproduction and lowering reproductive skew (S. J. Hodge et al. unpublished). Differential allocation of resources could also arise in the absence of individual recognition of young, if the offspring of dominant and subordinate females differ in their ability to compete for food or care. In banded mongooses (Mungos mungo), dominant females consistently produce larger offspring than subordinates, and larger offspring are better able to compete for care (S. J. Hodge et al. unpublished). This is likely to reduce the cost of subordinate reproduction to dominant females, helping to explain, at least in part, why multiple females regularly breed. Competitive differences between the offspring of dominant and subordinate females may also arise from the order in which females produce young, as older offspring often have an advantage in competition for resources (e.g. Norway rats, Rattus norvegicus: Mennella et al. 1990). Competitive differences of this kind may help to explain temporal variation in reproductive skew in meerkats (Suricata suricatta). In this species, reproductive skew is generally high, and dominant females use infanticide to control reproduction by subordinates (Clutton-Brock et al. 2001, Young & Clutton-Brock 2006). However, dominants are markedly less likely to kill the offspring of subordinate females if the subordinate gives birth after the dominant (Clutton-Brock et al. 2001, Young & Clutton-Brock 2006). As younger meerkat pups compete less effectively for care, these rare incidences of successful subordinate reproduction may reflect the fact that subordinate litters born after the dominant’s, or when the dominant is not breeding, have little detrimental effect on the dominant’s own reproductive success (S. J. Hodge, unpublished). Differences between mothers in the competitive ability of their offspring may be commonplace, and have the potential to affect the degree of skew, but as yet, this area has received little empirical or theoretical attention.
Understanding variation in reproductive skew Improved breeding success when subordinates breed
Dominants may gain little from disrupting subordinate breeding attempts if they actually produce more surviving offspring when breeding alongside subordinates than when subordinates are reproductively suppressed. There are a number of reasons why dominant females could be more successful when breeding alongside subordinates. The survival of the dominant female’s offspring may be higher when several females breed if the increase in offspring numbers provides protection from predation due to a “dilution” effect (e.g. black-tailed prairie dogs, Cynomys ludovicianus: Hoogland et al. 1989), or because rearing offspring with other females provides offspring thermoregulatory benefits (e.g. white-footed mice, Peromyscus leucopus: Jaquot & Vessey 1994). Co-breeding female mammals commonly suckle one another’s young, and there is evidence from several species that this communal nursing improves the growth and survival of pups (for reviews see Lewis & Pusey 1997, Hayes 2000). Avian dominant females may also benefit from breeding alongside subordinates if communally nesting females share incubation of the joint clutch (e.g. groove-billed anis, Crotophaga sulcirostris: Vehrencamp 1977) or if brood amalgamation allows females to invest less time in vigilance (e.g. com¨ st et al. 2002), both of which are likely to increase their time mon eider: O available for foraging. Breeding females may also join forces to defend offspring from predators (e.g. fat dormouse, Glis glis: Pilastro 1992), infanticidal males (e.g. ring-tailed lemurs, Lemur catta: Pereira & Weiss 1991; house mice, Mus musculus: Manning et al. 1995; lions, Panthera leo: Packer et al. 2001), or competitors for safe breeding sites (e.g. Polistes wasps: Gamboa 1978), all of which may improve offspring survival. Finally, females may benefit from breeding communally if the likelihood that they will die before their own young are reared is high; breeding alongside subordinates may therefore provide “assured fitness returns” for the dominant female if the subordinate will rear the communal brood should the dominant be predated (e.g. Polistes wasps: Gadagkar 1990; sweat bees, Megalopta genalis: Smith et al. 2007). The above examples suggest that there are a number of reasons why dominants might actually produce more surviving offspring when subordinates breed than when they are the sole breeder in their group. However, explicit comparisons of the fitness between dominant females breeding singly and those breeding communally are remarkably rare. Some studies have provided correlative evidence suggesting that communal breeding is advantageous (Table 15.1), but interpreting correlational studies of this kind is difficult because subordinates may be more likely to associate with good-quality females, or with females on good-quality territories. The finding that females have higher reproductive success when co-breeding may therefore arise
447
448
S. J. Hodge because of the dominant’s (or territory’s) superior quality, rather than because of the presence of the breeding subordinate per se (Tibbetts & Reeve 2003). Perhaps the best investigations of the costs and benefits of co-breeding to date come from experimental work on rodents (Table 15.1) and social insects (Heinze & Oberstadt 2003, Lopez-Vaamonde et al. 2003, Tibbetts & Reeve 2003) in which the number of subordinate co-breeders was experimentally manipulated and the reproductive output of dominant and subordinate females measured (Table 15.1). However, as many of these studies compared females breeding alone with those experimentally paired with another female, the treatments differed in group size. As a consequence it is often unclear whether any effects are due to differences in the number of breeders per se, or to differences in the number of individuals in the group. This distinction was made clear in a study of prairie voles (Microtus ochrogaster) in which experimental comparisons were made between the growth of pups born to females rearing pups alone, females rearing pups with a non-breeding subordinate, and females rearing pups communally with another breeding (and hence lactating) female. Females that bred communally produced pups that grew faster than those of females that bred alone or in groups with non-breeding subordinate females (Hayes & Solomon 2004). This work indicates that the presence of a breeding subordinate, rather than the presence of a subordinate per se, had the strongest influence on pup growth rates. Given the critical importance of understanding the costs and benefits experienced by dominants when cobreeding, further empirical studies of this kind should be prioritized. This work will be most valuable if skew is experimentally varied while keeping group size constant, or if breeder and territory quality are controlled statistically. Improved group stability when subordinates breed
The fitness costs that dominants experience when subordinates breed may also be low, favoring lower reproductive skew, if the presence of subordinates improves group productivity, and subordinates are less likely to disperse if they are allowed a share of reproduction. The idea that dominants allow subordinates “staying incentives” of this kind in return for cooperation is central to concession models of reproductive skew (Vehrencamp 1979, 1983). While a number of empirical studies have tested the predictions of these models (e.g. Jamieson 1997, Reeve & Keller 2001, Haydock & Koenig 2003, Langer et al. 2004, Heg et al. 2006a), there is little conclusive empirical evidence to support the key assumption that females are less likely to disperse if they are allowed a share of reproduction (Clutton-Brock 1998). The lack of data linking a subordinate’s likelihood of dispersal with her reproductive share is
450
449
Table 15.1 Examples of vertebrate studies which have compared the costs and benefits of communal breeding Is co-breeding beneficial in Species
comparison to
Type of benefit
breeding alone?
investigated
Data type
always change?
Reference
Yes (but only for
Reproductive
Correlation
Yes
Vehrencamp
# chicks fledged
Correlation
Yes
# fledged chicks
Correlation
No (some singular Mumme
Groove-billed ani
(Crotophaga
Moorhen
(Gallinula chloropus) Yes (but only
sulcirostris)
alpha female)
Did group size
rate per year
1978 McRae 1996
for senior female Acorn woodpecker
(Melanerpes formicivorus)
No (but costs may be offset over lifetime)
White-footed mouse
(Peromyscus
No difference
per female per
breeders were
year # pups weaned
et al. 1988
in groups) Correlation
leucopus)
No (some singular Wolff 1994 breeders were in groups)
Deer mouse
(Peromyscus
No difference
# pups weaned
Correlation
maniculatus)
No (some singular Wolff 1994 breeders were in groups)
House mouse
(Mus musculus)
Yes
LRS
Experiment
Yes
Ko¨nig 1994
Wood mouse
(Apodemus
No (but in some
# weaned and
Experiment
Yes
Gerlach &
sylvaticus)
cases costs may
pup growth
Bartmann
be offset by
2002
indirect benefits) Prairie vole
(Microtus
Yes (for both
ochrogaster)
Pup growth
Experiment
females)
No (some singular Hayes & breeders were
Solomon 2004
in groups) Tuco-tuco
(Ctenomys sociabilis)
Banded mongoose (Mungos mungo)
No (but costs may be offset over lifetime) Yes (but not in large communal groups)
# pups weaned
Correlation
Yes
Lacey 2004
Correlation (but
No (some singular Gilchrist 2006
per female and LRS Independent pups per female
group and territory
breeders were
quality controlled
in groups)
statistically) Fat dormouse
(Glis glis)
No difference
# pups weaned
Correlation
Yes
Experiment
No (some singular Ebensperger
per female Degu
(Octodon degus)
No difference (but females Pup growth producing second born
et al. 1996
and # weaned
breeders were
litter may suffer a cost) Common vole
(Microtus arvalis)
No
Pilastro
et al. 2007
in groups) # daughters surviving to breed
Correlation
Yes
Boyce & Boyce 1988
The costs and benefits of co-breeding have also been studied experimentally in insects, in particular Polistes wasps (e.g. Tibbetts & Reeve 2003), Leptothorax ants (e.g. Heinze & Oberstadt 2003), bumblebees (Lopez-Vaamonde et al. 2003), and beetles (e.g. Scott 1997, Heg et al. 2006b). LRS, Lifetime reproductive success.
Understanding variation in reproductive skew due, in part, to the difficulty of establishing whether subordinates are allowed to breed more because they have a high chance of successful dispersal or because of other factors that influence both subordinate reproduction and dispersal (e.g. age, experience, and competitive ability: Clutton-Brock 1998). There is, however, some evidence to support the related idea that dominants use reproduction as a “peace incentive” to appease potential challengers (Reeve & Ratnieks 1993, Reeve & Keller 1997). Subordinate female Polistes dominulus wasps were found to be less likely to challenge a returning dominant if they had previously reproduced (Cant et al. 2006). If dominant females do allow subordinates a share of reproduction in return for peaceful cooperation, then we might expect reproductive skew to be lower when subordinates have a higher chance of breeding independently, as dominants may be required to concede a larger share of reproduction to retain the subordinate’s cooperation (Vehrencamp 1983). Several studies have investigated the role that “ecological constraints” (Hatchwell & Komdeur 2000) on a subordinate’s chance of breeding independently can have on reproductive skew, and several have shown that skew is lower when constraints are low, both within (Jamieson 1997, Haydock & Koenig 2003) and across species (Bourke & Heinze 1994). However, caution must be exercised when investigating the influence of ecological constraints on reproductive skew, as ecological constraints may influence both the amount of suppression that subordinates are willing to tolerate if they remain with the dominant, and also whether a subordinate chooses to remain with the dominant at all (Emlen 1982, Hatchwell & Komdeur 2000). Ecological constraints on independent breeding could therefore make skew appear higher simply because there are more individuals with whom to share reproduction, rather than because subordinates are more willing to tolerate suppression (Magrath et al. 2004). Perhaps the best investigation of the influence of ecological constraints on reproductive skew to date comes from experimental studies on bees (Exoneura nigrescens: Langer et al. 2004) and cichlids (Neolamprologus pulcher: Heg et al. 2006a), in which the availability of independent breeding opportunities was manipulated while keeping group size constant. In both of these studies, the experimental manipulation of independent breeding positions was found to have no influence on the way the reproduction was shared. Indirect fitness benefits
Even if dominants do suffer a direct fitness cost when subordinates breed, the loss in direct fitness may be offset by indirect fitness benefits derived from the reproductive success of subordinate relatives. This could be of particular importance in promoting lower reproductive skew in cooperative
451
452
S. J. Hodge and communally breeding species where female group members are commonly close kin (Emlen 1995), and is a key component of many theoretical models. While numerous empirical studies have investigated whether correlations between reproductive skew and relatedness exist (e.g. Haydock & Koenig 2002, 2003, Langer et al. 2004, Liebert & Starks 2006), comparatively few have investigated the extent to which indirect fitness benefits arising from subordinate reproduction may compensate dominants for any associated reduction in their own reproductive success. The best way to achieve this is to quantify the direct and indirect fitness payoffs to dominants of breeding alongside a subordinate relative and compare this to those accrued when the subordinate does not breed: i.e. compare the reproductive success of dominants when breeding alone (DA) with the reproductive success of dominants when breeding with subordinates (DS þ rSD, where DS is the reproductive success of the dominant when the subordinate breeds, r is the relatedness between dominant and subordinate, and SD is the reproductive success of the subordinate when breeding alongside the dominant). The direct and indirect fitness payoffs that arise when co-breeding with females of differing relatedness versus breeding alone were compared in this way in the wood mouse (Apodemus sylvaticus). Experimentally manipulating group composition revealed that females who bred with a daughter suffered a small loss of direct fitness that was compensated by their indirect fitness payoff. However, females who bred with sisters or unrelated females suffered a larger drop in direct fitness, which could not be compensated by increases in indirect fitness (Gerlach & Bartmann 2002). Further empirical tests of this kind are required to fully understand the role that indirect fitness benefits may play in maintaining low reproductive skew. Long-term fitness benefits
If dominant females produce fewer offspring when breeding alongside other females, and this cost is not immediately offset by indirect benefits, selection may still favor dominants who tolerate subordinate reproduction if co-breeding females produce more offspring over their lifetimes than females that breed alone. This could arise if co-breeding females have longer breeding lifespans, perhaps because they invest less in rearing each offspring, or if they are less likely to be depredated (Ebensperger et al. 2007). Such long-term benefits of co-breeding are thought to maintain joint-nesting in acorn woodpeckers (Melanerpes formicivorus), as co-breeding females have a higher annual survival rate than females who are the only breeder within their group, which means the lifetime breeding success of co-breeding females is similar to that of females who breed singularly (Mumme et al. 1988). Long-term benefits have also been invoked to explain communal breeding in colonial tuco-tucos
Understanding variation in reproductive skew (Ctenomys sociabilis), where differences in the survival of solitary versus communally breeding females are thought to offset the direct cost of breeding in a social group (Lacey 2004). Few studies have considered the influence that longitudinal benefits of co-breeding could have on reproductive skew, perhaps because this requires long-term data that may not be easily available. While theory has begun to incorporate the likely long-term benefits that subordinates may gain from breeding associations (e.g. the possibility of territory inheritance: Ragsdale 1999, Kokko & Ekman 2002), theoretical consideration of the benefits that dominants may gain from subordinate reproduction over their lifetimes could prove profitable. Incomplete control over subordinate reproduction
I have discussed an array of situations where dominants may suffer little or no net fitness cost when subordinates breed, and may therefore gain little from suppressing subordinate reproduction, favoring lower reproductive skew. Where dominant females do suffer a net fitness cost, selection would be expected to favor dominants who monopolize reproduction by disrupting the breeding attempts of subordinates. However, even if dominants do suffer a fitness cost when subordinates breed, skew may still be relatively low if dominants are constrained in their ability to disrupt subordinate reproduction. Under these circumstances, the extent of reproductive skew is likely to depend upon the balance of power in reproductive competition between dominants and subordinates (Beekman et al. 2003). Below, I briefly discuss some of the ways in which dominants may attempt to control subordinate reproduction, and the tactics subordinates may use to evade this control. The most direct means by which dominants can control subordinate reproduction is to kill any offspring that subordinates produce, a behavior that has been observed in a variety of taxa (Vehrencamp 1977, Mumme et al. 1983, Hoogland 1985, Bourke, 1994, Clutton-Brock et al. 1998, Macedo et al. 2004). Killing the offspring of subordinates will only be an effective means of control, however, if dominants can be confident that they will not inadvertently kill their own young in the process (Hager & Johnstone 2004). For this reason, such tactics may be more successful when dominants have not yet produced young of their own (Vehrencamp 1977, Mumme et al. 1983, Clutton-Brock et al. 1998) or when individual offspring can be easily recognized (Beekman et al. 2003). Dominants could also exert control over subordinates before subordinates produce young by evicting those females that are most likely to breed from the group. Indeed, dominant female meerkats and banded mongooses evict females that are pregnant (Hodge 2003, Young et al. 2006), and some socialinsect queens prevent females from entering the nest if they have had contact
453
454
S. J. Hodge with an unrelated male (Bull et al. 1998). Dominants may also employ tactics that allow them to suppress subordinate reproductive attempts while retaining them in the group. For example, in ponerine ants, breeding queens mutilate subordinates to make them incapable of breeding (Monnin & Ratnieks 2001), and in meerkats, dominants compromise subordinate fertility through repeated stressful attacks (Young et al. 2006, Chapter 14 in this volume). Dominants may be constrained in their ability to use these tactics to monopolize reproduction, however, if subordinates develop ways of evading control, or if subordinates develop tactics of their own to reduce the dominant’s reproductive share. Subordinates may prevent dominants from killing their offspring by giving birth in a separate den or nest (e.g. black-tailed prairie dogs: Hoogland et al. 1989; lions: Pusey & Packer 1994), by synchronizing offspring production with the dominant female (Vehrencamp 1977, Mumme et al. 1983, Agrell et al. 1998, Ebensperger 1998), or by scrambling odor cues to prevent dominants removing their eggs (e.g. social insects: Monnin & Ratnieks 2001). Subordinates could also evade dominants’ control by becoming aggressive or forming alliances with other group members, increasing the costs that dominants experience when attempting to prevent them from breeding (e.g. chacma baboons, Papio cynocephalus: Palombit et al. 1997). In some species, subordinates have also been shown to use control tactics of their own to reduce the reproductive success of the dominant female. For example, workers in some social-insect societies destroy the eggs of queens (Bourke & Franks 1995), skew in the broods of some joint-nesting birds is the outcome of mutual egg tossing by all females (Vehrencamp 1977, Mumme et al. 1983), and subordinate female meerkats kill the offspring of dominants (Young & CluttonBrock 2006). Given the potential for evasion and control tactics among subordinates, if dominants do suffer a net fitness cost when subordinates breed, the level of reproductive skew will be determined largely by the relative ability of each female to enforce her own optimum: the balance of “power” (Beekman et al. 2003). Despite the critical importance of understanding how such power struggles are resolved, however, our knowledge of the limitations of the various tactics employed by dominants and subordinates remains poor (see Chapter 14). Understanding the way that power is distributed is also essential if we are to identify which theoretical models are applicable to a given system, as models of reproductive skew differ in their assumptions regarding which individuals have control over reproduction and how complete that control is (Johnstone 2000, Magrath et al. 2004). The main difficulty in investigating the extent of dominant control empirically is that it is often impossible to distinguish between situations where dominants “allow” subordinates a share of
Understanding variation in reproductive skew reproduction and those where there are constraints on the dominants’ ability to control. Mutual aggression cannot be used as evidence that dominants do not have complete control, as aggression may simply be the means through which dominance is established and tested (Emlen 1999). To further our understanding of the causes of variation in reproductive skew, it is therefore essential that we develop ways of empirically comparing the ability of dominants and subordinates to skew the distribution of reproduction in their favor.
Directions for future empirical research Much of the empirical investigation into the causes of reproductive skew to date has concentrated on testing the predictions of different skew models. While it is essential that reproductive-skew models are tested empirically to establish their biological relevance, testing the predictions of skew models will only be a valuable exercise if the study species is already known to satisfy the model’s key assumptions (Magrath et al. 2004). For example, it makes little sense to test the predictions of a model in which complete dominant control is assumed, in a species in which the dominant female does not have complete control over subordinate reproduction, as even if the patterns of skew do conform to the model’s predictions, they may do so for reasons quite different to the logic on which the model is based. For this reason, attempting to distinguish between concession and tug-of-war models using their differing predictions is likely to be of little value, as their underlying assumptions differ completely, so that only one category of model, if any, will be appropriate to test in a given species. Switching the focus of empirical research towards fully understanding the causes of reproductive skew in the study system in question, rather than blindly testing the predictions of competing models, will both advance our fundamental understanding of the causes of reproductive skew and allow more appropriate tests of existing theory. It will also allow factors likely to influence skew that are not currently incorporated into existing models to be identified, helping to bridge the gap between theory and reality. In discussing the key processes likely to influence reproductive skew, I have highlighted a number of areas that warrant investigation if we are to fully understand the causes of reproductive skew in a given system. In particular, empirical studies should consider: (1) the factors that influence a female’s breeding success regardless of interference from other group members (e.g. age, body condition, experience, and access to unrelated breeding partners); (2) the costs and benefits of co-breeding to dominant and subordinate individuals; (3) the extent to which dominants have complete control of subordinate
455
456
S. J. Hodge reproduction; and (4) the extent to which the answers to the above questions vary in different social and ecological situations. Below I outline these four key steps for empirical research in more detail (represented graphically in Figure 15.1). Would subordinates breed at comparable rates to dominants in the absence of dominant control?
A sensible starting point when attempting to understand the causes of variation in reproductive skew is to investigate the extent to which the reproductive success of subordinates is determined by factors other than reproductive conflict with the dominant. If, for example, the subordinate female does not have access to unrelated breeding partners, or if her physical condition and/or experience mean that successful reproduction is unlikely, skew may tend to be high regardless of the interests and behavior of the dominant (see Chapter 14). Experimentation can be powerfully employed to investigate these issues. For example, in Damaraland mole rats (Cryptomys damarensis: Cooney & Bennett 2000) and pine voles (Microtus pinetorum: Brant et al. 1998), the experimental addition of unrelated breeding partners to the group increased the likelihood that subordinates bred, even in the presence of the dominant female. These experiments indicate that inbreeding avoidance at least partly explains the low incidences of subordinate reproduction in these species. Although experimental work of this kind will not always be feasible, an understanding of the availability of unrelated breeding partners, and the influence of age, condition, and experience on the probability of a female breeding, will be an essential starting point when attempting to understand the underlying reasons for variation in reproductive skew (Wasser & Barash 1983). This will be of particular importance when attempting to test reproductive-skew models, as these causes of variation in reproductive skew are rarely incorporated into skew theory. Does the dominant suffer a net fitness cost when subordinates breed?
One of the biggest challenges facing researchers attempting to test reproductive-skew models is understanding whether one party has complete control over reproduction. This requires distinguishing between those situations where dominants “allow” subordinates a share of reproduction and those where dominants are unable to prevent subordinates from breeding. Perhaps the only way to fully understand who has control over reproduction (and hence which models are appropriate) is to look at whose optimum is favored: do dominants benefit when subordinate females reproduce, or is their own reproductive success lowered? If dominants do suffer a reduction in direct
Understanding variation in reproductive skew
INFLUENCED BY: - Subordinate female condition, age, experience - Access to unrelated breeding partners
(1) Would subordinates breed at comparable rates to dominants in the absence of dominant control?
