Animal Behaviour: Evolution and Mechanisms

Animal Behaviour: Evolution and Mechanisms

Peter Kappeler (Ed.)

Animal Behaviour: Evolution and Mechanisms

123

Prof. Dr. Peter Kappeler University of Göttingen Dept. of Sociobiology/Anthropology & CRC Evolution of Social Behavior & German Primate Center Behavioral Ecology & Sociobiology Unit Kellnerweg 6 37077 Göttingen Germany [email protected]

ISBN 978-3-642-02623-2 e-ISBN 978-3-642-02624-9 DOI 10.1007/978-3-642-02624-9 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2010922317 c 2010 Springer-Verlag Berlin Heidelberg  This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typesetting: Data supplied by the author Cover Design: WMX Design GmbH, Heidelberg Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Preface

The study of animal behaviour has become one of the fastest growing biological disciplines in recent decades. This development can be easily inferred, for example, from the steady increase in the total number of publications on any aspect of animal behaviour, in particular also in journals with a more general readership (e.g. Nature, Proceedings of the Royal Society or Current Biology), the ever-increasing number of participants at international conferences (e.g. IEC or ISBE), and from the growing numbers of students choosing courses in this field. This development has several causes, of which I find three particularly compelling. First, it is increasingly being appreciated that behaviour is the crucial level at which an individual’s genotype and phenotype interface with the environment. Recognising behaviour as the main mechanism animals employ to ascertain their homeostasis, growth, survival and reproduction therefore provides a deep understanding of organismal integration and adaptation. Second, the astonishing success of the study of animal behaviour also has importantly to do with the intellectual flexibility and methodological inter-disciplinarity required for comprehensive analyses of behaviour. Today, students of behaviour are jacks-of-all-trades; importing, applying and improving methods from many neighbouring disciplines, such as molecular genetics, physiology or micro-electronics, as well as concepts and theories from less obvious sources, such as economics or sociology, for example. Finally, Charles Darwin’s theory of natural and sexual selection provide the study of animal behaviour with a powerful and firm theoretical framework that many closely-related disciplines (e.g. neurobiology) are lacking. This increase in the number of studies published in a growing number of ever more specialised journals and the application of new concepts, methods and technologies also has frustrating consequences, however. Except perhaps for a few exceptional colleagues, no one today is really able to develop and maintain an active research programme and to read all interesting and important publications and books that appear every month. We are increasingly forced to specialise and to restrict our attention to a few topics or taxa, despite a much larger intellectual curiosity. Most readers with a PhD will be familiar with these constraints set by increasing administrative loads, constantly changing teaching obligations as well as new types of

VI

Preface

expectations of our employers about our publication and grant acquisition records. This tendency to specialise backfires, however, when new generations of students need to be introduced comprehensively to all aspects of the study of animal behaviour, when they ask for background information to pursue their own personal curiosity, or when they ask great questions in a lecture course or seminar. One solution to this problem is to identify a useful introductory text or up-to-date review. In practice, however, textbooks tend to focus more on general principles than on current research, and reviews are typically written for specialists, and are, hence, of limited use if you want to keep abreast the literature in a broader field. A collection of authoritative reviews written by active leaders in their respective fields can fill this gap if they specifically address a non-specialist readership (i.e. not the closest peers), summarise and explain recent developments, and if they provide a forward-looking perspective for interested students as well as their closer colleagues. In 1978, John Krebs and Nick Davies began providing just this type of guidance for the then latest and fastest-growing field of animal behaviour: behavioural ecology. Their subsequent four edited volumes have informed and influenced several generations of students and academic mentors alike by providing a useful and stimulating basis for graduate seminars, a competent source of reference for non-specialists as well as a source of inspiration for newcomers and experts alike. The study of animal behaviour has made enormous leaps forward since the publication of the last Krebs and Davies volume in 1997 (Blackwell). Others have also sensed the void left by the non-continuation of their series (Danchin et al. 2008: ‘Behavioural Ecology’ Oxford University Press; Westneat and Fox 2010: ‘Evolutionary Behavioral Ecology’ Oxford University Press), but the field has become so wide that no single volume can do justice to the existing diversity of behavioural research projects any more. A recent trend among volumes of this kind appears to have been the concentration on a regional (i.e. either francophone or anglo-american) set of peers and their work; perhaps because they share certain preferred formats for teaching. The current volume attempts to fill a similar niche by featuring the state of the art in the study of animal behaviour in central continental Europe, where ethology has its deepest roots. In addition, because of space limitations, the contributions to this volume only represent a subset of current major research topics, but cover all recent international developments. For historical reasons, ethologist in German-speaking countries have mainly been interested in mechanisms, but the number of researchers embracing ultimate questions has been growing steadily. The title of this volume was therefore chosen to reflect this development. However, all authors were requested to address both ultimate and proxi-

Preface

VII

mate aspects in the presentation of their respective topics, and some chapters also deal with one or both of Tinbergen’s other questions. The contributions to this volume are organised into four broad sections. Note, however, that several chapters would also fit comfortably under a different header, as indicated by numerous cross-references among chapters. Communication and cognition continue to be central topics in the study of animal behaviour. In chapter 1, Martin Schaefer reviews evolutionary and functional aspects of visual communication. He emphasises the fact that visual signals not only play important roles in several functional domains of animals, but also in the communication between plants and animals. Using examples from both areas, he discusses models of signal design and evolution, concluding that environmental and other ecological factors need to be considered explicitly for a more comprehensive understanding of communication systems. Claudia Fichtel and Marta Manser deal with vocal communication in chapter 2. They focus on communication beyond the traditional senderreceiver paradigm and argue forcefully that bystanders also perceive vocal signals exchanged among members of social groups. This point is underscored by their review of empirical studies on audience effects and eavesdropping. Furthermore, group coordination provides a particular context, where individuals have to address several or all members of their social unit simultaneously, and where the traditional dyadic communication model fails. This review should therefore inspire much exciting new research on communication from a network perspective. In chapter 3, Dustin Penn and Joachim Frommen address kin recognition as a functionally important aspect of social and sexual behaviour. Following a much-needed conceptual clarification of the main concepts, they focus on the various mechanisms and signals involved in the recognition of kin. They also discuss central theoretical aspects in the evolution of kin recognition mechanisms as well as their genetic underpinnings and consequences. Their contribution concludes with a forward-looking perspective, identifying main problems and areas of kin recognition research requiring further work. In chapter 4, Mario Pahl, Jürgen Tautz and Shaowu Zhang use the honeybee as a model system to illustrate the fascinating sensory and cognitive abilities of animals with small nervous systems. They introduce the honeybee’s sensory world and summarise experimental work on their various cognitive abilities, including categorisation, rule learning and contextdependent learning. These sometimes stunning abilities provide instructive examples of how domain-specific cognitive faculties are linked to the specific ecological and social challenges these social insects face.

VIII

Preface

In the final chapter of this section, Kurt Kotrschal, Isabella Scheiber and Katharina Hirschenhauser use the highly-structured societies of greylag geese to provide a fascinating comparative perspective on social cognition. These geese show striking convergence in several fundamental features of their social systems with many mammals, including a femalebonded clan structure and elaborate patterns of mutual social support. Mates paired for life form the basic social units in this species and they exhibit stunning hormonal synchrony. They contribute to social stability through a particular pattern of social support that is reminiscent of some other vertebrate societies with individualised long-term relationships. Contributions to the second part revolve around the two central problems facing members of animal societies: conflict and cooperation. In chapter 6, Jürgen Heinze reviews sources of conflict and conflict resolution in social insect societies. Long regarded as harmonious superorganisms, societies of eusocial animals have only recently been recognised as harbouring internal strife as well. Using the queen-worker conflict about sex allocation and the struggle for dominance as examples, Heinze illustrates the sources and nature of such conflicts and reviews the behavioural mechanisms used to minimise or to resolve them. The important role played by inclusive fitness considerations in these conflicts highlights the interaction between ultimate and proximate factors in this context. In chapter 7, Judith Korb turns to cooperation with social insect societies as an evolutionary puzzle. Her review illuminates how relatedness and mechanisms that make cheating costly act in concert to minimise cheating and to stabilise cooperation. She goes on to show that similar mechanisms favour cooperation at other levels of biological organisation and advocates the use of a multilevel selection approach to study this and other evolutionary problems at all levels of the biological hierarchy. The subsequent chapter by Redouan Bshary is concerned with another extreme: cooperation among unrelated individuals. He introduces game theory as a particularly powerful and biologically satisfying approach to studying the problem of cooperation because it can explicitly deal with the many contingencies of real life. Using the well-studied example of marine cleaning mutualism between members of different species, he goes on to show how this approach is both necessary and useful in explaining cooperation. The tragedy of the commons provides another example where this approach might be fruitfully applied. In chapter 9, Gerald Kerth looks at group-decision making as a particular example of group-level cooperation that is potentially hampered by inter-individual conflict. How animals reconcile the potentially conflicting demands of group performance and individual interests has been subject to much recent modelling. A concise review of these models is combined

Preface

IX

with the results of empirical studies, including Kerth’s studies of Bechstein bat societies, to identify fundamental rules in animal groups of varying complexity and composition. The specific suggestions for future research in this field offered by Kerth will promote a more comprehensive approach towards studying group-decisions by incorporating additional intrinsic and social variables known to shape other aspects of social behaviour. In the final contribution to this section, Fritz Trillmich reviews the scope for cooperation and conflict in the context of parental care. Successful reproduction is often jeopardised by conflicts between mates, between parents and offspring, and among siblings. Using primarily examples from mammals, where some of these conflicts are particularly pronounced, Trillmich uses Tinbergen’s four questions to examine the nature and causes of these conflicts, the leeway for cooperation as well as the mechanisms modulating parental care. His chapter provides yet another example of the benefits of addressing a particular problem from all four perspectives, demonstrating in this case how consideration of all levels contributes to a much richer understanding of a complex problem. The third part of this volume revolves around problems of sex and reproduction. Chapter 11 by Wolf Blanckenhorn provides a broad opening perspective on these topics by outlining formal approaches to the study of sexual and natural selection, both in the wild and in the laboratory. He emphasises the benefits of using standardised selection measures in achieving a comprehensive picture of both proximate and ultimate explanations of a particular phenomenon. His illustration of how to obtain these measures from a variety of sources and how to use them in meta-analyses should facilitate and encourage a much wider use of this approach in the study of behaviour. Chapter 12 by Nils Anthes reminds most of us that sexual conflict and other sexual behaviours are not limited to the all too familiar separate-sex animals. In simultaneous hermaphrodites, male and female reproductive functions reside within the same body, providing an interesting twist for the study of sexual strategies. Anthes reviews a growing body of literature on hermaphrodites from an array of taxa that reveals fascinating evidence for the occurrence of mate choice, sexual conflict and post-copulatory selection in these animals. How hermaphrodites reconcile their intrinsic conflicts between their male and female function provides another prime example for organismal fine-tuning in face of fundamental problems, such as successful reproduction. Bart Kempenaers and Emmi Schlicht deal with an intuitively much more familiar problem in chapter 13: the optimisation of reproductive success through extra-pair behaviour in pair-living species. Focussing on a vast literature on birds, these authors first explore sources of variation in extra-

X

Preface

pair behaviour among species and populations. As extra-pair copulations in pair-living species constitute a puzzle for evolutionary biologists, they examine the associated potential costs and benefits from the perspectives of both sexes, identifying effects in the expected direction, i.e. males can increase offspring number and females offspring quality through extra-pair activities. Kempenaers and Schlicht also explore hitherto neglected aspects of extra-pair behaviour by illuminating its effects on the strength of sexual selection. In chapter 14, Bernhard Kraus and Robin Moritz focus on the polyandrous mating behaviour of social hymenoptera. The mating of hymenopteran queens with multiple males is of fundamental importance for these animals because of its genetic consequences for their societies where conflict and cooperation are so finely tuned to prevailing relatedness patterns. The authors review all major hypotheses proposed to explain the evolution of polyandry and subsequently explore the consequences of genetic variation for the social organisation of colonies. Using the honeybee as one of the best-studied model systems for these problems, they illustrate how genetic analyses have revolutionised our understanding of social insect societies. In chapter 15, Jutta Schneider and Lutz Frommhage provide an additional challenge to conventional sex roles by dissecting the mating strategies of male spiders that invest heavily in paternity enhancement with one (or very few) female(s), rather than trying to maximise access to many females. A small group of spiders is characterised by such exceptional males, and therefore provides an instructive example to reflect on traditional sex roles and any sex-specific rules that have emerged from studies of more ‘normal’ animals. The authors outline the fascinating diversity of mating strategies among spiders and explain how aspects of their genital morphology ultimately influence female (cannibalism) and male (monogyny) reproductive strategies. In the last chapter of this part, Wolfgang Goymann and Heribert Hofer use the theme of this volume to examine the relationships and interactions among mating systems, social behaviour and hormones. They show that testosterone is an important proximate factor involved in the regulation of different mating systems across species and that it also influences individual mating decisions in some species. Additional hormones (oxytocin and arginine vasopressin) are involved in pair-bond formation and the expression of different mating systems; in one particularly well-studied case, changes in one hormone receptor gene have been shown to elicit a cascade of massive changes in social behaviour. A key conclusion of this chapter resounds the title of this volume: proximate and ultimate aspects of a particular behavioural phenomenon are best studied in combination.

Preface

XI

Contributions to the final section of this volume deal with various aspects of behavioural variation. In chapter 17, Norbert Sachser and Sylvia Kaiser focus on the development of behavioural variation among and within individuals. First, they examine the role of genetic polymorphism and gene-environment interactions in generating individual variation in behavioural traits. In the second part of their chapter, Sachser and Kaiser summarise our current knowledge about the modulation of behavioural development by social factors. Focussing on mammals and proximate mechanisms, they show how the social environment of a pregnant female affects her offspring’s behaviour later in life, and how variation in social influences produces lasting effects on the behaviour of a developing individual throughout adolescence. As argued by the authors, such variability appears advantageous from an evolutionary perspective, as individuals are rapidly adapted to prevailing social environments. In chapter 18, Michael Taborsky and Jane Brockmann focus on alternative reproductive tactics as widespread and diverse adaptations in the struggle for successful reproduction that also represent a form of behavioural variation. They examine alternative reproductive tactics from the perspective of different allocation decisions in response to trade-offs in reproduction or life history optima. They summarise the known diversity of alternative reproductive tactics, describe the proximate mechanisms underlying them and discuss the evolutionary consequences of their coexistence. This phenomenon provides a compelling example for both the sensitivity of sexual selection towards exploiting any possibility for achieving reproductive success, as well as for the organismal integration required to build and sustain different prototypes of the same sex and species. In recent years, inter-individual variation in behaviour has received much attention. In chapter 19, Ralph Bergmüller reviews these new developments, detailing both, evolutionary and proximate aspects. He first provides a welcome discussion of terms, concepts and definitions before looking at the two sides of the animal personality coin: constraint and flexibility. This research topic also provides a good opportunity to illuminate the genetic bases of various behavioural traits, and Bergmüller summarises our current understanding of these and other proximate mechanisms. Consistent individual differences also provide a unique opportunity to study behavioural development as well as the fitness consequences of any differences. The author’s up-to-date summary of all these aspects provides useful orientation in a new and rapidly growing field of research, and it contains plenty of suggestions for interesting future research projects for those entering this field. Behavioural variation is not only rampant among individuals, but also among populations. In chapter 20, Carel van Schaik focuses on social

XII

Preface

learning as a ubiquitous mechanism contributing to this variation. He reviews and defines the various forms of social learning and analyses their social and cognitive preconditions. Social transmission of novel skills or behaviours can lead to cultural differentiation among populations of the same species, and van Schaik reviews the known examples of animal cultures in the second part of his chapter. His analysis indicates that animal cultures may be much more widespread, but the sophistication found in our own species remains unparalleled in nature. His chapter may well provide a basis for a much-needed general theory of animal culture. In the final chapter of this volume, Peter Kappeler and Cornelia Kraus provide an extensive overview of levels and causes of behavioural variability. They begin by reviewing the paradigm shifts that the study of behavioural variation has experienced since the early days of ethology. They go on to explore the different hierarchical levels of behavioural variation above the species level, within species, among individuals and within individuals over time and highlight the most important mechanisms constraining, maintaining or generating behavioural variability. One important conclusion emerging from this overview is that phylogenetic constraints on behaviour are surprisingly pervasive, even in species with higher cognitive abilities.

Acknowledgements I wish to thank a number of people for their contribution to this volume. First, the authors of the chapters of this volume have been extremely cooperative and disciplined. Because the importance of contributions to edited volumes cannot (easily) be assessed through ‘impact factors’, writing book chapters has gravitated towards the lower end of everyone’s professional priority list. However, the contributors to this volume have given this project top priority; having written and revised their chapters within just 8 months. In addition, most authors also peer-reviewed at least one other chapter. I thank all of them for making this project possible and for keeping it on schedule. I am equally grateful to the external experts who improved the quality of every single chapter with their constructive comments. Because reviewing a chapter for an obscure book project is even lower on anybody’s priority list, I want to sincerely thank the following colleagues for their professional altruism: Elizabeth Adkins-Regan, Theo Bakker, Tim Birkhead, Dan Blumstein, Jacobus Boomsma, Lars Chittka, James Curley, Melanie Dammhahn, Mark Elgar, Matthew Grober, Rachel Kendal, Eric Keverne, Peter Klopfer, Hannah Kokko, Rolf Kuemmerli,

Preface

XIII

Janet Leonard, Randolf Menzel, Nico Michiels, Doug Mock, Allen Moore, Bryan Neff, Kees van Oers, Carsten Schradin, Thomas Seeley, Dirk Semmann, Lotta Sundström, Zuleyma Tang-Martinez, David Westneat, Andrew Whiten and four anonymous referees. My sincere thanks are due to Stefanie Wolf, my editor at Springer Verlag, for her support of this project from the beginning, for her help with administrative matters, and for her patience and understanding. Finally, words are not enough to acknowledge the crucial role of Ulrike Walbaum in the production of this volume. She single-handedly formatted all texts, re-drew most figures and double-checked every single reference. Thank you Ulli – this book is really your ‘second baby’! Göttingen, January 2010

Peter Kappeler

Contents

Part I Communication and cognition Chapter 1 Visual communication: evolution, ecology, and functional mechanisms .......................................... 3 H. MARTIN SCHAEFER Chapter 2 Vocal communication in social groups...................................................... 29 CLAUDIA FICHTEL AND MARTA MANSER Chapter 3 Kin recognition: an overview of conceptual issues, mechanisms and evolutionary theory .................................................................................... 55 DUSTIN J. PENN AND JOACHIM G. FROMMEN Chapter 4 Honeybee cognition ................................................................................... 87 MARIO PAHL, JÜRGEN TAUTZ AND SHAOWU ZHANG Chapter 5 Individual performance in complex social systems: the greylag goose example....................................................................... 121 KURT KOTRSCHAL, ISABELLA B.R. SCHEIBER AND KATHARINA HIRSCHENHAUSER Part II Conflict and cooperation Chapter 6 Conflict and conflict resolution in social insects ..................................... 151 JÜRGEN HEINZE

XVI

Contents

Chapter 7 Social insects, major evolutionary transitions and multilevel selection .. 179 JUDITH KORB Chapter 8 Cooperation between unrelated individuals – a game theoretic approach ....................................................................... 213 REDOUAN BSHARY Chapter 9 Group decision-making in animal societies............................................. 241 GERALD KERTH Chapter 10 Parental care: adjustments to conflict and cooperation............................ 267 FRITZ TRILLMICH Part III Sex and reproduction Chapter 11 The quantitative study of sexual and natural selection in the wild and in the laboratory .............................................................. 301 WOLF BLANCKENHORN Chapter 12 Mate choice and reproductive conflict in simultaneous hermaphrodites ........................................................................................ 329 NILS ANTHES Chapter 13 Extra-pair behaviour ................................................................................ 359 BART KEMPENAERS AND EMMI SCHLICHT Chapter 14 Extreme polyandry in social Hymenoptera: evolutionary causes and consequences for colony organisation.............. 413 F. BERNHARD KRAUS AND ROBIN F.A. MORITZ

Contents

XVII

Chapter 15 Monogynous mating strategies in spiders................................................ 441 JUTTA SCHNEIDER AND LUTZ FROMHAGE Chapter 16 Mating systems, social behaviour and hormones .................................... 465 WOLFGANG GOYMANN AND HERIBERT HOFER Part IV Behavioural variation Chapter 17 The social modulation of behavioural development ................................ 505 NORBERT SACHSER AND SYLVIA KAISER Chapter 18 Alternative reproductive tactics and life history phenotypes................... 537 MICHAEL TABORSKY AND H. JANE BROCKMANN Chapter 19 Animal personality and behavioural syndromes...................................... 587 RALPH BERGMÜLLER Chapter 20 Social learning and culture in animals ..................................................... 623 CAREL P. VAN SCHAIK Chapter 21 Levels and mechanisms of behavioural variability.................................. 655 PETER M. KAPPELER AND CORNELIA KRAUS Index ........................................................................................................ 685

Contributors

NILS ANTHES Animal Evolutionary Ecology Institute for Evolution and Ecology University of Tübingen, D [email protected]

CLAUDIA FICHTEL Dept. of Behavioral Ecology and Sociobiology German Primate Center (DPZ) Göttingen, D [email protected]

RALPH BERGMÜLLER Eco-Ethologie University of Neuchâtel Neuchàtel, CH [email protected]

LUTZ FROMHAGE Group Behavioural Biology Biocenter Grindel University of Hamburg Hamburg, D [email protected]

WOLF BLANCKENHORN Institute of Evolutionary Biology and Environmental Studies University of Zürich-Irchel Zürich, CH [email protected] H. JANE BROCKMANN Dept. of Biology University of Florida Gainesville, USA [email protected] REDOUAN BSHARY Eco-Ethologie Institut de Biologie Université de Neuchâtel Neuchâtel, CH [email protected]

JOACHIM G. FROMMEN Konrad Lorenz Institute for Ethology Austrian Academy of Sciences Vienna, A [email protected] WOLFGANG GOYMANN Dept. Behavioural Neurobiology MPI for Ornithology Seewiesen, D [email protected] JÜRGEN HEINZE Institute of Zoology University of Regensburg, D [email protected]

XX

Contributors

KATHARINA HIRSCHENHAUSER Dept. Behavioural Neurobiology MPI for Ornithology Seewiesen, D [email protected] HERIBERT HOFER Leibniz Institute for Zoo and Wildlife Research Berlin, D [email protected] SYLVIA KAISER Dept. of Behavioural Biology University of Münster Münster, D [email protected]

JUDITH KORB Behavioral Biology University of Osnabrück Osnabrück, D [email protected] KURT KOTRSCHAL Konrad Lorenz Research Station Grünau, A & Dept. of Behavioural Biology University of Wien, A [email protected] F. BERNHARD KRAUS Molecular Ecology Institute of Biology University of Halle-Wittenberg Halle (Saale), D [email protected]

PETER M. KAPPELER Dept. of Behavioral Ecology and Sociobiology German Primate Center (DPZ) & Dept. for Sociobiology and Anthropology University of Göttingen Göttingen, D [email protected]

CORNELIA KRAUS Dept. for Sociobiology and Anthropology University of Göttingen Göttingen, D [email protected]

BART KEMPENAERS Dept. Behavioural Ecology and Evolutionary Genetics MPI for Ornithology Seewiesen, D [email protected]

MARTA MANSER Animal Behaviour Institute of Zoology University of Zürich Zürich, CH [email protected]

GERALD KERTH Animal Behaviour Institute of Zoology Zürich, CH [email protected]

ROBIN F.A. MORITZ Institute of Biology University of Halle-Wittenberg Halle (Saale), D [email protected]

Contributors

XXI

MARIO PAHL BEEgroup, Biocenter Würzburg University Würzburg, D [email protected]

JUTTA SCHNEIDER Group Behavioural Biology Biocenter Grindel University of Hamburg Hamburg, D [email protected]

DUSTIN J. PENN Konrad Lorenz Institute for Ethology Austrian Academy of Sciences Vienna, A [email protected]

MICHAEL TABORSKY Dept. Behavioural Ecology Institute of Zoology University of Bern Hinterkappelen, Bern, CH [email protected]

NORBERT SACHSER Dept. of Behavioural Biology University of Münster Münster, D [email protected]

JÜRGEN TAUTZ BEEgroup, Biocenter Würzburg University Würzburg, D [email protected]

H. MARTIN SCHAEFER Dept. of Animal Ecology and Evolutionary Biology University of Freiburg, D [email protected]

FRITZ TRILLMICH Behavioral Biology University of Bielefeld [email protected]

ISABELLA B.R. SCHEIBER Konrad Lorenz Research Station Grünau, A [email protected]

CAREL P. VAN SCHAIK Anthropological Institute and Museum University of Zürich Zürich, CH [email protected]

EMMI SCHLICHT Dept. Behavioural Ecology and Evolutionary Genetics MPI for Ornithology Seewiesen, D [email protected]

SHAOWU ZHANG Visual Sciences Group Research School of Biology Canberra ACT 2601, AU [email protected]

Part I Communication and cognition

Chapter 1

Visual communication: evolution, ecology, and functional mechanisms H. MARTIN SCHAEFER

ABSTRACT Throughout their lives, animals gather and weigh information to decide upon alternative states. Many of the decisions in social interactions, mate choice, and intrasexual competition rely upon information transfer in visual communication between senders and receivers. Furthermore, animals as well as plants also engage in interspecific visual communication such as warning displays and the attraction of mutualists. In the first part of this chapter, I discuss how the sensory and neuronal processes involved in detecting and recognising visual stimuli can influence the evolution of the design of informative visual traits. Based upon this functional understanding I review two models of signal evolution, sensory exploitation and sensory drive, in the second part of the chapter. The sensory exploitation model predicts that those signals evolve that are more efficient in stimulating the sensory or perceptual system of the receiver. The sensory drive model predicts that the sensory systems of animals, the signals they use, and their habitats are evolutionarily coupled. Finally, I discuss several case studies from plant-animal communication and sexual selection in birds to illustrate the point that many colours are strongly influenced by the environment. As such, there is a need to integrate ecology into the primarily evolutionary concept of communication theory in order to understand the evolutionary dynamics and diversity of communication systems.

1.1 Introduction Communication is ubiquitous and key to the organisation of behaviour. Cells communicate with each other, members of social groups communicate with each other, and members of different species often do the same. Communication is therefore widespread and occurs at different organisa-

4

H. Martin Schaefer

tional levels. It involves participants that differ widely in the tightness of their association. While there are many definitions of communication, most authors agree that communication involves the provisioning of information by a sender and the evaluation of this information in decision making by a receiver. We need to know then what information is. In a colloquial sense, information can be defined as the reduction in the uncertainty of the receiver about alternative conditions of the sender (see Shannon and Weaver 1949). Yet, it is often difficult to measure whether the trait of a sender used in communication actually reduces the uncertainty of the receiver. Communication is often manipulative, and there are clearly communication systems, such as camouflage, that do not reduce the uncertainty of the receiver. Here I adhere to the definition of communication by Dawkins and Krebs (1978), according to which communication is an action of a sender that influences the sensory system of a receiver so that the receiver changes its behaviour to the benefit of the sender. This definition circumvents the contentious issue of information. It states that by communicating, the sender tries to manipulate the receiver so that its behaviour increases the fitness of the sender. Clearly, communication can result in the transfer of information. In this case it is the transfer of private information that the senders has but the receivers is uncertain about. Despite this asymmetry in the control of information flux, the evolution of communication is essentially a coevolutionary process (Maynard Smith and Harper 1995). Note that this coevolutionary process can be one of antagonistic evolution, that is an arms race between senders and receivers. Yet, many authors maintain that communication, although often manipulative, benefits on average the senders and receivers of information (e.g., Maynard Smith and Harper 1995). Note the use of ‘on average’ in the previous sentence. Whereas senders almost always benefit from communication – otherwise they will be directly selected against and communication will not be evolutionarily stable – , there are cases in which receivers do not benefit, e.g., when they are exploited by the senders. Exploitation of communication is widespread; examples range from permanently rewardless flowers that exploit pollinators for their reproduction to bola spiders that chemically mimic the sex pheromone of their prey. Yet, if receivers do not benefit from responding on average to a given stimulus, they will stop doing so and thereby disrupt the communication system. Accordingly, if males did not obtain on average fitness benefits by responding to the female sex pheromone (one sender), the bola spider (second sender) could not exploit the inter-sexual communication system between the moths.

Visual communication: evolution, ecology, and functional mechanisms

5

The exploitation of pollinators by rewardless flowers and of male moths by bola spiders documents that communication often consists of several different senders and receivers. Thus, even intraspecific communication can involve different species, e.g., potential mates and predators that respond to the same signals and whose selection pressures shape the design of the signal. Even within a species, the sexes can represent different receivers. For example, a trade-off between epigamic signals to females and antagonist signals to males in the red-collared widowbird Euplectes ardens entails that the evolution of one signal is not independent from the evolution of the other signal (Andersson et al. 2002). Although it is conceivable that very specialised communication systems may only involve one sender and one receiver (e.g., chemical communication between mates that requires specific receptors), eavesdropping is common and the very nature of conspicuous communication makes it particularly prone to exploitation by predators. As such, many communication systems consist of multiple senders and receivers, whose relative abundance shapes the selective pressures on the design of communication. Although this network character of communication is increasingly acknowledged, communication has typically been viewed as a binary sender and receiver game. These binary games allow to elucidate fundamental properties of information transfer which can then be extended to more complex communication system and signalling networks at all levels of biological organisation (Skyrms 2009). Signals are the vehicles of communication. They are defined as morphological, physiological, or behavioural characteristics that are maintained by natural selection because they convey information to other organisms (Otte 1974). Otte’s definition is very useful even if we avoid the term ‘convey information to’ and use ‘manipulate other organisms’ instead. This definition is useful because it provides an evolutionary perspective that builds on the selective pressures that shape traits. The term signal is often defined more loosely as a trait that conveys information or that elicits a response in a receiver (Wiley 1994). However, traits may convey information without being primarily selected for that function (see BOX 1.1). For example, body size can reliably indicate the fighting ability of an individual without being selected for a role in communication (Maynard Smith and Harper 1995). Similarly, kestrels (Falco tinnunculus) detected vole scent marks in UV light but not in light without a UV component. Accordingly, they can use UV reflectance as visible cue that is a giveaway of the presence of voles in a given area (Viitala et al. 1995). Although informative, body size and UV reflectance in these examples are not signals because they are not selected to manipulate the opponents and kestrels, respectively.

6

H. Martin Schaefer BOX 1.1 Distinguishing between signals and cues can be challenging Even if a trait conveys information that alters the behaviour of a receiver to the benefit of both sender and receiver, it is not necessarily a signal. This is because the selective pressure of the receiver can feasibly be overridden by those of other selective agents, such as abiotic factors. This is a common theme in visual signals because pigments often fulfil multiple roles. For example, it is well known that melanins may protect against high irradiance and play a role in thermoregulation. The role of these non-communicative functions is not well understood, mainly because hypotheses on the adaptive value of non-communicative functions are not directly pitted against the hypothesis of an adaptive communicative role. This drawback leads to an underestimation of the complexity of communication systems. Plant-animal interactions provide a good example of the complexity of communication. Anthocyanins are plant pigments that impart red, blue, or black colouration to plant tissue. Anthocyanins are at the same time powerful dietary antioxidants that contribute to the health of consumers (Kong et al. 2003, Stinzig and Carle 2004). This is the main reason why there is considerable ongoing interest to increase the contents of bioactive anthocyanins, and more generally flavonoids and other plant phenols, in human nutrition. Surprisingly, however, the evolutionary ecology of anthocyanins, particularly in relation to plant-animal interactions has been a neglected research topic although anthocyanins are the major pigments in fruit and flower displays. Dark colouration in fruits is imparted by anthocyanins. Birds can thus visually evaluate the contents of anthocyanins as dietary antioxidants in fruits (Schaefer et al. 2008a). Owing to the necessary link between dark fruit colouration and anthocyanin contents, dark fruit colouration is best seen as an index of anthocyanin contents. The frugivorous blackcap (Sylvia atricapilla) selects artificial food according to its anthocyanin contents. Importantly, selecting food containing anthocyanins improves the humoral immune response of these birds (Catoni et al. 2008). Birds are thus apparently able to select fruits using their colour as a reliable signal of health benefits. They can thus attend to fruit colouration to increase their intake of flavonoids for self-medication (Catoni et al. 2008, Schaefer et al. 2008a). Conceivably, plants could increase their reproductive success by communicating antioxidant rewards to seed dispersers. If birds select fruits according to their health benefits and colours, the important evolutionary question is whether fruit colouration evolved as a signal to communicate these health benefits to seed dispersers. Examining fruit colour in the binary signalling game between plants and seed dispersers might suggest so because the health benefits of flavonoids occur not only in birds but more widely also in mammals. Yet, fruits are not only consumed by seed dispersers but also by fruit predators that consume fruit pulp without dispers-

Visual communication: evolution, ecology, and functional mechanisms

7

ing the seeds. Bacteria and fruit-rot fungi are actually the most ubiquitous fruit consumers that can have strong, detrimental effects on plant fitness (Janzen 1977, Cipollini and Stiles 1993). Importantly, high contents of anthocyanins strongly inhibited fungal growth in vitro and in damaged grapes reducing fungal growth by 95% (Schaefer et al. 2008b). Thus, the selective pressure of both mutualistic animals (seed dispersers) and antagonists (fruitrot fungi) are aligned to select for increased anthocyanin contents. Unless it is shown that selection for communication overrides selection for defence, anthocyanin contents is best thought of as a cue, not a signal.

In order to understand the evolution of communication it is important to differentiate between traits that are selected for their communicative role, i.e., signals, and those which are not. Traits that are not primarily selected for a communicative function and that can be informative are best viewed as cues (Otte 1974, Seeley 1989). Note my emphasis on the word primarily, a word that was missing in the previous definitions of cues. It is becoming increasingly clear that traits, including those used in communication, are shaped by multiple selective pressures (BOX 1.1). Pigments, for example, can function as signals and they are often concomitantly selected for their role in thermoregulation. It is therefore essential to evaluate the relative selective pressures acting on a given trait. I therefore conclude that signals are primarily selected for their communicative role and can thus evolve to manipulate the receiver, whereas the evolution of cues towards increasing manipulation is constrained because they are primarily selected for a non-communicative role. The kestrel example (above) testifies to the difficulties of distinguishing between cues and senders if distinct receivers attend to the same trait. Obviously, scent marks are very likely selected to communicate with other voles and thus represent signals. Their visual properties, however, have certainly not been selected to communicate to kestrels which therefore attend to cues when foraging on voles. Thus, if predators eavesdrop (‘cue in’) on sexually selected signals, these traits constitute signals to mates but cues to predators. Finally, a trait that is necessarily informative – e.g., if body size is an unfakeable indication of fighting ability – is best viewed as an index of a certain quality (Maynard Smith and Harper 1995). In the following, I will briefly review the sensory (Sect. 1.3) and neuronal basis (Sect. 1.4) for colour vision to show how animals’ visual sensitivities may influence the design of visual signals (Sect. 1.2). Based on this information, I evaluate key concepts in signal evolution in the following sections. Although most concepts that I discuss in this chapter also pertain more widely to communication in other sensory modalities, I will focus on

8

H. Martin Schaefer

the evolution and the proximate mechanisms underlying visual communication.

1.2 Visual sensitivities Most animals are able to extract information from visual traits using two different aspects, achromatic information, that is variation in brightness or the intensity of reflected light from a given surface, and chromatic information on colour. Variation in brightness is perceived on the scale of white to black and often referred to as luminance if analysed according to the sensory abilities of animals. Humans perceive saturation and hue as chromatic aspects of visual information. Saturation describes the colours’ similarity to a chromatically neutral shade such as grey (or white or black); colours that contain no or very little grey are deeply saturated, whereas a grey with a little tint of colour has a low saturation (see Kelber et al. 2003 for a thorough review). In contrast, hue defines colour differences that are related to colour categories, e.g., red, yellow, and green. While there is no evidence that animals perceive hue and saturation as humans do, there is some evidence that bumblebees have a perceptual dimension of saturation (Lunau et al. 1996). Furthermore, experiments on chickens show that they categorise colours and generalise across them (that is treating stimuli that can be differentiated as equivalent) in a way similar to humans (Jones et al. 2001). In order to extract colour information, an animal needs at least two distinct receptors that differ in their spectral sensitivities because this is a prerequisite to distinguish between colours that do not differ in brightness. Animals can thus be characterised according to the number of distinct receptors that are found in their retina and that are used for colour vision. They vary from monochromats, dichromats (many mammals), trichromats (primates, many insects), to tetrachromats (most fishes yet known, most birds and lizards, some insects; see Fig. 1.1). Importantly, behavioural experiments are required to assess the dimensionality of colour vision of a given animal. For example, mantis shrimps are exceptional in possessing 16 morphologically different photoreceptors, but the dimensionality of their colour vision is currently unknown and likely involves fewer dimensions (Marshall et al. 2007). Understanding the spectral sensitivities of animals allows modelling the initial steps of colour perception. In general, visual neurons may either sum photoreceptor signals (in achromatic vision), or compare them by some type of inhibitory interaction (in chromatic vision) to give the ratio

Visual communication: evolution, ecology, and functional mechanisms a

1.2

9

UV

1.0

Sensitivity

0.8 0.6 0.4 0.2 0.0

b

1.2

UV

1.0

Sensitivity

0.8 0.6 0.4 0.2 0.0 300

400

500

600

700

Wavelength [nm]

Fig. 1.1 The spectral sensitivities of animals differ. Birds are tetrachromatic, possessing four cone types that differ in spectral sensitivities (a), whereas many insects such as bees are trichromatic, possessing three different cone types (b). The normalised spectral sensitivities of the cones are illustrated by the white lines. Colour shade indicates colour variation as seen by human eyes except for the UV which humans do not perceive. Note that bees are less sensitive to red light than birds (see Sect. 1.3.2).

or difference of receptor signals (Kelber et al. 2003). It is important to acknowledge that colour perception is much more complicated than these first basic steps of neuronal coding that are currently estimated in visual modelling. In particular, colour is not only in the eye of the beholder, but

10

H. Martin Schaefer

involves higher cognitive processes. As yet, the cognitive component of signal evaluation is still a black box in most non-verbal animals and represents one of the most important frontiers in our current understanding of signal evolution; not only in visual communication but more widely in all sensory modalities (Chittka and Brockmann 2005). Despite these obvious limitations, the use of colour vision models is a powerful tool allowing for tests of specific predictions about the selective pressures upon signal design. This explains their increasingly widespread use to explain the design of communication in aposematism, sexual selection, camouflage, and interspecific interactions between mutualists or antagonists (e.g., Heiling et al. 2003, Cuthill et al. 2005, Endler et al. 2005, Darst et al. 2006, Schaefer et al. 2007).

1.3 Sensory selection on signal design 1.3.1 Conspicuousness Colour vision modelling enables us to examine the basic assumption of signal theory that signal receivers should select for signals that are more conspicuous and easier to detect. This assumption seems intuitively plausible and it has been supported by theoretical modelling (Schluter and Price 1993). However, there are few tests of this assumption under natural conditions (Schaefer et al. 2006, Cazetta et al. 2009). Such tests would be desirable because the relationship between detectability and fitness benefits is unlikely to be constant in distinct environments and communication systems. It can even vary within communication systems. For example, Cazetta et al. (2009) analysed whether fruit detection is a function of the detectability of different fruit colours. According to their preliminary evidence from four different colours, the relationship between detectability and fruit consumption is linear in low contrasting fruits and asymptotic in strongly contrasting fruits. Animals can show behavioural modifications that increase the conspicuousness of their sexually selected signals. Manakins – small, neotropical forest-dwelling birds – have a lek system where males congregate at traditional mating arenas to compete for females. Males of the golden-collared manakin (Manacus vitellinus) show a particular behaviour in that they clear the court in which they are displaying. Colour vision modelling revealed that the plumage patches used during courtship are more contrasting against the clear court than they are against the typical leaf litter of the forest floor (Uy and Endler 2004). The behavioural modification may thus act to increase the conspicuousness of males to females, particularly be-

Visual communication: evolution, ecology, and functional mechanisms

11

cause, apart from colour contrasts, clearing the court results in a less variable background that further enhances the transmission of visual signals. 1.3.2 Selection by multiple receivers A cornerstone of sexual selection theory is that sexual and natural selection are two opposing forces shaping animal traits. Indeed, Endler (1980) demonstrated in his seminal study that guppy pigmentation is mediated by a balance between sexual selection driving the evolution of colour signals towards maximized conspicuousness for conspecifics and natural selection selecting against conspicuous individuals. This is because both mates and predators respond more strongly to conspicuous, i.e., contrasting colours. Increased viability costs associated with increased signalling thus provide one possible mechanism to ensure the evolutionary stability of reliable information transfer. Private communication channels are one way to resolve the evolutionary trade-off between natural and sexual selection in animal colouration. Private communication channels occur if prey uses colour signals that are more conspicuous to conspecifics than they are to predators. Note the relative difference; private communication channels are not invisible to predators, predators are just less sensitive to them. A prerequisite for private communication channels is that prey and predators differ in their visual sensitivities. This situation is common because many predators belong to different taxonomic classes than their prey species. Even if prey and predators belong to the same class, they may have distinct visual abilities. For example, the spectral sensitivity of the short-wave receptor of songbirds is shifted towards the ultraviolet, whereas it is apparently shifted towards the violet part of the spectrum in hawks and crows (Ödeen and Hảstad 2003). This difference explains why, according to vision modelling, the plumage colours of European songbird species are more conspicuous to their conspecifics than to their avian predators (Håstad et al. 2005). Similarly, the UV ornamentation of male swordtails (Xiphorus spp.) increases their attractiveness to females but not to their predator, the Mexican tetra (Astyanax mexicanus), that is less sensitive to UV (Cummings et al. 2003). It remains unresolved, however, how widespread such private communication channels really are, mainly because the visual sensitivities of many species are not known and it is doubtful how well current visual models predict their visual abilities (Stevens and Cuthill 2007). Differences in sensory abilities are also important in interactions between multiple species that are ecologically more similar than predators and potential mates. Such interactions among multiple species are com-

12

H. Martin Schaefer

monly found in plant-animal interactions where plants interact with suites of mutualistic and antagonistic species. Differences in the visual abilities of animals may result in divergent selective pressures of these agents. For example, monkeyflowers (Mimulus spp.) are pollinated by trichromatic bees and tetrachromatic hummingbirds. Introgressing different flower colour alleles into near iso-genic lines of two Mimulus species, Bradshaw and Schemske (2003) showed that a shift in floral colour resulted in a concomitant change in pollinator spectrum (hummingbirds vs. bees) that could, in theory, be initiated by a single mutation. This is an extremely interesting example because the sensory ecology of pollinators predicts their reaction towards the floral colour shift. Pink flowers are pollinated by bumblebees which are not very sensitive to variation in the red part of the spectrum. Hummingbirds, however, are very sensitive to red and pollinate predominantly the red flowers. Thus, a single mutation can be adaptive, leading to premating isolation owing to differential attraction of pollinators. The magnitude of this single mutation is explicable by considering the sensory ecology of different signal receivers. This, in turn, allows for understanding the repeated evolution of pollination by hummingbirds from ancestral insect-pollinated species in the North American flora. The examples above document how natural selection upon both senders and the sensory systems of receivers shapes the design of signals. This effect is also visible on different time scales. On a developmental time scale, tadpoles develop different colours depending on the presence and type of predators (Touchon and Warkentin 2008). While the induced colour changes of tadpoles may have been adaptive, rendering them less conspicuous to the two types of predators (dragonfly larvae and fish), the adaptiveness has not been demonstrated in this study. The most rapid predator-induced colour changes occur in species like chameleons that change colour within milliseconds or seconds. Chameleons show adaptive, predator-specific colour change if faced with predators that differ in visual abilities. They show stronger background matching if faced with an avian predator of acute colour vision relative to a snake with less developed colour vision (Stuart-Fox et al. 2008).

1.4 Selection by lower neuronal processes Not only the spectral sensitivities of animals, but also their neuronal and cognitive processes have selected for the design of colour signals. Although neuronal processes involved in the detection of signals and their recognition form a continuum, I tentatively discriminate between higher

Visual communication: evolution, ecology, and functional mechanisms

13

and lower neuronal processes. The latter represent the first steps of neuronal coding of the output of receptor cells, whereas the former are involved in cognitive processes (see Sect. 1.5). Major advances in recent years have demonstrated that lower neuronal processes can shape the selective pressures that predators exert upon the design of camouflage traits. There are many forms of camouflage including background matching (where prey matches the background in colour, form and shape), disruptive colouration, and masquerade (where prey resembles an uninteresting specific item of the background such as a stick or a leaf; Fig. 1.2). Because these different camouflage techniques have been recently reviewed (e.g., Stevens 2007), I will focus here only upon the effects of neuronal coding. Disruptive colouration thwarts detection by sporting contrasting patterns on the body outline (Cott 1940, Merilaita 1998). These contrasting patterns are particularly effective in hindering detection of a prey by exploiting edge detection mechanisms in early visual processes (Stevens 2007). Edge detection mechanisms enhance the sensitivity of predators to contrasts because these are naturally associated with the boundaries of objects (Osorio and Srinivasan 1991). Contrasting marginal patterns that create false boundaries and mask the shape of the body can thereby conceal the body outline (Cuthill et al. 2005). Edge detection mechanisms are probably optimised for detecting targets against the heterogeneous background of natural scenes. For example, a background of leaves varies dramatically, that is over 3 log units, in brightness (Sumner and Mollon 2000, Regan et al. 2001). Variation in brightness is particularly pronounced if there is a mosaic of sunlit spots and shadows created by three-dimensional objects. Thus, many organisms apparently segregate images based upon such high-contrast borders. As a consequence, edge detection is mediated primarily by achromatic contrasts because achromatic variation typically exceeds, and is therefore more reliable than, variation in chromatic contrasts (Osorio et al. 1999). Conversely, chromatic contrasts are particularly valuable for identifying objects against an achromatically variable background such as foliage (see Sect. 1.6). Marginal markings in prey that create either chromatic or achromatic contrasts can yield effective disruptive colouration that camouflages its bearer (Schaefer and Stobbe 2006, Stevens et al. 2006). Interestingly, both studies working on artificial butterfly-like models revealed that disruptive colouration is effective, albeit less so, even if the colours do not match those of the background. However, two further studies showed with different patterns that increasing contrast to the background decreases the survival probabilities of disruptive prey (Stevens et al. 2008a, Stobbe and Schaefer 2008). Thus, the extent to which disruptive colouration enables

14 a

d

H. Martin Schaefer b

c

e

Fig. 1.2 Animals use different visual techniques of protective camouflage to evade predators. Toxic animals often defend themselves with highly contrasting colours (a), whereas undefended prey can evade predators by background matching like the amphibian in (b) or disruptive colouration that is characterised by contrasting marginal colour patterns that disrupt the body outline (c). Disruptive colouration is thus very different from the border profile enhancement seen in (d). Contrasting spots are another form of protective colouration (e). The traditional hypothesis is that they startle predators through eye mimicry but recent experiments revealed that they rather serve as protective colouration instead by diverting the attention of the predator away from the shape of the body (Stevens et al. 2008b).

the bearer to exploit a larger diversity of habitats than background matching is not well resolved and may depend on form and location of the contrasting patterns. The eyespots that many butterflies but also fish (and possibly birds such as pygmy owls) sport (Fig. 1.2) are another visual feature that is apparently selected by the design of receptive fields in the retina of predators. The traditional hypothesis is that such eye-like spots mimic the eyes of a larger animal and thereby startle potential predators. However, such circular, contrasting spots are highly effective in stimulating the centre-surround arrangement of visual fields in the retina of vertebrates (Stevens et al. 2007). Detailed experiments showed that large size and higher number of spots increase their protective value, irrespective of the level of resemblance to

Visual communication: evolution, ecology, and functional mechanisms

15

real eyes. Conspicuous patches of different shapes serve to camouflage prey probably by diverting the attention away from the shape of the body (Stevens et al. 2008b). Thus, neuronal perceptual mechanisms, rather than eye mimicry, seem to explain the design of wing and fin spots.

1.5 Selection by higher neuronal processes The higher neuronal circuitry involves cognitive processes which are often unknown and difficult to study in non-verbal animals (Chittka and Brockmann 2005). Unravelling the evolutionary significance of higher brain processes presents a particularly exciting research topic. These processes can influence selection upon colour patterns through various mechanisms. These include perceptual errors, shifts in preferences for stimuli that are caused by discrimination learning (peak-shift phenomena), attention limited by the rate of information processing, speed-accuracy trade-offs, and the retrieval and devaluation rate of memory properties (Dukas 2002, ten Cate and Rowe 2007, Chittka et al. 2009). The retrieval function of the memory is best studied in the context of foraging. For example, bees choose among floral colours depending on their experience with these colours 24 hours ago. The retrieval of colour memories in the honeybee thus follows a circadian rhythm that apparently matches the circadian rhythm of nectar and pollen production in flowers (Zhang et al. 2006, Prabhu and Cheng 2008). Bees thus possess a Zeitgedächtnis (time memory) allowing them to adaptively fine-tune their foraging behaviour to the peak availability of nectar associated with visual stimuli (see also Pahl et al. this volume). Associative learning can influence colour selection also at different time scales. Similar aged garden warblers (Sylvia borin) that were either handraised or caught as immatures in the wild differed in their ability to assess sugar rewards in foods independent of colour stimuli. Birds of both groups initially preferred red food over orange food. Although all birds were kept on the same diet for a period of seven months, wild-caught birds achieved higher sugar intakes when confronted with alternative foods of distinct sugar rewards whose colouration switched between experiments (Schaefer et al. 2008c). The higher nutrient intake occurred because wild-caught birds devaluated the colour information of preceding experiments more quickly. These experiments thus document that previous experience has long-lasting effects on food choice even in simple foraging situations where foods only vary in the single dimension of sugar contents. The examples of bees (above) and garden warblers both document that a better

16

H. Martin Schaefer

understanding of associative learning is needed for evaluating how animals respond to visual stimuli. In the following, I will examine two models of signal evolution that allow for a better understanding of the staggering diversity of visual signals that is inherent to almost every communication system from flowering plants on an Alpine meadow to the visual (and acoustical) diversification of epigamic signals in an avian rainforest community. Doing so allows us to amend the primarily evolutionary concept of signal theory with the ecology of the species involved.

1.6 Sensory trade-offs and sensory exploitation An exciting line of research suggests that the diversity of signals can, at least partly, be explained by distinct signalling strategies. In general, traits that convey information consist of a design (or efficacy) and a content component (Guilford and Dawkins 1991). The design component describes the proximate, sensory aspects of a trait discussed above, whereas a trait’s content describes the information (message) that it conveys (or not). Thus, signals could feasibly diverge to either maximise their conspicuousness (design component) or their reliability (content component). Modelling signal evolution in a binary sender receiver game where signals differed in efficacy and reliability yielded that – in theory – this trade-off could explain the diversity of signals because the efficacy component will vary according to environmental conditions (Schluter and Price 1993). Studying the plumage characteristics in three bird species, Andersson (2000) suggested that such trade-offs indeed occur in avian plumage. There is also supporting evidence from other signalling systems. Closely related toxic poison frogs differ in their warning colouration. Increasing either conspicuousness or toxicity affords equivalent avoidance by predators; suggesting that such alternative strategies of warning displays can explain the diversity of warning colouration (Darst and Cummings 2006). Similarly, the colours of fleshy fruits in a Venezuelan rainforest diverged along the dichotomy of either content- or efficacy-related colours. The most contrasting fruit colour (red), did not indicate the contents of fruit pulp, whereas chromatic variation in other fruit colours indicated the sugar, protein, and tannin contents of fruits (Schaefer and Schmidt 2004, Schmidt et al. 2004). An intriguing trade-off between relative detectability and relative nutritional returns also occurs in red and black fruit colours. Both colours are imparted by anthocyanins as plant pigments, and many fruits change pig-

Visual communication: evolution, ecology, and functional mechanisms

17

mentation from red to black during the ripening process owing to a quantitative change in anthocyanin contents (e.g., blackberries; see BOX 1.1). During fruit ripening, the appearance of red pigmentation is associated with low contents of anthocyanins, whereas the black pigmentation of ripe fruits is imparted by further accumulation of anthocyanins. The ripening of fruits entails thus a shift from a chromatic, red cue to a predominantly achromatic black one in the ripe fruit. This shift is important because it has dramatic consequences for the signal-noise ratio of fruits. As explained earlier, the background of most terrestrial scenes is highly heterogenic in brightness. Against such a background, it is very difficult to detect a primarily achromatic target, simply because it is concealed among background noise. These theoretical considerations predict that ripe fruits, which are made to be eaten by seed dispersers, are actually more difficult to detect than pre-ripe fruits. Indeed, fish crows (Corvus ossifragus) detected red fruits from a larger distance (60% increase) compared with similar sized black fruits (Schaefer et al. 2006). Given that birds strongly prefer feeding on the ripest fruits available (Schaefer and Schaefer 2006) delaying the maturation of red midripe fruits could be adaptive for plants because it increases the detectability of the entire fruit display. This is a strategy similar to the retention of pollinated flowers that increase the overall attractiveness of the plant for pollinators (Weiss 1991). The relatively lower detectability of ripe fruits is thus not explicable by a communicative function but rather by the defensive properties of anthocyanins in ripe fruits (see BOX 1.1). How can we envision the evolutionary stability of efficacy-related signals that provide little information? One important model is the sensory exploitation model that has been developed in the context of sexual selection, but that is equally applicable to other communication systems, such as plant-animal communication (Schaefer and Ruxton 2009). This model posits i) that female preferences for traits are explicable because certain traits are more effective in stimulating the female sensory systems, and ii) that traits are more likely to evolve that match pre-existing biases in the females sensory system (Basolo 1990, Ryan et al. 1990, Endler and Basolo 1998). Natural selection upon sensory systems maintains sensory biases. For example, four species of avian seed dispersers and pollinators have inherent preferences for contrasting fruit or flower displays, respectively, even if these are presented at close distance to the animals (Schmidt et al. 2004, Naug and Arathi 2007). However, it is not only the magnitude of contrast that influences foraging decisions. In a series of elegant experiments Lunau et al. (1996) demonstrated that the magnitude of contrasts explained the initial approach to artificial flowers (i.e., detection), but that

18 a

H. Martin Schaefer b

c

Fig. 1.3 Many orchids that do not provide nutritional rewards to pollinators are polymorphic in colour. The rare white morph of the unrewarding orchid Orchis mascula (a) is very attractive to pollinators. Interestingly, if white table tennis balls are positioned among the common purple morph (b), they are as efficient in attracting pollinators as the rare white morph even though they do not resemble flowers. The attraction of insects to white table tennis balls (c) provides a particular convincing example of how senders might exploit sensory biases of receivers (here pollinators) to forage on contrasting targets (Dormont et al. 2009; photos © Bertrand Schatz).

the actual floral colour influenced the decision whether or not to land on a given flower (i.e., a preference). Preferences for contrasting food sources are probably adaptive for pollinators and seed dispersers because fruits and flowers differ from their predominantly leafy green background in colour, scent and shape. Reacting to colour contrasts thus presents one important proximate mechanism for animals to locate and identify food resources. However, these preferences for contrasting colours can also be exploited by signallers. For example, permanently unrewarding flowers are often highly contrasting and thereby attractive for pollinators (Fig. 1.3). Indeed, the floral displays of nectarless flowers are often described as flamboyant, but the hypothesis that unrewarding flowers are characterised by more contrasting pigmentation than rewarding flowers has apparently not been tested. Thus, using an efficacyrelated signalling strategy may be evolutionarily stable because it exploits inherent biases of the sensory system of the receivers. Senders can exploit sensory biases of other animals or biases that occur when the sensory input is processed in the brain of an animal. The latter are thus biases that arise during information processing. There are several related models on signal evolution that can be distinguished according to their relative emphasis on the exploitation of sensory vs. higher brain proc-

Visual communication: evolution, ecology, and functional mechanisms

19

esses (Endler and Basolo 1998). Regardless of the origin of a bias, models on the exploitation of pre-existing biases allow to predict why certain signals evolve if they include an explicit phylogenetic approach. They can thus predict the sequence of events in signal evolution. For example, biases for larger size in mates are widespread in the animal kingdom potentially driving the evolution of larger-sized ornaments (Ryan and Keddy-Hector 1992, Rosenthal and Evans 1998). A prerequisite for the pre-existing biases model is thus to show that the preference for a trait is ancestral occurring in species lacking that trait, and that the trait is developed in the predicted direction in related species. This has been demonstrated in systems as diverse as acoustic communication in frogs and visual inspection of morphological traits in swordtail fishes (Xiphorus spp.) and short-tailed widowbirds (Euplectes axiliaris) (Basolo 1990, Ryan et al. 1990, Pryke and Andersson 2002). Thus, pre-existing biases are a powerful mechanism that may explain the evolution of acoustic and visual signals.

1.7 Sensory drive The sensory drive model is an ecological model to explain the diversity of signals. This model predicts that the sensory system of animals, their signals and signalling behaviour, and habitat choice have all been coupled over evolutionary times (Endler 1992). Since environmental differences affect the transmittance and perception of signals (the efficacy component), this model predicts geographical differences in signal perception and signalling behaviour. Elegant tests of how sensory constraints associated with predictable environmental differences drive signal evolution, and even speciation, are found in aquatic environments. In five closely related dichromatic fish species, habitat choice entails an inherent trade-off in colour and luminance detection that leads to differences in the spectral tuning of these fish. The trade-off is associated with water depth but primarily driven by variation in background colour and brightness. Species living in habitats with high variance in luminance evolved sensory systems and signals that favour chromatic detection (Cummings 2007). Sensory drive can result in strong divergent selection. For example, in Lake Victoria cichlid species, divergent selection along an environmental gradient of water clarity and depth-mediated light gradients drives divergent evolution of the visual system (opsin genes). In the lake there are related cichlid species pairs where males differ in having either a red or blue nuptial colouration. Although these fishes are geographically sympatric, they are better thought of as parapatric with adjoining depth ranges in the

20

H. Martin Schaefer

lake. The red species always occurs at greater mean water depth. Divergent selection upon the visual systems of these fishes evidenced by higher genetic variation at the opsin gene loci compared to that of neutral loci is leading to incipient speciation (Terai et al. 2006). The contrasting selection regimes lead to divergence in male breeding colouration and female preferences in this system, but also in sticklebacks (Boughman 2001). Importantly, several populations of cichlid species along the steep environmental gradient of water clarity are reproductively isolated, showing that sensory drive may lead to speciation even in the absence of strict geographic isolation (Terai et al. 2006, Seehausen et al. 2008). Evidence for sensory drive is also found in terrestrial communication systems. Mesic and xeric habitats differ in the quality of ambient light because they differ in the amount of light filtered by the vegetation (Endler 1993). Allopatric Anolis lizard populations from mesic and xeric habitats differ in dewlap colouration, a sexually selected signal, so that each population sports a colour that increases detectability in its specific habitat (Leal and Fleishman 2004). Environmental differences in the conspicuousness of an epigamic trait might likewise explain unidirectional introgression in the hybrid zone of golden-collared (Manacus vitellinus) and whitecollared (M. candei) manakins. Each species’ collar is better detectable in areas where both are allopatric, whereas in the hybrid zone the conspicuousness of the golden collar exceeds that of the white collar. These results suggest that sensory mechanisms can explain the unidirectional spread of yellow plumage across the hybrid zone as well as its slowed movement beyond it (Uy and Stein 2007).

1.8 Environmental influence on communication The examples above provide powerful evidence for how environmental differences may drive visual perception, signal evolution and, ultimately, speciation. Abiotic factors can accelerate population differentiation and sensory drive because pigments often fulfil dual roles in thermoregulation and communication. Visual traits can therefore be particularly prone to respond to environmental differences, which may partly explain the diversification of visual traits. Environmental determination of colour traits occurs in many communication systems, for example, avian plumage traits that are involved in mate choice (Hadfield et al. 2007, Norris et al. 2007). A large cross-fostering experiment by Hadfield and colleagues uncoupled genetic from environmental effects because some blue tit (Cyanistes caeruleus) chicks of each

Visual communication: evolution, ecology, and functional mechanisms

21

family were raised by their parents and others by foster parents. Heritability of the sexually selected ultraviolet/blue cap of the species and of the yellow, carotenoid-based breast feathers was low (h2 = 0.10 and 0.07, respectively). None of the traits indicated components of offspring fitness. This is an important study for evaluating alternative models of sexual selection because it shows that female blue tits were unlikely to obtain indirect benefits (through high quality offspring) by choosing among males based upon the colouration of these plumage traits. In the good gene model, females obtain indirect genetic benefits for their offspring by mating with the most ornamented male because ornaments can be costly handicaps and thereby indicate the genetic quality of their bearer (Zahavi 1975). The good gene model is often contrasted with models on direct benefits that females may accrue by mating with a partner that invests more in parental duties. The low heritability of and strong environmental influence on plumage traits in the blue tit do not support indirect models of sexual selection (Hadfield et al. 2006). Thus, these studies raise fundamental questions about the reliability of information transfer in communication and, more generally, about the common usage of inferring genetic patterns from phenotypic data. I will now examine two case studies of plant-animal interactions. These are informative for two reasons. The first study shows that environmental determination of visual traits does not necessarily compromise their reliability. The second study nicely summarises why the textbook wisdom that floral colours are an adaptation to communicating to biotic dispersal agents does not fully appreciate the evolutionary trajectories of communication. The colouration of the bracts in the infructescences of black elder (Sambucus nigra) is a good example of an environmentally driven communication system (Fig. 1.4). In forest gaps, individuals have red bracts, whereas the bracts remain green in individuals growing in forest shade. Bract colouration is a phenotypically plastic trait because individuals can sport both colour phenotypes if they receive direct sunlight on only one side. Red colouration often develops in plant tissue as a protection against excess sun light. Since individuals in gaps receive more light, they also produce sweeter fruits, forging a link between bract colouration and fruit quality. In this communication system, fruit quality is quantitatively indicated by the colour contrasts between bracts and the background; the higher the contrasts, the sweeter the fruits. Hence, red pedicels concomitantly increase the detectability and the reliability of information exchange in black elder. Blackcaps attend to bract colouration when consuming fruits and prefer fruits from red bracts. Yet, since bract colouration is an environmentally determined trait, it should be regarded as a cue until there is evidence

22

H. Martin Schaefer

Fig. 1.4 Fruit-eating blackcaps preferentially consume black elder fruits from infructescences with red pedicels (illustrated by the arrow) because red pedicel pigmentation is reliably associated with sweeter fruits.

that the selective pressure of seed dispersing birds overrides that of excess irradiance (Schaefer and Braun 2009). This study shows that environmental coupling of the design of a visual trait and of the content it indicates can lead to reliable, environmentally determined information transfer. Differences in flower colour are traditionally interpreted as an adaptation towards divergent selection by pollinators; an idea that can be traced to the late 18th century (Sprengel 1793). Spatial differentiation in the blue and white colour morph of Linanthus parryae has become a model system in evolutionary biology. The colour polymorphism in this species is caused by a single gene. The sole pollinator, a melyrid beetle, does apparently not differentiate between colour morphs. Applying his model of isolation by distance, Wright (1943) concluded that spatial differentiation in L. parryae was caused by random drift. Schemske and Bierzychudek (2007) studied

Visual communication: evolution, ecology, and functional mechanisms

23

genetic differentiation in this species. They found no differentiation in allozyme markers in populations with distinct floral colour. They therefore concluded that spatial differentiation in floral colour is explicable by natural selection, a conclusion supported because each morph fared best in transplant experiments in their local habitats. Thus, the colour polymorphism in this species is explicable by pleiotropy, i.e. by the multiple effects that a single gene has on distinct phenotypes (colour and local performance) and not by divergent selection upon communication to pollinators.

1.9 Conclusion Throughout the chapter, I have stressed the need to account for the functional mechanisms underlying visual communication in order to understand the evolutionary trajectories of visual communication. Today, researchers can build upon a quickly expanding knowledge on the sensory processes of signal reception that form the basis for the models of sensory exploitation and sensory drive. Both of these models on signal evolution have emphasised the role of sensory biases and have led to the increasing use of visual modelling in evolutionary and ecological studies. Incorporating the evolutionary significance of higher neuronal processes into current models on signal evolution represents one of the most important current challenges for a functional understanding of communication. Another current frontier is to address how the often significant environmental influence on visual traits alters their design and information content. Substantial environmental influence on signalling can destabilise communication systems that transmit information on the genetic quality of the signaller. A good example of such a communication system is provided by the model of indirect genetic benefits in sexual selection theory that underlies the paradigm of the handicap principle. In other communication systems, however, the linkage between phenotypic visual trait and phenotypic quality may be largely driven by the environment. Future studies are needed to test this conjecture as well as to assess how widely the handicap principle explains the evolutionary stability of communication systems.

Acknowledgements I thank Martin Stevens, Lars Chittka and Peter Kappeler for many very constructive comments and Bertrand Schatz for permission to use his photos in Fig. 1.3. All other photos © Veronika Schaefer and Martin Schaefer.