NO
HIGH SKEW
YES
INFLUENCED BY: - Effect of subordinate reproduction for dominant breeding success - Effect of subordinates on dominant survival - Relatedness between females
(2) Do dominants suffer a net fitness cost when subordinates breed?
NO
LOW SKEW
YES
INFLUENCED BY: - Subordinate opportunities for independent reproduction - Relatedness between females
(3) NO
Does the dominant have complete control over subordinate reproduction?
SKEW WILL DEPEND ON DISTRIBUTION OF POWER
YES
HIGH SKEW
Figure 15.1. Schematic representation of an empirical approach to investigate the causes of variation in reproductive skew.
reproductive success that is not offset by increases in indirect or future fitness returns, then this would strongly suggest that subordinate reproduction does not arise from dominants with complete control adaptively “conceding” a reproductive share, but instead from constraints in the dominants’ ability to control. Attempts to compare the fitness of dominants in a given model system under different levels of reproductive skew must control for confounding factors such as variance in territory quality and group size that may affect the fitness of the dominants. Multifactorial statistics therefore represent a valuable tool to facilitate this process. Another way to resolve this problem is to
457
458
S. J. Hodge experimentally prevent subordinates from breeding in a species where skew is usually low (for example, by restricting access to mating partners, or hormonally blocking conception). An experiment of this kind would allow the success of dominants when breeding in high- and low-skew situations to be compared while keeping group size constant, and would also allow the effects of changes in skew on the behavior of dominant and subordinate individuals to be assessed (Cant 2006). This is a particularly powerful approach, as it would allow an experimental test of whether subordinates become aggressive or leave the group if their share of reproduction is reduced: an assumption key to theoretical approaches to skew based on reproductive concessions (Vehrencamp 1983, Reeve & Ratnieks 1993, Reeve & Keller 1997). Attempts to compare the costs and benefits of co-breeding versus singular breeding will be especially revealing if they quantify the direct and indirect fitness costs and benefits gained by both dominants and subordinates. As well as providing a quantitative comparison of the fitness of dominants and subordinates in high- and low-skew situations, this approach will also allow the underlying reasons for variation in reproductive skew to be ascertained (Box 15.1). For example, do dominants produce more offspring when subordinates breed because subordinate co-breeders assist in rearing young, or do dominants suffer a direct fitness cost when breeding alongside subordinates which is offset by the inclusive fitness benefits gained from breeding with close relatives? To what extent does the dominant have complete control over subordinate reproduction?
If investigation of the costs and benefits of subordinate reproduction to both parties reveals that the dominant female does suffer a net fitness cost when subordinates breed, then reproduction by subordinates is likely to arise through constraints on dominant control. Under these circumstances, the degree of reproductive skew is likely to represent the outcome of a tactical power struggle between dominants and subordinates, each seeking to maximize their own reproductive success. A detailed investigation of the mechanisms that dominants use to control subordinates, and of the ways in which subordinates may evade this control, would therefore help us to understand the effectiveness and limitations of different control tactics, which would allow us a better understanding of the way power is distributed. The amount that dominants and subordinates invest in maximizing their reproductive share could be influenced by a number of factors, such as the degree of relatedness between females and the availability of independent breeding opportunities,
Understanding variation in reproductive skew and these will need to be investigated if the distribution of power is to be fully understood. How do answers to the above questions vary in different social and ecological situations?
While I have discussed two broad processes that may promote low skew in any given species (dominants suffering little or no cost when subordinates breed, and dominants suffering constraints on their ability to disrupt subordinate reproduction), it is clear that both processes will act in tandem to determine the extent of reproductive skew. As both the dominant’s optimum level of skew and her ability to enforce it are likely to vary with social and ecological conditions (such as group size, food availability, and female condition), fully understanding the causes of variation in reproductive skew will require an understanding of how the answers to each of the above questions vary over time as these conditions change.
Directions for future theoretical research In this chapter I have discussed some of the key processes likely to influence reproductive skew in animal societies, focusing in particular on the importance of the costs and benefits of subordinate reproduction to dominant individuals, and the ability of each party to maximize her own reproductive share. While some factors that I have discussed are incorporated into current skew theory, others are not, and two areas in particular would be profitable for future theoretical study. First, most current theoretical models assume that reproductive skew will be the product solely of competition between individuals, and do not allow for the likelihood that females will commonly differ in their ability to produce offspring, even in the absence of competition. The many systems in which female reproductive success is influenced by their age, condition, experience, or access to unrelated breeding partners therefore lie outside the bounds of existing reproductive skew models. The development of models that incorporate variation in the capacity of different females to breed would consequently make reproductive-skew theory applicable to a far wider range of species. Second, important insights into the evolution of reproductive skew could be generated by allowing the initial fitness cost that dominant females experience when subordinates breed to vary. Most models make the assumption that reproduction by the subordinate female involves an initial direct fitness cost to the dominant (i.e. dominant direct fitness is a decreasing function of P, where P
459
460
S. J. Hodge is the reproductive share allocated to the subordinate), a cost which must be offset by the direct and indirect benefits of allowing subordinates to breed for low skew to be stable (Keller & Reeve 1994, Johnstone 2000, Magrath et al. 2004). However, if the initial cost is negligible (for example because resources necessary to rear offspring are relatively abundant, or the offspring of dominant females have priority access to resources), then the predicted outcome is likely to differ substantially, as fewer benefits will be required to make low reproductive skew stable (S.J. Hodge et al. unpublished). This could be explored generally, by developing theoretical models that allow the initial costs that dominants experience when subordinates breed to vary. This could also be investigated more specifically, by incorporating differences in the way resources are allocated between the offspring of dominant and subordinate females (for example by allowing the competitive ability of offspring of dominant and subordinate females to vary, or allowing mothers to care preferentially for their own young).
Conclusions In this chapter I have moved away from existing theoretical arguments about reproductive skew and considered afresh the causes of low reproductive skew. My aim was first to clarify the broad array of processes likely to influence the way that reproductive success is distributed in animal societies, and then to use this information to develop a set of priorities for future empirical research that will improve our understanding of the underlying causes of variation in reproductive skew. I have highlighted four main areas that should be considered by any empirical study attempting to investigate reproductive skew. First, empiricists should investigate the roles that factors other than competition between individuals play in precipitating reproductive skew, such as a female’s age, condition, experience, and access to unrelated breeding partners. Second, key insights will be generated by investigating the costs and benefits of co-breeding to dominant and subordinate group members (Box 15.1, Table 15.1). This will not only allow a better understanding of why skew varies, but will also help to assess whether dominant females have complete control over subordinate reproduction, by identifying whose optimum is closer to the observed level of skew. Third, empirical research should investigate the tactics that both parties use to maximize their reproductive success, as this will allow an understanding of the effectiveness and limitations of different control and evasion tactics. Finally, investigating the extent to which answers to the above three questions vary with changing social and ecological conditions will provide an understanding of how the optimum
Understanding variation in reproductive skew levels of reproductive skew vary for both parties over time. I believe that prioritizing empirical research that addresses these questions will advance our understanding of the broader processes likely to influence reproductive skew, both within and across animal societies. This will not only allow more rigorous testing of existing theoretical approaches to understanding reproductive skew, but will also help to assess how widely skew theory matches reality, and allow the refinement of skew theory to incorporate these empirical advances. Acknowledgments I would like to thank Reinmar Hager and Clara B. Jones for inviting me to contribute to this volume, and all of the members of the Large Animal Research Group at the University of Cambridge for countless discussions about reproductive skew. Matthew Bell, Mike Cant, Tim Clutton-Brock, Dieter Lukas, Andy Radford, and Andy Young provided valuable comments on this chapter, for which I am very grateful. References Agrell, J., Wolff, J. O., & Ylonen, H. (1998). Counter-strategies to infanticide in mammals: costs and consequences. Oikos, 83, 507–517. Beekman, M., Komdeur, J. & Ratnieks, F. L. W. (2003). Reproductive conflicts in social animals: who has power? Trends in Ecology and Evolution, 18, 277–282. Bell, M. B. V. (2007). Communication, cooperation and conflict in banded mongooses. Unpublised Ph.D. thesis, University of Cambridge. Bertram, B. C. R. (1979). Ostriches recognise their own eggs and discard others. Nature, 279, 233–234. Bertram, B. C. R. (1992). The Ostrich Communal Nesting System. Princeton, NJ: Princeton University Press. Bourke, A. F. G. (1994). Indiscriminate egg cannibalism and reproductive skew in a multiple-queen ant. Proceedings of the Royal Society of London B, 255, 55–59. Bourke, A. F. G. & Franks, N. R. (1995). Social Evolution in Ants. Monographs in Behavior and Ecology. Princeton, NJ: Princeton University Press. Bourke, A. F. G. & Heinze, J. (1994). The ecology of communal breeding: the case of multiple – queen leptothoracine ants. Philosophical Transactions of the Royal Society of London B, 345, 359–372. Boyce, C. C. K. & Boyce, J. L. (1988). Population biology of Microtus arvalis. 1. Lifetime reproductive success of solitary and grouped breeding females. Journal of Animal Ecology, 57, 711–722. Brant, C. L., Schwab, T. M., Vandenbergh, J. G., Schaefer, R. L., & Solomon, N. G. (1998). Behavioural suppression of female pine voles after replacement of the breeding male. Animal Behaviour, 55, 615–627.
461
462
S. J. Hodge Brown, J. L. (1987). Helping and Communal Breeding in Birds: Ecology and Evolution. Princeton, NJ: Princeton University Press. Bull, N. J., Mibus, A. C., Norimatsu, Y., Jarmyn, B. L., & Schwarz, M. P. (1998). Giving your daughters the edge: bequeathing reproductive dominance in a primitively social bee. Proceedings of the Royal Society of London B, 265, 1411–1415. Buston, P. M., Reeve, H. K., Cant, M. A., Vehrencamp, S. L., & Emlen, S. T. (2007). Reproductive skew and the evolution of group dissolution tactics: a synthesis of concession and restraint models. Animal Behaviour, 74, 1643–1654. Cant, M. A. (1998). A model for the evolution of reproductive skew without reproductive suppression. Animal Behaviour, 55, 163–169. Cant, M. A. (2006). A tale of two theories: parent–offspring conflict and reproductive skew. Animal Behaviour, 71, 255–263. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Cant, M. A., English, S., Reeve, H. K., & Field, J. (2006). Escalated conflict in a social hierachy. Proceedings of the Royal Society of London B, 273, 1471–2954. Clutton-Brock, T. H. (1989). Mammalian mating systems. Proceedings of the Royal Society of London B, 236, 339–372. Clutton-Brock, T. H. (1998). Reproductive skew, concessions and limited control. Trends in Ecology and Evolution, 13, 288–292. Clutton-Brock, T. H., Albon, S. D., & Guinness, F. (1988). Reproductive success in male and female red deer. In T. H. Clutton-Brock, ed., Reproductive Success: Studies of Individual Variation in Contrasting Breeding Systems. Chicago, IL: University of Chicago Press, pp. 325–343. Clutton-Brock, T. H., Brotherton, P. N. M., Russell, A. F., et al. (2001). Cooperation, control, and concession in meerkat groups. Science, 291, 478–481. Clutton-Brock, T. H., Brotherton, P. N. M., Smith, R., et al. (1998). Infanticide and expulsion of females in a cooperative mammal. Proceedings of the Royal Society of London B, 265, 2291–2295. Cooney, R. & Bennett, N. C. (2000). Inbreeding avoidance and reproductive skew in a cooperative mammal. Proceedings of the Royal Society of London B, 267, 801–806. Ebensperger, L. A. (1998). Strategies and counter strategies to infanticide in mammals. Biological Reviews of the Cambridge Philosophical Society, 73, 321–346. Ebensperger, L. A., Hurtado, M. J., & Leo´n, C. (2007). An experimental examination of the consequences of communal versus solitary breeding on maternal condition and the early postnatal growth and survival of degu, Octodon degus, pups. Animal Behaviour, 73, 185–194. Emlen, S. T. (1982). The evolution of helping. I. An ecological constraints model. American Naturalist, 119, 29–39. Emlen, S. T. (1995). An evolutionary theory of the family. Proceedings of the National Academy of Sciences of the USA, 92, 8092–8099. Emlen, S. T. (1996). Reproductive sharing in different types of kin associations. American Naturalist, 148, 756–763.
Understanding variation in reproductive skew Emlen, S. T. (1999). Reproductive skew in cooperatively breeding birds: an overview of the issues. In N. J. Adams & R. H. Slotow, eds., Proceedings of the 22nd International Ornithological Congress. Johannesburg: BirdLife South Africa, pp. 2922–2931. Gadagkar, R. (1990). Evolution of eusociality: the advantage of assured fitness returns. Philosophical Transactions of the Royal Society of London B, 329, 17–25. Gamboa, G. J. (1978). Intraspecific defense: advantage of social cooperation among paper wasp foundresses. Science, 199, 1463–1465. Gerlach, G. & Bartmann, S. (2002). Reproductive skew, costs, and benefits of cooperative breeding in female wood mice (Apodemus sylvaticus). Behavioral Ecology, 13, 408–418. Gilchrist, J. S. (2006). Reproductive success in a low skew, communal breeding mammal: the banded mongoose, Mungos mungo. Behavioral Ecology and Sociobiology, 60, 854–863. Hager, R. & Johnstone, R. A. (2004). Infanticide and control of reproduction in cooperative and communal breeders. Animal Behaviour, 67, 941–949. Hatchwell, B. J. & Komdeur, J. (2000). Ecological constraints, life history traits and the evolution of cooperative breeding. Animal Behaviour, 59, 1079–1086. Haydock, J. & Koenig, W. D. (2002). Reproductive skew in the polygynandrous acorn woodpecker. Proceedings of the National Academy of Sciences of the USA, 99, 7178–7183. Haydock, J. & Koenig, W. D. (2003). Patterns of reproductive skew in the polygynandrous acorn woodpecker. American Naturalist, 162, 277–289. Hayes, L. D. (2000). To nest communally or not to nest communally: a review of rodent communal nesting and nursing. Animal Behaviour, 59, 677–688. Hayes, L. D. & Solomon, N. G. (2004). Costs and benefits of communal rearing to female prairie voles (Microtus ochrogaster). Behavioral Ecology and Sociobiology, 56, 585–593. Heg, D., Bergmuller, R., Bonfils, D., et al. (2006a). Cichlids do not adjust reproductive skew to the availability of independent breeding options. Behavioral Ecology, 17, 419–429. Heg, D., Heyl, S., Rasa, O. A. E., & Peschke, K. (2006b). Reproductive skew and communal breeding in the subsocial beetle Parastizopus armaticeps. Animal Behaviour, 71, 427–437. Heinze, J. & Oberstadt, B. (2003). Costs and benefits of subordinate queens in colonies of the ant, Leptothorax gredleri. Naturwissenschaften, 90, 513–516. Hodge, S. J. (2003). The evolution of cooperation in the communal breeding banded mongoose. Unpublished Ph.D. thesis, University of Cambridge. Hoogland, J. L. (1985). Infanticide in prairie dogs: lactating females kill offspring of close kin. Science, 230, 1037–1040. Hoogland, J. L., Tamarin, R. H., & Levy, C. K. (1989). Communal nursing in prairie dogs. Behavioral Ecology and Sociobiology, 24, 91–95. Jamieson, I. G. (1997). Testing reproductive skew models in a communally breeding bird, the pukeko, Porphyrio porphyrio. Proceedings of the Royal Society of London B, 264, 335–340.
463
464
S. J. Hodge Jaquot, J. J. & Vessey, S. H. (1994). Non-offspring nursing in the white-footed mouse, Peromyscus leucopus. Animal Behaviour, 48, 1238–1240. Jarman, P. J. (1974). The social organisation of antelopes in relation to their ecology. Behaviour, 48, 215–267. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Keller, L. & Reeve, H. K. (1994). Partitioning of reproduction in animal societies. Trends in Ecology and Evolution, 9, 98–103. Kokko, H. (2003). Are reproductive skew models evolutionarily stable? Proceedings of the Royal Society of London B, 270, 265–270. Kokko, H. & Ekman, J. (2002). Delayed dispersal as a route to breeding: territorial inheritance, safe havens, and ecological constraints. American Naturalist, 160, 468–484. Konig, B. (1994). Components of lifetime reproductive success in communally and solitarily nursing house mice: a laboratory study. Behavioral Ecology and Sociobiology, 34, 275–283. Lacey, E. A. (2004). Sociality reduces individual direct fitness in a communally breeding rodent, the colonial tuco-tuco (Ctenomys sociabilis). Behavioral Ecology and Sociobiology, 56, 449–457. Langer, P., Hogendoorn, K., & Keller, L. (2004). Tug-of-war over reproduction in a social bee. Nature, 428, 844–847. Lessells, C. M. (1991). The evolution of life histories. In J. R. Krebs & N. B. Davies, eds., Behavioural Ecology: an Evolutionary Approach. Oxford: Blackwell, pp. 32–68. Lewis, S. E. & Pusey, A. E. (1997). Factors influencing the occurrence of communal care in plural breeding mammals. In N. G. Solomon & J. A. French, eds., Cooperative Breeding in Mammals. Cambridge: Cambridge University Press, pp. 335–363. Liebert, A. E. & Starks, P. T. (2006). Taming of the skew: transactional models fail to predict reproductive partitioning in the paper wasp Polistes dominulus. Animal Behaviour, 71, 913–923. Lopez-Vaamonde, C., Koning, W., Jordan, W. C., & Bourke, A. F. G. (2003). No evidence that reproductive bumblebee workers reduce the production of new queens. Animal Behaviour, 66, 577–584. Macedo, R. H. F., Cariello, M. O., Graves, J., & Schwabl, H. (2004). Reproductive partitioning in communally breeding guira cuckoos, Guira guira. Behavioral Ecology and Sociobiology, 55, 213–222. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Magrath, R. D., Johnstone, R. A., & Heinsohn, R. G. (2004). Reproductive Skew. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 157–176. Manning, C. J., Dewsbury, D. A., Wakeland, E. K., & Potts, W. K. (1995). Communal nesting and communal nursing in house mice, Mus musculus domesticus. Animal Behaviour, 50, 741–751.
Understanding variation in reproductive skew McRae, S. B. (1996). Family values: costs and benefits of communal nesting in the moorhen. Animal Behaviour, 52, 225–245. Mennella, J. A., Blumberg, M. S., McClintock, M. K., & Moltz, H. (1990). Interlitter competition and communal nursing among Norway rats: advantages of birth synchrony. Behavioral Ecology and Sociobiology, 27, 183–190. Monnin, T. & Ratnieks, F. L. W. (2001). Policing in queenless ponerine ants. Behavioral Ecology and Sociobiology, 50, 97–108. Mumme, R. L., Koenig, W. D., & Pitelka, F. A. (1983). Reproductive competition in the communal acorn woodpecker: sisters destroy each others eggs. Nature, 306, 583–584. Mumme, R. L., Koenig, W. D., & Pitelka, F. A. (1988). Costs and benefits of joint nesting in the acorn woodpecker. American Naturalist, 131, 654–677. ¨ st, M. & Back, A. (2003). Spatial structure and parental aggression in eider broods. O Animal Behaviour, 66, 1069–1075. ¨ st, M., Mantila, L., & Kilpi, M. (2002). Shared care provides time-budgeting O advantages for female eiders. Animal Behaviour, 64, 223–231. Packer, C., Pusey, A. E., & Eberly, L. E. (2001). Egalitarianism in female African lions. Science, 293, 690–693. Palombit, R. A., Seyfarth, R. M., & Cheney, D. L. (1997). The adaptive value of ‘friendships’ to female baboons: experimental and observational evidence. Animal Behaviour, 54, 599–614. Pereira, M. E. & Weiss, M. L. (1991). Female mate choice, male migration, and the threat of infanticide in ringtailed lemurs. Behavioral Ecology and Sociobiology, 28, 141–152. Pilastro, A. (1992). Communal nesting between breeding females in a free-living population of fat dormouse (Glis glis L). Bollettino di Zoologia, 59, 63–68. Pilastro, A., Missiaglia, E., & Marin, G. (1996). Age-related reproductive success in solitarily and communally nesting female dormice (Glis glis). Journal of Zoology, 239, 601–608. Pusey, A. E. & Packer, C. (1994). Non-offspring nursing in social carnivores: minimizing the costs. Behavioral Ecology, 5, 362–374. Ragsdale, J. E. (1999). Reproductive skew theory extended: the effect of resource inheritance on social organization. Evolutionary Ecology Research, 1, 859–874. Reeve, H. K. & Keller, L. (1997). Reproductive bribing and policing evolutionary mechanisms for the suppression of within-group selfishness. American Naturalist, 150, 42–58. Reeve, H. K. & Keller, L. (2001). Tests of reproductive-skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Ratnieks, F. L. W. (1993). Queen–queen conflicts in polygynous societies: mutual tolerance and reproductive skew. In L. Keller, ed., Queen Number and Sociality in Insects. Oxford: Oxford University Press, pp. 45–85. Russell, A. F. (2004). Mammals: comparisons and contrasts. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 210–227.
465
466
S. J. Hodge Scott, M. P. (1997). Reproductive dominance and differential ovicide in the communally breeding burying beetle Nicrophorus tomentosus. Behavioral Ecology and Sociobiology, 40, 313–320. Sherman, P. W., Lacey, E. A., Reeve, H. K., & Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108. Smith, A. R., Wcislo, W. T., & O’Donnell, S. (2007). Survival and productivity benefits to social nesting in the sweat bee Megalopta genalis. Behavioral Ecology and Sociobiology, 61, 1111–1120. Snowdon, C. T. (1996). Infant care in cooperatively breeding species. Advances in the Study of Behaviour, 25, 643–689. Tibbetts, E. A. & Reeve, H. K. (2003). Benefits of foundress associations in the paper wasp Polistes dominulus: increased productivity and survival, but no assurance of fitness returns. Behavioral Ecology, 14, 510–514. Vehrencamp, S. L. (1977). Relative fecundity and parental effort in the communally nesting anis, Crotophaga sulcirostris. Science, 197, 403–405. Vehrencamp, S. L. (1978). The adaptive significance of communal nesting in groovebilled anis (Crotophaga sulcirostris). Behavioral Ecology and Sociobiology, 4, 1–33. Vehrencamp, S. L. (1979). The roles of individual, kin and group selection in the evolution of sociality. In P. Marler & J. G. Vandenbergh, eds., Handbook of Behavioral Neurobiology: Social Behavior and Communication. New York, NY: Plenum Press, pp. 351–394. Vehrencamp, S. L. (1983). A model for the evolution of despotic versus egalitarian societies. Animal Behaviour, 31, 667–682. Wasser, S. K. & Barash, D. P. (1983). Reproductive suppression among females: implications for biomedicine and sexual selection theory. Quarterly Review of Biology, 58, 513–538. Wolff, J. O. (1994). Reproductive success of solitary and communally nesting whitefooted mice and deer mice. Behavioral Ecology, 5, 206–209. Young, A. J. & Clutton-Brock, T. H. (2006). Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biology Letters, 2, 385–387. Young, A. J., Carlson, A. A., Monfort, S. L., et al. (2006). Stress and the suppression of subordinate reproduction in cooperatively breeding meerkats. Proceedings of the National Academy of Sciences of the USA, 103, 12005–12010.
16
On the evolution of reproductive skew: a genetical view w . e d w i n ha r r i s an d r e i n m a r h a g e r
Summary Prevailing reproductive-skew theory is dominated by a game-theoretic approach to modeling optimal behavior. This approach has led to the development of many specific model variants exploring the evolutionary causes of reproductive partitioning in animal groups as well as motivating the empirical investigation of specific model predictions. Here, we discuss a novel quantitative-genetics view of the evolution of reproductive skew, where individual reproductive success and associated phenotypic traits are influenced by genetic interactions amongst social consorts. We first briefly review empirical evidence of genetic variation and constraints on social phenotypes relating to reproductive skew. Second, we develop a biological argument for viewing reproductive skew as the result of social competition and discuss potential gains of using a genetic framework to investigate social evolution. Finally, we describe general predictions and outline a program of study for investigating reproductive skew in this context. We hope that this fresh perspective on the evolution of reproductive skew offers ample scope for exciting future research complementing existing theory. In particular a genetical approach allows posing and investigating questions that could not be answered using traditional game-theory models, for example about past selection patterns, the evolutionary scope for traits that affect skew, and thus the range of behavioral strategies available to group members. Applying a quantitative-genetics approach to understanding patterns of reproductive skew may ultimately yield additional insights into the evolution of sociality.
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
467
468
W. E. Harris, R. Hager The optimal skew An enduring problem in biology has been to explain the evolutionary transition from the solitary lifestyle to that of animal societies (Maynard-Smith & Szathmary 1995). One approach to understanding the evolution of social groups is to investigate the sharing of reproductive benefits amongst group members. Sometimes reproductive benefits are inequitably partitioned; such reproductive skew is extreme in the social insects, and Darwin noted that these are “cases of special difficulty on the theory of natural selection” (Darwin 1872, p. 404). The key question at issue in understanding the evolution of animal sociality is why individuals remain in a group when they receive a relatively small portion of reproductive benefits (e.g. Nonacs et al. 2006). One possible explanation for this phenomenon is given by Hamilton’s rule, which defines total reproductive success as the sum of direct individual fitness and the fitness of relatives with whom genes are shared, and which was formulated to predict conditions under which cooperative behavior can evolve by kin selection (Hamilton 1964). Over the past several decades, Hamilton’s ideas have provided the theoretical background for research on reproductive skew, which has identified additional sources of variation contributing to reproductive skew such as the availability of resources or the spatial dispersion of suitable habitats (Emlen 1982, Keller & Reeve 1994, Sherman & Lacey 1995, Johnstone 2000, Reeve & Keller 2001, Cant 2006). Ultimately, the empirical phenomenon of reproductive skew is an example of social competition that is thought to have contributed to the evolution of animal sociality (West-Eberhard 1979, 1983, Frank 2006). Current theory explaining reproductive skew consists largely of game-theoretic models and aims to predict the evolutionarily stable level of reproductive partitioning in a group as the result of interacting individuals whose behavioral strategies depend on specific ecological and social parameters such as the relatedness among the breeders, the ability to win aggressive contests, and ecological constraints on independent reproductive success outside of a group (Johnstone 2000, Reeve & Keller 2001). Here, an individual’s behavioral strategy depends on the strategy adopted by others in the population; observed strategy sets are assumed to be the resulting endpoint of evolution at which no alternative strategy can enter the population and spread. In many skew models, the conditions are investigated under which social groups are either stable or prone to disbanding, thus predicting the formation of stable social groups and associated levels of skew (but see Johnstone et al. 1999, Reeve & Emlen 2000). For example, concession models predict that reproductive skew should decrease as relatedness between breeders increases (e.g. Johnstone 2000, Chapter 1 in this volume). Thus, game-theory models of reproductive skew can be used to
Evolution of reproductive skew: a genetical view generate powerful predictions about the phenotypic and genetic outcome of past selection (Reeve & Keller 2001; see also Kokko 2003). An alternative approach to understanding adaptive evolution is that of quantitative genetics, where the focus is on explaining variation in phenotypes by examining the underlying genetic and environmental variation. This view of evolution has a long history in evolutionary biology (Fisher 1930, Wright 1930, Haldane 1932, Dobzhansky 1951), which continues in its application to a wide range of biological problems relevant to reproductive skew such as social interactions amongst group members (e.g. group formation and stability), competitive interactions amongst social group members (e.g. acquisition and control of local resources), parental care (e.g. division of labor), and reproductive trade-offs (e.g. investment into direct versus indirect reproductive success: see Lande & Arnold 1983, Boake 1994, Shuster & Wade 2003, Wolf 2003, Wade 2007). The genetical approach allows direct examination of the evolutionary potential for such phenotypes within a predictive framework (Houle 1992). For example, quantitative genetics tools may be used to directly quantify constraints on phenotypic evolution due to genetic and environmental variation and covariation (for example, a negative genetic correlation amongst fitnessrelated traits can constrain evolutionary change). Complementarily, this approach can yield insight into patterns of past selection and coevolution amongst traits by examining the magnitude and sign of trait correlation. Recent theoretical advances investigating indirect genetic effects have been developed to examine the influence of the social environment itself, which may contain genes, on trait evolution. While the advent of these tools is not new, their application to problems in behavioral ecology is relatively recent, perhaps because of poor accessibility and understanding of the general mathematics underlying the theory of evolutionary genetics (Lande & Arnold 1983, Boake 1994, Boake et al. 2002). Thus, the quantitative genetic approach allows examination of the role of genetic constraint on phenotypic change within or between populations and also allows the inference of past selection on complex phenotypes (Gomulkiewicz & Kirkpatrick 1992, Moore & Boake 1994). Here, we develop an argument for the application of quantitative genetics tools to the study of reproductive skew, to complement existing skew theory. Recent advances in our understanding of how complex genetic effects can impact on traits expressed in the social environment offer a novel perspective on the evolution of such traits. Complex traits may be viewed in a genetic context as the phenotypic outcome of social interactions between individuals; in this respect the use of quantitative genetics is not different from interpretations of current reproductive-skew theory. However, a key difference in the genetic perspective is that the outcome of social interactions may be
469
470
W. E. Harris, R. Hager quantified in terms of the product of the interacting genotypes and the evolutionary potential arising due to the effects of an individual’s genotype on the phenotype of another individual (i.e. an indirect genetic effect – see below). This contrasts with the predominant approach taken by current skew models, where the focus is on the phenotype, ignoring the underlying genotype (e.g. Grafen 1991, Moore & Boake 1994, Gomulkiewicz 1998, Orzack & Sober 2001). Diverse taxa provide evidence of a multivariate basis of fitness arising due to social interactions (e.g. Wolf 2003, Silk 2007). The contribution to variation in behavioral traits may be partitioned into different source components: additive genetic variation, variation caused by traits expressed by social partners that themselves have a genetic basis (i.e. indirect genetic effects, or IGEs: Wolf et al. 1998, Wolf 2003, Muir 2005), non-biotic environmental effects, and the interaction effect amongst genotypes and the environment. For example, a study in Drosophila using different social competitive regimes (i.e. larvae were reared with individuals to which they were related to different degrees ranging from unrelated to full siblings in a competitive environment) demonstrated that IGEs accounted for 18% of phenotypic variance in body size, compared to 34% for direct genetic effects (Wolf 2003). This shows the large influence of IGEs on variation in traits that are affected by social competition. As above, of particular importance for our reasoning here is the consideration of how genes expressed in social consorts can modulate trait expression in a focal individual. The crucial point here with regards to reproductive skew is that the unit of the social group itself has a genetic basis, and can therefore itself evolve (Wolf et al. 1998, 1999). The fitness consequences in systems expressing reproductive skew are those most likely to exhibit social coevolution of relevant traits because of the individual fitness consequences of skew itself (e.g. the indirect genetic effect of a social consort’s genotype on individual fitness). Thus, for example, the expression of helping behavior in a social group might be influenced by the genotype of reproductives, a hypothesis which is easily tested. Further, within families, genes underlying the expression of behaviors leading to reproductive status, for example body size or aggressiveness, advantageous in contests with subordinates (e.g. polistine wasps: Reeve 1991), will have coevolved with genes underlying the expression of helping behaviors (Emlen 1982). In essence, we suggest exploring the multivariate genetic basis of complex phenotypes (Moore et al. 1997, 2002, Wolf et al. 1999, Wolf 2003, Harris et al. 2008) resulting in reproductive skew. Our objective in this chapter is not to present a comprehensive quantitativegenetics model of reproductive skew, but rather to outline recently developed genetic theory of evolution in the social environment and to make suggestions for its application to reproductive skew. We begin by outlining a program of
Evolution of reproductive skew: a genetical view research, with a brief definition of reproductive skew in quantitative-genetics terms. We then argue that the study of reproductive skew can benefit immediately by the application of genetics tools, (1) by examining the evolutionary potential due to the coevolution of traits influencing reproductive skew in social groups (e.g., the genetic covariation between dominant genotype and expression of subordinate phenotype), and (2) by examining the basis of constraint acting on the evolution of reproductive skew, for example by testing the concept of absolute ecological constraint on traits involved in social interactions resulting in reproductive skew. We hope to motivate theoretical and empirical studies using quantitative genetics tools for hypothesis testing, complementing current skew research. Finally, we aim to compare and contrast the questions that may be addressed using traditional game-theory models with those that may be addressed by a genetical approach, with the goal of progressing our understanding of the evolution of reproductive skew in animal societies. A genetical model of reproductive skew Genetics, covariance, and constraint
Behaviors relevant to the evolution of reproductive skew, such as aggressive behaviors associated with dominant, reproductively active individuals (e.g. in polistine wasps, Reeve 1991, Cant et al. 2006) or subordinate behaviors associated with offspring provisioning (e.g. helping to rear the dominant’s offspring: Cant & Field 2001) can be thought of in genetic terms as phenotypes. In this context, reproductive skew may be viewed as the sum of individual fitness consequences due to social interactions in a group. One way to describe the suite of factors contributing to phenotypic variation (morphological or behavioral) in such traits is to statistically partition the effects specifically attributable to genotypic variation and environmental variation respectively. This may be modelled as zi ¼ gi þ ei
ð16:1Þ
where zi is the phenotypic value of some trait i, gi is the degree to which genetic variation contributes to the expression of trait i (i.e. most simply, the additive genetic component, which may be referred to as the direct genetic effect or DGE), and ei is the random environmental component directly contributing to expression (Lynch & Walsh 1998). For example, i might be the level of aggression expressed as a function of genotype and the environment. Thus, the genotype and environment of an individual can be used to predict trait expression. This view of phenotypic evolution is far from new (Fisher 1930), but has been revolutionary as a tool in our understanding of the evolution of
471
472
W. E. Harris, R. Hager quantitative traits, especially as a predictive tool to estimate the evolutionary potential for change in traits under selection (Arnold & Wade 1984a, 1984b). In order to develop our argument as applicable to reproductive skew, here we describe a basic quantitative genetic framework. Equation 16.1 may be extended to describe many individuals; statistically this is represented as a series of estimates of genetic variance for each trait and the covariance of each trait shared with each other trait under consideration (Lynch & Walsh 1998), which is referred to as the genetic variance–covariance matrix, or simply the Gmatrix. This may be represented in matrix form as Dz ¼ Gb
ð16:2Þ
where Dz represents mean generational change in a vector of traits, G is the additive genetic variance–covariance matrix of the traits, and b is a vector of selection gradients influencing the traits (see Arnold & Wade 1984a). Equation 16.2 can then be expanded to Dz1 ¼ G11 b1 þ G12 b2 þ G1j bj
ð16:3Þ
where Dz1 is the mean generational change for trait 1, G11 is the additive genetic variance for trait 1, G12 is the genetic correlation between traits 1 and 2, bj is the selection gradient acting on trait j, and so on. Here, each selection gradient describes the partial regression coefficient of selection acting directly on each trait, holding correlated traits constant (Lande 1979, Lande & Arnold 1983, Arnold & Wade 1984a, 1984b). Thus, the outcome of selection on a trait depends on the net correlated strength of selection across all relevant traits on lifetime fitness. A key point arising from Equation 16.3 is that the evolutionary potential for change of a given trait may depend on constraint imposed by the genetic correlations it shares with other traits. Such correlations may facilitate or constrain trait change, depending on the magnitude and sign of the genetic correlation and the strength of the correlated response to selection (e.g. Gratten et al. 2008). This predictive framework is highly relevant to traits affecting the distribution of reproduction in a group. For example, the outcome of selection on the expression of aggressive behavior (e.g. towards subordinates in the context of breeding attempts or dominance displays) could be constrained by a negative genetic correlation with the expression of parental-care behaviors exhibited by individuals acting in a non-reproductive helper role (e.g. superb fairy-wrens, Malurus cyaneus: Cockburn 1998). In this case, the expression of both aggression and parental-care behaviors may be positively correlated with individual fitness. However, if genes increasing the expression of aggressiveness also tend to decrease the expression of parental care (i.e. a negative genetic
Evolution of reproductive skew: a genetical view correlation between these traits exists), then the outcome of selection on one trait will only be predicted in the multivariate context described above. Many current game-theory models of reproductive skew incorporate environmental constraints by determining the costs and benefits of staying in a group as a helper or leaving a group to pursue individual reproductive opportunities (e.g. Reeve 2000, Reeve & Shen 2006). Environmental effects are typically incorporated into quantitative genetic models of trait expression in that the environment can have direct or indirect effects on the expression of fitness-related traits. One important difference to game-theory models is that the environmental effects may depend on the genotype. In quantitative genetic models, trait expression can be influenced independent of genotype due to the presence of a direct environmental effect, and differentially depending on genotype due to the presence of a genotype-by-environment interaction effect. The genotype-by-environment interaction is given by zi ¼ gi þ ei þ gi ei
ð16:4Þ
where gi * ei is the interaction between genotype and environment that contributes to variation in trait zi. Here, for example, a “poor environment” in some genotypes might effect expression of behaviors and physiology associated with non-reproductive helper status (e.g. mole-rats: O’Riain & Faulkes 2008; see also Emlen 1982). Indirect genetic effects in the social environment
A further refinement of the quantitative-genetics view of evolution is to consider the role of indirect genetic effects (IGEs) in modulating trait variation and evolution in the social environment (Wolf et al. 1998). The environment experienced during the life history of an individual usually consists, at least in part, of interactions with many other individuals such as group members, parents, siblings or predators. By expanding the models above, we can quantify the extent to which the phenotype of a focal individual is influenced by the genotypes of individuals in its environment: zi ¼ gi þ ez0 i
ð16:5Þ
where ez0 i is the IGE on trait value zi as a result of a social interaction. Here, the prime indicates an individual other than the focal individual (e.g. the effect of aggressiveness of a social partner in the environment on aggressiveness in a focal individual). Thus IGEs arising from the genotype of a social competitor contributing to a skew-relevant trait (e.g. aggressiveness) can be measured directly (Figure 16.1). Furthermore, because the environment contributing to individual trait variation has a genetic basis, it can itself evolve and be
473
474
W. E. Harris, R. Hager
Figure 16.1 Sources of variation for a hypothetical phenotypic trait zi. Variation in trait i is affected by genetic variation gi (the direct genetic effect) and ei (the random environmental component directly contributing to trait expression). The indirect genetic effect ez0j caused by z0j (the phenotype of individual j), also contributes to trait variation in the focal phenotype. The indirect genetic effect ez0j in itself is composed of a genetic g0j and an environmental component e0j .
heritable with respect to individual fitness (Wolf et al. 1998, Wolf, 2000, 2003). This paradigm has been used to estimate the effect of selection, for example, in social traits relevant to parent–offspring interactions and maternal care (Cheverud 1984, Wolf & Hager 2006) and competition for resources (Muir 2005, Wolf et al. 2007, Harris et al. 2008). Thus, while the theoretical framework exists to examine the genetic effect for a variety of different kinds of social interactions on fitness, to our knowledge these tools have not yet been used to address problems in social evolution relevant to reproductive skew. This framework has also been expanded to incorporate kin selection theory, e.g. in the context of maternal effects and parental care (Cheverud 1984, Wolf et al. 1998, Wolf 2003): Dz ¼ ðGDD þ GDI ÞbD þ rðGII þ GDI ÞbD
ð16:6Þ
where GDD is the direct effect (DE) of additive genetic variance, GDI is the genetic covariance between DE and IGE, GII is the indirect additive genetic variance, and r is the coefficient of relatedness between the socially interacting individuals. Here, for example, it would be possible to estimate the expression of subordinate behaviors due to genotype of the focal individual and genotype of social partners (e.g. a dominant reproductive) while taking into account shared genes that are common by descent. Thus, analogous to above, it is possible to directly estimate the effect of the genetic covariance between focal genotype and IGE (i.e. GDI) on trait evolution by measuring the degree to which genes in the social environment affect both individuals and their social partners. As in Equation 16.3, the sign of GDI indicates the direction of trait change, Dz, effected by selection on both socially interacting individuals.
Evolution of reproductive skew: a genetical view Conclusion From the equations above a number of predictions follow that are relevant to reproductive-skew theory. First, the degree of evolutionary change possible in a given social trait will depend on potential constraint or facilitation determined by the nature of the genetic correlation (and in fact, the genetic variance–covariance matrix of all relevant traits). For example, the evolution of social aggression could constrain the evolution of cooperative behavior, depending on the magnitude of the (negative) genetic correlation between the traits. In other words, the evolving parameter space (i.e. heritable phenotypes in the quantitative genetic framework) depends on the genetic architecture of the social environment (including absolute and indirect environmental effects). To explore an example from concession models of skew where the dominant “concedes” a share of reproduction to retain the subordinate (Johnstone 2000), the dominant’s ability to do so may be constrained by its aggressiveness (and hence group stability would be predicted to be relatively low). On the other hand, the propensity to disperse could be positively correlated with expressed aggression, and thus an individual that is less aggressive would be more likely to disperse rather than stay in the group, everything else being equal (e.g. Parker 1974). A quantitative genetic analysis might thus yield insight into the limitations of the “parameter space,” i.e. the range of evolutionary responses possible for interacting individuals. Game-theory models generally assume that the whole range of parameter values can be explored and may be part of individual strategies. Therefore, quantitative genetic analysis may not only provide a more accurate picture of the evolution of skew but can also be used to test assumptions and complement the game-theory modeling approach concerning the range of possible strategies, given a specific genetic architecture. Second, partitioning the components of variance contributing to variation in trait expression will enable us to determine the importance of environmental constraints on the evolution of reproductive skew, both absolute constraints on relevant phenotypes and the interaction of environment and specific genotypes. This will shed light on whether ecological constraints play an important role for skew-relevant traits and for within-group aggression, as is often assumed, for example, in concession models (Johnstone 2000). Moreover, some compromise models, such as the tug-of-war models, assume that ecological constraints do not directly affect the subordinate’s share, in contrast to concession models. While these two models’ categories differ in a number of other assumptions, being able to assess the importance of ecological constraints might guide in the choice of model, and also inform future modeling work.
475
476
W. E. Harris, R. Hager Third, trait expression depends on the nature and magnitude of indirect genetic effects in a given social environment. As we argue above, understanding variation in relevant traits is vital to understand the evolution of reproductive skew by allowing the direct estimation of constraints on, for example, variation in the expression of reproductive dominant and subordinate behaviors due to direct and indirect genetic effects and environmental effects. Because the potential for evolutionary change may be constrained or facilitated depending on the magnitude and sign of the genetic covariance between the DGEs and IGEs for such traits correlated to reproductive skew, understanding the evolutionary potential for such traits is critical to understanding their persistence under selection (i.e. stable social groups exhibiting skew). Here, the degree of relatedness between social partners can influence trait expression as well, due to the covariance of shared alleles and the DGE · IGE interaction. Thus the expression of aggressiveness, and its evolutionary potential for change, is predicted to vary depending on the quantitative genetic underpinnings of trait expression in the social environment. The brief conceptual outline of a quantitative-genetics perspective on the evolution of skew presented here offers an alternative view of the causes of reproductive partitioning in animal groups. The estimation of the specific genetic or environmental contribution to trait expression within the framework we discuss here may be made using standard quantitative genetic breeding designs appropriate to the trait and source of variation of interest (Lynch & Walsh 1998) or by the use of pedigree data (e.g. Muir 2005). The reaction-norm approach to trait expression, especially in behavioral ecology, may also be used in an experimental framework (e.g. Schlichting & Pigliucci 1998). Thus, while new studies are required to investigate the ultimate causes of trait expression, quantitative genetics tools hold promise in enabling us to explore the results of past selection and constraints on the evolution of skewrelevant traits in the future. We hope that the introduction we present here is a precursor to the development of specific, genetically explicit models and empirical research investigating the evolution of social traits involved in reproductive skew, in order to gain further insight into the causes of reproductive skew and, ultimately, animal sociality.
Acknowledgments We would like to thank Per Smiseth and Clara B. Jones for helpful comments on a previous draft of this chapter.