24

H. Martin Schaefer

References Andersson S (2000) Efficacy and content in avian colour signals. In: Espmark Y, Amundsen T, Rosenqvist G (eds) Animal Signals: Signalling and Signal Design in Animal Communication. Tapir Academic Press, Trondheim, Norway, pp 47-60 Andersson S, Pryke SR, Örnborg J, Lawes MJ, Andersson M (2002) Multiple receivers, multiple ornaments, and a trade-off between agonistic and epigamic signaling in a widowbird. Am Nat 160:683-691 Basolo AL (1990) Female preference predates the evolution of the sword in swordtail fish. Science 250:808-810 Boughman JW (2001) Divergent sexual selection enhances reproductive isolation in sticklebacks. Nature 411:944-948 Bradshaw HD Jr, Schemske DW (2003) Allele substitution at a flower colour locus produces a pollinator shift in monkeyflowers. Nature 426:176-178 Catoni C, Schaefer HM, Peters A (2008) Fruit for health: the effect of flavonoids on humoral immune response and food selection in a frugivorous bird. Funct Ecol 22:649-654 Cazetta E, Schaefer HM, Galetti M (2009) Why are fruits colorful? The relative importance of achromatic and chromatic contrasts for detection by birds. Evol Ecol 23:233-244 Chittka L, Brockmann A (2005) Perception space – the final frontier. PLOS Biol 3:e137, doi:10.1371/journal.pbio.0030137 Chittka L, Skorupski P, Raine NE (2009) Speed-accuracy tradeoffs in animal decision making. Trends Ecol Evol 24:400-407 Cipollini ML, Stiles EW (1993) Fruit rot, antifungal defense, and palatability of fleshy fruits for frugivorous birds. Ecology 74:751-762 Cott HB (1940) Adaptive Coloration in Animals. Methuen, London Cummings ME (2007) Sensory trade-offs predict signal divergence in surfperch. Evolution 61:530-545 Cummings ME, Rosenthal GG, Ryan MJ (2003) A private ultraviolet channel in visual communication. Proc R Soc Lond B 270:897-904 Cuthill IC, Stevens M, Sheppard J, Maddocks T, Párraga CA, Troscianko TS (2005) Disruptive coloration and background pattern matching. Nature 434:72-74 Darst CR, Cummings ME (2006) Predator learning favours mimicry of a lesstoxic model in poison frogs. Nature 440:208-211 Darst CR, Cummings ME, Cannatella DC (2006) A mechanism for diversity in warning signals: conspicuousness versus toxicity in poison frogs. Proc Natl Acad Sci USA 103:5852-5857 Dawkins R, Krebs JR (1978) Animal signals: information or manipulation. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 282-309

Visual communication: evolution, ecology, and functional mechanisms

25

Dormont L, Delle-Vedove R, Bessière JM, Hossaert-McKey M, Schatz B (2009) Rare white-flowered morphs increase the reproductive success of common purple morphs in a food-deceptive orchid. New Phytol 185:300-310 Dukas R (2002) Behavioural and ecological consequences of limited attention. Philos Trans R Soc Lond B 357:1539-1547 Endler JA (1980) Natural selection on color patterns in Poecilia reticulata. Evolution 34:76-91 Endler JA (1992) Signals, signal conditions, and the direction of evolution. Am Nat 139:S125-S153 Endler JA (1993) The color of light in forests and its implications. Ecol Monogr 63:1-27 Endler JA, Basolo AL (1998) Sensory ecology, receiver biases and sexual selection. Trends Ecol Evol 13:415-420 Endler JA, Westcott DA, Madden JR, Robson T (2005) Animal visual systems and the evolution of color patterns: sensory processing illuminates signal evolution. Evolution 59:1795-1818 Guilford T, Dawkins MS (1991) Receiver psychology and the evolution of animal signals. Anim Behav 42:1-14 Hadfield JD, Burgess MD, Lord A, Phillimore AB, Clegg SM, Owens IPF (2006) Direct versus indirect sexual selection: genetic basis of colour, size and recruitment in a wild bird. Proc R Soc Lond B 273:1347-1353 Hadfield JD, Nutall A, Osorio D, Owens IPF (2007) Testing the phenotypic gambit: phenotypic, genetic and environmental correlations of colour. J Evol Biol 20:549-557 Håstad O, Victorsson J, Ödeen A (2005) Differences in color vision make passerines less conspicuous in the eyes of their predators. Proc Natl Acad Sci USA 102:6391-6394 Heiling AM, Herberstein ME, Chittka L (2003) Crab-spiders manipulate flower signals. Nature 421:334 Janzen DH (1977) Why fruits rot, seeds mold, and meat spoils. Am Nat 111:691713 Jones CD, Osorio D, Baddeley RJ (2001) Colour categorization by domestic chicks. Proc R Soc Lond B 268:2077-2084 Kelber A, Vorobyev M, Osorio D (2003) Animal colour vision – behavioural tests and physiological concepts. Biol Rev 78:81-118 Kong J-M, Chia L-S, Goh N-K, Chia T-F, Brouillard R (2003) Analysis and biological activities of anthocyanins. Phytochemistry 64:923-933 Leal M, Fleishman LJ (2004) Differences in visual signal design and detectability between allopatric populations of Anolis lizards. Am Nat 163:26-39 Lunau K, Wacht S, Chittka L (1996) Colour choices of naive bumble bees and their implications for colour perception. J Comp Physiol A 178:477-489 Marshall J, Cronin TW, Kleinlogel S (2007) Stomatopod eye structure and function: a review. Arthropod Struct Dev 36:420-448 Maynard Smith J, Harper DGC (1995) Animal signals: models and terminology. J Theor Biol 177:305-311

26

H. Martin Schaefer

Merilaita S (1998) Crypsis through disruptive coloration in an isopod. Proc R Soc Lond B 265:1059-1064 Naug D, Arathi HS (2007) Receiver bias for exaggerated signals in honeybees and its implications for the evolution of floral displays. Biol Lett 3:635-637 Norris DR, Marra PP, Kyser TK, Ratcliffe LM, Montgomerie RA (2007) Continent-wide variation in feather colour of a migratory songbird in relation to body condition and moulting locality. Biol Lett 3:16-19 Ödeen A, Hảstad O (2003) Complex distribution of avian color vision systems revealed by sequencing the SWS1 opsin from total DNA. Mol Biol Evol 20:855-861 Osorio D, Srinivasan MV (1991) Camouflage by edge enhancement in animal coloration patterns and its implications for visual mechanisms. Proc R Soc Lond B 244:81-85 Osorio D, Miklósi A, Gonda Z (1999) Visual ecology and perception of coloration patterns by domestic chicks. Evol Ecol 13:673-689 Otte D (1974) Effects and functions in the evolution of signaling systems. Annu Rev Ecol Syst 5:385-417 Prabhu C, Cheng K (2008) One day is all it takes: circadian modulation of the retrieval of colour memories in honeybees. Behav Ecol Sociobiol 63:11-22 Pryke SR, Andersson S (2002) A generalized female bias for long tails in a shorttailed widowbird. Proc R Soc Lond B 269:2141-2146 Regan BC, Julliot C, Simmen B, Viénot F, Charles-Dominique P, Mollon JD (2001) Fruits, foliage and the evolution of primate colour vision. Philos Trans R Soc Lond B 356:229-283 Rosenthal GG, Evans CS (1998) Female preference for swords in Xiphophorus helleri reflects a bias for large apparent size. Proc Natl Acad Sci USA 95:4431-4436 Ryan MJ, Keddy-Hector A (1992) Directional patterns of female mate choice and the role of sensory biases. Am Nat 139:S4-S35 Ryan MJ, Fox JH, Wilczynski W, Rand AS (1990) Sexual selection for sensory exploitation in the frog Physalaemus pustulosus. Nature 343:66-67 Schaefer HM, Braun J (2009) Reliable cues and signals of fruit quality are contingent on the habitat in black elder (Sambucus nigra). Ecology 90:1564-1573 Schaefer HM, Ruxton GD (2009) Deception in plants: mimicry or perceptual exploitation? Trends Ecol Evol 24:676-685 Schaefer HM, Schaefer V (2006) The fruits of selectivity: how birds forage on Goupia glabra fruits of different ripeness. J Ornithol 147:638-643 Schaefer HM, Schmidt V (2004) Detectability and content as opposing signal characteristics in fruits. Proc R Soc Lond B 271:S370-S373 Schaefer HM, Stobbe N (2006) Disruptive coloration provides camouflage independent of background matching. Proc R Soc Lond B 273:2427-2432 Schaefer HM, Levey DJ, Schaefer V, Avery ML (2006) The role of chromatic and achromatic signals for fruit detection by birds. Behav Ecol 17:784-789 Schaefer HM, Schaefer V, Vorobyev M (2007) Are fruit colors adapted to consumer vision and birds equally efficient in detecting colorful signals? Am Nat 169:S159-S169

Visual communication: evolution, ecology, and functional mechanisms

27

Schaefer HM, McGraw K, Catoni C (2008a) Birds use fruit colour as honest signal of dietary antioxidant rewards. Funct Ecol 22:303-310 Schaefer HM, Rentzsch M, Breuer M (2008b) Anthocyanins reduce fungal growth in fruits. Nat Prod Commun 3:1267-1272 Schaefer HM, Spitzer K, Bairlein F (2008c) Long-term effects of previous experience determine nutrient discrimination abilities in birds. Front Zool 5:4, doi:10.1186/1742-9994-5-4 Schemske DW, Bierzychudek P (2007) Spatial differentiation for flower color in the desert annual Linanthus parryae: was Wright right? Evolution 61:25282543 Schluter D, Price T (1993) Honesty, perception and population divergence in sexually selected traits. Proc R Soc Lond B 253:117-122 Schmidt V, Schaefer HM, Winkler H (2004) Conspicuousness, not colour as foraging cue in plant-animal interactions. Oikos 106:551-557 Seehausen O, Terai Y, Magalhaes IS, Carleton KL, Mrosso HDJ, Miyagi R, van der Sluijs I, Schneider MV, Maan ME, Tachida H, Imai H, Okada N (2008) Speciation through sensory drive in cichlid fish. Nature 455:620-626 Seeley TD (1989) The honey bee colony as a superorganism. Am Sci 77:546-553 Shannon CE, Weaver W (1949) The Mathematical Theory of Communication. University of Illinos Press, Urbana Skyrms B (2009) Evolution of signalling systems with multiple senders and receivers. Philos Trans R Soc Lond B 364:771-779 Sprengel CK (1793) Das entdeckte Geheimnis in der Natur im Bau und in der Befruchtung der Blumen. Weldon and Wesley, New York Stevens M (2007) Predator perception and the interrelation between different forms of protective coloration. Proc R Soc Lond B 274:1457-1464 Stevens M, Cuthill IC (2007) Hidden messages: are ultraviolet signals a special channel in avian communication? BioScience 57:501-507 Stevens M, Cuthill IC, Windsor AMM, Walker HJ (2006) Disruptive contrast in animal camouflage. Proc R Soc Lond B 273:2433-2438 Stevens M, Hopkins E, Hinde W, Adcock A, Connolly Y, Troscianko T, Cuthill IC (2007) Field experiments on the effectiveness of 'eyespots' as predator deterrents. Anim Behav 74:1215-1227 Stevens M, Graham J, Winney IS, Cantor A (2008a) Testing Thayer's hypothesis: can camouflage work by distraction? Biol Lett 4:648-650 Stevens M, Hardman CJ, Stubbins CL (2008b) Conspicuousness, not eye mimicry, makes ‘eyespots’ effective antipredator signals. Behav Ecol 19:525-531 Stinzig FC, Carle R (2004) Functional properties of anthocyanins and betalains in plants, food, and in human nutrition. Trends Food Sci Tech 15:19-38 Stobbe N, Schaefer HM (2008) Enhancement of chromatic contrast increases predation risk for striped butterflies. Proc R Soc Lond B 275:1535-1541 Stuart-Fox D, Moussalli A, Whiting MJ (2008) Predator-specific camouflage in chameleons. Biol Lett 4:326-329 Sumner P, Mollon JD (2000) Chromaticity as a signal of ripeness in fruits taken by primates. J Exp Biol 203:1987-2000

28

H. Martin Schaefer

ten Cate C, Rowe C (2007) Biases in signal evolution: learning makes a difference. Trends Ecol Evol 22:380-387 Terai Y, Seehausen O, Sasaki T, Takahashi K, Mizoiri S, Sugawara T, Sato T, Watanabe M, Konijnendijk N, Mrosso HDJ, Tachida H, Imai H, Shichida Y, Okada N (2006) Divergent selection on opsins drives incipient speciation in Lake Victoria cichlids. PLOS Biol 4:e433, doi:10.1371/journal.pbio.0040433 Touchon JC, Warkentin KM (2008) Fish and dragonfly nymph predators induce opposite shifts in color and morphology of tadpoles. Oikos 117:634-640 Uy JAC, Endler JA (2004) Modification of the visual background increases the conspicuousness of golden-collared manakin displays. Behav Ecol 15:10031010 Uy JAC, Stein AC (2007) Variable visual habitats may influence the spread of colourful plumage across an avian hybrid zone. J Evol Biol 20:1847-1858 Viitala J, Korpimäki E, Palokangas P, Koivula M (1995) Attraction of kestrels to vole scent marks visible in ultraviolet light. Nature 373:425-427 Weiss MR (1991) Floral colour changes as cues for pollinators. Nature 354:227229 Wiley RH (1994) Errors, exaggeration, and deception in animal communication. In: Real LA (ed) Behavioral Mechanisms in Evolutionary Ecology. University of Chicago Press, Chicago, pp 157-189 Wright S (1943) An analysis of local variability of flower color in Linanthus parryae. Genetics 28:139-156 Zahavi A (1975) Mate-selection – a selection for a handicap. J Theor Biol 53:204214 Zhang S, Schwarz S, Pahl M, Zhu H, Tautz J (2006) A honey bee knows what to do and when. J Exp Biol 209:4420-4428

Chapter 2

Vocal communication in social groups CLAUDIA FICHTEL AND MARTA MANSER

ABSTRACT Vocal communication plays a particularly important role in the regulation of social interactions and in the coordination of activities in many mammals and birds that are organised into social groups. Previous research on the function and evolution of vocal signals has mainly considered dyadic interactions of a signaller and its addressed receiver. However, in social groups it is likely that additional individuals attend to dyadic communication and that they use this information to their own benefit, sometimes at a cost to the signaller. To improve existing communication models, benefits and costs of vocal communication caused by bystanders must therefore also be considered. Here we discuss vocal communication in social groups and identify the effects of additional individuals on signalling interactions, concentrating on audience effects, eavesdropping and group coordination. First, a review of the existing literature reveals that the presence of an audience, i.e., additional individuals within the signalling range, clearly affects the outcome of communicative interactions, and that individuals modulate their signalling behaviour according to the presence of bystanders or a particular category of bystanders in a variety of contexts. Second, social knowledge acquired by eavesdropping on the communicative network within a group influences not only future actions, but can also provide individual benefits for eavesdroppers, whereas mutual eavesdropping can structure cooperation and alliance formation, and, hence, contribute to long-term group stability. Third, communicative networks also provide a means to facilitate the maintenance of group cohesion and decisionmaking processes. In conclusion, cost-benefit analyses at the level of dyadic interactions reveal clear differences with communication networks, where repeated interactions with multiple partners are considered. Future communication models and empirical studies should therefore consider the composition of the entire communication network as well as the effects of repeated interactions to fully understand signalling interactions in social groups.

30

Claudia Fichtel and Marta Manser

2.1. Introduction Communication is of central importance in the life of social animals. Due to their far-reaching effects on social relationships, survival and reproduction, communication systems are among the key research models to understand biological processes at the ultimate and proximate level. Animals communicate with each other using different sensory modalities (Bradbury and Vehrencamp 1998, Schaefer this volume, Penn and Frommen this volume); here we focus on acoustic signals. Communication in the broadest sense is defined as the interaction between a sender delivering a signal to a receiver attending to it and subsequently changing its behaviour (Bradbury and Vehrencamp 1998, Maynard Smith and Harper 2003). The three elements involved in communication – sender, signal and receiver – are connected by an optimising principle to make the signal as efficient as possible (Smith 1981, Bradbury and Vehrencamp 1998). Signals are ‘an act or structure that alters the behaviour of other organisms, because of that effect, and which is effective because the receiver’s response has also evolved’ (Maynard Smith and Harper 2003). In contrast, cues are ‘a feature of the world, animate or inanimate, that can be used by an animal as a guide to future action’ (Maynard Smith and Harper 2003). They are a by-product of an animal’s behaviour, such as body postures or noises inadvertently produced while eating, fleeing or mating, from which other individuals may in turn gain information about the specific context or traits of the sender. Cues are also given in the absence of receivers and are not selected for communication because they are inadvertently provided (Danchin et al. 2004). Both, signals and cues, provide information for recipients. Signal transmission has traditionally been viewed as a dyadic interaction between a sender and a receiver (Bradbury and Vehrencamp 1998). However, in social groups where multiple potential senders and receivers are present, communication occurs within a network consisting of many individuals (McGregor and Dabelsteen 1996, McGregor 2005). The fact that information is accessible to a broad audience (public information) allows individuals not only to gather information while observing communication between other individuals (social information) but also to acquire novel information by social learning (see van Schaik this volume). Social information can be gained by eavesdropping: an act in which third-party bystanders extract information from signalling interactions of others (McGregor and Dabelsteen 1996, McGregor 2005). Although the acquisition of social information is faster and achieved at lower costs than acquiring this information through individual experience, it bears the risk of ac-

Vocal communication in social groups

31

BOX 2.1 Definitions and terminology Audience or bystander: individual that is present, but does not take actively part in communicative interactions between others Audience effect: changes in the signalling behaviour caused by the presence of other individuals, of which the signaller is aware Communication: interaction between an individual (sender) delivering a signal to another individual (receiver) using it to make a decision Communication network: a set of interacting individuals within the signal range Cue: a feature of the world, animate or inanimate, that can be used by an animal as a guide to future action; not selected for communication Eavesdropping: the act of extracting information from signalling interactions between other individuals Information: property of a source – entity or a process – eliciting a reaction of a receiver in a biologically functional manner Interceptive eavesdropping: an individual benefits by intercepting signals intended for another individual, usually at a cost to a signaller Receiver: individual perceiving a signal or cue and likely altering its behaviour because of that Sender: individual producing a signal or a cue Signal: any act or structure that alters the behaviour of other organisms, which has been selected for communication Social eavesdropping: when individuals gather information on other individuals by attending to their signalling interaction (McGregor and Dabelsteen 1996, Bradbury and Vehrencamp 1998, Jablonka 2002, Maynard Smith and Harper 2003)

quiring false information (Laland 2004, Danchin et al. 2004, Dabelsteen 2005, Bonnie and Earley 2007). In addition, communication in stable social networks includes the potential for individuals to integrate information from past interactions into future actions, which in turn may have consequences for the costs and benefits of any given signalling action. Moreover, in stable social groups communication appears to facilitate the coordination of activities to achieve tasks that would not be possible were they to act alone (Sumpter and Brännström 2008). Acoustic communication is of special interest in studying communication networks because acoustic signals are typically loud, conspicuous and travel over long range and can hence be perceived by more individuals than the intended receiver. The production of acoustic signals is to a degree flexible and can be adjusted to the audience. Acoustic signals are used in a variety of contexts, including territorial defence, mating, parent-

32

Claudia Fichtel and Marta Manser

offspring interactions, predation, regulation of social interaction and coordination of group movements (Bradbury and Vehrencamp 1998); all contexts in which bystanders may gain important information for their future actions. Acoustic signals may convey basic information about the sender, such as its individual and group identity, age, sex, size, condition and motivational state, i.e. its tendency to react in a specific way (Hauser 1996, Bradbury and Vehrencamp 1998). Most of the time specific call types also seem to be related to the sender’s external context (reviewed in Manser 2009), such as contact calls while foraging, or travelling calls while moving (Boinski 1991). Moreover, some signals refer to specific attributes of objects or events in the external environment of the caller, as shown for food recruitment calls (Hauser 1998, Bugnyar et al. 2001, DiBitetti 2003, Gros-Louis 2004a) and functionally referential alarm calls in several primates (Seyfarth et al. 1980, Pereira and Macedonia 1991, Zuberbühler et al. 1999, Fichtel and Kappeler 2002), other mammals (Owings and Leger 1980, Slobodchikoff et al. 1991, Manser 2001, Manser et al. 2001) and birds (Evans et al. 1994, Griesser 2008). These signals function in a variety of ways to affect the behaviour of receivers. In territorial defence or mate attraction, signals are addressed at a broad class of receivers and may function to repel competitors and/or attract mates, respectively (Reby et al. 1999, Kitchen 2004, Catchpole and Slater 2008). In these situations, signal variation might be associated with the quality of the sender and signals from particular individuals are considered to be honest (Zahavi 1977, Zahavi and Zahavi 1997). Signals may also reflect variation in the affective state of the sender (Morton 1977, Marler 1992, Fichtel et al. 2001), and they may function to communicate what the sender is going to do next. In this chapter, we emphasise the role of communication networks in species living in stable social groups where individuals have many opportunities to monitor social interactions of several conspecifics simultaneously and repeatedly. In such groups, acoustic signals are usually also received by several individuals other than the addressed recipient, which in turn may have a direct impact on signalling interactions or their outcome. Moreover, communication interactions may have delayed and long-lasting influences on individual behaviour. We therefore focus on two related aspects; namely how the presence of other individuals influences the production of signals, and how additional receivers filter out the important information when they are not primarily addressed. Identifying the potential costs and benefits of the presence or influence of additional individuals as audience or eavesdroppers of a communicative act will help to illuminate to what extent these boundary conditions influence signalling behaviour. To this end, we first examine examples in which network members influ-

Vocal communication in social groups

33

ence the outcome of a signalling interaction (audience effect) and subsequently focus on situations in which others gain information by eavesdropping (social and interceptive eavesdropping). Third, we illustrate the use of signals that are explicitly addressed at several recipients and may facilitate group cohesion or decision-making processes. Finally, we discuss cost-benefit analyses of vocal communication within networks by the present models and suggest future directions for research that will enhance our understanding of animal societies in general.

2.2 Audience effects With the production of a signal, information is available to all individuals present within the signal range, and signallers often behave differently, depending on the presence and identity of others (Marler et al. 1986). This adjustment of behaviour, the so-called audience effect, requires that the caller is aware of the presence of specific receivers, and it indicates that signals are produced strategically, rather than involuntarily. Audience effects have first been described in chickens (Gallus gallus) emitting alarm and food calls. Cocks preferably emit food calls in the presence of hens, sometimes even when no food is available (Marler et al. 1986). Furthermore, both sexes emit terrestrial alarm calls, but males only emit aerial alarm calls when conspecifics are close by (Karakashian et al. 1988). A similar audience effect on callers has since been described in several other species (see below). Principally, we can distinguish between contexts, in which simply the presence of a receiver triggers calling, and contexts, in which bystanders attend to interactions between signallers and receivers (eavesdrop), and may gather information to make decisions about future actions. In contexts, in which the targeted receivers also act as audience, different categories have been identified. Signallers have been described to be affected in their call production by: i) the mere presence of a receiver; ii) the presence of a specific receiver category; or iii) the specific context the receiver finds itself in. The comparison between situations in which a signaller either emits a call when a receiver is present or not allows to test whether individuals have voluntary control over call production, or whether call production is an involuntary response to a specific context (Evans 1997). This question has mainly been investigated in the context of the production of alarm calls. Calls that are typically directed towards a predator and/or conspecifics, so-called mobbing calls, are emitted independent of whether an audience of conspecifics is nearby or not, whereas

34

Claudia Fichtel and Marta Manser

alarm calls emitted to warn group members are typically produced if a conspecific is nearby (Thomas langurs, Presbytis thomasis: Wich and Sterck 2003; yellow mongoose, Cynictis penicillata: Le Roux et al. 2008). Audience effects have also been described for interactions between species, where fork-tailed drongos (Dicrurus adsimilis) adjust their alarm call production when foraging with a group of pied babblers (Turdoides bicolour: Ridley et al. 2007). Not only the mere presence of a receiver influences the calling behaviour of animals, but also the identity or the behaviour of the audience. For example, in chickens, cocks adjust their production of alarm and food calls depending on the sex of conspecifics present (Marler et al. 1986). In Belding’s ground squirrels (Spermophilus belding), alarm calls are more likely emitted when close kin is nearby (Sherman 1977). In capuchin monkeys (Cebus capucinus), individuals finding food are more likely to emit calls when high-ranking individuals are close by than low-ranking ones (GrosLouis 2004b). Rumble vocalisations of elephants (Loxodonta africana) cause a decrease in the distance between individuals, and this effect is enhanced among highly affiliated social partners (Leighty et al. 2008). In banded mongoose (Mungos mungo), pups adjust their calling rate to the identity of their targeted providers (Bell 2008). Two recent studies suggested that callers also take the context or behaviour of their addressed audience into account and adjust calling accordingly. In blue monkeys (Cercopithecus mitis) males appear to take the level of danger experienced by the other group members into account when emitting alarm calls (Papworth et al. 2008). Similarly, Thomas langurs do not stop giving alarm calls when threatened by a predator until every single group members has responded with an alarm call, suggesting that they keep track of who has responded and who has not (Wich and de Vries 2006). Audience effects in triadic contexts involve the sender, a receiver and a bystander. In chimpanzees (Pan troglodytes), victims of aggression tend to modify the acoustic structure of their screams to exaggerate the situation, but only in the presence of individuals that are capable of intervening and supporting the caller (Slocombe and Zuberbühler 2007). Moreover, female chimpanzees are more likely to produce copulation calls in the presence of high-ranking males, but they are less likely to call when high-ranking females are nearby, suggesting that adjustment of calling rate may serve to reduce competition among females (Townsend et al. 2008). Avoidance of potential conflicts by adjusting call rate has also been suggested for nomadic male lions (Panthera leo), who benefit by roaring to enhance social ties with potential cooperators. However, in the presence of resident males they call less often, thereby concealing their own presence to avoid poten-

Vocal communication in social groups

35

tial costs of escalated conflicts by eavesdropping on resident males (Grinnell and McComb 2001). Although most studies on audience effects have dealt with the production of alarm and food calls, and some have focussed on aggressive interactions, the examples above show that these effects can occur in almost any context of communication. Furthermore, some recent studies revealed that not only the production of the signals, but also the response of receivers to these signals is affected by the presence of additional receivers. In rhesus macaques (Macaca mulatta), the outcome of mother-infant interactions was influenced by the potential risk of aggression posed by bystanders (see BOX 2.2, Semple et al. 2009). The likelihood of mothers allowing crying infants access to the nipple was higher in the presence of high-risk bystanders than during control situations or in the presence of low-risk bystanders. Finally, the presence of additional signallers can also influence a communicative act. The begging behaviour of nestlings in several bird species (Wright and Leonard 2002) and that of mammalian pups has been shown to be affected by siblings that either directly compete with each other and cause an increase in begging intensity (Price et al. 1996), or that cooperate and decrease begging intensity when others are present (Bell 2007). Therefore, in communication networks with more than just a single actor and reactor, both the production of signals and the response to them can be affected by presence of additional individuals. Audience effects on signal interactions have presumably evolved depending on the benefits and costs imposed on the signaller or intended receivers by the presence of bystanders. The examples above indicate that by taking the presence of bystanders, or of a particular category of bystanders, into account, individuals involved in signal interactions can benefit by either enhancing or inhibiting the production of signals. Contexts for enhancing signal production include the attraction of mates (Marler et al. 1986) or recruitment to food sources (Gros-Louis 2004b). Contexts for inhibiting signal production include aggressive interactions either within a group (e.g., infant crying in rhesus macaques), between groups (e.g., lions avoiding the attention of resident males), but also between species (e.g., killer whales, Orcinus orca, reducing vocal activity while hunting, see below). This perspective clearly shows the flexibility of animals in the production of acoustic signals, but also the necessity to consider communication within the existing social network, and not only as the sum of dyadic interactions, to understand the variation of signaller’s and receiver’s behaviour in similar contexts.

Claudia Fichtel and Marta Manser BOX 2.2 Audience affects the outcome of mother-infant interactions in rhesus macaques Research question: Infants in rhesus macaques (Macaca mulatta) cry to get access to their mother’s nipple. The question was whether crying is acoustically aversive, and therefore bystanders exhibit aggression towards the mother or the infant to stop the crying. It was predicted that mothers should respond more often to crying by allowing infant access to their nipples in the presence of potentially aggressive animals.

100

a)

Percentage of bouts in which access was given

Percentage of bouts in which access was given

Methods: 9 adult rhesus females that had dependent infants were observed in a free-ranging population on Cayo Santiago, Puerto Rico. All crying bouts of a focal mother’s infant were recorded and quantified for: i) the number and identity of bystanders within 2 m of the focal mother; ii) the occurrence of aggression towards the mother or the infant during the crying bout; iii) the type of aggression and the identity of the aggressive animal; iv) whether contact was made with the nipple. Each crying bout was then classified according to the presence of i) no bystanders; ii) low-risk bystanders; or iii) high-risk bystanders.

Percentage of bouts in which access was given

36

80 60 40 20 0

100

No bystanders

Low-risk bystanders

100

b)

80 60 40 20 0

No bystanders

High-risk bystanders

c)

80 60 40 20 0

Low-risk bystanders

High-risk bystanders

Fig. 2.1 a-c) Percentage of crying bouts in which mothers gave crying infants access to the nipple, depending on the presence and type of bystanders.

Vocal communication in social groups

37

Results: Crying was associated with increased aggression from bystanders towards mothers and infants. In particular high-risk bystanders affected the outcome of the crying bout more than low-risk bystanders (Fig. 2.1 a-c). Mothers were less likely to yield access to the nipples when no bystanders were present (39.4% of crying bouts), compared to situations when low-risk (53.5%) or at least one high-risk bystander (81.8%) were present. Conclusions: This study showed the influence of bystanders on a receiver, rather than on the signaller producing the call. It emphasises the point that the effect of additional receivers on all parties involved in a signalling interaction have to be considered to fully understand the social dynamics in group-living animals. (Semple et al. 2009)

2.3 Eavesdropping Eavesdropping is defined as the act of extracting information from signalling interactions between other individuals (McGregor and Dabelsteen 1996, McGregor 2005). Two classes of eavesdropping have been identified: social and interceptive eavesdropping (McGregor and Dabelsteen 1996, Peake 2005). Social eavesdropping occurs when individuals gather information on other individuals by attending to their signalling interaction. This form of eavesdropping has important consequences for individuals that interact regularly with several individuals and have the opportunity to monitor social and communicative interactions of other group members. In contrast, during interceptive eavesdropping, an individual benefits by intercepting signals intended for another individual, usually at a cost to the signaller. Predators and parasites commonly use interceptive eavesdropping to locate prey (Zuk and Kolluru 1998, Mougeot and Bretagnolle 2000, Bernal et al. 2007, Fichtel in press a), and competitors rely on it to oust rivals from a resource (Peake 2005). However, prey eavesdropping on vocalisations emitted by predators in turn resulted in adaptations of predators, such as transient killer whales (Orcinus orca), reducing their vocal activity while hunting on marine mammals, which have sensitive underwater hearing within the frequency range of killer whales (Deeke et al. 2005). Thus, the ability to eavesdrop and the knowledge that other species may eavesdrop (audience effect) provide a basis for the evolution of mutually fine-tuned behavioural adaptations.

38

Claudia Fichtel and Marta Manser

2.3.1 Social eavesdropping Eavesdropping plays an important role in assessing other individual’s quality in contexts such as territorial conflict or mating competition. By conducting playback experiments, it has been shown that male eavesdroppers modulate their agonistic decisions based on information obtained from listening to aggressive signalling interactions between unfamiliar males (Naguib and Todt 1997, Naguib et al. 1999, Peake et al. 2001, Mennill and Ratcliffe 2004a, Paz-y-Mino et al. 2004), where eavesdropping males usually respond more aggressively to the perceived winner (Naguib and Todt 1997, Naguib et al. 1999, Peake et al. 2001, Mennill and Ratcliffe 2004b, McGregor 2005). Females also use information transmitted during malemale aggressive interactions to assess the suitability of future males, and they prefer to mate with the winner of an observed contest (Otter et al. 1999, Menill et al. 2002, McGregor 2005). In social groups, eavesdropping is important in the acquisition of social knowledge that influences the formation of strategic alliances. Individuals of social groups form long-lasting stable relationships based on kinship, dominance or temporary affiliation (e.g., friendships in primates) of unrelated individuals (Krause and Ruxton 2002, Kotrschal et al. this volume). In order to compete successfully within such groups, it is advantageous for individuals to recognise who outranks whom, who is closely bonded to whom, and who is likely to be allied to whom. The ability to adopt a third party’s perspective and to discriminate among social relationships that exist among others would seem to be of great selective benefit (Cheney and Seyfarth 2005). Experiments revealed that primates attend extensively to vocalisations in social interactions among other group members. For example, in chacma baboons (Papio hamadryas ursinus), dominant individuals produce threat-grunts and subordinates screams during aggressive interactions. Mimicking such fights with playback experiments, Cheney and colleagues (1995) demonstrated not only that baboons pay attention to these vocal interactions, but also that simulated rank reversals elicited more attention than sequences that were consistent with the dominance hierarchy, indicating that bystanders keep track of the dominance hierarchy by regularly monitoring vocal interactions of group members. Bergman and colleagues (2003) additionally showed that baboons are able of evaluating aggressive vocal interactions not only on the basis of rank but also kinship (see BOX 2.3). Moreover, eavesdropping on conflicts involving kin enhances the opportunity to support relatives in these conflicts, which in turn may have long-term consequences on relative’s dominance rank (Engh et al. 2009). Thus, eavesdropping may have played an important proximate

Vocal communication in social groups

39

role in the evolution of alliance formation and cooperation (Cheney and Seyfarth 2005, Covas et al. 2007, Bshary this volume). Eavesdropping on vocal utterances of group members allows individuals not only to gather information by listening to dyadic interactions but also by combining information gained from vocalisations of individuals that do not directly interact with each other. In baboons, dominant males monopolise matings by forming consortships with oestrous females that are characterised by mate guarding and close proximity. Baboons usually produce grunts when travelling and foraging and females produce copulation calls while mating. Crockford and colleagues (2007) mimicked a temporal separation of the consorting male and its respective female by broadcasting the consort male’s grunts from one speaker and the female’s copulation calls from another speaker 40 m apart, simulating that she is copulating with another male. Lower-ranking males that were tested with this playback experiment approached the speaker with the female calls more often than after control trials, indicating that eavesdropping on other’s vocalisation might be one strategy by which male baboons achieve sneaky matings. In other species with stable social networks, such as African elephants, social knowledge seems to provide fitness benefits for receivers. African elephants live in complex fission-fusion societies comprising different matrilineal family units led by the oldest female. Playback experiments with contact calls revealed that adult females are familiar with calls of around 100 other individuals in the population, being able to discriminate between calls on the basis of how often they associate with the caller (McComb et al. 2000). However, families differ in how good they are in this task. Linking reproductive success of the family unit with the discriminatory ability of the oldest females revealed that families with older matriarchs have higher reproductive success, some of which might be attributable to superior social knowledge (McComb et al. 2001, McComb and Reby 2005). These examples illustrate that social knowledge acquired by eavesdropping on the communicative network within a group can provide fitness benefits on eavesdroppers. Mutual eavesdropping can structure cooperation and alliance formation, and, hence, contribute to long-term group stability. Moreover, at the community level, eavesdropping has important consequences for the evolution of anti-predator strategies. Some species eavesdrop on other prey’s alarm calls. Such heterospecific alarm call recognition occurs between taxa and is usually characterised by mutual recognition of both species’ alarm calls. Heterospecific recogniton of alarm calls has been demonstrated in many social birds and mammals, but also in a non-vocal reptile, the Galápagos marine iguana (Amblyrhynchos cristatus), or non- social species such as Gunther’s dik-dik (Madoqua

40

Claudia Fichtel and Marta Manser BOX 2.3 Hierarchical classification by rank and kinship in baboons Research question: Groups of chacma baboons (Papio hamadryas ursinus) are composed of a number of different matrilines arranged in a stable, linear dominance hierarchy in which all female members of one matriline outrank or are outranked by all female members of another matriline. Ranks are typically stable, often for 20 or more years, but when they do change entire families rise in rank over others. The question was whether baboons are able to evaluate rank and kinship of aggressively interacting females simultaneously, and whether they recognise that some rank reversals may have greater social importance than others. Methods: A1 - A2 - A3 B1 - B2 - B3 C1 - C2 - C3

Fig. 2.2 Schematic illustration of three matrilines, each including three females. Matriline/family A is high-ranking, B middle and C is low-ranking. Within families, female 1 is high, 2 middle, and 3 is low-ranking.

Playback experiments were designed to mimic a fight between two females. During aggressive interactions dominant females produce threat-grunts whereas subordinate females produce screams (A1 threat-grunts => A3 screams). To simulate a rank reversal threat-grunts and screams of different ranking females were broadcast from a hidden loudspeaker in an inconsistent order (A3 threat-grunts => A1 screams). 19 adult females living in the Okavango Delta in Botswana (all of known matrilineal kinship) were tested with 3 different playback trials: (1) (2) (3)

within family rank reversal (A3 threat-grunts => A1 screams) between family rank reversal (B3 threat-grunts => A1 screams) no-reversal control consistent with the rank order (A1 threat-grunts => A3 screams)

Based on video-recordings, the time females spent looking towards the speaker was measured.

Results: Individuals spent more time looking towards the speaker after the presentation of betweenfamily rank reversals than after within-family rank reversals or the control. Time spent looking towards the speaker did not differ after the presentation of within-family rank reversal and the control, however (Fig. 2.3).

Duration of looking [sec]

Vocal communication in social groups

41

5 4 3 2 1 0

Between family

Within family

No reversal control

Conclusions: Baboons are able to classify others simultaneously according to both individual rank and kinship. The selective pressures imposed by complex societies may therefore have favoured cognitive skills that allow monitoring rank reversals by eavesdropping on aggressive interactions of group members. (Bergman et al. 2003)

guentheri; Seyfarth and Cheney 1990, Shriner 1998, Fichtel 2004, Vitousek et al. 2007, Magrath et al. 2007, Lea et al. 2008, Müller and Manser 2008). In addition, in mixed-species associations, in which groups of different species regularly travel and forage together, eavesdropping on the other’s species alarm calls provides mutual benefits. Because mixedspecies groups often consist of species that preferentially use different habitat strata, each species benefits by the complementary sensitivity in predator detection (Fichtel in press b) and extended foraging niches (Gautier-Hion et al. 1983, McGraw and Bshary 2002, Wolters and Zuberbühler 2003). Thus, the ability to eavesdrop is one of the prerequisites for the formation of mixed-species associations.

2.4 Group cohesion and group coordination Living in groups has many benefits but also generates conflicts due to individual differences in needs depending on age, sex or social status (Alexander 1974, Bertram 1978, van Schaik 1983, Zemel and Lubin 1995). To maintain group cohesion and social stability despite these conflicts, individuals need to synchronise and coordinate their activities such as foraging, resting, social interactions and collective movements (Conradt and Roper 2003, 2007, Rands et al. 2003, Kerth et al. 2006, Kerth this volume). In the context of maintaining group cohesion and coordination, many animals produce vocal signals (Boinski and Garber 2000, Radford

42

Claudia Fichtel and Marta Manser

2004, Trillmich et al. 2004). These calls have usually been labelled as contact calls and are widespread among social mammals and birds. Contact calls may serve as a ‘location marker’ to announce the caller’s spatial position and to regulate spacing between group members, but also to signal group identity. These calls are usually given at high rates when groups move or forage and appear to be addressed at a group of targets; however, in some cases the same calls are also addressed at specific targets. Many social species produce contact calls when travelling, foraging and resting (Struhsaker 1967, Boinski 1993, Rendall et al. 2000, Radford 2004, Trillmich et al. 2004, Koda et al. 2008). For example, group-living pied babblers regularly produce so-called chucks while foraging to maintain cohesion, but also to regulate spacing of potential foraging competitors (Radford and Ridley 2006). Contact calls appear to be addressed at a class of targets and in many primate species, such as chacma baboons or Japanese macaques (Macaca fuscata), call rate increases when the risk of becoming separated from the group is increased (Rendall et al. 2000, Koda et al. 2008). From the sender’s perspective these calls appear to signal the senders’ location. However, from the receiver’s perspective, their function is less well understood, and contact calls may modulate the receiver’s tendency to approach or to avoid individuals. Thus, contact calls appear to have an important function within the signalling network of a group in the maintenance of cohesion, but the proximate mechanism driving the contact function are not entirely clear. In contrast, the function of the same type of contact calls seems to vary when they are addressed at a specific target. Many primates, such as baboons or macaques, produce contact calls when approaching other individuals, and the likelihood of a subsequent peaceful interaction is higher when approaches are accompanied by a contact call. During social interactions, contact calls seem to communicate the intention to behave peacefully towards others and may function as generic commitments conveying information about what animals will do next (Bauers 1983, Masataka 1989, Cheney et al. 1995, Rendall et al. 2000, Silk et al. 2000, Silk 2001). Thus, according to the audience, i.e. a group of targets or specific targets, the function of contact calls varies in some species from maintaining group cohesion to regulating social interactions. In addition, many species produce vocalisations to indicate their readiness to travel or to initiate group movements (reviewed by Boinski 2000). So-called travel calls (Boinski 1991) have been described in birds and mammals but mainly in primates (reviewed in Fichtel et al. in press). For example, group-living birds, such as green woodhoopoes (Phoeniculus purpureus) and domestic geese (Anser domesticus), or primates, such as squirrel and capuchin monkeys, use specific calls to initiate group move-

Vocal communication in social groups

43

ments and to recruit group members (Boinski 1993, Radford 2004). Some species also combine visual and acoustic displays to initiate movements, Barbary macaques (Macaca sylvanus) shake twigs or drum on dead wood (Mehlmann 1996), and elephants produce a specific rumble-call in combination with ear flagging (Poole et al. 1988). However, some species do not produce a specific travel call to initiate movements, but exchange contact calls during the decision-making processes before the group’s departure. For example, Guinea baboons (Papio papio) exchange contact barks to coordinate themselves before the group splits up into subgroups or fusions again (Byrne 1981). Verreaux’s sifakas (Propithecus verreauxi) do not produce a specific travel call to initiate movements, but contact calls before leaping off, suggesting that these are involved in the decision-making process of the group’s departure (Trillmich et al. 2004). Thus, the use of vocal signals that are directed at a broad audience and used to initiate and coordinate group movements seems to be widespread in group-living animals and might be an important mechanism of decision-making processes of movements (see also Kerth this volume). However, detailed empirical data on information transfer by acoustic signals driving decision-making processes are rare and represent a promising future research area in communication networks. Within a signalling network, animals can also benefit from coordinated activity via passive transfer of cue-based information of acoustic signals produced as a by-product of other behaviours. For example, rough-toothed dolphin (Steno bredanensis) subgroups often synchronise their swimming behaviour in order to save energy caused by positive hydrodynamic effects. During synchronised swimming, usually only one group member produces echo-locating signals and the other members are silent to avoid ambiguous echo scenery, thereby eavesdropping on the signalling individual, which may lead to further energy savings (Götz et al. 2006). Moreover, lesser bulldog bats (Noctilia albiventris) produce echo-locating calls when prey has been detected and is being attacked. Playback experiments revealed that these inadvertently produced cues attract other group members, facilitating social foraging and potentially more efficient exploitation of patchily distributed food resources (Dechmann et al. 2009). Moreover, inadvertently produced cues may also serve as alarm signals. In crested pigeons (Ocyphaps lophotes), which have modified flight feathers that cause a distinct whistle when flying off, it has been shown that other group members use these cues as alarm signals (Hingee and Magrath 2009). Thus, these examples illustrate that the production of acoustic signals or cues provides public information for group members, which in turn may facilitate group cohesion and coordination of activities.

44

Claudia Fichtel and Marta Manser

2.5 Costs-benefit analyses of signalling networks in social groups Communication evolves due to direct or indirect benefits to the signaller and receiver by the production of a signal, with the net gain potentially differing between them. The fact that senders adjust their calling behaviour depending on receivers acting as audience or eavesdroppers, suggests that uncontrolled signalling induces costs. It also indicates that to understand the evolution of communication we do not only have to consider a dyadic interaction between a sender and a receiver, but also the network of additional animals exposed to the signal’s transmission. The costs and benefits of adjusting signalling may be more affected in the production of costly rather than cheap signals and when the variation between receivers is large (Searcy and Nowicki 2005). In social groups, where a sender is likely to be surrounded by individuals differing in status, age, sex, reproductive stage or condition, the variation of the net gain to adjust signalling will depend on the size and heterogeneity of the group. Furthermore, the likelihood of interacting repeatedly with the same individuals is high, which can affect the ESS in signalling behaviour compared to one-time interactions (Silk et al. 2000). Cost-benefit analyses and theoretical models focusing on dyadic actor-reactor situations (e.g. Johnstone 1998) are thus unlikely to reflect the full complexity of communication in group-living species, where multiple senders and receivers interact repeatedly with each other. The early models of communication considering only dyadic one-time interactions between a single signaller and a receiver may explain the basic underlying mechanisms. However, more recent models in which the influence of additional individuals and repeated interactions on signalling interactions has been considered in specific contexts indicated that these factors can influence the outcome importantly. For example, in the context of begging, it has been shown that the presence of other individuals, in this case siblings, can affect the begging behaviour and cause different ESS in calling behaviour, depending on whether multiple signallers compete with each other (Godfray 1995) or cooperate (Johnstone 2004) while addressing the same receiver(s). For low-cost signals that have been suggested to evolve only when the benefits to both the signaller and the receiver are similar, a game-theoretical model that allowed individuals to interact repeatedly, revealed that low-cost signals can also evolve when conflicts of interests exists between sender and listener (Silk et al. 2000). Sumpter and Brännström (2008) suggested that costly signalling can evolve through synergy. By communicating within a network the group becomes more effective than the sum of their parts (see BOX 2.4). In addition, theoretical

Vocal communication in social groups BOX 2.4 Mechanistic model of foraging: synergy effect due to communication Research question: Foraging success and colony size correlate in cliff swallows (Petrochelidon pyrrhonota, Fig. 2.4; Brown and Brown 1996). Both, the amount of food collected by the parents per foraging trip and the frequency of trips increased with colony size. When these factors were combined, they appeared to add up to a linear increase in per capita success with colony size. The question, therefore, is why foraging success increases with colony size? Fig. 2.4 The amount of food collected per parent cliff swallow increases with colony size.

Amount of food collected per foraging trip [g]

1.0 66

0.8 0.6

27 43

15 37

9 31

0.4

3

0.2 1

10

100

1000

10000

Colony size

Model: General framework of the mechanism of how a group of communicating foragers finds a food source. A group of n individuals consists of x individuals that are informed about the location of the food source. This also means that x is a measure of group productivity, because it is the number of successful foragers at any point, and x/n equals per capita success because it represents the proportion of informed individuals in the group. The rate per individual of finding the feeder is given by the function f(x). The rate per individual of losing the food source is given by g(x). Both functions f and g differ depending on the recruitment mechanism used by the modeled species. Therefore the rate of change of the number of individuals going to the feeder is:

dx  f ( x)(n  x)  g ( x) x dt Solving

→

(t denotes time)

f ( x*)(n  x*)  g ( x*) x *

(x* is the equilibrium, or long-term, number of individuals going to the feeder). The change of this equilibrium with n determines how foraging efficiency changes with group size.

45

46

Claudia Fichtel and Marta Manser Conclusions: If we assume that social communication increases the rate at which individuals find the food source, then f(x*) increases and g(x*) decreases. As a direct consequence, each individual will profit from signalling about the food location as the feedback loop increases the total gain more than the summed benefits, and this effect may explain the pattern found in cliff swallows. (Sumpter and Brännström 2008)

models including multiple senders and receivers suggest that the main function of signalling networks is to facilitate collective action, and to guide group decisions that lead to coordinated action (Skyrms 2009). The different empirical studies reporting effects of additional individuals within the active signal range on signalling behaviour clearly emphasise the fact that variation in the composition of communication networks needs to be considered in cost-benefit analyses. However, existing theoretical models have only integrated additional senders and receivers to a limited extent. This may have two reasons: first, communication networks have only been recently emphasised (McGregor and Dabelsteen 1996), and secondly, dynamic analyses including multiple senders and receivers or repeated interactions result in complex signalling games that are mathematically challenging (Skyrms 2009). In addition, empirical studies that not only identify the dyadic interactions, but also describe the outcome of them in relation to the network and repeated interactions are still missing. For example, the benefits of attracting others by advertising a food source and partially sacrificing the caller’s own share (Gyger et al. 1986, Hauser and Marler 1993, Evans et al. 1994, Wilkinson and Boughman 1998, Di Bitetti 2003, Gros-Louis 2004b) have been intensely debated. Food callers may benefit by increasing their inclusive fitness, or decreasing predation risk or reciprocation (Elgar 1986, Heinrich and Marzluff 1991, Wilkinson and Boughman 1998). However, Sumpter and Brännström (2008) suggested that by communicating within a network, animals can increase their rate of finding food and costly signalling can evolve through synergy, where the gain for the group members by communicating with each other becomes more than the sum of their parts. Furthermore, the example of baboons, which use their memory of recent interactions to make inferences about whether a call is being directed at themselves or at some other individual (Engh et al. 2006), indicates the need to investigate signalling behaviour as a function of repeated interactions. Thus, the composition of the communication network and the repeated use of signals, as well as coordinated activities in social groups, clearly influence when and how they are used, in addition to the factors identified to affect dyadic interactions.

Vocal communication in social groups

47

2.6 Conclusions and outlook for future studies Vocal communication in social groups does not only include dyadic signalling interactions, but is affected by the presence of additional individuals within the signal range. To understand the function and evolution of specific signals, we therefore have to consider the entire communication network. Such an approach may reveal that benefits and costs must not be constant, but instead reflect non-linear functions of group size and composition. This insight furthers a better understanding of why different forms of social communication evolved in different types of groups (Sumpter and Brännström 2008). Adopting this perspective also allows the formulation of empirically testable predictions. Without knowing the mechanism of communication and how they depend upon the number, identity or quality of interacting individuals, we cannot predict whether or not a particular form of communication will evolve. It is important to measure the benefit function in a wide range of contexts, considering different group sizes and different group compositions (homogenous versus heterogeneous). Together with models that include multiple interacting individuals, we will be able to identify which factors and interactions drive the evolution of variation in vocal communication in social groups. Moreover, more empirical studies are required to identify the effects of social communication within the networks and to compare the outcomes of dyadic interactions without any bystanders to multiple interactions among several or all group members. The comparison of signalling behaviour in small groups versus large groups on group decision-making processes and the function of contact calls will permit identification of the effect of the presence of multiple individuals. Furthermore, eavesdropping and its influence on future actions have only been investigated in a few contexts and species, but it is likely to affect the dynamics of social groups extensively. Determining whether it is simply the number of individuals involved in a signalling interaction or their identity will also contribute to our understanding of communication in social groups.

Acknowledgements We thank Peter Kappeler for the invitation to contribute to this book. Additionally,we are grateful to Dan Blumstein, Peter Kappeler, Martin Schaefer and Robert Seyfarth for helpful comments on an earlier version or parts of this chapter.

48

Claudia Fichtel and Marta Manser

References Alexander RD (1974) The evolution of social behavior. Annu Rev Ecol Syst 5:325-383 Bauers KA (1993) A functional analysis of staccato grunt vocalizations in the stumptailed macaque (Macaca arctoides). Ethology 94:147-161 Bell MBV (2007) Cooperative begging in banded mongoose pups. Curr Biol 17:717-721 Bell MBV (2008) Receiver identity modifies begging intensity independent of need in banded mongoose (Mungos mungo) pups. Behav Ecol 19:1087-1094 Bergman TJ, Beehner JC, Cheney DL, Seyfarth RM (2003) Hierarchical classification by rank and kinship in baboons. Science 302:1234-1236 Bernal XE, Page RA, Rand AS, Ryan MJ (2007) Cues for eavesdroppers: do frog calls indicate prey density and quality? Am Nat 169:409-415 Bertram BCR (1978) Living in groups: predators and prey. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 64-96 Boinski S (1991) The coordination of spatial position: a field study of the vocal behaviour of adult female squirrel monkeys. Anim Behav 41:89-102 Boinski S (1993) Vocal coordination of troop movement among white-faced capuchin monkeys, Cebus capucinus. Am J Primatol 30:85-100 Boinski S (2000) Social manipulation within and between troops mediates primate group movement. In: Boinski S, Garber PA (eds) On the Move: How and Why Animals Travel in Groups. University of Chicago Press, Chicago, pp 421-469 Boinski S, Garber PA (2000) On the Move: How and Why Animals Travel in Groups. University of Chicago Press, Chicago Bonnie KE, Earley RL (2007) Expanding the scope for social information use. Anim Behav 74:171-181 Bradbury JW, Vehrencamp SL (1998) Principles of Animal Communication. Sinauer Associates, Sunderland/MA Bugnyar T, Kijne M, Kortschal K (2001) Food calling in ravens: are yells referential signals? Anim Behav 61:949-958 Byrne RW (1981) Distance vocalizations of Guinea baboons (Papio papio): an analysis of function. Behaviour 78:283-312 Catchpole CK, Slater PJB (2008) Birdsong: Biological Themes and Variations. University of Chicago Press, Chicago Cheney DL, Seyfarth RM (2005) Social complexity and the information acquired during eavesdropping by primates and other animals. In: McGregor PK (ed) Animal Communication Networks. Cambridge University Press, Cambridge, pp 583-603 Cheney DL, Seyfarth RM, Silk JB (1995) The responses of female baboons (Papio cynocephalus ursinus) to anomalous social interactions: evidence for causal reasoning? J Comp Psychol 109:134-141

Vocal communication in social groups

49

Conradt L, Roper TJ (2003) Group decision-making in animals. Nature 421:155158 Conradt L, Roper TJ (2007) Democracy in animals: the evolution of shared group decisions. Proc R Soc Lond B 274:2317-2326 Covas R, McGregor PK, Doutrelant C (2007) Cooperation and communication networks. Behav Proc 76:149-151 Crockford C, Wittig RM, Seyfarth RM, Cheney DL (2007) Baboons eavesdrop to deduce mating opportunities. Anim Behav 73:885-890 Dabelsteen T (2005) Public, private or anonymous? Facilitating and countering eavesdropping. In: McGregor PK (ed) Animal Communication Networks. Cambridge University Press, Cambridge, pp 38-62 Danchin E, Giraldeau L-A, Valone TJ, Wagner RH (2004) Public information: from nosy neighbors to cultural evolution. Science 305:487-491 Dechmann DKN, Heucke SL, Giuggioli L, Safi K, Voigt CC, Wikelski M (2009) Experimental evidence for group hunting via eavesdropping in echolocating bats. Proc R Soc Lond B 276:2721-2728 Deecke VB, Ford JKB, Slater PJB (2005) The vocal behaviour of mammal-eating killer whales: communicating with costly calls. Anim Behav 69:395-405 DiBitetti MS (2003) Food-associated call of tufted capuchin monkeys (Cebus apella nigritus) are functionally referential signals. Behaviour 140:565-592 Elgar MA (1986) House sparrows establish foraging flocks by giving chirrup calls if the resources are divisible. Anim Behav 34:169-174 Engh AL, Hoffmeier RR, Cheney DL, Seyfarth RM (2006) Who, me? Can baboons infer the target of vocalizations? Anim Behav 71:381-387 Engh AL, Hoffmeier RR, Seyfarth RM, Cheney DL (2009) O brother, where art thou? The varying influence of older siblings in rank acquisition by female baboons. Behav Ecol Sociobiol 64:97-104 Evans CS (1997) Referential signals. In: Owings DH, Beacher MD, Thompson NS (eds) Perspectives of Ethology. Plenum Press, New York, pp 99-143 Evans CS, Evans L, Marler P (1993) On the meaning of alarm calls: functional reference in an avian vocal system. Anim Behav 46:23-38 Fichtel C (2004) Reciprocal recognition in sifaka (Propithecus verreauxi verreauxi) and redfronted lemur (Eulemur fulvus rufus) alarm calls. Anim Cogn 7:45-52 Fichtel C (in press a) Costs of alarm calling: lemur alarm calls attract fossas. Lemur News Fichtel C (in press b) Predation on primates. In: Mitani JC, Call J, Kappeler PM, Palombit R, Silk J (eds) The Evolution of Primate Societies. University of Chicago Press, Chicago Fichtel C, Kappeler PM (2002) Anti-predator behavior of group-living Malagasy primates: mixed evidence for a referential alarm call system. Behav Ecol Sociobiol 51:262-275 Fichtel C, Hammerschmidt K, Jürgens J (2001) On the vocal expression of emotion. A multi-parametric analysis of different states of aversion in the squirrel monkey. Behaviour 138:97-116

50

Claudia Fichtel and Marta Manser

Fichtel C, Pyritz LW, Kappeler PM (in press) Coordination of group movements in non-human primates. In: Boos M, Kolbe M, Ellwart S, Kappeler PM (eds) Coordination in Human and Non-human Primate Groups. Springer, Heidelberg Gautier-Hion A, Quris R, Gautier J-P (1983) Monospecific vs polyspecific life: a comparative study of foraging and antipredatory tactics in a community of Cercopithecus monkeys. Behav Ecol Sociobiol 12:325-335 Godfray HCJ (1995) Signaling of need between parents and young: parentoffspring conflict and sibling rivalry. Am Nat 146:1-24 Götz T, Verfuss UK, Schnitzler H-U (2006) ‘Eavesdropping’ in wild roughtoothed dolphins (Steno bredanensis)? Biol Lett 2:5-7 Griesser M (2008) Referential calls signal predator behavior in a group-living bird species. Curr Biol 18:69-73 Grinnell J, McComb K (2001) Roaring and social communication in African lions: the limitations imposed by listeners. Anim Behav 62:93-98 Gros-Louis J (2004a) The function of food-associated calls in white-faced capuchin monkeys, Cebus capucinus, from the perspective of the signaller. Anim Behav 67:431-440 Gros-Louis J (2004b) Responses of white-faced capuchins (Cebus capucinus) to naturalistic and experimentally presented food-associated calls. J Comp Psychol 118:396-402 Gyger M, Karakashian SJ, Marler P (1986) Avian alarm calling: is there an audience effect? Anim Behav 34:1570-1572 Hauser MD (1996) The Evolution of Communication. Bradford/MIT Press, Cambridge Hauser MD (1998) Functional referents and acoustic similarity: field playback experiments with rhesus monkeys. Anim Behav 55:1647-1658 Hauser MD, Marler P (1993) Food-associated calls in rhesus macaques (Macaca mulatta). II. Costs and benefits of call production and suppression. Behav Ecol 4:206-212 Heinrich B, Marzluff JM (1991) Do common ravens yell because they want to attract others? Behav Ecol Sociobiol 28:13-21 Hingee M, Magrath R (2009) Flights of fear: a mechanical wing whistle sounds the alarm in a flocking bird. Proc R Soc Lond B 276:4173-4179 Jablonka E (2002) Information: its interpretation, its inheritance, and its sharing. Philos Sci 69:578-605 Johnstone RA (1998) Efficacy and honesty in communication between relatives. Am Nat 152:45-58 Johnstone RA (2004) Begging and sibling competition: how should offspring respond to their rivals? Am Nat 163:388-406 Karakashian SJ, Gyger M, Marler P (1988) Audience effect on alarm calling in chickens (Gallus gallus). J Comp Psychol 102:129-135 Kerth G, Ebert C, Schmidtke C (2006) Group decision making in fission-fusion societies: evidence from two-field experiments in Bechstein’s bats. Proc R Soc Lond B 273:2785-2790

Vocal communication in social groups

51

Kitchen DM (2004) Alpha male black howler monkey responses to loud calls: effect of numeric odds, male companion behaviour and reproductive investment. Anim Behav 67:125-139 Koda H, Shimooka Y, Suguira H (2008) Effects of caller activity and habitat visibility on contact call rate of wild Japanese macaques (Macaca fuscata). Am J Primatol 70:1055-1063 Krause J, Ruxton GD (2002) Living in Groups. Oxford University Press, Oxford Laland KN (2004) Social learning strategies. Learn Behav 32:4-14 Lea AJ, Barrera JP, Tom LM, Blumstein DT (2008) Heterospecific eavesdropping in a nonsocial species. Behav Ecol 19:1041-1046 Leighty KA, Soltis J, Wesolek CM, Savage A (2008) Rumble vocalizations mediate interpartner distance in African elephants, Loxodonta africana. Anim Behav 76:1601-1608 Le Roux A, Cherry MI, Manser MB (2008) The audience effect in a facultatively social mammal, the yellow mongoose, Cynictis penicillata. Anim Behav 75:943-949 Magrath RD, Pitcher BJ, Gardner JL (2007) A mutual understanding? Interspecific responses by birds to each other’s aerial alarm calls. Behav Ecol 18:944-951 Manser MB (2001) The acoustic structure of suricates’ alarm calls varies with predator type and the level of response urgency. Proc R Soc Lond B 268:2315-2324 Manser MB (2009) The generation of functionally referential and motivational vocal signals in mammals. In: Brudzynsky SM (ed) Handbook of Mammalian Vocalization. Elsevier, New York, pp 479-488 Manser MB, Bell MB, Fletcher LB (2001) The information that receivers extract from alarm calls in suricates. Proc R Soc Lond B 268:2485-2491 Marler P (1992) Functions of arousal and emotion in primate communication: a semiotic approach. In: Nishida T, McGrew WS, Marler P, Pickford M, de Waal FBM (eds) Topics in Primatology. Vol 1: Human Origins. University of Tokyo Press, Tokyo, pp 225-233 Marler P, Dufty A, Pickert R (1986) Vocal communication in the domestic chicken. II. Is a sender sensitive to the presence and nature of a receiver? Anim Behav 34:194-198 Masataka N (1989) Motivational referents of contact calls in Japanese macaques. Ethology 80:265-273 Maynard Smith J, Harper D (2003) Animal Signals. Oxford University Press, Oxford McComb K, Reby D (2005) Vocal communication networks in large terrestrial mammals. In: McGregor P (ed) Animal communication networks. Cambridge University Press, Cambridge, pp 372-389 McComb K, Moss C, Sayialel S, Baker L (2000) Unusually extensive networks of vocal recognition in African elephants. Anim Behav 59:1103-1109 McComb K, Moss C, Durant SM, Baker L, Sayialel S (2001) Matriarchs as respositories of social knowledge in African elephants. Science 292:491-494

52

Claudia Fichtel and Marta Manser

McGraw WS, Bshary R (2002) Association of terrestrial mangabeys (Cercocebus atys) with arboreal monkeys: experimental evidence for the effects of reduced ground predator pressure on habitat use. Int J Primatol 23:311-325 McGregor PK (2005) Animal Communication Networks. Cambridge University Press, Cambridge McGregor PK, Dabelsteen T (1996) Communication networks. In: Kroodsma DE, Miller EH (eds) Ecology and Evolution of Acoustic Communication in Birds. Cornell University Press, Ithaca/NY, pp 409-425 Mehlman PT (1996) Branch shaking and related displays in wild Barbary macaques. In: Fa JE, Lindburg DG (eds) Evolution and Ecology of Macaque Societies. Cambridge University Press, Cambridge, pp 503-526 Mennill DJ, Ratcliffe LM (2004a) Do male black-capped chickadees eavesdrop on song contests? A multi-speaker playback experiment. Behaviour 141:125-139 Mennill DJ, Ratcliffe LM (2004b) Overlapping and matching in song contests of black-capped chickadees. Anim Behav 67:441-450 Mennill DJ, Ratcliffe LM, Boag PT (2002) Female eavesdropping on male song contests in songbirds. Science 296:873 Morton ES (1977) On the occurrence and significance of motivation-structural rules in some bird and mammal sounds. Am Nat 111:855-869 Mougeot F, Bretagnolle V (2000) Predation as a cost of sexual communication in nocturnal seabirds: an experimental approach using acoustic signals. Anim Behav 60:647-656 Müller CA, Manser MB (2008) The information banded mongooses extract from heterospecific alarms. Anim Behav 75:897-904 Naguib M, Todt D (1997) Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Anim Behav 54:1535-1543 Naguib M, Fichtel C, Todt D (1999) Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proc R Soc Lond B 266:537-542 Otter K, Mcgregor PK, Terry AMR, Burford FRL, Peake TM, Dabelsteen T (1999) Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proc R Soc Lond B 266:13051309 Owings DH, Leger DW (1980) Chatter vocalizations of California ground squirrels: predator-social-role specificity. Z Tierpsychol 54:163-184 Papworth S, Böse A-S, Barker J, Schel AM, Zuberbühler K (2008) Male blue monkeys alarm call in response to danger experienced by others. Biol Lett 4:472-475 Paz-y-Miño G, Bond AB, Kamill AC, Balda RP (2004) Pinyon jays use transitive inference to predict social dominance. Nature 430:778-781 Peake TM (2005) Eavesdropping in communication networks. In: McGregor PK (ed) Animal Communication Networks. Cambridge University Press, Cambridge, pp 13-37 Peake TM, Terry AMR, McGregor PK, Dablesteen T (2001) Male great tits eavesdrop on simulated male-to-male vocal interactions. Proc R Soc Lond B 268:1183-1187

Vocal communication in social groups

53

Pereira ME, Macedonia JM (1991) Ringtailed lemur anti-predator calls denote predator class, not response urgency. Anim Behav 41:543-544 Poole JH, Payne K, Langbauer WR Jr, Moss CJ (1988) The social contexts of some very low frequency calls of African elephants. Behav Ecol Sociobiol 22:385-392 Price K, Harvey H, Ydenberg R (1996) Begging tactics of nestling yellow-headed blackbirds, Xanthocephalus xanthocephalus, in relation to need. Anim Behav 51:421-435 Radford AN (2004) Vocal coordination of group movement by green woodhoopoes (Phoeniculus purpureus). Ethology 110:11-20 Radford AN, Ridley AR (2006) Recruitment calling: a novel form of extended parental care in an altricial species. Cur Biol 16:1700-1704 Rands SA, Cowlishaw G, Pettifor RA, Rowcliffe JM, Johnstone RA (2003) Spontaneous emergence of leaders and followers in foraging pairs. Nature 423:432-434 Reby D, Cargnelutti B, Hewison AJM (1999) Contexts and possible functions of barking in roe deer. Anim Behav 57:1121-1128 Rendall D, Cheney DL, Seyfarth RM (2000) Proximate factors mediating ‘contact’ calls in adult female baboons (Papio cynocephalus ursinus) and their infants. J Comp Psychol 114:36-46 Ridley AR, Child MF, Bell MBV (2007) Interspecific audience effects on the alarm-calling behaviour of a kleptoparasitic bird. Biol Lett 3:589-591 Searcy WA, Nowicki S (2005) The Evolution of Animal Communication: Reliability and Deception in Signalling Systems. Princeton University Press, Princeton Semple S, Gerald MS, Suggs DN (2009) Bystanders affect the outcome of motherinfant interactions in rhesus macaques. Proc R Soc Lond B 276:2257-2262 Seyfarth RM, Cheney DL (1990) The assessment by vervet monkeys of their own and another species’ alarm calls. Anim Behav 40:754-764 Seyfarth RM, Cheney DL, Marler P (1980) Vervet monkey alarm calls: semantic communication in a free-ranging primate. Anim Behav 28:1070-1094 Sherman PW (1977) Nepotism and the evolution of alarm calls. Science 197:1246-1253 Shriner WM (1998) Yellow-bellied marmot and golden-mantled ground squirrel responses to heterospecific alarm calls. Anim Behav 55:529-536 Silk JB (2001) Grunts, girneys, and good intentions: the origins of strategic commitment in nonhuman primates. In: Nesse R (ed) Evolution and the Capacity for Commitment. Russell Sage Press, New York, pp 138-157 Silk JB, Kaldor E, Boyd R (2000) Cheap talk when interests conflict. Anim Behav 59:423-432 Skyrms B (2009) Evolution of signalling systems with multiple senders and receivers. Philos Trans R Soc Lond B 364:771-779 Slobodchikoff CN, Kiriazis J, Fischer C, Creef E (1991) Semantic information distinguishing individual predators in the alarm calls of Gunnison’s prairie dogs. Anim Behav 42:713-719

54

Claudia Fichtel and Marta Manser

Slocombe KE, Zuberbühler K (2007) Chimpanzees modify recruitment screams as a function of audience composition. Proc Natl Acad Sci USA 104:1722817233 Smith WJ (1981) Referents of animal communication. Anim Behav 29:1273-1275 Struhsaker TT (1967) Auditory communication among vervet monkeys (Cercopithecus aethiops). In: Altmann SA (ed) Social Communication among Primates. University of Chicago Press, Chicago, pp 281-384 Sumpter DJT, Brännström A (2008) Synergy in social communication. In: d’Ettore P, Hughes DP (eds) Sociobiology of Communication: An Interdisciplinary Perspective. Oxford University Press, Oxford, pp 191-208 Townsend SW, Deschner T, Zuberbühler K (2008) Female chimpanzees use copulation calls flexibly to prevent social competition. PLoS ONE 3:e243, doi:10.1371/journal.pone.0002431 Trillmich J, Fichtel C, Kappeler PM (2004) Coordination of group movements in wild Verreaux’s sifakas (Propithecus verreauxi). Behaviour 141:1103-1120 van Schaik CP (1983) Why are diurnal primates living in groups? Behaviour 87:120-144 Vitousek MN, Adelman JS, Gregory NC, St Clair JJH (2007) Heterospecific alarm call recognition in a non-vocal reptile. Biol Lett 3:632-634 Wich SA, de Vries H (2006) Male monkeys remember which group members have given alarm calls. Proc R Soc Lond B 273:735-740 Wich SA, Sterck EHM (2003) Possible audience effect in Thomas langurs (primates; Presbytis thomasi): an experimental study on male loud calls in response to a tiger model. Am J Primatol 60:155-159 Wilkinson GS, Boughman JW (1998) Social calls coordinate foraging in greater spear-nosed bats. Anim Behav 55:337-350 Wolters S, Zuberbühler K (2003) Mixed-species associations of Diana and Campbell’s monkeys: the costs and benefits of a forest phenomenon. Behaviour 140:371-385 Wright J, Leonard ML (2002) The Evolution of Begging: Competition, Cooperation and Communication. Kluwer, Dordrecht Zahavi A (1977) The cost of honesty (further remarks on the handicap principle). J Theor Biol 67:603-605 Zahavi A, Zahavi A (1997) The Handicap Principle: A Missing Piece of Darwin’s Puzzle. Oxford University Press, New York Zemel A, Lubin Y (1995) Inter-group competition and stable group sizes. Anim Behav 50:485-488 Zuberbühler K (2008) Audience effects. Curr Biol 18:R189-R190 Zuberbühler K, Cheney DL, Seyfarth RM (1999) Conceptual semantics in a nonhuman primate. J Comp Psychol 113:33-42 Zuk M, Kolluru GR (1998) Exploitation of sexual signals by predators and parasitoids. Q Rev Biol 73:415-438

Chapter 3

Kin recognition: an overview of conceptual issues, mechanisms and evolutionary theory DUSTIN J. PENN AND JOACHIM G. FROMMEN

‘all good kumrads you can tell by their altruistic smell.’ (e.e. cummings)

ABSTRACT Kin recognition (KR) is the ability to identify or distinguish kin from nonkin, and it is thought to be an important driving force in the evolution of social and sexual behaviour. Here, we provide an introduction to KR, including an overview of the main debates, the underlying mechanisms and evolutionary analyses. First, we examine the many evolving definitions for KR, as these have caused some confusion and debate. We explain why retaining both broad and narrow definitions can be instructive for thinking about the problem. Second, we provide examples of the different types of KR from empirical studies, ranging from the use of spatial cues to selfinspection and green-beard genes. We also suggest a classification scheme of the different mechanisms based on whether they are considered to be KR in the broad or in the narrow sense. Third, we consider the key components necessary for most or all KR mechanisms, and explain why the central problem for any recognition mechanism is to balance the risk of acceptance versus rejection errors. Fourth, we summarise theoretical analyses addressing the evolution of nepotism through kin recognition, and the maintenance of genetic polymorphisms controlling KR. Fifth, we examine evolutionary analyses of apparent KR failures, errors, and mistakes. Finally, we suggest some of the main challenges that need to be addressed in future KR research.

56

Dustin J. Penn and Joachim G. Frommen

3.1 Introduction Individuals in many social species behave altruistically, and in extreme cases may even sacrifice their own reproduction for the group, as with eusocial insects. To explain the remarkable altruism found in some species, W.D. Hamilton (1964) suggested his theory of inclusive fitness which clarifies the conditions under which altruism is expected to evolve: ‘The social behaviour of a species evolves in such a way that in each distinct behaviour-evoking situation the individual will seem to value his neighbors’ fitness against his own according to the coefficients of relationship appropriate to that situation’ (p. 19). This principle, known as Hamilton’s rule, shows that genetic relatedness is one of the keys to understanding altruism. This insight provided an important new theoretical approach to studying cooperation and other social behaviours (genes’-eye view of evolution; Dawkins 1976), and inspired the development of sociobiology (Wilson 1975). Inclusive fitness theory is arguably one of the most important advance in our understanding of natural selection since the modern evolutionary synthesis (Griffin and West 2002). Hamilton (1964) also suggested that when selection favours helping kin, mechanisms may evolve that allow individuals to discriminate kin from non-kin. Further, he (1964:25) proposed a thought experiment in which a gene evolves that enables ‘the perception of the presence of like genes in other individuals’ and it also helps other individuals carrying the genes, and in this way, it helps copies of itself. He acknowledged that such a ‘super gene’, or what is now known as a green-beard gene (Dawkins 1976), is highly improbable, but nevertheless it is instructive for obtaining insights into how altruism evolves. Hamilton’s ideas sparked much interest in testing whether and how animals recognise kin (see Holmes 2004) and the role that kin recognition (KR) plays in cooperation, inbreeding avoidance, and other kin biases (BOX 3.1). Before Hamilton, KR research had previously focused on parent-offspring recognition, whereas inclusive fitness theory broadened interest in social interactions among collateral kin (e.g., siblings, nieces, nephews, aunts and uncles). Subsequent work has shown that organisms in a wide variety of taxa – from single-celled organisms to humans – are able to distinguish kin from non-kin (e.g., Fletcher and Michener 1987, Waldman et al. 1988, Hepper 1991a). KR may also influence the behaviour of gametes, the haploid phase of the life cycle (i.e., sperm-egg and sperm-sperm interactions; Moore and Moore 2002), and fetal-placental interactions (Haig 1996, Summers and Crespi 2005).