Evolution of reproductive skew: a genetical view References Arnold, S. J. & Wade, M. J. (1984a). On the measurement of natural and sexual selection: Applications. Evolution, 38, 720–734. Arnold, S. J. & Wade, M. J. (1984b). On the measurement of natural and sexual selection: Theory. Evolution, 38, 709–719. Boake, C. R. B. (1994). Quantitative Genetic Studies of Behavioral Evolution. Chicago, IL: University of Chicago Press. Boake, C. R. B., Arnold, S. J., Breden, F., et al. (2002). Genetic tools for studying adaptation and the evolution of behavior. American Naturalist, 160, S143–S159. Cant, M. (2006). A tale of two theories: parent–offspring conflict and reproductive skew. Animal Behaviour, 71, 255–263. Cant, M. A. & Field, J. (2001). Helping effort and future fitness in cooperative animal societies. Proceedings of the Royal Society of London B, 268, 1959–1964. Cant, M. A., Llop, J. B., & Field, J. (2006). Individuald variation in social aggression and the probability of inheritance: theory and a field test. American Naturalist, 167, 837–852. Cheverud, J. M. (1984). Evolution by kin selection: a quantitative genetic model illustrated by maternal performance in mice. Evolution, 38, 766–777. Cockburn, A. (1998). Evolution of helping behaviour in cooperatively breeding birds. Annual Review of Ecology and Systematics, 29, 141–177. Darwin, C. (1859). The Origin of Species by Means of Natural Selection. London: Senate. Dobzhansky, T. (1951). Genetics and the Origin of Species. New York, NY: Columbia University Press. Emlen, S. T. (1982). The evolution of helping. I. An ecological constraints model. American Naturalist, 119, 29–39. Fisher, R. A. (1930). The Genetical Theory of Natural Selection. Oxford: Oxford University Press. Frank, S. A. (2006). Social selection. In C. W. Fox & J. B. Wolf, eds., Evolutionary Genetics: Concepts and Case Studies. New York, NY: Oxford University Press, pp. 350–363. Gomulkiewicz, R. (1998). Game theory, optimization, and quantitative genetics. In L. A. Dugatkin & H. K. Reeve, eds., Game Theory and Animal Behavior. Oxford: Oxford University Press, pp. 283–303. Gomulkiewicz, R. & Kirkpatrick, M. (1992). Quantitative genetics and the evolution of reaction norms. Evolution, 46, 390–411. Grafen, A. (1991). Modelling in behavioural ecology. In J. R. Krebs & N. B. Davies, eds., Behavioural Ecology: an Evolutionary Approach. Oxford: Blackwell, pp. 5–31. Gratten, J., Wilson, A., McRae, A., et al. (2008). A localized negative genetic correlation constrains microevolution of coat color in wild sheep. Science, 319, 318–320. Haldane, J. (1932). The Causes of Evolution. London: Longmans, Green. Hamilton, W. D. (1964). The genetical evolution of social behaviour. I. Journal of Theoretical Biology, 7, 1–16.
477
478
W. E. Harris, R. Hager Harris, W. E., McKane, A. J., & Wolf, J. B. (2008). The maintenance of heritable variation through social competition. Evolution, 62, 337–347. Houle, D. (1992). Comparing evolvability and variability of quantitative traits. Genetics, 130, 195–204. Johnstone, R. A. (2000). Models of reproductive skew: a review and synthesis. Ethology, 106, 5–26. Johnstone, R. A., Woodroffe, R., Cant, M., & Wright, J. (1999). Reproductive skew in multimember groups. The American Naturalist, 155, 315–331. Kokko, H. (2003). Are reproductive skew models evolutionarily stable? Proceedings of the Royal Society of London B, 270, 265–270. Keller, L. & Reeve, H. K. (1994). Partitioning of resources in animal societies. Trends in Ecology and Evolution, 9, 98–102. Lande, R. (1979). Quantitative genetic analysis of multivariate evolution, applied to brain: Body size allometry. Evolution, 33, 402–416. Lande, R. & Arnold, S. J. (1983). The measurement of selection on correlated characters. Evolution, 37, 1210–1226. Lynch, M. & Walsh, B. (1998). Genetics and Analysis of Quantitative Traits. Sunderland, MA: Sinauer Associates. Maynard-Smith, J. & Szathmary, E. (1995). The Major Transitions in Evolution. New York, NY: Oxford University Press. Moore, A. J. & Boake, C. R. B. (1994). Optimality and evolutionary genetics: complementary procedures for evolutionary analysis in evolutionary biology. Trends in Ecology and Evolution, 9, 69–72. Moore, A. J., Brodie, E. D. & Wolf, J. B. (1997). Interacting phenotypes and the evolutionary process. I. Direct and indirect genetic effects of social interactions. Evolution, 51, 1352–1362. Moore, A. J., Haynes, K. F., Preziosi, R. F., & Moore, P. J. (2002). The evolution of interacting phenotypes: Genetics and evolution of social dominance. American Naturalist, 160, S186–S197. Muir, W. M. (2005). Incorporation of competitive effects in forest tree or animal breeding programs. Genetics, 170, 1247–1259. Nonacs, P., Liebert, A., & Atarks, P. (2006). Transactional skew and assured fitness return models fail to predict patterns of cooperation in wasps. American Naturalist, 167, 467–480. O’Riain, M. J. & Faulkes, C. G. (2008). African mole-rats: eusociality, relatedness and ecological constraints. In J. Karb & J. Heinze, eds., Ecology of Social Evolution. Berlin: Springer, pp. 207–224. Orzack, S. H. & Sober, E. (2001). Adaptationism and Optimality. Cambridge: Cambridge University Press. Parker, G. A. (1974). Assessment strategy and the evolution of fighting behaviour. Journal of Theoretical Biology, 47, 223–243. Reeve, H. K. (1991). The social biology of Polistes. In K. G. Ross & K. R. Matthews, eds., The Social Biology of Wasps. Ithaca, NY: Cornell University Press, pp. 99–148.
Evolution of reproductive skew: a genetical view Reeve, H. K. (2000). A transactional theory of within-group conflict. American Naturalist, 155, 365–382. Reeve, H. K. & Emlen, S. T. (2000). Reproductive skew and group size: an N-person staying incentive model. Behavioral Ecology, 11, 640–647. Reeve, H. K. & Keller, L. (2001). Tests of reproductive skew models in social insects. Annual Review of Entomology, 46, 347–385. Reeve, H. K. & Shen, S.-F. (2006). A missing model in reproductive skew theory, the bordered tug-of-war. Proceedings of the National Academy of Sciences of the USA, 103, 8430–8434. Schlichting, C. D. & Pigliucci, M. (1998). Phenotypic Evolution: a Reaction Norm Perspective. Sunderland, MA: Sinauer Associates. Sherman, P. & Lacey, A. (1995). The eusociality continuum. Behavioural Ecology, 6, 102–108. Shuster, S. M. & Wade, M. J. (2003). Mating Systems and Strategies. Princeton, NJ: Princeton University Press. Silk, J. (2007). The adaptive value of sociality in mammalian groups. Philosophical Transactions of the Royal Society B, 362, 539–559. Wade, M. J. (2007). The co-evolutionary genetics of ecological communities. Nature Reviews Genetics, 8, 185–195. West-Eberhard, M. J. (1979). Sexual selection, social competition and evolution. Proceedings of the American Philosophical Society, 123, 222–234. West-Eberhard, M. J. (1983). Sexual selection, social competition and speciation. Quarterly Review of Biology, 58, 155–183. Wolf, J. (2003). Genetic architecture and evolutionary constraint when the environment contains genes. Proceedings of the National Academy of Sciences of the USA, 100, 4655–4660. Wolf, J., Brodie, E., & Moore, A. (1999). Interacting phenotypes and the evolutionary process. II. Selection resulting from social interactions. American Naturalist, 153, 254–266. Wolf, J. B. (2000). Indirect genetic effects and gene interactions. In J. B. Wolf, E. D. Brodie, & M. J. Wade, eds., Epistasis and the Evolutionary Process. New York, NY: Oxford University Press, pp. 158–176. Wolf, J. B. & Hager, R. (2006). A maternal–offspring coadaptation theory for the evolution of genomic imprinting. PLoS Biology, 4, 2238–2243. Wolf, J. B., Brodie, E. D., Cheverud, J. M., Moore, A. J., & Wade, M. J. (1998). Evolutionary consequences of indirect genetic effects. Trends in Ecology and Evolution, 13, 64–69. Wolf, J. B., Royle, N. J., & Harris, W. E. (2007). The capture of heritable variation for genetic quality through social competition. Genetica, 134, 89–97. Wright, S. (1930). The genetical theory of evolution: a review. Journal of Heredity, 21, 349–356.
479
17
Social conflict resolution, life history, and the reconstruction of skew bernard j. crespi
Summary Models of reproductive skew have provided useful conceptual frameworks for analyzing social conflict and cooperation in groups of reproductively totipotent individuals, in that they specify explicitly how aspects of ecology and genetic relatedness can generate within-group variation in behavior and reproduction. The main outcome of many years of development and application of these models, however, is a growing consensus that transactional models apply to few if any real situations, and that the models cannot be critically evaluated because assembling sufficient quantitative information to allow critical tests of their assumptions and predictions is not feasible. The “top-down” approach of skew modeling, which makes strong yet unsubstantiated assumptions to extract explanations from data, can be contrasted with a “bottom-up” approach, which involves inference of convergences in sets of diverse social, demographic, and life-history traits across highly diverse taxa, and analyses of fine-scale divergences in small sets of traits between conspecific populations and closely related species. The bottom-up approach explicitly recognizes that each population and taxon exhibits a constellation of more or less similar evolutionarily interrelated traits, especially those that affect (1) the ecological circumstances that underly benefits of dividing labor, (2) the opportunities, costs, and benefits of using force (taking control of behavior away), coercion (cost imposition or repression), or persuasion (providing benefits) to modify the expression of conflicts of interest, and (3) the life-history trade-offs and feedbacks that coevolve with social adaptations. I
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes, ed. Reinmar Hager and Clara B. Jones. Published by Cambridge University Press. ª Cambridge University Press 2009.
480
Social conflict resolution, life history, and the reconstruction of skew illustrate this approach with examples of social convergences among insects, birds, and mammals (including humans, Homo sapiens) that yield insights into selective pressures, and present a model based on such convergences for the role of cooperative breeding in the origin and expansion of modern humans. Finally, I argue that analyses based on wide-scale convergence and fine-scale divergence will ultimately demonstrate the degree to which social animals and social systems fall into a relatively small set of more or less discrete categories with regard to how ecology, relatedness, behavior, and reproduction are inter related – categories that can usefully be subjected to the development of robust, predictive, category-specific models that seek to explain variation in skew. A professor must have a theory as a dog must have fleas. H. L. Mencken
Introduction Grand unified theories are a mainstay of hope for scientific progress. Such theories integrate, subsume, or sometimes vitiate previous ideas, and provide explicit, elegant, and powerful predictive frameworks, as relativity and quantum mechanics have demonstrated in physics. In biology, our theories are more circumscribed, contingent, and specific, with the only true universalities being natural selection itself, and maximization of inclusive fitness. But beyond these fundamental processes, we are drawn to develop and apply theory that is as simple, as general, and as predictive as possible within any given domain – from, for example, sex ratios to mating systems and ecological constraints. More general models are more useful in having wider taxonomic scope, and simpler models can be better in being more intuitive and explicable with regard to causes and effects; by contrast, more specific and realistic models are more directly useful for explaining and predicting in any given application, and more precise models are amenable to more rigorous tests. However, there are strong trade-offs between model characteristics that prevent any given model from optimizing across all criteria: for example, more general models will tend to be less realistic and less precise (Levins 1966, Matthewson & Weisberg 2008). Perhaps most importantly in biology, models that make stronger assumptions can allow us to simplify a problem and make stronger predictions, thus potentially extracting more information from data. For example, maximumparsimony methods assume that phylogenetic change is rare and they thus allow the inference of unique historical sequences – which are, or course, wrong to the extent that change is not rare. Statistical tests based on parsimony may also be relatively powerful compared, for instance, to likelihoodbased tests, but they are also more likely to suffer type 1 error.
481
482
B. J. Crespi Reproductive-skew theory is simple, general, and apparently powerful. Social interactions involving conflict and cooperation among totipotent individuals are stripped by theory to bare bones – relatedness, ecological benefits, dominance, and reproduction – and small sets of assumptions can yield wide ranges of predictions. Skew-model reasoning has been with us for about 30 years, and models more or less specific to given situations and taxa have proliferated for about 15 years. Where and how has this approach been useful, and what are the alternatives for addressing the same questions? I will evaluate the usefulness of skew models not just in terms of fitting predictions to data, but more generally in the context of how we think about social cooperation and conflict, and how we then study it. In this chapter I first consider the complexity of our problem and the dimensionality of the questions that are being addressed, and how suites of social and life-history traits evolve. “Skew,” or variation in reproduction within groups, is only one trait, and only one way of characterizing social systems (Crespi & Yanega 1995, Sherman et al. 1995, Crespi 2005). How can we maximize the conceptual power of this term? How useful is a skew perspective and focus, compared to other viewpoints? Second, I suggest that we divide and conquer the analytic complexities of social cooperation and conflict from the animals up. I thus contrast the topdown modeling framework of skew theory with bottom-up, comparative approaches that are designed to ultimately build robust foundations for integrative and predictive social theory. This book provides a wealth of examples that help to illustrate the convergences among taxa that yield strong insights into selective pressures and trajectories, and components for future models. I will focus in particular on the coevolution of human life history and reproduction, as exemplified in components of the chapters by Jones (Chapter 4) and Cant et al. (Chapter 2) in this volume. I also attempt to draw some general inferences regarding what I see as a core question in social evolution: specification of what factors determine how different forms of conflict are resolved, or remain ongoing.
A complex problem
For every complex problem there is an answer that is clear, simple, and wrong. H. L. Mencken Models per se cannot be wrong, being “if–then” statements set in the rigorous language of mathematics. But the if components can be wrong in being inapplicable or inaccurate for particular situations, and the then components
Social conflict resolution, life history, and the reconstruction of skew can be problematic to evaluate if the predictions are not sufficiently precise and specific to the model under evaluation, compared to alternatives that are more or less specified. Strong-inference tests (this volume, Chapter 10, Taborsky) require risky falsifiability, which can seldom be achieved in evolution and ecology without comprehensive knowledge concerning a large number of variables, many of which are challenging to measure (Magrath & Heinsohn 2000). One example of limitations on the specificity of predictions is provided by Koenig et al. (this volume, Chapter 9) who demonstrate positive correlations of skew with relatedness, ecological constraints, and grouping benefits among species and populations of birds. Such patterns are generally consistent with detailed predictions of skew models involving concessions, but they are also generally consistent with simple predictions from inclusive-fitness theory alone (such that individuals are unlikely to give up reproduction to nonrelatives) and ecological constraints theory (see also Clutton-Brock 2006). As Koenig et al. in Chapter 9 describe, a key component of the models is which behaviors, such as concessions, transactions, coercion, and direct aggression, actually modulate the reproductive success of group members. Does any dominant animal willingly give up reproduction to another one, in direct return for help? Or do relatedness, ecological benefits of group living and cooperation, and variation among populations and species in opportunity or ability to coerce or force other individuals to cooperate ultimately drive the patterns that are observed? Whatever the ultimate explanation, the results are encouraging because they show clear signals of convergence between diverse avian taxa (in sets of covarying traits), and divergence among populations, that are generally consistent with inclusive-fitness theory and ecologicalconstraints theory. Skew theory developed from the trinity of perceived explanations for social cooperation that dominated this field in the 1970s and 1980s: mutualism (Lin & Michener 1972), parental manipulation (Alexander 1974), and genetic relatedness (Hamilton 1964). These have each come to be seen not as alternatives, but as key factors that must be integrated, in the context of social and sexual conflicts within cooperative systems or societies (Trivers 1972, 1974, Trivers & Hare 1976, Haig and Graham 1991), and in the context of specific mechanisms whereby cooperation evolves without being eroded by cheating and conflicts (Lehmann & Keller 2006, Bergmu¨ller et al. 2007, West et al. 2007). The degree to which conflicts are resolved or ongoing, and the preconditions and selective forces underlying resolution or overtness of conflict, have become central issues with regard to the development of models and the analysis of behavior (Ratnieks et al. 2006) – to such an extent that they may
483
484
B. J. Crespi underly the categorization of social systems and their coevolved patterns of behavior and reproduction. But our understanding of the nature and functions of apparent conflict is rudimentary at best – for example, Hart & Ratnieks (2005) point out that aggression, a core behavior in social groups, may represent: (1) (2) (3) (4)
physical domination, which directly suppresses the reproduction of subordinates and pre-empts effective challenges; a response by a dominant individual to more or less obvious testing of the dominant by one or more subordinates; a signal of vigor and vitality of the dominant, indicating that resistance is physically futile; or, I might add, a signal of health and reproductive ability, in the context of remaining an efficient egg-layer who provides sufficient inclusive fitness returns to subordinates that they are not selected to challenge.
Most generally, where the interests of individuals conflict, we expect selection for one party to win, with the outcome of such selection mediated by relative power and resource-holding potential (Parker 1974, Beekman et al. 2003), leverage (Lewis 2002), and other asymmetries, inclusive fitness benefits, costs of competition, pleiotropic effects (Foster et al. 2004), and other factors, most notably the alternative options of individuals who lose. Taken together, these factors should determine whether conflict is ongoing and overt or resolved. Under potential or ongoing conflict, one individual can influence another in three ways (Brown et al. 1997): (1) (2)
(3)
taking control of their behavior away (“force,” such as killing, eviction, egg eating, infanticide, or direct physical alterations); imposing costs to alter their behavior (“coercion,” such as punishment, threats of punishment, or aggression via queen policing) (Clutton-Brock & Parker 1995, Crespi & Ragsdale 2001, Ratnieks & Wenseleers 2008); or providing benefits to alter their behavior (“persuasion,” such as reproductive concessions, reciprocally altruistic behavior, or other forms of mutualistic benefit) (e.g. Bergmu¨ller et al. 2007, West et al. 2007).
The degree to which force, coercion, or persuasion mechanisms operate depends upon the phenotypes, life histories, and other idiosyncracies of the species under consideration, but selection should, all else being equal, favor force or threat of force, as it is most effective in generating control, and most efficient in mimimizing costs to self or benefits to others. Force or threat of force may be found most commonly in strongly asymmetric interactions, such
Social conflict resolution, life history, and the reconstruction of skew as between parents and offspring or between individuals who are otherwise old versus young, and its effectiveness should notably favor the evolution of self-suppression of reproduction, and helping, as in marmosets (Callithrix jacchus: Abbott et al., Chapter 12) and workers in eusocial forms. Symmetric interactions may more commonly involve coercive interactions, with asymmetries in fighting ability (“resource-holding potential”) and the relative value of winning versus losing being the prime determinants of outcome, as in classical models of animal contests (Parker 1974), but with inclusive-fitness considerations across colony or group duration as key additional factors. Ratnieks & Wenseleer (2008) describe how some forms of collective force or coercion, such as worker policing, may also be more powerful than kin selection in driving the evolution of altruistic self-sacrifice in social insects. Finally, the degree to which imposition or provision of costs versus benefits results in a more or less stable or cooperative group must be some function of the ecological costs and benefits of group living, cooperation, and division of labor into reproductives versus helpers. For some taxa, such as bumblebees and vespine wasps, stability and mechanisms of control shift during a season, such that as the colony’s end nears, workers depose the queen or engage in anarchic reproduction (e.g. Bourke 1994); these species may provide insights into the apparent evolutionary shift from physical domination by the queen, in smallcolony eusocial species, to apparent chemical signaling of queen vigor, in largecolony forms (Alexander et al. 1991, Keller & Nonacs 1993, Bourke 1994). To the extent that opportunity and ability of dominants to repress the reproduction of other individuals, via aggression or any other means, varies idiosyncratically among social groups involving relatives (especially where options such as leaving are severely circumscribed), skew becomes more or less inaccurately predictable from any model that does not include accurate and specific assumptions. Relative fighting ability, which mediates skew in tugof-war models (Nonacs 2007), is only one of the factors that modulate the degree of attempted and successful repression observed in any given case; Crespi & Ragsdale (2000) describe a skew-based model predicated on various forms of cost that dominants impose on subordinates, to manipulate the costs and benefits of the subordinate’s options in their favor. Opportunities for such coercion, and its inclusive fitness benefits, should vary widely between taxonomic groups that differ in their habitats, life histories, and phenotypic asymmetries among individuals, which may yield comparative associations that are broadly consistent with inclusive-fitness theory and ecologicalconstraints theory, but not consistent with a model that does not take account of the specific factors affecting how conflicts are pre-empted, resolved, or ongoing.
485
486
B. J. Crespi Patterns of conflict and resolution are also situated in the larger context of such key factors as life-history trade-offs, inheritance, incest avoidance, lifespans, and effects of group size (Koenig & Dickinson 2004), any of which may critically affect trajectories of social evolution. Considering all of these issues together, we presumably need a family or hierarchy of models that make different assumptions regarding control of behavior and reproduction, conflicts, and other key variables for any given taxon or group. At one extreme, we have models that are so general that their predictions are nearly untestable because too much is assumed or neglected; several authors have recently put optimal-skew models in this category (Magrath & Heinsohn 2000, Kokko 2003, Nonacs 2007, Chapter 15 in this volume). At the other extreme we have detailed models tailored to specific species or sets of very similar taxa (e.g. Stephens et al. 2005), which are more testable but yield few insights into the social world at large. Where, in between, is an optimum for progress in using models to understand social cooperation? Models of optimal skew are based on two of the three possible mechanisms above: reproductive concessions, which are a form of persuasion, or mutual attempted coercion, via impositions of costs in tugs-of-war. To the extent that factors other than contest ability determine variation in reproduction, and where leaving the group or independent reproduction represent poor options due to extremely low success (e.g. Clouse 2001), skew is mediated by factors that are not currently encompassed in an optimal-skew framework, but may, in theory, be predictable from knowing mechanisms of control, relatedness, and ecological costs and benefits, in the larger context of life history.
Taxonomic divide and conquer: ecology, phenotypes, life histories, and trade-offs Models represent a “top-down” approach, whereby pre-existing hypotheses are formalized. Such hypotheses, of course, are originally developed from comparative observations of nature that yield apparent convergences, whereby similarities across diverse, independent lineages yield evidence of adaptation. Convergences are “bottom-up”, from taxa to hypothesis, and models and novel data should undergo cycles of reciprocal, illuminating interaction, as clearly illustrated by progress in the study of cooperatively breeding birds and mammals (Koenig & Dickinson 2004, Clutton-Brock 2006). Skew models have motivated the collection of particular forms of data, designed to assess their predictions. Other forms of data and other questions have, as a result, been neglected to some degree. In this section I discuss the use of comparative and phylogenetic methods in developing and testing ideas
Social conflict resolution, life history, and the reconstruction of skew salient to explaining variation among populations and species in social behavior, social system, and skew. My main goals are to discover which variables, currently missing from skew models and approaches, would be most useful to further incorporate, to assess how to infer more accurate assumptions regarding control of behavior, and to develop novel perspectives on evaluating the roles of ecology, relatedness, and control in social evolution. Convergence and divergence
Analyses of convergence appear to be ingrained in the human patternrecognizing, rationalizing psyche: we observe a wide variety of taxa, note that some appear quite similar for some traits even though they are taxonomically far removed, and devise hypotheses to explain why. Formal, statistical, comparative tests of social evolution have been surprisingly rare, but the ones that exist have been disproportionately important and influential (Faulkes et al. 1997, Arnold & Owens 1998) in that they serve to isolate a small number of specific selective factors that apply across entire large clades. More usually, informal tests involve comparisons of taxa that are so far removed as to make the application of independent contrasts impossible, such as between mole-rats and ants (Hart & Ratnieks 2005), birds and wasps (Brockmann 1997), or aphids, thrips, termites, and shrimp (Crespi 2007). Such analyses are predicated on the wild improbability of sets of social variables covarying in exactly the same way in independently evolved groups that differ profoundly in other aspects of their evolutionary histories; for example, redcockaded woodpeckers (Picoides borealis) and Austroplatypus beetles both burrow into living trees to create a costly, extremely valuable breeding habitat, and both represent the apparent sole example of high-skew societies in their respective clades (Kent & Simpson 1992, Ligon & Burt 2004). Similarly, male parental care, and joint female nesting or oviposition, appear to have evolved convergently in joint-nesting birds and some arthropods (Tallamy 2001, Vehrencamp & Quinn 2004). Data compiled by Koenig et al. (this volume, Table 9.3), in conjunction with recent work by Fanelli et al. (2008) allow such an analysis for skew-theory models. Fanelli et al. (2008) noted that, in three studies of primitively eusocial wasps, aggression from subordinates is negatively correlated with reproductive skew, such that high levels of aggression are associated with low skew, and the reverse. For the species they studied, Fanelli et al. (2008) also showed that neither a tug-of-war nor a concession model could explain the observed patterns of association among sets of variables. Data from Koenig et al. (this volume, Table 9.3) similarly show that lower skew is apparently associated with overt competition: among species of cooperatively breeding birds, four (57%) of
487
488
B. J. Crespi seven species without obvious competition exhibit moderate or high skew, compared to only one (14%) of seven species with obvious competition that did so (p ¼ 0.09 by Fisher’s exact test, two-tailed). This comparison is only suggestive, but it does imply that coercion and aggression may unidirectionally influence levels of skew, across diverse taxa, in ways that are not encompassed by existing models (Fenelli et al. 2008). Indeed, to the extent that the nature of the mechanisms and selective pressures mediating control over reproduction determine levels of skew, current skew models are assuming what should instead be considered as a key outcome of any analyses, as also discussed by Hodge (this volume, Chapter 15). Comparative analyses based on statistical or broadly inferential methods provide broad-scale results, usually for effects of one or two variables of interest. Such studies are therefore severely limited in that social traits are embedded in complex causal networks involving effects on sociality from competitors, predators, parasites, and other ecological factors, all interacting with mating systems, life-history traits, and trade-offs (Crespi 2007) – in such networks, the pull of selection on any single trait ramifies throughout the entire evolving system, with diverse effects. There are two main ways to deal analytically with such multidimensional variation within species. First, among-population studies within species, as compiled for birds by Koenig et al. (this volume, Chapter 9), allow inference of which sets of social and other traits covary between populations, and whether the inferred causes of among-population social differences are the same or similar between different species. Such studies benefit greatly from the fact that much variation is “held equal” within species, or among very closely related species (e.g. Doerr & Doerr 2006), although the directionality of coincident changes can be difficult or impossible to infer. Directionality can, however, be inferred from a second method, whereby a considerable number of traits is mapped onto a specieslevel phylogeny, to analyze sequences of fine-scale divergent changes in social characters and their putative causes and consequences. This “divergence” approach is illustrated in the analysis by Chapman & Crespi (2008) of social evolution in Australian gall thrips, where a species-level phylogeny, coupled with data on skew, demography, behavior, and ecology for social and related species, allowed inference of gall size and efficacy of defensive behavior as key determinants of the degree to which soldiers reproduce, in a system without overt dominance. More generally, fine-scale analyses of transitions between social systems allow inference of which transitions are actually observed, compared to which are theoretically possible, and inference of how specific traits, such as monogamy or particular values of relatedness, may potentiate
Social conflict resolution, life history, and the reconstruction of skew transitions between systems, such as high relatedness at the inferred origin of soldiers in Acacia gall thrips (Chapman et al. 2000), or ecological shifts at the origins or losses of cooperative breeding in birds (Ligon & Burt 2004). Such changes provide information about which variables may be necessary versus sufficient for certain changes, such as high ecological constraints being necessary but not necessarily variable between related species with and without cooperative breeding (Doerr et al. 2007). Among arthropods, transitions can be inferred from taxonomic and phylogenetic data for shifts between maternal care and communal care, between maternal care and cooperative breeding (groups with totipotent helpers and breeders), and, unidirectionally, from cooperative breeding to eusociality – with no transitions inferred at all between communal breeders (groups where all individuals help and attempt to breed) and cooperative breeding (Crespi 1996, 2007). To the extent that this pattern also appears to apply in groups of vertebrates, such as clades of birds with cooperative (singular) breeding versus plural breeding (Ligon & Burt 2004, Vehrencamp & Quinn 2004), societies with low skew may be fundamentally different from those with moderate or high skew, presumably due to differences in the selective pressures related to the establishment or costs and benefits of dominance. This idea can be evaluated further by the eventual expansion of Tables 9.3 and 9.4 in this volume (Koenig et al.), and via fine-scale phylogenetic studies of specific vertebrate groups. The four main methodological scales for the analysis of skew and social evolution are organized in Table 17.1. The complementarity of their strengths and weaknesses should help to motivate studies of the same taxa at multiple scales, for the same questions. As regards analyzing causes of skew, withinspecies studies will clearly be most effective at isolating the selective pressures and other factors underlying dominance and control of reproduction via persuasion, coercion, or force where interests conflict, but among-species studies will be required to evaluate the generality of predictions based on any given conception of who controls reproduction, how, and to what extent. This methodological exercise also underscores the tension between analyses of social systems based on the static, stable states predicted by models, and analyses of the dynamics of change between systems such as cooperative breeding and eusociality, or large-colony forms and small-colony forms (Bourke 1999, 2007, Crespi 2004, Dickinson & Hatchwell 2004), which may involve self-reinforcing positive feedbacks. Might such dynamics also mediate major transitions within social systems, such as between alternative-state highand low-skew societies in cooperative breeders, where reproduction is controlled by different means? More generally, the evolution of new mechanisms
489
490
B. J. Crespi Table 17.1 Methods for the analysis of reproductive skew and social evolution at different taxonomic and phylogenetic scales exhibit complementary strengths and weaknesses with regard to their feasibility, inferential power, and generality Within
Among
Fine-scale
Broad-scale
populations,
populations,
phylogenetic
analysis of
Approach
within species
within species
divergence
convergence
Description
Collection of data Collection of
Collection of data Collection of data
on behavior,
same data on
from complete
from large
skews, and
behavior, skew,
sets of closely
numbers of
other traits in
and other traits
related species,
species in
single
from two or
in phylogenetic
phylogenetic
populations
more
context
context
populations Strengths
Allows for
Allows for
Allows for
Allows for robust
experimental
inference of
inference of
statistical
tests and
how some
trajectories of
inference of
explicit tests of
social traits
change in sets
convergences;
model
covary, with
of traits, which
high generality
assumptions
other traits
can be used to
of results; high
and predictions;
held more or
infer causation;
feasibility of
high feasibility;
less constant;
moderate;
data collection
can measure
high feasibility;
generality of
if few traits are
many traits
if such
results
included
Low generality of Low to moderate Moderate
Difficult to
variation exists can measure many traits Weaknesses
results
generality of
feasibility; can
include many
results, hard to
measure only
traits for all
infer cause and
low to
species; large,
effect without
moderate
robust
data on
numbers of
phylogeny
directionality,
traits for each
needed
or experiments
species
for repression of competition between otherwise cooperative entities, such as fair meiosis, queen or worker policing, or coercively imposed monogamy in humans (Homo sapiens), represents a primary cause for the evolution of cooperation, on a par with relatedness as a proximate cause of cooperation across all major groups of organisms (Frank 2003).
Social conflict resolution, life history, and the reconstruction of skew Life histories and trade-offs
Analyses of divergence and convergence, designed to ultimately yield results that underpin the next generation of models, require as their most basic step the selection of variables to measure, beyond the most obvious ones such as relatedness, size, dominance, aggression, and individual reproductive success. One fundamental suite of variables, which has been considered in some models of social evolution but has yet to be comprehensively considered, is life-history traits, and most importantly the trade-offs that they may entail. Social behavior, like all behavior, evolves embedded in the context of life history, with schedules of reproduction, mortality, and, here, helping, driving variation in reproductive success subject to the trade-offs that vary among taxonomic groups. A central role for trade-offs between helping and mortality, and consequent effects on opportunities for inheritance and long tenure as a reproductive, may, for example, help to explain some broad patterns in the social evolution of three main groups of animals, as illustrated in Figure 17.1. In each of these groups, life histories and social behavior evolve in the context of basic necessary resources – the nest, territory, hive, burrow, or gall that serves as a nursery and nexus for cooperative resource exploitation (Alexander et al. 1991). For factory-fortress species, such as some thrips, aphids, termites, and shrimp, the basic necessary resource is quite special in providing combined, defensible food, shelter and nursery for a lifetime or more. Here, we observe high skew, and variation in skew among species with reproductive totipotency, but there is also a general lack of apparent dominance and aggression over reproduction, perhaps due in part to the high relatedness and common monogamy imposed by claustral habitats, coupled with extreme ecological benefits of cooperation in such a special habitat, and low or weak trade-offs between helping and current or future reproduction. The distinction between cooperative breeders and eusocial forms may thus be mediated extrinsically, by life history, especially by habitat duration in relation to individual lifespan. Cooperation in factory-fortress species is notable in that it appears to be highly predictable from a small set of necessary and sufficient ecological and phenotypic conditions (Crespi 1994, Queller & Strassmann 1998), perhaps because conflicts are reduced when relatedness and ecological benefits of cooperation are sufficiently high. In Hymenoptera and other social animals that forage outside of a nest, skew evolves in the context of strong trade-offs between helping and survival, which may drive transitions from cooperative breeding to eusociality if colonies are
491
492
B. J. Crespi
Figure 17.1 Animals with cooperative breeding or eusociality can be categorized into three main modes, differing in the major selective forces and evolutionary responses that underly transitions between social systems. Birds and mammals are highly heterogeneous, but only one evolutionary scenario is provided here, for simplicity.
large and live sufficiently longer than do individuals (Alexander et al. 1991). By contrast, in small-colony forms with reproductive totipotency among some set of individuals, the evolution of competition among potential reproductives may be mediated by life-historical considerations such as chances for inheritance (Ragsdale 1999), which are some function of colony lifespan in relation to individual lifespan, given options of helping or “waiting” (e.g. Tsuji & Tsuji 2005), and the individual and colony-level costs of supercedure, which decline towards the end of the season as colonies of some bumblebee and vespine species descend into selfish and matricidal semi-chaos (Crespi 1992, Bourke 1994). Here, monopolization of reproduction, and prediction of skew, can apparently become strong functions of life history coupled with relatedness. Birds and mammals differ from arthropods for a suite of phenotypic and ecological traits salient to the evolution of cooperative breeding, most notably long lifespans, large size relative to the scale of the habitat, high costs of reproduction for females, inability to greatly improve or expand nest sites to enhance the benefits of dividing labor, and the lack of structures such as the
Social conflict resolution, life history, and the reconstruction of skew sting that can favor heroic, high-benefit nepotism (Alexander et al. 1991). Aside from these differences, vertebrate cooperative breeders encompass a huge ecological and behavioral range, for which unitary explanations of skew and cooperation are unrealistic – like invertebrates, their causal linkages between ecological, behavioral, and reproductive variation presumably fall into some set of selective clusters that are more or less discrete. At the largest comparative scale, however, vertebrates are apparently subject to relatively weak trade-offs between helping and reproducing – indeed, in some species, and perhaps especially in males (Cockburn 1998) helping can increase personal reproduction over the long term. Inheritance of reproductive and dominance status is likely a much stronger selective force in vertebrates – as for the smallcolony invertebrates that most resemble them (e.g. Sumner et al. 2002, Tsuji & Tsuji 2005), and mutualistic benefits of larger group size itself, in the contexts of territoriality defense and predation risk, are also of comparatively strong import in birds and mammals (Clutton-Brock 2002, 2006). All of these factors, and reduced benefits from the division of reproductive labor, should tend to work against strong trade-offs between helping and reproduction, and favor retention of totipotency and lower skew than in invertebrates (Alexander et al. 1991). By contrast, in vertebrates, trade-offs of maintenance with reproduction, and current with future reproduction, may remain strong in cooperatively breeding forms, whereas in many invertebrates these trade-offs are reduced in strength either by the special nature of the habitat, or by the benefits of dividing labor. Transitions between social systems may commonly involve major shifts in the shapes of life-history trade-off curves, such as stronger trade-offs between work and survival or reproduction in incipient hymenopteran workers, weaker trade-offs between reproduction and survival in incipient queens, and, at least potentially, lower costs of reproduction in female reproductives under avian or mammalian cooperative breeding. Consideration of the three main modes of cooperatively breeding animals, in the context of explaining skew from phenotypes, ecology, and behavior, suggests that a key factor missing from most skew models in particular, most models of communal breeding, cooperative breeding, and eusociality in general, and most comparative analyses of social evolution, is the structure of life-history trade-offs. These structures may be especially useful in that they integrate myriad ecological and demographic selective pressures, and coevolve closely with the costs and benefits that ensue from helping versus reproducing. The presence, strengths, and forms of such trade-offs, in the context of colony and individual lifespans, have been considered before by some authors (e.g. Queller 1994a, 1996, Hardling & Kokko 2003, Tsuji & Tsuji 2005, Young et al. 2005, Field & Cant 2007), but families of models that comprehensively
493
494
B. J. Crespi join life-history considerations with social-evolution theory are in their infancy (e.g. Bourke 2007).
Cooperative breeding in the origin and evolution of modern humans Unusual taxa can provide unusual and powerful comparative insights, as seen, for example, by analyses of naked mole-rats (Heterocephalus glaber) from the perspective of insect sociality (Sherman et al. 1991). Hypotheses for the selective pressures that drove the evolution of perhaps the most unusual vertebrate of all, modern humans, have focused predominantly on genes, brains, and tools and ratcheting cultural change (e.g. Kaplan & Robson 2002, Wolpert 2003, Bradley 2008). Such scenarios have largely neglected some key aspects of evolutionary ecology, notably breeding systems and life history, that affect the demographic bases of the large-scale population expansions that were more or less coincident with the evolution of modern human behavior (Mellars 2006, Templeton 2007). Humans can be considered as cooperative breeders, albeit strange ones (Foster & Ratnieks 2005). We share a set of basic traits with some other primates (e.g. Sellen 2007), such as extensive parental care, as well as alloparental help from other individuals, mainly female kin, but we are also more or less unique for a small set of characteristics: (1) the presence of menopause, grandmothering, and a long lifespan, (2) high costs of infant production and feeding, due largely to the extreme fatness of human babies and their big brains, (3) despite such costs, the “stacking” of children, such that a given reproductive female and her helpers care for multiple highly dependent young via short inter-birth intervals compared to other apes, (4) a long duration of childhood, with alloparental care also provided by older, less dependent, pre-reproductive offspring, mainly females (Kuzawa 1998, Kaplan 2000, Hawkes 2003, Gurven & Walker 2006, Sellen 2007, Quinlan & Quinlan 2008, Robson & Wood 2008). How might these traits be related, in the context of selective pressures mediating the evolution of cooperative breeding? What insights might humans provide into cooperative breeding and reproductive skew more generally? Summers (2005) comprehensively applied reproductive-skew theory and its components to the evolution of despotism versus egalitarianism in the mating systems of human males, over three main historical periods: (1) largely egalitarian hunter–gathering, (2) the rise of resource-concentrating agriculture, despotic dominance hierarchies, and higher skew, and (3) modern societies where reproductive options are partially curtailed by the socially imposed repression of law. His analysis shows that historically, high male skew is broadly associated with coercion via punishment in conjunction with the
Social conflict resolution, life history, and the reconstruction of skew ability of some males, with their kin and subordinate allies, to control the most basic of necessary resources, food supplies, labor, and coercive physical power, and thus gain cultural and social dominance. By contrast, in modern societies with socially imposed, albeit serial, monogamy, male skew may be determined more by persuasion of females with the benefits of material reproductive resources, as control of behavior has generally shifted towards female interests. The evolution of skew and life history in human females appears to be an older story than for males, and a more important one for major transitions in human evolution (Figure 17.2). The simplest concatenation of the four traits listed above has selection for large brains at the fulcrum, and selection for life-history and behavioral mechanisms that concentrate female reproduction into an almost insectan queenlike specialized period of about 20 years between a long childhood and grandmaternal nepotism (Pavard et al. 2007), with clear physical signals of endocrine-mediated reproductive potential for individual females (Jasienka et al. 2004). Cant & Johnstone (2008) and Cant et al. (this volume, Chapter 2) provide evidence from theory, primatology, and anthropology that the evolution of menopausal “self-suppression” of reproduction was mediated by a combination of life-history trade-offs increasing in strength with age (Hawkes 2003, Shanley et al. 2007) in combination with conflicts of interest between young immigrant females and older-generation resident females that are resolved in the young female’s interests, due to her insensitivity to the costs imposed on the older female by her breeding, and her relatively high reproductive value. An additional key conflict of interest, not considered in the Cant and Johnstone (2008) model, arises between a mother and her son upon the death of a mother’s mate (Figure 17.3), which is likely to occur when a female nears the critical age for a decision to either allocate alloparental care to a son’s children, or have additional children of her own. The son would benefit greatly, in terms of family resources such as food and labor, expectations for inheritance, and higher relatedness to beneficiaries, by alloparental care from the mother, but the mother may commonly benefit from reproduction with a new mate. Emlen (1995) describes the common presence of severe conflict in this situation among non-human vertebrates, and in humans, the strength and resolution of conflicts is likely to depend upon a variety of cultural and demographic factors, including the strength of life-history trade-offs involving help, survival, and current versus future reproduction. When the original father is still alive, such intra-family discord would be expected to be even more complex though perhaps less dramatic, including a range of predictable conflicts between the father, mother, son, and his mate over reproduction by the parental versus filial generations, and over the relative extents of alloparental and parental care engaged in by all parties.
495
496
B. J. Crespi Longer lifespan
Population expansion; increased competition between and within groups, increased cooperation within groups
Shorter interbirth intervals, higher r
Increased acquisition and transfer of skills
Selection for increased alloparental care by grandmothers, males, offspring Reduced juvenile and adult mortality selection for delayed senescence
Figure 17.2 An increase in the prevalence and strength of alloparental care may have driven some of the major changes concomitant to the origin of modern humans. One result of this suite of changes has been the concentration of female fertility into a relatively short, “queenlike” interval of concentrated reproduction. Under this model, population expansion and increased levels of competition and cooperation are postulated to coevolve with the entire suite of behavioral and demographic traits shown on the right. This model is consistent with hypotheses for selective forces underlying the origin of modern humans developed by Alexander (1989), Kaplan (2000), and Flinn et al. (2005), yet it also posits a central role for increased alloparental care, under low female reproductive skew, allowing female humans to reduce the strength of life-history trade-offs that formerly constrained high reproductive output under conditions of high investment in each offspring.