Kin recognition: conceptual issues, mechanisms and evolutionary theory

57

BOX 3.1 Kin biases occur with and without kin recognition There are many ways that organisms treat kin differentially from non-kin. Such kin biases can be developmental, physiological, or behavioural responses and may include either helping or avoiding kin. Many species display parental care, which can enhance offspring fitness under certain conditions (see Trillmich this volume). Furthermore, conditional helping for kin is common, and such nepotism results in kin selection. An extreme case of altruistic helping is the evolution of sterile workers in eusocial insects (Hughes et al. 2008). Kin biases also include kin aggregations and discriminating admittance to social groups. In addition to helping kin, some species avoid competition with kin (West et al. 2002), or show conditional aggression toward non-kin (cannibalism and infanticide) (Pfennig et al. 1993, Manica 2002). Many species show inbreeding avoidance, which most likely functions to avoid the deleterious consequences of inbreeding (inbreeding depression) (Crnokrak and Roff 1999). However, more studies are needed to determine whether natural selection favours intermediate levels of outbreeding due to harmful consequences of extreme outbreeding (optimal outbreeding). Kin biases do not require KR, however. For example, sex-biased dispersal may function to avoid inbreeding, and it does not require an ability to distinguish kin versus non-kin. It has been suggested that kin biases that do not involve KR should be called ancillary kin bias (Halpin 1991, Tang-Martinez 2001) or non-discriminatory kin bias (Barnard et al. 1991). However, the term ancillary implies a bias of secondary importance, which is not the case, and non-discriminatory kin bias may be confusing since non-discriminatory implies no bias (see BOX 3.2). We suggest that such biases might instead be referred to as kin biases without kin recognition.

The trouble is that kin recognition research has been plagued with terminological, conceptual, and methodological issues, including debates over the meaning of the term kin recognition (Waldman et al. 1988, Gamboa et al. 1991). These debates came to a head when Grafen (1990) evaluated the field, and made several controversial objections. He argued that there is a paucity of evidence for ‘true kin recognition’, which he defined as systems ‘whose use and function is to assess the kinship of conspecifics’. Grafen concluded that there was only one convincing example of true kin recognition (i.e., Grosberg and Quinn 1986) because previous studies failed to rule out possible artifacts or incidental biases arising from individual, group, and species recognition abilities. Nevertheless, Grafen described a verbal model for how a true kin recognition system might evolve and maintain itself. His paper triggered a flurry of criticism

58

Dustin J. Penn and Joachim G. Frommen

(Blaustein et al. 1991, Byers and Bekoff 1991, Stuart 1991) and replies (Barnard et al. 1991, Grafen 1991a,b,c). There is still no consensus over these issues, even after 20 years, though there have been many important advances. In this chapter, we provide an overview of kin recognition, and focus particularly on the following issues. First, we consider the various definitions for KR (and kin discrimination), and the terminological debates that have created much controversy and confusion. Second, we examine the proximate mechanisms controlling KR, and we suggest a novel classification of the mechanisms. Third, we consider how KR evolves, and summarise results from theoretical analyses aimed to understand the origins and maintenance of nepotism through KR. Fourth, we summarise evolutionary ideas proposed to explain apparent KR failures. Finally, we suggest some of the central problems that need to be addressed in future research on KR.

3.2 What is kin recognition? Sorting out terminological confusion Determining whether animals recognise their kin or not depends on how one defines kin recognition. In the broad sense, kin recognition refers to the ability to identify, distinguish or classify kin from non-kin, regardless of the mechanism or evolutionary functions (descriptive definition). However, the problem is that there are several other definitions (Table 3.1). For example, kin recognition is also defined as the ability to distinguish and the differential treatment of kin versus non-kin (Sherman and Holmes 1985, Waldman 1987). This operational definition is practical, but it muddles two potentially distinct processes, recognition – which is a proximate mechanism – versus kin-biased actions (Byers and Bekoff 1986). Recognition does not necessarily lead to differential responses toward kin, ‘just as recognising a fruit as an orange does not necessarily lead us to eating it’ (Barnard 1991), and kin biases are not necessarily due to kin recognition (see BOX 3.1). The problem is that different researchers use different definitions, not only for kin recognition (KR), but also for kin discrimination (KD) and kin bias (KB), sometimes even in the same paper. Therefore, it is useful to examine how these terms cause confusion (BOX 3.2). To try to reduce the ambiguity and confusion, some researchers propose using more strict or narrow definitions for KR (Table 3.1). However, the problem is that few seem to agree on a definition, and which definition a researcher uses depends upon whether they emphasise a particular mechanism, adaptive function, or evolutionary history. For example, some sug-

Kin recognition: conceptual issues, mechanisms and evolutionary theory

59

Table 3.1. Evolving definitions for kin recognition (KR). 1. Descriptive = ability to identify or distinguish kin versus non-kin, regardless of mechanisms or functions 2. Operational = ‘…differential treatment of conspecifics differing in genetic relatedness’ (Sherman and Holmes 1985; considered ‘KB’ here) 3. Mechanistic = ability to distinguish kin versus non-kin using genetic similarity or any cues that are correlated with kinship (Holmes 2004) = ‘the process by which individuals assess the genetic relatedness of conspecifics to themselves or others based on their perception of traits expressed by or associated with these individuals’ (Waldman et al. 1988) = ability to distinguish kin versus non-kin among conspecifics, including familiarity, but excluding spatial and temporal cues (Halpin 1991, TangMartinez 2001) = unobservable neural process of classifying individuals as kin (Byers and Bekoff 1986, Barnard 1991, Tang-Martinez 2001) = ability to distinguish and respond differentially to kin (Hepper 1991a; called ‘KD’ here) 4. Cognitive Mechanism = ‘the (externally, at least) unobservable neural process of classing individuals as kin’ (Barnard 1991, Barnard et al. 1991, Tang-Martinez 2001, Griffin and West 2002) 5. Adaptive Function = any mechanism that functions to recognise kin, regardless of the underlying mechanisms 6. Mechanism and Adaptive Function = ability to distinguish genetic similarity in conspecifics, which currently functions for this purpose (so-called true kin recognition; Grafen 1990) 7. Mechanism, Origins and Adaptive Function = ability to distinguish kin versus non-kin among conspecifics that originally evolved and currently functions for this purpose (Tang-Martinez 2001)

gest restricting the term KR to particular mechanisms, such as direct recognition of conspecific phenotypes (Halpin 1991, Tang-Martinez 2001), neural and cognitive mechanisms (Barnard 1991, Tang-Martinez 2001), or the ability to detect genetic similarity (Grafen 1990). Narrowly defining

60

Dustin J. Penn and Joachim G. Frommen BOX 3.2 Coming to terms with kin recognition, kin discrimination, and kin bias There are several definitions for kin recognition (KR), kin discrimination (KD), and kin biases (KB), which cause much confusion and debate. The word recognition refers to the ability to recall and identify someone from past experience or knowledge (from the Latin recognoscere, or to know again). Thus, Halpin (1986) suggested that strictly speaking this term should only be used for the ability to identify or distinguish familiar or previously known individuals, as the recognition of a stranger is ‘a logical impossibility’. Halpin points out that recognition is often used synonymously with discrimination (from the Latin discriminare or to divide), and suggests that discrimination is a preferable term for the ability to distinguish individual conspecifics. The problem, however, is that the word discrimination has three different definitions in English: (1) perceptive abilities to distinguish or make fine distinctions; (2) biased treatment of individuals based on their membership to a group; (3) or both. This ambiguity explains why KD is defined in different ways: (1) the ability to distinguish kin from non-kin (which we call KR); (2) differential treatment of kin versus non-kin (which we call KB, but it is also used as an operational definition for KR; Hepper 1991b); (3) both an ability to distinguish and differential treatment of kin versus non-kin. i.e., some use all three interchangeably (Holmes and Sherman 1983); (4) kin biases due to KR (Barnard 1991, Barnard et al. 1991). It is little wonder that KD causes so much confusion, especially since authors often slip back and forth among the common usages. The figure below summarises the relationships between the various definitions and terms, and how we define them in this paper (also see the Glossary). (1) The ability to identify or distinguish kin versus non-kin = KR (but also called KD)

(4)

(2) differential treatment of kin versus non-kin = KB (but also called KR, KD)

(3) the ability to distinguish and the differential treatment of kin versus non-kin KR and KB = KD (4) differential treatment of kin versus non-kin due to the ability to distinguish them KB based on KR = KD

KR based on a particular mechanism is unnecessarily restrictive, however. Restricting KR to mechanisms that rely on individual phenotypic cues, for example, might rule out other interesting possibilities, such as KR based on extended phenotypes (e.g., acquired commensal microbes or the shape of a birds’ nest, etc.). Limiting KR to cognitive or neural mechanisms ex-

Kin recognition: conceptual issues, mechanisms and evolutionary theory

61

cludes bacteria, plants, unicellular social amoebae, and colonial marine invertebrates (unless cognition is broadly defined to include these taxa). Restricting KR to genetic similarity detection (Grafen 1990) could unnecessarily define the topic out of existence. Also, in many circumstances, environmentally determined cues may provide even better indicators of kinship than genetic ones (Gamboa et al. 1991). As Hamilton (1975) pointed out, ‘kinship should be considered just one way of getting this positive regression of genotype in the recipient, and that it is this positive regression that is vitally necessary for altruism’. If a species has evolved parent, offspring, or sibling recognition, then it is difficult to understand why such behaviours should not be considered KR (after all, offspring are kin) just because they use mechanisms other then genetic similarity detection. Grafen (1990) also proposed that the term KR should be restricted to mechanisms that specifically function to recognise kin. Testing the consequences of KR is necessary to distinguish between adaptive versus incidental KR. However, some ethologists caution against labeling behaviours by their presumed functions or proximate causes, as these assumptions may later turn out to be incorrect (and a behaviour may have more than one function), and instead recommend using descriptive terms. Tang-Martinez (2001) goes further and proposes reserving the term KR only for mechanisms that originally evolved as well as currently function to recognise kin. This idea follows from the suggestion that the term adaptation should be reserved for traits that are shaped by natural selection for their current role, and the term exaptation for traits that originally evolved for something other (Gould and Vrba 1982). However, history-laden definitions for adaptation are impractical and misleading for several reasons (Reeve and Sherman 2001). For example, it is impossible to discern where an adaptation ends and an exaptation begins. How much of the current function of a trait must differ from its original role for it to be considered to be an exaptation? Adaptations generally evolve by modifying or coopting existing traits, and therefore, all traits are exaptations if we go back far enough in time. So, while more work is needed on the functions (Blaustein et al. 1991) and phylogenetic origins of KR (Crampton and Hurst 1994), there are practical reasons to advocate descriptive over mechanistic, functional or history-laden definitions. To demonstrate KR, some researchers insist that in addition to showing an ability to distinguish kin versus non-kin, one must also work out the mechanism (though there is no consensus about which mechanism, as we address further in the next section), its adaptive functions, and its evolutionary origins. This is hardly practical. A number of definitions for KR have been proposed, and there are advantages in retaining both broad and

62

Dustin J. Penn and Joachim G. Frommen

narrow definitions (as is the case for many terms in biology, such as heritability). We suggest that strict definitions of KR should be treated as hypotheses about the underlying mechanisms and functions. It is instructive to note that similar terminological problems plagued the study of social learning, as Galef (1976) pointed out over 30 years ago: ‘Some investigators prefer purely descriptive terms even though these can obscure differences in the mechanisms underlying surface similarities in behavioural interaction.… Others utilise terminology reflecting hypothesised underlying mechanisms mediating observed behavioural interaction…, and there are those who employ operationally defined categories…’. He lamented that the explosion of new definitions and labels has added nothing to our understanding of the mechanisms or the adaptive functions for social learning. This is probably why Hamilton (1975) suggested that ‘it seems on the whole preferable to retain a more flexible use of terms’.

3.3 Proximate mechanisms: how do organisms recognise kin? In this section, we provide an overview and examples of the various mechanisms through which organisms identify or distinguish kin versus non-kin. We also suggest a classification scheme for these mechanisms based on the broad and narrow definitions for KR. Then, we consider a general model describing the key components that are thought to comprise all or most KR mechanisms. 3.3.1 Types of kin recognition mechanisms There are several different mechanisms that enable animals to distinguish kin versus non-kin, and several classifications of KR mechanisms have been proposed. The most widely cited scheme (e.g., Sherman and Holmes 1985, Waldman 1987) proposes four categories: (1) recognition alleles; (2) familiarity (associative learning); (3) phenotype matching; and (4) spatially-based mechanisms. These categories are criticised because they are not based on consistent criteria, and some mechanisms (2 and 4) are not considered to be KR by researchers who reject the broad definition for KR. The alternative proposals, however, are rather complicated (Barnard et al. 1991, Tang-Martinez 2001) or too simple (e.g., it has been suggested to exclude spatially-based mechanisms, and combine the rest into two categories, recognition alleles and learning; Barnard et al. 1991, Tang-Martinez

Kin recognition: conceptual issues, mechanisms and evolutionary theory

63

Table 3.2. Kin recognition mechanisms can be classified according the types of cues used and matching rules (i.e., the referents used for recognition). Shading shows broad and more narrow definitions. 1. Contextual cues R uses spatial, temporal, and state-dependent cues 2. Phenotypic cues A. Direct familiarity R compares S at time 1 to S at time 2 (e.g., associative learning of individually-distinctive cues) B. Without direct familiarity (i) Indirect familiarity R compares S to K (e.g., associative learning of familial-distinctive cues) (ii) Self inspection R compares S with R (e.g., compare individually-distinctive cues through habituation/dishabituation) (iii) Green-beard genes R compares S with R (no prior experience necessary) R = receiver making the evaluation, S = sender being evaluated, and K = likely kin to R

2001). Thus, the proposed classifications of KR mechanisms vary depending upon how researchers define kin recognition. Therefore, we propose a classification of the types of KR based on broad or narrow-sense definitions (Table 3.2). We first examine examples of KR in the broad-sense (spatially-based mechanisms and familiarity), and then in the narrow-sense (indirect familiarity, self-inspection, and green-beard genes). Another distinction used for classifying KR mechanisms is based on the types of cues, which can be divided into two categories: indirect, contextual cues versus direct, phenotypic cues (Waldman 1988). Since direct, phenotypic cues are further divided into direct versus indirect familiarity (see below; Wyatt 2003), we only use the terms direct and indirect for this later distinction. 3.3.1.1 Contextual cues (spatial, temporal and other nonphenotypic mechanisms) Many species rely on spatial, temporal, or other contextual cues to distinguish kin versus non-kin (see Table 3.3). For example, many birds, mam-

64

Dustin J. Penn and Joachim G. Frommen

mals and fish rely on location to recognise offspring, and use the rule: ‘any young in my nest are mine’. For example, male house mice (Mus musculus) normally kill pups, but after mating with a female, they no longer attack her pups (Elwood and Ostermeyer 1984). They thus use mating as a contextual cue to assess the probability of being related to the pups (statemediated recognition). Female moorhens (Gallinula chloropus) destroy any eggs they find in their nest if they have not begun to lay eggs, and they apparently use their knowledge of their own egg-laying to bias investment into their own offspring (McRae 1996). Such decision rules are not necessarily all-or-none. In dunnocks (Prunella modularis), males allocate parental care based on the amount of sexual access they had to a mate, which correlates with offspring paternity (Davies et al. 1992, Hartley et al. 1995). Some contextual cues may be easily exploited by cheaters, but there may be no cheat-proof mechanisms. Extreme examples are birds that care for a Table 3.3 Kin recognition (KR) has been found in a wide variety of species, and these are some arbitrarily chosen examples from different taxa. Taxa

Species

Behaviour

Reference

1. Contextual cues Burying beetles (Necrophorus vespilloides)

Mothers kill larvae that hatch too early to be own young

Müller and Eggert 1990

Birds

Bank swallows (Riparia riparia)

Treat all chicks in their nest as their own, as long as chicks are younger than 14 days

Beecher et al. 1981

2. Phenotypic cues a. Direct familiarity Predatory mites (Phytoseiulus persimilis)

Cannibalise non-kin and unfamiliar kin, but not familiar kin

Schausberger 2007

Birds

Barnacle geese (Branta leucopsis)

Breed near sister, but only when they were born in the same year

van der Jeugd et al. 2002

Mammals

Barbary macaques (Macaca sylvanus)

Avoid mating with individuals with whom they were familiarised during early life

Kuester et al. 1994

Broad sense KR

Insects

Broadest sense KR

Insects

Kin recognition: conceptual issues, mechanisms and evolutionary theory

Broadest sense KR

b. Without direct familiarity (i) Indirect familiarity Prefer to shoal with unfamiliar kin, based on learning of sibling’s cues

Frommen et al. 2007a,b

Fishes

Pelvicachromis taeniatus

Both sexes prefer familiar as well as unfamiliar kin as mates

Thünken et al. 2007a,b

Mammals

House mice (Mus musculus)

Males avoid mating with females carrying MHC genes of the family with which they are reared

Yamazaki et al. 1988

(ii) Self inspection Plants

Sea rockets (Cakile edentula)

Grow additional roots when share soil with non-kin

Dudley and File 2007

Birds

Peafowl (Pavo cristatus)

Peacocks, raised with non-kin, lek with relatives

Petrie et al. 1999

Amphibians

Clawed frogs (Xenopus leavis)

Preferred to group with siblings with which they shared MHC haplotypes to those with no MHC haplotypes in common

Villinger and Waldman 2008

Prefer forming chimeras with related cells

Queller et al. 2003

Broad sense KR

Three-spined stickleback (Gasterosteus aculeatus)

Narrow sense KR

Fishes

65

(iii) Green-beard genes Social amoebae

Dictyostelium purpureum

cuckoo’s chick in their nest. Furthermore, contextual cues may not allow recognising kin outside the respective context. Because conspecific phenotypic cues are not involved, and mistakes are possible, some researchers argue that species that utilise contextual mechanisms do so because they lack the ability to recognise kin (Barnard et al. 1991, Tang-Martinez 2001). Indeed, classifying contextual cues as KR seems an odd use of the term recognition, but perhaps such cues are commonly utilised in most other types of recognition systems. Contextual mechanisms are KR in the broadest sense of the term, and some may instead prefer to consider them as kin bias without kin recognition (see BOX 3.1).

66

Dustin J. Penn and Joachim G. Frommen

3.3.1.2 Phenotypic cues Some suggest that the term KR should be restricted for mechanisms that rely on direct cues from conspecifics, including chemical, acoustic, visual, behavioural or other phenotypes (Barnard et al. 1991, Tang-Martinez 2001). These may be features of individuals or groups (e.g., colony odours). The first type of phenotypic recognition we address is another example of KR in the broad sense. A. Direct familiarity (also called ‘prior association’) There is a great deal of evidence that animals rely on direct familiarity to recognise kin (e.g., Table 3.3). This mechanism can be found in many species where young stay together during the first days of their life. Here, the phenotypes of individuals who are kin – or likely to be kin – are individually learned, classified and later on treated as related or not. Close kin are often more familiar than non-kin, and a wide variety of species use previous association to recognise kin (e.g., Westermarck 1891, Porter et al. 1981, Dewsbury 1982, Holmes and Sherman 1982). They follow a rule: treat familiar individuals as kin because they are likely to be kin. Individuals may become classified as kin if they become familiar at particularly sensitive times during ontogeny. For example, humans and other mammals negatively imprint on individuals with which they are reared, and subsequently avoid mating as adults. They follow the rule: avoid mating with individuals I grow up with in my family. This so-called Westermarck effect was the first mechanism discovered to control kin discrimination in humans (BOX 3.3). Tang-Martinez (2001) suggests that kin biases based on familiarity should be regarded as individual rather than kin recognition. These are not mutually exclusive alternatives, however, and it seems moot to debate whether to call such behaviours kin biases based on individual recognition or kin recognition based on individual familiarity. As familiarity is not considered to be KR by some researchers (but instead, it is seen as something that must be controlled to test KR), we consider it to be KR in the broad sense. Regardless, determining how animals distinguish kin versus non-kin among unfamiliar individuals is clearly one of the more challenging problems in this field. There are at least two types of mechanisms that do not require familiarity, and we consider these to be KR in the narrow sense.

Kin recognition: conceptual issues, mechanisms and evolutionary theory

67

BOX 3.3 Kin recognition in Homo sapiens

In his autobiography, Hamilton (1996) commented that he ‘did not anticipate the degree of relevance to humans that the findings eventually proved to have’. However, he surely realised that incest avoidance is a topic that has puzzled thinkers since ancient times (Porter 1991). Like other animals, humans show kin biases in a wide array of contexts, including incest avoidance, parental care, and cooperative behaviour. The mechanisms they use are multifarious, too. Maybe the best known is direct familiarity (also known as the Westermarck effect). Westermarck (1981) described that small children are negatively imprinted on their age mates, avoiding them as partners when reaching adulthood. Similar results were found by Lieberman et al. (2007) who showed that kinship is estimated by the perinatal association of the proposed sibling with the individual’s genetic mother as well as through the duration of time siblings live together. Many studies show that human offspring are able to recognise their mothers through volatile scent cues, and vice versa (Porter 1991). There is suggestive evidence that humans are also capable of recognising unfamiliar kin as well, and self-reference may play a role. For example, DeBruine (2002) found people trust a stranger’s face more when it has been morphed with their own than when it was left unchanged. Familiarity was ruled out by using morphs of celebrities; only selfresemblance mattered. The use of self-reference in human KR was further supported by a study that found that dizygotic as well as monozygotic twins preferred pictures of faces that were merged with their own over faces that were merged with the face of the twin (Bressan and Zucchi 2009). B. Without direct familiarity (1) Indirect familiarity (also called ‘phenotype matching’) Individuals can potentially recognise kin even among unfamiliar individuals by comparing the phenotypes of putative kin with those of known (familiar) kin: individuals who resemble their own kin are treated as related. There are examples of this KR mechanism in many species in which the young stay together during the first days of their life (Table 3.3). For example, in brood-caring fish, fry have ample opportunities to learn cues (e.g., smell) of their siblings or parents before they leave the nest or cave. Later on in life they use these cues to

68

Dustin J. Penn and Joachim G. Frommen

recognise not only their nest mates, but also relatives raised in other broods. This mechanism not only allows recognising unfamiliar siblings, but also more distantly related kin like half-siblings or cousins (see BOX 3.4). Its weakness is that in species with close contact between kin and non-kin during maturation (e.g., in colonial species) or with high amount of stolen fertilisations, individuals may learn the BOX 3.4 MHC genes: candidate genetic loci for kin recognition cues The genes of the major histocompatibility complex (MHC) are a multigene family of highly polymorphic loci that control immunological self/nonself recognition, and are suspected to play a role in KR (Brown and Eklund 1994). MHC genes influence odour and mating preferences in a variety of species, including mice, humans, fish, and frogs (Yamazaki et al. 1988, Penn and Potts 1998, Milinski 2006). MHC-dependent mating preferences may function to increase or optimise offspring MHC-heterozygosity (e.g., Penn and Potts 1998, Penn et al. 2002), avoid inbreeding, or both (Penn 2002). Female mice prefer to nest communally with sisters, or MHC-identical individuals when sisters are unavailable (Manning et al. 1992). Also, female mice are more likely to retrieve pups if the pups are MHC-identical to the dam (Yamazaki et al. 2000). It is unclear how MHC genes influence odour, though several candidate volatiles have been identified (Willse et al. 2005, 2006, Novotny et al. 2007), and MHC-derived peptides have been shown to be detected through the vomeronasal organ (Leinders-Zufall et al. 2004). Interestingly, house mice avoid mating with individuals carrying MHC genes that are identical to the foster family with which they are reared (familial imprinting; Yamazaki et al. 1988, Penn and Potts 1998). This behaviour may provide a more effective mechanism for avoiding sib matings than self-inspection. For example, consider the MHC-genotypes of closely related mice, shown below (most individuals are heterozygous since MHC genes are polymorphic). If ac relies entirely on self-inspection, she will risk mating with 0.25 of her siblings (bd) and 0.5 of her half-siblings (de and df). In contrast, by using familial imprinting, ac can avoid mating with all full-siblings (bc, ad, bd), all half-siblings (ce, cf, de, df), and half of all cousins (Penn and Potts 1999). By imprinting on a variety of phenotypic cues of family members, this may provide a highly effective mechanism to avoid inbreeding (i.e., indirect familiarity). ab

ac

bc

cd

ad

ef

bd

ce

cf

de

df

Kin recognition: conceptual issues, mechanisms and evolutionary theory

69

cues of relatives as well as non-relatives and treat all individuals bearing these cues similarly. This mechanism is often called phenotype matching, and is seen categorically distinct from familiarity, but the crucial difference is that with phenotype matching the animal learns self- or familial- rather than individually-distinctive cues (Porter 1988, Barnard et al. 1991, Tang-Martinez 2001). (2) Self-inspection (also called ‘self-referent phenotype matching’ and the ‘armpit effect’) It has been suggested that to recognise kin, animals might inspect their own scent, voice or other phenotypic cues and compare how well they match with other individuals, and use the similarity to recognise kin (armpit effect, Dawkins 1976). There are some, though still surprisingly few candidates for KR through self-inspection (Table 3.3), and these are often found in unicellular organisms, fungi and plants, rather than animals (reviewed in Tsutsui 2004; Table 3.3). Mateo and Johnston (2000) reported evidence for self-referent phenotype matching in golden hamsters (Mesocricetus auratus). However, their study is criticised for several reasons, e.g., not ruling out post-partum and earlylife familiarisation with close kin through fetal olfaction (e.g., Heth and Todrank 2001, Mateo and Johnston 2001, Hare et al. 2002, Mateo and Johnston 2003). Self-inspection may evolve when the opportunity to learn relative’s cues is limited or unreliable due to a lack of family members to use for comparison, or due to the presence of unrelated individuals in the same nest or burrow. Self-inspection is often viewed as categorically distinct from other learning mechanisms, yet it is arguably a type of phenotype matching, in which the cues used for comparison are not learned from conspecifics but from self. Self-inspection might also be considered to be a type of familiarity (associative learning) (though if it involves habituation to one’s own cues, this is non-associative rather than associative learning). Oddly, self-reference matching is widely considered a form of genetic KR in the strict sense (recognition alleles, see below) even though genetically determined cues may or may not be involved. A problem with the self-inspection hypothesis is that it is unclear how it might be empirically distinguished from green-beard effects. (3) Green-beard genes (also called ‘recognition alleles’, and ‘genetic kin recognition’) As mentioned in the introduction, when speculating on how animals might evolve mechanisms to recognise their kin, Hamilton (1964) postulated the evolution of a super gene that is able to recognise and help copies of itself in other individuals (Haig 1996), and this model be-

70

Dustin J. Penn and Joachim G. Frommen

came known as a green-beard gene (Dawkins 1976, 1982). A greenbeard gene helps copies of itself by encoding three traits: (1) influencing the expression of a rare phenotypic trait, such as a green beard; (2) enabling individuals to recognise the trait in others; and (3) acting altruistically towards individuals carrying the trait (or antagonistically towards individuals that do not). In other words, the same gene influences all these traits simultaneously through pleiotropy. Green-beard effects may be controlled by single or multiple loci. If their effects are controlled by multiple loci, this might be considered a genetic KR system (Grafen 1990). The distinction between green-beard effects and other types of genetic KR mechanisms is fuzzy in theoretical models (Rousset and Roze 2007). Green-beard genes are often dismissed as being far-fetched; even Hamilton recognised that it is unlikely that a single gene could ever influence all these traits. Some argue that green-beard genes are unlikely because they are genetic outlaws, as they help individuals bearing similar cues who are sometimes unrelated, this leads to intragenomic conflicts, which favours the evolution of suppressor genes (the parliament of the genome; Alexander and Borgia 1978, Helantera and Bargum 2007). Others argue that green-beard genes are not necessarily outlaws, as the rest of the genome will reap any benefits obtained by a green-beard gene (Ridley and Grafen 1981). The first report for a green-beard gene was found in a study on spiteful behaviour in fire ants (Solenopsis invicta; Keller and Ross 1998), though these findings have been challenged (Vander Meer and Alonso 2002). Subsequently, several candidate green-beard effects have been reported, including cooperation among budding yeast (Smukalla et al. 2008), social amoeba (Queller et al. 2003) and lizards (Sinervo et al. 2006), parent-offspring recognition at the placenta (Haig 1996, Summers and Crespi 2005), altruistic sperm pairing (Moore and Moore 2002), and self-incompatibility loci in plants (reviewed in Tsutsui 2004). As mentioned above, the main problem with the green-beard gene hypothesis is empirically distinguishing it from self-inspection: it requires eliminating all phenotypic cues that an organism could obtain about its own phenotypic cues, which seems difficult if not impossible. Therefore, many evolutionary researchers combine them together into tag-based kin recognition (e.g., Axelrod et al. 2004, Antal et al. 2009), as we will see in the next section. In summary, there is much evidence that many species can distinguish kin and non-kin, and some evidence to support all of the proposed mechanisms. Furthermore, the various types of KR mechanisms are not mutually exclusive, as they may be used by different individu-

Kin recognition: conceptual issues, mechanisms and evolutionary theory

71

als within a species, or they may change during development, age, or an individuals’ condition or context (Neff and Sherman 2002, Mateo 2004). 3.3.2 Key components of KR mechanisms Even though there are several types of KR mechanisms, it has been suggested that all KR systems must have several key components or design features to enable individuals to recognise and respond differentially to kin (e.g., Sherman and Holmes 1985, Waldman 1987; Fig. 3.1). From this perspective, KR is not a single system, but rather a device or contraption with many components. First, KR requires a sender that produces kinship cues (also known as markers, labels, signatures, and tags), which are phenotypic traits produced by the organism, such as chemical, visual, or acoustic signatures, that provide information about kinship. Kinship cues must be highly variable among individuals – and also consistent or developmentally stable within each individual, and the variability of cues must be genetic or correlated with genetic relatedness (e.g., see BOX 3.4). KR cues may function specifically to indicate relatedness, or more likely they are produced incidentally, with no special function for communicating relatedness. For example, Todrank and Heth (2003) suggest that animals produce odour signals (which are usually a blend of compounds or ‘signature mixtures’ sensu Wyatt 2003) that have a wide range of functions, from individual to kin to species recognition. The broadest definition of KR, as we have seen, also includes indirect spatial, temporal or other contextual cues associated with kinship, as well as direct phenotypic cues, though these would not be an example of kin communication (see below). Second, recognition begins with sensory detection (or perception) of kinship cues, such as through olfaction, visual, hearing, or tactile mechanisms by receivers (also called discriminators). Usually, only one type of sensory modality is studied, though a combination may be involved (multimodal processing. Next, recognition requires a phenotype matching mechanism: a hypothetical device that compares a sender’s cues to a template, an internal model or representation of cues from self or other referent, and evaluates the similarity between them, as necessary for classification (Lacy and Sherman 1983). A template may be learned from self (selfinspection) or from other referents, even though phenotype matching is often seen as categorically distinct from self-inspection. Recognition errors occur due to excess over-lap in cues, and one of the central problems for KR systems is to optimise the risk of acceptance errors (Type I or false

72

Dustin J. Penn and Joachim G. Frommen

Sender: (1) Cues Display chemical, visual, acoustic information e.g., scent marking

Contextual information (e.g., spatial cues)

Receiver: (2) Recognition Perception (sensory detection) e.g., olfaction Phenotypic matching e.g., self-inspection

(3) Decision-rules / Heuristics e.g., avoid similar smelling mates

Receiver’s physiological state and ecological context

(4) Action e.g., inbreeding avoidance

Fig. 3.1 Kin recognition (or kinship communication) involves several key components, from (1) the production of cues by senders to (2) recognition, (3) decisions, and (4) actions by receivers. Recognition per se includes cue detection and phenotypic matching, which are used for making decisions regarding actions. Spatial, temporal, and other contextual cues may be used, at least for KR in the broadest sense. Self-inspection (shown above) and indirect familiarity are examples of KR mechanisms in the narrow sense. Decision rules involve acceptance thresholds and simple rules-of-thumb that enable receivers to evaluate the likely relatedness of conspecifics. Decisions may also integrate contextual information about the receivers’ own age, condition, physiological state, and ecological constraints to weigh the perceived costs and benefits (adapted from Sherman and Holmes 1985, Waldman et al. 1988; mice drawn by Shawn Meagher).

positives), due to accepting ‘undesirable recipients’, with rejection errors (Type II or false negatives), due to rejecting ‘desirable recipients’ (Reeve 1989, Beecher 1991, Sherman et al. 1997).

Kin recognition: conceptual issues, mechanisms and evolutionary theory

73

Third, after recognition, receivers use this information for making decisions, which are cognitive or other information processing devices that control behavioural and other responses (or store the information into memory). Adaptive decisions also integrate contextual information about the receivers’ internal state or motivation, and other relevant information. For example, plains spadefoot toad tadpoles (Spea bombifrons) cannibalise other tadpoles and they eat more non-kin than kin, except for when they are very hungry; then they become indiscriminate (Pfennig et al. 1993). The importance of the context-dependence of acceptance thresholds should not be underestimated. Fourth, receivers’ decisions enable them to take the appropriate action, which may be developmental, physiological or behavioural responses that result in differential treatment of kin (e.g., Liebert and Starks 2004). For example, this occurs when a mother rejects a foreign offspring because she detects that it has previously been licked or labeled by another female. Actions can be all-or-none, such as when there is a critical threshold above which all senders are accepted and below which they are rejected. However, whether receivers take action or not depends on the relative rates of interaction with and the fitness consequences of accepting and rejecting desirable and undesirable recipients (Reeve 1989). As we will see later, recognition does not necessarily lead to discrimination if, for example, the perceived costs outweigh the benefits. This summary of KR components differs somewhat from previous versions (Sherman and Holmes 1985, Waldman 1987). For example, we do not consider cues and actions as components of recognition per se, even though cues are necessary for recognition to occur. We divide recognition and decision-rules into separate components, because they may be mechanistically distinct. Referring to all these steps as components of recognition overlooks the communication aspects of the system, and particularly selection on the sender’s cues (Beecher 1991). Therefore, this model is more accurately described as components of kinship communication rather than a KR system, and communication is not necessarily honest. This is an important point for evolutionary analyses of KR, which we address in the next sections.

3.4 Evolutionary analyses of kin recognition KR research has always been grounded on evolutionary analyses of social behaviour, and particularly Hamilton’s ideas about nepotism (see Holmes 2004). In this section, we provide an overview of theoretical analyses that

74

Dustin J. Penn and Joachim G. Frommen

have aimed to understand how KR recognition evolves, and in particular, how natural selection maintains the diversity of genes controlling KR (in the narrow sense). These analyses have only considered the evolution of KR in mediating cooperation (nepotism) so far, and not inbreeding avoidance or other kin biases, and they have only examined self-inspection (or green-beard effects), and not considered indirect familiarity. 3.4.1 Evolutionary origins and maintenance of genetic kin recognition It has been difficult to understand how KR might evolve and be maintained by natural selection. Crozier (1986) made the first mathematical model of the evolution of KR by examining its role in nepotism, and he concluded that it is unstable, and inevitably drives itself to extinction. He found that individuals bearing common phenotypic cues (markers) have more opportunities to find and engage in cooperative activities than those bearing rare markers, and therefore, they gain greater fitness. The common markers become even more common over time, and eventually all individuals have the same marker. Once a marker no longer provides an indicator of kinship, there is no benefit to KR. In short, rare alleles will be at a disadvantage in finding cooperators, which eliminates the marker diversity, and any benefits of KR. Nevertheless, Crozier (1987) suggested that genetic variation in KR cues might be maintained by piggy backing on other loci under balancing selection, such as inbreeding avoidance, parentoffspring recognition, individual recognition, or a completely different source of selection, such as parasites. Grafen (1990) pointed out that Crozier’s model assumed that social interactions are always beneficial as there are no cheats (social parasites that display a marker and exploit other’s cooperation), and therefore, there is no benefit to limiting interactions to kin. Rare markers provide better indicators of kinship, which would favour greater altruism among individuals carrying rare markers. He concluded that ‘cheating maintains genetic polymorphisms at the matching locus because common alleles are hit harder when cheating arises, because they are too trusting’. On the other hand, if the marker and conditional helping evolve independently, then cheaters, displaying the correct marker without providing help, might undermine the system. Axelrod et al. (2004) conducted a simulation in which heritable cues of relatedness, which they called tags, can coevolve with discrimination and other strategies based on these indicators. They found that conditional or discriminating altruism can be maintained in a stable equilibrium.

Kin recognition: conceptual issues, mechanisms and evolutionary theory

75

Jansen and van Baalen (2006) made a similar simulation based on a recognisable cue (coloured beards) to study the evolutionary dynamics of beard colour polymorphism (beard chromodynamics). They found that complete linkage between the matching and conditional helping loci leads to highly unstable dynamics, and to the rapid loss of tag variation (beard colours). However, when they allowed low recombination, beard colour and conditional helping were maintained at intermediate frequency in the population. This is in contrast to the assumption that tight coupling (pleiotropy or close linkage) is necessary for green-beard effects. Thus, it appears that cooperative genes must continuously change their tags to avoid exploitation by defectors or social parasites, much like red queen dynamics in host-parasite interactions. Rousset and Roze (2007) constructed a model using two-loci that control matching and helping, and they examined how various rates of recombination and dispersal (spatial population structure) affect the maintenance of polymorphisms. They found that selection usually eliminates KR and helping due to the benefits for common markers, just as Crozier (1986) found. Yet, when they reduced dispersal and recombination between the matching and helping loci, the polymorphisms were maintained: rare markers benefit from greater levels of helping, just as Grafen (1990) suggested. Increasing mutation rates help to maintain polymorphisms at both loci, as mutations supply new alleles into the system as they go extinct, but this still requires restrictive conditions. Perhaps the most interesting finding is that adding an extrinsic selective advantage for rare alleles, such as negative-frequency dependent selection from parasites, effectively maintains KR and conditional helping (Gardner and West 2007), as Crozier (1986) also suggested. More recently, Antal et al. (2009) used an analytical approach to model the evolution of conditional cooperation based on tag-based recognition of phenotypic similarity (green-beard or self-referent effects). Unlike previous models, they found that the evolution of cooperation does not require spatial structure. They also found that cooperation is more likely to evolve if the strategy mutation rate is controlled by one or few genes, and the phenotypic tags are encoded by many loci. In summary, nepotism mediated by KR may be prone to drive itself to extinction, but not always. It can be maintained under certain conditions, namely linkage between loci and subdivided population structure, and from extrinsic sources of balancing selection – such as from parasites. Its red queen dynamic nature suggests that its occurrence will vary over time and space, i.e., among populations and even within populations of the same species. We should expect genetic KR driven nepotism to be found in species with low dispersal and recombination. Evolutionary analyses

76

Dustin J. Penn and Joachim G. Frommen

help explain why MHC genes, which control immunological recognition of parasites, are also implicated in KR (Gardner and West 2007; BOX 3.3). Although these models suggest that nepotism through genetic KR is unlikely to evolve without spatial structure or other restrictive conditions, a recent analysis suggests otherwise (Antal et al. 2009). 3.4.2 Apparent kin recognition failures There are many examples in which animals do not seem to recognise their kin, even when it would seem to be in their interest to do so. For example, when male birds are cuckolded, they do not eject the extra-pair offspring from the nest. It is instructive to consider apparent KR failures (Beecher 1991, Sherman et al. 1997), and the evolutionary hypotheses to explain them. First, KR (or KB) may not occur because the costs of discrimination outweigh the benefits. There are surely potential tradeoffs for any kinbiased behaviours, so that the optimal response may sometimes require taking no action or discrimination, which might be mistaken as a recognition failure. For example, it has been suggested that the benefits of rejecting extra-pair offspring by male birds may not be worth the costs of making recognition mistakes, and erroneously killing their own offspring (Kempenaers and Sheldon 1996). There may be other costs for males that discriminate against extra-pair offspring, such as risk of desertion by their mate. It is instructive to consider that the body’s immune system functions to recognise parasites; however, a lack of immune responses to a parasite is not necessarily due to a recognition failure. Mounting an immune response can help resolve an infection, but it also has costs, especially in case of autoimmunity and other forms of immunopathology. Since immune recognition has detrimental side-effects, selection sometimes favours immune tolerance to infection. Similarly, just as selection favours optimal immunity and tolerance rather than maximal responsiveness, we should expect kin discrimination responses to be optimised to balance the potential fitness costs and benefits. This decision problem faced by receivers has been studied in several theoretical models (Reeve 1989, Beecher 1991), but it has not been empirically tested to our knowledge. Second, KR may not occur because signalers are under selection to conceal their identity to potential receivers (anti-recognition hypothesis) (Beecher 1991). For example, fathers may not recognise their own offspring because in species with high rates of extra-pair paternity, offspring may be favoured to conceal their identity so that they do not resemble their genetic fathers (also called neonatal identify deception) (Beecher 1991,

Kin recognition: conceptual issues, mechanisms and evolutionary theory

77

Pagel 1997). Most analyses assume that signalers honestly signal their identity, even if only inadvertently, and do not consider the possibility that individuals evolve mechanisms to disguise themselves. Yet, one theoretical analysis shows that selection can favour concealment over a wide range of conditions (Johnstone 1997). Moreover, it shows that kin discrimination will not evolve just because it benefits receivers: it depends on whether honest signature cues are also beneficial to signalers. This is called the communication problem, as it involves the evolution of the sender’s cues, as well as the receiver’s KR abilities (Johnstone 1997). Similarly, the failure of the immune system to recognise invading pathogens is sometimes due to pathogens evolving mechanisms to escape immune recognition. Interestingly, such deception games may be required for selection to maintain cooperation through KR (Grafen 1990), as we previously discussed. Finally, a failure to recognise kin may be maladaptive due to insufficient time to evolve and adapt to new environmental conditions (evolutionary lag-time hypothesis; see Dawkins 1982), or as we have seen, selection from kin-biased behaviours eroding the genetic polymorphisms required to maintain itself (Crozier 1987). Such maladaptive ideas are hypotheses of last resort however, as testing requires ruling out the functional hypotheses.

3.5 Conclusions and future directions Many species have the ability to distinguish kin from non-kin, which may function to facilitate helping kin, avoiding inbreeding, or other kin-biased behaviours. Determing whether a species utilises KR or not depends on how one defines this term, and there is much debate over how to define it. It might be useful to replace the term KR with KD or perhaps kin detection, kin identification, or kinship communication, though these alternatives have their own problems. We suggest that it is useful to retain broad definitions for KR, and treat the strict definitions as hypotheses about the mechanisms and functions. There is also much debate about how to classify the various types of KR mechanisms, and as a pragmatic approach, we suggest classifying the mechanisms according to whether they fit the broad or narrow-sense definitions of KR. Finally, recent theoretical studies are helping to explain the evolution of nepotism through KR, as well as apparent KR mistakes, errors and failures.

78

Dustin J. Penn and Joachim G. Frommen

There are still many unresolved problems regarding KR, and here, we suggest a few ideas for future research on both mechanistic and evolutionary questions: (1) The proximate mechanisms through which KR affects KB’s are unclear, and more work is needed here. Efforts to find candidate genetic loci controlling KR could narrow the search by focusing on linked loci that are highly polymorphic, and control immune recognition to pathogens and parasites, and not only MHC genes (BOX 3.4). (2) The adaptive functions of KR are not well understood (Blaustein et al. 1991). KD is thought to increase the indirect fitness of the individual through cooperation; however, experimental evidence for this hypothesis is still scarce, opening ample opportunities for future studies. (3) It has been suggested that somatic incompatibility (or allorecognition) systems of colonial organisms might function for KR, and that these have multiple phylogenetic origins (Crampton and Hurst 1994). These ideas deserve further analyses. (4) It is often assumed that a species has either evolved KR or not, and when different studies report mixed results the positive evidence is treated with scepticism. Yet, theoretical models predict much intraspecific variation in KR abilities within and among populations, as well as dynamic changes over time and according to the varying ecological constrains. For example, the report of green-beard genes in fire ants (Keller and Ross 1998) was not supported in a different study (Vander Meer and Alonso 2002) perhaps due to geographic variation (also see Liebert and Starks 2004). Future studies should examine such variability, and also how mechanisms may change during development, age, condition or context (Neff and Sherman 2002, Mateo 2004). (5) There are several models on the evolution of genetic KR, but what are the conditions in which natural selection favours the evolution of KR through familiarity, indirect familiarity and other types of learning? (6) Theoretical analyses of the evolution of KR are restricted to the evolution of conditional helping (nepotism), and future models need to consider inbreeding avoidance and other potential benefits (and preferences for mating with kin to increase inclusive fitness benefits).

Kin recognition: conceptual issues, mechanisms and evolutionary theory

GLOSSARY Green-beard gene: a postulated gene that simultaneously influences the development of a characteristic phenotype (cue), the ability to recognise this cue, and also helping conspecifics bearing this cue or label (or harming those who do not). Genetic relatedness: estimated as the proportion of genes shared, or the probability of alleles at a randomly chosen locus, between two individuals being shared due to common descent (Sewall Wright’s coefficient of relationship r). There is no absolute measure of relatedness among individuals, as it is relative to the population mean, which can result in negative relatedness. Thus, relatedness is not simply a measurement of genetic similarity between individuals, as it depends on the population structure (Griffin and West 2002). Heuristics: simple rules-of-thumb for detection, recognition, making classifications or other decisions that are efficient and accurate under most circumstances, even though they can also lead to errors. These appear to be used in all KR mechanisms. Inclusive fitness: an individual’s own reproductive success (conventional fitness) plus its effects on the reproductive success of its relatives, each one weighed by the coefficient of relatedness (Dawkins 1982). Kin: individuals that are genetically related due to common descent (e.g., offspring, siblings, cousins, etc.). Kinship: a special case of genetic similarity in which the probability of individuals sharing an allele at a particular locus depends upon their distance in a path of common descent (Grafen 1990). Kin bias (KB): differential treatment of kin versus non-kin, which may be due to kin recognition or not (Barnard 1991). Kin discrimination (KD): in the broad sence, the ability to distinguish (KR) and the differential treatment (KB) of kin versus non-kin. In the narrow sense, KB due to KR (Barnard 1991). Kin recognition (KR): in the broad sense, the ability to identify, distinguish and classify kin versus non-kin, though there are several narrow versions (see Table 3.1). We use the term in this broad sense, unless otherwise indicated. Kin selection: natural selection due to interactions among kin, such as nepotistic helping behaviours. There are several common misunderstandings about kin selection (see Dawkins 1979), which is why Hamilton disliked this term. Nepotism: a form of conditional helping, in which altruism is provided to close kin (parental care is a special case of nepotism). This behaviour is often referred to as kin selection. Recognition alleles: in the broad sense, genes that control KR (or at least the production of cues and ability to recognise kinship cues). In the literature this term is used for green-beard genes, self-inspection, or both.

79

80

Dustin J. Penn and Joachim G. Frommen

Acknowledgements We are very grateful to Z. Tang-Martinez, who offered particularly thoughtful comments on the manuscript. We thank A. Hettyey, T. Thünken, S.A. Baldauf, S.M. Zala, an anonymous referee, as well as many students for their helpful comments on earlier versions, S. Meagher for the illustrations of mice’ and R. Hengsberger for assistance with editing figures and tables. Finally, we thank P.M. Kappeler for editing the volume and inviting us to write a chapter.

References Alexander RD, Borgia G (1978) Group selection, altruism, and the levels of organization of life. Annu Rev Ecol Syst 9:449-474 Antal T, Ohtsuki H, Wakeley J, Taylor PD, Nowak MA (2009) Evolution of cooperation by phenotypic similarity. Proc Natl Acad Sci USA 106:8597-8600 Axelrod R, Hammond RA, Grafen A (2004) Altruism via kin-selection strategies that rely on arbitrary tags with which they coevolve. Evolution 58:1833-1838 Barnard CJ (1991) Kinship and social behavior: the trouble with relatives. Trends Ecol Evol 6:310-312 Barnard CJ, Hurst JL, Aldhous P (1991) Of mice and kin: the functional significance of kin bias in social behavior. Biol Rev 66:379-430 Beecher MD (1991) Success and failures of parent-offspring recognition in animals. In: Hepper PG (ed) Kin Recognition. Cambridge University Press, Cambridge, pp 95-124 Beecher MD, Beecher IM, Hahn S (1981) Parent-offspring recognition in bank swallows (Riparia riparia). II. Development and acoustic basis. Anim Behav 29:95-101 Blaustein AR, Bekoff M, Byers JA, Daniels TJ (1991) Kin recognition in vertebrates: what do we really know about adaptive value? Anim Behav 41:10791083 Bressan P, Zucchi G (2009) Human kin recognition is self-rather than familyreferential. Biol Lett 5:336-338 Brown JL, Eklund A (1994) Kin recognition and the major histocompatibility complex: an integrative review. Am Nat 143:435-461 Byers JA, Bekoff M (1986) What does ‘kin recognition’ mean? Ethology 72:342345 Byers JA, Bekoff M (1991) Development, the conveniently forgotten variable in ‘true kin recognition’. Anim Behav 41:1088-1090 Clarke FM, Faulkes CG (1999) Kin discrimination and female mate choice in the naked mole-rat Heterocephalus glaber. Proc R Soc Lond B 266:1995-2002 Crampton WGR, Hurst LD (1994) True kin recognition, in the form of somatic incompatibility, has multiple independent origins. Anim Behav 47:230-234

Kin recognition: conceptual issues, mechanisms and evolutionary theory

81

Crnokrak P, Roff DA (1999) Inbreeding depression in the wild. Heredity 83:260270 Crozier RH (1986) Genetic clonal recognition abilities in marine invertebrates must be maintained by selection for something else. Evolution 40:1100-1101 Crozier RH (1987) Genetic aspects of kin recognition: concepts, models, and synthesis. In: Fletcher DJC, Michener CD (eds) Kin Recognition in Animals. Wiley, New York, pp 55-73 Davies NB, Hatchwell BJ, Robson T, Burke T (1992) Paternity and parental effort in dunnocks Prunella modularis  how good are male chick-feeding rules. Anim Behav 43:729-745 Dawkins R (1976) The Selfish Gene. Oxford University Press, New York Dawkins R (1979) Twelve misunderstandings of kin selection. Z Tierpsychol 51:184-200 Dawkins R (1982) The Extended Phenotype: the Gene as the Unit of Selection. WH Freeman, Oxford DeBruine LM (2002) Facial resemblance enhances trust. Proc R Soc Lond B 269:1307-1312 Dewsbury DA (1982) Avoidance of incestuous breeding between siblings in two species of Peromyscus mice. Biol Behav 7:157-169 Dudley SA, File AL (2007) Kin recognition in an annual plant. Biol Lett 3:435438 Elwood RW, Ostermeyer MC (1984) Does copulation inhibit infanticide in male rodents? Anim Behav 32:293-294 Fletcher DJC, Michener CD (1987) Kin Recognition in Animals. John Wiley, New York Frommen JG, Luz C, Bakker TCM (2007a) Kin discrimination in sticklebacks is mediated by social learning rather than innate recognition. Ethology 113:276282 Frommen JG, Mehlis M, Brendler C, Bakker TCM (2007b) Shoaling decisions in three-spined sticklebacks (Gasterosteus aculeatus) – familiarity, kinship and inbreeding. Behav Ecol Sociobiol 61:533-539 Galef BG Jr (1976) Social transmission of acquired behavior: a discussion of tradition and social learning in vertebrates. Adv Stud Behav 6:77-100 Gamboa GJ, Reeve HK, Holmes WG (1991) Conceptual issues and methodology in kin-ecognition research: a critical discussion. Ethology 88:109-127 Gardner A, West SA (2007) Social evolution: the decline and fall of genetic kin recognition. Curr Biol 17:R810-R812 Gould SJ, Vrba ES (1982) Exaptation – a missing term in the science of form. Paleobiology 8:4-15 Grafen A (1990) Do animals really recognise kin? Anim Behav 39:42-54 Grafen A (1991a) Kin vision? A reply to Stuart. Anim Behav 41:1095-1096 Grafen A (1991b) A reply to Blaustein et al. Anim Behav 41:1085-1087 Grafen A (1991c) A reply to Byers and Bekoff. Anim Behav 41:1091-1092 Griffin AS, West SA (2002) Kin selection: fact and fiction. Trends Ecol Evol 17:15-21

82

Dustin J. Penn and Joachim G. Frommen

Grosberg RK, Quinn JF (1986) The genetic control and consequences of kin recognition by larvae of a colonial marine invertebrate. Nature 322:456-459 Haig D (1996) Gestational drive and the green-bearded placenta. Proc Natl Acad Sci USA 93:6547-6551 Halpin ZT (1986) Individual odors among mammals: origins and functions. Adv Stud Behav 16:39-70 Halpin ZT (1991) Kin recognition cues of vertebrates. In: Hepper PG (ed) Kin Recognition. Cambridge University Press, Cambridge, pp 220-258 Hamilton WD (1964) The genetical evolution of social behaviour I. J Theor Biol 7:1-16 Hamilton WD (1975) Innate social aptitudes of man: an approachfrom evolutionary genetics. In: Fox R (ed) ASA Studies 4: Biosocial Anthropology. Malaby Press, London, pp 133-153 Hamilton WD (1996) Narrow Roads of Gene Land. WH Freeman, Oxford Hare JF, Sealy SG, Underwood TJ, Ellison KS, Stewart RLM (2002) Evidence of self-referent phenotype matching revisited: airing out the armpit effect. Anim Cogn 6:65-68 Hartley IR, Davies NB, Hatchwell BJ, Desrochers A, Nebel D, Burke T (1995) The polygynandrous mating system of the alpine accentor, Prunella collaris. II. Multiple paternity and parental effort. Anim Behav 49:789-803 Helanterä H, Bargum K (2007) Pedigree relatedness, not greenbeard genes, explains eusociality. Oikos 116:217-220 Hepper PG (1991a) Kin Recognition. Cambridge University Press, Cambridge Hepper PG (1991b) Recognizing kin: ontogeny and classification. In: Hepper PG (ed) Kin Recognition. Cambridge University Press, Cambridge, pp 259-288 Heth G, Todrank J (2001) Self referencing in hamsters: comment from Heth and Todrank. Trends Ecol Evol 16:74 Holmes WG (2004) The early history of Hamiltonian-based research on kin recognition. Ann Zool Fennici 41:691-711 Holmes WG, Sherman PW (1982) The ontogeny of kin recognition in two species of ground squirrels. Am Zool 22:491-517 Holmes WG, Sherman PW (1983) Kin recognition in animals. Am Sci 71:46-55 Hughes WOH, Oldroyd BP, Beekman M, Ratnieks FLW (2008) Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science 320:1213-1216 Jansen VAA, van Baalen M (2006) Altruism through beard chromodynamics. Nature 440:663-666 Johnstone RA (1997) Recognition and the evolution of distinctive signatures: when does it pay to reveal identity? Proc R Soc Lond B 264:1547-1553 Keller L, Ross KG (1998) Selfish genes: a green beard in the red fire ant. Nature 394:573-575 Kempenaers B, Sheldon BC (1996) Why do male birds not discriminate between their own and extra-pair offspring? Anim Behav 51:1165-1173 Kuester J, Paul A, Arnemann J (1994) Kinship, familiarity and mating avoidance in Barbary macaques, Macaca sylvanus. Anim Behav 48:1183-1194

Kin recognition: conceptual issues, mechanisms and evolutionary theory

83

Lacy RC, Sherman PW (1983) Kin recognition by phenotype matching. Am Nat 121:489-512 Leinders-Zufall T, Brennan P, Widmayer P, Chandramani P, Maul-Pavicic A, Jager M, Li X-H, Breer H, Zufall F, Boehm T (2004) MHC class I peptides as chemosensory signals in the vomeronasal organ. Science 306:1033-1037 Lieberman D, Tooby J, Cosmides L (2007) The architecture of human kin detection. Nature 445:727-731 Liebert AE, Starks PT (2004) The action component of recognition systems: a focus on the response. Ann Zool Fennici 41:747-764 Manica A (2002) Filial cannibalism in teleost fish. Biol Rev 77:261-277 Manning CJ, Wakeland EK, Potts WK (1992) Communal nesting patterns in mice implicate MHC genes in kin recognition. Nature 360:581-583 Mateo JM (2004) Recognition systems and biological organization: the perception component of social recognition. Ann Zool Fennici 41:729-745 Mateo JM, Johnston RE (2000) Kin recognition and the ‘armpit effect’: evidence of self-referent phenotype matching. Proc R Soc Lond B 267:695-700 Mateo JM, Johnston RE (2001) Self referencing in hamsters: reply from Hauber and Sherman. Trends Ecol Evol 16:74-75 Mateo JM, Johnston RE (2003) Kin recognition by self-referent phenotype matching: weighing the evidence. Anim Cogn 6:73-76 McRae SB (1996) Family values: costs and benefits of communal nesting in the moorhen. Anim Behav 52:225-245 Milinski M (2006) The major histocompatibility complex, sexual selection, and mate choice. Annu Rev Ecol Evol Syst 37:159-186 Moore T, Moore HD (2002) Marsupial sperm pairing: a case of ‘sticky’ green beards? Trends Ecol Evol 17:112-113 Müller JK, Eggert A-K (1990) Time-dependent shifts between infanticidal and parental behavior in female burying beetles: a mechanism of indirect motheroffspring recognition. Behav Ecol Sociobiol 27:11-16 Neff BD, Sherman PW (2002) Decision making and recognition mechanisms. Proc R Soc Lond B 269:1435-1441 Novotny MV, Soini HA, Koyama S, Wiesler D, Bruce KE, Penn DJ (2007) Chemical identification of MHC-influenced volatile compounds in mouse urine. I: quantitative proportions of major chemosignals. J Chem Ecol 33:417434 Pagel M (1997) Desperately concealing father: a theory of parent-infant resemblance. Anim Behav 53:973-981 Penn DJ (2002) The scent of genetic compatibility: sexual selection and the major histocompatibility complex. Ethology 108:1-21 Penn DJ, Potts WK (1998) MHC-disassortative mating preferences reversed by cross-fostering. Proc R Soc Lond B 265:1299-1306 Penn DJ, Potts WK (1999) The evolution of mating preferences and major histocompatibility complex genes. Am Nat 153:145-164 Penn DJ, Damjanovich K, Potts WK (2002) MHC heterozygosity confers a selective advantage against multiple-strain infections. Proc Natl Acad Sci USA 99:11260-11264

84

Dustin J. Penn and Joachim G. Frommen

Petrie M, Krupa A, Burke T (1999) Peacocks lek with relatives even in the absence of social and environmental cues. Nature 401:155-157 Pfennig DW, Reeve HK, Sherman PW (1993) Kin recognition and cannibalism in spadefoot toad tadpoles. Anim Behav 46:87-94 Porter RH (1988) The ontogeny of sibling recognition in rodents: superfamily Muroidea. Behav Gen 18:483-494 Porter RH (1991) Mutual mother-infant recognition in humans. In: Hepper PG (ed) Kin Recognition. Cambridge University Press, Cambridge, pp 413-432 Porter RH, Tepper VJ, White DM (1981) Experimental influence on the development of huddling preferences and ‘sibling’ recognition in spiny mice. Dev Psychobiol 14:375-382 Queller DC, Ponte E, Bozzaro S, Strassmann JE (2003) Single-gene greenbeard effects in the social amoeba Dictyostelium discoideum. Science 299:105-106 Reeve HK (1989) The evolution of conspecific acceptance thresholds. Am Nat 133:407-435 Reeve HK, Sherman PW (2001) Optimality and phylogeny: a critique of current thought. In: Orzack SH, Sober E (eds) Adaptationism and Optimality. Oxford University Press, Oxford, pp. 64-113 Ridley M, Grafen A (1981) Are green beard genes outlaws? Anim Behav 29:954955 Rousset F, Roze D (2007) Constraints on the origin and maintenance of genetic kin recognition. Evolution 61:2320-2330 Schausberger P (2007) Kin recognition by juvenile predatory mites: prior association or phenotype matching? Behav Ecol Sociobiol 62:119-125 Sherman PW, Holmes WG (1985) Kin recognition: issues and evidence. In: Hölldobler B, Lindauer M (eds) Experimental Behavioral Ecology and Sociobiology. Gustav Fischer-Verlag, Stuttgart, pp 437-460 Sherman PW, Reeve HK, Pfennig DW (1997) Recognition systems. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 69-98 Sinervo B, Chaine A, Clobert J, Calsbeek R, Hazard L, Lancaster L, McAdam AG, Alonzo S, Corrigan G, Hochberg ME (2006) Self-recognition, color signals, and cycles of greenbeard mutualism and altruism. Proc Natl Acad Sci USA 103:7372-7377 Smukalla S, Caldara M, Pochet N, Beauvais A, Guadagnini S, Yan C, Vinces MD, Jansen A, Prevost MC, Latge J-P, Fink GR, Foster KR, Verstrepen KJ (2008) FLO1 is a variable green beard gene that drives biofilm-like cooperation in budding yeast. Cell 135:726-737 Stuart RJ (1991) Kin recognition as a functional concept. Anim Behav 41:10931094 Summers K, Crespi B (2005) Cadherins in maternal-foetal interactions: red queen with a green beard? Proc R Soc Lond B 272:643-649 Tang-Martinez Z (2001) The mechanisms of kin discrimination and the evolution of kin recognition in vertebrates: a critical re-evaluation. Behav Proc 53:21-40 Thünken T, Bakker TCM, Baldauf SA, Kullmann H (2007a) Active inbreeding in a cichlid fish and its adaptive significance. Curr Biol 17:225-229

Kin recognition: conceptual issues, mechanisms and evolutionary theory

85

Thünken T, Bakker TCM, Baldauf SA, Kullmann H (2007b) Direct familiarity does not alter mating preference for sisters in male Pelvicachromis taeniatus (Cichlidae). Ethology 113:1107-1112 Todrank J, Heth G (2003) Odor-genes covariance and genetic relatedness assessments: rethinking odor-based ‘recognition’ mechanisms in rodents. Adv Stud Behav 32:77-130 Tsutsui ND (2004) Scents of self: the expression component of self/nonself recognition systems. Ann Zool Fennici 41:713-727 van der Jeugd HP, van der Veen IT, Larsson K (2002) Kin clustering in barnacle geese: familiarity or phenotype matching? Behav Ecol 13:786-790 Vander Meer RK, Alonso LE (2002) Queen primer pheromone affects conspecific fire ant (Solenopsis invicta) aggression. Behav Ecol Sociobiol 51:122-130 Villinger J, Waldman B (2008) Self-referent MHC type matching in frog tadpoles. Proc R Soc Lond B 275:1225-1230 Waldman B (1987) Mechanisms of kin recognition. J Theor Biol 128:159-185 Waldman B (1988) The ecology of kin recognition. Annu Rev Ecol Syst 19:543571 Waldman B, Frumhoff PC, Sherman PW (1988) Problems of kin recognition. Trends Ecol Evol 3:8-13 West SA, Pen I, Griffin AS (2002) Cooperation and competition between relatives. Science 296:72-75 Westermarck EA (1891) The History of Human Marriage. Macmillian, London Willse A, Belcher AM, Preti G, Wahl JH, Thresher M, Yang P, Yamazaki K, Beauchamp GK (2005) Identification of major histocompatibility complexregulated body odorants by statistical analysis of a comparative gas chromatography/mass spectrometry experiment. Anal Chem 77:2348-2361 Willse A, Kwak J, Yamazaki K, Preti G, Wahl JH, Beauchamp GK (2006) Individual odortypes: interaction of MHC and background genes. Immunogenetics 58:967-982 Wilson EO (1975) Sociobiology: The New Synthesis. Havard University Press, Cambridge/MA Wyatt TD (2003) Pheromones and Animal Behaviour: Communication by Smell and Taste. Cambridge University Press, Cambridge Yamazaki K, Beauchamp GK, Kupniewski D, Bard J, Thomas L, Boyse EA (1988) Familial imprinting determines H-2 selective mating preferences. Science 240:1331-1332 Yamazaki K, Beauchamp GK, Curran M, Bard J, Boyse EA (2000) Parentprogeny recognition as a function of MHC odortype identity. Proc Natl Acad Sci USA 97:10500-10502

Chapter 4

Honeybee cognition MARIO PAHL, JÜRGEN TAUTZ AND SHAOWU ZHANG

ABSTRACT Behavioural research on honeybees has shown that bees are not the simple, hardwired creatures they were once believed to be. Bees display perceptual and ‘cognitive’ abilities that are rich, complex and flexible. In this chapter, we begin a review of these abilities with a brief introduction of the bee’s sensory equipment. Next, we describe several experimental approaches to bee behaviour. As this review is not intended to be exhaustive, we focus on behavioural experiments on free-flying honeybees. The studies described here investigate complex forms of learning and navigation, and mark important steps in understanding the processes underlying the bee’s remarkable behaviours.

4.1 Introduction Animals are constantly exposed to a stream of sensory information that must be successfully harnessed to improve their chances of survival and reproduction. The way in which animals go about doing this, i.e. the mechanisms by which animals acquire, process, store, and act on information from the environment (Shettleworth 1998; see also Giurfa 2003), can be termed cognition. This chapter focuses on the latter part of this sequence of events, namely the manner in which animals respond to information in the environment through observable behaviours, such as foraging, navigation, mate recognition and other forms of decision making. We begin by outlining the reasons why the honeybee (Apis mellifera) is eminently suited for such investigations into animal cognition. We then describe the sensory world of the honeybee, as this background information is highly relevant to the types of behaviour that this insect can produce. The bulk of the chapter presents the results of research on the impressive behavioural repertoire of honeybees in relation to the sensory information

88

Mario Pahl, Jürgen Tautz and Shaowu Zhang

available to them, under both laboratory and natural conditions, focusing on more recent behavioural studies that demonstrate that these insects also possess cognitive abilities that were once only ascribed to vertebrates. We also decided to concentrate on studies that have made use of free-flying bees, as they go further towards simulating situations that match the behaviour of bees in nature. Naturally, we recognise the important and fascinating contributions made by studies using tethered bees in experimental paradigms like the proboscis extension reflex (such as Bitterman et al. 1983, Maleszka and Helliwell 2001, Giurfa and Malun 2004), but an indepth review of these studies is beyond the scope of the current chapter.

4.2 Why study honeybees? Studies of animal cognition have traditionally focused on vertebrates like rats, monkeys and birds, i.e. animals that possess relatively large brains and are evolutionarily not too distant from humans. A bee brain, on the other hand, contains less than a million neurons (Witthöft 1967), about five orders of magnitude less than a human brain with an estimated 100 billion neurons (Williams and Herrup 1988). However, it would be a mistake to assume that higher neuron numbers and larger brains are a prerequisite for complex learning, which is defined as any learning other than simple associative or Pavlovian learning and operant conditioning (Giurfa 2003). Many of the experimental paradigms and behavioural assays developed for vertebrates have now been adapted to insects. One example is the delayed match-to-sample (DMTS) paradigm, which was independently developed in 1959 by Blough and Konorski for pigeons, and later used to study memory in honeybees (see Sect. 4.4.1; Blough 1959, Konorski 1959). Honeybees are social insects with a sophisticated system of division of labour, in which each bee flexibly carries out different tasks, depending on its age and the colony’s needs. Arranged in the order of occurrence, these are: cell cleaning, capping the brood, caring for the brood, serving in the queen’s court, receiving nectar, production of honey, removal of detritus, pollen packing, comb building, ventilation, entrance guarding and foraging. They fly fast and with precision, navigate large distances to food sources, communicate those food locations and potential new nest sites using a symbolic dance language, and efficiently manipulate different flowers to extract nectar and pollen (von Frisch 1967, Seeley 1995). This lifestyle makes honeybees highly suitable organisms for studying the principles of learning, memory and navigation. When trained to fly into a Y-maze to memorise one of two opposing stimuli, a bee can learn a new

Honeybee cognition

89

odour during just a few rewarded visits to a food source, a new colour in about 5 visits, a new pattern in about half a day (after 20-30 visits), and a new route to a food source in only 3-4 visits (Zhang 2006). The underlying navigational skills that make this behaviour possible require efficient information processing and storage mechanisms which, in turn, allow bees to display perceptual and ‘cognitive’ abilities that are surprisingly rich, complex and flexible (Menzel and Mueller 1996, Srinivasan et al. 1998, Collett and Collett 2002, Giurfa 2003, Zhang and Srinivasan 2004a,b).

4.3 The honeybee’s sensory world As the acquisition of useful information from the outside world is the first step in our cognitive sequence of events defined above, it is pertinent to consider the two main sensory systems of the honeybee – the eyes and the antennae. Most of the honeybee brain is devoted to processing the sensory input from these structures (Brandt et al. 2005), and it follows that vision and olfaction play critically important roles in honeybee cognition. These two systems are briefly described below. 4.3.1 The compound eye The honeybee’s compound eyes cover a large part of the head, and consist of approximately 6,000 ommatidia each (Tautz 2008). Compound eyes have, compared to human lens eyes, a low visual acuity, because each ommatidium contributes 1 pixel to the picture the bee perceives (in contrast to a single complete picture on the retina formed by a lens). Thus, a bee has to inspect a flower closely to be able to see its exact shape. The bee’s temporal resolution, however, is extremely high, and well suited for fast flight maneuvers (Srinivasan et al. 1999). Bees have trichromatic vision provided by UV, blue and green receptors (Menzel and Blakers 1976), which shift their visual spectrum into the shorter wavelengths, compared to human colour vision. This means that honeybees cannot see red (which appears black to them), but they can see UV light. The short wavelength UV receptor is also involved in perceiving linearly polarised light (Rossel and Wehner 1986). Because many blossoms carry UV-reflecting patterns on the petals, this information can be used to discriminate flowers (Chittka et al. 1994). The green receptor is, apart from colour vision, used for edgeand movement detection (Lehrer et al. 1990), including distance estimation by optic flow (Chittka and Tautz 2003).