Under a model considering effects of alloparental care in the evolution of modern humans, long childhoods may be beneficial both to offspring themselves, in terms of cognitive preparation for the social complexities of adulthood (e.g. Flinn & Ward 2005, Flinn et al. 2007) and physiological preparation for the rigors of breeding in females (Ellison 2003), and to mothers, via help that they receive from older children (Hrdy 2005a, 2005b). Mothers are expected to be in a strong position to coerce help via manipulation of daughters (Alexander 1974), with mother–daughter conflict in this context increasing as the daughter matures. Regardless of the details of the mechanisms underlying human life-history shifts, the outcome of reproductive stacking of costly offspring by mothers, potentiated by combined help from husbands, daughters, and grandmothers (Sear & Maceb 2008) was apparently a demographic breakthrough, raising the human intrinsic rate of increase far beyond that of even much smaller-brained
Social conflict resolution, life history, and the reconstruction of skew
Figure 17.3 Higher male than female mortality rates, and the tendency for males to reproduce with females younger than themselves, create conditions where the mate of a female relatively commonly dies when she has more or less grown offspring yet remains capable of further child-bearing. These circumstances are expected to generate conflict between the female (in a rectangle) and her son (circled) over whether the female (a) forgoes personal reproduction and engages in alloparental care of the son’s children (dashed line), or (b) pairs with another male and produces more of her own children, who will be her son’s maternal half-sibs. Resolution or persistence of this conflict should be mediated by the ability of each party to persuade, coerce, or force the other, and by the forms of the life-history trade-offs that determine the relative costs and benefits of alternative behaviors related to helping, maintenance, and current versus future reproduction. Under scenario (a), older females gain a form of “assured fitness return”, as in some eusocial Hymenoptera (Queller 1989, Gadagkar 1990), in that high-benefit investment opportunities are available in the absence of personal reproduction, and parental investment in relatives will continue after their death. To the degree that sons are philopatric (Cant & Johnstone 2008) and can commonly “win” in such conflicts, the selective forces represented by this scenario should, like relatedness asymmetries under common family resources (Cant & Johnstone 2008, Chapter 2 in this volume), favor the evolution of menopause and long lifespan along the lineage leading to modern humans.
large primates (Hrdy 2005a, 2005b). The rest, as they say, may have been history – due in large part to the evolution of confluences of interest among different parties in the successful reproduction of the 20–40-year-old queenlike females in their family group, and a generally egalitarian solution to the social conflicts underlying skew. Low skew in human females, despite the clear potential benefits of helping, was likely also driven by the high energetic costs of human reproduction (Cant & Johnstone 1999) (such that females benefit relatively more from the reproduction of relatives), group augmentation
497
498
B. J. Crespi effects (Kokko et al. 2001) (whereby extended family groups benefit from larger size, at least to some degree: Quinlan & Flinn 2005), and the separation of the timing of periods for helping and breeding, across the life history. As in birds (Cockburn 1998), female humans may, under this view, help more for inclusive fitness benefits of producing young, while males help more in the context of maximizing personal reproduction, under more or less strong trade-offs between parental effort and mating effort (Strassmann 1981); this hypothesis is also consistent with evidence for a lack of benefits to grandchildren due to the local presence of grandfathers (Lahdenpera et al. 2007). Are humans uniquely unique (Alexander 1990) as regards the evolution of their cooperative breeding system? Apparently not, in at least two regards. First, by the “supersaturation” model for population-level effects of the evolution of kin-based cooperative polygamy in birds (bell miner, Manorina melanophrys: Dickinson & Hatchwell 2004), an increase in helping should lead to higher group productivity, increased carrying capacity, larger intrinsic rate of increase, larger benefits of inheritance and coalition-forming, and stronger within-group and between-group competition; in turn, such strong competition may lead to strong group-level selection processes (Wright 2007) and nested hierarchies of social organization, as found, for example, in bell miners (Manorina melanophrys: Dickinson & Hatchwell 2004) as well as humans. These considerations suggest that cooperative breeding and stacking of offspring in humans represent a key innovation that, coevolving with large brains, high parental investment, and strong between-group competition, generated a positive-feedback loop driving human population increases as well as cognitive capacities (Crespi 2004, Flinn et al. 2005). Once established, such a process involves a strong component of selection among groups, which in turn selects for within-group cooperation and against within-group competition (Lahti & Weinstein 2005). In humans, as in some carnivore and insectivore cooperative breeders (e.g. Courchamp et al. 1999), it is the social group itself, and not just the territory or habitat, that represents the most basic core resource mediating survival and reproduction. But most importantly, in these groups humans can cooperate to compete, such that population expansions encompass not just the colonization of initially human-vacant territory, but also forms of warfare at ever-increasing scales of organization (Alexander 1989, 1990). Second, the only other primates with a clear-cut alloparental breeding system, callitrichids (marmosets and tamarins), show a number of notable convergences with humans, such as relatively high costs of child rearing (due here to obligate twinning), the relatively-common presence of infanticide by mothers of their own infants under conditions when little alloparental help is available (Hrdy 1999, p. 180), and “honest” physiological self-suppression of
Social conflict resolution, life history, and the reconstruction of skew females that help (Abbott et al. Chapter 12 in this volume). Marmosets are also known as rare primate examples of potentially fast-breeding “colonizer” species – suggesting that, as in humans, helping has evolved in the context of strong demographic benefits to the family and social group. Did humans colonize the globe due to a combination of ecological dominance with the demographic feedbacks that follow from enhanced cooperative breeding? An important difference between marmosets and humans, of course, is that marmoset helpers tend to be young (older siblings of the helped offspring) yet potentially reproductive. By contrast, evidence of menopause in other animals is restricted to a few long-lived taxa, such as some whales with matrilineal kin groups (Foote 2008); here, the selective pressures involved have yet to be analyzed in enough detail for strong comparative inferences to be drawn, but benefits from accumulated knowledge and information – neural capital, in addition to more direct reproductive benefits, may, as in humans (Kaplan 2000, Gurven et al. 2006), also underlie selection for post-reproductive life. How can this scenario for the role of cooperative breeding in the evolution of modern humans be evaluated? Data from molecular phylogenetics and fossils dates the large-scale expansion of modern humans out of Africa to about 60 000–50 000 years ago (Mellars 2006, Fagundes et al. 2007), which is tens of thousands of years after the origin of anatomical modernity but generally coincident with paleontological evidence for a major shift about 50 000 years ago towards a higher proportion of older individuals among human fossils (Caspari & Lee 2004). Mellars (2006) associates the population expansion with roughly concomitant improvements in tool technology, and with evidence for symbolic thought as exemplified by bodily adornments. Certainly a better, more reliable food supply would have been a prerequisite for producing more babies faster, although more alloparental help in provisioning, and the evolution of a life history more like the highly productive, super-organismal social insects, may also have provided economic benefits even in the absence of tools that allowed the hunting and gathering of substantially more food. More generally and directly, these ideas can be evaluated via comparative studies of traditional human societies, testing predictions of theory for the evolution of cooperative breeding that were developed for less unusual creatures. Conclusions The essence of science is that it is always willing to abandon a given idea, however fundamental it may seem to be, for a better one. H. L. Mencken Skew models were one of the most innovative and insightful developments in the study of cooperative breeding, in that they explicitly integrated all of the
499
500
B. J. Crespi core determinants of Hamilton’s inclusive-fitness equation in the context of individuals forming social groups under varying degrees of personal reproduction. I would argue however, as have others (Magrath & Heinsohn 2000, Kokko 2003, Nonacs 2006, 2007, Hodge Chapter 15 in this volume), that their simple assumptions of control are unjustified and their predictions are too general, in most cases, to be of much practical use for the planning or interpretation of empirical studies. The study of social evolution is extremely difficult compared to other enterprises in behavior, ecology, and evolution, in that quantitative understanding requires measuring effects of alternative behaviors on lifetime personal and inclusive fitness, in networks of multiple – normally more than two – kin and non-kin that differ in age, power, and other phenotypic traits. Students of social cooperation have spawned extremely detailed terminology systems to classify behavior (e.g. Bergmu¨ller et al. 2007), and constellations of more and less general models (e.g. Nonacs 2007), but only for a few species, such as meerkats (Suricata suricatta: e.g. Stephens et al. 2005), have long-term field studies accumulated enough information for robust interpretations and predictions of social-behavioral interactions in their lifehistorical contexts. These are too few to compare, and we are left with a vast mosaic of partial information on hundreds of diverse animal taxa to which we can apply classical and statistical comparative methods. The main problems then become what variables to measure, and what specific questions to address, in animals at what taxonomic scales for the most effective progress. If the last 15 years of skew models have told us little else of general significance, they have made clear that the nature and extent of control over reproduction in cooperative social groups must be specified before the selective pressures underlying a social system can be well understood. Such analyses are set in the wider framework of how conflicts of varying forms evolve, be they parent versus offspring, maternal gene versus paternal gene, male versus female, queen versus worker, worker versus worker, breeder versus breeder, or breeder versus helper, and these diverse forms of conflict should evolve under common general rules (Queller 1994b, Cant 2006). The degree to which conflict resolution or persistence is predictable across taxa, compared, for example, to the predictability of ecological effects, remains to be determined, but repression of competition by force or coercion has proven wide-scale predictive ability (Frank 2003), as does simple dominance based on resource-holding potential (Parker 1974). Skew modeling has also made clear the central importance of life history and demography in social evolution, which can now be extended to incorporate various patterns of life-history trade-offs that may characterize convergent social patterns across broad or narrow taxonomic groups. Once
Social conflict resolution, life history, and the reconstruction of skew mechanisms and processes of social control are better understood, and the lifehistorical architecture of social-reproductive interactions and systems has been sketched out, skew models may again become an approach of choice to structure the search for causes of social convergence and divergence. We can thus seek to develop unified theories of social evolution as grand and general as nature happens to have made them, yet no more. Acknowledgments I am very grateful to Reinmar Hager and Clara B. Jones for inviting me to write this article, to Kyle Summers for helpful comments and insights, and to NSERC for financial support. References Alexander, R. D. (1974). The evolution of social behavior. Annual Review of Ecology and Systematics, 5, 325–383. Alexander, R. D. (1989). Evolution of the human psyche. In P. Mellars & C. Stringer, eds., The Human Revolution: Behavioral and Biological Perspectives on the Origins of Modern Humans. Edinburgh, Edinburgh University Press, pp. 455–513. Alexander, R. D. (1990). How did humans evolve? Reflections on the uniquely unique species. University of Michigan Museum of Zoology Special Publication, 1, 1–38. Alexander, R. D., Noonan, K., & Crespi, B. J. (1991). The evolution of eusociality. In P. W. Sherman, J. Jarvis, & R. D. Alexander, eds., The Biology of the Naked Mole Rat. Princeton, NJ: Princeton University Press, pp. 3–44. Arnold, K. E. & Owens, I.P.F. (1998). Cooperative breeding in birds: a comparative test of the life history hypothesis. Proceedings of the Royal Society of London B, 265, 739–745. Beekman, M., Komdeur, J., & Ratnieks, F. L. W. (2003). Reproductive conflicts in social animals: who has power? Trends in Ecology and Evolution, 18, 277–282. Bergmu¨ller, R., Johnstone, R. A., Russell, A. F., & Bshary, R. (2007). Integrating cooperative breeding into theoretical concepts of cooperation. Behavioral Processes, 76, 61–72. Bourke, A. F. G. (1994). Worker matricide in social bees and wasps. Journal of Theoretical Biology, 167, 283–292. Bourke, A. F. G. (1999). Colony size, social complexity and reproductive conflict in social insects. Journal of Evolutionary Biology, 12, 245–257. Bourke, A. F. G. (2007). Kin selection and the evolutionary theory of aging. Annual Review of Ecology, Evolution and Systematics, 38, 103–128. Bradley, B. J. (2008). Reconstructing phylogenies and phenotypes: a molecular view of human evolution. Journal of Anatomy, 212, 337–353. Brockmann, H. J. (1997). Cooperative breeding in wasps and vertebrates: the role of ecological constraints. In J. Choe & B. Crespi, eds., Evolution of Social Behavior in Insects and Arachnids. Ithaca, NY: Cornell University Press, pp. 348–371.
501
502
B. J. Crespi Brown, W. D., Crespi, B. J., and Choe, J. C. (1997). Sexual conflict and the evolution of mating systems. In J. Choe & B. Crespi, eds., The Evolution of Mating Systems in Insects and Arachnids. Cambridge: Cambridge University Press, pp. 352–377. Cant, M. A. (2006). A tale of two theories: parent–offspring conflict and reproductive skew. Animal Behaviour, 71, 255–263. Cant, M. A. & Johnstone, R. A. (1999). Costly young and reproductive skew in animal societies. Behavioral Ecology, 10, 178–184. Cant, M. A. & Johnstone, R. A. (2008). Reproductive conflict and the separation of reproductive generations in humans. Proceedings of the National Academy of Sciences of the USA, 105, 5332–5336. Caspari, R. & Lee, S. H. (2004). Older age becomes common late in human evolution. Proceedings of the National Academy of Sciences of the USA, 101, 10895–10900. Chapman, T., Crespi, B. J., Schwarz, M., & Kranz, B. (2000). High relatedness and inbreeding at the origin of eusociality in Australian gall thrips. Proceedings of the National Academy of Sciences of the USA, 97, 1648–1650. Chapman, T. W. & Crespi, B. J. (2008). The evolutionary ecology of eusociality in Australian gall thrips: a “model clades” approach. In J. Korb & J. Heinze, eds., Ecology of Social Evolution. Princeton, NJ: Princeton University Press, pp. 57–83. Clouse, R. (2001). Some effects of group size on the output of beginning nests of Mischocyttarus mexicanus (Hymenoptera: Vespidae). Florida Entomologist, 84, 418– 425. Clutton-Brock, T. (2002). Breeding together: kin selection and mutualism in cooperative vertebrates. Science, 296, 69–72. Clutton-Brock, T. H. (2006). Cooperative breeding in mammals. In P. M. Kappeler & C. P. van Schaik, eds., Cooperation in Primates and Humans: Mechanisms and Evolution. Berlin: Springer, pp. 173–190. Clutton-Brock, T. H. & Parker, G. A. (1995). Punishment in animal societies. Nature, 373, 209–216. Cockburn, A. (1998). Evolution of helping behavior in cooperatively breeding birds. Annual Review of Ecology and Systematics, 29, 141–177. Courchamp, F., Clutton-Brock, T. H., & Grenfell, B. (1999). Inverse density dependence and the Allee effect. Trends in Ecology and Evolution, 14, 405–410. Crespi, B. J. (1992). Cannibalism and trophic eggs in subsocial and eusocial insects. In M. Elgar & B. J. Crespi, eds., Cannibalism: Ecology and Evolution Among Diverse Taxa. Oxford: Oxford University Press, pp. 176–213. Crespi, B. J. (1994). Three conditions for the evolution of eusociality: are they sufficient? Insectes Sociaux, 41, 395–400. Crespi, B. J. (1996). Comparative analysis of the origins and losses of eusociality: causal mosaics and historical uniqueness. In E. Martins, ed., Phylogenies and the Comparative Method in Animal Behavior. Oxford: Oxford University Press, pp. 253–287. Crespi, B. J. (2004). Vicious circles: positive feedback in major evolutionary and ecological transitions. Trends in Ecology and Evolution, 19, 627–633.
Social conflict resolution, life history, and the reconstruction of skew Crespi, B. J. (2005). Social sophistry: logos and mythos in the forms of cooperation. Annales Zoologici Fennici, 42, 569–572. Crespi, B. J. (2007). Comparative evolutionary ecology of social and sexual systems: water-breathing insects come of age. In E. Duffy & M. Thiel, eds., Evolutionary Ecology of Crustacea. Oxford: Oxford University Press, pp. 442–460. Crespi, B. J. & Ragsdale, J. E. (2000). A skew model for the evolution of sociality via manipulation: why it is better to be feared than loved. Proceedings of the Royal Society of London B, 267, 821–828. Crespi, B. J., Morris, D. C., & Mound, L. A. (2004). Evolution of Ecological and Behavioural Diversity: Australian Acacia Thrips as Model Organisms. Canberra: Australian Biological Resources Study & CSIRO Entomology. Crespi, B. J. & Yanega, D. (1995). The definition of eusociality. Behavioral Ecology, 6, 109–115. Dickinson, J. L. & Hatchwell, B. J. (2004). Fitness consequences of helping. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 48–66. Doerr, E. D. & Doerr, V. A. J. (2006). Comparative demography of treecreepers: evaluating hypotheses for the evolution and maintenance of cooperative breeding. Animal Behaviour, 72, 147–159. Doerr, E. D., Doerr, V. A. J. & Safran, R. J. (2007). Integrating delayed dispersal into broader concepts of social group formation. Behavioural Processes, 76, 114–117. Ellison, P. T. (2003). Energetics and reproductive effort. American Journal of Human Biology, 15, 342–351. Emlen, S. T. (1995). An evolutionary theory of the family. Proceedings of the National Academy of Sciences of the USA, 92, 8092–8099. Fagundes, N. J., Ray, N., Beaumont, M., et al. (2007). Statistical evaluation of alternative models of human evolution. Proceedings of the National Academy of Sciences of the USA, 104, 17614–17619. Fanelli, D., Boomsma, J. J., & Turillazzi, S. (2008). Subordinate wasps are more aggressive in colonies with low reproductive skew. Animal Behaviour, 75, 879–886. Faulkes, C. G., Bennett, N. C., Bruford, M. W., et al. (1997). Ecological constraints drive social evolution in the African mole-rats. Proceedings of the Royal Society of London B, 264, 1619–1627. Field, J. & Cant, M. A. (2007). Direct fitness, reciprocity and helping: a perspective from primitively eusocial wasps. Behavioural Processes, 76, 160–162. Flinn, M. V., Geary, D. C. & Ward, C. V. (2005). Ecological dominance, social competition, and coalitionary arms races: why humans evolved extraordinary intelligence. Evolution and Human Behavior, 26, 10–46. Flinn, M. V. & Ward, C. V. (2005). Evolution of the social child. In B. Ellis & D. Bjorklund, eds., Origins of the Social Mind: Evolutionary Psychology and Child Development. London, Guilford Press, pp. 19–44. Flinn, M. V., Quinlan, R. L., Ward, C. V., & Coe, M. K. (2007). Evolution of the human family: cooperative males, long social childhoods, smart mothers, and extended
503
504
B. J. Crespi kin networks. In C. Salmon & T. Shackelford, eds., Family Relationships. Oxford: Oxford University Press, pp. 16–38. Foote, A. D. (2008). Mortality rate acceleration and post-reproductive lifespan in matrilineal whale species. Biology Letters, 4, 189–191. Foster, K. R., Shaulsky, G., Strassmann, J. E., Queller, D. C., & Thompson, C. R. (2004). Pleiotropy as a mechanism to stabilize cooperation. Nature, 431, 693–696. Foster, K. R. & Ratnieks, F. L. W. (2005). A new eusocial vertebrate? Trends in Ecology and Evolution, 20, 363–364. Frank, S. A. (2003). Perspective: repression of competition and the evolution of cooperation. Evolution, 57, 693–705. Gadagkar, R. (1990). Evolution of eusociality: the advantage of assured fitness returns. Philosophical Transactions of the Royal Society of London B, 329, 17–25. Gurven, M. & Walker, R. (2006). Energetic demand of multiple dependents and the evolution of slow human growth. Proceedings of the Royal Society of London B, 273, 835–841. Gurven, M., Kaplan, H., & Gutierrez, M. (2006). How long does it take to become a proficient hunter? Implications for the evolution of extended development and long life span. Journal of Human Evolution, 51, 454–470. Haig, D. & Graham, C. (1991). Genomic imprinting and the strange case of the insulin-like growth factor-II receptor. Cell, 64, 1045–1046. Hamilton, W. D. (1964). The genetical evolution of social behaviour. Journal of Theoretical Biology, 7, 1–52. Hardling, R. & Kokko, H. (2003). Life-history traits as causes or consequences of social behaviour: why do cooperative breeders lay small clutches? Evolutionary Ecology Research, 5, 691–700. Hart, A. G. & Ratnieks, F. L. W. (2005). Crossing the taxonomic divide: conflict and its resolution in societies of totipotent individuals. Journal of Evolutionary Biology, 13, 383–395. Hawkes, K. (2003). Grandmothers and the evolution of human longevity. American Journal of Human Biology, 15, 380–400. Hrdy, S. (1999). Mother Nature: a History of Mothers, Infants, and Natural Selection. New York, NY: Pantheon. Hrdy, S. B. (2005a). Evolutionary context of human development: the cooperative breeding model. In C. S. Carter, L. Ahnert, K. E. Grossmann, et al., eds., Attachment and Bonding: a New Synthesis. Report of the 92nd Dahlem Workshop. Cambridge, MA: MIT Press, pp. 9–32. Hrdy, S. B. (2005b). Comes the child before the man: how cooperative breeding and prolonged postweaning dependence shaped human potential. In B. S. Hewlett and M. E. Lamb, eds., Hunter–Gatherer Childhoods. New Brunswick, NJ: Transaction, pp. 65–91. Jasienska, G., Ziomkiewicz, A., Ellison, P. T., Lipson, S. F., & Thune, I. (2004). Large breasts and narrow waists indicate high reproductive potential in women. Proceedings of the Royal Society of London B, 271, 1213–1217.
Social conflict resolution, life history, and the reconstruction of skew Kaplan, H. (2000). A theory of human life history evolution: Diet, intelligence, and longevity. Evolutionary Anthropology, 9, 156–185. Kaplan, H. S. & Robson, A. J. (2002). The emergence of humans: the coevolution of intelligence and longevity with intergenerational transfers. Proceedings of the National Academy of Sciences of the USA, 99, 10221–10226. Keller, L. & Nonacs, P. (1993). The role of queen pheromones in social insects: queen control or queen signal? Animal Behaviour, 45, 787–794. Kent, D. S. & Simpson, J. A. (1992). Eusociality in the beetle Australoplatypus incompertus (Coleoptera: Platypodidae). Naturwissenschaften, 79, 86–87. Koenig, W. D. & Dickinson, J. L. eds., (2004). Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press. Kokko, H., Johnstone, R. A., & Clutton-Brock, T. H. (2001). The evolution of cooperative breeding through group augmentation. Proceedings of the Royal Society of London B, 268, 187–196. Kokko, H. (2003). Are reproductive skew models evolutionarily stable? Proceedings of the Royal Society of London B, 270, 265–270. Kuzawa, C. W. (1998). Adipose tissue in human infancy and childhood: an evolutionary perspective. American Journal of Physical Anthropology, Suppl 27, 177–209. Lahdenpera, M., Lummaa, V., Helle, S., Tremblay, M., & Russell, A. F. (2004). Fitness benefits of prolonged post-reproductive lifespan in women. Nature, 428, 178–181. Lahdenpera, M., Russell, A. F., & Lummaa, V. (2007). Selection for long lifespan in men: benefits of grandfathering? Proceedings of the Royal Society of London B, 274, 2437–2444. Lahti, D. C. & Weinstein, B. S. (2005). The better angels of our nature: group stability and the evolution of moral tension. Evolution and Human Behavior, 26, 47–63. Lehmann, L. & Keller, L. (2006). The evolution of cooperation and altruism: a general framework and a classification of models. Journal of Evolutionary Biology, 19, 1365–1376. Levins, R. (1966). The strategy of model building in population biology. In E. Sober, ed., Conceptual Issues in Evolutionary Biology. Cambridge, MA: MIT Press, pp. 18–27. Lewis, R. J. (2002). Beyond dominance: the importance of leverage. The Quarterly Review of Biology, 77, 149–164. Ligon, J. D. & Burt, D. B. (2004). Evolutionary origins. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 5–34. Lin, N. & Michener, C. D. (1972). Evolution of sociality in insects. Quarterly Review of Biology, 47, 131–159. Magrath, R. D. & Heinsohn, R. G. (2000). Reproductive skew in birds: models, problems and prospects. Journal of Avian Biology, 31, 247–258. Mellars, P. (2006). Why did modern human populations disperse from Africa ca. 60,000 years ago? A new model. Proceedings of the National Academy of Sciences of the USA, 103, 9381–9386.
505
506
B. J. Crespi Matthewson, J. & Weisbergy, M. (2008). The structure of tradeoffs in model building. Synthese (published online, July 2008). Nonacs, P. (2006). The rise and fall of transactional skew theory in the model genus Polistes. Annales Zoologici Fennici, 43, 443–455. Nonacs, P. (2007). Tug-of-war has no borders: it is the missing model in reproductive skew theory. Evolution, 61, 1244–1250. Parker, G. A. (1974). Assessment strategy and the evolution of animal conflicts. Journal of Theoretical Biology, 47, 223–243. Pavard, S., Koons, D. N. & Heyer, E. (2007). The influence of maternal care in shaping human survival and fertility. Evolution, 61, 2801–2810. Queller, D. C. (1989). The evolution of eusociality: reproductive head starts of workers. Proceedings of the National Academy of Sciences of the USA, 86, 3224–3226. Queller, D. C. (1994a). Extended parental care and the origin of eusociality. Proceedings of the Royal Society of London B, 256, 105–111. Queller, D. C. (1994b). Male–female conflict and parent–offspring conflict. American Naturalist, 144, S84–S99. Queller, D. C. (1996). The origin and maintenance of eusociality: the advantage of extended parental care. In S. Turillazzi & M. J. West-Eberhard, eds., Natural History and Evolution of Paper Wasps. Oxford: Oxford University Press, pp. 218–234. Queller, D. C. & Strassmann, J. E. (1998). Kin selection and social insects. BioScience, 48, 165–175. Quinlan, R. J. & Flinn, M. V. (2005). Kinship and reproduction in a Caribbean community. Human Nature, 16, 32–57. Quinlan, R. J. & Quinlan, M. B. (2008). Human lactation, pair-bonds, and alloparents; a cross-cultural analysis. Human Nature, 19, 87–102. Ragsdale, J. E. (1999). Reproductive skew theory extended: the effect of resource inheritance on social organization. Evolutionary Ecology Research, 1, 859–874. Ratnieks, F. L. W., Foster, K. R., & Wenseleers, T. (2006). Conflict resolution in insect societies. Annual Review of Entomology, 51, 581–608. Ratnieks, F. L. W. & Wenseleers, T. (2008). Altruism in insect societies and beyond: voluntary or enforced? Trends in Ecology and Evolution, 23, 45–52. Robson, S. L. & Wood, B. (2008). Hominin life history: reconstruction and evolution. Journal of Anatomy, 212, 394–425. Sear, R. & Maceb, R. (2008). Who keeps children alive? A review of the effects of kin on child survival. Evolution and Human Behavior, 29, 1–18. Sellen, D. W. (2007). Evolution of infant and young child feeding: implications for contemporary public health. Annual Review of Nutrition, 27, 123–148. Shanley, D. P., Sear, R., Mace, R., & Kirkwood, T. B. (2007). Testing evolutionary theories of menopause. Proceedings of the Royal Society of London B, 274, 2943–2949. Sherman, P. W., Jarvis, J. U. M., & Alexander, R. D. eds., (1991). The Biology of the Naked Mole-Rat. Princeton, NJ: Princeton University Press.
Social conflict resolution, life history, and the reconstruction of skew Sherman, P. W., Lacey, E. A., Reeve, H. K., & Keller, L. (1995). The eusociality continuum. Behavioral Ecology, 6, 102–108. Stephens, P. A., Russell, A. F., Young, A. J., Sutherland, W. J., & Clutton-Brock, T. H. (2005). Dispersal, eviction, and conflict in meerkats (Suricata suricatta): an evolutionarily stable strategy model. American Naturalist, 165, 120–135. Strassmann, B. I. (1981). Sexual selection, paternal care, and concealed ovulation in humans. Ethology and Sociobiology, 2, 31–40. Summers, K. (2005). The evolutionary ecology of despotism. Evolution and Human Behavior, 26, 106–135. Sumner, S., Casiraghi, M., Foster, W., & Field, J. (2002). High reproductive skew in tropical hover wasps. Proceedings of the Royal Society of London B, 269, 179–186. Tallamy, D. W. (2001). Evolution of exclusive paternal care in arthropods. Annual Review of Entomology, 46, 139–165. Templeton, A. R. (2007). Genetics and recent human evolution. Evolution, 61, 1507– 1519. Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell, ed., Sexual Selection and the Descent of Man. Chicago, IL: Aldine, pp. 1871–1971. Trivers, R. L. (1974). Parent–offspring conflict. American Zoologist, 14, 249–264. Trivers, R. L. & Hare, R. (1976). Haplodiploidy and the evolution of the social insects. Science, 191, 250–263. Tsuji, K. & Tsuji, N. (2005). Why is dominance hierarchy age-related in social insects? The relative longevity hypothesis. Behavioral Ecology and Sociobiology, 58, 517–526. Vehrencamp, S. L. & Quinn, J. S. (2004). Avian joint laying systems. In W. D. Koenig & J. L. Dickinson, eds., Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press, pp. 177–196. West, S. A., Griffin, A. S., & Gardner, A. (2007). Evolutionary explanations for cooperation. Current Biology, 17, R661–R672. Wolpert, L. (2003). Causal belief and the origins of technology. Philosophical Transactions of the Royal Society of London A, 361, 1709–1719. Wright, J. (2007). Cooperation theory meets cooperative breeding: exposing some ugly truths about social prestige, reciprocity and group augmentation. Behavioural Processes, 76, 142–148. Young, A. J., Carlson, A. A., & Clutton-Brock, T. H. (2005). Trade-offs between extraterritorial prospecting and helping in a cooperative mammal. Animal Behaviour, 70, 829–837.
507
Taxonomic index
Acorn woodpecker 237, 239, 242, 245, 247, 250, 449, 452
Brachyteles arachnoides 97 Brown jay 242, 245, 247, 250
African striped mouse 97
Brown lemur 97
Alpine marmot 97
Brown long-eared bat 97
Alouatta seniculus 172
Brown skua 250
Ansell’s mole-rat
Buteo galapagoensis 250
Anseranas semipalmata 201 Antelocarpa americana 97 Aphelocoma
449 Cryptomys anselli hottentotus 55 Cryptomys damarensis 72, 370, 375, 380 Ctenomys sociabilis 449, 453
Callithrix jacchus 136, 337, 340, 353, 388
coerulescens 228
Callitrichinae 338
ultramarina 4
Carmague stallion 97
Cyanocorax morio 242, 245, 247, 250 Daubentonia
Apis mellifera 4
Capuchin monkey 185
Apodemus sylvaticus 97, 449,
Catharacta antarctica 250
Deer mouse 449
Cebus capucinus 185
Degu 449
Cercopithecus aethiops 97
Drosophila melanogaster
452 Arabian babbler 242, 246, 247, 250 Acrocephalus sechellensis 247 Arthropods 489 Austroplatypus sp. 487
Cervus elaphus 54 Cleaner fish 20
Baboon, Hamadryas 44 Savannah 138
madagascarensis 55, 60
470 Dunnock 236, 247, 250
Cheirogaleus medius 60–1 Chimpanzee 32, 42, 66, 97, 170, 175, 178 Common marmoset 136, 337, 340, 353, 388 Common moorhen 241, 242, 247
Elephant 31 Elephant seal 54 Elk 54 Equids 198, 199 Equus caballus 97 Eulemur coronatus 55
Balaenoptera musculus 31
Common vole 449
Eulemur fulvus 55, 97, 157
Barbary macaque 97
Coquerel’s mouse lemur 60
Eulemur macaco 57
Bathyergidae 370
Corcorax melanorhamphos
Eulemur mongoz 55
Bechstein’s bat 97
508
Crotophaga sulcirostris 247,
242, 246, 247, 250
Blue whale 31
Colobus guereza 172
Fat dormouse 449
Bonobo 38, 42, 55, 65–6
Crocuta crocuta 55, 66, 97
Fat-tailed dwarf lemur 60–1
Taxonomic index Fish 267
Jay, Mexican 4
Florida scrub-jay 228
Julidochromus ornatus 270, 284
Galapagos hawk 250 Gallinula chloropus 241, 449
382, 386 Common 55, 377, 380, 389 Damaraland 72, 376, 378,
Killer whale 25, 40
Gallinula mortieri 247 Garnett’s greater bushbaby
naked 55, 70, 376, 378,
381, 387 Moorhen 241, 449
Lemur catta 57, 63–5, 97
Mountain gorilla 97
Lion 54, 68, 97
Moustached tamarind 97
Gentle lemurs 59
Long-tailed manakin 250
Mungos mungo 4, 97, 446,
Glis glis 449
Loxodonta sp. 31
55
Globicephala sp. 25, 40 Gray mouse lemur 55, 61–3 Grey seal 97
449 Mus musculus domesticus
Macaca mulatta 97, 136, 175
449 Myotis bechsteinii 97
Goby, coral dwelling 21
Macaca sylvanus 97
Gorilla 38, 172, 175, 178, 184
Magpie goose 247
Naked mole rat 370
Gorilla beringei beringei 97,
Malurus cyaneus 234
Neolamprologus
136, 138 Gorilla gorilla 38, 172, 175, 178 Groove-billed ani 247, 250, 449 Guira cuckoo 247 Guira guira 247 Haematopus sp. 247 Halichoerus grypus 97 Hanuman langur 139, 172 Hapalemur griseus 55, 59 Helogale parvula 4 Heterocephalus glaber 55, 70, 370, 376, 378, 382, 386 Homo sapiens 24, 25, 27, 34, 36, 38, 43, 97
Mandrill 97, 175, 177
multifasciatus 270, 280,
Mandrillus sphinx 97, 175, 177
282 Neolamprologus pulcher 270,
Marmota marmota 97, 420 Mediterranean ocellated
287 Neolamprologus savoryi 270
wrasse 268 Melanerpes formicivorus
Octodon degus 449
237, 239, 242, 245, 247,
Odocoileus virginiamus 97
250, 449, 452
Orcinus orca 25, 40
Marmota flaviventris 97, 119 Meerkat 4, 97, 418, 426,
Otolemur garnettii 55 Ovis aries 97 Oystercatcher 247
446 Megaptera novaeangliae 97
Pan
Meriones unguiculatus 414
paniscus 38, 42, 55, 65–6
Microcebus murinus 55, 57,
troglodytes 32, 42, 66, 97,
60, 61–3
170, 175, 178
Honey bee 4
Microtus arvalis 449
Paper wasp 312
Horses 196–224
Microtus ochrogaster 448,
Papio
House mouse 449
449
cyocephalus 138
Hover wasp 314
Mirounga angustirostris 54
hamadryas 44
Human 24, 25, 27, 34, 36, 38,
Mirza coquereli 55, 60, 74
Phaner furcifer 55
Mongolian gerbil 414
Panthero leo 54, 97
Mongoose,
Paragobidon xanthosomus
43, 97, 494 Humpback whale 97 Hyaena, spotted 55
dwarf 4
Hymenoptera 491
banded 4, 97, 446, 449 Mole rat,
Indri indri 57
African 370
21 Pelvicachromis pulcher 268, 270, 278 Peromyscus leucopus 449
509
510
Taxonomic index Peromyscus maniculatus 449
Rhabdomys pumilio 97
Philocolobus sp. 44
Rhesus macaques 97, 136,
Picoides borealis 487 Pilot whale 25, 40
138, 174, 175
Plecotus auritus 97 Polistes sp. 312, 317 Prairie vole 448, 449 Presbytis entellus 139 Prunella modularis 236, 247,
246, 247, 250 Turkey 237, 250
Ring-tailed lemur 63–5, 97 Varecia variegata 57
Porphyrio porphyrio 234, 241, 242, 247, 250
Turdoides squamiceps 242,
Saguinus mystax 97
Vervet monkey 97
Semnopithecus entellus 172 Sericornis frontalis 242, 246, 250
West-African cichlid 268 White-browed scrubwren 242, 246, 250
Seychelles warbler 247
White-footed mouse 449
Soay sheep 97
White-tailed deer 97
Pronghorn antelope 97
Spotted hyaena 66, 97
White-winged chough 242,
Propithecus verreauxi 55, 57,
Stenogastrinae 314
250
73, 74 Pukeko 234, 241, 242, 247, 250
Suricata suricatta 4, 97, 418, 426, 446
273, 287
487 Red colobus 44 Reindeer 54
Yellow-bellied marmot 97, 119, 222 Yellow baboon 97
Red-cockaded woodpecker Red howler monkey 172
Woolly spider monkey 97
Superb fairy-wren 234 Symphodus ocellatus 268,
Rangifer tarandus 54
246, 247, 250 Wood mouse 97, 449, 452
Taiwan yuhina 247, 250 Tasmanian native hen 247,
Yuhina brunneiceps 247, 250
250 Tuco tuco 449, 453
Zebra (plains) 196–224
Subject index
Ache people of Paraguay 27, 29 acorn woodpecker (Melanerpes formicivorus) 173, 449, 452 social system 229, 231, 232, 234, 237, 238–9, 239–40 tests of concession skew theory 242–3, 245, 247, 251 African elephant (Loxodonta africana) 31 African mole-rats (Bathyergidae) 369–90 aridity food distribution hypothesis 375–6 cooperative breeding 370
phylogeny 371, 372 proximate factors maintaining high skew 377–84 reproductive skew 72–3 reproductive suppression and skew 380–4 skew among social
collaris) 288 alpine marmot (Marmota marmota) 420 alternative reproductive tactics (ARTs) models, cooperative
skew in lifetime reproductive success 379–80 skew theory and empirical data 385–9 social and reproductive diversity 370–4 ultimate factors leading to cooperative breeding 375–7
kin structure and
alaotrensis) 56, 59–60 alpine accentor (Prunella
377–84
evolutionary relationships 372
(Hapalemur griseus
mole-rat groups
eusocial species 370, 371–2 between genera 371,
Alaotran gentle lemur
see also common mole-rat; Damaraland
reproduction in fish 288–9 Antarctic fur seal (Arctocephalus gazella) 94 Arabian babbler (Turdoides squamiceps) 237, 243, 246, 249, 252 aridity food distribution hypothesis 375–6 Aristotle 27 Asiatic wild ass (Equus hemionus) 198
mole-rat; mashona
Austroplatypus beetles 487
reproductive skew
mole-rat; naked
avian social organization
380–4
mole-rat
limitations of skew models 385 molecular phylogenetics studies 370 origin and adaptive radiation 371, 372
African wild ass (Equus africanus) 198 African wild dog (Lycaon pictus) 352, 360 aggression, interpretations of 484
application of skew theory 227–58 bisexual groups 228–9 bordered tug-of-war model 229, 258 conflict in cooperatively breeding birds 228
511
512
Subject index avian social organization (cont.) cooperative polyandry 229 cooperative polygynandry 229
tests of skew theory in birds 240–56 aye-aye (Daubentonia madagascarensis) 55, 56 Azorean rock-pool blenny
cooperative polygyny 229
(Parablennius
definition of co-breeders
sanguinolentus) 271, 278
231 definition of non-breeding helpers 231 extra-group parentage 234–6
baboons (Papio spp.) 44, 54, banded mongoose (Mungos
group size and sociality
453
potential breeders 230–4 interspecific tests of skew theory (meta-analysis) 253–6 intraspecific tests of concession skew theory 240–53 issues related to measuring reproductive skew 230–40 members of cooperatively breeding groups 228–9 models of reproductive skew 229 non-breeding helpers 228 non-independence of paternity 239–40 null models 239–40 potential versus actual reproductive roles 230–4 relatedness asymmetries in co-breeders 238–9 search for a unified theory of social evolution 230 sexual conflict 236–7
90–2 brown jay (Cyanocorax morio) 173, 234, 239 tests of concession skew theory 243, 245–6, brown skua (Catharacta antarctica) 251
138, 184, 351, 454 mungo) 4, 419, 446, 450,
237–8
adaptive significance
248, 251 B index of skew 117, 141
goal of skew theory 229–30
incest avoidance in
brain enlargement,
bargaining theory 11–13 battleground models 308–9 bee (Exoneura nigrescens) 451 behavioral flexibility, adaptive significance 92 behavioral phenotypes in mammals 92 behavioral roles see yellowbellied marmot bell miner (Manorina melanophrys) 498 birds, cooperative breeding see avian social organization black-and-white colobus (Colobus guereza) 172 black lemur (Eulemur macaco) 56, 57 black rhinoceros (Diceros bicornis) 94 black-tailed prairie dog (Cynomys ludovicianus) 352, 447, 454 blue whale (Balaenoptera musculus) 31 bonobo (Pan paniscus) 38, 55, 57, 65–6, 172 bordered tug-of-war model 229, 258
Callitrichinae 138, 498, 359–60 extreme female reproductive skew 338 see also lion tamarins; marmosets; tamarins chacma baboon (Papio cynocephalus) 454 Cheirogaleus medius (fat-tailed dwarf lemur) 56, 60–1 chimpanzee (Pan troglodytes) 30, 34, 35, 38 fission-fusion societies 171–2 inter-group aggression 170–1 reproductive skew 176, 178–9 cichlids Julidochromis ornatus 270, 272, 274–5, 284–5 Neolamprologus multifasciatus 270, 272, 274–5, 282–3, 285, 288 Neolamprologus pulcher 270, 271, 272, 274–5, 280–2, 285, 287, 288, 290, 451 Pelvicachromis pulcher 268, 271, 274–5, 278–80, 285, 288
Subject index see also cooperative reproduction in fish commitment model 359 common brown lemur (Eulemur fulvus) 55, 73 common eider (Somateria mollissima) 446 common marmoset (Callithrix jacchus) 136, 409, 337–61 causes of female reproductive skew 357–9 comparative aspects of reproductive skew 359–61 dominant females 340–1 extreme reproductive skew in the Callitrichinae 338
post-ovulatory reproductive failure 351 pre-conception reproductive suppression 343–51 proximate regulation of reproductive skew 342–59 resource competition hypothesis for infanticide 352–7 social organization 339–42 suppression of ovulation in subordinates 345–51 common mole-rat (Cryptomys hottentotus hottentotus) 55, 56 kin structure and
free-living groups 339–41
reproductive skew
infanticide as an agent of
380–1
selection 359 infanticide by females 352–7, 358–9 inhibition of sexual behavior in subordinates 343–4 intra-group aggression levels 340–1 laboratory groups 340–1, 341–2 mechanisms of reproductive skew 338–9 models of reproductive skew 338 neural and neuroendocrine suppression of ovulation 349–51 post-conception regulatory mechanisms 351–7, 358–9
maintenance of reproductive skew 377–8
alloparental care 267, 268–70 alternatives to skew theory 285–90 application of skew theory 270–85 approaches to the use of models 290–4 brood-care helpers 269–70 cooperation between bourgeois and satellite males 268 dynamic modeling approach 289–90 enforced cooperation approach 287–8 factors affecting level of reproductive skew 285 forms of male cooperation 267 interaction between theory and empiricism 290–4 joint defense of spawning site or territory 267, 268
phylogeny 371, 372
kin-selection theory 285–6
skew theory and empirical
male skew in
data 389 see also African mole-rats common moorhen (Gallinula chloropus) 241, 242, 247, 409, 449 common vole (Microtus arvalis) 450 compromise models of reproductive skew 8–10 concession models of reproductive skew 5–8, 9–10, 307 cooperative breeding, role in the evolution of modern humans 494–9 cooperative reproduction in fish 265–94
polygynandrous species 278–80 models of alternative reproductive tactics 288–9 polygynandrous breeding groups 268 reciprocity models 286–7 reproductive skew among group members 280–5 reproductive skew among polymorphic males 278–80 reproductive skew between bourgeois and satellite males 273–8
513
514
Subject index cooperative reproduction in fish (cont.) unraveling underlying mechanisms 290–4 variation in reproductive patterns 267 cooperatively breeding birds see avian social organization Coquerel’s mouse lemur
discriminate infanticide model 359 domestic cat (Felis catus) 168 dominant role in compromise models 8, 9–10 in transaction models 7–8, 9–10 dominants active interference in
(Mirza coquereli) 55, 56,
subordinate
60, 74
reproduction 402–3,
coral-dwelling goby (Paragobidon xanthosomus) 21 costly young model 308
405–6, 407, 415–20 incomplete control of reproduction 443, 444, 453–5
crowned lemur (Eulemur
dunnock (Prunella modularis)
coronatus) 55, 56
229, 248, 250, 288 dwarf mongoose (Helogale
Damaraland mole-rat (Cryptomys damarensis) 72, 360, 373, 409, 410
parvula) 4, 290, 358, 360 dynamic modeling approach, cooperative
eusociality 370, 371–2
reproduction in fish
kin structure and
289–90
reproductive skew 381–2 maintenance of reproductive skew 378 phylogeny 371, 372 reproductive suppression 381–2 skew theory and empirical data 387–9 see also African mole-rats Darwin, Charles 468 Daubentonia madagascarensis (aye-aye) 55, 56 deer mouse (Peromyscus maniculatus) 449 degu (Octodon degus) 450 demographic factors and reproductive skew in primates 172–4
focus on underlying causes of skew 455–9 framework for future empirical research 455–9 incomplete control of reproduction by dominants 443, 444, 453–5 need for investigation of underlying causes 441 potential causes of low reproductive skew 442–55 potential confounding factors 440 reduced costs of subordinate reproduction to dominants 443–53 testing predictions of skew models 440–1 endocrine effects, physiological
ecological constraints theory 483 effective-number-ofbreeding-individuals skew index (S3) 141 elephant seal (Mirounga angustirostris) 54 empirical research on reproductive skew 439–61 assumptions underlying skew models 440 difficulties in testing skew models 440 effects of factors other than competition 442–55 evasion and control tactics by subordinates 454
suppression in vertebrates 398–401 endogenous (natural) Stackelberg solution 42 endothermy, adaptive significance 90, 91 energy limitation and reproductive allocation patterns 98–101 enforced cooperation approach, cooperative reproduction in fish 287–8 environmental heterogeneity 84–104 adaptive significance of behavioral flexibility 92 adaptive significance of endothermy 90, 91
Subject index adaptive significance of
reproductive allocation
relative brain
patterns of males and
enlargement 90–2
females 98–101
applicability of reproductive skew models 84–6 behavioral phenotypes in mammals 92 costs and benefits of social behavior 88–90 definition 86 differing male and female reproductive strategies 88 energy limitation and reproductive allocation patterns 98–101 eutherian adaptive complex 87–97 eutherian evolutionary selection pressures 86–7 evolutionary predisposition to cope with 94 female emancipation 89 influence on factors affecting reproductive skew 86 influence on group size and structure 87–8 influence on reproductive skew 93–7 influence on reproductive suppression 101–3 phenotypic hitchhiking 93 polyphenisms (environmentally switched alternative phenotypes) 92 range of reproductive skew
responsive switching in social mammals 91–2 stochastic phenotype switching 91–2 totipotency of individuals 92 Equid societies age and reproductive success in females 202–4 age and reproductive success in males 208–9 factors affecting female reproductive success 202–8 factors affecting male reproductive success 208–13 female behavioral alternatives and reproductive skew 216–17 fission-fusion sociality 198 harem pattern of sociality 198 male behavioral alternatives and reproductive skew 213–16 male routes to adulthood 211–13 management impact on