90

Mario Pahl, Jürgen Tautz and Shaowu Zhang

4.3.2 The antennae Olfaction is important for navigation, foraging and social behaviours in the hive. Floral scents guide honeybees to profitable flowers to forage on, and they can amplify the spatial memories of those locations. Once a scent is associated with a food location, re-encountering this scent inside the hive (for example from a returning forager who has been foraging on the same flower species) is enough to trigger the navigational memories, and make the bee revisit the food site (Johnson and Wenner 1966, Reinhard et al. 2004a). Inside the hive, the queen pheromone prevents worker bees from laying eggs and rearing other queens, as long as the active queen is young and healthy. Nurse bees can distinguish between queen and worker eggs based on their scent, and destroy eggs from other worker bees (see Ratnieks and Wenseleers 2008, Tautz 2008, and Heinze this volume for details on worker policing). Since each hive has its own special bouquet of hydrocarbons on the bees’ cuticulae, they can easily distinguish between their nestmates and foreign bees. The 60,000 olfactory receptor cells on the antennae are able to discriminate a nearly unlimited number of odours (Vareschi 1971). Interestingly, the sensillae are concentrated on the right antenna, and bees respond to odours better when they are trained on their right antenna. This kind of functional lateralisation is a phenomenon well known in humans and other vertebrates, and has recently been shown in the honeybee as well (Letzkus et al. 2006).

4.4 Experimental approaches The visual and olfactory capabilities mentioned above make the honeybee a flexible model organism for the study of cognition and behaviour, one that can be subjected to a variety of experimental techniques, both in the laboratory and in the field. Some of the more widely used techniques are described below. 4.4.1 Learning in maze experiments Bees can be easily trained to forage at artificial sucrose feeders, even several kilometres from the hive (von Frisch 1967). Each visiting bee can be individually marked, and its behaviour can then be observed at the feeder and inside the hive (for example in Collett and Baron 1994). Experiments such as this are very useful in identifying and separating out the many fac-

Honeybee cognition a

b

Matching stimuli

A’

91

B’

Baffle

Sample stimulus A d

% Correct decisions

100

A

B

A’

B’

75

50 1

2

3

4

Training block

5

6

Probability of response

c

1.0 P 0.5

0.0

U

P

6

7

U 1

2

3

4

5

Trial

8

Fig. 4.1 a. Delayed match-to-sample (DMTS) setup in a Y-maze. In order to get a reward, the bee has to look at the sample stimulus A, store it in working memory, and recall it when deciding for one of the two matching stimuli A’ or B’. The baffle prevents the bee from seeing both sample and matching stimulus at the same time. A’ leads to a reward in this example. b. Proboscis extension response (PER) in harnessed bees. Touching the antennae with sucrose solution (US) elicits proboscis extension in naive bees, while an odour does not. Bees are conditioned to associate an odour (CS) with a sucrose reward (US). The odour alone leads to proboscis extension in the trained bee. c. Learning curve of a DMTS experiment. * denotes a choice level significantly different from random choice. Adapted from Gross et al. (2009). d. Learning curve of a PER experiment. Probability of response to the CS in a group of subjects given paired training (P) and in a group given unpaired training (U) followed by paired training (U → P). Adapted with permission from Bitterman et al. (1983). See text for details. Pictures in b courtesy of Helga R. Heilmann, BEEgroup.

tors that honeybees have to take into account when, for instance, embarking on a foraging trip to a food source. Artificial feeders can also be placed behind a stimulus at the end of a maze (Fig. 4.1a), in which the bee has to make one or more decisions for a visual stimulus or odour to gain access to

92

Mario Pahl, Jürgen Tautz and Shaowu Zhang

the feeder. The maze can be simply Y-shaped, when a single discrimination task is to be performed (for example, Srinivasan and Lehrer 1988), or more complex, when the bee is trained to navigate an extensive maze according to symbolic cues or rules (see Sect. 4.5.2). A widely used setup to investigate the principles of learning and working memory is the delayed match-to-sample (DMTS) paradigm. Established in more traditional animals models in cognitive studies, such as monkeys (Damato et al. 1985), dolphins (Herman and Gordon 1974) and pigeons (Roberts 1972), it proved extremely useful in honeybee research as well. In most DMTS experiments, the animal is presented with a sample stimulus A, followed by a delay, and is then presented with two matching stimuli, A’ and B’. In order to receive a reward, the animal has to memorise the sample stimulus, and after a short delay (e.g. 5 sec), choose the matching stimulus identical to the sample; A’ in this case. If B is the sample stimulus, the animal has to choose matching stimulus B’ to receive a reward (Fig. 4.1a, c). The animal then has to show the ability to transfer the learned matching ability to novel stimuli, which it has not encountered during the training phase. Both long-term and working memory are required to successfully complete this task – long-term memory for the matching rule, and working memory for the sample pattern to be matched on each visit to the maze. 4.4.2 Navigation in tagged free-flying honeybees While it is relatively easy to observe marked honeybees at an artificial feeder or at the hive, it is nearly impossible to study the behaviour of honeybees in mid-flight with the naked eye. Human observation can also only be carried out for a few hours at a time, which makes it difficult to gauge the behaviour of large numbers of foragers over a long study period, such as days or weeks. It is precisely to overcome such difficulties that some researchers have turned to miniature signaling devices that can be attached to the thorax of individual bees, thereby allowing their behaviour to be monitored automatically. One such technique involves the use of harmonic radar, with which the exact trajectories of individuals can be monitored over short periods of time. This technique has proven valuable in studies of honeybee navigation, particularly in the investigation of homing mechanisms (Menzel et al. 2005, Reynolds et al. 2007). More recently, radio frequency ID (or RFID) tags have also been used to study foraging behaviour in honeybees (Streit et al. 2003, Fuchs et al. 2006). Each tag is coded with an individual ID which is logged by a receiver every time a tagged honeybee passes near it. This technology seems

Honeybee cognition

93

ideally suited for investigating various problems, such as the effect of the weather or the visual landscape on honeybee homing and foraging abilities. 4.4.3 Classical conditioning in harnessed bees Classical conditioning is a form of associative learning that was first demonstrated in dogs by Ivan Pavlov. In this experimental paradigm, the bee is harnessed in a tube so that the antennae and mouth parts can move freely (Fig. 4.1b). When the antennae are touched with a sucrose solution (unconditioned stimulus, US), the bee extends its proboscis to drink the sucrose. Simple, non-associative forms of learning, such as habituation and sensitisation, can be studied in this way. Odourants do not elicit this reaction in untrained animals. However, when an odour (conditioned stimulus, CS) is presented just a few seconds before the sucrose, the bee forms an association between the US and the CS (forward pairing), and the odourant (CS) alone can elicit the PER in the next test (Bitterman et al. 1983, see Fig. 4.1d for a learning curve). So far, this is purely associative learning, and no higher processing is required. However, honeybees can not only be trained to perform elementary distinctions involving a rewarded (A+) and an unrewarded odour (B–, where + and – represent rewarded and non-rewarded stimuli, respectively). They can solve configural learning tasks as well, where pure elementary associations are insufficient to succeed. In a biconditional discrimination experiment, involving four odours A, B, C and D, bees could be trained to react to only two binary combinations of four odours (AB+, CD+) but not to the other combinations (AC–, BD–, see Hellstern et al. 1995). Although the reward frequency during the training for each stimulus was equal, the bees formed configural associations, enabling them to react appropriately to the odour combinations. Honeybees can solve positive patterning tasks as well, where the animals learn to react to the combination of two odours (AB+), but not to the single odours (A–, B–). The same is true for negative patterning (AB–, A+, B+, see Deisig et al. 2001). The PER setup is an excellent method to bring advanced learning and memory processes into a laboratory situation. The harnessed bee setup also enables scientists to apply electrophysiological and optophysiological techniques in vivo (Giurfa 2007), while the bee is behaving in a PER experiment, to trace memory and learning to brain areas and individual neurons. However, as stated earlier, the remainder of this chapter will be devoted to studies involving free-flying bees.

94

Mario Pahl, Jürgen Tautz and Shaowu Zhang

4.5 Cognitive mechanisms underlying behaviour All animals must make decisions throughout their lives when interacting with their surroundings, and the honeybee is no exception. Decisionmaking is necessary inside the hive, when, for instance, following a waggle dance, and outside the hive, when locating a food source, actively foraging at the food source, or returning to the hive. We will therefore discuss in this section studies on three important mechanisms that inform decisionmaking in honeybees, as evidenced by behaviour observed in free-flying animals. 4.5.1 Categorisation When honeybees are foraging, they need to remember routes to and from different food sources. There can be little doubt that they use some kind of neural ‘snapshot’ to remember and recognise landscapes and landmarks on these routes (Collett and Cartwright 1983, Judd and Collett 1998). However, it is hard to believe that the bee’s brain is capable of storing so many different images of the environment in its restricted memory. Categorisation is an information processing strategy in which objects or events are grouped together into categories, so that a similar response can be made to all members of the category (Keller and Schoenfeld 1950, Troje et al. 1999). Thus, an organism is not restricted to respond only to stimuli it has already encountered, but it can develop a set of appropriate reactions to all stimuli that match certain criteria (Wasserman 1993). Since categorisation is the basis for any identification and classification task, it has enormous biological relevance, particularly for a foraging honeybee. It would be especially interesting to know the extent to which invertebrates are able to employ this ability in spite of their miniature central nervous systems (see Prete 2004 for some examples). 4.5.1.1 Learning pattern properties The lifestyle of honeybees requires them to remember a number of different patterns, such as the shape of the hive, shapes of nectar-bearing flowers, and shapes of prominent landmarks. Thus, one would expect that honeybees have evolved a way to compress image information and to extract the general identifying features of a pattern, such as its shape, orientation, symmetry, colour and size. Indeed, honeybees have been shown to be able to extract orientation (van Hateren et al. 1990), radial and circular symmetry (Horridge and Zhang 1995) as well as bilateral symmetry, including the

Honeybee cognition

95

orientation axis (Giurfa et al. 1996, Horridge 1996), of visual patterns, and to learn these features to solve decision-making tasks. Other characteristics of objects, such as colour and size, can be extracted and memorised as well, without having to store the entire image (Horridge et al. 1992, Ronacher 1992). Recently, using differential conditioning in free-flying bees, Dyer et al. (2005) have shown that honeybees are even capable to discriminate between, and recognise images of, human faces. 4.5.1.2 Categorisation of natural objects In 2004, Zhang and colleagues investigated whether bees can categorise similar natural images together. They trained bees in a multiple choice maze (Fig. 4.2a, b) to distinguish between four different types of natural scenes, and to group them into four distinct categories: Category F consisted of images of flowers that were star-shaped and of different colours; Category f comprised images of flowers that were nearly circular in shape and of different colours; Category P consisted of images of plant stems of various shapes; and Category L was composed of images of landscapes (Fig. 4.2c). Within each category, individual images differed in details of shape, texture, and sometimes colour. Bees were trained in a multiple choice maze in a DMTS task, in which they were shown a picture as a sample stimulus in the entrance chamber (C1). To continue through the maze, they had to fly through a small hole in the middle of the sample picture to enter chamber two (C2). The back wall of C2 was transparent, with a 3-cm hole in its centre. The small aperture restricted the bees’ flight speed, and the transparent wall allowed them to see four additional pictures (or comparison stimuli) on the rear wall of chamber three (C3, see Fig. 4.2a, b). During training, the sample stimuli and the four comparison stimuli were all from Group 1. Groups 1-4 each consisted of one unique example of each of the four categories; all of them are shown in Fig. 4.2c. If the bee chose the correct test stimulus in C3, she received a reward of sugar solution from a feeder that was placed in the reward box behind that stimulus, by landing on and crawling through a tube in the centre of the stimulus (Fig. 4.2). In transfer tests, Zhang et al. investigated whether the trained bees could match a sample stimulus from one group to a stimulus of the same category from a different group. They had to, for example, match the yellow star-shaped flower F1 in group 1 to the blue star-shaped flower F3 in group 3 (Fig. 4.2c and 4.3b). In these tests, the sample was always a stimulus from Group 1, but the comparison stimuli were from Group 2, 3 or 4 (Fig. 4.2c). The bees performed very well in these transfer tests. In fact, in each case, the bees showed a clear and significant preference for the test stimu-

96

Mario Pahl, Jürgen Tautz and Shaowu Zhang a) Lateral view

b) Frontal view

R1,2 C1

C2

C3 R3,4

1

2

3

4

20 cm

c) Stimuli Group 1 F1

f1

P1

L1

F2

f2

P2

L2

F3

f3

P3

L3

F4

f4

P4

L4

Group 2

Group 3

Group 4

Fig. 4.2 The multiple choice maze in the categorisation experiment. a shows the multiple choice maze in lateral view. b shows a frontal view of the choice area and the arrangement of stimuli. c illustrates all the training (group 1) and testing (groups 2-4) stimuli. Adapted with permission from Zhang et al. (2004).

lus that belonged to the same category as the sample (Fig. 4.3a, b). Particularly noteworthy is the transfer test using Group 4, in which the comparison stimuli were entirely novel (Fig. 4.3c, Group 4). These stimuli had never been used in the training phase, or in the learning tests or transfer tests. Again, the bees performed very well at chosing the test stimulus that was in the same category as the sample (Fig. 4.3c).

Honeybee cognition

***

n =10 bees

*** ***

***

***

***

n =10 bees

* ***

n =10 bees

***

***

n =10 bees

***

***

n =10 bees

n =10 bees

*** *

*

*

**

n =10 bees

***

* ***

*

***

n =10 bees

***

***

n =10 bees

***

***

n =10 bees

***

* ***

***

***

97

***

* ***

n =10 bees

***

*** ** ***

n =10 bees

***

***

Fig. 4.3 Results of the categorisation experiment. These transfer tests examined the ability of bees to categorise sample stimuli from group 1, to test stimuli of the same category in (A) group 2, (B) group 3 and (C) the novel group 4. For each group, the bars show the relative preferences for the four test stimuli when the sample was F1, f1, P1 or L1, as shown underneath the abscissa. In each panel, N denotes the number of bees that were tested in each experiment. Asterisks denote statistically significant differences from the random choice level of 25% (broken horizontal lines). ***P < 0.001; **P < 0.01; *P < 0.05. Black asterisks refer to levels significantly greater than 25%, and red asterisks to levels significantly lower than 25%. O denotes P > 0.05. Values are means ± S.E.M. In each case, the bees chose the test stimulus that belonged to the same category as the sample. Adapted with permission from Zhang et al. (2004).

98

Mario Pahl, Jürgen Tautz and Shaowu Zhang

The results of the transfer tests with novel stimuli (Fig. 4.3c) show that the bees performed very well at chosing the novel test stimulus that was in the same category as the sample. The honeybees exhibited the same response to novel stimuli that differed greatly in their individual, low-level features. That is, bees treat these highly variable stimuli as equivalent. These findings suggest that the honeybee possesses an ability to group similar visual stimuli into categories, similar to that of vertebrates (Zhang et al. 2004). 4.5.2 Rule learning The ability to learn general rules for dealing with often-encountered situations is adaptive for an animal, as it would remove the need to repeatedly assess every situation and try out new strategies each time. This is particularly true for situations where the appropriate response can be determined by attending to real or abstract cues, and is followed up by an application of a learned set of behaviours. Honeybees are good candidates for the study of the acquisition of abstract rules; after all, the waggle dance is the most complex abstract communication systems in the animal kingdom outside human language. Several studies have shown that honeybees can not only learn visual and olfactory stimuli as signals that indicate a particular action to be performed, but that they can also be trained to learn other abstract rules that are specific to particular experimental situations. 4.5.2.1 Learning principles of symbolic matching In 2001, Giurfa and colleagues showed that honeybees are capable of learning the concept of ‘sameness’ in a DMTS task, and the concept of ‘difference’ in a delayed non-matching task (DNMTS). The bees were trained in a Y-maze as shown in Fig. 4.1a. In one experiment, a group of bees was trained to match gratings: when the sample was a horizontal grating, the bees had to choose the horizontal comparison stimulus to get a reward; if the sample was a vertical grating, it had to choose the vertical stimulus. The position of the comparison stimuli was interchanged regularly to prevent the bees from simply learning to fly to the left or right. The bees learned this task well. The trained group was then tested on their ability to transfer their learned rules to colours. When a yellow sample was presented, they had to choose the yellow comparison stimulus, and when blue was presented as a sample, they had to choose blue. The bees were able to immediately solve the new task, although they had never encountered colours during their training.

Honeybee cognition

99

Honeybees can also learn to match odours in the same setup. Furthermore, bees trained to match odours can immediately transfer the learned matching ability to colours, showing that the concept of ‘matching’ can be transferred even across sensory modalities. Finally, bees can learn the concept of ‘difference’ as well, when they are trained to choose the nonmatching stimulus rather than the matching one. These tasks involve working memory, when the bee has to remember the sample stimulus during the short delay, and long-term memory, where the general rule ‘choose the matching pattern similar to the sample at the entrance’ is stored. A more complex variant of this paradigm is the symbolic delayed match-to-sample (SDMTS) task. Here, the sample and matching stimuli are never identical; the correct choice is either arbitrarily designed by the experimenter (i.e. if the sample is blue, choose the horizontal grating as matching stimulus, and if the sample is yellow, choose the vertical grating, see Zhang et al. 1999), or share a symbolic feature, e.g. number of items presented, and the animal has to extract the common number to get a reward (see Gross et al. 2009 and Sect. 4.5.2). Why would honeybees have evolved the ability to solve such tasks? Considering their foraging behaviour, we find that honeybees are often required to match memorised images to actual scenes. Collett and Wehner suggested that, when travelling repeatedly between a food source and the hive, foraging insects use a series of visual snapshots acquired en route to find their way. By comparing the stored image to the actual scene, the insect is able to ascertain whether it is on the right path to its goal (Collett and Kelber 1988, Wehner et al. 1990, 1996, Judd and Collett 1998). When bees forage at a meadow with many different flowers, they visit several blossoms of the same plant species before they reach their nectar load of ~50 mg and head home. Flower constancy is beneficial not only for the plants, ensuring pollination by pollen of the same species, but also for the bees, who become more efficient when learning to handle the same type of blossom repeatedly (Chittka et al. 1999). It also allows a bee to learn that a particular plant species produces nectar at a particular time of day. Thus, matching their memorised representation of a profitable species several times on a foraging trip improves foraging efficiency (see also Sect. 4.5.1). 4.5.2.2 Visual working memory in decision making In 2005, Zhang and colleagues set out to investigate the role of working memory in decision making in honeybees. They used a modified version of a Y-maze (shown in Fig. 4.4) with a long tunnel attached in front of the decision chamber, so that the distance between the sample and the matching stimuli (d1 in Fig. 4.4) could be varied between 25 and 575 cm. A

100

Mario Pahl, Jürgen Tautz and Shaowu Zhang

0.9

d2

0.8

Entrance

d1 A’

Choice frequency

B’

Sample A/B

R

0.7 A

B

A’

B’

0.6 0.5 0.4

y = 0.48e 0

2.5

-0.53t

+ 0.50

2

r = 0.999

5.0 7.5 Delay [sec]

10.0

12.5

Fig. 4.4 Maze setup, visual stimuli and results of the working memory experiment. The bee flies through the entrance into the tunnel (d2), and encounters sample A or B. Then it travels distance d1 between the sample and the decision chamber (1st circle). Here, it makes a decision between stimuli A’ and B’ to get a reward (R). B’ is rewarded in this example. The choice frequency drops to random level (dashed line) at a delay of 8.68 sec (d1 = 475 cm). Adapted with permission from Zhang et al. (2005).

longer distance between the sample and the decision chamber means a longer flight time for the bee, and thus the delay time could be varied between 1.24 ± 0.1 sec at 25 cm and 11.41 ± 1.69 sec at 575 cm. They trained the bees with a set of blue gratings oriented 45º (A) and 135º (B) from the horizontal (Fig. 4.4). During training, the distance between the sample pattern and the choice chamber was 25 cm. When the bees’ choice frequency for the correct pattern reached a plateau at 75 ± 3%, tests with greater distances and thus longer delay times were carried out. The results show that a visual pattern can be kept in working memory for about 8 sec; longer delays lead to random choices in the DMTS task (dashed line in Fig. 4.4). The information stored in working memory decays exponentially as a function of time (see Fig. 4.4 for equation). There is a surprising symmetry to Chittka’s findings on the memory dynamics for flower choice in bumblebees. He and his colleagues found that the flower constancy was excellent at intervisit intervals of up to 2 sec, dropped to half-maximum accuracy at 3-4 sec, and had decayed completely at 9 sec (Chittka et al. 1997, 1999). Zhang et al. also found that, when the DMTS task involves two samples (one relevant, the other irrelevant), bees can be trained to learn to use the relevant sample to perform the task if the relevant sample is always at a fixed position, or the relevant sample always has the same place in the sequence of presentation (always first or always second). Bees that have learned to use the relevant sample and to ignore the irrelevant sample can generalise this learning, and apply it to novel sets of sample and compari-

Honeybee cognition

101

son stimuli that they have not previously encountered. The findings point to a remarkably robust, and yet plastic working memory in the honeybee (Zhang et al. 2005). 4.5.2.3 Number-based visual generalisation The numerical abilities of many vertebrate species such as raccoons (Davis 1984), dolphins (Kilian et al. 2003), monkeys (Brannon and Terrace 2000), songbirds (Hunt et al. 2008) and even salamanders (Uller et al. 2003), have been investigated in the scientific literature. However, there are few convincing accounts of invertebrate numerical competence. Honeybees, by virtue of their other impressive cognitive feats, are a prime candidate for investigations of this topic. In a recent experiment, Gross et al. (2009) used the DMTS paradigm to test the limits of the honeybees’ ability to match two visual patterns solely on the basis of the shared number of elements in the two patterns. A group of about 20 bees was trained in a modified Ymaze apparatus to perform a basic DMTS task (Fig. 4.1a). After the bees had learned the DMTS rule with an accuracy of about 70% correct decisions, as shown in the learning curve in Fig. 4.1c, they were tested on novel sets of visual stimuli containing 2 and 3 blue dots in random configurations. The bees solved this task with an average accuracy of 74%, providing a first hint that the bees may be using the element number as a cue to solve the task (Fig. 4.5a). The bees were then tested whether they could transfer the rule ‘match the number of items’ to a totally new set of stimuli. Once again, the bees were able to convincingly match the sample and choice patterns, even when they contained two versus three yellow stars, or lemons (Fig. 4.5b). Then, an additional level of abstraction was introduced by making the elements of the sample and choice patterns different. Now, the bees encountered a sample pattern of three blue dots, for instance, which they had to match to a choice pattern composed of three yellow lemons, again in random configurations. Here, too, the bees performed remarkably well, using the number of items to identify the rewarded pattern (Fig. 4.5c). Next, a test was performed to investigate if the bees could transfer their ability to discriminate between two and three, to arrays of three and four items, the latter being a value they had not previously encountered. The bees could successfully carry out a three-to-three match, when the competing stimulus contained four elements. However, they were not able to consistently do a four-to-four match (Fig. 4.5d). Thus, there seemed to be a limit to their ability to extrapolate to higher numerosities: their performance in discriminating four versus five, five versus six and four versus six was also not above chance in all tests (Fig. 4.5e-g). Control experiments confirmed

102

Mario Pahl, Jürgen Tautz and Shaowu Zhang

a Choice n = 19 n = 17 pattern 1.00 Choice frequency

***

***

b

***

n = 25

***

***

n = 18

c

***

n = 16

***

***

n = 17

d

***

0.75 0.50 0.25

0.00 Sample pattern

Choice frequency

e Choice n = 15 pattern 1.00

n = 15

f

n = 12

n = 12

g

n = 14

n = 13

0.75 0.50 0.25

0.00 Sample pattern

Fig. 4.5 Results of the numerosity discrimination experiment. Shown are the transfer tests with various pattern configurations. The pattern below each pair of bars is the sample and that above each bar is the choice pattern; the y-axis gives the choice frequency. The data present the pooled first choices (from each foraging trip) of individual bees. a The configuration of dots on the sample and choice patterns is randomised. b The blue dot patterns in A are replaced with yellow stars, to see if bees can transfer their matching ability to different, unknown stimuli. c The sample and choice patterns are composed of two different elements. d-f Bees trained to discriminate between two and three are tested on patterns with d three and four elements, e four and five elements, f five and six elements, g four and six elements. n = number of bees per condition. Error bars show standard errors. ***denotes statistically significant difference at p < 0.001, **denotes p < 0.01, *denotes p < 0.05 and O denotes p > 0.05. Adapted with permission from Gross et al. (2009).

Honeybee cognition

103

that the bees were not using cues such as scent, the colour or the exact configuration of the visual elements, the combined area or edge length of the elements, or illusory contours formed by the elements (see Gross et al. 2009 for details). An estimation of relative numerical quantity could be extremely useful to foraging honeybees: combining information about the degree of stomach distension along with the number of flowers visited on a foraging trip could provide bees with an index of the profitability of a food source. Honeybees can recognise images of complex natural scenes (Dyer et al. 2008), and may be able to use them as potential landmarks. The number of landmarks encountered on a foraging trip, or found near the hive, could be useful in navigation (Chittka and Geiger 1995, Skorupski and Chittka 2006, Dacke and Srinivasan 2008). Number generalisation could also help in estimations of the number of blossoms on a branch and/or, the number of bees on a blossom, thereby allowing a forager to decide whether to forage at that location or whether to search for a new one. 4.5.2.4 Negotiating a complex maze The classical model animals in vertebrate cognition research, i.e. rats, mice and pigeons (Pick and Yanai 1983, Dale 1988) have often been tested for the ability to learn navigating complex mazes. Apart from the numerous studies using Y-maze setups, few studies exist that investigate the ability of invertebrates to negotiate complex mazes, where more than one or two correct decisions are required to reach the reward. In 1996, Zhang et al. conducted a study investigating the ability of honeybees to fly through a variety of complex mazes in search of a sugar reward in the presence or absence of a number of visual cues. Each maze consisted of a 4x5 matrix of identical cubic boxes with a hole in the centre of each side wall. The experimenters created different paths through the maze by leaving some holes open and blocking others, so that the bees had to fly through a sequence of 9-14 boxes, including 3-5 decision boxes, to reach the reward (Fig. 4.6a, c). To exclude the possible role of external landmarks, the position and orientation of the maze were changed regularly. The experimenters also swapped the position of all boxes before each test to exclude possible olfactory cues left by the bees in training. a) Negotiating a maze by following a colour cue. In the first series of experiments, the bees’ ability to fly through a maze following a colour cue was tested. The cue in this case was a 4x4-cm green mark placed under the correct exit hole in each decision box. After the initial training, in which a feeder was moved step by step along the correct path in the first three

104

Mario Pahl, Jürgen Tautz and Shaowu Zhang

a

b

Reward 18

19

20

13

14

15

16

9

10

11

12

5

6

7

8

1

2

3

4

80 Test 1

Frequency [%]

17

60 40

0

C1

C2

C3

Category

d

Reward 17

18

19

20

13

14

15

16

9

10

11

12

5

6

7

8

1

2

3

4

20

Entrance

0

Colour mark Correct path Incorrect path

Control

20

Entrance c

Test 2

Test 3

Frequency [%]

80

Test 4

Test 5

C4

Control

60 40

C1

C2

Category

C3

C4

Right turn Left turn Decision boxes on correct path

Fig. 4.6 Learning to negotiate complex mazes using visual cues. a shows the maze setup in series 1. The green colour marks indicate the correct path. b shows the results of series 1. In test 1, the bees are following the trained route. In test 2, the maze is reconfigured and the bees have to follow the green patches on a novel route. Control bees were tested in a maze without colour cues. c shows the maze setup in series 2. Colours indicate the direction of turn. d shows the results of series 2. Test 3 shows the results for the trained route. In tests 4 and 5, the maze was rearranged, and the bees are using the learnt turning rules in a novel route. Adapted with permission from Zhang et al. (1996).

boxes, it was placed in the last box. During this initial training, the bees had the chance to learn that the green patch indicated the correct exit hole. After each bee had visited the feeder in the last box at least once, it was placed in the feeder cage behind the last box.

Honeybee cognition

105

Immediately thereafter, each individual bee’s performance was tested. Performance was scored in 4 categories (C1-C4). C1 included flights in which the bee made no mistake and flew directly to the goal, in the flights in C2, the bee turned back and retraced her path one or more times but stayed on the correct path, C3 contained flights in which the bee made one or more mistakes but still arrived at the goal, and the bees in C4 did not reach the goal at all within 5 minutes. The bees that had learned to follow the colour mark in the first three boxes were immediately able to follow the same cue through the whole maze (Test 1 in Fig. 4.6b). When the maze was rearranged and the path changed, the bees were still able to follow the cue to the goal (Test 2 in Fig. 4.6b). This means that the bees could generalise the sign tracking rule they had learned in the first maze, and used it successfully to fly through a new maze. Their performance in Test 1 and 2 was significantly better than a control group of bees with the same training in an unmarked maze with an unfamiliar route to the goal. b) Negotiating a maze by learning a symbolic cue. Can bees learn to negotiate a maze by using symbolic cues as well? To address this question, Zhang et al. trained bees in the same way as above. In this new maze, a left or right turn in the decision boxes was indicated by a colour cue at the back of the box (Fig. 4.6c). Thus, the correct exit was not directly indicated by a colour mark, but the colours were symbols signaling the direction of turn. Blue indicated a right turn, and green indicated a left turn. The bees learned this task as well, as shown in Fig. 4.6d, Test 3. Again, the bees were able to transfer the learned symbolic rules to new maze configurations, and followed novel paths to the goal (Fig. 4.6d, Tests 4 and 5). The performance in all tests was as high as in the previous experiment, and significantly better than the control group. When Zhang et al. trained bees in a third experiment, in a maze where no cues were given, the bees did not learn the route on their own. They had to be trained stepwise along the entire path, but still performed significantly better than the control group. Thus, bees can be taught to negotiate a maze, if the training is done in a series of simple steps. Learning was fastest when the bees were given a set of rules which they can follow, and they learned symbolic cues just as rapidly as direct cues. The bees also demonstrated the ability to transfer the learned rules to novel but analogous contingencies in the spatial and the chromatic domain (Zhang et al. 1996).

106

Mario Pahl, Jürgen Tautz and Shaowu Zhang

4.5.3 Context-dependent learning Contextual cues are dependent on the environment and the situation. They can facilitate memory retrieval, when the context in which the memory was encoded is replicated. Thus, context cues help to carve up the world into distinct regions, and help animals to cope with possible confusions (Colborn et al. 1999, Fauria et al. 2002, Cheng 2005, Dale et al. 2005). Honeybees are able to flexibly change their preference for a visual pattern according to the context in which a task is carried out (Zhang et al. 2006). Two recent studies have demonstrated the simultaneous use of up to three contextual cues in the honeybee. They investigated whether the task at hand (foraging or homing), the location (different foraging areas), and the time of day (morning or afternoon), could be utilised as contextual cues by bees. 4.5.3.1 Task and time as contextual cues In this study, Zhang et al. investigated the effect of time of day (morning or afternoon), task (foraging and homing), as well as both parameters simultaneously, as contextual cues in modulating bees’ preference for a visual pattern. Bees were trained to visit an artificial feeder in a Y-maze (like Fig. 4.1a but without the sample pattern), where they had to make a foraging decision between visual stimulus A and B to get a reward. After feeding, the bees returned to their hive, which had two entrances. Each carried a visual stimulus, A and B, and only one entrance was open at a time. Here, the bees had to choose the correct visual stimulus to gain access to the hive and unload the sugar water they had foraged in the maze. We carried out three series of experiments to investigate these interactions. In the first series of experiments, the bees were trained to choose A at the feeder and at the hive in the morning. In the afternoon, following a midday break, they had to choose B at the feeder and at the hive. During the breaks (midday and night), the maze was closed, and the visual stimuli were removed from the hive, and both entrances were open. The bees could easily learn this task and preferred A in the morning and B in the afternoon, respectively. The results indicate that bees can reverse their preference following midday breaks, as well as overnight breaks, at the feeder and at the hive. Time acted as a contextual cue in this experiment. For the second series of experiments, a new group of honeybees was trained to choose stimulus A at the feeder, and B at the hive. The bees could learn this task as well, switching their preference in just a few minutes, depending on whether they were going out to forage or were returning to the hive. The task at hand was acting as context cue in this case.

Honeybee cognition Feeder side

107

Hive side

Choice frequency for stimulus A

1.00

A

A

n = 15

0.75

n = 15 n = 15

n = 13 0.50

0.25

0.00

n = 14

n = 14 n = 15

n = 15 Training 5

7

9

11

Training 13

Time of day

15

17

Training 7

9

11

Training 13

15

17

19

Time of day

Fig. 4.7 Task as contextual cue. Results of experimental series 3 at the feeder and the maze. The visual patterns were black/white horizontal vs. vertical gratings at the feeder and at the hive. The horizontal grating at the feeder and the vertical grating at the hive were the rewarded patterns in the morning, whereas the vertical grating at the feeder and the horizontal grating at the hive were rewarded in the afternoon. Choice frequencies for the horizontal grating are given. Results show that the trained bees significantly reverse their pattern preference at the feeder and at the hive entrance following midday breaks, as well as overnight breaks. There is a significantly different pattern preference at the feeder and at the hive entrance within each testing period. The modulation of the average choice frequency, with reference to the horizontal grating, could be approximated very well by a sinusoidal curve with a frequency of 0.52, i.e. a period of 12 h. The phase of the sinusoidal curve at the hive was shifted 180° with reference to the feeder. *** denotes p < 0.001. See text for further details. Adapted with permission from Zhang et al. (2006).

The third series of experiments was the most demanding one for the experimental bees. Results are shown in Fig. 4.7. A new group of foragers was trained to choose stimulus A in the maze and stimulus B at the hive in the morning. The same bees then had to choose stimulus B at the feeder and A at the hive in the afternoon to be rewarded. Stimulus A was a horizontal black and white grating, and B was a vertical grating. They solved this task as well, demonstrating that trained bees can reverse their pattern preference at the feeder and at the hive entrance depending on the time of day, and at the same time reverse their preference between foraging and homing, and choose opposite patterns at the feeder and at the hive. The training thus imposed a learnt pattern preference on the bees’ daily cir-

108

Mario Pahl, Jürgen Tautz and Shaowu Zhang

cadian rhythm (Fig. 4.7). This study demonstrates that forager bees possess a sophisticated memory and are able to remember tasks within a temporal context. Honeybees can thus ‘plan’ their activities in time and space, and use context to determine which action to perform and when (Zhang et al. 2006). 4.5.3.2 Time of day and location as contextual cues In one context cue study, Pahl and colleagues investigated how the colour, shape and location of patterns could be memorised within a time frame. Bees were trained to visit two Y-mazes, one of which presented yellow vertical (rewarded) versus horizontal (non-rewarded) gratings at one site in the morning, while another presented blue horizontal (rewarded) versus vertical (non-rewarded) gratings at another site in the afternoon. The bees learned the correct decisions in the two mazes easily. They also had no problems solving transfer tests in the learning mazes, in which the colour cues of the visual patterns were removed, but the location cue, the orientation of the visual patterns and the temporal cue still existed. Now, in order to remove the location cue, the two training mazes were removed and a single maze constructed in a neutral location between the previous mazes, where the bees had never received any training. Three kinds of transfer tests were performed, each in the morning and in the afternoon, in the novel maze (Fig. 4.8). In test 1, the location cue was removed, but other contextual cues, i.e. the colour and orientation of the visual patterns, as well as the temporal cue, still existed. The bees could immediately solve the task at the new location (Fig. 4.8a). In test 2, the location cue and the orientation cue of the visual patterns were removed, but the colour cue and temporal cue still existed. Here, the bees preferred the yellow grating in the morning and the blue grating in the afternoon (Fig. 4.8b). In test 3, the location cue and the colour cue of the visual patterns were removed, but the orientation cue and the temporal cue still existed. Now, the bees preferred the vertical black grating in the morning and the horizontal black grating in the afternoon. The results of this experiment revealed that honeybees can recall their memory of the rewarded visual patterns by using spatial and temporal information, separately and simultaneously. In the learning tests, the bees reached an average performance (morning and afternoon sessions taken together) of 83% correct choices. Setting this as a baseline, we can compare the difficulty of the transfer tests, and thus determine the relative importance of the different cues for the bees. In the transfer tests in the neutral location, the bees reached their best average performance of 91% in experiment 2, the colour discrimination task (Fig. 4.8b). The performance in this transfer test was even better than

Honeybee cognition

a

Choice frequency for stimulus

1.00

b

n = 18

n=9

morning

afternoon

109

c

n = 19

n = 11

n = 12

n = 13

0.75

0.50

0.25

0.00

morning afternoon

morning afternoon

Fig. 4.8 Time as contextual cue. Results of the transfer tests in the neutral location. a Test 1 with yellow and blue training patterns. The bees preferred the yellow vertical stimulus in the morning, and the blue horizontal stimulus in the afternoon. b Test 2 with yellow and blue patterns in the same orientation. The bees preferred the yellow stimulus in the morning, and the blue stimulus in the afternoon. c Test 3 with black patterns. The bees preferred the vertical grating in the morning, and the horizontal grating in the afternoon. n denotes number of individual bees in each test; bars are means ± S.E.M.. Adapted with permission from Pahl et al. (2007).

that of the learning tests, regardless of the missing location and pattern orientation cues. Thus, colour seems to be the most important visual cue for honeybee choice behaviour. These findings are consistent with previous reports that honeybees learn a new colour after about five visits, whereas they require 20–30 visits to learn a pattern (Zhang and Srinivasan 1994). Using the training patterns at maze C in transfer test 1, the bees performed about the same as in the training mazes. Here, the only missing cue was the maze location. This cue seems to have had almost no effect on the bees’ choice performance in small scale navigation, when other contextual cues were available. Pattern colour, shape and the time of day were enough to allow a baseline level of performance at a new location. When the colour cue and the location cue were both taken out in test 3, the bees’ average performance was reduced to 72% (Fig. 4.8c). These results indicate that the shape cue is more difficult for the bees to use than the colour cue. Pattern orientation or, in nature, the shape of different flowers, is thus more important than location for the bees’ choice behaviour once they

110

Mario Pahl, Jürgen Tautz and Shaowu Zhang

have reached their feeding site. The bees clearly used the former to distinguish between the patterns in all experiments except the colour discrimination task, where pattern shape was unavailable. Applying these findings to the natural situation, the colour and shape of flowers are the most important visual cues used by bees to choose between different flower species. When visiting different feeding sites, or different patches of flowers, they can recall their memory of the most rewarding species in conjunction with the time of day, and thus find the most profitable food source even at a new location. Recently, Prabhu and Cheng showed that honeybees trained on colours for only one day can also use temporal information in decision making. This indicates that bees might have a natural tendency towards inferring a circadian pattern on unclear reward situations (Prabhu and Cheng 2008a). They also found that, when similar training is conducted with odours, bees prefer only the odour that was most recently associated with a reward (Prabhu and Cheng 2008b). Thus, bees have two ways to assess the reward probability when encountering a novel situation, which may produce conflicting interests. However, in a forager bee’s natural foraging environment, colour and scent are seldom experienced separately from each other. More experiments are required to examine how the bees’ circadian preference pattern is modulated when color and scent are coupled. The bees’ ability to integrate elements of circadian time, place and visual stimuli is akin to episodic-like memory (Clayton and Dickinson 1998) and we have therefore named this kind of memory ‘circadian timed episodic-like memory’ (Pahl et al. 2007). 4.5.3.3 Associative learning and recall In humans, a smell or a sound can easily trigger a vivid memory of an associated event in the past, even when it involves a different sensory modality and the episode occurred a long time ago. When the smell of sunscreen reminds us of a day at the beach during last summer, our brain displays a cross-modal associative recall; it links a scent to a visual scene and a location. In 1998, Srinivasan and co-workers could show that honeybees link sights to smells, and thus show that insect brains are capable of crossmodal associative recall as well (Srinivasan et al. 1998, Zhang et al. 1999). When foraging, cross-modal associative recall could facilitate the search for a food source in honeybees. For example, the scent of lavender could initiate a search for blue flowers (Zhang 2006), and canola scent could lead to a preference for yellow blossoms. So far, the data show that honeybees could potentially use cross-modal recall, but do bees actually use it in everyday foraging in a natural envi-

Honeybee cognition

111

ronment? To investigate this question, Reinhard and colleagues designed an experiment to test whether floral scents can induce recall of navigational and visual memories in a natural foraging terrain (Reinhard et al. 2004a,b). They trained individually marked bees to forage at several scented sugar feeders at different outdoor locations, each 50 m from the hive. In the first experiment, feeder one carried rose scent, and feeder two carried lemon scent. After two days of training, the bees were tested by replacing the feeders with empty, unscented jars. Then, for eight minutes each, rose and lemon scents were alternately blown into the hive by a small fan. During each scent blowing interval, the experimenters noted the number of marked bees arriving at each test feeder. The foragers leaving the hive during the rose scent interval showed a significant preference for the feeder that was previously rose-scented and vice versa. Repeating the experiment with different scent combinations, such as rose and almond or lemon and almond, led to similar results. The injected scent triggered navigational memories previously acquired during the training. Foraging for nectar and pollen is crucial for the survival of the hive. Foraging bees cover a huge area of potential foraging sites, and when a single forager finds a particularly rewarding site, it needs to convey the information to its nestmates. In his classic studies, von Frisch (1967) reported the famous honeybee dance language, in which foragers communicate distance and direction of a food source to each other, in order to recruit more foragers to a rewarding site. A recent study has even hinted at the possibility that two different bee species, Apis cerana cerana and Apis mellifera ligustica, can understand each other’s dances (Su et al. 2008). This study demonstrates an additional mechanism that could be used by bees to recruit large numbers of bees to a good food source. The scent of the returning bee, as well as the taste of the nectar brought to the hive, can trigger navigational, visual or olfactory memories in other experienced bees, and facilitate recruitment to and utilisation of the food source. 4.5.4 Other cognitive mechanisms used in navigation To properly service the needs of their colony, forager bees need to accurately locate profitable food sources, and efficiently transport nectar and pollen back to the hive. The mechanisms by which bees achieve this are numerous, and we are only beginning to obtain an in-depth understanding of these processes. It has long been known that the polarisation pattern on the sky provides bees compass information, even when the sun is hidden by clouds (von Frisch 1967, also see Wehner 2001 for a recent review). In both ants and

112

Mario Pahl, Jürgen Tautz and Shaowu Zhang

honeybees, this information can be combined with odometric information about angular and linear movements – the path integration mechanism – which provides the animal with information about its current position relative to the point of departure (Wehner 2003). Honeybees also use landmarks in their visual environment to find a location where they have previously foraged (Gould 1987, Collett and Kelber 1988). The memory of a landmark, which indicates the location of a food source, is stored in the context of the journey to that food source, i.e. along with its expected distance from the hive (Cartwright and Collett 1983), and also with other landmarks that immediately precede the flight target (Chittka et al. 1995a). The use of multiple landmarks to break a journey into several segments also has the effect of increasing the accuracy of navigation across long distances (Srinivasan et al. 1997). However, as honeybees can forage in both novel and familiar terrain, it should come as no surprise that they can use their path integration and landmark-based navigation systems flexibly: Chittka and colleagues have shown that when foraging by familiar landmarks, honeybees are able to suppress their path integration system, even when those landmarks are displaced. When forced to forage in a novel location without learnt landmarks, they use path integration to navigate back to the hive (Chittka et al. 1995b). Honeybees can store at least two different sets of visual stimuli as landmarks that have the potential to trigger memories of two different foraging routes (Zhang et al. 1999). These associative groupings of landmarks can also be used flexibly, as encountering any one stimulus out of the group can facilitate the recall of the other stimuli. The advantage of such a system would be the following: if a forager, who is attempting to retrace a learnt route indicated by a set of landmarks, becomes somehow disoriented, she could still potentially re-orient herself with respect to the route as long as she encounters any one of those landmarks while searching, regardless of its place in the learnt sequence of landmarks. The distance from the hive to a food source can also be precisely estimated by honeybees, using the phenomenon of optic flow, which is the movement of the image of an animal’s surroundings, as detected by its eyes, as the animal changes position. Two experiments have demonstrated that it is such image motion that tells a bee how far it has travelled, in both naturalistic (Esch and Burns 1995) and controlled (Srinivasan et al. 2000) situations. In the first study, it was shown that honeybees that flew from a hive to a feeder, both placed on the roofs of tall buildings, underestimated their distance travelled, as indicated by their waggle dances. This was because flying at a great height reduces the amount of optic flow produced by the image of the ground. In the second experiment, bees that were made to fly into a narrow tunnel with black and white vertical stripes on its walls

Honeybee cognition

113

(to produce maximum image motion), signaled large distances in their dances, even though they had only flown about 12 metres. Here, the reverse happened: the narrow tunnel amplified the optic flow registered by the bees’ eyes, leading them to overestimate the distance travelled. Finally, there has been some debate in the literature concerning the use of map-based strategies in navigation by honeybees, as opposed to the simpler, route learning mechanisms (e.g. with the help of landmarks) discussed above. Using harmonic radar, Menzel et al. found that displaced bees were able to set a new course at this arbitrary point, and choose to fly to at least two locations (an artificial feeder or the hive) from there. This result is suggestive of a map-like organisation of spatial memory in navigating honeybees (Menzel et al. 2005). Recently, however, Wray et al. (2008) demonstrated that potential foragers inside the hive will not reject dances that advertise food resources in improbable locations, such as the middle of a lake – argueing against the possibility that bees possess a spatial map of the hive’s surroundings. In another study, bees fitted with RFID transmitters, and released at distant locations up to 11 km away, where they would probably never have foraged, were recorded as having returned to the hive, sometimes after as long as 48 hours (Pahl et al. in prep.). However, the rate and speed at which bees returned to the hive was strongly affected by the direction in which the bees were released. This indicates that bees make use of either prominent landmarks or landscape features to successfully navigate in some locations around the hive, but are greatly challenged when displaced to new locations.

4.6 Conclusions Successful foraging is crucial for the survival of a honeybee colony, and foragers act as its sensory units. Honeybees constantly integrate environmental information, in order to maximise individual foraging efficiency, and also adjust their behaviour according to the colony’s needs (Seeley 1995). 100 million years of co-evolution with angiosperm plants (Hu et al. 2008) have resulted in sensory systems that are well-adapted to this task. The studies reviewed here show a rich repertoire of adaptive behaviours, some of which were earlier believed to be restricted to bigger-brained animals. Many of the results cannot be explained by simple associative or operant learning, and thus demonstrate that complex learning is possible even in small brains. Honeybees face complex choices inside and outside the hive, and the studies discussed in this chapter illuminate some of the mechanisms involved in their decision making.

114

Mario Pahl, Jürgen Tautz and Shaowu Zhang

We showed that honeybees are capable of forming and using categories, in which objects can be grouped together. A honeybee is thus able to use a similar response to all stimuli that match one category, and it does not have to learn an appropriate response for each new stimulus. In experimental setups, honeybees are able to categorise the shared features of artificial patterns, as well as those of natural objects. Rules help honeybees to cope with often recurring situations, e.g. repeatedly visiting the same plant for foraging. They can learn to use visual and olfactory stimuli as an indicator of what to do next; in a complex maze setup, they learn abstract rules in order to successfully navigate the maze. In a DMTS setup, honeybees quickly learn the delayed-matching rule, which requires not only long-term memory to remember the rule itself, but also short-term memory to recall the sample pattern after the delay. Context learning is another way to flexibly cope with motivational and environmental changes. During the outbound foraging flight to a food source, the foraging bee has to recall a different route memory than on the inbound homing flight. The different tasks at hand, ‘foraging’ and ‘homing’, set the memories in different contexts, thus avoiding confusion. When plant species differ in their peak nectar secretion times, the time of day can separate the competing plants into different contexts. Thus, honeybees learn to visit one profitable plant in the morning, and a different one in the afternoon. Such a circadian-timed episodic-like memory allows honeybees to remember tasks within a temporal context, and to ‘plan’ their foraging trips most efficiently. Cross-modal associative recall can motivate experienced forager bees inside the hive to fly out and search for a particular flower, whose scent is borne by the bee advertising that food source through a waggle dance. This highly adaptive mechanism leads to faster recruitment to profitable food sources, and thus enhances the hive’s competitiveness. Honeybees navigate successfully over several kilometres when collecting different resources, and when swarming to a new nest site. Through the dance language, the foraging bees communicate the direction and distance of those locations to their hive mates. Honeybees accomplish the impressive task of finding such food sources with the help of their polarisation compass, a path integration system and local and global landmark memories. There is currently mixed evidence in favour of a map-like spatial memory in honeybees. Further experiments are needed to investigate this controversial question. Throughout this chapter, we have presented research that demonstrates that honeybees are not the simple reflexive automata that they were once assumed to be. The cognitive capabilities of an animal seem to be largely

Honeybee cognition

115

governed by what it needs to pursue its lifestyle, and not by the presence or absence of a backbone.

Acknowledgements We are grateful for Aung Si’s valuable help with the manuscript, and for Peter Kappeler’s constructive comments on an earlier version. MP is supported by a grant of the German Excellence Initiative to the Graduate School of Life Sciences, University of Würzburg.

References Bitterman ME, Menzel R, Fietz A, Schäfer S (1983) Classical conditioning of proboscis extension in honeybees (Apis mellifera). J Comp Psychol 97:107119 Blough DS (1959) Delayed matching in the pigeon. J Exp Anal Behav 2:151-160 Brandt R, Rohlfing T, Rybak J, Krofczik S, Maye A, Westerhoff M, Hege H-C, Menzel R (2005) Three-dimensional average-shape atlas of the honeybee brain and its applications. J Comp Neurol 492:1-19 Brannon EM, Terrace HS (2000) Representation of the numerosities 1-9 by rhesus macaques (Macaca mulatta). J Exp Psychol Anim Behav Process 26:31-49 Cartwright BA, Collett TS (1983) Landmark learning in bees – experiments and models. J Comp Physiol A 151:521-543 Cheng K (2005) Context cues eliminate retroactive interference effects in honeybees (Apis mellifera). J Exp Biol 208:1019-1024 Chittka L, Geiger K (1995) Can honey bees count landmarks? Anim Behav 49:159-164 Chittka L, Tautz J (2003) The spectral input to honeybee visual odometry. J Exp Biol 206:2393-2397 Chittka L, Shmida A, Troje N, Menzel R (1994) Ultraviolet as a component of flower reflections, and the colour perception of hymenoptera. Vision Res 34:1489-1508 Chittka L, Geiger K, Kunze J (1995a) The influences of landmarks on distance estimation of honey bees. Anim Behav 50:23-31 Chittka L, Kunze J, Shipman C, Buchmann SL (1995b) The significance of landmarks for path integration in homing honeybee foragers. Naturwissenschaften 82:341-343 Chittka L, Gumbert A, Kunze J (1997) Foraging dynamics of bumble bees: correlates of movements within and between plant species. Behav Ecol 8:239-249 Chittka L, Thomson JD, Waser NM (1999) Flower constancy, insect psychology, and plant evolution. Naturwissenschaften 86:361-377

116

Mario Pahl, Jürgen Tautz and Shaowu Zhang

Clayton NS, Dickinson A (1998) Episodic-like memory during cache recovery by scrub jays. Nature 395:272-274 Colborn M, Ahmad-Annuar A, Fauria K, Collett TS (1999) Contextual modulation of visuomotor associations in bumble-bees (Bombus terrestris). Proc R Soc Lond B 266:2413-2418 Collett TS, Baron J (1994) Biological compasses and the coordinate frame of landmark memories in honeybees. Nature 368:137-140 Collett TS, Cartwright BA (1983) Eidetic images in insects: their role in navigation. Trends Neurosci 6:101-105 Collett TS, Collett M (2002) Memory use in insect visual navigation. Nat Rev Neurosci 3:542-552 Collett TS, Kelber A (1988) The retrieval of visuo-spatial memories by honeybees. J Comp Physiol A 163:145-150 Dacke M, Srinivasan MV (2008) Evidence for counting in insects. Anim Cogn 11:683-689 Dale RH (1988) Spacial memory in pigeons on a four-arm radial maze. Can J Psychol 42:78-83 Dale K, Harland DP, Manning-Jones A, Collett TS (2005) Weak and strong priming cues in bumblebee contextual learning. J Exp Psychol 208:65-74 D’Amato MR, Salmon DP, Colombo M (1985) Extent and limits of the matching concept in monkeys (Cebus apella). J Exp Psychol Anim Behav Process 11:35-51 Davis H (1984) Discrimination of the number three by a raccoon (Procyon lotor). Anim Learn Behav 12:409-413 Deisig N, Lachnit H, Giurfa M, Hellstern F (2001) Configural olfactory learning in honeybees: negative and positive patterning discrimination. Learn Mem 8:70-78 Dyer AG, Neumeyer C, Chittka L (2005) Honeybee (Apis mellifera) vision can discriminate between and recognise images of human faces. J Exp Biol 208:4709-4714 Dyer AG, Rosa MGP, Reser DH (2008) Honeybees can recognise images of complex natural scenes for use as potential landmarks. J Exp Biol 211:1180-1186 Esch HE, Burns JE (1995) Honeybees use optic flow to measure the distance to a food source. Naturwissenschaften 82:38-40 Fauria K, Dale K, Colborn M, Collett TS (2002) Learning speed and contextual isolation in bumblebees. J Exp Biol 205:1009-1018 Fuchs S, Kralj J, Tautz J (2006) Radio frequency identification tags (RFID) for monitoring of worker bees infested with Varroa destructor or with Nosema apis at the hive entrance. Apidologie 37:643-645 Giurfa M (2003) Cognitive neuroethology: dissecting non-elemental learning in a honeybee brain. Curr Opin Neurobiol 13:726-735 Giurfa M (2007) Behavioral and neural analysis of associative learning in the honeybee: a taste from the magic well. J Comp Physiol A 193:801-824 Giurfa M, Malun D (2004) Associative mechanosensory conditioning of the proboscis extension reflex in honeybees. Learn Mem 11:294-302

Honeybee cognition

117

Giurfa M, Eichmann B, Menzel R (1996) Symmetry perception in an insect. Nature 382:458-461 Giurfa M, Zhang SW, Jenett A, Menzel R, Srinivasan MV (2001) The concepts of ‘sameness’ and ‘difference’ in an insect. Nature 410:930-933 Gould JL (1987) Honey bees store learned flower-landing behaviour according to the time of day. Anim Behav 35:1579-1581 Gross HJ, Pahl M, Si A, Zhu H, Tautz J, Zhang SW (2009) Number-based visual generalisation in the honeybee. PLoS ONE 4:e4263, doi:10.1371/journal.pone.0004263 Hellstern F, Wüstenberg D, Hammer M (1995) Contextual learning in honeybees under laboratory conditions. In: Elsner N, Menzel R (eds) Learning and Memory. Proceedings of the 23rd Göttingen Neurobiology Conference. Vol. I. Thieme Verlag, Stuttgart, abstract 30 Herman LM, Gordon JA (1974) Auditory delayed matching in the bottlenose dolphin. J Exp Anal Behav 21:19-26 Horridge GA (1996) The honeybee (Apis mellifera) detects bilateral symmetry and discriminates its axis. J Insect Physiol 42:755-764 Horridge GA, Zhang SW (1995) Pattern vision in honeybees (Apis mellifera): flower-like patterns with no predominant orientation. J Insect Physiol 41:681688 Horridge GA, Zhang SW, Lehrer M (1992) Bees can combine range and visual angle to estimate absolute size. Philos Trans R Soc Lond B 337:49-57 Hu S, Dilcher DL, Jarzen DM, Taylor DW (2008) Early steps of angiospermpollinator coevolution. Proc Natl Acad Sci USA 105:240-245 Hunt S, Low J, Burns KC (2008) Adaptive numerical competency in a foodhoarding songbird. Proc R Soc Lond B 275:2373-2379 Johnson DL, Wenner AM (1966) A relationship between conditioning and communication in honey bees. Anim Behav 14:261-265 Judd SPD, Collett TS (1998) Multiple stored views and landmark guidance in ants. Nature 392:710-714 Keller FS, Schoenfeld WN (1950) Principles of Psychology. Appleton-CenturyCrofts, New York Kilian A, Yaman S, von Fersen L, Güntürkün O (2003) A bottlenose dolphin discriminates visual stimuli differing in numerosity. Learni Behav 31:133-142 Konorski J (1959) A new method of physiological investigation of recent memory in animals. Bull Acad Pol Sci Biol 7:115-117 Lehrer M, Srinivasan MV, Zhang SW (1990) Visual edge detection in the honeybee and its chromatic properties. Proc R Soc Lond B 238:321-330 Letzkus P, Ribi WA, Wood JT, Zhu H, Zhang SW, Srinivasan MV (2006) Lateralization of olfaction in the honeybee Apis mellifera. Curr Biol 16:1471-1476 Maleszka R, Helliwell P (2001) Effect of juvenile hormone on short-term olfactory memory in young honeybees (Apis mellifera). Horm Behav 40:403-408 Menzel R, Blakers M (1976) Colour receptors in the bee eye – morphology and spectral sensitivity. J Comp Physiol A 108:11-13 Menzel R, Mueller U (1996) Learning and memory in honeybees: from behavior to neural substrates. Annu Rev Neurosci 19:379-404

118

Mario Pahl, Jürgen Tautz and Shaowu Zhang

Menzel R, Greggers U, Smith A, Berger S, Brandt R, Brunke S, Bundrock G, Hülse S, Plümpe T, Schaupp F, Schüttler E, Stach S, Stindt J, Stollhoff N, Watzl S (2005) Honey bees navigate according to a map-like spatial memory. Proc Natl Acad Sci USA 102:3040-3045 Pahl M, Zhu H, Pix W, Tautz J, Zhang SW (2007) Circadian timed episodic-like memory – a bee knows what to do when, and also where. J Exp Biol 210:3559-3567 Pick CG, Yanai J (1983) Eight arm maze for mice. Int J Neurosci 21:63-66 Prabhu C, Cheng K (2008a) One day is all it takes: circadian modulation of the retrieval of colour memories in honeybees. Behav Ecol Sociobiol 63:11-22 Prabhu C, Cheng K (2008b) Recency preference of odour memory retrieval in honeybees. Behav Ecol Sociobiol 63:23-32 Prete FR (2004) Complex Worlds from Simpler Nervous Systems. MIT Press, Cambridge/MA Ratnieks FLW, Wenseleers T (2008) Altruism in insect societies and beyond: voluntary or enforced? Trends Ecol Evol 23:45-52 Reinhard J, Srinivasan MV, Guez D, Zhang SW (2004a) Floral scents induce recall of navigational and visual memories in honeybees. J Exp Biol 207:43714381 Reinhard J, Srinivasan MV, Zhang SW (2004b) Olfaction: scent-triggered navigation in honeybees. Nature 427:411 Reynolds AM, Smith AD, Reynolds DR, Carreck NL, Osborne JL (2007) Honeybees perform optimal scale-free searching flights when attempting to locate a food source. J Exp Biol 210:3763-3770 Roberts WA (1972) Short-term memory in pigeon: effects of repetition and spacing. J Exp Psychol 94:74-83 Ronacher B (1992) Pattern-recognition in honeybees: multidimensional-scaling reveals a city-block metric. Vision Res 32:1837-1843 Rossel S, Wehner R (1986) Polarization vision in bees. Nature 323:128-131 Seeley TD (1995) The Wisdom of the Hive. The Social Physiology of Honey Bee Colonies. Harvard University Press, Cambridge/MA Shettleworth SJ (1998) Cognition, Evolution, and Behavior. Oxford University Press, New York Skorupski P, Chittka L (2006) Animal cognition: an insect’s sense of time? Curr Biol 16:R851-R853 Srinivasan MV, Lehrer M (1988) Spatial acuity of honeybee vision and its spectral properties. J Comp Physiol A 162:159-172 Srinivasan MV, Zhang SW, Bidwell NJ (1997) Visually mediated odometry in honeybees. J Exp Biol 200:2513-2522 Srinivasan MV, Zhang SW, Zhu H (1998) Honeybees link sights to smells. Nature 396:637-638 Srinivasan MV, Poteser M, Kral K (1999) Motion detection in insect orientation and navigation. Vision Res 39:2749-2766 Srinivasan MV, Zhang SW, Altwein M, Tautz J (2000) Honeybee navigation: nature and calibration of the ‘odometer’. Science 287:851-853

Honeybee cognition

119

Streit S, Bock F, Pirk CWW, Tautz J (2003) Automatic life-long monitoring of individual insect behaviour now possible. Zoology 106:169-171 Su S, Cai F, Si A, Zhang SW, Tautz J, Chen S (2008) East learns from West: Asiatic honeybees can understand dance language of European honeybees. PLoS ONE 3:e2365. doi:10.1371/journal.pone.0002365 Tautz J (2008) The Buzz about Bees: Biology of a Superorganism. Springer, Berlin Troje NF, Huber L, Loidolt M, Aust U, Fieder M (1999) Categorical learning in pigeons: the role of texture and shape in complex static stimuli. Vision Res 39:353-366 Uller C, Jaeger R, Guidry G, Martin C (2003) Salamanders (Plethodon cinereus) go for more: rudiments of number in an amphibian. Anim Cogn 6:105-112 van Hateren JH, Srinivasan MV, Wait PB (1990) Pattern recognition in bees: orientation discrimination. J Comp Physiol A 167:649-654 Vareschi E (1971) Duftunterscheidung bei der Honigbiene – Einzelzell-Ableitungen und Verhaltensreaktionen. Z Vergl Physiol 75:143-173 von Frisch K (1967) The Dance Language and Orientation of Bees. Harvard University Press, Cambridge/MA Wasserman EA (1993) Comparative cognition: beginning the second century of the study of animal intelligence. Psychol Bull 113:211-228 Wehner R (2001) Polarization vision – a uniform sensory capacity? J Exp Biol 204:2589-2596 Wehner R (2003) Desert ant navigation: how miniature brains solve complex tasks. J Comp Physiol A 189:579-588 Wehner R, Bleuler S, Nievergelt C, Shah D (1990) Bees navigate by using vectors and routes rather than maps. Naturwissenschaften 77:479-482 Wehner R, Michel B, Antonsen P (1996) Visual navigation in insects: coupling of egocentric and geocentric information. J Exp Biol 199:129-140 Williams RW, Herrup K (1988) The control of neuron number. Annu Rev Neurosci 11:423-453 Witthöft W (1967) Absolute Anzahl und Verteilung der Zellen im Hirn der Honigbiene. Zoomorphology 61:160-184 Wray MK, Klein BA, Mattila HR, Seeley TD (2008) Honeybees do not reject dances for ‘implausible’ locations: reconsidering the evidence for cognitive maps in insects. Anim Behav 76:261-269 Zhang SW (2006) Learning of abstract concepts and rules by the honeybee. Int J Comp Psychol 19:318-341 Zhang SW, Srinivasan MV (1994) Prior experience enhances pattern discrimination in insect vision. Nature 368:330-332 Zhang SW, Srinivasan MV (2004a) Exploration of cognitive capacity in honeybees: higher functions emerge from a small brain. In: Prete FR (ed) Complex Worlds from Simpler Nervous Systems. MIT Press, Cambridge/MA, pp 41-74 Zhang SW, Srinivasan MV (2004b) Visual Perception and Cognition in Honeybees. In: Chalupa LM, Werner JS (eds) The Visual Neurosciences. MIT Press, Cambridge/MA, pp 1501-1513

120

Mario Pahl, Jürgen Tautz and Shaowu Zhang

Zhang SW, Bartsch K, Srinivasan MV (1996) Maze learning by honeybees. Neurobiol Learn Mem 66:267-282 Zhang SW, Lehrer M, Srinivasan MV (1999) Honeybee memory: navigation by associative grouping and recall of visual stimuli. Neurobiol Learn Mem 72:180-201 Zhang SW, Srinivasan MV, Zhu H, Wong J (2004) Grouping of visual objects by honeybees. J Exp Biol 207:3289-3298 Zhang SW, Bock F, Si A, Tautz J, Srinivasan MV (2005) Visual working memory in decision making by honey bees. Proc Natl Acad Sci USA 102:5250-5255 Zhang SW, Schwarz S, Pahl M, Zhu H, Tautz J (2006) Honeybee memory: a honeybee knows what to do and when. J Exp Biol 209:4420-4428

Chapter 5

Individual performance in complex social systems: the greylag goose example KURT KOTRSCHAL, ISABELLA B.R. SCHEIBER AND KATHARINA HIRSCHENHAUSER

ABSTRACT Convergent social structures can be found in taxa that split a long time ago, for example more than 230 Mio years ago as in the case of mammals and birds. Such convergence is explained by common selection regimes, as all social systems are shaped by sex-specific tactics and strategies to optimise their reproductive success. In addition, the major social mechanisms, brain and physiology, are highly conserved throughout the vertebrates. Manoeuvring social contexts tends to be energetically costly and, hence, favours efficient decision-making. Therefore, at least in vertebrates, complex social systems generally select for social cognition. As an example for social convergence between mammals and birds, we introduce the surprisingly complex social system of greylag geese, featuring components such as a female-bonded clan structure, long parent-offspring relationships, as well as elaborate and highly functional patterns of mutual social support. Our results show that partners in reproductively successful goose pairs are in hormonal synchrony and provide social support to each other. We suggest that social support may be a major structuring principle of other social systems with long-term individualized and valuable partnerships as well. In general, individual performance in social systems is determined by the interplay between proximate mechanisms and ultimate functions.

122

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

5.1 Introduction: being social is the default setting in mammals and birds Vertebrate bodies, physiologies and minds are shaped by individual relationships with others. Being social may be useful, rewarding, disastrous, pleasant or intimidating, but it is certainly unavoidable in most species. Therefore, individuals and species cannot be understood via their adaptations to the physical and ecological environments alone, but their social dimension needs to be considered as well. In contrast to simple groups, such as aggregations that coordinate in space and time (Krause and Ruxton 2002), many social systems are relatively complex, i.e. individuals live with a partner and/or a group permanently or seasonally, and they engage in specific, valuable and durable relationships. In fact, the essence of all social life is to optimise one’s own fitness whilst keeping one’s own social web intact (Fig. 5.1). This insight points at the importance of the individual styles of stress coping and conflict behaviour (Aureli and de Waal 2000, Aureli et al. 2002, Koolhaas et al. 2007) for sociality. At least in homoeothermic vertebrates, sociality is probably the default option. This assumption is supported by a number of structures and mechanisms, which feature prominently in vertebrate brain and physiology (Fig. 5.2), including the ‘social behaviour network’ (Goodson 2005, McGregor 2005), bonding mechanisms (Curley and Keverne 2005, Goodson et al. 2009), basic emotional systems (Panksepp 1998, 2005) and the two stress

Cognition and emotionality

Individual social performance and fitness Social context / embedding

Coping with social stress

Fig. 5.1 Schematic representation of the major interacting faculties relevant for individual decision making and fitness in complex social systems: Individual stress coping is determined by individual behavioural phenotype (including emotionality), social embedding and social experiences past. This simple model is also applicable for ontogeny, particularly epigenetics.

Individual performance in complex social systems

123

Brain

Genetic background

NCl

Maternal effects Parenting style

SAA

HPA

PC

Stress systems

Adrenals

Cardiovascular system

Fig. 5.2 Diagram of some of the components of a ‘common social toolbox’ contributing to explain some of the parallels in social organisation between birds, such as geese, and mammals. Brain and major peripheral elements are shown that are involved in homoeothermic vertebrate social behaviour, mediating social bonding/attachment, social support, as well as basic emotional systems. The ‘socio-sexual diencephalic and tegmental network’ (Goodson 2005; horizontal hatching) governs the ‘instinctive’ social behaviour, which is controlled by the prefrontal cortex (PC) in mammals and by the posterior forebrain caudolateral nidopallium (NCl) in birds. These brain systems communicate tightly with the two stress axes, the hypothalamo-pituitary axis (HPA; pituitary in black, adrenal cortex vertically hatched) and the sympathico-adrenergic axis (SAA, adrenal core dotted). The setpoints regulating the interactions of these systems are shaped during ontogeny by genetic background, maternal effects, parenting styles and ongoing social interactions, thereby, social styles may be epigenetically passed on over generations.

coping systems, which were conservatively maintained over evolutionary times, the fast sympathetico-adrenomedullary (SAS) and slow hypothalamic-pituitary-adrenocortical (HPA) stress axes (Sapolsky 1992, von Holst 1998, Summers 2002). In fact, the greatest modulation of these (anti-)stress systems generally occurs in social contexts (DeVries et al. 2003, McEwen and Wingfield 2003, Wascher et al. 2008a,b). Since the groundbreaking work of Hamilton (1964) and Wilson (1975), biologists have begun to incorporate dynamic social components into their concepts. Essential input was provided by experimental psychologists and

124

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

primatologists (de Waal 1982, Byrne and Whiten 1988, Whiten and Byrne 1997). Social behaviour is difficult to deal with scientifically, not the least because individual behaviour in a social net depends on what others do, or even on their mere presence (McGregor 2005). Moreover, social behaviour, social embedding and social experience feed back onto the brain structures and physiological mechanisms involved, contributing to individual phenotypic plasticity (Fig. 5.2). Emotional arousal caused by virtually any social encounter comes with surges in systemic glucocorticoids, which affect individual energetics, learning and memory formation via dose-dependent effects on the nervous system (McEwen and Sapolsky 1995), and will also modulate an individual’s disposition to respond to future challenges (Sapolsky et al. 2000, Kruk et al. 2004). In many social systems, the experience of previous social interactions, as well as the presence of social allies, will affect the balance between key hormones, such as oxytocin, glucocorticoids and androgens, as well as neurotransmitters. This, in turn, will affect individual motivation and the likelihood of engaging and succeeding in prospective social interactions (e.g., McGregor et al. 1997, Oliveira et al. 1998, Peake et al. 2001, 2002). Via epigenetic inheritance such socially induced changes may even extend over generations (Roemer et al. 1997, Champagne and Curley 2005, Daisley et al. 2005, Bertin et al. 2008, 2009).