rank and reproductive success in males 209–11 skew dynamics 217–19, 220–4 social, ecological and demographic features 199–202 social stability and reproductive success in females 206–8 social stability and reproductive success in males 211–13 sociality and reproductive skew 196–224 study groups 198–9 estrous overlap and reproductive skew in primates 172–4 estrous synchrony and reproductive skew 168 Eulemur coronatus (crowned lemur) 55, 56 Eulemur fulvus (common brown lemur) 55, 73 Eulemur fulvus mayottensis (Mayotte brown lemur) 57 Eulemur fulvus rufus (redfronted lemur) 57 Eulemur macaco (black lemur) 56, 57 Eulemur mongoz (Mongoose lemur) 55, 56
inequality and skew
eusociality index 115
219–20
eutherian adaptive complex
patterns of reproductive inequality and skew 202 patterns of sociality 198 rank and reproductive
in eutherian societies
success in females
94–7
204–6
87–97 and reproductive skew 93–7 significance of behavioral flexibility 92 significance of endothermy 90, 91
515
516
Subject index eutherian adaptive complex (cont.) significance of relative brain enlargement 90–2 eutherian societies, range of reproductive skew 94–7 evolution of modern humans, role of cooperative breeding 494–9 evolution of reproductive skew behaviors as phenotypes 471 complex genetic traits 469–70 determining the role of
phenotypic evolution 471–3 predictions from a quantitative genetic approach 475–6 quantitative genetic approach 469–76 quantitative genetic framework 472–4 trait expression and indirect genetic effects 476 evolution of sociality see social evolution evolution of the eutherian evolutionary predisposition to cope with environmental
constraints 475
heterogeneity 94
sociality 468
extended priority-of-access (POA) model 169
evolutionary constraints imposed by trait correlations 472–3 game-theory models 468–9 genetic variancecovariance matrix 472 genetical model of reproductive skew 471–4 Hamilton’s rule 468 indirect genetic effects 470 indirect genetic effects in the social environment 473–4, 476 kin selection theory 468, 474 limits of the range of possible strategies 475 multivariate genetic basis of complex phenotypes 469–70
lemurs 55–65, 70–2, 73–4 reproductive skew among spotted hyenas 66–72, 74 social dominance and reproductive skew 54–5 fish see cooperative reproduction in fish Florida scrub-jay (Aphelocoma coerulescens) 228 fork-marked lemur (Phaner furcifer) 55, 57
mammals 86–7
environmental evolution of animal
reproductive skew among
Gala´pagos hawk (Buteo galapagoensis) 240, 250, 288 game-theory models 307–8 evolution of reproductive skew 468–9 Garnett’s greater bushbaby
fat dormouse (Glis glis) 447, 450 fat-tailed dwarf lemur (Cheirogaleus medius) 56, 60–1 female-biased dispersal, consequences of 37–44 female choices, effects on male reproductive skew 168–9 female-dominated mammalian societies 53–75 applicability of skew models 70–2 definition of female dominance 54–5 predictions from skew models 70–2 reproductive skew among bonobos 65–6
(Otolemur garnettii) 55, 56 genetical model of reproductive skew 471–4 behaviors as phenotypes 471 determining the role of environmental constraints 475 evolutionary constraints imposed by trait correlations 472–3 genetic variancecovariance matrix 472 indirect genetic effects in the social environment 473–4, 476 limits of the range of possible strategies 475 phenotypic evolution 471–3
Subject index predictions related to skew theory 475–6 quantitative genetic framework 472–4 trait expression and indirect genetic effects 476 golden-crowned sifaka (Propithecus tattersalli) 56 golden-headed lion tamarin (Leontopithecus chrysomelas) 414 golden lion tamarin (Leontopithecus rosalia) 358, 409
dispersal costs and reproductive skew 155–6 ecological influences on skew 142–3, 154–5 effects of female choice on male reproductive skew 158 estimating and analysing
environmental heterogeneity honeybee (Apis mellifera) 4 horse (Equus caballus), sociality and reproductive skew 196–224 house mouse (Mus musculus) 94, 447, 449 hover wasps (Stenogastrinae),
levels of skew 142,
nesting biology 314
146–50
see also primitively
factors affecting male reproductive skew 138–59 future mating prospects
eusocial wasps human (Homo sapiens) reproductive skew 25–6 role of cooperative
gorilla (Gorilla gorilla) 34, 38
and reproductive skew
breeding in evolution
“grandmother” hypothesis
153–4
494–9
for menopause evolution 28–30 gray gentle lemur (Hapalemur griseus) 55 Grevy’s zebra (Equus grevyi) 198 grey mouse lemur (Microcebus murinus) 55, 56, 57, 61–3 groove-billed ani (Crotophaga sulcirostris) 248, 250, 447, 449 group size and structure, influence of environmental heterogeneity 87–8
life-history/population influences on skew 143, 144 measuring reproductive skew 140–1 predation risk and skew 143–4 reproductive skew in the Ramnagar population 144–6 results and discussion 144–58 role of infanticide in reproductive skew 151–3
Guira cuckoo (Guira guira) 248
skew predictor variables
Hadza people of Tanzania 27
social organization 139–40
hamadryas baboon (Papio
within-group aggression
147–50
hamadryas) 44 Hamilton’s rule 285–6, 468 Hanuman langur (Semnopithecus entellus) 172 calculating levels of skew 141–2 choice of skew index 140–1
157–8 Hapalemur griseus (gray gentle lemur) 55 Hapalemur griseus alaotrensis
inbreeding avoidance in vertebrates 409 inclusive-fitness theory 483 indri (Indri indri) 56, 57 infanticide 416 discriminate infanticide model 359 infanticide by females as an agent of selection 359 common marmoset 352–7, 358–9 resource competition hypothesis 352–7 insects see primitively eusocial wasps Japanese macaque (Macaca fuscata) 35, 169 Julidochromis ornatus (cichlid) 270, 272, 274–5, 284–5
(Alaotran gentle lemur) 56, 59–60 hetergeneous environmental regimes see
kiang (Equus kiang) 198 killer whale (Orcinua orca) 24, 25, 40, 44
517
518
Subject index kin selection theory 468, 474 cooperative reproduction in fish 285–6 kin structure and reproductive skew, African mole-rats 380–4 kinship dynamics model of menopause evolution 37–44 !Kung people of the Kalahari 27
improved breeding success
Mayotte brown lemur
when subordinates
(Eulemur fulvus
breed 443, 447–8 improved group stability when subordinates breed 443–4, 448, 451 incomplete control by dominants 443, 444, 453–5 indirect fitness benefits of subordinate reproduction 444, 451–2
Lemur catta (ring-tailed lemur) 56, 57, 63–5 lemurs female dominance 55 reproductive skew 55–65, 70–2, 73–4 see also particular species life history traits and tradeoffs as variables in skew models 491–9 lion (Panthera leo) 54, 94, 447
long-term fitness benefits
tamarin (Leontopithecus chrysomelas) 414 golden lion tamarin (Leontopithecus rosalia) 358, 409 long-tailed mannakin
potential causes 442–55 reduced costs of subordinate reproduction to dominants 443–53 relative abundance of resources 443, 445
445–6 effects of factors other than competition 442–55 evasion and control tactics by subordinates 454
dominant 453, 454 female reproductive skew 358, 359–60 inbreeding avoidance 409
of subordinates 418–19, 426 temporal variation in reproductive skew 446 menopause evolution 24–44 adaptive explanations 28–30 consequences of
magpie goose (Anseranas
female-biased
semipalmata) 247
dispersal 37–44
male reproductive skew,
costs of co-breeding 33–7
Hanuman langurs 138–59 mandrill (Mandrillus sphinx) 175, 177–8 mantled howler monkey 89
of dominants 443,
136, 290 control of reproduction by
reproductive suppression
452–3
254
allocation to offspring
meerkat (Suricata suricatta) 4,
infanticide 352
(Alouatta palliata palliata)
differential resource
ocellatus) 268, 269, 273–8, 287, 288
reproduction 444,
(Chiroxiphia linearis) 252, low reproductive skew
wrasse (Symphodus
of subordinate
lion tamarins golden-headed lion
mayottensis) 57 Mediterranean ocellated
marmosets (Callithrix spp.) 94, 485 see also common marmoset marmot (Marmota marmota) 136 see also alpine marmot; yellow-bellied marmot mashona mole-rat (Cryptomys darlingi) 409 see also African mole-rats
date of emergence of menopause 26–7 demography and kin selection across the lifespan 37–40 “grandmother” hypothesis 28–30 kinship dynamics model 37–44 “mother” hypothesis 28–9 phylogenetic inertia argument 30–1 physiological constraints argument 30–1 rapid reproductive senescence 27–8
Subject index rate of follicular attrition 31 relatedness asymmetries across the lifespan 37–40 relatedness asymmetries and conflict resolution 40–2 reproductive competition perspective 33–7 reproductive overlap in humans 34–7 reproductive skew in human societies 25–6 selection pressure argument 32–3 Mexican jay (Aphelocoma ultramarina) 4 Microcebus berthae (pygmy mouse lemur) 57 Microcebus murinus (gray
comparative and phylogenetic approaches 486–94 complexity of modeling social systems 482–6 compromise models 8–10 concession models 5–8, 9–10, 307 control issues 9–10 definition of reproductive skew 4 dominant role in
mouse lemur) 55, 56, 60, 74 models of reproductive skew 3–21 applicability to femaledominated mammals 70–2 application to primate societies 166–7, 169–70 assumptions about reproductive conflict 4–5 bargaining theory 11–13 bottom-up comparative approaches 486–94 classification 307–12
307–8 selection of variables to measure 491–4 synthetic models of skew 11–21, 311 testing predictions in primates 136–59 transactional models 5–8,
transaction models 7–8, 9–10 evaluation of usefulness 499–501 future directions for research 459–60 in heterogeneous
edwardsi) 56
sociality 135–6 restraint models 5–8, 9–10,
9–10 dominant role in
61–3
Mirza coquereli (Coquerel’s
quantifying features of
top-down approach 486
genetical model 471–4
(Propithecus diadema
models 10–11
compromise models 8,
mouse lemur) 55, 56, 57, Milne-Edwards sifaka
problems with synthetic
environmental regimes 84–6 incorporation of outside options 10–21 influence on study of social systems 486–7 interpretations of aggression 484 life history traits and trade-offs as variables 491–9 limitations of current models 482–6 measures of inequality in breeding success 4 outside option principle 11–13
9–10 tug-of-war models 8–9, 10–19, 308, 309–11 types of models and their limitations 481–2 underlying assumptions 307–12 ways that one individual can influence another 484–5 mole-rats see African molerats Mongolian gerbil (Meriones unguiculatus) 414 mongoose lemur (Eulemur mongoz) 55, 56 moorhen (Gallinula chloropus) 241, 242, 247, 409, 449 “mother” hypothesis for menopause evolution 28–9 mountain gorilla (Gorilla beringei) 136, 138, 172, 176, 178 mountain zebra (Equus zebra) 197, 198
patterns of conflict and resolution 483–6 problems with a top-down approach 486
naked mole-rat (Heterocephalus glaber) 4, 55, 56, 89, 92, 94, 136, 373, 409, 494
519
520
Subject index naked mole-rat (cont.) eusociality 370, 371–2 kin structure and reproductive skew
ovulation, suppression in subordinates 345–51 oystercatcher (Haematopus ostralegus) 237, 247
382–4 maintenance of reproductive skew 378–9 phylogeny 371, 372 reproductive suppression 382–4 skew theory and empirical data 386–7 see also African mole-rats Neolamprologus multifasciatus (cichlid) 270, 272, 274–5, 282–3, 285, 288 Neolamprologus pulcher
Pelvicachromis pulcher (cichlid) 268, 271, 274–5, 278–80, 285, 288 Phaner furcifer (fork-marked lemur) 55, 57 phenotypic hitchhiking 93 pheromones, physiological
349–51 446 Nyakyusa people of Tanzania 36
307–8 optimization models 308 orangutan 35 ostrich (Struthio camelus) 446 Otolemur garnettii (Garnett’s greater bushbaby) 55, 56 outside option principle 11–13, 311 outside options incorporation into models
analyses of skew 179–83
active interference by dominants 402–3, 405–6, 407, 415–20 adaptive framework 406 causes 402–6 definition 398 effects of a same-sex dominant 411–14 presence of a dominant 408–11 evidence for endocrine effects 398–401 evolution of physiological restraint 414–15 future directions for research 421–4 implications for reproductive skew 424–7
of reproductive skew
inbreeding avoidance 409
10–21
infanticide 416
situations where they influence skew 16–21
405–15 pied kingfisher (Ceryle rudis) 409 pilot whales (Globicephala spp.) 24, 25, 40, 44 plains zebra (Equus burchelli) 196–224 Polistes wasps 447, 451
effects unrelated to the one-shot sequential models
subordinate restraint 402,
vertebrates 416–17 phylogenetic comparative
vertebrates 397–427
Norway rat (Rattus norvegicus)
socially-induced stress
Pliny 27
274–5, 280–2, 285, 287,
suppression of ovulation
pheromones 416–17
suppression in
physiological suppression in
288, 290, 451
403–4
417–20 peace-incentive models 307
(cichlid) 270, 271, 272,
neural and neuroendocrine
subordinates 398–401,
link to compromised fertility in
nesting biology 312–14 see also primitively eusocial wasps polyphenisms (environmentally switched alternative phenotypes) 92 prairie vole (Microtus ochrogaster) 448, 450 primate societies application of reproductive skew models 166–7, 169–70 case studies of the causes of reproductive skew 174–9 causes of male reproductive skew 167–83 consequences of male reproductive skew 184–6 demographic factors and reproductive skew 172–4 effects of estrous synchrony 168 estrous overlap and reproductive skew 172–4
Subject index extended priority-of-access model 169 female choices and reproductive skew 168–9 phylogenetic comparative analyses of skew 179–83 priority-of-access model 167–9 relatedness among males and skew 173, 174 reproductive skew in chimpanzees 176, 178–9 reproductive skew in mandrills 175, 177–8 reproductive skew in mountain gorillas 176, 178 reproductive skew in rhesus macaques 174, 175, 177 sexually transmitted disease and
comparison of skew in insects and vertebrates 326–8 definition of primitively dominant control over reproduction 315 ecological constraints and skew 319–21 empirical studies of reproductive skew
and models 316–26 levels of reproductive skew 317 likely sources of reproductive conflict 314–15 nesting biology of hover wasps 314 nesting biology of Polistes wasps 312–14 peace incentives 322–3 possible determination of skew by convention 325–6
testing predictions of skew models 136–59, 172–4 within-group relatedness and skew 184–5 primitively eusocial bee (Halictus rubicundus) 314 primitively eusocial wasps 305–28 assumptions about acts of aggression 325 assumptions underlying models of skew 307–12 classification of models of
pygmy mouse lemur (Microcebus berthae) 57 quantitative genetic approach, evolution of reproductive skew 469–76 queueing by subordinates 138–9
316–18 lack of fit between data
185–6 skew models 170–2
234, 241, 242, 247, 251
eusocial 306
reproductive skew testing assumptions of
pukeko (Porphyrio porphyrio)
predictions from skew theory 314–15 subordinate inheritance of dominance 321–2 testing reproductive skew theory 312–26 threat of escalated conflict 323–4 priority-of-access (POA) model 158–9, 167–9 Propithecus diadema edwardsi (Milne-Edwards sifaka) 56
reciprocity models, cooperative reproduction in fish 286–7 recruiter-joiner model of reproductive skew 72 red-cockaded woodpecker (Picoides borealis) 409, 487 red colobus (Piliocolobus rufomitratus) 44 red deer (Cervus elaphus) 54, 445 red-fronted lemur (Eulemur fulvus rufus) 57 red howler monkey (Alouatta seniculus) 172 reindeer (Rangifer tarandus) 54 relatedness among males and skew in primates 173, 174 reproductive allocation patterns in social mammals 98–101 reproductive competition and menopause evolution 33–7 reproductive conflict
Propithecus tattersalli (golden-
resolution
crowned sifaka) 56
and relatedness
Propithecus verreauxi
reproductive skew
(Verreaux’s sifaka) 55,
307–12
56, 57–9, 73, 74
asymmetries 40–2 kinship dynamics models 37–44
521
522
Subject index reproductive conflict resolution (cont.) see also menopause; models of reproductive skew reproductive overlap in humans 34–7 reproductive regulation by
within-group relatedness 184 ring-tailed lemur (Lemur catta) 56, 57, 63–5, 447 ruffed lemur (Varecia variegata) 56, 57 “rules for responding” type of model 311
dominant females 342–59 reproductive skew definition 4 high-skew societies 4 in human societies 25–6 low-skew societies 4 measures of inequality in breeding success 4 reproductive strategies,
comparative analyses 487–90 comparative and phylogenetic approaches 486–94 complexity of the modeling problem 482–6 evaluation of usefulness
S3 (effective-number-ofbreeding-individuals skew index) 141 savannah baboon (Papio cynocephalus) 138, 184 sexual behavior, inhibition in subordinate females 343–4 sexually transmitted disease,
of skew models 499–501 influence of skew models on study 486–7 interpretations of aggression 484 life history traits and trade-offs 491–9 limitations of current
differences between
effects of reproductive
models 482–6
males and females 88
skew in primates 185–6
patterns of conflict and
reproductive suppression and skew in African molerats 380–4 in heterogeneous environments 101–3 reproductive suppression model 358 research see empirical research on reproductive skew; theoretical research on reproductive skew resolution models 309–11 resource competition hypothesis for infanticide 352–7 responsive switching in social mammals 91–2 restraint models of reproductive skew 5–8, 9–10, 307–8 rhesus macaque (Macaca mulatta) 31, 136, 138 reproductive skew 174, 175, 177
Seychelles warbler (Acrocephalus sechellensis) 234, 249 skew indices 115, 116–17, 141 skew models see models of reproductive skew skew theory, goal of 229–30 Smith’s longspur (Calcarius pictus) 288 social behavior costs and benefits 88–90 differing benefits for females and males 98–101 social dominance and reproductive skew 54–5 social evolution analyses of convergence and divergence 487–90 analyses of skew and social evolution 487–90 bottom-up comparative approaches 486–94
resolution 483–6 problems with a top-down approach 486 role of cooperative breeding in modern humans 494–9 search for a unified theory 230 selection of variables to measure 491–4 ways that one individual can influence another 484–5 socially induced stress, physiological suppression in vertebrates 417–20 spider monkeys (Ateles spp.) 172 spotted hyena (Crocuta crocuta) 55 factors affecting male reproductive success 69–70
Subject index reproductive skew among males 69–70
stable levels of competitive effort 14
reproductive skew and female dominance 66–72, 74 social system 56 Stackelberg models 42, 307–8 Stenogastrinae (hover wasps), nesting biology 314 see also
Taiwan yuhina (Yuhina brunneiceps) 234, 235, 237, 248, 250 tamarins (Saguinus spp.) 94 Tasmanian native hen 250 theoretical research on reproductive skew, future directions 459–60
subordinate reproductive restraint 402, 405–15 subordinates, evasion and
totipotency of individuals 92 transactional models of reproductive skew 5–8, 9–10 tuco-tuco (Ctenomys sociabilis)
control tactics 454
450, 452–3
superb fairy-wren (Malurus
tug-of-war models of
cyaneus) 234, 472 sweat bee (Megalopta genalis)
reproductive skew 8–9, 10–19, 308, 309–11
447 Symphodus ocellatus (Mediterranean ocellated wrasse) 268, 269, 273–8, 287, 288 synthetic models of skew
theory 11–13 “concession” zone 14–15 evaluation of outside options 13–14
Varecia variegata (ruffed lemur) 56, 57 Verreaux’s sifaka (Propithecus verreauxi) 55, 56, 57–9, 73, 74
wasps western gorillas, withingroup relatedness 184 white-browed scrubwren (Sericornis frontalis) 234,
implications for skew
244, 246, 251, 253 white-faced capuchin
modeling process 13
monkey (Cebus capucinus)
outside option principle
94, 185
11–13
white-footed mouse
“restraint” zone 15
(Peromyscus leucopus) 447,
results 16–18
449
situations where outside
primates 184–5 wood mouse (Apodemus sylvaticus) 449, 452 yellow-bellied marmot (Marmota flaviventris) 114–31 B index of skew 117 choice of skew index 115, 116–17 definition of a behavioral role 115 definition of a group 117–18 fitness consequences 122 functional implications of results 130–1 individual- versus groupdirected behaviors 129–30 non-reproductive behaviors 120–1 possible consequences of
wasps see primitively eusocial
group breakup 16 17–19
and reproductive skew in
group composition 120
11–21, 311 application of bargaining
within-group relatedness,
(Gallinula mortierii) 249,
wasps switching 91–2
gallapavo) 228, 237–8, 252, 254
primitively eusocial stochastic phenotype
wild turkey (Meleagris
white-winged chough
options influence
(Corcorax melanorhamphos)
skew 16–21
238, 243, 246, 249, 251
behavioral roles 118–19 results and discussion 122–31 scout role 121, 128–9 sentinel role 115, 124, 128 significance of social skew 114, 115 skew calculation 121–2 social skew study 115–31 study animals and study site 119–20 use of skew to study roles 115
523