5.2 Social efficiency and social competence Ultimately, individual success in any biological system is defined by the number of reproductively successful offspring produced. In addition to being adapted for coping optimally with the physical and ecological environment, success in a social system will depend upon ‘social efficiency’ and ‘social competence’, which may be considered as two sides of the same coin. Socially competent individuals will reach their immediate (social) goals with a minimal agonistic and other behavioural effort and with minimal damage, if any, to its long-term valuable relationships. Thereby, socially competent individuals will also be efficient, i.e. they will reach their immediate (social) goals with a minimum of energetic investment. Hence, individual performance and the extent of embedding in the social web should be fitness-relevant, because energy saved in interactions should be available for other domains, such as reproduction. Social competence and efficiency will probably become the more important, the more complex a social system is. Social performance may depend upon sex, behavioural phenotype, early socialisation and individual experience and life

Individual performance in complex social systems

125

history state, on partner behaviour, audience, and season, just to name a few of the most important factors.

5.3 Cognition-based complex social systems There is no consent on how to define a complex social system. Social systems may either be considered complex if they result in complex patterns and outcomes, such as in Hymenoptera, or if they feature intreractions, which rely at least partly on complex interactions and cognition, such as mutual, long-term valuable dyadic relationships (among mates, parents and offspring, unrelated social allies, friends, etc.), adaptive and varied patterns of conflict resolution (Aureli and de Waal 2000, Aureli et al. 2002), the ability to track higher-level relationships (Judge and Mullen 2005), social support (Sapolsky 1993, von Holst 1998, Abbott et al. 2003, Goymann and Wingfield 2004), distinctive differences in individual behavioural phenotypes and associated social roles (Sih et al. 2004), individual contextdependent flexibility, the forming of traditions, or ‘cultural divergence’ (van Schaik this volume) via social learning, etc. Such an understanding of social complexity includes flexibility in behavioural tactics and hence, is based on cognitive mechanisms, rather than on strategies, i.e., inherited decision rules (Whiten and Byrne 1997, Taborsky and Brockman this volume). As a result, vertebrate societies will show a different kind of complexity than insect societies, with the prevalence of strategies in the latter (see Heinze this volume) and tactics in the former. At least in the homoeothermic vertebrates, social complexity seems to select for large brains and complex cognition (Dunbar 1995, Dunbar and Bever 1998; but see Barrett and Henzi 2005, Barrett et al. 2007, Beauchamp and Fernández-Juricic 2004, Healy and Rowe 2007).

5.4 Complex social systems in birds: how large a brain does one need? For a long time, birds have been held incapable of complex cognition because of their mainly ‘striatal’ forebrain. Today, it is recognised that birds have a proportion of pallial forebrain equivalent to the mammalian cortex (Jarvis et al. 2005). Moreover, the nidopallium caudolaterale in birds is functionally equivalent to the mammalian prefrontal cortex, which is involved in higher functions such as controlling of instinctive impulses, reasoning, all kinds of decision making, generating socially appropriate be-

126

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

haviour and conscious thought (Divac et al. 1994, Güntürkün 2005). Hence, the striking parallels of social organisation in birds and mammals may be based on similar (both, homologous and convergent) cognitive mechanisms (Emery and Clayton 2004a). Some birds, such as geese, are indeed able social learners (Fritz and Kotrschal 2002), and the learning and reasoning abilities of birds closely parallel those of mammals (Huber et al. 2001, Huber 2002, 2007, Schloegl et al. 2009). In fact, advanced cognitive abilities were recently shown in the ‘brainy’ crow family (Dally et al. 2006, Bugnyar 2007) and in some parrots (Pepperberg 1999, Iwaniuk and Nelson 2003). These abilities include mental time travelling (Clayton and Dickinson 1998, Clayton et al. 2001, Raby et al. 2007), advanced cooperative abilities (Huber et al. 2008), technical and physical understanding (Auersperg et al. 2009), self-recognition in a mirror (Prior et al. 2008), and theory of mind (i.e. knowing about others’ minds; Dally et al. 2006, Bugnyar 2007, Kotrschal et al. 2007). Despite the fact that relatively smaller-brained birds express social complexities, which rival those of some mammals, there is still scepticism to what extent this may be based on cognitive abilities (Emery et al. 2007; but see Scheiber et al. 2008). In support of the involvement of social cognition, greylag geese show differentiated heart rate responses when watching third party interactions (Wascher et al. 2009) and perform remarkably well in a transitive inference task (Weiß et al. 2009; i.e. a well established operant procedure, in which the ordering of a set of arbitrary stimuli can be inferred from a series of dyadic comparisons; Bond et al. 2003). This does not imply, however, that all social interactions appearing complex require high-level cognition (Barrett et al. 2007, Healy and Rowe 2007).

5.5 The greylag goose society Concepts of social complexity have been discussed mainly in the mammalian, notably primate literature (see de Waal and Tyack 2003 for a review). However, recent evidence indicates considerable mammal-like social complexity in birds as well, for example in corvids, parrots and greylag geese (Emery and Clayton 2004b, Kotrschal et al. 2007, Scheiber et al. 2008). Greylag geese (Anser anser), in general, are long-lived and long-term monogamous birds, showing family cohesion over years and a female-centred clan organisation, elaborate post-conflict behaviour and distinct behavioural and physiological patterns related to social support; they form social traditions, assume personality-related individual social roles and may even be knowledgeable of third party relationships. The basics of the greylag

Individual performance in complex social systems

127

BOX 5.1 Greylag goose social system and biology Greylag geese (Anser anser) are long-lived (more than 14 years in the wild, Hudec and Rooth 1970) and form long-term monogamous pairs (Lorenz 1988), which are the elementary social unit of a flock. Even at a balanced flock sex ratio, males may engage in pair-bonds with other males, which is a male tactic to maintain high rank (Kotrschal et al. 2006). Excess females will attach as a ‘secondary’ female to a heterosocial pair. Trios of two males and one female are formed when a female attaches to a homosocial male pair. Social units maintain spatial proximity and perform the ‘triumph ceremony’ together (Fischer 1965, Radesäter 1974). Individual dominance rank is conditional on social relationships (Lamprecht 1986, 1987, Lorenz 1988, Black et al. 2007): highest ranking are family units, where family members acquire the gander’s rank. Pairs without offspring rank intermediate, singletons rank lowest. Family members provide active and passive social support to each other (Weiß and Kotrschal 2004, Scheiber et al. 2005a, 2009a,b). Interindividual behavioural variation exists also with regards to social roles based on personalities (Pfeffer et al. 2002, Kralj-Fiser et al. 2007). Geese flock from summer into late winter to avoid predation (Lazarus 1978, Lazarus and Inglis 1978, Kotrschal et al. 1992). As facultative migrants, flocks may temporarily or permanently cease their migratory traditions when mild winters afford sufficient food supply (Rutschke 1982, 1997, Kear 1990). After returning from their wintering grounds, flocks disintegrate into pairs. At this time, greylag females form loose breeding colonies. The non-breeders stay together or aggregate for moulting in May and June. Females lay up to 8 eggs, which they alone incubate for 28 days. Young females tend to dump eggs into the nests of experienced geese (Weigmann and Lamprecht 1991). During incubation, the gander stays close, but does not approach the nest unless it needs defending. Goslings and females leave the nest approximately 24 – 48 hours after hatching, when also the gander joins again. Usually the pair raises their young in proximity of other families. Goslings need to be brooded regularly by the female for their first 3 weeks of life. They forage for themselves immediately after leaving the nest, but pay attention to the parental beaks, thereby forming feeding traditions (Fritz and Kotrschal 2002). A gander defends female and goslings against other flock members or predators. Approximately 5 weeks after hatching, parents start moulting their wing feathers. Parents regain the ability to fly at the time their young fledge. Then families form summer foraging flocks with non-breeders, increasingly integrating into winter flocks in autumn. For example, up to 20.000 greylag geese from all over Europe aggregate at Neusiedlersee. They may stay there until cold spells or snowfall inhibits access to food at harvested fields around the lake. Only then geese leave for their wintering areas at the Mediterranean coast of Tunisia (Rutschke 1982, 1997).

128

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

BOX 5.2 The Grünau flock of greylag geese Konrad Lorenz established a free-flying, non-migratory flock of greylag geese in the Upper Austrian valley of the river Alm in 1973. The flock is unrestrained and roams the valley between the Konrad Lorenz research station (KLF) and lake Almsee 10 km to the South, which is used as a night roost. For most of the year, the flock can be found on the meadows surrounding the KLF during the day, where it is provided with supplemental (summer) or sustaining (winter) amounts of food twice a day. From breeding (March) to moult (June/July) the flock disintegrates and the geese spread to various locations in the valley, most of which are accessible to researchers. All individuals are marked with coloured leg rings and show neither behavioural nor physiological (heart rate) responses when being approached to a distance of 1.5 m by humans they are acquainted with. Individual life histories and social relationships have been monitored on a daily to weekly basis since the establishment of the flock. Flock size varied between 120 and 170 individuals over the years. Every year up to 10% of the individuals are lost to natural predators, mainly red foxes (Vulpes vulpes; Hemetsberger 2001). Except for the lack of a migratory tradition, the seasonal patterns of the KLF flock closely resemble that of wild geese (Kotrschal and Hemetsberger 1995, Rutschke 1997, Hirschenhauser et al. 1999b). After break-up of the winter flock, females nest at various locations between the KLF and Almsee. They either build nests in relatively predator-safe breeding boxes, or they nest in reeds, where predation pressure is high: approximately 10 females and clutches are lost per year. This leads to a male-biased sex ratio in the flock (up to 70%) and fosters a tradition of breeding in safe boxes (Weiß et al. 2008). Approximately 80% of the hatched goslings die within the first 2 weeks of age due to predation or parasite infections. In general, fledging success increases with maternal age (Hemetsberger 2001). Although the number of heterosexual pairs increased from 16 in 1973 to over 40 in 2000, the number of pairs with fledged young remains remarkably constant at 4.4 ± 2.9 (SD) per year, with a total of 16.5 ± 11.4 fledglings per year. This barely balances the fledged geese lost to predators and dispersal (Hemetsberger 2001). Over the years, 20-30% of the geese in the flock have been hand-raised in sibling groups, resulting in individuals accessible for experiments. Eggs for hand-raising are artificially incubated and frequently originate from outside populations. Thereby, the gene pool is increased and inbreeding is counterbalanced. Shortly after hatching, hand-raised geese are in contact with the flock to ensure proper socialisation. After fledging they fully integrate into the flock and are socially indistinguishable from goose-raised geese, except for a tendency to maintain stable homosocial pair-bonds between brothers (Kotrschal et al. 2006).

Individual performance in complex social systems

129

social system are given in BOX 5.1 and in some detail, in the following, where we will guide the reader through this social system from an individual’s perspective and, as much as possible, also from a proximate point of view (sensu Tinbergen 1963). We will focus on the pair-bond, i.e. behavioural and hormonal synchrony between pair partners, as a cornerstone for coping with the demands of a complex social life. Finally, we will discuss individual social efficiency based on our long-term heart rate data as a proxy for emotional involvement and social energetics, and will emphasise stress reduction by social support as a key factor of the greylag social system. These results are mainly based on our long-term studies of a flock at Grünau (BOX 5.2). 5.5.1 The start of a social career: mothers and eggs Individual dispositions to behave in a certain way in the social domain (‘personality’ or ‘individual behavioural phenotype’: i.e., influence of individual emotionality, stress coping and bonding style, plus experience) are genetically heritable and can be epigenetically affected by ‘maternal effects’ (Bergmüller this volume), as well as by early socialisation and parenting (Sachser and Kaiser this volume). In fact, bird mothers differ in how much they invest into single eggs (Bertin et al. 2008, Groothuis and Schwabl 2008). Egg size varies between females and generally decreases with laying order, the latter being interpreted as an evolutionary preadjustment to varying conditions during raising of young (Groothuis et al. 2005). However, there is also differential investment in regard to the qualitative components of the eggs. The amount of carotenoids, for instance, will affect the chicks’ immune system (Saino et al. 2003), whereas the amount of androgens will affect the behaviour of the offspring. The more androgens the yolk contains, the more vigorously the chick will beg (Schwabl et al. 1997), the more aggressive it may be (Groothuis and Carere 2005) and the more its behavioural phenotype will be shifted towards being proactive (sensu Koolhaas et al. 1999, e.g., Daisley et al. 2005, Bertin et al. 2009). In response to environmental conditions, females may epigenetically adjust offspring phenotype, mainly via hormones in the yolk (Daisley et al. 2005, Groothuis and Carere 2005, Bertin et al. 2008, 2009). Proactive individuals, in contrast to reactive ones, are relatively bold and aggressive in social and non-social contexts, are prone to form routines and scrounge from others and should be less socially dependent and hence, are likely to disperse. Proactives are also generally less stressed by social isolation than reactives. In agreement with the producer-scrounger

130

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

paradigm (Barnard and Sibly 1981), we found that the reactives, rather than the proactives tended to be the innovators, in a group of unmanipulated greylag goslings (Pfeffer et al. 2002): individuals, which had higher baseline levels of glucocorticoids at the age of two weeks solved an operant task faster than others at 6 weeks of age and were innovative producers of food at 10 months of age. Hence, the expression of social performance and social roles seems to involve some predisposition by individual behavioural phenotype (Kralj-Fiser et al. 2007), which is probably mediated to some degree by the maternal deposition of steroid compounds in the yolk (Daisley et al. 2005, Groothuis and Schwabl 2008). 5.5.2 The first year The precocial goslings are faced with social, physical and ecological challenges beginning with their first day after leaving the nest. As an effect of high yolk steroid levels, particularly high levels of corticosterone metabolites are excreted up to 20 days after hatching (McNabb et al. 1998, Frigerio et al. 2001a). Right after hatching, goslings need to be kept safe from predators and apart from other families to prevent fusion with gosling groups from other families (Lorenz 1988, Kalmbach et al. 2005). This is mainly achieved by the male, who also shields his female from being disturbed by other geese when brooding the goslings (Lorenz 1988). Consequently, the males’ corticosterone levels, as well as their involvement in agonistic encounters, increase after hatching of their goslings (Kotrschal et al. 1998, Swoboda et al. submitted). Males reduce their agonistic involvement at approximately 6-7 weeks after hatching. From then on, goslings gradually increase their participation in aggressive encounters towards fledging. In fact, they win agonistic encounters against larger and older individuals, with the active help of their fathers (Scheiber 2009a), but generally lose such agonistic interactions if not supported. Thus, before fledging, the main beneficiaries of social support are the goslings: they tend to win encounters and show low faecal corticosterone metabolites levels (Swoboda et al. submitted). But do parents also benefit from the presence of their young? Contrary to many avian species, goose parents remain with their young for an extended period of time, i.e. approximately one year, after the goslings fledge (‘primary family’). Such a relatively long-lasting family cohesion is uncommon among birds and seems to be not only in the interest of the offspring. Shortly before the goslings fledge, parents have re-grown their wing feathers after moult. Now, particularly females seem to benefit from passive social support in the company of their young. Females did not excrete

Individual performance in complex social systems

131

elevated glucocorticoid metabolites when experimentally exposed to a social density situation (Scheiber et al. 2005a) when they were accompanied by at least three fledged young. Thereby, the fledged young appear to be the ‘physiological helpers’ of their parents (Scheiber et al. 2009a), sensu ‘helpers at the nest’ in communally breeding species (Brown 1987), because females who successfully raised young in one year will enter the next breeding season in better condition than females which failed. Energetic costs of reproduction are usually higher for females than for males (Adkins-Regan 2005). Any activation of the stress systems is energetically costly and frequent/chronic activation may negatively affect an individual’s reproductive success (Bowman et al. 1978, Mendl et al. 1992) and health (von Holst 1998). Therefore, for the females it is particularly crucial to respond adequately and economically to stressors (Nephew et al. 2003). The young also benefit from active social support when still integrated in their family. In case of lost agonistic interactions, they may reverse a prior defeat experience, when family members retaliate by repeatedly attacking the opponent which previously had won against the primarily targeted offspring/sibling (Scheiber et al. 2009b). This may be one of the reasons why juvenile geese rank much higher in the flock hierarchy than expected from their age and size. Such benefits of being socially supported by social allies are probably also the reason for female kin to maintain loose clan bonds (Weiß et al. 2008). 5.5.3 The second year If parents fail to reproduce in their next breeding attempt a year later, the young of previous years may rejoin their family for a second or even third year (‘secondary or tertiary families’, Lorenz 1988). To determine the potential functionality of this, we investigated active and passive social support mechanisms in secondary families (Scheiber et al. 2009a). Contrary to primary families, we were unable to detect any form of active social support in secondary families during agonistic interactions with other flock members, but – similar to primary families – the beneficiaries were the adult and subadult females, with respect to a dampened glucocorticoid modulation (i.e. ‘passive’ social support). For subadult males, ‘motivational components’ (Lamprecht 1986) are most likely the main explanation why they join their parents for a second year: via enhanced agonistic motivation, subadult males benefit from winning agonistic interactions, thereby climbing higher in rank than would be possible outside their family unit (Weiß et al. 2008). When accepting last year’s young, costs for parents are probably negligible. Therefore, joining

132

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Fig. 5.3 Goose Beleriand (right) threatened and attacked by another female, Lanzelot (middle). Her partner Boston (left back) and an unrelated goose, Nazca (left front) provide active social support by co-threatening the opponent. Such an attack may result in a ‘serial attack’ (Scheiber et al. 2009b), in which family members will repeatedly attack one particular opponent. These serial attacks are thought to serve in reversing previous losses, reinforce a losing experience on the side of the opponent or signal the agonistic potential of a family to other flock members. Photo © Brigitte Weiß.

a secondary family when no appropriate mate is available during their first year after fledging may be viewed as an alternative tactic for subadult geese (Scheiber et al. 2009a). 5.5.4 Adult years: dynamic clan structure and associated benefits for females In most mammals, male-biased dispersal and female natal philopatry (Greenwood 1980, Waser and Jones 1983) are the rule; in most cases, their social groups are matrilines, with high levels of female relatedness (Coltman et al. 2003). Most birds, on the other hand, show male philopatry, and in some species neighbourhood male groups are closely related (Höglund et al. 1999, Petrie et al. 1999, Pirtney et al. 1999, Shorey et al. 2000). These differences probably root in the different possibilities in mammals

Individual performance in complex social systems

133

and birds for investing in offspring. Waterfowl (Anatidae) pose a notable exception, where female-biased natal philopatry is the rule, while males usually follow their mate to her hatching area for breeding (Waldeck et al. 2008). Barnacle geese (Branta leucopsis) were shown to nest close to their parents and sisters when breeding on the same island. However, when parents and sisters nested on different islands, females stayed close to their same-age sisters (van der Jeugd et al. 2002). In greylag geese, adult female sisters, but not brothers, were shown to seek each other’s company during rest (Frigerio et al. 2001b). Even when already paired, females were found close to their sisters. The same seems to be true for adult daughters and their mothers, although in this case spatial proximity was less pronounced than in sister-sister pairs (Scheiber and Weiß submitted). A number of hypotheses have been generated to explain the femalecentered population structure in waterfowl (e.g., Eadie et al. 1988, Chesser 1991, Avise et al. 1992, Regehr et al. 2001, van der Jeugd et al. 2002, McKinnon et al. 2006, Waldeck et al. 2008), but stress-reduction through the permanent presence of social allies has seemingly been overlooked. We suggest that social support may be one of the driving forces behind this female-centred social organisation. Any reduction of female (social) stress should be advantageous, particularly during the incubation phase of the breeding cycle and while the goslings need to be brooded (see above). Indeed, females, but not males, enjoyed dampened glucocorticoid responses during stressful situations when part of a secondary family (Scheiber et al. 2009a). The observed female-centred clan structures are a much less determining factor in social organisation than, for example, in macaques (de Ruiter and Geffen 1998, Kappeler and van Schaik 2002) or baboons (Silk et al. 2003, 2009), probably because maternal rank is strictly heritable in macaques, whereas in geese, female rank mainly depends on the male partner. 5.5.5 Pair formation, pair-bond maintenance and hormonal synchrony between pair partners The major function of monogamy is biparental care (Black 1996), the effectiveness of which is probably supported by at least some behavioural and physiological coordination and synchronisation between pair partners. In fact, assortative mating with respect to age and size was shown to optimise lifetime reproductive success in barnacle geese (Choudhury et al. 1996), whereas reproducing with non-preferred partners, as well as divorce may have negative fitness consequences (Black et al. 1996, Bluhm and Gowaty 2004, Angelier et al. 2007). Particularly in monomorphic species

134

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

with long-term pair-bonds, mate choice is inevitably a two-way decision process as each individual is both, the chooser and a chosen (AdkinsRegan 2005). Also, the maintenance of a pair-bond is the result of the ‘bilateral’ interactions between pair partners and of the ‘multilateral’ interactions between the pair and its physical and social environment. Therefore, in search for behavioural and physiological correlates of successful reproduction it may be revealing to shift the unit of analysis from the individual to the pair. For example, goose pairs with a high degree of within-pair testosterone co-variation (TC; i.e. the degree of seasonal testosterone covariation between the male and the female pair partner) produced more offspring in the year sampled, as well as over their lifetime than pairs low in TC (Hirschenhauser et al. 1999a). But why are androgens particularly relevant in social careers? Throughout the vertebrates, male testosterone (T) regulates and responds to courtship, sexual and agonistic behaviour (Goymann and Hofer this volume). Thus, individual T responsiveness is a function of the stability of the social environment (Wingfield et al. 1990). Furthermore, T is the main physiological mediator shifting male investment towards male-male aggressiveness and constrained paternal care. The maintenance of elevated T levels bears costs (Wingfield et al. 2001), pointing at two fundamental phenomena: i) a male must be sufficiently healthy to afford both, T responses to certain behaviours and the behavioural performance itself. Thus, T-related behaviours and traits efficiently advertise fitness and immunocompetence, particularly because T may be regarded a physiological handicap (Folstad and Karter 1992, Zahavi and Zahavi 1997); ii) the male should benefit from the expression of post-conflict T responses. Perhaps T facilitates the likelihood of winning in future agonistic interactions (Oliveira et al. 2009). In female birds, androgens generally mediate sexual receptivity and prebreeding territoriality (Gill et al. 2007). This seems to be a general phenomenon, as also in female mammals androgens were linked to sexual desire and receptivity, which is supported by the fact that peaks occur around ovulation (Longcope 1986, Alexander and Sherwin 1993). Androgens in both sexes may be of gonadal, as well as adrenal origin (Lee and Bahr 1994, Rodríguez-Maldonado et al. 1996, Boonstra et al. 2008). In geese, the androgen metabolites we measure in droppings from female geese are predominantly of gonadal origin (Hirschenhauser et al. 2010). Female androgens as indicators of sexual motivation contribute to explain the parallel androgen levels in the male and female partners of a successful pair. The within-pair testosterone-co-variation (TC) studies in greylag geese (Hirschenhauser et al. 1999a, Weiß et al. 2010) provide one of the rare examples in which steroid patterns were linked with long-term reproductive success (Adkins-Regan 2005). TC predicted both, the pair’s reproductive

Individual performance in complex social systems

135

output, as well as individual fitness of both pair partners. We suggest different contributions to a pair’s TC from females and males: the female needs to assemble with a complementary and compatible male partner able to adapt behaviourally and physiologically to her needs as related to breeding. A male’s continuous and flexible androgen responsiveness to the female seems essential for the pair’s joint status and performance within the flock. This certainly includes both partners’ stress responsivity to winning or losing a pair-bond challenge during frequent ‘soap operas’ in the mating season (Hirschenhauser et al. 2000). Surprisingly, however, pair-bond duration was not related to hormonal synchrony between pair partners (Hirschenhauser et al. 1999a). Meanwhile, we know from analyses of further 53 greylag goose pairs that hormonal synchrony between pair partners may even decrease over time spent together. Simple ageing of the pair partners could not explain this pattern, as experienced older females had higher TC ranks with their pair partners than younger females (Weiß et al. 2010).

Fig. 5.4 A wing-shoulder fight of two ganders. Individuals grasp each other by the shoulder and violently strike the opponent with the knob at the frontal edge of their wings. Such severe fights may last seconds to minutes and usually occur in competition for partners. Such interactions will cause maximum activity of the stress axes. Particularly losers of such interactions would benefit from social support provided by their partners (‘consolation’). Photo © Brigitte Weiß.

136

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Experiments with domestic geese revealed the female contribution to a pair’s TC: females with preferred partners had higher seasonal androgen peaks during egg formation and later on, were more likely to initiate incubation than females paired at random; however, mate choice per se did not explain TC variation. When the same individuals were allowed to freely attach and interact in a large flock, the TC-ranks we previously observed in pairs were generally reduced. Thus, TC is responsive and vulnerable to the social environment (Hirschenhauser et al. 2010). Also in greylag geese, TC must be viewed as a physiological correlate of the current social condition of the pair in the flock context rather than a pair-specific, lasting and stable trait (Weiß et al. 2010). In sum, it seems that partners, which do not match early on, probably never will. Yet, on top of this, compatible partners still need a favourable social environment to express high TC. 5.5.6 Social energetics Winning or a losing in a single encounter does not have to be particularly stressful, whereas repeated challenge or defeat may result in chronic stress load. Whether social stress affects dominants or subordinates to a greater degree is context-dependent (Kotrschal et al. 1998, Sapolsky 2002, Abbott et al. 2003, Goymann and Wingfield 2004). As in other species, individual geese differ in how they cope with social conflicts, reflected in individual differences of glucocorticoid responsiveness (Koolhaas et al. 1999, Goymann and Wingfield 2004, Kralj-Fiser et al. 2007). The modulation of both stress axes, SAS and HPA, correlates with oxygen consumption; hence, social conflict may be energetically expensive (McEwen and Wingfield 2003). A prime physiological parameter to study social stress is heart rate (HR), which is highly variable and influenced by virtually all factors relevant for an individual (Bastian 1984). Particularly important HR modulators are physical activity, season and time of the day (Moen 1978, Dressen et al. 1990, Boyd et al. 1999, Arnold et al. 2006, Nilsson et al. 2006), as well as ambient temperature and thermoregulation (Bartholomew et al. 1962, Müller 1982, Arnold et al. 2006). Generally, physical activity received the most attention in the context of HR, because of its clear connection to energy expenditure (Wieser 1986). Consequently, most studies in birds focused on flight, which is among the energetically most demanding behaviours. In greater white-fronted geese (Anser albifrons), for example, HR has been shown to rise from 112 beats per minute (bpm) during rest to a maximum of 446 bpm during flight (Ely et al. 1999). Our data from 25 greylags implanted with HR transmitters (for a detailed description of

Individual performance in complex social systems

137

technology and procedures see Wascher et al. 2009) revealed that in greylags the mean HR at rest is around 85 bpm, while it rises to around 400 bpm before and during take-off (Wascher et al. 2008a). Particularly striking are data from bystanding geese not moving themselves, but watching either social or non-social events: HR increases in these observers were greater when they watched social rather than non-social events (mean maximal heart rates 170 bpm versus 146 bpm, Wascher et al. 2008b) and HR varied with the intensity and duration of an agonistic interaction they watched, as well as with the identity and rank of the opponents involved in the interactions (Wascher et al. 2009). Based on these data, we conclude that geese may have some understanding of third-party relationships. We also recorded the long-term patterns of HR and body temperature. We found circannual and diurnal variation, with lows during early winter and late night (yearlong, midnight to morning) and highs in spring and summer and during the day (Kotrschal et al. unpublished data). This reflects the seasonal modulation of metabolic rate adapted to ecological conditions and social tasks (Kotrschal et al. 1998, Hirschenhauser et al. 2000). On top of this, HR was 13% higher in the females than in the males and pair-bonded individuals showed lower HR during potentially stressful situations than singletons (Wascher et al. 2008b). From these long-term data we concluded that roughly 30% of the total individual energy may be spent in the social domain and that inter-individual variation in that respect is substantial (Kotrschal et al. unpublished). For example, during the 28day incubation period we found great individual differences in mean HR of five females, ranging from 60 bpm to 180 bpm. This indicates a three-fold difference in metabolic rate and energy expense between these females, even in an evident ‘standard’ situation, such as incubation (Scheiber et al. 2006). 5.5.7 Social support among clan members Social interactions are known to be among the most potent stressors (DeVries et al. 2003) and greylag geese are no exception (Wascher et al. 2008b). But this is counterbalanced via the support social allies may provide. The greylag geese social system is characterised by the integration of social units, such as pairs, families and female-centred clans into a flock, which includes agonistic interactions between and social support within these social units. Although the advantages of group living (Krause and Ruxton 2002) certainly outweigh the costs of social stress, we have demonstrated that social life does not come cheap. Individuals in a flock differ with respect to their personalities and related to that, in their physiology

138

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

and basic metabolism (Koolhaas et al. 2007). Still, for coping effectively with the challenges of sociality they all need long-term allies. With a competent and matching partner they will be able to budget their energy efficiently and, thereby, not only reproduce successfully, but also reap benefits from social support through their offspring. Hence, we propose that social support per se is a core factor in stabilising long-term monogamous pairbonds, family cohesion and female-centred clans in greylags.

5.6 Relating to other complex social systems: more questions than answers Social organisation, social structure and mating system have been defined as distinct components of a social system (Kappeler and van Schaik 2002). In fact, basic social organisation, monogamous pairs, female-centred clans, male dominance systems, multi-male groups, etc. may differ greatly between species. Some of the causation of the basic structuring of social systems is reasonably well understood. For instance, the multi-male groups in chimpanzees (Pan troglodytes), and to some extent, pack structure in wolves (Canis lupus), seem to be related to the often deadly skirmishes with neighbouring groups at the territorial borders. The lion harem and infanticide system (as well as mammalian infanticide systems in general) results from the constraints imposed by the male reproductive interests (Packer et al. 1990). And the prevalence of monogamy in birds is certainly linked to the greater opportunities for males to invest in offspring directly. However, the causation of other social structuring remains enigmatic, for instance the female dominance system in most lemurs (Kappeler and Kraus this volume) or the rigid female hierarchies in macaques and baboons (Cheney and Seyfarth 2008). The greylag example shows that passive social support affects the energetic efficiency of social life, making it a prime candidate for structuring this and other vertebrate social systems. Because the ability to profit from social support depends on personality and attachment style (Kotrschal et al. 2009), not all individuals will be affected by social support in the same way. We still lack data from other species, but social support may affect individuals in diverse social systems and examples exist in species from all vertebrate taxa. Lionesses, for example, may mutually support each other for stress reduction and also may support their males to ensure long residency, thereby decreasing the likelihood of infanticide over their lifetime (Packer et al. 1990). As another example, grooming networks in baboons become more exclusive in times of social instability when individuals evi-

Individual performance in complex social systems

139

dently focus on partners, which are able social supporters (Cheney and Seyfarth 2008). Thus, social stress reduction may be a hidden mechanism in many complex vertebrate social systems. Moreover, stress reduction via social support and activation of the oxytocin system dampens the HPAglucocorticoid cascade in a number of ways (DeVries et al. 2003), thereby decreasing anxiety and enhancing individual sociability. By this means, social support positively affects individual sociability. In the past, mainly the direct links between social performance and fitness have been in the focus of quantitative behavioural biology. In contrast, the fitness relevance of feedbacks and dependencies between social behaviour, structure and organisation, as well as the underlying physiological dynamics is known, at best, at a qualitative level. Knowledge within the research levels, proximate, ultimate, ontogeny and evolutionary history (Tinbergen 1963) is advancing rapidly, but linking the levels is necessary for the systemic understanding of social evolution. This remains still a widely open field.

References Abbott DH, Keverne EB, Bercovitch FB, Shively CA, Mendoza SP, Saltzman W, Snowdon CT, Ziegler TE, Banjevic M, Garland T Jr, Sapolsky RM (2003) Are subordinates always stressed? A comparative analysis of rank differences in cortisol levels among primates. Horm Behav 43:67-82 Adkins-Regan E (2005) Hormones and Animal Social Behavior: Monographs in Behavior and Ecology. Princeton University Press, Princeton Alexander GM, Sherwin BB (1993) Sex steroids, sexual behavior, and selection attention for erotic stimuli in women using oral contraceptives. Psychoneuroendocrinology 18:91-102 Angelier FH, Moe B, Clement-Chastel C, Bech C, Chastel O (2007) Corticosterone levels in relation to change of mate in black-legged kittiwakes. Condor 109:668-674 Arnold W, Ruf T, Kuntz R (2006) Seasonal adjustment of energy budget in a large wild mammal, the Przewalski horse (Equus ferus przewalskii). II. Energy expenditure. J Exp Biol 209:4566-4573 Auersperg AM, Gajdon GK, Huber L (2009) Kea (Nestor notabilis) consider spatial relationships between objects in the support problem. Biol Lett 5:455-458 Aureli F, de Waal FBM (2000) Natural Conflict Resolution. University of California Press, Berkeley Aureli F, Cords M, van Schaik CP (2002) Conflict resolution following aggression in gregarious animals: a predicitive framework. Anim Behav 64:325-343 Avise JC, Alisauskas RT, Nelson WS, Ankney CD (1992) Matriarchal population genetic structure in an avian species with female natal philopatry. Evolution 46:1084-1096

140

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Barnard CJ, Sibly RM (1981) Producers and scroungers: a general model and its application to captive flocks of house sparrows. Anim Behav 29:543-550 Barrett L, Henzi SP (2005) The social nature of primate cognition. Proc R Soc Lond B 272:1865-1875 Barrett L, Henzi SP, Rendall D (2007) Social brains, simple minds? Does social complexity really require congnitive complexity? Philos Trans R Soc Lond B 362:561-575 Bartholomew GA, Hudson JW, Howell TR (1962) Body temperature, oxygen consumption, evaporative water loss, and heart rate in poor-will. Condor 64:117-127 Bastian H-V (1984) Die Änderung der Herzfrequenz als Maß der Erregung – eine Literaturübersicht. Vogelwarte 32:226-233 Beauchamp G, Fernández-Juricic E (2004) Is there a relationship between forebrain size and group size in birds? Evol Ecol Res 6:833-842 Bertin A, Richard-Yris A-M, Houdelier C, Lumineau S, Möstl E, Kuchar A, Hirschenhauser K, Kotrschal K (2008) Habituation to humans affects yolk steroid levels and offspring phenotype in quail. Horm Behav 54:396-402 Bertin A, Hirschenhauser K, Kotrschal K (2009) Trans-generational influence of human disturbances in Japanese quail: egg quality influences male social and sexual behavior. Ethology 115:879-887 Black JM (1996) Partnerships in Birds: The Study of Monogamy. Oxford University Press, Oxford Black JM, Choudhury S, Owen M (1996) Do barnacle geese benefit from lifelong monogamy? In: Black JM (ed) Partnerships in Birds: The Study of Monogamy. Oxford University Press, Oxford, pp 91-117 Black JM, Prop J, Larsson K (2007) Wild Goose Dilemmas: Population Consequences of Individual Decisions in Barnacle Geese. Branta Press, Groningen Bluhm CK, Gowaty PA (2004) Social constraints on female mate preferences in mallards, Anas platyrhynchos, decrease offspring viability and mother productivity. Anim Behav 68:977-983 Bond AB, Kamil AC, Balda RP (2003) Social complexity and transitive inference in corvids. Anim Behav 65:479-487 Boonstra R, Lane JE, Boutin S, Bradley A, Desantis L, Newman AEM, Soma KK (2008) Plasma DHEA levels in wild, territorial red squirrels: seasonal variation and effects of ACTH. Gen Comp Endocrinol 158:61-67 Bowman LA, Dilley SR, Keverne EB (1978) Suppression of oestrogen induced LH surges by social subordination in talapoin monkeys. Nature 275:56-58 Boyd IL, Bevan RM, Woakes AJ, Butler PJ (1999) Heart rate and behaviour of fur seals: implications for measurement of field energetics. Am J Physiol 276:H844-H857 Brown JL (1987) Helping and Communal Breeding in Birds. Princeton University Press, Princeton Bugnyar T (2007) An integrative approach to the study of ‘theory-of-mind’-like abilities in ravens. Jpn J Anim Psychol 57:15-27

Individual performance in complex social systems

141

Byrne RW, Whiten A (1988) Machiavellian Intelligence: Social Expertise, and the Evolution of Intellect in Monkeys, Apes, and Humans. Clarendon Press, Oxford Champagne FA, Curley JP (2005) How social experiences influence the brain. Curr Opin Neurobiol 15:704-709 Cheney DL, Seyfarth RM (2008) Baboon Metaphysics: The Evolution of a Social Mind. University of Chicago Press, Chicago Chesser RK (1991) Gene diversity and female philopatry. Genetics 127:437-447 Choudhury S, Black JM, Owen M (1996) Body size, reproductive success and compatibility in barnacle geese. Ibis 138:700-709 Clayton NS, Dickinson A (1998) Episodic-like memory during cache recovery by scrub jays. Nature 395:272-274 Clayton NS, Griffiths DP, Emery NJ, Dickinson A (2001) Elements of episodiclike memory in animals. Philos Trans R Soc Lond B 356:1483-1491 Coltman DW, Pilkington JG, Pemberton JM (2003) Fine-scale genetic structure in a free-living ungulate population. Mol Ecol 12:733-742 Curley JP, Keverne EB (2005) Genes, brains and mammalian social bonds. Trends Ecol Evol 20:561-567 Daisley NJ, Bromundt V, Möstl E, Kotrschal K (2005) Enhanced yolk testosterone influences behavioral phenotype independent of sex in Japanese quail chicks Coturnix japonica. Horm Behav 47:185-194 Dally J, Emery NJ, Clayton NS (2006) Food-caching scrub jays keep track of who was watching when. Science 312:1662-1665 de Ruiter JR, Geffen E (1998) Relatedness of matrilines, dispersing males and social groups in long-tailed macaques (Macaca fascicularis). Proc R Soc Lond B 265:79-87 DeVries AC, Glasper ER, Detillion CE (2003) Social modulation of stress responses. Physiol Behav 79:399-407 de Waal FBM (1982) Chimpanzee Politics: Power and Sex among Apes. Harper and Row, New York de Waal FBM, Tyack PL (2003) Animal Social Complexity: Intelligence, Culture and Individualized Societies. Harvard University Press, Cambridge/MA Divac I, Thibault J, Skageberg G, Palacios JM, Dietl MM (1994) Dopaminergic innervation of the brain in pigeons. The presumed ‘prefrontal cortex’. Acta Neurobiol Exp 54:227-234 Dressen W, Grün H, Hendrichs H (1990) Radio telemetry of heart rate in male tammat wallabies (Marsupialia: Macropodidae): temporal variations and behavioural correlates. Austr J Zool 38:89-103 Dunbar RIM (1995) Neocortex size and group size in primates: a test of the hypothesis. J Hum Evol 28:287-296 Dunbar RIM, Bever J (1998) Neocortex size predicts group size in carnivores and some insectivores. Ethology 104:695-708 Eadie JM, Kehoe FP, Nudds TD (1988) Pre-hatch and post-hatch brood amalgamation in North American Anatidae. A review of hypotheses. Can J Zool 66:1709-1721

142

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Ely CR, Ward DH, Bollinger KS (1999) Behavioral correlates of heart rates of free-living greater white-fronted geese. Condor 101:390-395 Emery NJ, Clayton NS (2004a) The mentality of crows: convergent evolution of intelligence in corvids and apes. Science 306:1903-1907 Emery NJ, Clayton NS (2004b) Comparing the complex cognition of birds and primates. In: Rogers LJ, Kaplan GT (eds) Comparative Vertebrate Cognition: Are Primates Superior to Non-Primates? Springer, Berlin, pp 3-55 Emery NJ, Seed AM, von Bayern AMP, Clayton NS (2007) Cognitive adaptations of social bonding in birds. Philos Trans R Soc Lond B 362:489-505 Fischer H (1965) Das Triumphgeschrei der Graugans (Anser anser). Z Tierpsychol 22:247-304 Folstad I, Karter AJ (1992) Parasites, bright males, and the immunocompetence handicap. Am Nat 139:603-622 Frigerio D, Möstl E, Kotrschal K (2001a) Excreted metabolites of gonadal steroid hormones and corticosterone in greylag geese (Anser anser) from hatching to fledging. Gen Comp Endocrinol 124:246-255 Frigerio D, Weiß BM, Kotrschal K (2001b) Spatial proximity among adult siblings in greylag geese (Anser anser): evidence for female bonding? Acta Ethol 3:121-125 Fritz J, Kotrschal K (2002) On avian imitation: cognitive and ethological perspectives. In: Nehaniv CL, Dauterhahn K (eds) Imitation in Animals and Artifacts. MIT Press, Boston, pp 133-155 Gill SA, Alfson ED, Hau M (2007) Context matters: female aggression and testosterone in a year-round territorial neotropical songbird (Thryothorus leucotis). Proc R Soc Lond B 274:2187-2194 Goodson JL (2005) The vertebrate social behavior network: evolutionary themes and variations. Horm Behav 48:11-22 Goodson JL, Schrock SE, Klatt JD, Kabelik D, Kingsbury MA (2009) Mesotocin and nonapeptide receptors promote estrildid flocking behavior. Science 325:862-866 Goymann W, Wingfield JC (2004) Allostatic load, social status and stress hormones: the costs of social status matter. Anim Behav 67:591-602 Greenwood PJ (1980) Mating systems, philopatry and dispersal in birds and mammals. Anim Behav 28:1140-1162 Groothuis TGG, Carere C (2005) Avian personalities: characterization and epigenesis. Neurosci Biobehav Rev 29:137-150 Groothuis TGG, Schwabl H (2008) Hormone-mediated maternal effects in birds: mechanisms matter but what do we know of them? Philos Trans R Soc Lond B 363:1647-1661 Groothuis TGG, Müller W, von Engelhardt N, Carere C, Eising C (2005) Maternal hormones as a tool to adjust offspring phenotype in avian species. Neurosci Biobehav Rev 29:329-352 Güntürkün O (2005) The avian prefrontal cortex. Curr Opin Neurobiol 15:686-693 Hamilton WD (1964) The genetical evolution of social behaviour, I and II. J Theor Biol 7:1-52

Individual performance in complex social systems

143

Healy SD, Rowe C (2007) A critique of comparative studies of brain size. Proc R Soc Lond B 274:453-464 Hemetsberger J (2001) Die Entwicklung der Grünauer Graugansschar seit 1973. [The demographic development of Konrad Lorenz’ flock of greylag geese in Grünau/Austria since 1973.] In: Kotrschal K, Müller G, Winkler H (eds) Konrad Lorenz und seine verhaltensbiologischen Konzepte aus heutiger Sicht. Filander, Fürth, pp 249-260 Hirschenhauser K, Möstl E, Kotrschal K (1999a) Within-pair testosterone covariation and reproductive output in greylag geese Anser anser. Ibis 141:577586 Hirschenhauser K, Möstl E, Kotrschal K (1999b) Seasonal patterns of sex steroids determined from feces in different social categories of greylag geese (Anser anser). Gen Comp Endocrinol 114:67-79 Hirschenhauser K, Möstl E, Wallner B, Dittami J, Kotrschal K (2000) Endocrine and behavioural responses of male greylag geese (Anser anser) to pairbond challenges during the reproductive season. Ethology 106:63-77 Hirschenhauser K, Weiß BM, Haberl W, Möstl E, Kotrschal K (2010) Female androgen patterns and within-pair testosterone compatibility in domestic geese (Anser domesticus). Gen Comp Endocrinol 165:195-203 Höglund J, Alatalo RV, Lundberg A, Rintamäki PT, Lindell J (1999) Microsatellite markers reveal the potential for kin selection on black grouse leks. Proc R Soc Lond B 266:813-816 Huber L (2002) Clever birds: keas learn through observation. Interpret Bird Bull 3:57-59 Huber L (2007) Emulation learning: the integration of technical and social cognition. In: Nehaniv CL, Dautenhahn K (eds) Imitation and Social Learning in Robots, Humans and Animals: Behavioural, Social and Communicative Dimensions. Cambridge University Press, Cambridge, pp 427-439 Huber L, Rechberger S, Taborsky M (2001) Social learning affects object exploration and manipulation in keas, Nestor notabilis. Anim Behav 62:945-954 Huber L, Gajdon GK, Federspiel I, Werdenich D (2008) Cooperation in keas: social and cognitive factors. In: Itakura S, Fujita K (eds) Origins of the Social Mind: Evolutionary and Developmental Views. Springer, Tokyo, pp 99-119 Hudec K, Rooth J (1970) Die Graugans. Westarp-Wissenschaft, Magdeburg Iwaniuk AN, Nelson JE (2003) Developmental differences are correlated with relative brain size in birds: a comparative analysis. Can J Zool 81:1913-1928 Jarvis ED, Güntürkün O, Bruce L, Csillag A, Karten H, Kuenzel W, Medina L, Paxinos G, Perkel DJ, Shimizu T, Striedter G, Wild JM, Ball GF, Dugas-Ford J, Durand SE, Hough GE, Husband S, Kubikova L, Lee DW, Mello CV, Powers A, Siang C, Smulders TV, Wada K, White SA, Yamamoto K, Yu J, Reiner A, Butler AB (2005) Avian brains and a new understanding of vertebrate brain evolution. Nat Rev Neurosci 6:151-159 Judge PG, Mullen SH (2005) Quadratic postconflict affiliation among bystanders in a hamadryas baboon group. Anim Behav 69:1345-1355

144

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Kalmbach E, van der Aa P, Komdeur J (2005) Adoption as a gosling strategy to obtain better parental care? Experimental evidence for gosling choice and agedependency of adoption in greylag geese. Behaviour 142:1515-1533 Kappeler PM, van Schaik CP (2002) Evolution of primate social systems. Int J Primatol 23:707-740 Kear J (1990) Man and Wildfowl. Poyser, London Koolhaas JM, Korte SM, de Boer SF, van der Vegt BJ, van Reenen CG, Hopster H, de Jong IC, Ruis MA, Blokhuis HJ (1999) Coping styles in animals: current status in behavior and stress-physiology. Neurosci Biobehav Rev 23:925935 Koolhaas JM, de Boer SF, Buwalda B, van Reenen K (2007) Individual variation in coping with stress: a multidimensional approach of ultimate and proximate mechanisms. Brain Behav Evol 70:218-226 Kotrschal K, Hemetsberger J (1995) Social constraints in feeding and individual specialization for transportable food in greylag geese (Anser anser). Ökol Vögel 17:157-163 Kotrschal K, Hemetsberger J, Dittami J (1992) Vigilance in a flock of semi-tame greylag geese (Anser anser) in response to approaching eagles (Haliaeetus albicilla and Aquila chrysaetos). Wildfowl 43:215-219 Kotrschal K, Hirschenhauser K, Möstl E (1998) The relationship between social stress and dominance is seasonal in greylag geese. Anim Behav 55:171-176 Kotrschal K, Hemetsberger J, Weiß BM (2006) Making the best of a bad situation: homosociality in male greylag geese. In: Sommer V, Vasey PL (eds) Homosexual Behaviour in Animals: An Evolutionary Perspective. Cambridge University Press, Cambridge, pp 45-76 Kotrschal K, Schlögl C, Bugnyar T (2007) Lektionen von Rabenvögeln und Gänsen: Instinkt oder Verstand? Biol Unserer Zeit 37:366-374 Kotrschal K, Schöberl I, Bauer B, Thibeaut A-M, Wedl M (2009) Dyadic relationships and operational performance of male and female owners and their male dogs. Behav Proc 81:383-391 Kralj-Fiser S, Scheiber IBR, Blejec A, Möstl E, Kotrschal K (2007) Individualities in a flock of free-roaming greylag geese: behavioral and physiological consistency over time and across situations. Horm Behav 51:239-248 Krause J, Ruxton GD (2002) Living in Groups. Oxford University Press, Oxford Kruk MR, Halász J, Meelis W, Haller J (2004) Fast positive feedback between the adrenocortical stress response and a brain mechanism involved in aggressive behavior. Behav Neurosci 118:1062-1070 Lamprecht J (1986) Structure and causation of the dominance hierarchy in a flock of bar-headed Geese (Anser indicus). Behaviour 96:28-48 Lamprecht J (1987) Female reproductive strategies in barheaded geese (Anser indicus): why are geese monogamous? Behav Ecol Sociobiol 21:297-305 Lazarus J (1978) Vigilance, flock size and domain of danger size in the whitefronted goose. Wildfowl 29:135-145 Lazarus J, Inglis IR (1978) The breeding behaviour of the pink-footed goose: parental care and vigilant behaviour during the fledging period. Behaviour 65:62-88

Individual performance in complex social systems

145

Lee KA, Bahr JM (1994) Utilization of substrates for testosterone and estradiol-17 beta production by small follicles of the chicken ovary. Domest Anim Endocrinol 11:307-314 Longcope C (1986) Adrenal and gonadal androgen secretion in normal females. Clin Endocrinol Metab 15:213-228 Lorenz K (1988) Hier bin ich – wo bist Du? Ethologie der Graugans. Piper, München McEwen BS, Sapolsky RM (1995) Stress and cognitive function. Curr Opin Neurobiol 5:205-216 McEwen BS, Wingfield JC (2003) The concept of allostasis in biology and biomedicine. Horm Behav 43:2-15 McGregor PK (2005) Animal Communication Networks. Cambridge University Press, Cambridge McGregor PK, Dabelsteen T, Holland J (1997) Eavesdropping in a territorial songbird communication network: preliminary results. Bioacoustics 8:253254 McKinnon L, Gilchrist HG, Scribner KT (2006) Genetic evidence for kin-based female social structure in common eiders (Somateria mollissima). Behav Ecol 17:614-621 McNabb AFM, Scanes CG, Zeman M (1998) The endocrine system. In: Starck JM, Ricklefs RE (eds) Avian Growth and Development: Evolution within the Altricial-Precocial Spectrum. Oxford University Press, London, pp 174-202 Mendl M, Zanella AJ, Broom DM (1992) Physiological and reproductive correlates of behavioural strategies in female domestic pigs. Anim Behav 44:11071121 Moen AN (1978) Seasonal changes in heart rates, activity, metabolism and forage intake of white-tailed deer. J Wildl Manage 42:715-738 Müller EF (1982) Heart rate in slow loris (Nycticebus coucang). Daily variation, effects of ambient temperature and role of oxygen transport in blood. Folia Primatol 38:250-258 Nephew BC, Kahn SA, Romero LM (2003) Heart rate and behavior are regulated independently of corticosterone following diverse acute stressors. Gen Comp Endocrinol 133:173-180 Nilsson A, Ahman B, Norberg H, Redbo I, Eloranta E, Olsson K (2006) Activity and heart rate in semi-domesticated reindeer during adaptation to emergence feeding. Physiol Behav 88:116-123 Oliveira RF, McGregor PK, Latruffe C (1998) Know thine enemy: fighting fish gather information from observing conspecific interactions. Proc R Soc Lond B 265:1045-1049 Oliveira RF, Silva A, Canário AVM (2009) Why do winners keep winning? Androgen mediation of winner but not loser effects in cichlid fish. Proc R Soc Lond B 276:2249-2256 Packer C, Scheel D, Pusey AE (1990) Why lions form groups: food is not enough. Am Nat 136:1-19 Panksepp J (1998) Affective Neuroscience: The Foundations of Human and Animal Emotions. Oxford University Press, Oxford

146

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

Panksepp J (2005) Affective consciousness: core emotional feelings in animals and humans. Conscious Cogn 14:30-80 Peake TM, Terry AMR, McGregor PK, Dabelsteen T (2001) Male great tits eavesdrop on simulated male-to-male vocal interactions. Proc R Soc Lond B 268:1183-1187 Peake TM, Terry AMR, McGregor PK, Dabelsteen T (2002) Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proc R Soc Lond B 269:1925-1929 Pepperberg IM (1999) The Alex Studies: Cognitive and Communicative Abilities of Grey Parrots. Harvard University Press, Cambridge/MA Petrie M, Krupa A, Burke T (1999) Peacocks lek with relatives even in the absence of social environmental cues. Nature 401:155-157 Pfeffer K, Fritz J, Kotrschal K (2002) Hormonal correlates of being an innovative greylag goose, Anser anser. Anim Behav 63:687-695 Pirtney SB, MacColl ADC, Lambin X, Moss R, Dallas JF (1999) Spatial distribution of genetic relatedness in a moorland population of red grouse (Lagopus lagopus scoticus). Biol J Linn Soc 68:317-331 Prior H, Schwarz A, Güntürkün O (2008) Mirror-induced behavior in the magpie (Pica pica): evidence of self-recognition. PLoS Biol 6:e202, doi:10.1371/ journal.pbio.0060202 Raby CR, Alexis DM, Dickinson A, Clayton NS (2007) Planning for the future by western scrub-jays. Nature 445:919-921 Radesäter T (1974) On the ontogeny of orienting movements in the triumph ceremony of two species of geese (Anser anser L. and Branta canadensis L.). Behaviour 50:1-15 Regehr HM, Smith CM, Arquilla B, Cooke F (2001) Post-fledging broods of migratory Harlequin ducks accompany females to wintering areas. Condor 103:408-412 Rodríguez-Maldonado E, Velázquez PN, Juárez-Oropeza MA, Pedernera E (1996) Steroid metabolism in theca externa cells from preovulatory follicles of domestic hen (Gallus domesticus). Gen Comp Endocrinol 101:173-179 Roemer I, Reik W, Dean W, Klose J (1997) Epigenetic inheritance in the mouse. Curr Biol 7:277-80 Rutschke E (1982) Stability and dynamics in the social structure of the greylag goose (Anser anser L.). Aquila 89:39-55 Rutschke E (1997) Wildgänse. Lebensweise, Schutz, Nutzung. Parey, Berlin Saino N, Ferrari R, Romano M, Martinelli R, Møller AP (2003) Experimental manipulation of egg carotenoids affects immunity of barn swallow nestlings. Proc R Soc Lond B 270:2485-2489 Sapolsky RM (1992) Neuroendocrinology of the stress response. In: Becker S, Breedlove SM, Crews D (eds) Behavioral Endocrinology. MIT Press, Cambridge/MA, pp 287-324 Sapolsky RM (1993) Endocrinology alfresco: psychoendocrine studies of wild baboons. Recent Prog Horm Res 48:437-468

Individual performance in complex social systems

147

Sapolsky RM (2002) Endocrinology of the stress response. In: Becker JB, Breedlove SM, Crews D, McCathy MM (eds) Behavioral Endocrinology, 2nd edn. MIT Press, Cambridge/MA, pp 409-450 Sapolsky RM, Romero LM, Munck AU (2000) How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr Rev 21:55-89 Scheiber IBR, Weiß BM (submitted) ‘Girls united’: spatial proximity among adult female lineal kin in greylag geese (Anser anser). Scheiber IBR, Weiß BM, Frigerio D, Kotrschal K (2005a) Active and passive social support in families of greylag geese (Anser anser). Behaviour 142:15351557 Scheiber IBR, Kralj S, Kotrschal K (2005b) Sampling effort/frequency necessary to infer individual acute stress responses from fecal analysis in greylag geese (Anser anser). Ann NY Acad Sci 1046:154-167 Scheiber IBR, Wascher CAF, Kotrschal K (2006) Biparental care in greylag geese: heart rate as a measure of parental contribution. J Ornithol 147:245-246 Scheiber IBR, Weiß BM, Hirschenhauser K, Wascher CAF, Nedelcu IT, Kotrschal K (2008) Does ‘relationship intelligence’ make big brains in birds? Open Biol J 1:6-8 Scheiber IBR, Kotrschal K, Weiß BM (2009a) Benefits of family reunions: social support in secondary greylag goose families. Horm Behav 55:133-138 Scheiber IBR, Kotrschal K, Weiß BM (2009b) Serial agonistic attacks by greylag goose families (Anser anser) against the same opponent. Anim Behav 77:1211-1216 Schloegl C, Dierks A, Gajdon GK, Huber L, Kotrschal K, Bugnyar T (2009) What you see is what you get? Exclusion performances in ravens and keas. PLoS ONE 4:e6368, doi:10.1371/journal.pone.0006368 Schwabl H, Mock DW, Gieg JA (1997) A hormonal mechanism for parental favouritism. Nature 386:231 Shorey L, Pirtney SB, Stone J, Höglund J (2000) Fine-scale genetic structuring on Manacus manacus leks. Nature 408:352-353 Sih A, Bell AM, Johnson JC, Ziemba RE (2004) Behavioral syndromes: an integrative overview. Q Rev Biol 79:241-277 Silk JB, Alberts SC, Altmann J (2003) Social bonds of female baboons enhance infant survival. Science 302:1231-1234 Silk JB, Beehner JC, Bergman, Crockford C, Engh AL, Moscovice LR, Wittig RM, Seyfarth RM, Cheney DL (2009) The benefits of social capital: close social bonds among female baboons enhance offspring survival. Proc R Soc Lond B 276:3099-3104 Summers CH (2002) Social interactions over time, implications for stress responsiveness. Integr Comp Biol 42:591-599 Swoboda R, Kotrschal K, Scheiber IBR (submitted) Ontogenetic changes of agonistic behaviour and physiology in greylag goose families. Tinbergen N (1963) On aims and methods of ethology. Z Tierpsychol 20:410-433 van der Jeugd HP, van der Veen IT, Larsson K (2002) Kin clustering in barnacle geese: familiarity or phenotype matching? Behav Ecol 13:786-790

148

Kurt Kotrschal, Isabella B.R. Scheiber and Katharina Hirschenhauser

von Holst D (1998) The concept of stress and its relevance for animal behavior. Adv Stud Behav 27:1-131 Waldeck P, Andersson M, Kilpi M, Öst M (2008) Spatial relatedness and brood parasitism in a female-philopatric bird population. Behav Ecol 19:67-73 Wascher CAF, Arnold W, Kotrschal K (2008a) Heart rate modulation by social contexts in greylag geese (Anser anser). J Comp Psychol 122:100-107 Wascher CAF, Scheiber IBR, Kotrschal K (2008b) Heart rate modulation in bystanding geese watching social and non-social events. Proc R Soc Lond B 275:1653-1659 Wascher CAF, Scheiber IBR, Weiß BM, Kotrschal K (2009) Heart rate responses to agonistic interactions in greylag geese, Anser anser. Anim Behav 77:955961 Waser PM, Jones WT (1983) Natal philopatry among solitary mammals. Q Rev Biol 58:355-390 Weigmann C, Lamprecht J (1991) Intraspecific nest parasitism in bar-headed geese, Anser indicus. Anim Behav 41:677-688 Weiß BM, Kotrschal K (2004) Effects of passive social support inn juvenile greylag geese (Anser anser): a study from fledging to adulthood. Ethology 110:429-444 Weiß BM, Kotrschal K, Frigerio D, Hemetsberger J, Scheiber IBR (2008) Birds of a feather stay together: extended family bonds and social support in greylag geese (Anser anser). In: Columbus F (ed) Family Relations: Behavioral, Psychological and Sociological Aspects. Nove Science Publishers, New York, pp 69-88 Weiß BM, Kehmeier S, Mikolasch S, Schlögl C (2009) Transitive inference in greylag geese. Primate Rep (special issue, february):38 Weiß BM, Kotrschal K, Möstl E, Hirschenhauser K (2010) Social and life history correlates of hormonal partner compatibility in greylag geese (Anser anser). Behav Ecol 21:138-143 Whiten A, Byrne RW (1997) Machiavellian Intelligence. II. Extensions and Evaluations. Cambridge University Press, Cambridge Wieser W (1986) Bioenergetik. Energietransformation bei Organismen. Thieme Verlag, Stuttgart Wilson EO (1975) Sociobiology: The New Synthesis. Harvard University Press, Cambridge/MA Wingfield JC, Hegner RF, Dufty AM Jr, Ball GF (1990) The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. Am Nat 136:829-846 Wingfield JC, Lynn SE, Soma KK (2001) Avoiding the ‘costs’ of testosterone: ecological bases of hormone-behavior interactions. Brain Behav Evol 57:239251 Zahavi A, Zahavi A (1997) The Handicap Principle: A Missing Piece of Darwin’s Puzzle. Oxford University Press, New York

Part II Conflict and cooperation

Chapter 6

Conflict and conflict resolution in social insects JÜRGEN HEINZE

ABSTRACT The major transitions in evolution are characterised by cooperation and division of labour among biological entities. Such transitions have led to the evolution of complex genomes from independent oligonucleotides, of eukaryotic cells from independent prokaryotes and proto-eukaryotes, of multicellular organisms from multiple cells, and of complex animal societies from individuals. For higher-level entities to function as a unit and to compete with other such higher-level entities it is fundamental that withingroup conflict among its constituents is resolved. Indeed, selfishness by lower-level entities, e.g., genes, organelles, and cells, appears to be prevented or at least minimized by higher-level mechanisms. Insect societies have often been regarded as harmonious superorganisms, in which all individuals cooperate to increase the reproductive output of the society as a whole. Recent research on social insects has revealed the occurrence of similar conflict of interest among group members, however. For example, individual workers may have selfish interests concerning the origin of male offspring, which go against the interests of other workers and the whole society. Theoretical and empirical studies suggest that many insect societies are structured by a fine-tuned balance between the interests of individuals and the group, and that this balance is maintained by a complex system of behavioural mechanisms, including mutual surveillance, policing, and punishment. The aims of this chapter are to review the various conflicts of interest that may occur in social insects, in particular ants, as well as the behavioural and chemical mechanisms by which conflict is resolved. Furthermore, I will summarise our recent understanding of insect societies as governed by kin selected, ‘enforced altruism’.

152

Jürgen Heinze

Table 6.1 Major types of conflict in the societies of ants, bees, and wasps.

1

Antagonists

Subject of conflict

Predicted outcome

Observed outcome

Queens ↔ workers

Sex allocation ratio

Under worker control: higher female-bias at low queen numbers and queen mating frequencies; under queen control: equal investment in both sexes

Variable: evidence for pure queen, pure worker and mixed control in different species

Queens ↔ workers

Origin of males

From relatedness alone: worker reproduction in monogynous, monandrous societies; queen reproduction in polygynous, polyandrous societies

Usually queen monopoly

Worker ↔ worker

Origin of males in queenless colonies

Workers strive for producing own sons

In small colonies: formation of hierarchies, top-ranking workers lay eggs

Queen ↔ queen

Partitioning of reproduction

Reproductive skew varies with relatedness, probability of successful dispersal and other factors

High skew in a minority of specialists inhabiting patchy habitats; variation in relatedness relatively unimportant

Queen ↔ queen

Production of sexuals under lethal local mate competition

Queens aim at increasing their own fitness by producing males earlier and in larger numbers

Queens in polygynous colonies of Cardiocondyla sp. produce males earlier than queens in monogynous colonies1; sex ratio less female-biased in polygynous colonies

Larvae ↔ adults

Caste differentiation

More female larvae strive to develop into sexuals than is optimal for colony as a whole

Workers control caste fate and kill surplus queens in stingless bees

Different genetic lineages

Production of sexuals

Workers prefer sexuals of their own genetic lineage

Evidence for nepotistic queen rearing only in Formica ants2

Male ↔ male

Access to female sexuals

Yamauchi et al. (2006), 2Hannonen and Sundström (2003)

Male fighting in a few ant genera with local mating

Conflict and conflict resolution in social insects

153

6.1 Conflict and cooperation in insect societies Ants, bees, and wasps have long fascinated mankind, and their societies have been idealized since Biblical times as models of virtue and perfect social organisation, where all leadership belongs to females, where everyone abides by the law, and where all cooperate harmoniously for the common good of the society (e.g., Drouin 2005). However, the conception of life in insect societies has shifted over time, following general trends and fashions of science, and our contemporary understanding of what is going on inside a bee hive or an ant mound is quite different from the image of noble communitarianism commonly praised in the 19th and early 20th century (Sleigh 2007). The acceptance of insect societies as supra-individual biological units (‘superorganisms’) has waxed and waned over the last century, and, although colony-level phenomena have never lost their attraction for researchers, the last few decades have been characterised by an increasing interest in conflict among selfish individuals and how these disruptive forces are counterbalanced by mutual manipulation and policing: the superorganism as a police state (Whitfield 2002; Table 6.1). The amount of theoretical and empirical data on potential and actual conflict within insect societies makes it difficult to summarise all that is known on a few pages. In this chapter I can therefore not provide a complete summary of conflict and conflict resolution in social insects; instead, focusing on ants, I will summarise our current understanding of a few important types of conflict involving queens and workers. Other types of conflict, e.g., potential nepotism in the rearing of female sexuals in societies with multiple queens, conflict concerning caste determination between female larvae and adult workers, and conflict involving males, have been reviewed elsewhere (e.g., Heinze 2004, Ratnieks et al. 2006).

6.2 Levels of selection and conflict resolution Cooperation is a powerful driving force in evolution and has repeatedly facilitated the formation of new levels of complexity in the hierarchy of biological organisation. Many biological units have evolved through the grouping together of simpler units, be it in the formation of chromosomes through the assembly of independently replicating oligonucleotides, the evolution of the eukaryotic cell through endosymbiotic associations, and the origin of multicellularity through cooperation among individual cells that failed to separate after division. In these ‘major transitions in evolution’ (Maynard Smith and Szathmáry 1995), lower-level units often be-

154

Jürgen Heinze

come specialised for complementary tasks and forego their autonomous reproduction. This may result in the emergence of new properties and the creation of new levels of selection. Rather than competing for resources to reproduce autonomously, the constituents of higher-order units may form a community of interest, which interacts as a single individual with the environment and competes with other such higher-order units (Michod 1996). The benefits from cooperation do not preclude the existence of selfishness in lower-level units, which may strive to achieve short-term benefits from overexploiting the common pool of resources. This results in the well known ‘tragedy of the commons’, as in the long run selfishness may damage the integrity of the higher-level unit and, hence, also the selfish lowerlevel units themselves (Hardin 1968, Rankin et al. 2007). For example, mobile, repetitive nucleotide sequences, such as transposons and retrotransposons, may replicate independently and increase the number of sequence copies at the cost of the efficiency of the entire genome; defective, but fast replicating mitochondria, such as the ‘petite’ mutant of yeast, may accumulate and reduce the fitness of the whole cell; and selfish genetic cell lineages, such as tumours and cancers, have well-known detrimental effects for the whole organism (e.g., Burt and Trivers 2006). The evolution of individuality of higher level units requires that between-unit selection overrides within-unit selection and that within-unit conflict must be suppressed. Cellular and genetic mechanisms, such as DNA methylation, RNA interference, unicellular inheritance of organelles, single cell stages in early development, and tumor suppression genes, have been interpreted as defense mechanisms, through which higher-level entities punish or prevent harmful selfishness by genes, organelles, and cells (Maynard Smith and Szathmàry 1995, Michod 1996, Leigh 1999). For example, the mechanisms of a fair meiosis usually guarantee that homologue alleles have an equal chance of being transmitted into a gamete, which to some extent prevents selfishness. Similarly, uniparental inheritance helps to avoid costly conflict among genetically distinct lineages of organelles within the host (Burt and Trivers 2006). Social insects (bees, wasps, ants, and termites) are the classic examples for yet another major transition – from solitary animals to ‘eusociality’ (defined as a life history with cooperative brood care, overlapping generations, and reproductive division of labour: Wilson 1971). In insect societies, queens (and, in termites, kings) specialise in reproduction, whereas workers engage in brood care, foraging, nest defense, and nest maintenance. Reproductive division of labour allows the evolution of striking group-level adaptations, such as fungus cultivation, aphid farming, optimised trails free of traffic congestion, and perfectly tempered bee hives. This specialisation is certainly one of the causes of the unrivaled ecologi-

Conflict and conflict resolution in social insects

155

cal and evolutionary success of social insects (Wilson 1971, Seeley 1985, Moritz and Southwick 1992, Hölldobler and Wilson 1990, 2009). Reproductive division of labour requires close collaboration among nestmates, and any disharmony may threaten the success of the group as a whole. Nevertheless, the societies of social Hymenoptera are not less conflictridden than other higher-level units. Instead, their organisation reflects a fine-tuned balance between the interests of individual group members and the society as a whole. Due to efficient mechanisms of conflict prevention and resolution, potential conflict rarely breaks open and actual conflict with violent antennation, biting, or stinging can be observed only under special conditions (Ratnieks 1988, Heinze et al. 1994, Bourke 1999, Heinze 2004, Ratnieks et al. 2006). Some of these conflicts resemble those well known from other group-living animals and concern the partitioning of reproductive rights. They arise, for example, when multiple potential egg layers co-occur in a society and fight for reproductive dominance. In addition, the unusual genetic family structure of social Hymenoptera resulting from haplodiploid sex determination (males develop from unfertilised eggs and are haploid, females develop from fertilised eggs and are diploid) introduces additional types of conflict, which are absent from diploid animals: queen-worker conflict about sex allocation and queen-worker conflict about the origin of males.

6.3 Kin selection explains both conflict and cooperation in social animals Most researchers agree that inclusive fitness theory (Hamilton 1964a) provides the best explanation for the evolution of eusociality. Through foregoing reproduction and helping relatives to produce and raise extra offspring, workers indirectly increase their inclusive fitness, because the offspring of relatives carry the helper’s genes in copies that are identical by descent. If the total offspring produced exclusively through the helper’s efforts carry more copies of the helper’s genes than would have been present in all of the helper’s own young, helping can be favoured. Hamilton (1964a) formalised this model in his famous inequality rB > C, where B represents the benefits of helping (the extra offspring the beneficiary of the help produces only because of the assistance provided by the helper) and C the costs of helping (i.e., the helper’s offspring not produced because of helping). r is the coefficient of relatedness, i.e., the probability of the alleles at a randomly chosen locus in helper and recipient being identical by descent from a common ancestor. The inclusive fitness of a helper then consists of its di-

156

Jürgen Heinze

rect fitness, i.e., its personal reproductive success excluding all effects of its social environment, plus the indirect fitness it gains through the surplus offspring produced by the beneficiary of the help, weighted by the coefficient of relatedness between helper and beneficiary. This looks like a simple summation, but measuring inclusive fitness in practice can be rather tricky (e.g., Queller 1996, West et al. 2007a). Kinship (r) is of critical importance in inclusive fitness theory, which is therefore often referred to as ‘kin selection theory’. It must not be forgotten, however, that inclusive fitness theory ascribes equally prominent roles to ecological and behavioural costs and benefits. Kinship among two individuals can be quantified relatively easily from pedigrees or with population genetic methods by determining the genetic similarity between two individuals relative to that between two randomly drawn individuals from a carefully chosen reference population (e.g., Queller 1994). In contrast, B and C are notoriously difficult to determine. Several authors have therefore negated the usefulness of inclusive fitness theory and advanced alternative approaches (Wilson 2005, 2008, Wilson and Hölldobler 2005, Wilson and Wilson 2007). One of these alternatives is multilevel selection theory, in which selection is partitioned in within-group and between-group components. Kin and multilevel selection are mathematically congruent and essentially two sides of the same coin (e.g., Queller 1992a,b, Korb and Heinze 2004, Foster et al. 2006, Lehmann et al. 2007, West et al. 2007b, Hölldobler and Wilson 2009), as evident from Price’s covariance method for the hierarchical analysis of natural selection (Price 1970). The ongoing discussion about how the evolution of altruism and eusociality is explained best therefore boils down to the semantic issue about which term is more appropriate. The recurrent outbreak of debate about the adequacy of inclusive fitness theory is probably kindled in part by longstanding misinterpretations of the role of haplodiploidy in the evolution of eusociality (Dawkins 1979, Foster et al. 2006). Several independent evolutionary origins of eusociality exist within the Hymenoptera, an insect order characterised by haplodiploid sex determination. Haplodiploidy creates special degrees of relatedness (Table 6.2), which were thought to facilitate the evolution of eusociality: females are more closely related to their sisters than to their own offspring and therefore appear to achieve higher fitness gains from helping their mothers to produce additional sisters than by producing own daughters (Hamilton 1964b). This simple conclusion quickly turned out to be wrong, because the disproportionally low relatedness of females to their brothers counterbalances the disproportionately high fitness gains arising from the close relatedness among sisters (e.g., Bourke and Franks 1995, Crozier and Pamilo 1996). In any case, the recent rejection of inclusive fitness theory because

Conflict and conflict resolution in social insects

157

Table 6.2 Life-for-life relatedness (regression relatedness multiplied by the sexspecific reproductive values of individuals, i.e., 1 for females, 0.5 for males) of a Hymenopteran female to its relatives.

Life-for-life relatedness of a Hymenopteran female to its Mother

0.5

Daughter

0.5

Sister

0.75

Half-sister

0.25

Son

0.5

Brother

0.25

Nephew

0.375

Half-nephew

0.125

not all haplodiploids are eusocial and not all eusocial species are haplodiploid is wrong for two reasons. First, it does not necessarily follow from the observation that a certain trait promotes the evolution of eusociality that all organisms exhibiting this trait are inevitably eusocial. Second, and more importantly, kin selection does not rely on haplodiploidy but works whenever the genotypes of helper and beneficiary are statistically associated, so that Hamilton’s rule is satisfied. Male haploidy might promote the evolution of eusociality in other ways, e.g., through facilitating the spread of mutations causing maternal care and reproductive altruism (Linksvayer and Wade 2005, Teyssèdre et al. 2006). Nevertheless, the multiple evolutions of sterile workers in the Hymenoptera presumably relies also on those ecological and behavioural traits that are believed to underlie the evolution of eusociality in termites and molerats: the commonness of brood care and stable nest sites in solitary ancestors (e.g., Bourke and Franks 1995, Korb and Heinze 2008). Furthermore, the sting of aculeate Hymenoptera is a potent weapon for brood defense and may additionally favour group living in that helping becomes easier (e.g., Starr 1985). Finally, the unusual life-long monogamy, characteristic of all social insects, might have been a fundamental pre-adaptation in the evolution of altruistic worker castes, because it stabilises the relatedness of helpers to the brood and cost-benefit ratios (Boomsma 2007).

158

Jürgen Heinze

6.4 Haplodiploidy and queen-worker conflict about sex allocation Inclusive fitness theory predicts that haplodiploidy creates a conflict between queens and workers about how much resources are invested in female vs. male sexuals (sex allocation ratio; e.g., Trivers and Hare 1976, Mehdiabadi et al. 2003). Because of the relatedness asymmetries resulting from haplodiploidy, workers in a colony with a single, singly-mated queen (monogyny, monandry) can maximize their inclusive fitness by allocating three times more resources to female sexuals than to males. The ‘fitness return’ workers receive from both sexes, given by a worker’s life-for-life relatedness (see glossary) to a sexual, multiplied by the relative mating success of its sex, will then be equal end evolutionarily stable (e.g., Bourke and Franks 1995, Crozier and Pamilo 1996). In contrast, queens are equally related to their male and female offspring (Table 6.2) and gain most from a balanced sex ratio. The resulting conflict about sex allocation may be less pronounced when colonies contain multiple, related queens (polygyny) or a multiply mated queen (polyandry). For example, when a queen mates multiply, some of the female sexuals will be sisters of a given worker and others will be its halfsisters. While the relatedness of the worker to its brothers remains unchanged, its average relatedness to female sexuals will decrease with increasing queen mating frequency and approach the relatedness to brothers. Queen-worker conflict about sex allocation ratios therefore may vanish at high mating frequencies (Table 6.2). When queen number and mating frequency vary among colonies of a population, workers may be selected to rear either only male or female sexuals, depending on the genetic structure of their own colony relative to the average colony structure in the population, which leads to ‘split sex ratios’ (Boomsma and Grafen 1990, Ratnieks and Boomsma 1997). For example, workers will benefit from the exclusive production of males, if their colony has an aboveaverage number of queens. Note that this does not require that workers have information on the average colony structure in the population. Instead, selection will optimise the response of workers to the structure of their own colony over time. Data summarised in the seminal paper by Trivers and Hare (1976) and many subsequent studies in the heydays of sex ratio research suggested worker control of sex ratios, particularly in ants, and thus provided powerful evidence for the importance of inclusive fitness and kin selection in social evolution. Workers have been shown to manipulate the primary sex ratio produced by the queen by selective culling of male brood (Sundström

Conflict and conflict resolution in social insects

159

et al. 1996) or overfeeding of female larvae so that a larger proportion grows into female sexuals (Hammond et al. 2002). During the last years, however, the accumulation of new data has considerably blurred this previously clear picture. Overall, variation in relatedness asymmetry and queen number has been found to explain only 21% of the variance in sex allocation ratios (Meunier et al. 2008). Worker control therefore appears to be far from complete and a variety of factors other than relatedness may introduce considerable noise in sex allocation (Kümmerli and Keller 2009). These include ecological parameters and physiological or behavioural tactics of the queens. For example, female sexuals are more costly and colonies endowed with fewer resources might therefore concentrate on the production of the cheaper males (Nonacs 1986, Deslippe and Savolainen 1995); queens may force workers to rear male-biased broods simply by limiting the number of fertilised eggs they lay (Passera et al. 2001); ongoing tugs-of-war between queens and workers may result in sex ratios intermediate between the optima of both castes (Reuter and Keller 2001); and sex ratios may be affected by queen age (Schwander et al. 2008). In addition, dispersal and colony founding tactics (e.g., solitary founding vs. colony fragmentation, local resource competition, Bourke and Franks 1995) and local mate competition (Hamilton 1967) strongly influence the production of sexuals. Local resource competition is common, wherever female sexuals show limited dispersal and compete for care by workers, food, or nest sites at or around the maternal nest (e.g., Pearcy and Aron 2006). When brothers compete locally for mating opportunities, mother queens in monogynous societies are predicted to produce just the number of males that guarantees that all their daughters are inseminated, resulting in a highly female-biased sex ratio. If multiple queens contribute to the brood, sex ratios are expected to be more balanced, as each queen will be selected to increase its chances of propagating copies of its genes also via its sons. Local mating in the nest and local mate competition, culminating in lethal combat among wingless males for access to female sexuals, have recently been documented in a few taxa of social Hymenoptera (BOX 6.1). Despite all this theoretical and empirical work, one fundamental aspect of sex ratio optimisation has attracted surprisingly little attention. The adaptive evolution of sex allocation ratios requires genetic variation in this trait. Because of the long lifespan of many social insects it is often difficult to obtain data on lifetime reproductive success. The heritability of sex ratios has therefore rarely been studied. Nevertheless, cross-breeding studies in the ant Cardiocondyla kagutsuchi indicate that heritable variation occurs (Frohschammer and Heinze 2009), which, though needed for adaptive evo-

160

Jürgen Heinze

BOX 6.1 Lethal male fighting in Cardiocondyla The genus Cardiocondyla is characterised by wingless males, which engage in lethal fighting for the monopolisation of matings with the female sexuals produced within a colony. According to local mate competition theory, mother queens in monogynous nests are predicted to produce only the number of males needed to inseminate all daughters. In contrast, in polygynous nests, queens are selected to raise the percentage of sons among their offspring to increase the chance that one of their own sons survives competition for mating. The number of fighter males among the first 20 sexuals produced was indeed significantly lower in single-queen than two-queen colonies of Cardiocondyla obscurior (mean 5.3 vs. 9.8). Colonies also produced a higher proportion of fighter males when eggs from alien polygynous colonies were added to focal monogynous colonies, but not when eggs from alien monogynous colonies were added. This outcome suggests that workers do not manipulate sex ratios, but that queens react to the presence of other queens by increasing the proportion of unfertilised eggs (Cremer and Heinze 2002). This interpretation could be confirmed through direct analysis of the ploidy of freshly laid eggs by fluorescence in situ hybridization (de Menten et al. 2005).

Fig. 6.1 Two wingless males of the ant Cardiocondyla obscurior locked in lethal combat about access to female sexuals (photo © Sylvia Cremer)

Conflict and conflict resolution in social insects

161

lution, might momentarily constrain the power of queens and workers to bias sex allocation.

6.5 Conflict about the origin of males The second fundamental conflict arising from haplodiploidy concerns the origin of males. Workers of most species of social Hymenoptera cannot mate and lack a spermatheca for the storage of sperm (e.g., Bourke and Franks 1995). Nevertheless, they usually have ovaries and can in principle produce haploid males from unfertilised eggs. Due to haplodiploidy, workers are more closely related to their own sons than to their brothers. In addition, in a monogynous, monandrous society they are even more closely related to the sons of other workers, i.e., their nephews, than to their brothers (Table 6.2). On the one hand, selection acting on the workers therefore favours the production of males by workers. On the other hand, the queen, being more closely related to her own sons than to her grandsons, is selected to prevent worker reproduction. As with the conflict about sex allocation, multiple mating and the presence of multiple queens change the average relatedness of workers to the male offspring of other workers and the queens in such a way that workers may oppose worker egg laying. For example, if a queen uses equal amounts of sperm from two males to fertilise her eggs (effective queen mating frequency me = 2), the work force will consist of sisters and halfsisters. The relatedness of workers to average worker-produced males will then be 1/2 (0.125 + 0.375) = 0.25 (Table 6.2). With increasing effective mating frequencies relatedness decreases towards 0.125 and workers will finally be more closely related to their brothers than to other worker-produced males. Queens might be capable of suppressing worker reproduction by aggression and eating their eggs if colonies are very small (e.g., Franks and Scovell 1983, Nakata and Tsuji 1996, Wenseleers et al. 2005), but physical queen control will break down in societies with hundreds or thousands of workers. In such colonies, the predictions from relatedness alone seem clear: males are expected to be the workers’ offspring in monogynous, monandrous societies and, if worker policing occurs, queen offspring in polygynous or polyandrous societies. Indeed, the proportion of males produced by workers in the Saxon wasp, Dolichovespula saxonica, was associated with queen mating frequency (Foster and Ratnieks 2000), but whether the prediction holds more generally for social insects is a contentious matter (Hammond and Keller 2004, Heinze 2004, 2008, Wenseleers and Ratnieks 2006a).

162

Jürgen Heinze

Maternity analyses using genetic markers, such as microsatellites, revealed that in most cases males are (almost) exclusively offspring of the queen, regardless of queen number and mating frequency. Significant worker reproduction appears to occur only in a minority of taxa and here again independent of genetic colony structure (e.g., Helanterä and Sundström 2007). As in the case of sex allocation, the exact values of relatedness are therefore probably less important in the regulation of who produces males than expected from theory. Instead, the rarity of worker reproduction might indicate that selfish egg laying is associated with high costs (Ratnieks 1988, Cole 1986). Indeed, egg laying workers do not forage or defend the nest, engage less in brood care, and, because each is most closely related to its own offspring, might waste energy in prolonged dominance interactions for egg laying rights with their nestmates (Gobin et al. 2003, Hartmann et al. 2003; see Sect. 6.6). Furthermore, workers, in particular in species with a pronounced caste polymorphism, have less ovarioles and a smaller abdomen than queens and can therefore only lay a few eggs. On the whole, worker egg laying thus reduces the total output of the colony in several ways, thereby also reducing the average inclusive fitness of workers. This reduction will be even more pronounced when workers are less closely related to an average worker-produced male than that to an average queen-produced male, i.e., in polygynous and / or polyandrous societies, but it appears to be large enough to prevent worker reproduction even in most monogynous and monandrous species (Hammond and Keller 2004, Heinze 2004). But still: each individual worker gains most from laying its own eggs, even though this might decrease the inclusive fitness of everybody else. The interest of the summed individuals therefore conflicts with the sum of the interests of the individuals – a clear entomological exemplification of Rousseau’s contrast between volonté générale and volonté de tous. What keeps individual workers from being selfish? This tragedy of the commons is resolved through ‘worker policing’, a phenomenon first observed in honey bees, but since then reported from a large number of wasps and ants. When a honey bee worker lays a haploid egg into an empty cell on the comb of a queenright colony, other workers quickly destroy and eat the egg (Ratnieks and Visscher 1989) and also attack the egg layer (Visscher and Dukas 1995). As worker egg laying is usually a rare event, and policing is therefore difficult to study in natural colonies, researchers have split insect colonies into two halves, one with a queen and the other without, and re-united the fragments after a worker had become reproductive in the queenless part. After reunification, workers that have become reproductive are fiercely attacked, expelled from the nest and even killed by their nestmates, both from the queenless and the queenright colony parts (e.g.,

Conflict and conflict resolution in social insects

163

Monnin and Peeters 1997, Kikuta and Tsuji 1999, Gobin et al. 1999, Liebig et al. 1999, Monnin and Ratnieks 2001, Hartmann et al. 2003). Similarly, the introduction of worker-laid, but not of queen-laid eggs results in egg eating (Foster et al. 2002, D’Ettorre et al. 2004). It appears that the threat of being policed in the presence of a queen is often enough to prevent workers from laying eggs. This ‘self-restraint’ probably explains why worker egg laying is rare despite of the lack of constant aggression among workers (Ratnieks 1988) The recognition of reproductives relies on particular odour cues, presumably hydrocarbons on the workers’ cuticle, which are believed to signal fertility. Non-reproductive workers and reproductive queens differ significantly in their hydrocarbon bouquets. Whereas such ‘queen pheromones’ were long regarded as inhibitive substances, through which queens suppress worker reproduction, they are now commonly understood as honest signals, through which queens reliably signal their fertility to the workers (Keller and Nonacs 1993). The odour of workers, which have recently begun to lay eggs in a policing experiment, first resembles the scent of non-reproductives but approaches that of the queen with increasing fecundity (Monnin 2006, Heinze and D’Ettorre 2009). Workers appear to police new reproductives as long as they are not yet fully fertile and differ in scent from the old reproductive, but the odour of new reproductives, which have laid eggs for several weeks, may be indistinguishable from that of an old reproductive. After reunification, such established new reproductives are often not policed by the workers, but instead old and new reproductives may fight and besmear each other with substances, which elicit aggression from other workers (e.g., Heinze et al. 1998, Monnin et al. 2002, Hartmann et al. 2005). Chemical marking of rivals has been referred to as ‘punishment’ and is considered to be different from policing, because it permanently prevents selfish behaviour – the selfish individual may be attacked and killed. Policing, e.g., through egg eating, usually prevents selfishness only temporarily (Monnin and Ratnieks 2001). The concepts of policing and punishment widely overlap with a third type of aggression among group members: dominance (Monnin and Ratnieks 2001, Frank 2003; see Sect. 6.6). Through policing, individuals repress competition among group members, increase group cohesion, and achieve indirect fitness gains. In contrast, the winners of dominance interactions increase their own potential direct fitness at the cost of other group members (Monnin and Ratnieks 2001). Recent observations show that aggression may occasionally serve both purposes. Workers are not equally likely to engage in policing (Frank 1996, van Zweden et al. 2007). Rather, the policing elite consist of those workers that are most likely to inherit the top rank in the hierarchy when the reproductive is removed. Policing there-

164

Jürgen Heinze

BOX 6.2 Dominance and policing in the ant Temnothorax unifasciatus Societies of the ant Temnothorax unifasciatus usually contain only a few dozen workers and a single, singly-mated queen. When colonies are split into two fragments, workers in the queenless part form dominance hierarchies by aggressive antennation and biting, and one or a few workers begin to lay eggs. When the two fragments are later reunited, reproductive workers are attacked both by individuals from the former queenless and the former queenright parts (worker policing). Aggression leads to the expulsion of reproductive workers or at least causes them to stop laying eggs. This consequence can directly be shown by providing workers in the queenless fragment with food stained with fat soluble dyes, which are incorporated into their eggs. By removing the queen from the reunited colony, it was shown that those workers, which were most active in policing reproductive workers after reunification, tended to become dominant and started laying eggs themselves. Policing therefore has a selfish component (Stroeymeyt et al. 2007). Becoming dominant is associated with a change in the bouquet of cuticular hydrocarbons. Queens and fertile workers are characterised by significantly higher abundances of heptacosane and nonacosane than nonlaying workers, but whether these linear alkanes serve as fertility signals needs to be shown (Brunner et al. 2009).

Fig. 6.2 Dominance interaction between two Temnothorax unifasciatus workers. Both workers are individually marked with loops of thin copper wire (photo © Bartosz Walter).

Conflict and conflict resolution in social insects

165

fore has a selfish component: by eliminating egg laying rivals, police workers at the same time fight for dominance and increase their present indirect fitness and their future direct fitness (Saigo and Tsuchida 2004, Wenseleers et al. 2005, Stroeymeyt et al. 2007; BOX 6.2). Worker policing is thus an adaptation on the level of the individual, rather than a group adaptation (Gardner and Grafen 2009).

6.6 Fighting for dominance: workers vs. workers, queens vs. queens In colonies without mated reproductives, workers may produce male or, in a few species with thelytokous parthenogenesis, female offspring from unfertilised eggs. Usually only a minority of the workers becomes reproductive while the other workers continue to engage in non-reproductive tasks. Division of reproductive labour appears to be associated with age, e.g., in the thelytokous Cape honey bee and the thelytokous ant Pristomyrmex punctatus (Crewe 1988), and is maintained by pheromones that presumably signal fecundity (Monnin 2006, Heinze and D’Ettorre 2009). Workers of many species with smaller colonies engage in dominance interactions after their colonies have lost their reproductives or when the worker hierarchy is disturbed, as in the policing experiments described above. Aggression reveals the existence of linear or near-linear hierarchies, in which the highest ranking workers later begin to lay eggs, whereas the low-ranking ones continue with the daily duties in the nest (e.g., Trunzer et al. 1999). In species with a limited caste dimorphism, such as queenless ants (see below) and wasps, dominance relations among workers may be clearly visible even in the presence of an egg layer (e.g., Ito and Higashi 1991, Deshpande et al. 2006). In contrast, the aggressive establishment and maintenance of rank orders in queenright colonies of highly eusocial insects might be selected against, because they are costly (Gobin et al. 2003) and disrupt the division of labour within the colony (Cole 1986). Nevertheless, careful observations allow predicting which worker will become dominant after orphaning. This suggests that rank relationships based on more subtle interactions already exist in the presence of the queen. For example, future dominants are often characterised by a higher rate of interaction with the queen (E. Brunner and J. Heinze, unpubl. data). Similar dominance interactions commonly occur also among young queens, which jointly initiate a new colony. In paper wasps and several species of ants, in which founding queens search for food, rank orders may be associated with a division of labour between the dominants, which lay

166

Jürgen Heinze

eggs and stay close to the brood pile, and the subordinates, which forage for food (Pardi 1946, Gadagkar 1980, Kolmer and Heinze 2000). In ant species, in which founding queens do not forage but rely on stored fat reserves and histolyzed body tissue to produce their first young, the initially peaceful coexistence and mutual help among queens are quickly replaced by severe fighting once the first workers have eclosed. The latter often take part in expelling or killing all but one of the queens (e.g., Strassmann 1989). In established multi-queen colonies, queens usually behave amicably or ignore each others’ presence, and rank orders among queens have been observed only in a small minority of species. Theory predicts that hierarchies and high reproductive skew, e.g., an unbalanced partitioning of reproduction among potential egg layers, may be evolutionarily stable when the success of dispersal and becoming reproductive away from the established nest is low and when relatedness among nestmates is high (e.g., Keller and Reeve 1994). The boreal ants of the genus Leptothorax, in which one dominant queen monopolises egg laying, are indeed specialists of patchy habitats, such as dry areas in spruce bogs. Low skew species, in which queens do not form hierarchies and contribute equally to the colony’s offspring, live in more homogeneous, extended habitats, such as the coniferous forests, which cover large parts of the Northern Hemisphere (Bourke and Heinze 1994). Dominance hierarchies are probably best known from social Hymenoptera with a secondarily reduced or lost caste polymorphism. For example, in several genera of the ant subfamily Ponerinae workers have retained a spermatheca and are in principle capable of mating and laying fertilised eggs. In Diacamma, Dinoponera, Streblognathus and other genera, mated workers (‘gamergates’) have completely replaced morphological queens (Peeters 1991). In these taxa, reproductive conflict is particularly pronounced, because high ranking workers can supersede the top-ranking gamergate, mate and then produce both daughters and sons (e.g., Ito and Higashi 1991, Peeters et al. 1992, Monnin and Peeters 1999, Monnin and Ratnieks 2001). Similar to these conflicts about reproductive rights among totipotent individuals is the potential conflict among developing females and the rest of the colony about which and how many larvae will develop into sexuals (Bourke and Ratnieks 1999). Hymenopteran larvae often lack the power to control their own development, because food is provided by the workers. Nevertheless, conflict about caste fate appears to be realised in stingless bees. Here, female larvae are in charge of their own food intake, because they develop in identical, mass-provisioned, sealed cells, and caste size dimorphism is low. A large percentage of female larvae develop into female sexuals, but many of them are later executed by the workers

Conflict and conflict resolution in social insects

167

(Wenseleers et al. 2004a). Interestingly, the proportion of larvae developing into excess female sexuals is smaller in those species in which males are produced by workers, presumably because of the higher relatedness of workers to worker-produced males (Wenseleers and Ratnieks 2004).

6.7 Proximate aspects of dominance and policing Dominance, punishment, and policing may involve overtly aggressive behaviour, such as biting, pulling of legs and antennae, immobilisation, and stinging, but in many cases ritualised aggression appears to be sufficient to clarify rank relationships. Antennation bouts, often referred to as antennal boxing, have been described from ants, wasps, and bees, and are probably the most common type of agonistic interaction. In contrast to the rather slow, inspective antennation commonly seen in insect societies, antennal boxing involves rapid strikes directed mostly towards the head of the opponent. When rank relations are clear, subordinates may respond to antennation by crouching with retracted antennae and legs and the dominant may step onto the subordinate with extended legs and antennae, assuming a ‘big game hunter’s position’ on top of it for several seconds. Antennal boxing closely resembles food begging and subordination is occasionally associated with the offering of food. Appeasing regurgitation of food might thus have provided the behavioural basis for the evolution of food exchange in social Hymenoptera (Liebig et al. 1997). When rank relations are not yet settled, antennation may elicit counterantennation and result in mutual antennal boxing, which may intensify into overt aggression, from elegant fencing with elongated, beak-like mandibles (Peeters and Hölldobler 1995) to ‘sumo wrestling’ with shaking contests (Gobin and Ito 2003). In Leptothorax ants, dominance interactions among workers usually remain ritualised and rarely lead to injuries, whereas dominance interactions among queens more frequently escalate and may result in the expulsion or death of an individual (e.g., Heinze and Smith 1990). Similarly, in Temnothorax ants, policing but not worker dominance may involve biting and stinging and result in the death of the attacked individual (E. Brunner and J. Heinze, unpubl. data). Antennation and biting serve to establish rank orders, but what determines, which individual becomes dominant and which becomes subordinate? Obviously, a positive feedback exists between social and reproductive dominance, in that physical aggression often ceases when one or several individuals have begun to lay eggs and signal their fertility by pheromones. This is nicely documented by an experiment in which only

168

Jürgen Heinze

the lowest ranking queens in colonies of a high-skew ant were allowed to become reproductive after overwintering, while the more dominant queens were kept under prolonged artificial hibernation. Dominants did not attack the now egg laying former subordinate when they were later transferred into the nest and instead were policed themselves by the workers (Ortius and Heinze 1999). An association between ovarian development and dominance is also indicated by the observation that workers in orphaned colonies of the ants Temnothorax nylanderi and T. affinis with ovaries consisting of three or four ovarioles achieved higher social and reproductive ranks than ‘normal’ workers with two ovarioles (Heinze et al. 1997). A promising avenue for untangling the interrelations between fertility and dominance is the surgical removal of ovaries, which allows separating social and reproductive dominance in paper wasps. Ovariectomized queens of paper wasps were still capable of socially dominating their nestmates but no longer prevented them from egg laying (Röseler and Röseler 1989). In addition to reproductive status, the age of contestants may also influence their rank. In perennial social insects, such as ants or many tropical wasps, the youngest individuals become dominant when an older reproductive is removed (Ito and Higashi 1991), whereas in annual, temperate wasp species old individuals dominate the young. This difference was explained by the relationship between individual and society life span. Because of the decline of annual societies in fall, younger individuals gain more from helping and saving the costs of fighting for dominance. In contrast, becoming dominant is particularly beneficial for young individuals in perennial societies because of the long future tenure (Tsuji and Tsuji 2005). Closely connected with this phenomenon is the peculiar inverse relationship between fecundity and longevity in perennial social insects. Whereas most animals show a trade-off between reproduction and life span, laying eggs apparently prolongs the life of perennial social insects (Heinze and Schrempf 2008), i.e., once an individual has become socially and reproductively dominant it will remain so for a long period. The evidence for an association between social status and other physical attributes, such as size or fluctuating asymmetry, is as yet ambiguous (Heinze and Oberstadt 1999, Gobin and Ito 2003, Clémencet et al. 2008), but the linearity of many insect hierarchies suggests that social experience might play a role in the establishment of rank orders. As in other animals, winners keep winning and losers keep losing, i.e., experience from the previous interaction feeds back on the performance in future fights (Dugatkin 1997). In accordance with models on hierarchy formation (Chase 1982), domination in Leptothorax colonies was most frequently followed by the attacker attacking a third individual (double dominance; J.

Conflict and conflict resolution in social insects

169

Heinze, unpubl. data). Winner and loser effects presumably depend on hormone titers, but the association between aggression, reproduction, and hormones are not yet fully elucidated: elevated brain levels of octopamine were associated with social dominance in bumble bees (Bloch et al. 2000) but with egg laying in the queenless ant Streblognathus peetersi (CuvillierHot and Lenoir 2006).

6.8 Conclusion: the insect society – superorganism or police state? This review clearly demonstrated that insect societies are often ridden with potential conflict. Three aspects might merit a few additional thoughts. First, most previous research in social insects has focused on individual types of conflict, but because inclusive fitness is affected by numerous factors, a more comprehensive approach might be needed. For example, how strongly female larvae of stingless bees strive to develop into female sexuals appears to be associated with the extent of male production by workers (Wenseleers and Ratnieks 2004), and conflict about sex allocation is tightly interwoven with conflict about reproductive allocation, i.e., the partitioning of resources into colony maintenance and the production of sexuals (Herbers et al. 2001), and workers that focus on egg laying instead of working have less opportunities to control and manipulate queen sex ratios. As always in evolution, it is impossible to simultaneously optimise all individual fitness components. Instead, trade-offs might lead to results that seem suboptimal when studied in isolation. In addition, phylogenetic inertia might lead to unexpected phenomena. The parthenogenetic ant Platythyrea punctata has retained ‘selfish policing’ from their sexual ancestors despite of the evolutionary futility of selfishness in clonal societies (Hartmann et al. 2003, K. Kellner, E. Brunner and J. Heinze, unpubl. data). The case of clonal Platythyrea also emphasizes a second point: the role of relatedness in the occurrence of conflict in insect societies appears to be considerably less pronounced than previously thought. This is evidenced by the comparatively low percentage of variance in sex allocation explained by variation in kinship (Meunier et al. 2008) and the common absence of worker reproduction from monogynous and monandrous insect societies (Hammond and Keller 2004). Note that this does not reduce the fundamental importance of inclusive fitness and kinship in the evolution and maintenance of eusociality. Instead, research on conflict and conflict resolution in insect societies again and again gives evidence of the robustness of predictions from inclusive fitness theory. Insect societies generally

170

Jürgen Heinze

consist of more or less extended families and are closed against unrelated individuals by efficient nestmate discrimination (e.g., Wilson 1971, Hölldobler and Wilson 1990). Like the single-cell stage in the early development of multicellular animals counteracts the accumulation of selfish,parasitic cell lineages (Maynard Smith and Szathmàry 1995), the single-queen stage in the early life of many social insects prevents intraspecific parasitism and the erosion of relatedness. Nevertheless, in particular in complex insect societies, the costs arising from individual selfishness for the inclusive fitness of average group members societies appear to outweigh the inclusive fitness effects of variation in relatedness. Third, individuals in insect societies are not inherently altruistic but their cooperation is achieved by policing, punishment, and dominance. Selfish behaviour, which threatens the efficiency of the colony and decreases the average inclusive fitness of others, is suppressed, and individuals are forced to cooperate. The more efficient mutual surveillance, the less is to be gained from cheating. The percentage of individuals that selfishly pursue their interests decreases with the likelihood of failure, e.g., due to policing (Wenseleers and Ratnieks 2006b). Likewise, workers in many ant species with large, polygynous colonies, in which old queens are readily replaced by new queens reared from the brood, have completely lost their ovaries, presumably because their chance of laying unfertilised eggs is extremely small (Bourke 1999, Wenseleers et al. 2004b). Once internal conflict has been repressed, group-level adaptations may evolve (Gardner and Grafen 2009). Seen from outside, such societies come closest to the traditional image of the harmonious superorganism, in which ‘individuality’ is achieved through the suppression of competition within the colony and which compete as single units with other such units in territorial disputes.

Acknowledgements Supported by Deutsche Forschungsgemeinschaft (He 1623/17). Peter Kappeler, Rolf Kümmerli and an anonymous referee made helpful comments on the manuscript.

Conflict and conflict resolution in social insects

171

GLOSSARY Gamergate: mated, fertile worker, in particular in ponerine ants Monogyny: one single mated egg layer (queen or gamergate) per colony Monandry: single-mating by female sexuals Polygyny: multiple mated egg layers (queens or gamergates) per colony; if reproduction is highly skewed due to dominance hierarchies among egg layers, apparent polygyny may in fact be ‘functional monogyny’. Polyandry: multiple mating by females; as paternity is often skewed, the pedigree-effective mate number (the reciprocal of the sum of the squared average proportional contributions of each male) reflects the influence of multiple mating on the genetic structure of the colony better than the actual number of mates. Relatedness: the coefficient of relatedness is defined as the proportion of alleles that two individuals share by common descent from a recent common ancestor. As pedigrees are usually not known for social insects, most studies estimate kinship from ‘regression relatedness’, the slope of the regression line obtained by regressing the average genotype (e.g., at a microsatellite locus) in individuals of type X (e.g., female sexuals), across all colonies, on the genotype in individuals of type Y (e.g., workers). Due to haplodiploidy, the genes of males contribute only 1/3 to the gene pool of future generations, whereas females contribute 2/3 (assuming that workers do not reproduce). Regression relatedness therefore needs to be multiplied by the sex-specific reproductive values of individuals (1 for females, 0.5 for males) to obtain the commonly used ‘lifefor-life’ relatedness as given in Table 6.1. In diploids, regression relatedness equals ‘life-for-life’ relatedness. Relatedness asymmetry: relatedness of workers to female sexuals / relatedness of workers to males; usually 3 in monogynous societies, but approaching 1 with increasing queen mating frequency or increasing number of related queens Sex allocation ratio: in social insects usually given as the amount of resources invested in female sexuals relative to that invested in all sexuals. Sex allocation ratio equals the numerical sex ratio when males and female sexuals are similar in size and equally costly to produce, but as female sexuals may be much larger and richer in fat than males (particular in ants), sex allocation ratios are often very different from numerical sex ratios.

172

Jürgen Heinze

References Bloch G, Simon T, Robinson GE, Hefetz A (2000) Brain biogenic amines and reproductive dominance in bumble bees (Bombus terrestris). J Comp Physiol A 186:261-268 Boomsma JJ (2007) Kin selection versus sexual selection: why the ends do not meet. Curr Biol 17:R673-R683 Boomsma JJ, Grafen A (1990) Intraspecific variation in ant sex ratios and the Trivers-Hare hypothesis. Evolution 44:1026-1034 Bourke AFG (1999) Colony size, social complexity and reproductive conflict in social insects. J Evol Biol 12:245-257 Bourke AFG, Franks NR (1995) Social Evolution in Ants. Princeton University Press, Princeton Bourke AFG, Heinze J (1994) The ecology of communal breeding: the case of multiple-queen leptothoracine ants. Philos Trans R Soc Lond B 345:359-372 Bourke AFG, Ratnieks FLW (1999) Kin conflict over caste determination in social Hymenoptera. Behav Ecol Sociobiol 46:287-297 Brunner E, Kroiss J, Heinze J (2009) Chemical correlates of reproduction and worker policing in a myrmicine ant. J Insect Physiol 55:19-26 Burt A, Trivers RL (2006) Genes in Conflict. The Biology of Selfish Genetic Elements. Harvard University Press, Cambridge/MA Chase ID (1982) Behavioral sequences during dominance hierarchy formation in chickens. Science 216:439-440 Clémencet J, Rome Q, Fédérici P, Doums C (2008) Aggressions and size-related fecundity of queenless workers in the ant Cataglyphis cursor. Naturwissenschaften 95:133-139 Cole BJ (1986) The social behavior of Leptothorax allardycei (Hymenoptera, Formicidae): time budgets and the evolution of worker reproduction. Behav Ecol Sociobiol 18:165-173 Cremer S, Heinze J (2002) Adaptive production of fighter males: queens of the ant Cardiocondyla adjust the sex ratio under local mate competition. Proc R Soc Lond B 269:417-422 Crewe RM (1988) Natural history of honey-bee mandibular gland secretions: development of analytical techniques and the emergence of complexity. In: Needham GR, Page RE Jr, Delfinado-Baker M, Bowman CE (eds) Africanized Honey Bees and Bee Mites. Wiley, New York, pp 149-158 Crozier RH, Pamilo P (1996) Evolution of Social Insect Colonies. Sex Allocation and Kin Selection. Oxford University Press, Oxford Cuvillier-Hot V, Lenoir A (2006) Biogenic amine levels, reproduction and social dominance in the queenless ant Streblognathus peetersi. Naturwissenschaften 93:149-153 Dawkins R (1979) Twelve misunderstandings of kin selection. Z Tierpsychol 51:184-200 de Menten L, Cremer S, Heinze J, Aron S (2005) Primary sex ratio adjustment by ant queens in response to local mate competition. Anim Behav 69:1031-1035

Conflict and conflict resolution in social insects

173

Deshpande SA, Sumana A, Surbeck M, Gadagkar R (2006) Wasp who would be queen: a comparative study of two primitively eusocial species. Curr Sci 91:332-336 Deslippe RJ, Savolainen R (1995) Sex investment in a social insect: the proximate role of food. Ecology 76:375-382 D’Ettorre P, Heinze J, Ratnieks FLW (2004) Worker policing by egg eating in the ponerine ant Pachycondyla inversa. Proc R Soc Lond B 271:1427-1434 Drouin J-M (2005) Ants and bees. Between the French and the Darwinian revolution. Ludus Vitalis XIII:3-14 Dugatkin LA (1997) Winner and loser effects and the structure of dominance hierarchies. Behav Ecol 8:583-587 Foster KR, Ratnieks FLW (2000) Facultative worker policing in a social wasp. Nature 407:692-693 Foster KR, Gulliver J, Ratnieks FLW (2002) Worker policing in the European hornet Vespa crabro. Insectes Soc 49:41-44 Foster KR, Wenseleers T, Ratnieks FLW (2006) Kin selection is the key to altruism. Trends Ecol Evol 21:57-60 Frank SA (1996) Policing and group cohesion when resources vary. Anim Behav 52:1163-1169 Frank SA (2003) Perspective: repression of competition and the evolution of cooperation. Evolution 57:693-705 Franks NR, Scovell E (1983) Dominance and reproductive success among slavemaking worker ants. Nature 304:724-725 Frohschammer S, Heinze J (2009) A heritable component in sex ratio and caste determination in a Cardiocondyla ant. Front Zool 6:27 Gadagkar R (1980) Dominance hierarchy and division of labour in the social wasp, Ropalidia marginata (Lep.) (Hymenoptera: Vespidae). Curr Sci 49:772775 Gardner A, Grafen A (2009) Capturing the superorganism: a formal theory of group adaptation. J Evol Biol 22:659-671 Gobin B, Ito F (2003) Sumo wrestling in ants: major workers fight over male production in Acanthomyrmex ferox. Naturwissenschaften 90:318-321 Gobin B, Billen J, Peeters C (1999) Policing behaviour towards virgin egg layers in a polygynous ponerine ant. Anim Behav 58:1117-1122 Gobin B, Heinze J, Strätz M, Roces F (2003) The energetic cost of reproductive conflicts in the ant Pachycondyla obscuricornis. J Insect Physiol 49:747-752 Hamilton WD (1964a) The genetical evolution of social behaviour. I. J Theor Biol 7:1-16 Hamilton WD (1964b) The genetical evolution of social behaviour. II. J Theor Biol 7:17-52 Hamilton WD (1967) Extraordinary sex ratios. Science 156:477-488 Hammond RL, Keller L (2004) Conflict over male parentage in social insects. PLoS Biol 2:e248, doi:10.1371/journal.pbio.0020248 Hammond RL, Bruford MW, Bourke AFG (2002) Ant workers selfishly bias sex ratios by manipulating female development. Proc R Soc Lond B 269:173-178

174

Jürgen Heinze

Hannonen M, Sundström L (2003) Worker nepotism among polygynous ants. Nature 421:910 Hardin G (1968) The tragedy of the commons. Science 162:1243-1248 Hartmann A, Wantia J, Torres JA, Heinze J (2003) Worker policing without genetic conflicts in a clonal ant. Proc Natl Acad Sci USA 100:12836-12840 Hartmann A, D’Ettorre P, Jones GR, Heinze J (2005) Fertility signaling – the proximate mechanism of worker policing in a clonal ant. Naturwissenschaften 92:282-286 Heinze J (2004) Reproductive conflict in insect societies. Adv Stud Behav 34:1-57 Heinze J (2008) The demise of the standard ant (Hymenoptera: Formicidae). Myrmecol News 11:9-20 Heinze J, D’Ettorre P (2009) Honest and dishonest communication in social Hymenoptera. J Exp Biol 212:1775-1779 Heinze J, Oberstadt B (1999) Worker age, size and social status in queenless colonies of the ant Leptothorax gredleri. Anim Behav 58:751-759 Heinze J, Schrempf A (2008) Aging and reproduction in social insects. Gerontology 54:160-167 Heinze J, Smith TA (1990) Dominance and fertility in a functionally monogynous ant. Behav Ecol Sociobiol 27:1-10 Heinze J, Hölldobler B, Peeters C (1994) Conflict and cooperation in ant societies. Naturwissenschaften 81:489-497 Heinze J, Puchinger W, Hölldobler B (1997) Worker reproduction and social hierarchies in Leptothorax ants. Anim Behav 54:849-864 Heinze J, Oberstadt B, Tentschert J, Hölldobler B, Bestmann HJ (1998) Colony specificity of Dufour gland secretions in a functionally monogynous ant. Chemoecology 8:169-174 Helanterä H, Sundström L (2007) Worker reproduction in Formica ants. Am Nat 170:E14-E25 Herbers JM, DeHeer CJ, Foitzik S (2001) Conflict over sex allocation drives conflict over reproductive allocation in perennial social insect colonies. Am Nat 158:178-192 Hölldobler B, Wilson EO (1990) The Ants. Harvard University Press, Cambridge/MA Hölldobler B, Wilson EO (2009) The Superorganism: The Beauty, Elegance, and Strangeness of Insect Societies. Norton, New York Ito F, Higashi S (1991) A linear dominance hierarchy regulating reproduction and polyethism of the queenless ant Pachycondyla sublaevis. Naturwissenschaften 78:80-82 Keller L, Nonacs P (1993) The role of queen pheromones in social insects: queen control or queen signal? Anim Behav 45:787-794 Keller L, Reeve HK (1994) Partitioning of reproduction in animal societies. Trends Ecol Evol 9:98-102 Kikuta N, Tsuji K (1999) Queen and worker policing in the monogynous and monandrous ant, Diacamma sp. Behav Ecol Sociobiol 46:180-189 Kolmer K, Heinze J (2000) Rank orders and division of labour among unrelated cofounding ant queens. Proc R Soc Lond B 267:1729-1734

Conflict and conflict resolution in social insects

175

Korb J, Heinze J (2004) Multilevel selection and social evolution of insect societies. Naturwissenschaften 91:291-304 Korb J, Heinze J (2008) Ecology of Social Evolution. Springer, Heidelberg Kümmerli R, Keller L (2009) Patterns of split sex ratio in ants have multiple evolutionary causes based on different within-colony conflicts. Biol Lett 5:713716 Lehmann L, Keller L, West S, Roze D (2007) Group selection and kin selection: two concepts but one process. Proc Natl Acad Sci USA 104:6736-6739 Leigh EG Jr (1999) Levels of selection, potential conflicts, and their resolution: the role of the ‘common good’. In: Keller L (ed) Levels of Selection in Evolution. Princeton University Press, Princeton, pp 15-30 Liebig J, Heinze J, Hölldobler B (1997) Trophallaxis and aggression in the ponerine ant, Ponera coarctata: implications for the evolution of liquid food exchange in the Hymenoptera. Ethology 103:707-722 Liebig J, Peeters C, Hölldobler B (1999) Worker policing limits the number of reproductives in a ponerine ant. Proc R Soc Lond B 266:1865-1870 Linksvayer TA, Wade MJ (2005) The evolutionary origin and elaboration of sociality in the aculeate Hymenoptera: maternal effects, sib-social effects, and heterochrony. Q Rev Biol 80:317-336 Maynard Smith J, Szathmáry E (1995) The Major Transitions in Evolution. WH Freeman, Oxford Mehdiabadi NJ, Reeve HK, Mueller UG (2003) Queens versus workers: sex-ratio conflict in eusocial Hymenoptera. Trends Ecol Evol 18:88-93 Meunier J, West SA, Chapuisat M (2008) Split sex ratios in the social Hymenoptera: a meta-analysis. Behav Ecol 19:382-390 Michod RE (1996) Cooperation and conflict in the evolution of individuality. II. Conflict mediation. Proc R Soc Lond B 263:813-822 Monnin T (2006) Chemical recognition of reproductive status in social insects. Ann Zool Fennici 43:515-530 Monnin T, Peeters C (1997) Oophagy of subordinate’s eggs in the monogynous queenless ant Dinoponera quadriceps. Naturwissenschaften 84:499-502 Monnin T, Peeters C (1999) Dominance hierarchy and reproductive conflicts among subordinates in a monogynous queenless ant. Behav Ecol 10:323-332 Monnin T, Ratnieks FLW (2001) Policing in queenless ponerine ants. Behav Ecol Sociobiol 50:97-108 Monnin T, Ratnieks FLW, Jones GR, Beard R (2002) Pretender punishment induced by chemical signalling in a queenless ant. Nature 419:61-65 Moritz RFA, Southwick EE (1992) Bees as Superorganisms: An Evolutionary Reality. Springer, Heidelberg Nakata K, Tsuji K (1996) The effect of colony size on conflict over maleproduction between gamergate and dominant workers in the ponerine ant Diacamma sp. Ethol Ecol Evol 8:147-156 Nonacs P (1986) Sex-ratio determination within colonies of ants. Evolution 40:199-204 Ortius D, Heinze J (1999) Fertility signaling in queens of a North American ant. Behav Ecol Sociobiol 45:151-159

176

Jürgen Heinze

Pardi L (1946) Ricerche sui Polistini. VII. La ‘dominazione’ e il ciclo ovarico annuale in Polistes gallicus L. Boll Ist Entomol Univ Bologna 15:25-84 Passera L, Aron S, Vargo EL, Keller L (2001) Queen control of sex ratio in fire ants. Science 293:1308-1310 Pearcy M, Aron S (2006) Local resource competition and sex ratio in the ant Cataglyphis cursor. Behav Ecol 17:569-574 Peeters C (1991) The occurrence of sexual reproduction among ant workers. Biol J Linn Soc 44:141-152 Peeters C, Hölldobler B (1995) Reproductive cooperation between queens and their mated workers: the complex life history of an ant with a valuable nest. Proc Natl Acad Sci USA 92:10977-10979 Peeters C, Billen J, Hölldobler B (1992) Alternative dominance mechanisms regulating monogyny in the queenless ant genus Diacamma. Naturwissenschaften 79:572-573 Price GR (1970) Selection and covariance. Nature 227:520-521 Queller DC (1992a) Quantitative genetics, inclusive fitness, and group selection. Am Nat 139:540-558 Queller DC (1992b) A general model for kin selection. Evolution 46:376-380 Queller DC (1994) Genetic relatedness in viscous populations. Evol Ecol 8:70-73 Queller DC (1996) The measurement and meaning of inclusive fitness. Anim Behav 51:229-232 Rankin DJ, Bargum K, Kokko H (2007) The tragedy of the commons in evolutionary biology. Trends Ecol Evol 22:643-651 Ratnieks FLW (1988) Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. Am Nat 132:217-236 Ratnieks FLW, Boomsma JJ (1997) On the robustness of split sex ratio predictions in social Hymenoptera. J Theor Biol 185:423-439 Ratnieks FLW, Visscher PK (1989) Worker policing in the honeybee. Nature 342:796-797 Ratnieks FLW, Foster KR, Wenseleers T (2006) Conflict resolution in insect societies. Annu Rev Entomol 51:581-608 Reuter M, Keller L (2001) Sex ratio conflict and worker production in eusocial Hymenotera. Am Nat 158:166-177 Röseler P-F, Röseler I (1989) Dominance of ovariectomized foundresses of the paper wasp, Polistes gallicus. Insectes Soc 36:219-234 Saigo T, Tsuchida K (2004) Queen and worker policing in monogynous and monandrous colonies of a primitively eusocial wasp. Proc R Soc Lond B 271:S509-S512 Schwander T, Humbert J-Y, Brent CS, Helms Cahan S, Chapuis L, Renai E, Keller L (2008) Maternal effect on female caste determination in a social insect. Curr Biol 18:265-269 Seeley TD (1985) Honeybee Ecology. Princeton University Press, Princeton Sleigh C (2007) Six Legs Better: A Cultural History of Myrmecology. Johns Hopkins University Press, Baltimore Starr CK (1985) Enabling mechanisms in the origin of sociality in the Hymenoptera – the sting’s the thing. Ann Entomol Soc Am 78:836-840

Conflict and conflict resolution in social insects

177

Strassmann JE (1989) Altruism and relatedness at colony foundation in social insects. Trends Ecol Evol 4:371-374 Stroeymeyt N, Brunner E, Heinze J (2007) ‘Selfish worker policing’ controls reproduction in a Temnothorax ant. Behav Evol Sociobiol 61:1449-1457 Sundström L, Chapuisat M, Keller L (1996) Conditional manipulation of sex ratios by ant workers: a test of kin selection theory. Science 274:993-995 Teyssèdre A, Couvet D, Nunney L (2006) Lower group productivity under kinselected reproductive altruism. Evolution 60:2023-2031 Trivers RL, Hare H (1976) Haplodiploidy and the evolution of the social insects. Science 191:249-263 Trunzer B, Heinze J, Hölldobler B (1999) Social status and reproductive success in queenless ant colonies. Behaviour 136:1093-1105 Tsuji K, Tsuji N (2005) Why is dominance hierarchy age-related in social insects? The relative longevity hypothesis. Behav Ecol Sociobiol 58:517-526 van Zweden JS, Fürst MA, Heinze J, D’Ettorre P (2007) Specialization in policing behaviour among workers in the ant Pachycondyla inversa. Proc R Soc Lond B 274:1421-1428 Visscher KP, Dukas R (1995) Honey bees recognize development of nestmates’ ovaries. Anim Behav 49:542-544 Wenseleers T, Ratnieks FLW (2004) Tragedy of the commons in Melipona bees. Proc R Soc Lond B 271:S310-S312 Wenseleers T, Ratnieks FLW (2006a) Comparative analysis of worker reproduction and policing in eusocial Hymenoptera supports relatedness theory. Am Nat 168:E163-E179 Wenseleers T, Ratnieks FLW (2006b) Enforced altruism in insect societies. Nature 444:50 Wenseleers T, Hart AG, Ratnieks FLW, Quezada-Euán JJG (2004a) Queen execution and caste conflict in the stingless bee Melipona beecheii. Ethology 110:725-736 Wenseleers T, Hart AG, Ratnieks FLW (2004b) When resistance is useless: policing and the evolution of reproductive acquiescence in insect societies. Am Nat 164:E154-E167 Wenseleers T, Tofilski A, Ratnieks FLW (2005) Queen and worker policing in the tree wasp Dolichovespula sylvestris. Behav Ecol Sociobiol 58:80-86 West SA, Griffin AS, Gardner A (2007a) Evolutionary explanations for cooperation. Curr Biol 17:R661-R672 West SA, Griffin AS, Gardner A (2007b) Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. J Evol Biol 20:415-432 Whitfield J (2002) The police state. Nature 416:782-784 Wilson EO (1971) The Insect Societies. Harvard University Press, Cambridge/MA Wilson EO (2005) Kin selection as the key to altruism: its rise and fall. Soc Res 72:159-168 Wilson EO (2008) One giant leap: how insects achieved altruism and colonial life. BioScience 58:17-25 Wilson EO, Hölldobler B (2005) Eusociality: origin and consequences. Proc Natl Acad Sci USA 102:13367-13371

178

Jürgen Heinze

Wilson DS, Wilson EO (2007) Rethinking the theoretical foundation of sociobiology. Q Rev Biol 82:327-348

Chapter 7

Social insects, major evolutionary transitions and multilevel selection JUDITH KORB

ABSTRACT The history of life is characterised by an increase in biological complexity from simple replicators to multicellular organisms. These major evolutionary transitions have in common that independent entities came together and cooperated and that finally a new entity was formed with a new fitness and a single evolutionary fate. Yet, the stable evolution of cooperation poses a classical Darwinian puzzle: Organisms compete over reproduction and selfish individuals that reap the benefits of the cooperation without paying the costs (cheaters) can invade a population of cooperators and drive the disappearance of cooperation. Social insects have become model organisms to study stable cooperation and how conflict between individuals is resolved. Here, I will first summarise what we have learned from social insect research about the evolution of stable cooperation. Besides little-studied ecological factors that determine the benefits and costs of cooperation, two common mechanisms to prevent the spread of cheaters have been identified: (i) common ancestry and aligned evolutionary interests mainly achieved through relatedness and (ii) enforcement mechanisms that make cheating costly. Then, I will show that similar mechanisms have evolved at other levels of the biological hierarchy that favour cooperation. Thirdly, I will present the multilevel selection approach, which promises to be a useful tool to study evolution at multiple selection levels. I will end by showing how a multilevel selection approach in future research might help to quantify benefits and costs of cooperation, so that insect societies and all major evolutionary transitions alike are being recognised as more than the sum of their components.

180

Judith Korb

7.1 Life on earth – a hierarchy of selection levels The history of life is characterised by an increase of biological complexity. While evolution need not intrinsically drive towards increasing complexity on a short time scale, one can nevertheless track a progression in sophistication over geological time. The rise in complexity from simple replicating molecules to multicellular organisms can be characterised as a series of ‘major transitions’ in evolution where replicating units came together to form a new, more complex unit. At least three such major transitions can be identified (Fig. 7.1; Maynard Smith and Szathmáry 1995). At the origin of life, independent replicators evolved into cooperatively replicating chromosomes that formed the first protocells (Szathmáry 2006). Around 1.4 billion years ago, eukaryotes orginated via endosymbioses from two independent prokaryotic ancestors (e.g. Margulis 1981). To date, the last clear major transition was the multiple and independent origins of multi-

CTTGGTACCGACTCGGACCCACTA GTAACGGCCGCCAGTGTGCTGGAA TTCGGCTTTTCTGCTTCGGTAACTA ACGTCATCACAGGGGCATTCCCTC CCCCGCTTATTCTTCAGCAAC

Fig. 7.1 Levels of complexity. This termite colony comprises a hierarchy of levels of cooperation. Genes cooperate in genomes, nuclear genes and cytoplasmic elements cooperate and build cells, cells cooperate and build multicellular organisms, and individuals cooperate and form colonies. At the top-most level, even colonies have their own ‘phenotype’: complex mounds arise through self-organised processes.

Social insects, major evolutionary transitions and multilevel selection

181

cellular life (e.g. Herron et al. 2009). Maynard Smith and Szathmáry (1995) also identified a fourth major transition, the evolution of eusocial groups in which individuals live in complex societies (see below). Each of these events constitutes an increase in structural complexity, as groups of individuals become individuals in their own right. Put another way, there is a transfer of fitness from the individuals making up the group to the group itself, a new entity is formed with a new fitness and a single evolutionary fate (Herron et al. 2009). The evolution of high-level complexity in this way makes considerable sense as it allows organisms to become larger, occupy new niches and perhaps dominate old ones. However, at the same time, the major transitions pose a major problem for evolutionary biology (Maynard Smith and Szathmáry 1995, Herron et al. 2009). 7.1.1 The problem of cooperation A key feature of each of the major transitions is the requirement for cooperation (see Glossary) between independent entities and this represents a classic problem for Darwinian thinking. Why should low-level units that compete among each other over reproduction invest in other individuals, especially when it compromises their own reproduction? At first glance, it appears that such behaviours will result in reduced passage of genes to the next generation and so reduced fitness. A necessary condition then for cooperation to evolve is that the benefits for a cooperative group must be larger than the costs paid by the constituent members (Fig. 7.2). Cooperation must increase the fitness of cooperators over selfish solitary organisms.

Evolutionary Dilemma A

B

Aco + Bco

Aco + Bch

Bch

Fig. 7.2 The evolutionary dilemma of cooperation. For cooperation to evolve each cooperator (indicated by the subscript co) must have a higher fitness (indicated by the size of the circle) than a solitary entity. Yet, this is not sufficient. If a cheater mutant arises, that invests less in the cooperation but still reaps the benefit of cooperation (indicated by the subscript ch), this mutant will be evolutionary favoured as it has a higher fitness than the cooperator. In the long-term, this will lead to the breakdown of cooperation.

182

Judith Korb

Yet, this condition is not sufficient. If a mutant arises among cooperators which invests less in cooperation but can still reap the benefit of cooperation, it can outreproduce cooperators because it gains the benefits of cooperation without paying the cost. Such ‘cheaters’ might then invade the population and drive the disappearance of cooperators and the breakdown of cooperation. Thus, for evolutionarily stable cooperation to evolve, mechanisms must exist which prevent the origin and spread of cheaters (see also Bshary this volume).

7.2 Social insects as model organisms for cooperation One group of organisms is especially suited to study social life, both its conflicts and conflict resolution mechanisms: eusocial insects, which appear themselves to be close to making a new evolutionary transition that transcends the level of the multicellular organism (Maynard Smith and Szathmáry 1995). Eusociality is defined by colony life in which only one or a few individuals reproduce while the large majority of individuals − the workers and soldiers − forego direct reproduction, at least temporarily, and help the others reproduce (Wilson 1971). Worker and soldier behaviour take many and extreme forms, ranging from helping with brood care, through foraging to the famously suicidal defense of the honeybee worker or of termite soldiers that explode upon contact with a predator. Such examples of cooperation − in particular suicidal behaviours − are linked to reduced reproductive fitness on the part of the workers/soldiers, an example of what is called biological altruism (see Glossary; West et al. 2007, Foster 2008). A range of insect species display eusociality, including the Hymenoptera (ants and some bees and wasps) and termites − and the less familiar social aphids and thrips (Chapman et al. 2008, Pike and Foster 2008). Why then do workers/soldiers so strikingly forgo their personal reproduction to help others raise offspring in these different lineages? With this question, we return to the central question in the study of major transitions in evolution, and it is for this reason − the altruism of workers and soldiers − that the social insects have become a model system for studying cooperation (e.g. Bourke and Franks 1995, Crozier and Pamilo 1996, Bourke 1997, Queller and Strassmann 1998, Korb and Heinze 2008a, Gadau and Fewell 2009). Several key factors favouring cooperation and altruism became evident from these studies, which I will briefly summarise in the following (for more details see also Heinze this volume, Kraus and Moritz this volume).

Social insects, major evolutionary transitions and multilevel selection

183

7.2.1 Explaining cooperation and altruism The key for explaining the evolution of cooperation and altruism is kin selection theory (Hamilton 1963, 1964, Maynard Smith 1964), the idea that genes are not only transmitted to the next generation via own offspring (direct fitness) but also through offspring of close relatives (indirect fitness). Note, kin selection is not a separate force to natural selection and its application is not only restricted to interactions between kin (see Hamilton 1963, 1964, Gardner and Foster 2008). Thus, the term inclusive fitness (direct and indirect fitness effects) theory might be more appropriate. Inclusive fitness theory is encapsulated in Hamilton’s rule (1963, 1964). It states that a behaviour is evolutionarily favoured if it results in a net increase of the inclusive fitness of the individual: br – c > 0 where: br comprises the indirect fitness component, composed of the fitness benefit (b) for the recipient multiplied by the coefficient of relatedness (r) between actor and recipient, and c being the direct fitness component measured as the actor’s direct fitness cost of the behaviour. From Hamilton’s rule it is easily derived that an altruistic behaviour that decreases the direct fitness of an individual can only be evolutionarily favoured if it increases its indirect fitness component sufficiently. Accordingly, insect societies where individuals behave altruistically are generally comprised of family groups with offspring helping their mother (social Hymenoptera, thrips, aphids) or parents (termites) to reproduce (e.g. Wilson 1971, Korb and Heinze 2008a, Gadau and Fewell 2009). Thus, the altruists gain indirect fitness benefits through raising siblings. At the same time, however, the degree of relatedness (r) alone is often not sufficient to explain the occurrence of altruism or details in the organisation of social life (Korb and Heinze 2008b, see below). For instance, variation in relatedness cannot explain the pattern of helping behaviour in the stenogastrine wasp, Liostenogaster flavolineata (Field 2008), or the occurrence of workers in the lower termite Cryptotermes secundus (Korb 2008; for more examples see Korb and Heinze 2008a). This is not surprising, given that Hamilton’s rule is composed of benefits (b) and costs (c) as well. Yet, ecological costs and benefits of cooperation have often been disregarded, also in standard textbooks, for example when attributing the evolution of eusociality in Hymenoptera erroneously to haplodiploidy alone (i.e. the fact that males derive from haploid eggs while females develop from fertilized diploid eggs; this haplodiploidy leads to relatedness asymmetries within colonies; Heinze this volume, Kraus and Moritz this vol-

184

Judith Korb

ume). One reason for this negligence is the paucity of empirical studies explicitly testing these costs and benefits (for exceptions see e.g., Nonacs and Reeve 1995, Reeve et al. 1998, Field et al. 1998, 2000, Korb 2006, Foster and Xavier 2007). Additionally, with the advent of molecular tools, r can easily be quantified but ecological factors hidden in the c and b term of Hamilton’s rule are much more difficult to measure. 7.2.2 Ecological factors favouring cooperation: sociality syndromes The limited data available, supplemented by data on cooperatively breeding vertebrates, have revealed a number of ecological factors that repeatedly show up as vital for the evolution of cooperation and altruism (resultig in social life; see Glossary) and three sociality syndromes have been identified, each with a unique combination of ecological factors (Korb and Heinze 2008b; Fig. 7.3). The importance of relatedness for the evolution of cooperation and altruism should depend on the syndrome to which a species belongs. From syndrome I to III, the importance of indirect fitness benefits through rais-

• Progressive food provisioning

Indirect fitness

allofeeding • Inbreeding impossible no inheritance opportunities Workers

Syndrome III:

Most ants,most social bees, foraging termite species

• No bonanza type food resource

part. progressive food provisioning allofeeding • Limited inbreeding limited inheritance opportunities • Less fortress defense Helpers

Syndrome II: Most wasps, most cooperatively breeding birds and mammals

• Bonanza type food resource

no allofeeding

• Inbreeding possible

inheritance opportunities • Fortress defense Soldiers

Syndrome I:

Aphids, thrips, wood-dwelling termites, naked mole-rat

Direct fitness

Fig. 7.3 Sociality syndromes. Shown are the three sociality syndromes positioned along the direct and indirect fitness axis. For more information see text.

Social insects, major evolutionary transitions and multilevel selection

185

ing offspring increases, while those of direct benefits mainly through nest inheritance decrease (Fig. 7.3). The importance of relatedness for the evolution of altruism is predicted to increase from syndrome I to III. It is important to recognise that each sociality syndrome is characterised by a set of ecological traits and that these traits determine the importance of direct versus indirect benefits for the evolution of sociality. For instance, the possibility to reduce the reproductive load of breeders through alloparental care determines the benefit of helping (Heinsohn 2004), or inbreeding avoidance limits inheritance of a territory or it leads to group instability (Emlen 1997). Sociality syndrome I: These species live in groups that are composed of totipotent individuals protected by altruistic defenders. They monopolise a long-lasting bonanza-type food source (Wilson 1971), which supports the co-existence of many individuals without selection for dispersal (Hamilton and May 1977). The totipotent individuals do not provide intensive alloparental care, probably because food is easily accessible to all individuals and alloparental care can hardly alleviate the reproductive burden of breeders substantially. Helpers seem to stay because the nest is a safe haven with plenty of food, and there is a substantial possibility of inheriting the natal breeding position. Moreover, the chances of founding an own nest independently are low due to high mortality risks during dispersal. The only truly altruistic individuals in these groups are soldiers (fortress defenders after Queller and Strassmann 1998), which mainly protect the resource against competitors and their nestmates against predators. Inbreeding occurs regularly in such societies and it seems to play an important role, not necessarily through increasing relatedness, but through reducing relatedness asymmetries in the haplodiploid thrips (Chapman et al. 2008) and stabilising groups over longer periods as heirs do not have to mate with unrelated partners (in contrast to vertebrates, Emlen 1997). Aphids, thrips, wood-dwelling termites and the naked mole rat belong to this group of social organisms. Sociality syndrome II: These species take an intermediate position between societies consisting of totipotent individuals of class (I) and those with altruistic, subfertile workers of class (III), and include social Hymenoptera with totipotent workers (e.g. wasps and queenless ants) and cooperatively breeding vertebrates. Helpers can gain indirect fitness benefits through alloparental care as well as direct benefits through inheriting the breeding position or by founding a nest of their own. The ecological factors favouring alloparental care in totipotent workers are identical to those for subfertile workers of class (III), namely the potential to provide costly

186

Judith Korb

help that frees the reproductives from provisioning their offspring. As they still have the opportunity to reproduce, it is expected that they adjust their degree of costly helping according to their opportunity to breed (Kokko and Johnstone 1999). Limited evidence suggests that this is indeed the case. Depending, for instance, on their rank in the society’s hierarchy, wasps and birds invest differentially in alloparental care (Field et al. 1999, Koenig and Dickinson 2004, Cant and Field 2005, Cant et al. 2006). Accordingly, the comparatively small group sizes in syndrome II species might be explained by several factors: (i) The opportunity of founding a nest independently, (ii) presumed decreasing benefits of helping with increasing group size (Michener 1964, Reyer 1984, Karsai and Wenzel 1998) and (iii) the limitation of food sources, which causes local resource competition. Sociality syndrome III: Species in this category comprise the classical social insects, i.e. most social Hymenoptera (such as the honeybee and most ants) and foraging termites with sterile or subfertile workers. They are characterised by intensive, altruistic alloparental care. They engage in costly helping with low chances of personal direct reproduction. Brood care usually involves progressive food provisioning (Strohm and Liebig 2008), which is costly to the reproductives and can be ‘handed over’ to workers. Consequently, reproductives can concentrate on egg-laying. These sociality syndromes cover (i) the classification of species along a eusociality gradient, where sociality was classified along the degree of altruistic helping (Sherman et al. 1995), (ii) the classification according to reproductive skew (in high-skew societies, actual reproduction is concentrated in one or a small set of individuals within a group, in low-skew societies reproduction is distributed more evenly among group members; Reeve 1998, Johnstone 2000), or (iii) the social trajectory approach in which a species social structure is viewed as the result of a trajectory of decisions individuals make about whether or not to disperse, whether to co-breed and so on (Helms Cahan et al. 2002). It might help to solve the longstanding debate in cooperatively breeding vertebrates whether their systems are driven by direct or indirect fitness benefits by showing that it can be both, and it aids to identify to what extent direct and indirect benefits are important. By identifying the ecological factors that are associated with each syndrome the different causes for social life are explicitly revealed. Note, these syndromes should not be considered as a new classification system, rather this approach aims at identifying various common combina-

Social insects, major evolutionary transitions and multilevel selection

187

tions of ecological factors that in concert seem to favour social life. Thus, apparently similar social organisations can be driven by different syndromes (e.g. ant and wood-dwelling termites), while similar ecological syndromes occur in phylogenetically distant groups (e.g. naked mole rat and wood-dwelling termites). 7.2.3 Conflict, enforcement and coercion The above examples explain cooperation and altruism as a ‘voluntary’ tactic (sensu Wenseleers and Ratnieks 2006). Yet, there have been recurrent discussions of whether they might not be the result of manipulation, either by parents or siblings (Alexander 1974, Michener and Brothers 1974, West-Eberhard 1975, Charnov 1978, Zimmerman 1983, Stubblefield and Charnov 1986, Roisin 1994). Recently these discussions received renewed interest under the question whether altruism in insect societies is voluntary or enforced (Wenseleers and Ratnieks 2006, Ratnieks and Wenseleers 2008). Most likely it is a bit of both. Coercion clearly can play a role in social evolution but it always does in a social context where relatedness is important. Thus, it probably is a mechanism which evolved secondarily within a social context (Boomsma 2007). In terms of inclusive fitness, coercion reduces the benefit of attempting to reproduce directly, relative to the benefit of rearing non-descendent kin (Lehmann and Keller 2006). Having addressed the benefits of cooperation above, I will now summarise evidence for enforcement in the following section. Behavioural observations show that insect societies are not the harmonious entities once supposed and that policing and punishment seem to be important mechanisms in solving within-colony conflicts in modern-day insect societies (Ratnieks and Visscher 1989, Ratnieks 1993, Liebig et al. 1999, Monnin and Ratnieks 2001, Foster et al. 2002, Hartmann et al. 2003, Hammond and Keller 2004, Wenseleers et al. 2004a,c, Bourke 2007). Additionally, new theoretical models suggested that the relatedness in present-day social Hymenoptera often is too low to account for their high degree of worker altruism, which might therefore only be explained by coercion (Wenseleers et al. 2004a,c, Ratnieks and Wenseleers 2008). Evidence for enforced altruism comes in several forms (according to Ratnieks and Wenseleers 2008). (i) Worker reproduction and policing. In many social Hymenoptera, workers can still produce unfertilized, male-destined eggs and, all else being equal, they are selected to do so because they are more related to their own sons than to either the sons of their mother queen or sister

188

Judith Korb

workers (Ratnieks 1988). The proportion of worker reproduction in a colony should depend on the costs of worker reproduction measured in colony efficiency (Ratnieks 1988, Wenseleers et al. 2004c). It decreases if workers reproduce instead of caring for the brood and/or if queens can produce more offspring than workers for the same investment. Based on these factors, models predicted that more workers should lay male eggs than is observed in present-day colonies and this restraint from egg-laying is interpreted as enforced altruism (Wenseleers and Ratnieks 2006, Ratnieks and Wenseleers 2008). Yet, there are few studies that actually measured the efficiency costs of worker reproduction in field colonies (Cole and Wiernasz 1999, Dampney et al. 2004). In societies with multiple-mated queens, such as the honeybee, relatedness considerations predict worker policing (i.e. the removal of worker laid eggs by other workers) because workers are more closely related to the queen’s sons than to their sister’s sons (Ratnieks 1988, Kraus and Moritz this volume). Indeed, worker policing occurs often under such situations (Ratnieks and Visscher 1989, Ratnieks 1993, Foster et al. 2002, Helenterä and Sundström 2007), but it also occurs in some species with a single-mated queen, suggesting that it can also have other benefits (Hartmann et al. 2003, Hammond and Keller 2004, Saigo and Tsuchida 2004). Evidence for coercion of egg-laying, and as a consequence reduced worker reproduction, comes from a comparative study of ten species. Species in which worker-laid eggs had a greater chance of being killed had smaller proportions of egg-laying workers (Wenseleers and Ratnieks 2006). Similarly, the presence of egg-laying workers in orphaned colonies (Monnin and Ratnieks 2001, Wenseleers and Ratnieks 2006) where policing is absent has been interpreted as evidence that workers are enforced to behave altruistically. (ii) Control over caste fate. In most social Hymenoptera, caste fate is under the control of the queen or workers. The food provided during larval development is critical for caste determination (Ratnieks et al. 2006). Yet, there are exceptions that support the idea that food control is evidence for coercion. In Melipona stingless bees, queens and workers are reared in identical, closed cells so that there are few possibilities for nestmate workers to influence caste fate during larval development (Bourke and Ratnieks 1999, Ratnieks 2001, Wenseleers and Ratnieks 2004). In some of these species, up to 20% of the larvae develop into queens. This is greatly in excess of the few needed to head a swarm, to replace a mother queen or to invade foreign colonies (Wenseleers et al. 2004b, Wenseleers and Ratnieks 2004). These extra

Social insects, major evolutionary transitions and multilevel selection

189

queens are later executed by the workers (Wenseleers et al. 2004b). A model showed that the number of these extra queens seems to reflect the optimal value from the developing workers’ point of view (Wenseleers et al. 2003, Ratnieks and Wenseleers 2008). Conversely, it was concluded that the lower numbers of new queens in species with worker control over caste fate is a result of coercion (Ratnieks and Wenseleers 2008). A similar phenomenon, where caste fate seems to be under the control of the developing individual, exists in drywood termites, the Kalotermitidae (Lenz 1985, Lenz et al. 1985). In contrast to social Hymenoptera, the workers of termites are not adults, but more or less differentiated (larval) immature stages (Roisin 2000, Korb and Hartfelder 2008). As hemimetabolous insects, these immatures are miniature-adults, which can perform all colony duties and are less dependent on alloparental care than the altricial grub-like larvae of social Hymenoptera (Korb 2007, 2008a). In wood-dwelling termites, to which the drywood termites belong, all individuals are totipotent to remain workers or to develop into soldiers or two kinds of reproductives (Roisin 2000, Korb and Katrantzis 2004, Korb and Hartfelder 2008): winged dispersing alates that found a new nest elsewhere or neotenic, replacement reproductives that inherit the natal breeding position when the same-sex reproductive of the colony dies (in contrast to social Hymenoptera, termite colonies are headed by a male and female reproductive).Thus, conflict over inheritance is predicted to occur when one of the natal reproductives dies (Korb 2005). As in Melipona species, the death of a natal reproductive can lead to an excess development of workers into replacement reproductives (Lenz 1985, Lenz et al. 1982, 1985). Yet, in the end only one pair survives and all other reproductives are being killed. This pattern seems to fit the Melipona example (Ratnieks et al. 2006, Ratnieks and Wenseleers 2008), but this might be a premature conclusion. Although caste fate is similarly ‘self-controlled’ in all Cryptotermes species, in many species only the lost reproductive is replaced without excess development of replacement reproductives. No general pattern in the mode of replacement exists (Lenz et al. 1982, Lenz 1985). Even closely-related sympatric species have different modes of replacement (J. Korb, unpubl. data). Thus, further studies are needed to address whether excess development of reproductives is really evidence for relief from coercion. Currently, it is assumed that there is little evidence for a significant role of parental manipulation in the origin of eusociality (Ratnieks and Wensel-

190

Judith Korb

eers 2008). Yet, coercion is considered to be a major cause of altruism in present-day insect societies. ‘Without coercion, altruism would still occur and societies would still function because family levels of relatedness are sufficient to cause most individuals to act altruistically. However, such societies would be more similar to queenless honey bee colonies or colonies of Melipona bees, in which a large proportion of the individuals tries to reproduce by developing into queens or by laying eggs’ (Ratnieks and Wenseleers 2008). In general, coercion should be more important in species with a higher degree of conflict between colony members. The potential for conflict depends on relatedness but even more important is whether individuals still can reproduce. Accordingly, the potential for the outbreak of conflict – and thus the evolution of coercion – should be most pronounced in sociality syndrome I and II species, where individuals can still gain direct fitness. In many syndrome III species conflicts might have been resolved through the evolution of sterility.

7.3 Social insects and the major transitions in evolution: Common mechanisms What do these insights from social insects tell us about the major evolutionary transitions? Why was cooperation evolutionarily stable and why did it lead to increasing evolutionary complexity? Studies on lower level systems (e.g. the transition to multicellularity studied in Volvocales: Kirk 1998, Michod 1999, Herron and Michod 2008; genomic conflicts: Hurst et al. 1996, Pomiankowski 1999) revealed striking similarities to the results on social insects. At least two common mechanisms prevent the spread of cheaters and favour stable cooperation among social insects: interactions among relatives align the evolutionary interests of individuals and enforcement mechanisms make cheating costly. In the following, I will show similar mechanisms at other levels of the biological hierarchy to illuminate their potential generality. 7.3.1 Common ancestry and aligned interests A major explanatory principle that promotes cooperation over cheating is genetic relatedness, whereby like genotypes preferentially interact. More formally, genetic relatedness is defined as the probability above chance that individuals have identical alleles at one or more loci (see Glossary). Preferential interaction by relatedness ensures that cooperator genotypes

Social insects, major evolutionary transitions and multilevel selection

191

will occur together and that cheater genotypes do the same preventing the exploitation of the former by the latter (Bourke and Franks 1995, Gilbert et al. 2007). In social insects, individuals live in families. Hence, individuals share any ‘cooperator/altruist’ alleles with a high probability through common ancestry and one individual can increase the frequency of its alleles that are carried to the next generation by helping the other individual to reproduce. Note that the important measure of relatedness that determines the evolutionary trajectory of an allele for a social action is relatedness at its locus (Bourke and Franks 1995). Similarly, the cells in multicellular organisms derive from a single cell and are thus identical clones. A notable exception are slime molds: In the slime mold, Dictyostelium, aggregates of independent single cells build a ‘slug’ and ‘fruiting body’ for dispersal when resources in the environment become rare (Queller et al. 2003). The cells of fruiting bodies seem to be closely related (Gilbert et al. 2007). Besides kin recognition, which leads to preferred aggregation of relatives (Mehdiabadi et al. 2006), cheating seems to be rare because of pleiotropic gene effects (Foster et al. 2004). Common descent is probably the most common mechanism to form groups of cooperators at the exclusion of cheaters. Yet, other mechanisms, such as green-beard alleles (see Glossary), habitat heterogeneity or population viscosity in general, may in principle lead to a positive assortment of cooperators (Dawkins 1976, Wilson and Dugatkin 1997, Pepper and Smuts 2002, Axelrod et al. 2004, Penn and Frommen this volume). But the latter mechanisms seem to be uncommon in nature (Grafen 2006). Yet the Gp-9 locus in the red fire ant, Solenopsis invicta, proves a rare example for a green-beard mechanism (Keller and Ross 1998). In multiple-queen colonies, all egg-laying queens are Bb heterozygotes at this locus. Workers that carry the b allele of this gene kill BB queens (which lack the b allele). This implies that allele Gp-9b is a green-beard allele that preferentially induces workers bearing the allele to kill all queens not bearing it. Not all major evolutionary transitions evolved from cooperation within a species: some have arisen as between-species cooperations. This includes the evolution of chromosomes, where ‘species’ are genes, and eukaryotes, which evolved from different prokaryotic ancestors (Maynard Smith and Szathmáry 1995, Cavalier-Smith and Chao 2003, Emelyanov 2003). The equivalent of common descent for cooperation between species is the linkage of the two cooperating partners during multiplication or reproduction (Fig. 7.4). If the reproductive success of both partners is coupled so that an increase in reproductive success of one partner is correlated with an increase in reproductive success of the other, stable cooperation can evolve (Fig. 7.4; Frank 2003).

192

Judith Korb

(a) Common ancestry and aligned future: Related entities with vertical inheritance

(b) No common ancestry and aligned future: Non related entities with vertical inheritance

(c) Common ancestry and not-aligned future: Related entities with horizontal inheritance

(d) No common ancestry and not-aligned future: Non related entities with horizontal inheritance

Fig. 7.4 The impact of common ancestry and aligned future interests for the evolutionary stability of cooperation. Fitness is indicated by the size of the circles; large open circles indicate that entities can invade or leave the group; large closed circles indicate that the group is closed; blue: cooperator; red: cheater that invades from outside; yellow: cheater mutant that arises within the cooperative group. (a) Common ancestry and aligned future (indicated by only closed circles) prevent the spread of cheaters. If a cheater mutant arises within a cooperative group (yellow circle in group with blue circles), it might have a higher within-group fitness (larger circle) than cooperators, but it decreases the fitness of the group (reduced size of enclosing circle). As the cheater cannot spread because of vertical transmission, this reduction in group fitness reduces its own direct fitness which selects against cheating. (b) No common ancestry (indicated by the 1. open circle) and aligned future (indicated by the 2. + 3. closed circle). Cheaters can easily invade the group from outside (red circle coming into the group circle), which reduces the group fitness. Similar to (a) cheaters will not easily spread because their fitness is linked to group fitness. (c) Common ancestry (1.+ 2. closed circle) and not-aligned future (3.+ 4. open circle). A cheater mutant can arise within a cooperative group (yellow circle in group with blue circles). It will spread and annihilate cooperation as it can leave the exploited group through horizontal transmission. (d) No common ancestry and not-aligned future (open circles only). Cheaters can easily invade the group from outside (red circle coming into the group circle), which reduces the group fitness. Cheating will spread and annihilate cooperation as the cheater can leave the exploited group through horizontal transmission.

Social insects, major evolutionary transitions and multilevel selection

193

This process can be studied in symbiotic associations such as fungus growing ants or termites (Korb and Aanen 2003, Mikheyev et al. 2006). Here, the transition from mutualism to parasitism is gradual, as each partner is selected to exploit the other. Yet, if reproduction of both partners is linked, there is an alignment of reproductive interests and interactions are more mutualistic. One mechanism to link reproduction is vertical inheritance, i.e. the transmission of the symbiont via hosts’ offspring (Fig. 7.4; Frank 2003). Alternatively, symbionts can be re-acquired each generation anew (horizontal inheritance), with symbionts reproducing independently from their hosts (e.g. as is the case for viruses, or disease in general; Fig. 7.4). Cytoplasmic elements of eukaryotes are transmitted via vertical inheritance (Cosmides and Tooby 1981, Hoekstra 1990). Both mitochondria and plastids are propagated to the next generation together with the nucleus in the germline. Interestingly, this inheritance is exclusively maternal, i.e. cytoplasmic elements are propagated only via eggs but not via sperm. Uniparental inheritance reduces the amount of genetic mixing of cytoplasmic elements from one generation to the next, and thus reduces competition. Hence, it is discussed to be an adaptation to reduce genetic conflict (Cosmides and Tooby 1981). (Note, at the same time uniparental inheritance creates new conflict, see below). Vertical inheritance aligns the reproductive interests of both partners because, if the symbiont harms its host, and thus decreases the host’s reproductive success, it indirectly harms itself (Fig. 7.4). Similar to relatedness, the linkage between both partners favours self-restraint. Furthermore, it provides a mechanism for cooperating partners to stay together and prevents infiltration by cheaters. In summary, for both stable cooperation within and between species, mechanisms are essential that (i) stably assort groups/pairs of cooperators in space and time and separate them from potential cheaters, and (ii) align the reproductive interests of cooperators over time. Common ancestry, through classical relatedness or vertical inheritance, seems to be a major mechanism that led to stable cooperation. Through common ancestry and linked/aligned future interests, new evolutionary levels can emerge (Fig. 7.4). 7.3.2 Conflict and conflict resolution Through common ancestry and aligned future, organisms appear as homogeneous entities whereupon selection acts. Yet, similar to social insects, they are not the harmonious entities that they were once considered to be. Evidence has accumulated indicating that conflict occurs at all levels below the individual (Burt and Trivers 2006).

194

Judith Korb

Nuclear genes in a diploid genome generally cooperate because each allele has an equal probability of being represented in a gamete. This fairness is guaranteed by meiosis, which aligns the interests of nuclear genes (Frank 2003). However, there are genes, segregation distorters, which cheat during meiosis or gametogenesis and are therefore present in more than half of the functional gametes. The most studied examples of segregation distorters are sd in Drosophila melanogaster (Temin et al. 1991), the t-haplotype in Mus musculus (Lyon 1991, Ardlie and Silver 1996, 1998) and sk in Neurospora fungi (Turner and Perkins 1991, van der Gaag et al. 2000). Segregation distorters that are present in sexual chromosomes (as the X chromosome in several Drosophila species) are denominated sexratio distorters, as they induce a sex-ratio bias in the offspring of the carrier. These genes cheat during transmission to the next generation. Homing endonuclease genes (HEG) convert their rival allele into a copy of themselves, and are thus inherited by nearly all meiotic daughter cells of a heterozygote cell (Gutz and Leslie 1976). They achieve this imbalance by encoding an endonuclease that breaks the rival allele. This break is repaired by using the sequence of the HEG as template. A second class of selfish genes cheats by multiplying themselves within the genome. For example, transposons are autonomously replicating genes that encode the ability to move to new positions in the genome and therefore accumulate in the genomes (reviewed in Craig et al. 2002). They can replicate themselves in spite of being detrimental to the rest of the genome. At the next level within eukaryotes, conflict also occurs between the cytoplasmatic elements and the nuclear genes. As detailed above, cytoplasmatic elements are transmitted to the next generation via maternal inheritance. Thus, they have no evolutionary interest in the production of male offspring, which do not transmit mitochondria and plastids. By contrast, the nuclear genome is transmitted through female and male offspring equally, and an evolutionary stable 1:1 sex ratio is favoured. Thus, conflict over sex ratios arises. One outcome of this conflict seems to be cytoplasmic male sterility (CMS) in many flowering plants: Most plants are hermaphrodites – they produce both pollen and ovules – but their mitochondria are only transmitted through ovules. So, mitochondrial mutants that abolish male production, causing CMS, are positively selected (Lewis 1941). Because nuclear genes are transmitted equally through pollen and ovules, a nuclear gene causing male sterility will spread only if it more than doubles female fertility. As a result of this conflict, the spread of CMS genes usually selects for nuclear suppressor genes that counteract the CMS gene and restore male fertility (e.g. Budar et al. 2003). Due to this co-evolutionary arms race, when nuclear genes have the upper hand, CMS genes are often hidden within a population. They only re-appear when

Social insects, major evolutionary transitions and multilevel selection

195

CMS genes are ‘uncoupled’ from their nuclear repressor genes during cross-population- or interspecific hybridisation (Budar et al. 2003). Mechanisms which align the reproductive interests of the selfish gene and the rest of the genome, or which make drive directly more costly for the selfish gene (e.g. through punishment) could result in conflict resolution. Yet, while the former seems to be widespread – for instance, the physical linkage of genes in chromosomes or the process of meiosis – direct punishment as is found in social insects seems to be rare, the closest equivalent being nuclear repressor alleles. At the inter-cell level of multicellular organisms, conflicts seem to be less common. Although cancer is a well-known example of cells replicating selfishly at the cost of the organism, lethal cancer is strongly selected against if expressed before cessation of reproduction. Reasons for the rarity of selfish cells in multicellular organisms include their common descent from a single cell and the early separation of germ and soma in animals. This leads to limited opportunities for cheater mutants to invade and to aligned interests of the cells of an organism. Notable exceptions are the canine transmissible venereal disease, and its analogue in the Tasmanian devils, the devil facial tumor disease, which are contagious tumors that are transmitted directly from individual to individual (e.g. Murgia et al. 2006, Pearse and Swift 2006). This horizontal transmission leads to non-aligned future interests (Fig. 7.4).

7.4 The level of selection debate: multilevel selection So far, we have seen that animals consist of a hierarchy of selection levels and that the entities at each level do not always agree in their evolutionary interests, so that conflicts can arise. The question then becomes: At which level does selection act? What is the unit of selection? There has been a longstanding debate about the level of selection (Reeve and Keller 1999). Before the 1960s it was not uncommon to regard species as the unit of selection (as reviewed by Williams 1966). Wynne-Edwards (1962) used adaptations at the group level to reason about the evolution of population size. He and many others suggested that self-restraint in reproduction (voluntary birth control) should evolve to prevent overexploitation of resources within populations because it is good for the survival of the species. Such ‘good for the species’ arguments are examples of the ‘old’ group selection, where behaviours were carelessly explained as group- or species-level benefits (Grafen 1984). As Williams (1966) pointed out, self-restraint will generally not be stable. As seen in the ‘tragedy of the commons’ (Hardin

196

Judith Korb

BOX 7.1 The Price equation and multilevel selection (after Wenseleers et al. in press) The Price equation (Price 1970, 1972) is the foundation for a universally applicable theory of selection (for details see Frank 1995, Wenseleers et al. in press). Consider a population of n entities indexed by j. These entities are usually individuals, but they can also be genes within genomes, cells, or social groups. Let wj be the absolute fitness of the j th entity (i.e. how many successful offspring entities it leaves in the next generation), and vj the fitness relative to the population average (wj/w ¯ ). The standard form of Price’s theorem states that the average change in the value of some trait z (Δz) from one generation to the next is given by  wj   wj   z  cov , z j   E  z j   cov v j , z j  E v j z j  w   w 



 



(1)

where the terms cov and E denote covariance (a measure of how two variables change together; here vj and zj) and expectation (arthimetric average), respectively, both taken over all entities in the population. The term Δzj is the change in the entity’s trait value zj between parent (zj) and offspring (zj’), where Δzjz = zj’ – zj. The first term in equation (1) generally reflects the effects of selection (Price 1970, Frank 1995, Okasha 2006). This can be illustrated by decomposing it into two separate components: cov(vj, zj) = βvz.Vz, with a leastsquares regression coefficient, βvz., and a statistical variance Vz. The regression term β describes the intensity of selection (whether the trait of interest z will increase or decrease the relative fitness v of the focal entity) and the trait-variance term V gives us the rate at which selection acts. The more variable the focal trait, the more possibilities for selection. The second term in equation (1), E(vj Δzj), captures all effects other than selection. Social evolution in terms of opposing selection can be analysed as follows: The standard Price equation (1) must be slightly changed in notation. We will use subscripts i and ij to refer to the i th group and the j th individual within group i, respectively. The evolutionary change in the average gene frequency g is a function of the mean fitness W and mean gene frequency G in the i th group w g  covwi , g i   Ei wi g i 

(2)

Equation (2) describes selection on the groups in our population. To account for selection on individuals within each group, the expectation term Ei(wiΔgi) can easily be expended to capture the effects of within-group selection

Social insects, major evolutionary transitions and multilevel selection









Ei wi g i   Ei cov i wij , g ij  E j.i wij g ij



197

(3)

where the right-hand side is a second version of the standard Price equation, but on the next lower level in the selective hierarchy, i.e. it describes withingroup selection. Substituting equation (3) into equation (2) yields:









w g  covwi , g i   Ei cov i wij , g ij  E j.i wij g ij



(4)

where the expectations and covariances are taken over their subscripts, with i standing for groups, ij standing for individual j of group i, and j.i for individuals j for a specified group i. The first covariance term captures the effects of the gene on group success, the second covariance term captures the effect of the gene on the relative success of individuals within a group, and the final term accounts for any effects caused by processes other than selection. One of the beauties of Price’s equation is that it can be expanded to include multiple levels of selection until all relevant levels are included. Doing so, and disregarding effects other then selection, the last term (Ej.i(wijΔgij)) can be set to zero. As the mean fitness W is always greater than zero, a gene for a social trait is selected for when







covwi , g i   Ei cov i wij , g ij   wi gi  V gi   wij gij.i  V gij .i  0

(5)

where the covariances are broken up into their constituent regression and variance terms. In this inequality, the two components reflect between-group and within-group (among individual) selection, respectively. Each level of selection comprises a selective response which equals the intensity of selection of that level weighed by the genetic variance present at that level. The between- and within-group genetic variances can be calculated using population genetics, namely Wright’s hierarchical F-statistics (Yang 1998). Importantly, however, they can also be expressed as a function of genetic relatedness, which links back to kin selection (Hamilton 1975).

1968), cheaters that reproduce at the expense of others will be evolutionarily favoured. This insight, together with inclusive fitness theory (Hamilton 1963, 1964), led to the general consensus that groups of individuals cannot be units of selection. Instead, as popularised by Dawkins (1976), genes are the ultimate units of selection and organisms can be viewed as higher-level ‘vehicles’ upon which selection acts. As a consequence, during the 1960s and 1970s, group selection became such a heretical concept that many evolutionary biologists stopped reading the actual literature (Wilson 1997, 2007). Parallel to this development, mainly Price (1970, 1972), Hamilton (1975) and Wade (1978) developed quantitative models, which are now often referred to as the ‘new’ group selection theory (common other terms:

198

Judith Korb

intrademic group selection, trait-group selection, levels of selection or multilevel selection; I will use the latter in the following). These models date back to some simple models by Wright and Haldane in the 1930s (Gardner and Foster 2008). In Hamilton’s (1996) autobiographical account, he recalls excitedly telling Price that ‘through a ‘group-level’ extension of his formula I now had a far better understanding of group selection and was possessed of a far better tool for all forms of selection acting at one level or at many than I had ever had before’ (Wilson 2007:173). These models showed that group selection can occur, yet ‘arguments must be applied carefully without neglecting competition between individuals within each of these units’ (Gardner and Foster 2008) as was done before in the case of the old group selection arguments. Using Price’s covariance theorem, this approach phrases social evolution in terms of selection within and between groups (BOX 7.1), rather than separating individual fitness into direct and indirect components. This partition can be very useful for conceptualising the potential tension between the interests of individuals and the group (Hamilton 1975). For example, it can be analysed whether social traits, which may be favoured at one level but counter-selected at another level, would still evolve. Interests of entities at the same levels are aligned if within-group variance is small compared to between-group variance. Mechanisms to reduce within-group variance include relatedness or any other mechanism that leads to the formation of groups composed of kind (see above). One of the beauties of the multilevel selection approach is that the same conceptual framework can be applied to all levels of the biological hierarchy, from genes to groups of organisms (Wilson 1997), because it can be mathematically formalised with Price’s covariance equations whenever there is a hierarchical cluster of groups (BOX 7.1). Addressing cooperation within species, this equation is mathematically equivalent to the inclusive fitness approach. It is important to note that the multilevel selection approach is not in contradiction to inclusive fitness or similar fitness approaches (e.g. neighbour modulated fitness). Rather, they are complementary perspectives to look at the same problem. It is sad that there is a continuing tendency to mistakenly assume that switching between the methods also means that different biological processes are at play (e.g. Wilson 1975, Sober and Wilson 1997, Fehr and Fischbacher 2003, Wilson and Hölldobler 2005; for a thorough treatise of the problem see: e.g. Foster et al. 2006). Each approach has its advantages and disadvantages (Wenseleers et al. in press). Some important caveats of the multilevel selection approach are that it still suffers from semantic difficulties (Okasha 2006, West et al. 2007) or that, when wrongly applied, nonsocial traits that are independent of group effects will be attributed to the

Social insects, major evolutionary transitions and multilevel selection

199

operation of between-group selection (Heisler and Damuth 1987, Gardner and Foster 2008). Probably the most important problem at present is that multilevel selection theory comprises a large amount of verbal discussion and a collection of mathematical models that appear to provide helpful insights but a conceptual unity is still missing (reviewed in Okasha 2006). There are two main advantages of adopting a multilevel selection perspective (Wenseleers et al. in press). First, it offers a possibility to analyse selection at more than two levels. Thus, it can cover tensions at all levels of the biological hierarchy and can make quantitative predictions about the outcome. Second, in social insects, it might offer a possibility to switch from the relatedness-centered research to studies that address those factors that are hidden in the c and b term of Hamilton’s rule. This might be especially promising as insect societies with their colony level phenomena (e.g. self-organisation) are clearly more than the sum of their components.

7.5 Conclusions, open questions and future research Insect societies are more than the sum of their components. Through the interaction with many individuals, more than additive benefits derive, which probably explains the impressive evolutionary and ecological success of social insects. These non-linear effects are also the basis for the efficiency costs that often seem to prevent the occurrence of conflicts (see above). Such colony-level phenomena (emergent properties) ‘equip’ a colony with new properties, including: (i) Division of labour. Division of labour allows specialisation with increased efficiency in the performance of tasks (e.g. Beshers and Fewell 2001). In social insects this specialisation is commonly reflected in the evolution of morphological castes, each caste being adapted for particular tasks. Due to division of labour, queens are released from brood care and can concentrate on egg-laying. An impressive example are queens of the fungus-growing termite Macrotermes bellicosus: they increase in body size more than 10fold as they age (Fig. 7.1). This physogastry is caused by an increase in the number of active ovarioles, so that queens become true ‘egg-laying’ machines that can produce more than 20 millions eggs daily! These queens can hardly move around, let alone care for their brood. This extreme specialisation is only possible because workers care for the brood and build a protective nest, and specialised soldiers defend them in case of danger. Such a specialisation might also release individuals from de-

200

Judith Korb

velopmental constraints (Roux et al. 2009). For instance, a trade-off between fecundity and defense (i.e. investment in weapons is associated with reduced fecundity) seems to be common in many male insects (Simmons and Emlen 2006). By devoting different individuals to different castes, this trade-off can be solved (Roux et al. 2009). Yet, the trade-off between defense and fecundity might be reflected in the caste composition of the colony; the caste for defense, the soldiers, are always sterile and the reproductives generally lack defensive adaptations. Thus, the trade-off might have shifted from the individual to the colony level (Roux et al. 2009). (ii) Information transfer. Living in social groups requires information transfer (communication) between group members. Indeed, social insects have evolved very elaborate types of communication. Communication during foraging is often achieved through trail pheromones (Kaib et al. 1982): chemical substances are deposited by foragers on the ground when a rewarding food item is found. Through positive feedback processes (i.e. each successful forager deposits the chemical blend), this can lead to the recruitment of nestmates to rewarding food sites (Bonabeau et al. 1997). Yet, communication via trail pheromones only works if there is a minimum number of foragers (Beekman et al. 2001), otherwise the pheromones evaporate too quickly to provide information. The equivalent to trail pheromones in honeybees is their ‘dance language’ (von Frisch 1946, 1967, Lindauer 1954, Dyer 2002, Grüter and Farina 2009, Pahl et al. this volume). Such ‘sophisticated’ communication mechanisms provide insect colonies with very efficient means to exploit food resources. Mass recruitment becomes possible that allows the exploitation of new food sources, inaccessible to solitary insects, as well as their monopolisation from neighbours. (iii) Building. The existence of a protective nest is a central characteristic of all social insects (Seger 1991). It is a necessary requirement to have a common place where individuals interact and the brood is reared. Yet, the nests of social insects are more than that. Through selforganisational processes elaborate ‘buildings’ can be constructed, especially in species with large colony sizes (Hansell 1984, Wenzel 1991, Heinrich 1993, Turner 2000; Fig. 7.1). Such mounds/nests protect the colony against predators and harsh environmental conditions, providing homeostatic conditions that are largely independent of environmental fluctuations, so that a year-round production of offspring is possible. An especially striking example is the fungus-cultivating termite M. bellicosus (Fig. 7.1). It has air-conditioned mounds with efficient mechanisms for the exchange of respiratory gases that provide constant temperatures, fluctuating less than 2°C daily and annually

Social insects, major evolutionary transitions and multilevel selection

201

(Lüscher 1961, Korb and Linsenmair 2000, Korb 2003). The construction and maintenance of such structures is only possible because the individuals live in large colonies. Although we are only beginning to understand how macroscopic features of insect buildings emerge, the key lies in self-organisation – the emergence of higher-level patterns and collective action from simple low-level behaviours (e.g. Bonabeau et al. 1997, Camazine et al. 2001). Similar to the social insects, the lower level selection units are also characterised by properties that are more than the sum of their components. Multicellular organisms and genomes share division of labour with social insects. In multicellular organisms, different organs are specialised in performing different tasks (e.g. the gonads for reproduction, the liver for disposal of metabolic waste or the skin for protection), as are the genes of chromosomes that encode these different functions (although in a less simple 1:1 correspondence). Interestingly, recent data at the genome level indicate that gene effects at the phenotype are also less additive than formerly thought. There are many non-linear epigenetic effects between the interactions of genes that determine the phenotype. This again reminds us of the interactions of nestmates in an insect colony, which result in new, emergent phenotypes at the colony level. Whether the genetic architecture of organisms and their phenotypes is governed by the same principles as insect colonies will be an exciting question for future researchers to address. The rapidly increasing technology for fast and cheap sequencing will accumulate enough genomic data to reveal patterns of organisation. Unraveling the mechanisms and algorithms between nestmate interactions and their resulting colony phenotypes might help to interpret the patterns observed at the genotype-phenotype transition. It might be these emergent properties that arise through the interaction of many ‘simple’ units that explain the evolutionary stability of the ‘major transitions in evolution’. These group-level benefits that provide the new, more complex level with beneficial attributes beyond those possible at the lower level, provide a high direct fitness benefit for cooperation as new niches can be occupied or old ones exploited more efficiently. These benefits have scarcely been measured. Taking a multilevel selection perspective and comparing the reproductive success of insect colonies with various traits might become an important aim for future studies. By manipulating the degree of cooperation and conflict within insect societies it is possible to measure the benefits and costs of cooperation at the colony level and to transfer them back to the individual. Thus, for instance, efficiency costs of conflicts can be quantified. Measuring these costs at the colony level − within a multilevel selection framework, which does not forget the indi-

202

Judith Korb

vidual level and the covariance (= relatedness, see BOX 7.1) between group members − it might be possible to determine the importance of ecological factors. During the last two decades, most social insect research was devoted to measuring the influence of relatedness on social evolution, and we have learned a lot. Yet, as Hamilton’s rule states that relatedness cannot explain everything, now might be the right time to consider and measure the costs and benefits of social interactions. Tsuji’s (1995) work on the ant Pristomyrmex pungens might serve as an illustrative example. In this species, workers reproduce pathenogenetically and a worker dimorphism exists with large reproducing workers having a higher within-colony fitness than small reproducing worker. However, large workers reduce a colony’s fitness at the between-colony level. Using a multilevel selection approach (contextual analysis) and partitioning of covariance Tsuji (1995) showed that selection on the proportion of foragers appeared to be responsible for the evolutionary maintenance of cooperative breeding in this species. A further promising approach in this context is provided by social network analyses. This analytical tool examines social groups as networks of nodes connected by social ties (e.g. Wey et al. 2008, Krause et al. 2009). Thus, network analyses are especially suited to characterise, analyse and compare interactions among colony members at the colony level, and they are probably an appropriate means to quantify emergent colony properties. Social insects have become model organisms for the study of social evolution. We have learned that relatedness through common ancestry is an important means to favour cooperation and altruism. It is an efficient mechanism that prevents cheating because like genotypes interact preferentially. Similar mechanisms that exclude cheaters and align the reproductive interests of group members can also be found at other levels of the evolutionary hierarchy. Future research might take a multilevel selection perspective to quantify the benefits and costs of cooperation, which often seem to have non-additive fitness consequences, so that insect societies and all major evolutionary transitions alike are being recognised as more than the sum of their components.

Acknowledgements I thank Peter Kappeler for the invitation to contribute to this volume. He, Kevin Foster and two anonymous referees also provided very helpful comments on the manuscript.

Social insects, major evolutionary transitions and multilevel selection

203

GLOSSARY Altruism: Carrying out an action that benefits a recipient (i.e. increases the evolutionary fitness of the recipient) at a cost to an individual’s own lifetime fitness (i.e. direct fitness) Cooperation: Carrying out an action that benefits a recipient without incurring a net cost to the actor Division of labour: Specialisation of cooperative labour in specific, circumscribed tasks Emergent properties: Emergence is the way complex systems can arise out of relatively simple interactions. Emergent properties can appear when a number of simple entities operate and form more complex behaviours as a collective. Green-beard gene: A gene that causes a phenotypic effect (such as the presence of a green-beard or any other conspicuous feature), allows the bearer of this feature to recognise it in other individuals, and causes the bearer to behave differently towards other individuals depending on whether or not they possess the feature. Relatedness: A genetic correlation between individual loci or organisms Social life: Living in a group with cooperative or altruistic interactions among individuals Social organisation: The characteristics of a social group, e.g. its sex ratio or its reproductive skew

References Alexander RD (1974) The evolution of social behavior. Annu Rev Ecol Syst 5:325-383 Ardlie KG, Silver LM (1996) Low frequency of mouse t haplotypes in wild populations is not explained by modifiers of meiotic drive. Genetics 144:17871797 Ardlie KG, Silver LM (1998) Low frequency of t haplotypes in natural populations of house mice (Mus musculus domesticus). Evolution 52:1185-1196 Axelrod R, Hammond RA, Grafen A (2004) Altruism via kin-selection strategies that reply on arbitrary tags with which they coevolve. Evolution 58:1833-1838 Beekman M, Sumpter DJT, Ratnieks FLW (2001) Phase transition between disordered and ordered foraging in Pharaoh’s ants. Proc Natl Acad Sci USA 98:9703-9706 Beshers SN, Fewell JH (2001) Models of division of labor in social insects. Annu Rev Entomol 46:413-440 Bonabeau E, Theraulaz G, Deneuborug JL, Aron S, Camazine S (1997) Selforganization in social insects. Trends Ecol Evol 12:188-193

204

Judith Korb

Boomsma JJ (2007) Kin selection versus sexual selection: why the ends do not meet. Curr Biol 17:R673-R683 Bourke AFG (1997) Sociality and kin selection in insects. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach, 4th edn. Blackwell, Oxford, pp 203-227 Bourke AFG (2007) Social evolution: community policing in insects. Curr Biol 17:R519-R520 Bourke AFG, Franks NR (1995) Social Evolution in Ants. Princeton University Press, Princeton Bourke AFG, Ratnieks FLW (1999) Kin conflict over caste determination in social Hymenoptera. Behav Ecol Sociobiol 46:287-297 Budar F, Touzet P, De Paepe R (2003) The nucleo-mitochondrial conflict in cytoplasmic male sterilities revisited. Genetica 117:3-16 Burt A, Trivers R (2006) Genes in Conflict. The Biology of Selfish Genetic Elements. Harvard University Press, Cambridge/MA Camazine S, Deneubourg JL, Franks NR, Sneyd J, Theraulaz G, Bonabeau E (2001) Self-Organization in Biological Systems. Princeton University Press, Princeton/NJ Cant MA, Field J (2005) Helping effort in a dominance hierarchy. Behav Ecol 16:708-715 Cant MA, Llop JB, Field J (2006) Individual variation in social aggression and the probability of inheritance: theory and a field test. Am Nat 167:837-852 Cavalier-Smith T, Chao EE-Y (2003) Phylogeny of choanozoa, apusozoa, and other protozoa and early eukaryote megaevolution. J Mol Evol 56:540-563 Chapman TW, Crespi BJ, Perry SP (2008) The evolutionary ecology of eusociality in Australian gall thrips: a ‘model clades’ approach. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 57-83 Charnov EL (1978) Evolution of eusocial behavior − offspring choice or parental parasitism. J Theor Biol 75:451-465 Cole BJ, Wiernasz DC (1999) The selective advantage of low relatedness. Science 285:891-893 Cosmides ML, Tooby J (1981) Cytoplasmic inheritance and intragenomic conflict. J Theor Biol 89:83-129 Craig NL, Craigie R, Gellert M, Lambowitz AM (2002) Mobile DNA II. ASM Press, Washington/DC Crozier RH, Pamilo P (1996) Evolution of Social Insect Colonies: Sex Allocation and Kin Selection. Oxford University Press, Oxford Dampney JR, Barron AB, Oldroyd BP (2004) Measuring the cost of worker reproduction in honeybees: work tempo in an ‘anarchic’ line. Apidologie 35:8388 Dawkins R (1976) The Selfish Gene. Oxford University Press, Oxford Dyer FC (2002) The biology of the dance language. Annu Rev Entomol 47:917949 Emelyanov VV (2003) Mitochondrial connection to the origin of the eukaryotic cell. Europ J Biochem 270:1599-1618

Social insects, major evolutionary transitions and multilevel selection

205

Emlen ST (1997) Predicting family dynamics in social vertebrates. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach, 4th edn. Blackwell, Oxford, pp 228-253 Fehr E, Fischbacher U (2003) The nature of human altruism. Nature 425:785-791 Field J (2008) The ecology and evolution of helping in hover wasps (Hymenoptera: Stenogastrinae). In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 85-107 Field J, Foster W, Shreeves G, Sumner S (1998) Ecological constraints on independent nesting in facultatively eusocial hover wasps. Proc R Soc Lond B 265:973-977 Field J, Shreeves G, Sumner S (1999) Group size, queuing and helping decisions in facultatively eusocial hover wasps. Behav Ecol Sociobiol 45:378-385 Field J, Shreeves G, Sumner S, Casiraghi M (2000) Demographic advantages in the evolution of eusociality. Nature 404:869-871 Foster KR (2008) Behavioral ecology: altruism. In: Encyclopedia of Ecology. Elsevier, Amsterdam, pp 154-159 Foster KR, Xavier JB (2007) Cooperation: bridging ecology and sociobiology. Curr Biol 17:R319-R321 Foster KR, Gulliver J, Ratnieks FLW (2002) Worker policing in the European hornet Vespa crabro. Insectes Soc 49:41-44 Foster KR, Shaulsky G, Strassmann JE, Queller DC, Thompson CRL (2004) Pleiotropy as a mechanism to stabilize cooperation. Nature 431:693-696 Foster KR, Wenseleers T, Ratnieks FLW (2006) Kin selection is the key to altruism. Trends Ecol Evol 21:57-60 Frank SA (1995) George Price’s contributions to evolutionary genetics. J Theor Biol 175:373-388 Frank SA (2003) Perspective: repression of competition and the evolution of cooperation. Evolution 57:693-705 Gadau J, Fewell JH (2009) Organization of Insect Societies: From Genes to Sociocomplexity. Harvard University Press, Cambridge/MA Gardner A, Foster KR (2008) The evolution and ecology of cooperation − history and concepts. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 1-36 Gilbert OM, Foster KR, Mehdiabadi NJ, Strassmann JE, Queller DC (2007) High relatedness maintains multicellular cooperation in a social amoeba by controlling cheater mutants. Proc Natl Acad Sci USA 104:8913-8917 Grafen A (1984) Natural selection, kin selection and group selection. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach, 2nd edn. Blackwell, Oxford, pp 62-84 Grafen A (2006) Various remarks on Lehmann and Keller’s article. J Evol Biol 19:1397-1399 Grüter C, Farina WM (2009) The honeybee waggle dance: can we follow the steps? Trends Ecol Evol 24:242-247 Gutz H, Leslie JF (1976) Gene conversion: a hitherto overlooked parameter in population-genetics. Genetics 83:861-866 Hamilton WD (1963) The evolution of altruistic behavior. Am Nat 97:354-356

206

Judith Korb

Hamilton WD (1964) The genetical evolution of social behaviour, I and II. J Theor Biol 7:1-52 Hamilton WD (1975) Innate social aptitudes of man: an approach from evolutionary genetics. In: Fox R (ed) Biosocial Anthropology. Malaby, London, pp 133-155 Hamilton WD (1996) Narrow Roads of Gene Land. WH Freeman, Oxford Hamilton WD, May RM (1977) Dispersal in stable habitats. Nature 269:578-581 Hammond RL, Keller L (2004) Conflict over male parentage in social insects. PLoS Biol 2:e248, doi:10.1371/journal.pbio.0020248 Hansell MH (1984) Animal Architecture and Building Behaviour. Longman, London Hardin G (1968) The tragedy of the commons. Science 162:1243-1248 Hartmann A, Wantia J, Torres JA, Heinze J (2003) Worker policing without genetic conflicts in a clonal ant. Proc Natl Acad Sci USA 100:12836-12840 Heinrich B (1993) The Hot-Blooded Insects. Strategies and Mechanisms of Thermoregulation. Harvard University Press, Cambridge/MA Heinsohn RG (2004) Parental care, load-lightening, and costs. In: Koenig WD, Dickinson JL (eds) Ecology and Evolution of Cooperative Breeding in Birds. Cambridge University Press, Cambridge, pp 67-80 Heisler IL, Damuth J (1987) A method for analyzing selection in hierarchically structured populations. Am Nat 130:582-602 Helenterä H, Sundström L (2007) Worker reproduction in Formica ants. Am Nat 170:E14-E25 Helms Cahan S, Blumstein DT, Sundström L, Liebig J, Griffin A (2002) Social trajectories and the evolution of social behaviour. Oikos 96:206-216 Herron MD, Michod RE (2008) Evolution of complexity in the volvocine algae: transitions in individuality through Darwin’s eye. Evolution 62:436-451 Herron MD, Hackett JD, Aylward FO, Michod RE (2009) Triassic origin and early radiation of multicellular volvocine algae. Proc Natl Acad Sci USA 106:3254-3258 Hoekstra RF (1990) Evolution of uniparental inheritance of cytoplasmic DNA. In: Maynard Smith J, Vida G (eds) Organizational Constraints on the Dynamics of Evolution. Manchester University Press, Manchester, pp 269-278 Hurst LD, Atlan A, Bengtsson BO (1996) Genetic conflicts. Q Rev Biol 71:317364 Johnstone RA (2000) Models of reproductive skew: a review an synthesis. Ethology 106:5-26 Kaib M, Bruinsma O, Leuthold RH (1982) Trail-following in termites: evidence for a multicomponent system. J Chem Ecol 8:1193-1205 Karsai I, Wenzel JW (1998) Productivity, individual-level and colony-level flexibility, and organization of work as consequences of colony size. Proc Natl Acad Sci USA 96:8665-8669 Keller L, Ross KG (1998) Selfish genes: a green beard in the red fire ant. Nature 394:573-575 Kirk DL (1998) Volvox: Molecular-Genetic Origins of Multicellularity and Cellular Differentiation. Cambridge University Press, Cambridge

Social insects, major evolutionary transitions and multilevel selection

207

Koenig WD, Dickinson J (2004) Ecology and Evolution of Cooperative Breeding in Birds. Cambridge University Press, Cambridge Kokko H, Johnstone RA (1999) Social queuing in animal societies: a dynamic model of reproductive skew. Proc R Soc Lond B 266:571-578 Korb J (2003) Thermoregulation and ventilation of termite mounds. Naturwissenschaften 90:212-219 Korb J (2005) Regulation of sexual development in termites: mutilation, pheromonal manipulation or honest signal? Naturwissenschaften 92:45-49 Korb J (2006) Limited food induces nepotism in drywood termites. Biol Lett 2:364-366 Korb J (2007) Workers of a drywood termite do not work. Front Zool 4:e7, doi:10.1186/1742-9994-4-7 Korb J (2008) The ecology of social evolution in termites. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 151-174 Korb J, Aanen DK (2003) The evolution of uniparental transmission of fungal symbionts in fungus-growing termites (Macrotermitinae). Behav Ecol Sociobiol 53:65-71 Korb J, Hartfelder K (2008) Life history and development − a framework for understanding developmental plasticity in lower termites. Biol Rev 83:295-313 Korb J, Heinze J (2008a) Ecology of Social Evolution. Springer, Heidelberg Korb J, Heinze J (2008b) The ecology of social life: a synthesis. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 245-259 Korb J, Katrantzis S (2004) Influence of environmental conditions on the expression of the sexual dispersal phenotype in a lower termite: implications for the evolution of workers in termites. Evol Dev 6:342-352 Korb J, Linsenmair KE (2000) Ventilation of termite mounds: new results require a new model. Behav Ecol 11:486-494 Krause J, Lusseau D, James R (2009) Animal social networks: an introduction. Behav Ecol Sociobiol 63:967-973 Lehmann L, Keller L (2006) The evolution of cooperation and altruism − a general framework and a classification of models. J Evol Biol 19:1365-1376 Lenz M (1985) Is inter- and intraspecific variability of lower termite neotenic number due to adaptive thresholds for neotenic elimination? − considerations from studies on Porotermes adamsoni (Froggatt) (Isoptera: Termopsidae). In: Watson JAL, Okot-Kotber BM, Noirot C (eds) Caste Differentaition in Scoial Insects. Pergamon Press, Oxford, pp 125-146 Lenz M, McMahan EA, Williams ER (1982) Neotenic production in Cryptotermes brevis (Walker): influence of geographical origin, group composition, and maintenance conditions (Isoptera: Kalotermitidae). Insectes Soc 29:148-163 Lenz M, Barrett A, Williams ER (1985) Reproductive strategies in Cryptotermes: neotenic production in indigenous and ‘tramp’ species in Australia (Isoptera: Kalotermitidae). In: Watson JAL, Okot-Kotber BM, Noirot C (eds) Caste Differentiation in Social Insects. Pergamon Press, Oxford, pp 147-164 Lewis D (1941) Male sterility in natural populations of hermaphrodite plants: the equilibrium between females and hermaphrodites to be expected with different types of inheritance. New Phytol 40:56-63

208

Judith Korb

Liebig J, Peeters C, Hölldobler B (1999) Worker policing limits the number of reproductives in a ponerine ant. Proc R Soc Lond B 266:1865-1870 Lindauer M (1954) Temperaturregulierung und Wasserhaushalt im Bienenstaat. Z Vgl Physiol 36:391-432 Lüscher M (1961) Air-conditioned termite nests. Sci Am 205:138-145 Lyon MF (1991) The genetic basis of transmission-ratio distortion and malesterility due to the tert-complex. Am Nat 137:349-358 Margulis L (1981) Symbiosis in Cell Evolution. Freeman and Co, San Fransisco Maynard Smith J (1964) Group selection and kin selection. Nature 201:1145-1147 Maynard Smith J, Szathmáry E (1995) The Major Transitions in Evolution. WH Freeman, Oxford Mehdiabadi N, Jack CN, Farnham TT, Platt TG, Kalla SE, Shaulsky G, Queller DC, Strassmann J (2006) Kin preference in a social microbe. Nature 442:881882 Michener CD (1964) Reproductive efficiency in relation to colony size in hymenopterous societies. Insectes Soc 11:317-342 Michener CD, Brothers DJ (1974) Were workers of eusocial hymenoptera initially altruistic or oppressed. Proc Natl Acad Sci USA 71:671-674 Michod RE (1999) Individuality, immortality, and sex. In: Keller L (ed) Levels of Selection in Evolution. Princeton University Press, Princeton, pp 53-74 Mikheyev AS, Mueller UG, Abbott P (2006) Cryptic sex and many-to-one coevolution in the fungus-growing ant symbiosis. Proc Natl Acad Sci USA 103:10702-10706 Monnin T, Ratnieks FLW (2001) Policing in queenless ponerine ants. Behav Ecol Sociobiol 50:97-108 Murgia C, Pritchard J, Kim S, Fassati A, Weiss R (2006) Clonal origin and evolution of a transmissible cancer. Cell 126:477-487 Nonacs P, Reeve HK (1995) The ecology of cooperation in wasps: causes and consequences of alternative reproductive decisions. Ecology 76:953-967 Okasha S (2006) Evolution and the Levels of Selection. Oxford University Press, Oxford Pearse AM, Swift K (2006) Allograft theory: transmission of devil facial-tumour disease. Nature 439:549 Pepper JW, Smuts BB (2002) A mechanism for the evolution of altruism among nonkin: positive assortment through environmental feedback. Am Nat 160:205-213 Pike N, Foster WA (2008) The ecology of altruism in a clonal insect. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 37-56 Pomiankowski A (1999) Intragenomic conflict. In: Keller L (ed) Levels of Selection in Evolution. Princeton University Press, Princeton, pp 121-152 Price GR (1970) Selection and covariance. Nature 277:520-521 Price GR (1972) Extension of covariance selection mathematics. Annal Hum Genet 35:485-490 Queller DC, Strassmann JE (1998) Kin selection and social insects. BioScience 48:165-175

Social insects, major evolutionary transitions and multilevel selection

209

Queller DC, Foster KR, Fortunato A, Strassmann JE (2003) Cooperation and conflict in the social amoeba, Dictyostelium discoideum. In: Kikuchi T, Azuma N, Higashi M (eds) Genes, Behaviors and Evolution of Social Insects. Hokkaido University Press, Sapporo, Japan, pp 173-200 Ratnieks FLW (1988) Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. Am Nat 132:217-236 Ratnieks FLW (1993) Egg-laying, egg-removal, and ovary development by workers in queenright honey-bee colonies. Behav Ecol Sociobiol 32:191-198 Ratnieks FLW (2001) Heirs and spares: caste conflict and excess queen production in Melipona bees. Behav Ecol Sociobiol 50:467-473 Ratnieks FLW, Visscher PK (1989) Worker policing in the honeybee. Nature 342:796-797 Ratnieks FLW, Wenseleers T (2008) Altruism in insect societies and beyond: voluntary or enforced? Trends Ecol Evol 23:45-52 Ratnieks FLW, Foster KR, Wenseleers T (2006) Conflict resolution in insect societies. Annu Rev Entomol 51:581-608 Reeve HK (1998) Game theory, reproductive skew, and nepotism. In: Dugatkin LA, Reeve HK (eds) Game Theory and Animal Behaviour. Oxford University Press, Oxford, pp 118-145 Reeve HK, Keller L (1999) Levels of selection: burying the units-of-selection debate and unearthing the crucial new issues. In: Keller L (ed) Levels of Selection in Evolution. Princeton University Press, Princeton, pp 3-14 Reeve HK, Peters JM, Nonacs P, Starks PT (1998) Dispersal of first ‘workers’ in social wasps: causes and implications of an alternative reproductive strategy. Proc Natl Acad Sci USA 95:13737-13742 Reyer H-U (1984) Investment and relatedness: a cost/benefit analysis of breeding and helping in the pied kingfisher (Ceryle rudis). Anim Behav 32:1163-1178 Roisin Y (1994) Intragroup conflicts and the evolution of sterile castes in termites. Am Nat 143:751-765 Roisin Y (2000) Diversity and evolution of caste patterns. In: Abe T, Bignell DE, Higashi M (eds) Termites: Evolution, Sociality, Symbiose, Ecology. Kluwer Academic Publishers, Dordrecht/NL, pp 95-119 Roux EA, Roux M, Korb J (2009) Selection on defensive traits in a sterile caste − caste evolution: a mechanism to overcome life-history trade-offs? Evol Dev 11:80-87 Saigo T, Tsuchida K (2004) Queen and worker policing in monogynous and monandrous colonies of a primitively eusocial wasp. Proc R Soc Lond B 271:S509-S512 Seger J (1991) Cooperation and conflict in social insects. In: Krebs JR, Davies NB (eds) Behavioural Ecology. An Evolutionary Approach. Blackwell, Cambridge/MA, pp 338-373 Sherman PW, Lacey EA, Reeve HK, Keller L (1995) The eusociality continuum. Behav Ecol 6:102-108 Simmons LW, Emlen DJ (2006) Evolutionary trade-off between weapons and testes. Proc Natl Acad Sci USA 103:16346-16351

210

Judith Korb

Sober E, Wilson DS (1997) Unto Others: The Evolution of Altruism. Harvard University Press, Cambridge/MA Strohm E, Liebig J (2008) Why are so many bees but so few digger wasps social? The effect of provisionning mode and helper efficiency on the distribution of sociality among the Apoidea. In: Korb J, Heinze J (eds) Ecology of Social Evolution. Springer, Heidelberg, pp 109-128 Stubblefield JW, Charnov EL (1986) Some conceptual issues in the origin of eusociality. Heredity 57:181-187 Szathmáry E (2006) The origin of replicators and reproducers. Philos Trans R Soc Lond USA 361:1761-1776 Temin RG, Ganetzky B, Powers PA, Lyttle TW, Pimpinelli S, Dimitri P, Wu C-I, Hiraizumi Y (1991) Segregation distortion in Drosophila melanogaster: genetic and molecular analyses. Am Nat 137:287-331 Tsuji K (1995) Reproductive conflicts and levels of selection in the ant Pristomyrmex pungens: contextual analysis and partitioning of covariance. Am Nat 146:586-607 Turner JS (2000) The Extended Organism: The Physiology of Animal-Built Structures. Harvard University Press, Cambridge/MA Turner BC, Perkins DD (1991) Meiotic drive in Neurospora and other fungi. Am Nat 137:416-429 van der Gaag M, Debets AJM, Oosterhof J, Slakhorst M, Thijssen JAGM, Hoekstra RF (2000) Spore-killing meiotic drive factors in a natural population of the fungus Podospora anserina. Genetics 156:593-605 von Frisch K (1946) Die Tänze der Bienen. Österr Zool Z 1:1-48 von Frisch K (1967) The Dance Language and Orientation of Bees. Harvard University Press, Cambridge/MA Wade MJ (1978) Kin selection: a classical approach and a general solution. Proc Natl Acad Sci USA 75:6154-6158 Wenseleers T, Ratnieks FLW (2004) Tragedy of the commons in Melipona bees. Proc R Soc Lond B 271:S310-S312 Wenseleers T, Ratnieks FLW (2006) Enforced altruism in insect societies. Nature 444:50 Wenseleers T, Ratnieks FLW, Billen J (2003) Caste fate conflict in swarmfounding social Hymenoptera: an inclusive fitness analysis. J Evol Biol 16:647-658 Wenseleers T, Hart AG, Ratnieks FLW (2004a) When resistance is useless: policing and the evolution of reproductive acquiescence in insect societies. Am Nat 164:E154-E167 Wenseleers T, Hart AG, Ratnieks FLW, Quezada-Euan JJG (2004b) Queen execution and caste conflict in the stingless bee Melipona beecheii. Ethology 110:725-736 Wenseleers T, Helanterä H, Hart A, Ratnieks FLW (2004c) Worker reproduction and policing in insect societies: an ESS analysis. J Evol Biol 17:1035-1047 Wenseleers T, Gardner A, Foster KR (in press) Social evolution theory: a review of methods and approaches. In: Székely T, Komdeur J, Moore AJ (eds) Social

Social insects, major evolutionary transitions and multilevel selection

211

Behaviour: Genes, Ecology and Evolution. Cambridge University Press, Cambridge Wenzel J (1991) Evolution of nest architecture. In: Ross KG, Matthews RW (eds) The Social Biology of Wasps. Cornell University Press, Cornell, Ithaka, pp 480-519 West SA, Griffin AS, Gardner A (2007) Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. J Evol Biol 20:415-432 West-Eberhard MJ (1975) The evolution of social behavior by kin selection. Q Rev Biol 50:1-33 Wey T, Blumstein DT, Shen W, Jordán F (2008) Social network analysis of animal behaviour: a promising tool for the study of sociality. Anim Behav 75:333-344 Williams GC (1966) Adaptation and Natural Selection: A Critique of Some Current Evolutionary Thoughts. Princeton University Press, Princeton Wilson EO (1971) The Insect Societies. Harvard University Press, Cambridge/MA Wilson DS (1975) A theory of group selection. Proc Natl Acad Sci USA 72:143146 Wilson DS (1997) Altruism and organism: disentangling the themes of multilevel selection theory. Am Nat 150:S122-134 Wilson DS (2008) Social semantics: toward a genuine pluralism in the study of social behaviour. J Evol Biol 21:368-373 Wilson DS, Dugatkin LA (1997) Group selection and assortative interactions. Am Nat 149:336-351 Wilson EO, Hölldobler B (2005) Eusociality: origin and consequences. Proc Natl Acad Sci USA 102:13367-13371 Wynne-Edwards VC (1962) Animal Dispersion in Relation to Social Behavior. Oliver and Boyd, Edinburgh Yang R-C (1998) Estimating hierarchical F-statistics. Evolution 52:950-956 Zimmerman RB (1983) Sibling manipulation and indirect fitness in termites. Behav Ecol Sociobiol 12:143-145

Chapter 8

Cooperation between unrelated individuals – a game theoretic approach REDOUAN BSHARY

ABSTRACT Cooperation between unrelated individuals has attracted a lot of research interest because the acts of mutual helping have to be reconciled with evolutionary theory and its focus on individual benefits. Various theoretical frameworks exist, and this chapter will focus specifically on the game theoretic approach. Game theory captures key ecological and life history features like longevity, mutual dependency or mobility in simple games in order to explore conditions that allow cooperative solutions. As behaviour is embedded into underlying strategies, game theory is particularly suited to investigate the conditionality of cooperative behaviour. A great variety of game theoretic concepts offer explanations why cooperation between two individuals may exist in nature. By providing selected examples and describing in detail a case study on marine cleaning mutualism I will argue that the variety of concepts is indeed necessary to capture the diversity of documented examples. In the next part of the chapter, concepts and evidence for cooperation in groups is presented. The focus is on the ‘tragedy of the commons’ and the conditions under which humans may solve the tragedy. At the end and in two boxes I highlight key topics of debate that warrant future research.

8.1 Introduction Natural selection favours behaviours that yield individuals on average a higher fitness than the population mean for alternative behaviours. Therefore, an intuitive reflection would suggest that any form of helping behaviour – a behaviour that increases the fitness of another individual – should be under negative selection. However, helping behaviours are frequently observed in nature, both unilateral helping where the helper reduces its

214

Redouan Bshary

own direct fitness (termed ‘altruism‘) and mutual helping where partners increase their direct fitness (termed ‘cooperation‘ within species and ‘mutualism’ between species, see BOX 8.1 for definitions of key terms). The widespread existence of helping offers scientists the wonderful challenge to explain helping within the framework of evolutionary theory. Hamilton’s theory of kin selection (Hamilton 1964) provides the basic explanation for altruistic behaviours, i.e. behaviours that reduce the lifetime direct fitness of the actor: altruism may evolve if the costs of helping (c) are outweighed by the benefits for the receiver (b) times the degree of genetic relatedness (r) between helper and receiver (r x b > c). Otherwise, altruistic behaviour is under negative selection. Altruism towards unrelated individuals is thus typically under negative selection (there is the possible exception that the recipient compensates a helper’s direct fitness loss by helping the helper’s relatives in return). In contrast to the straightforward conditions for altruism to evolve, explanations for the evolution of cooperative behaviour are diverse. The aim of this chapter is both to introduce and defend the diversity of cooperation concepts. The key challenge for the understanding of cooperative behaviour is to explain how a cooperative individual may assure that the partner(s) cooperates as well. The basic problem is best illustrated with the well-known prisoner’s dilemma game. In this game two players can either cooperate or defect. There are thus four possible behavioural combinations, leading to different payoff distributions between the two players. In the standard payoff matrix, both players receive three units for mutual cooperative behaviour and one unit for mutual defective behaviour. If one cooperates and one defects, the cooperator receives zero units and the cheater five units. Mutual cooperation is therefore better than mutual defection but a player receives more units if he defects than if he cooperates, independently of the partner’s behaviour: five units instead of three if the partner cooperates and one unit instead of zero if the partner defects. In other words, cooperative behaviour in the prisoner’s dilemma game is an investment: it reduces the actor’s immediate payoffs and increases the recipient’s immediate payoff. Any investment must yield more than compensating future benefits or there is selection against it. Therefore, it does not pay to cooperate in a single-round prisoner’s dilemma game, whereas defection is the only evolutionarily stable strategy (ESS, Maynard Smith 1982). The question of how conditional investments, and hence stable cooperation, may evolve in an iterated version of the prisoner’s dilemma game (Axelrod and Hamilton 1981, further references in Dugatkin 1997) has long distracted empiricists and theoreticians from the fact that the prisoner’s dilemma framework is just one out of many possible scenarios that describe the setting for cooperation in nature. In fact, already in the 1990s

Cooperation between unrelated individuals – a game theoretic approach

215

BOX 8.1 Terminology Helping: a behaviour that on average increases the direct fitness of a recipient and that has been under positive selection at least in part because of this positive effect on the recipient. Helping comprises both altruistic and cooperative behaviour. Altruistic behaviour: a behaviour that on average increases the direct fitness of a recipient and which reduces the direct fitness of the actor. Therefore, altruism is under positive selection only if it increases the actor’s inclusive fitness because of indirect fitness benefits. Note that the definition is about lifetime reproductive success. In contrast, social scientists define altruism in a proximate way, namely as other-regarding behaviour. Similarly, Trivers (1971) defines altruism as a short-term cost while he assumes that ultimately the act will increase the actor’s lifetime direct fitness. I use the term ‘investment‘ for such behaviours to avoid the confusion between short-term and lifetime effects. Cooperative behaviour: a behaviour that on average increases the direct fitness of a recipient and which on average also yields direct fitness benefits to the actor. The ‘on average’ is important because in many cases, for example in cooperative hunting, cooperating does not always yield success. Furthermore, West et al. (2007a) pointed out that cooperative behaviour should at least in part be selected because of the benefits to the recipient. This addition allows us to exclude cases like an elephant self-servingly producing dung and thereby benefiting dung beetles. Cooperation: the outcome of an interaction (or repeated interactions) where all participants on average increase their direct fitness. The distinction between individual behaviours and the outcome of interactions is essential for a game theoretic approach, which focuses on the conditionality of cooperative behaviour. Investment: a behaviour that increases the immediate payoff of a recipient and decreases the immediate payoff of the actor. Self-serving behaviour: a behaviour that increases the actor’s direct fitness independently of what the partner does. The effect on the partner’s fitness can be positive or negative. Public good: a public good is created if a helping behaviour benefits two or more recipients in addition to potential benefits to the actor. Tragedy of the commons (n-player prisoner’s dilemma): this phenomenon occurs if a contribution to a public good is an investment, which normally implies that the contribution is altruistic.

several authors (Dugatkin and Wilson 1991, Noë et al. 1991, Connor 1995) made the point that most examples of cooperation do not present solutions to an iterated prisoner’s dilemma game, and Bergstrom et al. (2003) con-

216

Redouan Bshary

cluded that the game does not capture the essential features of any case of interspecific mutualism. The iterated prisoner’s dilemma is only one of many frameworks that address cooperative behaviour in pairs, and further concepts are needed if cooperation takes place in groups of more than two individuals. In this chapter, I will first introduce the game theoretic approach to studying cooperation. I will then introduce the variety of game theoretic concepts that have been put forward in order to explain stable cooperation in pairs. The chapter will end with a presentation of concepts that may explain cooperation in n-player interactions. The concepts will be illustrated with selected empirical examples. In two separate boxes, I briefly introduce two highly contentious topics as additional food for thought for those who are interested in studying cooperation.

8.2 The game theoretic approach 8.2.1 Conditionality of behaviour: the key feature of the game theoretic approach Three approaches to studying cooperation are prominent in the biological literature: social evolution theory, ecological modelling, and game theory. The key difference between evolutionary game theory and the other two fields is that social evolution theory and ecological theory study the conditions under which cooperative behaviour is under positive selection, while evolutionary game theory embeds cooperative behaviour into underlying strategies and decision rules (Maynard Smith and Price 1973, Maynard Smith 1982). As long as behaviour is unconditional, behaviour and strategy mean the same thing, and the three approaches converge. The general question then is which ecological or life history variables promote unconditional cooperative behaviour. In contrast, if cooperative behaviour is just one option out of the behavioural repertoire of an individual for a given situation, only the game theoretic approach with its emphasis on strategies and decision rules can grasp the conditionality of cooperative behaviour. The question then is not any more whether cooperative behaviour is under positive selection but which decision rule specifying the conditions under which the individual behaves cooperatively is under positive selection. As a consequence of the conditionality of cooperative behaviour, the very same individual may sometimes cooperate and sometimes defect. Because of such flexible behaviour it is essential for the game theoretic approach to distinguish between the behaviour of an individual and the outcome of an interaction. Only if both players cooperate can one speak of cooperation (Dugatkin 1997, see BOX 8.1). Scientists embedded in social evolution

Cooperation between unrelated individuals – a game theoretic approach

217

theory do not distinguish between cooperation and cooperative behaviour (West et al. 2007a). 8.2.2 The methods of the game theoretic approach The behaviour of individuals in real-life interactions may be influenced by many variables, both proximately and due to evolutionary processes. These variables include aspects of the life history of a species, such as longevity, migration patterns, or the degree of overlap between generations. Other aspects concern the stakes and the degree of dependency on cooperation to achieve the goals. Furthermore, behaviour may be conditional on the identity of the partner, past experience, body condition, or aspects of the situation: number of previous encounters, likelihood of future interactions, presence or absence of bystanders, etc. The game theoretic approach attempts to capture these aspects in a synthetic way. Life history variables are translated into the probability of repeated interactions, and the stakes and dependencies are translated into a payoff matrix. The aspects of conditionality may influence behaviour through the likely number of interactions, the payoff matrix, and the partners’ past behaviour. The challenge is to identify the few key variables where the combination of parameter states of these key variables can be said to adequately describe the game structure. Bshary and Bronstein (2004) compiled an extensive list of key variables that may influence cooperation and mutualism. Among this list, the following four variables are probably the most essential ones: 1) Investment: is the act of helping an investment or does it provide immediate benefits to the actor independently of how partners behave? If the behaviour is an investment, further questions arise: 2) Payoff relations between ‘C’ (cooperating) and ‘D’ (defecting): if the investment yields foreseeable benefits because the recipient(s) will use the investment in a self-serving way which helps the investor as a by-product, unconditional investment will be selected (‘C > D’). If the benefits of an investment depend on return investments, then the highest payoff would be achieved by defecting a cooperative partner (‘D > C’). Under these circumstances, only conditional strategies (like ‘cooperate as long as the partner(s) cooperate(s), defect if the partner(s) defect’) may yield stable cooperation. 3) Number of interactions: cooperative behaviour may more readily evolve if partners interact often with each other rather than just a few times or even only once. 4) Partner choice: the possibility to exert choice may be a very powerful mechanism to prevent cheating because cheaters risk to be left alone or they will/may end up matched with each other.

218

Redouan Bshary

The alternative parameter states of the four variables can be combined freely to create different game structures. Theoreticians and empiricists can then explore which game structures yield cooperative outcomes and to assess how (with what partner control mechanisms) individuals can reduce the payoffs of defectors to a level that cooperative behaviour is selected for. Most models actually specify the conditions under which a cooperative strategy can be an evolutionarily stable strategy (ESS, Maynard Smith 1982). An ESS yields the highest payoff of all strategies considered in a population where everybody uses the strategy, which makes the ESS resistant to invasion by a rare mutant strategy. Identifying the game structure and cooperative ESS for the iterated prisoner’s dilemma game, one can note that cooperative behaviour is an investment because it reduces the actor’s immediate payoff independently of how the partner behaves. Also, it is clear that ‘D > C’ because a player’s ideal solution (yielding the highest payoff) is to always cheat an always cooperating partner. Furthermore, partners interact repeatedly (200 times in Axelrod and Hamilton 1981) and partner choice does not exist. For this specific parameter combination strategies that are based on ‘positive reciprocity‘ (Clutton-Brock and Parker 1995) – strategies that make individuals generally reward cooperative behaviours and as controlling mechanism defect in response to cheating – provide cooperative ESS (Nowak and Sigmund 1992, 1993; discussed in detail in Dugatkin 1997). If we change one of the four parameter states, we obtain a different game structure and therefore also different strategies with different control mechanisms that may emerge as cooperative solutions, as developed below (Sect. 8.3). 8.2.3 On the importance of the game theoretic approach The strength of game theory is that behaviour (loosely defined, so including physiological responses in plants or bacteria) is embedded in a strategy with decision rules that specify the conditions under which an individual cooperates, cheats, or responds to a partner’s behaviour in an appropriate way. If one assumes that the behaviour rather than the decision rules is under selection, game theory is of minor importance. This is because game theoreticians would capture the conditions with a payoff matrix that offers higher payoffs for cooperating than for defecting, making the answer trivial. In contrast, asking which ecological or life history parameters cause such a payoff matrix is very interesting. A long list of theoretical work has identified potential key variables like low migration, longevity, overlapping generations, or interdependency between group members (Nowak and May 1992, van Baalen and Rand 1998, Killingback et al. 1999, Kokko and

Cooperation between unrelated individuals – a game theoretic approach

219

Johnstone 1999, Taylor and Irwin 2000, Aviles 2002, further references in Lehmann and Keller 2006). The issue of unconditional helping will be a major focus of the part on n-player cooperation. Forms of helping that are not conditional on a partner’s behaviour are clearly widespread in nature: kin selection in the case of altruism, byproduct mutualism and positive pseudoreciprocity in the case of cooperation. They all have in common that as long as basic conditions are fulfilled (like Hamilton’s inequality in the case of altruism) cheating is never a profitable option. In by-product mutualism the act of helping provides benefits to the actor independently of how partners behave. Thus, the act of helping is a self-serving behaviour and the benefits to others a mere byproduct. Many examples of cooperation fit this scenario: a classic case is hunting in golden jackals, where individuals had a six-fold increase in hunting success in pairs, compared by singletons hunts (Lamprecht 1978). With such an increase in success probability, the best response to a partner hunting is to join the hunt (to cooperate) rather than to defect and gain access to a prey hunted by the partner with a comparatively low probability. Positive pseudoreciprocity is similar to by-product mutualism but differs with respect to the timing of benefits. In by-product mutualism, the act of helping provides immediate net benefits to the actor, whereas the benefits are delayed in positive pseudoreciprocity. Therefore, the act of helping becomes an investment and there is a certain risk that the benefits will never be accrued. However, that would be due to the recipient(s) dying rather than cheating: because the investment in principle enables a recipient to perform self-serving behaviour that benefits the donor as a by-product. From the outline above, it appears that cases of cooperation where helping is conditional on the behaviour of the partner might be relatively rare. However, relative payoffs for cooperating and cheating may change repeatedly during the life of individuals, either because ecological conditions fluctuate or because payoffs may depend on an individual’s state. For example, helping might be beneficial if an individual is in good condition but costly if an individual is in poor condition (Sherratt and Roberts 2001, Lotem et al. 2003). Such variability in payoffs should select for conditional behaviour. In any case, conditional cooperation has attracted great interest in the literature, probably because such cases provide a greater intellectual challenge. A further advantage of the game theoretic approach is that it can incorporate the coexistence of cooperative behaviour and cheating. Real life observations document such variation. Therefore, rather than asking under what conditions cooperative behaviour will become fixed in a population the more realistic question is to ask why variation in cooperative behaviour is maintained. There are several solutions proposed in the literature. One

220

Redouan Bshary

possibility is to assume that individuals make errors and hence do not always cooperate (Nowak and Sigmund 1992, 1993). Alternative scenarios explain variation in cooperative behaviour by including assumptions about condition dependency and frequency dependence (Sherratt and Roberts 2001, Lotem et al. 2003). Finally, McNamara et al. (2004) developed a scenario based on an iterated prisoner’s dilemma game with the assumption that the initial state of a population is variation in levels of cooperation. The result was a distribution of strategies that vary in the degree of cooperativeness.

8.3 Theoretical concepts for cooperation between two individuals The conditions that yield stable cooperation when interactions take place in pairs have been studied in much detail compared to interactions between n individuals. A large variety of theoretical concepts exists. It is often difficult to assess in what respect each concept is unique, or whether different terms actually mean the same thing. To give a non-exhaustive list: byproduct mutualism, pseudoreciprocity, partner choice, sanctions, power, punishment, threat of reciprocity, generalised reciprocity, tit-for-tat like reciprocity, pay-to-stay, indirect reciprocity, building up relationships, group augmentation, strong reciprocity, social prestige, and policing are all concepts that have been proposed to yield stable cooperation. My aim is not to present all these concepts in detail but rather to develop a scheme that logically links the most basic concepts in my view. The concepts differ with respect to the responses to a cheating partner that may reduce the cheater’s fitness, thereby causing selective pressure against cheating. I refer to such responses as ‘partner control mechanisms’. There are several key strategic answers to the basic question ‘why should an individual help another individual?’:  because the behaviour increases the direct fitness of the actor irrespective of the recipient’s actions  because the behaviour is an investment that causes self-serving responses that benefit the investor as a by-product  because the behaviour causes a return investment  because the behaviour causes the absence of negative responses Furthermore, one has to ask who provides the benefits to the actor/refrains from reducing the actor’s payoff. If it is the recipient, the link is said to be

Cooperation between unrelated individuals – a game theoretic approach

221

Terms from the literature

no

yes

By-product mutualism

+

Pseudoreciprocity

–

Sanction, power, partner choice

+

Social prestige

–

Pay-to-stay

+

‘Tit-for-tat’ like reciprocity

–

Punishment

+

Image scoring, GR

–

Policing, ‘strong reciprocity’

direct Pseudoreciprocity no

indirect

Costly response direct

yes Reciprocity

indirect

Fig. 8.1 A decision tree to determine nine basic cooperation concepts by identifying the state of four key parameters: investment (‘yes’ or ‘no’), costly response (‘yes’ or ‘no’), response by the recipient (‘direct’) or by someone else (‘indirect’), controlling behaviour a reward for cooperative behaviour (+) or a reduction of a cheater’s payoff (-). In the column to the right, terms from the literature are given. GR: generalized reciprocity. Adapted from Bshary and Bergmüller 2008.

‘direct’. If the benefits of helping are due to the behaviour of third parties, the link is said to be ‘indirect’. Note that the use of the terms ‘direct’ and ‘indirect’ has nothing to do with inclusive fitness theory: we only consider direct fitness benefits. All the strategic aspects mentioned above can be combined into a simple scheme with four parameters that can be in two different states (see Bshary and Bergmüller 2008). The combination of the states of the parameters explains what concept/partner control mechanism may explain why an individual helps another one. 1) Investment: yes or no. 2) short term costly response: yes or no. 3) Benefits: direct from receiver or indirect from third parties. 4) Benefits: reward (positive) or absence of negative responses. Nine basic concepts can be identified (Fig. 8.1). Several corresponding well known terms from the literature are also presented in Fig. 8.1. The advantage of this classification is that it specifies the similarities and the differences between the basic concepts, and that it allows precise definitions of concepts. To give two examples, punishment is a concept/control mechanism that causes an individual to invest into a recipient in order to avoid a costly response from the recipient that would reduce the individual’s payoff. Punishment is thus negative direct reciprocity (Clutton-Brock

222

Redouan Bshary

and Parker 1995). In contrast, sanctions as defined by Herre et al. (1999) are a form of negative direct pseudoreciprocity: subjects invest because failure to do so would cause self-serving response from the recipient that would reduce the actor’s payoff. The terms listed in Fig. 8.1 may often be used with different meanings in the literature. Therefore, the aim of Fig. 8.1 is not necessarily to define the terms in my favourite way, but to make clear how many terms we need to specify the most basic concepts. Further information about other parameters is needed to distinguish between more specific concepts. For example, partner choice differs from other forms of negative direct pseudoreciprocity by the assumption that an individual’s outside option is to find another partner if the current one does not behave cooperatively. In contrast, sanctions and power assume that the outside option is to simply terminate the interaction and not interact at all.

8.4 Empirical evidence that we need all the basic cooperation concepts for two-player interactions 8.4.1 General assessment Relatively few studies on cooperation between unrelated individuals use a game theoretic approach. Studies on mutualism typically focus on ecological questions rather than on the strategic behaviour of individuals. Therefore, it is currently impossible to rate the relative importance of the basic concepts by counting the number of examples that fit each concept. However, as long as there is at least one convincing example for a concept, its existence is already justified. Table 8.1 summarises convincing examples for all nine basic cooperation concepts for the situation where interactions take place between two individuals. As a general rule, there is considerably less convincing evidence for reciprocity concepts than for pseudoreciprocity concepts, in particular outside humans. It seems likely that the difference in abundance is real. Cooperation based on investments by one partner in order to gain by-product benefits or to avoid negative effects of self-serving behaviour seems to be inherently more stable than cooperation based on mutual investments (Bergstrom et al. 2003, Hammerstein 2003, Bshary and Bronstein 2004). The one important reservation against this conclusion is that it is notoriously difficult to demonstrate for a cooperative behaviour that it is contingent on cooperative behaviour of the partner. Furthermore, even if a contingency has been demonstrated the discussion may continue. A prime example is the preda-

Cooperation between unrelated individuals – a game theoretic approach

223

Table 8.1 Examples and references for the nine basic cooperation concepts where interactions always take place between two players. Concept/ control mechanism

Example

Reference

By-product mutualism

Golden jackals increase hunting success six fold if hunting in pairs rather than alone

Lamprecht 1978

Positive direct pseudoreciprocity

Leafcutter ants invest in farming fungi because the fungi’s self-serving response to grow healthy colonies yields by-product benefits to the ants as the latter harvest fungi for food

Mueller et al. 2005

Negative direct pseudoreciprocity

Leguminose plants stop growing roots in areas where rhizobia fail to fix nitrogen but rather invest in areas where nitrogen is fixed

Kiers et al. 2003

Positive indirect pseudoreciprocity

Cleaners give a better service to clients in the presence of bystanders because the latter self-servingly prefer cooperative cleaners

Bshary and Grutter 2006

Negative indirect pseudoreciprocity

Cichlid helpers contribute to help offspring to avoid self-serving eviction from the territory by the dominant

Bergmüller and Taborsky 2005

Positive direct reciprocity

Sticklebacks inspect predators in pairs where the level of cooperation (approach distance) of one fish depends on the cooperativeness of the partner

Milinski 1987, Milinski et al. 1990, 1997

Negative direct reciprocity

Resident clients chase cleaners in response to cheating, which causes cleaners to be more cooperative in future interactions

Bshary and Grutter 2005

Positive indirect reciprocity

Humans are willing to pay to help individuals they have observed helping others

Wedekind and Milinski 2000

Negative indirect reciprocity

Humans are willing to pay to reduce the payoff of individuals they have observed cheating others

Fehr and Fischbacher 2004

224

Redouan Bshary

tor inspection behaviour shown by various fish species. A close inspection yields information about the predator’s hunger level, which can then be used for one’s own foraging decisions (Häberli et al. 2005). Milinski (1987) demonstrated that single sticklebacks paired with a mirror that suggested the presence of a conspecific approached a predator more closely if the mirror image approached in parallel than if the mirror image lagged behind (due to a different positioning of the mirror). Milinski interpreted the behaviour as consistent with a tit-for-tat strategy during pair inspections. A further experiment clearly demonstrated that a stickleback’s current minimal approach distance is contingent on a partner’s approach behaviour during past joint inspections (Milinski et al. 1990). Most importantly, Milinski et al. (1997) could show that lagging behind in an inspecting pair significantly reduced the laggard’s predation risk at the expense of the leading partner. In the experiment, dead sticklebacks ‘performed’ predator inspections by remote control according to an algorithm that mimicked natural inspection. Despite this unique combination of evidence, colleagues still remained doubtful about the exact payoff matrix and the tit-for-tat interpretation (Hammerstein 2003). Thus, the reciprocity debate in behavioural ecology became similar to the ‘theory of mind’ debate in animal cognition (Heyes 1998): a final decisive proof is very difficult, while the absence of reciprocity or theory of mind is the default hypothesis. Below I will use marine cleaning interactions as a model mutualism complex to illustrate the main point of the first part of the chapter: there are many different game structures in nature and as a consequence many different control mechanisms that ensure that partners typically behave cooperatively. 8.4.2 On the diversity of game theoretic concepts found in marine cleaning mutualism Over 100 fish and shrimp species have been observed to inspect the surface and sometimes even the gills and the mouth of fishes (Côté 2000). These interactions apparently have evolved because the ‘cleaners’ remove ectoparasites from their ‘clients’ (Grutter 1997, Becker and Grutter 2004). There is variation among cleaner species with respect to their dependency on interactions for their diet, and most cleaner fish species only clean as juveniles (Côté 2000). The most likely evolutionary starting point for cleaning mutualism is that small fish that feed on invertebrates living on substrate included other fish species as substrate and thus removed their ectoparasites. This scenario suggests that the starting point was a by-

Cooperation between unrelated individuals – a game theoretic approach

225

product mutualism. By-product mutualism is inherently stable because the behaviours of partners are self-serving. The lack of investments makes cheating either impossible or unprofitable. Caribbean gobies of the genus Elacatinus, where member species are highly dependent on cleaning for their diet, still seem to be engaged in a by-product mutualism with their clients. There is no evidence that these cleaners manipulate client decisions or that clients use particular partner control mechanisms to prevent gobies from cheating (eating mucus instead of ectoparasites; Soares et al. 2008a,b). These observations suggest that cleaning gobies prefer ectoparasites over mucus and therefore invariably prefer to cooperate, which appears to be the case (Soares et al. unpubl. data). Clients then just have to leave in time to avoid excessive mucus feeding by the goby once parasites become rare/absent on their body. The situation is very different in the cleaner wrasse Labroides dimidiatus, the most important cleaner in the Indopacific. Individuals occupy small territories (‘cleaning stations’) and have about 2000 interactions per day with 100-500 clients belonging to about 30-50 species (Grutter 1997, Bshary 2001). Two findings make L. dimidiatus a particular interesting study object. First, they prefer client mucus over ectoparasites (Grutter and Bshary 2003). Therefore, there is a major conflict between cleaner and client over what a cleaner should eat, and clients have to make cleaners eat against their preference in order to receive a good service. Second, these cleaners often give tactile stimulation to clients with their pectoral and in particular pelvic fins, a behaviour that is used to manipulate client decisions and behaviour in various circumstances (Bshary and Würth 2001, Grutter 2004). The cleaners interact with a large diversity of client species, and clients differ with respect to key variables that affect the game structure of cleaner-client interactions. Therefore, the cleaners switch between several game structures during a chain of interactions with various client species. The most fundamental distinction with respect to game structure is to be made between reef fish clients and pelagic clients. Reef fishes have defined territories or home ranges and hence interactions with cleaners are potentially repeated. In contrast, truly pelagic fishes may effectively visit a cleaning station only once and then move on to new areas. Most cooperation concepts rely on repeated interactions, hence the one-off interactions between cleaners and pelagic clients are particularly intriguing. A similar problem appears to exist in the closely related cleaner wrasse L. bicolor, where the cleaners rove over large areas. Roving by potential cheaters hinders stable cooperation (Dugatkin and Wilson 1991), so the challenge is to explain why cooperation nevertheless persists. A possible explanation is that cleaners may largely cooperate in such one-off interactions because a

226

Redouan Bshary

prolonged search for ectoparasites may yield a higher payoff than a short interaction terminated prematurely by clients as a response to cheating by the cleaner (Johnstone and Bshary 2002). Currently, no published test of this concept exists. For the L. dimidiatus system, the client reef fish can be categorised mainly along two criteria: predatory/non-predatory and residents/visitors. Residents have small home ranges and hence access to one cleaning station only, while visitors have larger home ranges and hence access to > 1 cleaning stations. Interactions between cleaners and residents, visitors and predators have been studied in quite some detail. Trivers (1971) used interactions between cleaners and predators as one potential example to explain his concept of reciprocity: he proposed that predatory clients refrain from eating a cleaner that enters their mouth because the short-term benefits of such cheating would be outweighed by the cumulative long-term benefits of cooperating due to the cleaner repeatedly removing parasites. This hypothesis appears to be surprisingly difficult to test; it has not been done so far. Trivers (1971) was not concerned about cheating by cleaners because at that time it was not yet known that they would prefer to eat mucus. So both cleaners and predators have symmetrical behavioural options, i.e. they can cooperate or cheat, and cheating a cooperative partner appears to offer the highest short-term benefit. Despite these conditions that fit a prisoner’s dilemma payoff matrix, titfor-tat like reciprocity (cooperate during the first interaction and then typically cooperate as long as your partner cooperated, Axelrod and Hamilton 1981, Dugatkin 1997) cannot explain the mutualism between cleaners and predators. This is because a cheating predator effectively terminates the game, and hence annihilates the very reason why a tit-for-tat like strategy is effective: if the partner cheats, it risks that one cheats in return in the next interaction (Hammerstein and Hoekstra 1995). A cleaner that has been cheated by a predator cannot cheat by eating mucus during the next interaction because it is dead. The solution seems to be that cleaners have to provide such a good service that the payoff for cooperating for a predator is higher than the payoff for cheating. If a cleaner cheats, it is not a cleaner anymore but a food item. The predator’s ‘threat of reciprocity’ (Bshary and Bronstein 2004), which would be terminal, seems to keep the system stable: cleaners hardly ever cheat predators and provide them tactile stimulation much more frequently than they give tactile stimulation to non-predatory clients (Bshary and Würth 2001), in particular when predators are hungry (Grutter 2004). The vast majority of clients do not predate on fishes the size of a cleaner wrasse but instead feed on plankton, polyps or algae. These clients hence do not have the option to cheat a cleaner and the strategic options become

Cooperation between unrelated individuals – a game theoretic approach

227

asymmetric. One therefore has to ask how these clients make cleaners feed against their preference. Visitors apparently use their choice options as a means to control cleaner fish behaviour: if the service was good they are likely to return for their next inspection, but if the cleaner cheated, they are likely to visit another station (Bshary and Schäffer 2002). Models show that the risk of spending time without a partner, because cheating stops an interaction, may select for cooperative behaviour, in particular if the cheated partner may find a new partner with ease while the cheater does not (Ferrière et al. 2002, Johnstone and Bshary 2008). Such partner choice in a biological market (Noë et al. 1991) is a form of direct negative pseudoreciprocity: potential cheaters opt to cooperate in order to avoid that the partner makes a self-serving decision (it would abandon the cheater because the next partner would on average be more cooperative) that would reduce the cheater’s fitness as a by-product. Resident clients cannot simply switch to another cleaner because there is no alternative for them. Instead of switching, they rely on punishment sensu Clutton-Brock and Parker (1995) to make cleaners behave cooperatively: a cleaner that cheats risks that the client chases it. Chasing is a short-term investment by the client that inflicts energetic and opportunity costs on the cleaner, and which causes them to behave more cooperatively during their next interaction with the punisher (Bshary and Grutter 2002, 2005). In conclusion, cleaners usually cooperate with residents in order to avoid direct negative reciprocity. Residents do not only chase cleaners for cheating but they actually also chase immigrant cleaners (Bshary 2002a). Immigrant cleaners provide residents with a lot of tactile stimulation and largely refrain from inspecting and cheating. Thus, cleaners and residents build up relationships based on mutual investments, and only established relationships appear to yield mutual benefits through parasite removal for the client and food for the cleaner. Cleaners and visitors do not build up relationships in any visible way (Bshary 2002a). Partner switching and punishment typically cause a good service quality but sometimes these mechanisms fail: a minority of cleaners switches to a temporarily biting strategy, excepting predatory clients (Bshary 2002b). The best strategy for clients is to avoid any interactions with these biting cleaners, and indeed both visitors and residents appear to use image scoring as an additional partner control mechanism. They observe any ongoing interaction when they arrive at a station and typically invite for inspection if the observed interaction ends without apparent conflict, and avoid the cleaner if the observed interaction ends with a conflict (the current client flees or chases the cleaner).

228

Redouan Bshary

As a consequence of image scoring by clients, cleaners behave more cooperatively in the presence of such eavesdropping bystanders (McGregor 1993) than when alone with a client (Bshary and Grutter 2006). The scenario fits the social prestige idea (Zahavi 1995, Roberts 1998), which is the equivalent of positive indirect pseudoreciprocity: cleaners cooperate with a current client in order to make bystanding clients take a self-serving decision (invite inspection) that benefits the cleaner as a by-product. Until now we only know of humans showing the more complex form of positive indirect reciprocity based on image scoring (Wedekind and Milinski 2000), where every single act is an investment that is more than compensated for by the investment of bystanders (Nowak and Sigmund 1998). In conclusion, a large diversity of game structures occurs in the cleaning mutualism involving L. dimidiatus, where differences in the clients’ strategic options lead to the use of different mechanisms that contribute to the control of the foraging behaviour of cleaners. A final complication concerning cleaner-client interactions is that cleaners may often inspect the same client in an established pair of a male and typically the largest female of his harem. Such pairs do not stay together for life, but the female will eventually switch sex once she is big enough and become a competitor for the male (Robertson 1973). During pair inspections a client may still respond to cheating by one cleaner with evasive action, which reduces the payoff for the cooperating cleaner as well. A game theoretic model shows that the game between the two cleaners resembles an iterated prisoner’s dilemma for a wide parameter space, and ‘always cheat’ (biting immediately) is always an ESS for the game. In reality, however, clients seem to prefer cleaning stations with a cleaner pair (Bshary and Schäffer 2002), which becomes understandable once one realises that both model and data suggest that pairs provide a better service quality (a lower cheating rate per time unit inspection) than singletons (Bshary et al. 2008). Cleaner pairs thus appear to readily solve a problem cooperatively where ‘always cheat’ would be an ESS. The solution is asymmetric in the sense that females are more cooperative during pair inspections than males are, while the two sexes offer similar service qualities when inspecting alone. Such asymmetries are due to the larger males coerce females into showing particularly cooperative behaviour. Males often chase females if the latter cheated. This causes females to behave more cooperatively during future joint inspections, which in turn benefits not only future clients but also the males because they obtain more food (Raihani et al. 2009). Thus, males use their size advantage to increase their own payoff at the expense of their female partners.

Cooperation between unrelated individuals – a game theoretic approach

Pair inspections: reciprocity based on coercion?

Cleaner prefers ectoparasites



229

Predator: Threat of reciprocity

Cleaner prefers mucus Resident: Punishment Building up relationships Social prestige

All clients: By-product mutualism Visitor: partner switching Social prestige Pelagic client: Power? Social prestige?

Fig. 8.2 A summary of games/partner control mechanisms that have been described in marine cleaning mutualisms.

In summary, marine cleaning mutualism harbours an amazing variety of game structures and corresponding partner control mechanisms. Fig. 8.2 summarises this diversity.

8.5 n-player cooperation; theory and evidence It is generally assumed that n-player games are fundamentally different from two-player games. The essential feature of n-player games is that an individual’s contribution cannot be directed to specific group members or been withheld from some other group members. In other words, cooperative behaviour produces a public good. This critical assumption leads to the following problem for players: suppose you have three players, where two players cooperate and one cheats and the gains are shared equally between them. The cheater will have the highest payoff between the three of them. Unfortunately, a cooperative player that decides to cheat in response in the next round will not only produce a negative effect for the cheater, but also for the cooperative partner: not contributing to the public good affects/harms all group members. I prefer to make a distinction between a public good and the tragedy of the commons (BOX 8.1). I use the former

230

Redouan Bshary

term for the general condition where a cooperative behaviour benefits more than one recipient, where a tragedy of the commons is one possible payoff matrix (see next section). 8.5.1 Two basic forms of public goods It is important to distinguish two basic forms of public goods based on the payoffs that a cooperating individual receives from its contribution, independently of what other partners are doing. First, as assumed in most models, cooperating can be an investment where the individual contribution is larger than the individual returns due to the contribution. Under these conditions, investing is altruistic in evolutionary terminology, which means that only kin selection may explain helping under such circumstances, as long as r x b > c. Between unrelated individuals, the tragedy of the commons arises: the overall payoff would be highest if all contribute, but individual contribution is altruistic and hence selected against, with the result that cooperation breaks down. The question which extensions to the basic game may yield cooperative solutions to the tragedy of the commons has been the key focus in the cooperation literature on human behaviour. In a second payoff matrix for public goods games cooperating is a self-serving act because for each unit of contribution, the cooperating individual receives more than one unit in return. If the returns are immediate, the concept of by-product mutualism applies. If the contribution allows recipients to act in a self-serving way in the future that benefits the investor as a byproduct, then the concept of pseudoreciprocity applies. Thus, by-product mutualism and pseudoreciprocity are concepts that are not limited to 2player interactions but can easily be extended to n-player games. The challenge for empiricists is to adequately describe the payoffs in natural systems. This will be particularly difficult in systems where immediate payoffs of cooperating may fluctuate between self-serving and investment. 8.5.2 Unconditional benefits for contributing to a public good Few empirical examples have explicitly tested the idea that public goods may be based on immediate unconditional benefits (by-product mutualism) or delayed unconditional benefits (pseudoreciprocity). The idea that the benefits of group-living are due to the selfish herd (Hamilton 1971) fits well. Each individual joins the group for self-serving reasons, while all other group members benefit from the increase in dilution effects. Another important concept is group augmentation (Kokko and Johnstone 1999), which is well studied in meerkats (Suricata suricatta). Individual meerkats

Cooperation between unrelated individuals – a game theoretic approach

231

often act as sentinels in exposed places while the rest of the group forages. Detailed analyses demonstrated that individuals take the sentinel position only if they have fed well before and are therefore satiated. Only under these conditions is their best option to watch out for predators, hence increasing the probability of survival for all other group members (CluttonBrock et al. 1999). The sentinels benefit from the behaviour because their inclusive fitness depends crucially on group size. Also, cooperative breeding in meerkats seems to be based on self-serving behaviour. Many individuals contribute without any evidence for mutual control concerning relative contribution, which is again explained with the benefits of group augmentation (Clutton-Brock et al. 2000, Clutton-Brock 2002). A recent model by Sherratt et al. (2009) demonstrated that contributions to a public good can be further stabilised if a contributing individual receives an extra proportion of the benefits created by its contribution. For example, bacteria or spiders may communally digest host cells/prey (West et al. 2007b, Schneider and Bilde 2008), where the production of digestive enzymes constitutes a public good. The model assumes that each individual has priority of access to the food digested by its own enzymes just because of spatial proximity. Under these conditions, contributions to the public good exist under a wider parameter space even though individual contributions correlate negatively with the number of recipients (Sherratt et al. 2009). Both model and reality work without invoking kin selection, though kin selection reduces the conflict: Schneider and Bilde (2008) demonstrated that in social spiders, related individuals digest prey more efficiently than unrelated individuals do. By-product mutualism and pseudoreciprocity in n-player games can be found under a different name in the literature: ‘weak altruism‘ (Wilson 1990). However, as should be clear from above, this term is misleading. An important point to make is that while by-product mutualism and pseudoreciprocity are not exciting concepts for game-theory because cheating is never a profitable option, these concepts deserve plenty of attention. We need to know what ecological factors cause a situation where cooperative behaviour in groups is under positive selection without being contingent on the behaviour of others. The very same models that were mentioned in the section on unconditional helping in pairs may also apply to n-player games (Nowak and May 1992, van Baalen and Rand 1998, Killingback et al. 1999, Kokko and Johnstone 1999, Taylor and Irwin 2000, Aviles 2002, further references in Lehmann and Keller 2006).

232

Redouan Bshary

BOX 8.2 Hot debates in the cooperation literature I Inclusive fitness theory versus group selection/multi-level selection The most basic insight of Hamilton’s inclusive fitness theory is that a behaviour is under positive selection if it increases the actor’s inclusive fitness – the sum of the behaviour’s effect on the direct fitness and the indirect fitness – relative to the population average. Group selection theory measures the relative importance of within-group competition to between-group competition. Supporters of group selection theory argue that group selection is fundamentally different from inclusive fitness theory, while proponents of the latter concept argue that group selection theory and inclusive fitness theory can be transformed mathematically into each other. Therefore, the question is not which concept is right or wrong, but how useful one finds each alternative. My personal view is that inclusive fitness theory is simpler, easier to understand and asks precisely the question I am interested in: how a behaviour benefits an individual, which is the essential unit that makes reproductive decisions. I prefer to think about between-group competition as an ecological parameter that may select for helping among group members. Group selection theory does not answer the fundamental question whether an individual helps because of direct or indirect fitness benefits. Also, if an individual dies, then it is clear that its future fitness is zero, whereas group extinction does not automatically mean that all members died. Some might have migrated or were assimilated in the invading group; in humans, victorious warriors often reproduce with the local women. Group selection offers an interesting way of thinking, if the trait of interest is not simply the sum of individual behaviours, but if individuals adopt group standards. This is the case in cultural group selection. In humans, language, religious beliefs, or norms are better seen as group traits than as individual traits. And such traits can go extinct if a group goes extinct and eventual survivors adopt the trait of the winning group. Note, however, that also a cultural trait is still only under positive selection if it increases on average the inclusive fitness of individuals who express the trait. Some relevant recent literature: Boyd et al. 2003, Gintis et al. 2003, Grafen 2006, Lehmann and Keller 2006, Traulsen and Nowak 2006, Lehmann et al. 2007a, West et al. 2007a, Wilson and Wilson 2007.

8.5.3 Solving the tragedy of the commons Public goods, where individual contributions are investments and the benefits contingent upon the investment of others, have captured most attention in recent research on human cooperative behaviour. This body of research

Cooperation between unrelated individuals – a game theoretic approach

233

was sparked by the obvious discrepancy between the theoretical prediction that social dilemmas cannot be solved and the general perception that humans often manage to cooperate in large groups of unrelated individuals in their daily life. Experiments demonstrated that human bystanders readily accept costs to punish individuals who they observed cheating others, or to reward individuals who helped others (Fehr and Gächter 2002, Rockenbach and Milinski 2006, Egas and Riedl 2008, Gächter et al. 2008). Humans show punishment in response to cheating even in anonymous one-off interactions and such behaviour effectively causes punished individuals to behave more cooperatively in the future with other partners (Fehr and Gächter 2002). These results prompted the development of the cultural group selection concept (Boyd et al. 2003, Gintis et al. 2003), which states that strong competition between groups selected for cultural traits in humans that benefitted group survival. Recent efforts by biologists focussed on the translation of cultural group selection into inclusive fitness theory. Such a translation is important for evolutionary biologists because the original economic experiments (Fehr and Gächter 2002) demonstrated a mechanism (‘altruistic punishment‘) out of place: one can measure the amount of money people received in the experiment as a correlate of fitness and finds that punishment is under negative selection in one-off interactions in groups consisting of unrelated individuals. The analyses demonstrated that punishment in one-off interactions may only be under positive selection if the beneficiaries of the punishment are related to the punisher (Gardner and West 2004, Lehmann et al. 2007b). Alternatively, the mechanism underlying punishment, the stimulation of the punisher’s reward system in the brain (de Quervain et al. 2004), evolved in a different context and cannot be switched off even when humans know that the rational decision would be not to punish in one-off interactions. At the moment, however, the importance of punishment for the stabilisation of cooperation in n-player games is contentious, independently of the number of rounds played (Rockenbach and Milinski 2006, Egas and Riedl 2008, Nikiforakis 2008). One possibility that has not been tested yet is that punishers are feared and that this may yield benefits in other contexts. Along this line of argument, even seemingly needless aggression towards a partner may be theoretically explained with the gain of a reputation of being nasty, which may deter competitors (Johnstone and Bshary 2004). Currently, we only have one clear empirical demonstration of a functional explanation of why humans may solve the tragedy of the commons. The general idea is that a group of humans do engage both in two-player interactions and in public goods games and how they behave in any interaction may have implications for how their partners behave in future inter-

234

Redouan Bshary

actions, across the different types of interactions. In this scenario, contributors to public goods may gain an increase in their image score, and this image score could then yield benefits during two-player interactions. The idea of an image score was introduced by Alexander (1987). Nowak and Sigmund (1998) presented the first model that helping in two player interactions may evolve due to a resulting increase in the helper’s image score and the consequently increased probability that the helper will receive support from image scoring bystanders. The basic idea works well in humans (Wedekind and Milinski 2000). In an extension, the indirect reciprocity game was combined with a social dilemma game (Milinski et al. 2002). The results were that individuals who contributed to the public good received more help in the indirect reciprocity game and left the experiment with more money (= a higher fitness) than individuals who did contribute little to the public good. In an additional experiment, Semmann et al. (2005) demonstrated that investments do not need to be restricted to group members. Instead, participants who donated parts of their gain to charity also gained an increased image score as visible by the increased amount of help they received compared to participants who did not donate to charity. Finally, Sommerfeld et al. (2008) tackled the problem of how individuals acquire the necessary information if one accepts that one cannot observe all interactions where help is needed. They found that gossip is both a reliable indicator of other players’ cooperativeness and that humans readily use gossip information for their decisions on whether or not to help.

8.6 Conclusions and outlook  As cooperation between unrelated individuals occurs in a variety of game structures we need a great diversity of concepts to explain the evolution and persistence of cooperation in nature.  We face major unresolved questions that offer plenty of future research opportunities.  Cooperation theory is currently much more advanced than our empirical knowledge. This might partly be due to the theoreticians’ focus on reciprocity, which is most likely rare compared to pseudoreciprocity and by-product mutualism outside humans. The diversity of existing concepts should nevertheless encourage empiricists to study cooperation problems in their study systems.  Studying cooperation between n players provides a particular challenge for both theoreticians and empiricists, where a profound knowledge of

Cooperation between unrelated individuals – a game theoretic approach

235

the ecology may help to understand the conditions for the evolution and stability of cooperation in groups in nature.  Theoreticians typically focus on quite simple conditions and find quite sophisticated strategies as solutions. During the Hamilton lecture at the ISBE conference 2008, John McNamara proposed that the future might be to model complex worlds and likely find more rule-of-thumb type solutions. Such an approach may indeed offer a major step forward towards a deeper understanding of cooperation in nature. BOX 8.3 Hot debates in the cooperation literature II What is so special about human cooperation? Very few people would contest the notion that human cooperation is far more developed and complex than cooperation in other species: we reciprocate, we give other people an image score depending on how cooperatively they behave towards others, we may punish transgressors without being the victim of cheating and reward cooperators without having been the beneficiary, and we manage to cooperate in very large groups, including conditions that constitute a tragedy of the commons. Disagreement concerns the level at which differences to other animals exist. Some authors proposed that we need to extend evolutionary theory in order to explain human cooperation. The issue returns to the debate on whether cultural group selection can properly be represented with inclusive fitness theory (see BOX 8.2). In my view, the fundamental differences between cooperation in humans and other species are due to differences in underlying mechanisms. Thanks to their large brain, humans can negotiate efficiently using language. In addition, culture, norms, policing, and features of human life seem to have selected for subjective reward mechanisms that enable us to often avoid the pitfalls of immediate benefit maximisation and allow us to show otherregarding behaviours. The subjective reward systems are built on emotions and ultimately physiological reactions in the brain, which could at least in principle, evolve in any species but apparently rarely have. Nevertheless, I always wonder whether a policing ant ‘feels good’ about eating the egg laid by a worker (and how one could test that). Some relevant recent literature: Fehr and Gächter 2002, Milinski et al. 2002, Hammerstein 2003, de Quervain et al. 2004, Lehmann et al. 2007a, Bshary and Bergmüller 2008, Kappeler and Silk 2010.

236

Redouan Bshary

Acknowledgements I thank Manfred Milinski, Dirk Semmann, Peter Kappeler and an anonymous referee for comments. Ana Pinto extracted all the fish silhouettes for Fig. 8.2, where the picture of the goby and its client was taken by Karen Cheney. Funding was provided by the Swiss Science Foundation.

References Alexander RD (1987) The Biology of Moral Systems. Aldine de Gruyter, New York Avilés L (2002) Solving the freeloaders paradox: genetic associations and frequency-dependent selection in the evolution of cooperation among nonrelatives. Proc Natl Acad Sci USA 99:14268-14273 Axelrod R, Hamilton WD (1981) The evolution of cooperation. Science 211:13901396 Becker JHA, Grutter AS (2004) Cleaner shrimp do clean. Coral Reefs 23:515-520 Bergmüller R, Taborsky M (2005) Experimental manipulation of helping in a cooperative breeder: helpers ‘pay-to-stay’ by pre-emptive appeasement. Anim Behav 69:19-28 Bergstrom CT, Bronstein JL, Bshary R, Connor RC, Daly M, Frank SA, Gintis H, Keller L, Leimar O, Noë R, Queller DC (2003) Group report: interspecific mutualism – puzzles and predictions. In: Hammerstein P (ed) Genetic and Cultural Evolution of Cooperation. MIT Press, Cambridge, pp 241-256 Boyd R, Gintis H, Bowles S, Richerson PJ (2003) The evolution of altruistic punishment. Proc Natl Acad Sci USA 100:3531-3535 Brown JL (1983) Cooperation: a biologist’s dilemma. In: Rosenblatt JS (ed) Advances in the Study of Behaviour. Academic Press, New York, pp 1-37 Bshary R (2001) The cleaner fish market. In: Noë R, van Hooff JARAM, Hammerstein P (eds) Economics in Nature: Social Dilemmas, Mate Choice and Biological Markets. Cambridge University Press, Cambridge, pp 146-172 Bshary R (2002a) Building up relationships in asymmetric co-operation games between the cleaner wrasse Labroides dimidiatus and client reef fish. Behav Ecol Sociobiol 52:365-371 Bshary R (2002b) Biting cleaner fish use altruism to deceive image-scoring client fish. Proc R Soc Lond B 269:2087-2093 Bshary R, Bergmüller R (2008) Distinguishing four fundamental approaches to the evolution of helping. J Evol Biol 21:405-420 Bshary R, Bronstein JL (2004) Game structures in mutualistic interactions: what can the evidence tell us about the kind of models we need? Adv Stud Behav 34:59-101 Bshary R, Grutter AS (2002) Asymmetric cheating opportunities and partner control in a cleaner fish mutualism. Anim Behav 63:547-555

Cooperation between unrelated individuals – a game theoretic approach

237

Bshary R, Grutter AS (2005) Punishment and partner switching cause cooperative behaviour in a cleaning mutualism. Biol Lett 1:396-399 Bshary R, Grutter AS (2006) Image scoring and cooperation in a cleaner fish mutualism. Nature 441:975-978 Bshary R, Schäffer D (2002) Choosy reef fish select cleaner fish that provide high-quality service. Anim Behav 63:557-564 Bshary R, Würth M (2001) Cleaner fish Labroides dimidiatus manipulate client reef fish by providing tactile stimulation. Proc R Soc Lond B 268:1495-1501 Bshary R, Grutter AS, Willener AST, Leimar O (2008) Pairs of cooperating cleaner fish provide better service quality than singletons. Nature 455:964-967 Clutton-Brock TH (2002) Breeding together: kin selection and mutualism in cooperative vertebrates. Science 296:69-72 Clutton-Brock TH, Parker GA (1995) Punishment in animal societies. Nature 373:209-216 Clutton-Brock TH, O’Riain MJ, Brotherton PNM, Gaynor D, Kansky R, Griffin AS, Manser MB (1999) Selfish sentinels in cooperative mammals. Science 284:1640-1644 Clutton-Brock TH, Brotherton PNM, O’Riain MJ, Griffin AS, Gaynor D, Sharpe L, Kansky R, Manser MB, McIlrath GM (2000) Individual contributions to babysitting in a cooperative mongoose, Suricata suricatta. Proc R Soc Lond B 267:301-305 Connor RC (1995) Altruism among non-relatives: alternatives to the ‘Prisoner’s Dilemma’. Trends Ecol Evol 10:84-86 Côté IM (2000) Evolution and ecology of cleaning symbioses in the sea. Oceanogr Mar Biol 38:311-355 de Quervain DJ-F, Fischbacher U, Treyer V, Schellhammer M, Schnyder U, Buck A, Fehr E (2004) The neural basis of altruistic punishment. Science 305:12541258 Dugatkin LA (1997) Cooperation Among Animals: An Evolutionary Perspective. Oxford University Press, Oxford Dugatkin LA, Wilson DS (1991) Rover: a strategy for exploiting cooperators in a patchy environment. Am Nat 138:687-701 Dugatkin LA, Mesterton-Gibbons M, Houston AI (1992) Beyond the prisoner’s dilemma: towards models to discriminate among mechanisms of cooperation in nature. Trends Ecol Evol 7:202-205 Egas M, Riedl A (2008) The economics of altruistic punishment and the maintenance of cooperation. Proc R Soc Lond B 275:871-878 Fehr E, Fischbacher U (2004) Third-party punishment and social norms. Evol Hum Behav 25:63-87 Fehr E, Gächter S (2002) Altruistic punishment in humans. Nature 415:137-140 Ferrière R, Bronstein JL, Rinaldi S, Law R, Gauduchon M (2002) Cheating and the evolutionary stability of mutualisms. Proc R Soc Lond B 269:773-780 Gächter S, Renner E, Sefton M (2008) The long-run benefits of punishment. Science 322:1510 Gardner A, West SA (2004) Cooperation and punishment, especially in humans. Am Nat 164:753-764

238

Redouan Bshary

Gintis H, Bowles S, Boyd R, Fehr E (2003) Explaining altruistic behavior in humans. Evol Hum Behav 24:153-172 Grafen A (2006) Optimization of inclusive fitness. J Theor Biol 238:541-563 Grutter AS (1997) Spatiotemporal variation and feeding selectivity in the diet of the cleaner fish Labroides dimidiatus. Copeia 1997:346-355 Grutter AS (2004) Cleaner fish use tactile dancing behavior as a preconflict management strategy. Curr Biol 14:1080-1083 Grutter AS, Bshary R (2003) Cleaner wrasse prefer client mucus: support for partner control mechanisms in cleaning interactions. Proc R Soc Lond B 270:S242-S244 Häberli MA, Aeschlimann PB, Milinski M (2005) Sticklebacks benefit from closer predator inspection: an experimental test of risk assessment. Ethol Ecol Evol 17:249-259 Hamilton WD (1964) The genetical evolution of social behaviour, I and II. J Theor Biol 7:1-52 Hamilton WD (1971) Geometry for the selfish herd. J Theor Biol 31:295-311 Hammerstein P (2003) Why is reciprocity so rare in social animals? A protestant appeal. In: Hammerstein P (ed) Genetic and Cultural Evolution of Cooperation. MIT Press, Cambridge, pp 83-93 Hammerstein P, Hoekstra RF (1995) Mutualism on the move. Nature 376:121-122 Herre EA, Knowlton N, Mueller UG, Rehner SA (1999) The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends Ecol Evol 14:49-53 Heyes CM (1998) Theory of mind in nonhuman primates. Behav Brain Sci 21:101-134 Johnstone RA, Bshary R (2002) From parasitism to mutualism: partner control in asymmetric interactions. Ecol Lett 5:634-639 Johnstone RA, Bshary R (2004) The evolution of spite through indirect reciprocity. Proc R Soc Lond B 271:1917-1922 Johnstone RA, Bshary R (2008) Mutualism, market effects and partner control. J Evol Biol 21:879-888 Kappeler PM, Silk JB (2010) Mind the Gap: Tracing the Origins of Human Universals. Springer, Heidelberg Kiers ET, Rousseau RA, West SA, Denison RF (2003) Host sanctions and the legume-rhizobium mutualism. Nature 425:78-81 Killingback T, Doebeli M, Knowlton N (1999) Variable investment, the continuous prisoner’s dilemma, and the origin of cooperation. Proc R Soc Lond B 266:1723-1728 Kokko H, Johnstone RA (1999) Social queuing in animal societies: a dynamic model of reproductive skew. Proc R Soc Lond B 266:571-578 Lamprecht J (1978) The relationship between food competition and foraging group size in some larger carnivores. A hypothesis. Z Tierpsychol 46:337-343 Lehmann L, Keller L (2006) The evolution of cooperation and altruism – a general framework and a classification of models. J Evol Biol 19:1365-1376 Lehmann L, Keller L, West S, Roze D (2007a) Group selection and kin selection: two concepts but one process. Proc Natl Acad Sci USA 104:6736-6739

Cooperation between unrelated individuals – a game theoretic approach

239

Lehmann L, Rousset F, Roze D, Keller L (2007b) Strong reciprocity or strong ferocity? A population genetic view of the evolution of altruistic punishment. Am Nat 170:21-36 Lotem A, Fishman MA, Stone L (2003) From reciprocity to unconditional altruism through signalling benefits. Proc R Soc Lond B 270:199-205 Luce RD, Raiffa H (1957) Games and Decisions: Introduction and Critical Survey. Wiley, New York Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge Maynard Smith J, Price G (1973) The logic of animal conflict. Nature 246:15-18 McGregor PK (1993) Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philos Trans R Soc Lond B 340:237-244 McNamara JM, Barta Z, Houston AI (2004) Variation in behaviour promotes cooperation in the Prisoner’s Dilemma game. Nature 428:745-748 Milinski M (1987) Tit for tat in sticklebacks and the evolution of cooperation. Nature 325:433-435 Milinski M, Külling D, Kettler R (1990) Tit for tat: sticklebacks (Gasterosteus aculeatus) ‘trusting’ a cooperating partner. Behav Ecol 1:7-11 Milinski M, Lüthi JH, Eggler R, Parker GA (1997) Cooperation under predation risk: experiments on costs and benefits. Proc R Soc Lond B 264:831-837 Milinski M, Semmann D, Krambeck H-J (2002) Reputation helps solve the ‘tragedy of the commons’. Nature 415:424-426 Mueller UG, Gerardo NM, Aanen DK, Six DL, Schultz TR (2005) The evolution of agriculture in insects. Annu Rev Ecol Evol Syst 36:563-595 Nikiforakis N (2008) Punishment and counter-punishment in public good games: can we really govern ourselves? J Publ Econ 92:91-112 Noë R, van Schaik CP, van Hooff JARAM (1991) The market effect: an explanation for pay-off asymmetries among collaborating animals. Ethology 87:97118 Nowak MA, May RM (1992) Evolutionary games and spatial chaos. Nature 359:826-829 Nowak MA, Sigmund K (1992) Tit for tat in heterogeneous populations. Nature 355:250-253 Nowak MA, Sigmund K (1993) A strategy of win-stay, lose-shift that outperforms tit-for-tat in Prisoner’s Dilemma game. Nature 364:56-58 Nowak MA, Sigmund K (1998) Evolution of indirect reciprocity by image scoring. Nature 393:573-577 Raihani NJ, Grutter AS, Bshary R (2009) Punishers benefit from third-party punishment in fish. Science 327:171 Roberts G (1998) Competitive altruism: from reciprocity to the handicap principle. Proc R Soc Lond B 265:427-431 Robertson DR (1973) Sex change under the waves. New Sci 58:538-540 Rockenbach B, Milinski M (2006) The efficient interaction of indirect reciprocity and costly punishment. Nature 444:718-723

240

Redouan Bshary

Schneider JM, Bilde T (2008) Benefits of cooperation with genetic kin in a subsocial spider. Proc Natl Acad Sci USA 105:10843-10846 Semmann D, Krambeck H-J, Milinski M (2005) Reputation is valuable within and outside one’s own social group. Behav Ecol Sociobiol 57:611-616 Sherratt TN, Roberts G (2001) The importance of phenotypic defectors in stabilizing reciprocal altruism. Behav Ecol 12:313-317 Sherratt TN, Roberts G, Kassen R (2009) Evolutionary stable investment in products that confer both an individual benefit and a public good. Front Biosci 14:4557-4564 Soares MC, Bshary R, Cardoso SC, Côté IM (2008a) The meaning of jolts by fish clients of cleaning gobies. Ethology 114:209-214 Soares MC, Côté IM, Cardoso SC, Bshary R (2008b) On the absence of punishment, partner switching and tactile stimulation in the cleaning goby – client mutualism. J Zool Lond 276:306-312 Sommerfeld RD, Krambeck K-J, Milinski M (2008) Multiple gossip statements and their effect on reputation and trustworthiness. Proc R Soc Lond B 275:2529-2536 Taylor PD, Irwin AJ (2000) Overlapping generations can promote altruistic behavior. Evolution 54:1135-1141 Traulsen A, Nowak MA (2006) Evolution of cooperation by multilevel selection. Proc Natl Acad Sci USA 103:10952-10955 Trivers RL (1971) The evolution of reciprocal altruism. Q Rev Biol 46:35-57 van Baalen M, Rand DA (1998) The unit of selection in viscous populations and the evolution of altruism. J Theor Biol 193:631-648 Wedekind C, Milinski M (2000) Cooperation through image scoring in humans. Science 288:850-852 West SA, Griffin AS, Gardner A (2007a) Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. J Evol Biol 20:415-432 West SA, Diggle SP, Buckling A, Gardner A, Griffin AS (2007b) The social lives of microbes. Annu Rev Ecol Evol Syst 38:53-77 Wilson DS (1990) Weak altruism, strong group selection. Oikos 59:135-140 Wilson DS, Wilson EO (2007) Rethinking the theoretical foundation of sociobiology. Q Rev Biol 82:327-348 Zahavi A (1995) Altruism as a handicap – the limitations of kin selection and reciprocity. J Avian Biol 26:1-3

Chapter 9

Group decision-making in animal societies GERALD KERTH

ABSTRACT Individuals need to coordinate their activities to benefit from group living. Thus group decisions are essential for societies, especially if group members cooperate with each other. Models show that shared (democratic) decisions outperform unshared (despotic) decisions, even if individuals disagree about actions. This is surprising as in most other contexts, differences in individual preferences lead to sex-, age-, or kin-specific behaviour. Empirical studies testing the predictions of the theoretical models have only recently begun to emerge. This applies particularly to group decisions in fission-fusion societies, where individuals can avoid decisions that are not in their interest. After outlining the basic ideas and theoretical models on group decision-making I focus on the available empirical studies. Originally most of the relevant studies have been on social insects and fish but recently an increasing number of studies on mammals and birds have been published, including some that deal with wild long-lived animals living in complex societies. This includes societies where group members have different interests, as in most mammals, and which have been less studied compared to eusocial insects that normally have no conflict among their colony members about what to do. I investigate whether the same decision rules apply in societies with conflict and without conflict, and outline open questions that remain to be studied. The chapter concludes with a synthesis on what is known about group decision-making in animals and an outlook on what I think should be done to answer the open questions.

242

Gerald Kerth

9.1 Why are group decisions important in animal societies? Group decisions are important in all human and non-human societies (stable individualised groups) because group-living requires individuals to coordinate their activities (Bonabeau et al. 1997, Seeley and Buhrman 1999, Franks et al. 2002, Conradt and Roper 2003, 2005, List 2004, Conradt and List 2009). Group decisions are needed whenever group coordination is beneficial but individuals can choose between alternatives. Group deciTable 9.1 Different forms of animal sociality

Forms of animal sociality Aggregation

Anonymous assemblage of individuals. Aggregations are generally less stable than societies. Animals in aggregations show no social bonds and do not engage in cooperative or other affiliate social interactions. Examples include many fish swarms, large flocks of migrating birds that meet at night roosts, and other transient animal groups.

Society

Group of individuals that show social bonds, that may engage in cooperative behaviours, and that recognise each other individually or as members of the same society. Despite the occurrence of social bonds and cooperation, in societies with a heterogeneous demographic and genetic structure individuals are likely to differ in their interests because of individual differences in age, dominance, relatedness, reproductive status, or sex. Examples for heterogeneous societies from mammals include troops of primates, bands of many social carnivores, and colonies of some bat species.

Fission-fusion society

Group that shows all characteristics of a society – social bonds, cooperation, group or individual recognition – but which regularly splits into smaller subgroups (fission), which later merge again (fusion). Fission-fusion societies are found in elephants, dolphins and some other cetaceans, in some primates, in a few social carnivores, in some ungulates, and in several bat species.

Group decision-making in animal societies

243

sion-making occurs when an action by one, several, or all group members affects the behaviour of the entire group. Group-living animals have to communally decide when and where to move, where and on what to feed, and where to rest, roost, or nest (Byrne 2000, Franks et al. 2002, Conradt and Roper 2005, Seeley et al. 2006). Group decisions also occur in more or less transient aggregations of animals, such as fish swarms or large flocks of migrating birds (Conradt and List 2009). However, they are particularly important in stable societies where the fitness of individuals strongly depends on the behaviour of their group members, for instance because they show cooperative behaviour. Eusocial insects for example, depend on group decisions for the functioning of their colonies during foraging and nest-site choice (Seeley et al. 1991, 2006, Franks et al. 2002). Group decision-making may also be necessary to coordinate reproduction of group members where individuals benefit from synchronised breeding via reduced predation or infanticide risk, and from energetic benefits through communal warming (Conradt and Roper 2005). Consequently, information on how animals make group decisions has significant implications for our understanding of how their societies function, particularly with respect to cooperation and group stability. To understand decision-making in animal groups, we need to determine who makes the decisions, what factors affect the way in which they are made, and what decision-making processes are optimal under different circumstances (Conradt and Roper 2005). Before I introduce several of the models and empirical studies on group coordination in animals, it is necessary to define some of the terms used in this chapter. Table 9.1 defines three forms of animal sociality, while Table 9.2 defines the most important terms used to describe group decision-making in animals. Table 9.2 Definitions of terms used to describe group decision-making within groups of animals (based on Conradt and Roper 2005 and Conradt and List 2009)

Definitions Consensus

Binding agreement among group members on a decision.

Consensus cost

Cost to an individual group member that arises from not taking an action that would be optimal for the individual and instead complying with the consensus decision of the group.

244

Gerald Kerth

Table 9.2 Continued Coordination

Outcome of a group decision; can involve behaviour (e.g. joint travelling or communal nesting/roosting) or physiological states (e.g. synchronisation of reproduction or torpor) of a group.

Grouping benefits

Benefits that arise from staying in a group compared to a situation where an individual is on its own. Examples include energetic benefits from social thermoregulation or safety from predators from increased group vigilance.

Shared decisions

Decisions made by most or all group members; corresponds to democratic decision-making in humans. Shared decisions are ‘equally weighted’ if each group member has the same influence on the decision or ‘weighted’ if certain individuals have more influence, for example depending on their age, sex, information status, dominance, or individual needs.

Expression of preference

Signalling of individual preferences during the decisionmaking process; corresponds to voting in humans.

Informationpooling

Transfer of individual information to other group members during the decision-making process.

Leadership

If a single group member influences a group decision; may be restricted to a single decision and taken in turn by all or many of the group members (distributed leadership) or stable over several decisions (personalised leadership).

Quorum

Number of group members needed to make a group decision. Responses to a quorum threshold have been observed in many species making group decision, such as ants, honeybees, fish, and even humans.

Self-organisation

Group behaviour, which emerges in the absence of global control when individuals behave according to (simple) behavioural rules that are based on local information and that are influenced by the behaviour of neighbouring group members. Widespread in nature, occurs for example in moving fish swarms or bird flocks and even in humans, for example in pedestrians moving in a crowd.

Unshared decisions

Decisions made by one or a few (minority) of the group members; corresponds to despotic decision-making in humans.

Group decision-making in animal societies

245

9.2 Group decision-making: Basic ideas and theoretical models The particular group decision-making processes should optimise the benefits of living in groups for the involved individuals given the limits that conflicts between individuals impose. When group members share a common goal, such as finding a suitable place for communal nesting or a profitable food resource, optimal group decisions depend on the ability to compare the information available to different group members and to select the option that is best for all. Several verbal and mathematical models (Conradt and Roper 2003, List 2004, Simons 2004, Couzin et al. 2005) proposed that shared decisions, in which many individuals pool their individual information, are advantageous as they are more likely to be correct (optimal outcome for all group members) than decisions made by one or a few ‘leaders’. Mechanisms for making group decisions of this kind range from relatively simple self-organisation rules to highly evolved information-pooling behaviours. Self-organising processes during which complex group behaviours emerge through simple behavioural rules at the individual level are widespread among animals, and also exist in humans (Deneubourg and Goss 1989, Bonabeau et al. 1997, Camazine et al. 2001, Couzin and Krause 2003). Self-organisation occurs for example in cockroaches (Blatta germanica) seeking a shelter (Amé et al. 2006), army ants (Eciton burchelli) forming foraging lanes (Couzin and Franks 2003), bird flocks avoiding predators (Ballerini et al. 2008), and foraging fish swarms (Sumpter et al. 2008). Complex information-pooling behaviours, such as the honeybee’s (Apis mellifera) waggle dance, or ‘teaching’ in ants (Temnothorax albipenis), are found in eusocial insects selecting new communal nest-sites (Franks and Richardson 2006, Seeley et al. 2006). Although these different self-organisation processes and information-pooling behaviours differ in their complexity, all are regulated at the local level through positive and negative feedback processes that amplify or dampen the emergent group behaviours (Couzin 2008). Group decision-making is most complex if individual preferences vary depending on the age, sex, and social or reproductive status of the group members, and, thus, the best decision for the average group member is not optimal for all. A first model by Conradt and Roper (2003) suggested that in situations with conflicting interests, decisions made by a majority of group members should be beneficial as they avoid extreme outcomes by averaging over individual preferences, thereby keeping the consensus costs equally low for each individual. The model concludes that shared decisions, to which many or all group members contribute, outperform un-

246

Gerald Kerth

shared decisions by one or a few leaders, even in situations where individuals disagree about actions. This prediction was unexpected as in many other contexts different individual preferences lead to sex-, age-, status- or kin-specific behaviour (Krause and Ruxton 2002). For example, many ungulates as well as some bats and carnivores show social sexual segregation. Here, males and females live in separate groups because they have different activity budgets or feed on different food (Gompper et al. 1996, Conradt 1998, Ruckstuhl and Neuhaus 2002, Safi et al. 2007). While the initial model of Conradt and Roper’s (2003) could not explain how shared decisions evolved, more recent models by Conradt and Roper (2007, 2009) and Conradt et al. (2009) with an evolutionary approach overcame this problem by exploring different circumstances under which shared or unshared decisions are likely to evolve. They predicted that the evolution of shared decisions might depend on a variety of factors, including group size, the level of inter-individual conflict, and the modality of the group decision. Two factors were found to be most important: 1) the level of inter-individual conflict weighted against the grouping benefits that the individuals gain if they stay together; 2) whether the group decision-making process allows for a feasible compromise, which seems more likely for activity synchronisation (i.e., when to go) than for situations where a group has to decide about a spatial target (i.e., where to go to; Conradt and Roper 2009). Evolutionary models of this kind are very helpful if we want to understand ‘real world’ group decision-making in animal societies because they consider the diversity of situations under which animals have to make group decisions. One important assumption of most of the models dealing with group decision-making is that group members only gain grouping benefits if all of them maintain group cohesion (Conradt and Roper 2003, 2007, Rands et al. 2003, Couzin et al. 2005, Conradt et al. 2009). This assumption, however, is probably unrealistic for a substantial number of social species. In many animal societies, group members regularly form subgroups for certain time periods without leaving the society permanently (Aureli et al. 2008). Such fission-fusion behaviour (see BOX 9.1) could reflect situations where a beneficial consensus cannot be reached because interindividual conflicts or time constraints lead to a temporary group fission instead of a compromise (Franks et al. 2003, Kerth et al. 2006, King et al. 2008, Conradt et al. 2009). Alternatively, splitting into subgroups could also be an optimal group decision, if it is best for all individuals to temporarily forage or roost in smaller subgroups (Kerth et al. 2006). It will depend on the shape of the function describing the relationship between the grouping benefits and the size of a group whether individuals in subgroups can still gain the full range of grouping benefits.

Group decision-making in animal societies

247

BOX 9.1 Fission-fusion behaviour Fission-fusion behaviour is the temporary splitting and reformation of animal groups. Kummer (1968) originally introduced the term in his description of the social organisation of hamadryas baboons. Fission-fusion behaviour is widespread among animals, occurring in social insects, fish, birds, and mammals, including humans. Understanding the fission-fusion dynamics of animal groups is important as the extent to which groups can temporarily split into smaller social units has far reaching consequences for a species’ social organisation, mating system, and social interactions. Whether fissionfusion behaviour requires special communication and cognitive skills is still an open question (Aureli et al. 2008). Moreover, fission-fusion societies provide us with novel prospects to study group decisions (Kerth et al. 2006). The temporary formation of subgroups that better represent their preferences than the group as a whole could help individuals that disagree about an action to avoid a consensus without foregoing grouping benefits. The option of splitting into subgroups has largely been neglected in studies on group decision-making or has been treated as a non-adaptive outcome (e.g. Couzin et al. 2005). However, fission-fusion behaviour is important for our understanding of group coordination because whether or not group members are able to temporarily avoid group decisions that are not in their favour may strongly affect the way group decisions are made (Conradt and Roper 2005, Kerth et al. 2006). For the analyses of fission-fusion societies, association indices and social network analyses are available (Croft et al. 2008, Whitehead 2008). Both methods allow one to assess to what degree relatedness and shared interests – resulting from the same reproductive status, sex, age, or rank – explain the fission-fusion dynamics, and hence the social organisation of species (Kerth and König 1999, Archie et al. 2006, Sundareasan et al. 2007, Aureli et al. 2008, Kerth 2008, Sueur and Petit 2008). Such data are prerequisites for understanding cooperation and coordination in fissionfusion societies.

Conradt and Roper (2005) distinguish between consensus decisions in situations when all group members need to stay together and combined decisions when groups can temporarily split into subgroups. They argued that consensus and combined group decisions are conceptually different. However, several recent empirical studies suggest that consensus and combined decisions may result from the same decision-making process and the outcome of a group decision may depend more on the situation (i.e. the amount of conflict, and ultimately the ratio of grouping benefits versus consensus costs) than on the way decisions are made (e.g. Biro et al. 2006, Kerth et al. 2006, King et al. 2008).

248

Gerald Kerth

9.3 Group decision-making: empirical studies The first empirical studies on group decision-making in animals observed the coordination of travelling individuals in relatively conspicuous species, such as primates, ungulates, and birds (Kummer 1968, Wallraff 1978, Reinhardt 1983, Black 1988, Boinski and Campell 1995, Prins 1996). For some of these species, such as mountain gorillas (Gorilla gorilla beringei) or cattle (Bos indicus), the studies reported that a single individual more or less consistently led the group during travel (‘personalised leadership’; Schaller 1963, Reinhardt 1983). For other species, such as capuchin monkeys (Cebus capucinus) and African buffalos (Syncerus caffer), a ‘distributed leadership’, where many or all individuals contributed to the leading of a group, was described (Boinski and Campell 1995, Prins 1996). Although largely descriptive, these studies showed that the way group decisions are made could differ between species and sometimes even between different groups of the same species (Boinski and Campell 1995). In gorillas even a within-species contrast between unshared decisions about travelling destination and partially shared decisions about activity synchronisation was found (Stewart and Harcourt 1994). The observational field studies on mammals and birds were followed by a large number of experimental studies on social insects and fish, which were mostly carried out under controlled laboratory conditions (e.g. Seeley et al. 1991, Krause et al. 1992, Lachlan et al. 1998, Seeley and Buhrman 1999, 2001, Reebs 2000, Mallon et al. 2001, Franks et al. 2002, 2003, Pratt et al. 2002, Amé et al. 2006). Again, the majority of these studies explored group decisions over movements, either to a profitable food resource or a suitable shelter. Only recently, studies on wild mammals have started to combine field experiments and behavioural observations of recognisable individuals with the aim to investigate group decisions in heterogeneous animal societies (McComb et al. 2001, Kerth et al. 2006, King et al. 2008). These recent studies on group decision-making in mammals, birds, fish, and social insects (Fig. 9.1) illustrate the flexibility of animal group decision-making in situations with and without conflict.

Group decision-making in animal societies a

b

c

d

249

Fig. 9.1 Examples of animal societies that differ in how many individuals influence group decisions. (a) Groups of African elephants are led by a single female (Loxodonta africana; photo © C. Schradin). (b) Swarms of honeybees are led to a new nest-site by scouts that comprise about 5% of the colony members (Apis mellifera; photo © T.D. Seeley). (c) Colonies of house-hunting ants are led to a new nest-site by scouts that comprise about a third of the colony members (Temnothorax albipennis formerly Leptothorax albipennis; photo © N.R. Franks). (d) In groups of chacma baboons the dominant male decides where to feed, but leadership is shared during other travel decisions (Papio ursinus; photo © H. Peck/ZSL Tsaobis Baboon Project).

250

Gerald Kerth

9.3.1 Shared versus unshared decisions, and the importance of a quorum Whether group decisions are shared or unshared depends at least on four factors, which may affect the asymmetry of the influence of group members on the rest of the group: 1) the time available to make the decision (Franks et al. 2003), 2) the distribution of information on which the decision is based (List 2004, Simons 2004, Couzin et al. 2005), 3) the cognitive abilities of the organism (Mallon et al. 2001, McComb et al. 2001), and 4) the amount of conflict among the group members (Conradt and Roper 2005, Conradt and List 2009). Time constraints may override other factors influencing group decisions. When fast decisions are needed, all group members may benefit by following the decision of the best-informed individual (Laland 2004). Thus, situations requiring fast responses should favour unshared decisions. Indeed, in species that emit alarm calls to approaching predators, the first individual to spot the danger gives the call and causes the others to respond aptly (Sherman 1977). Moreover, when a decision has to be made almost immediately, individuals commonly copy the behaviour of their neighbour (Laland 2004). However, copying can also generate information-cascades that may lead to wrong decisions. Blindly copying the behaviour of others can therefore be maladaptive (Laland and Williams 1998, Danchin et al. 2004). The trade-off between decision accuracy and the time it takes to make a decision has been shown in ants searching for a new nest site. The more ants are involved in the decision, the more accurate a decision might become but it also takes longer to arrive at a consensus (Franks et al. 2009). Ward et al. (2008) recently showed experimentally that sticklebacks (Gasterosteus aculatus) were able to avoid maladaptive informationcascades by copying the travelling behaviour of others only if a minimum number (quorum) of conspecifics moved in the same direction. Finally, a study by List et al. (2009) looking at the group decisions of honeybees demonstrated that it depends on the interplay between inter-dependence and independence during the gathering and transfer of information in order to avoid information-cascades and to reach a consensus for the best nestsite instead. All these studies underline the importance of quorum responses during group decision-making in the common situation when a more or less graded trade-off exists between speed and accuracy (Sumpter and Pratt 2009).

Group decision-making in animal societies

251

9.3.2 Group decisions without conflict In situations in which group members share a common goal but vary in the kind of information they hold, theoretical models predict shared decisions (List 2004, Simons 2004, Couzin et al. 2005). The degree of sharing observed in empirical studies, however, differs between species. An example of a partially shared decision is the choice for a new nest-site in swarming honeybees (A. mellifera). Here, only about 5% of the colony members, the so-called scouts, decide where to lead the swarm (Seeley and Buhrman 1999). At the end of each summer, after a new queen has hatched, the old queen leaves the hive with a part of the original colony to move to a new nesting site, for example in a hollow tree. The scouts searching for new nesting-sites decide among each other which one is most suitable. By varying the intensity and duration of their waggle dance for a certain nesting site, the scouts signal their individual preferences. During this process it is not required that each scout has visited all potential nesting-sites. After the scouts have agreed on a site, they lead the swarm to the chosen place. The remaining 95% of the 3,000 – 10,000 swarm bees and the queen are not actively involved in the selection of a new home. Similarly, in the ant T. albipennis, which lives in relatively small colonies of about 200 – 500 individuals, colonies have to search for a new nest-site whenever their old nest, which is typically located in a fragile rock crevice or a nutshell, has been destroyed. In these ‘house-hunting’ ants about a third of the colony members are responsible for selection of and recruitment to a new nest-site (Pratt et al. 2002). Group decisions that involve only part of a group are also found in primates. In hamadryas baboons, only adult males are involved in the decisions about travel timing and direction (Kummer 1968), whereas in two lemur species specific females lead most of the time during travel (Erhart and Overdorff 1999). In contrast, in chacma baboons (Papio ursinus), Verreaux’s sifakas (Propithecus verreauxi), brown lemurs (Eulemur fulvus fulvus), and capuchin monkeys (C. capucinus), all adults more or less equally influence most travel decisions (Leca et al. 2003, Trillmich et al. 2004, Jacobs et al. 2008, Stueckle and Zinner 2008). Such inter-specific differences raise the question whether there are optimal proportions of individuals contributing to a group decision, depending on the situation. In ants, honeybees, African elephants (Loxodonta africana), and some primates the oldest or experienced individuals lead a social group (Seeley and Buhrman 1999, Mallon et al. 2001, Byrne 2000, McComb et al. 2001, Pratt et al. 2002). The long-term study of McComb and co-workers (2001) on elephants in Kenya provides a fine example of unshared group decision-making in animal societies, in which individuals differ in their experi-

252

Gerald Kerth

ence. Groups of elephants comprise several adult females and their offspring. Groups are led by the eldest female, the matriarch. The researchers used playback experiments to present different groups of elephants with calls of individuals that did not belong to their own group but lived in the same area. The experiment showed that the older a matriarch was, the better she was in discriminating between the calls of foreign individuals. This was the case because old matriarchs knew a large number of foreign elephants from previous encounters. McComb et al. (2001) also found that old matriarchs were most efficient in leading their groups away from those foreign individuals that were most aggressive and dangerous. This remarkable field study shows that it can be very advantageous to follow an experienced leader, and that leading a group may require special cognitive abilities if a lot of information has to be stored. More influence on decisions by experienced individuals has also been found in fish where information about food was experimentally restricted to a few members of a shoal (Krause et al. 1992, Reebs 2000, Couzin and Krause 2003, Laland 2004). This coincides with the model of Couzin et al. (2005), which predicts that in large groups of animals, such as shoal-living fish or swarming honeybees, a small proportion of informed individuals is sufficient to make an accurate decision about where to move. Finally, if conflict among group members is low and information is generally high, it may become less crucial to pool information, as each individual already makes sufficiently accurate decisions, and this might also favour unshared, fast decisions. In this case, any group member could initiate the group decision as suggested for some kinds of decisions in dolphins (Lusseau and Conradt 2009). 9.3.3 Group decisions with conflict In many circumstances, the interests of the members of a group vary, generating conflict over group decisions (Conradt and Roper 2005, 2009). Inter-individual conflict may be caused by differences in phenotype, such as age, sex, dominance status or reproductive and nutritional status. The level of inter-individual conflict also depends on the distribution of individual genotypes among group members. Colonies of eusocial insects that are founded by a single queen will often have lower conflict levels than groups of vertebrates where the members generally have a more variable degree of relatedness. Differences among individuals may also vary over time. Consequently, the level of conflict can differ depending on the situation. For example, female mammals are likely to experience less conflict outside the reproductive period than during lactation, when the physiological needs of mothers differ more from that of non-reproductive females (Kerth et al.

Group decision-making in animal societies

253

2001, Dall and Boyd 2004, Fischhoff et al. 2007). Several models (Rands et al. 2003, Conradt et al. 2009) predicted the identity of leading individuals according to the individual needs and this has been confirmed in several empirical studies (Krause et al. 1992, Fischhoff et al. 2007, Furrer et al. submitted). In their initial model, Conradt and Roper (2003) suggested that in most situations with conflict, shared decisions are favoured as subordinates are likely to resist an unfavourable ‘despotic’ decision of a dominant individual. However, a recent field study on chacma baboons (P. ursinus) suggests that the way group decisions are made does not only depend on the distribution of immediate individual payoffs of a decision, but also on the asymmetry of influence in a group and the opportunity of subordinates to avoid dominants that manipulate the outcome of a group decision in their favour (King et al. 2008). King and co-workers (2008) provided two chacma baboon groups with experimental food patches, where food was spatially more concentrated than in natural food patches. As a result, the dominant males could only monopolise food in the experimental patches. If a dominant male could obtain foraging benefits by leading his group to an experimental food patch he consistently did so, even though under other travel conditions leadership in chacma baboons is more widely distributed among group members (Stueckle and Zinner 2008). King et al. (2008) also investigated the influence of kinship, rank, and social bonds on the likelihood of following a despotic decision of the dominant individual. They observed two types of close-followers: The first type consisted of relatively high-ranking males that arrived soon after the dominant leader at an experimental patch. This assured them a disproportionately high share of the food, compared to the other, lower-ranking group members. The second type of close followers comprised lowranking animals that were highly socially affiliated with the leader. They were also very likely to closely follow the dominant male to an experimental patch, even though there they received a consensus cost from lower food intake compared to a natural food patch. Kinship had no influence on the likelihood to follow the leader closely. King et al. (2008) concluded that following behaviour in the presence of a conflict is more likely to occur when followers have strong social relationships with the leader. Longterm benefits of associating with a dominant male, such as protection from predators and lower infanticide risk (Stueckle and Zinner 2008), could explain why individuals followed in situations where this behaviour resulted in a considerable consensus cost (King et al. 2008). In the larger of the two groups, however, low-ranking individuals did not always follow the dominant male. On a few occasions, group fission occurred, resulting in the formation of two subgroups. One subgroup con-

254

Gerald Kerth

BOX 9.2 Group decisions over communal roosts in Bechstein’s bat fission-fusion societies Bechstein’s bats (M. bechsteinii) are forest-dwelling bats that can reach an age of 20 years. Females breed communally in colonies consisting of 10 – 45 adult females (Fig. 9.2). Females are philopatric and stay in their natal colony whereas males disperse in their first year of life and afterwards live solitarily during summer (Kerth and Morf 2004). As the philopatric females mate with different males each year, colonies comprise both, closely related (e.g. mother-daughter pairs) and genetically largely unrelated females (Kerth et al. 2002). Females do not give birth every year so that colonies comprise reproductive and non-reproductive females. Within their home range, colonies switch almost daily among communal day roosts (tree cavities and bat boxes) to avoid parasites and to select optimal roost temperatures (Kerth and König 1999, Kerth et al. 2001, Reckardt and Kerth 2007). Over the last 17 years, my co-workers and I collected data on the demography, reproductive success, survival, behaviour, and relatedness of more than 300 PIT-tagged individuals from four colonies living in forests near Würzburg, Germany. We found evidence for complex communication, group and individual recognition, flexible context-related social interaction, and benefits of sociality that include social thermoregulation and opportunities for information transfer about suitable roosts (Kerth and König 1999, Kerth et al. 2001, 2002, Kerth and Reckardt 2003, Safi and Kerth 2003, Kerth 2008, Pretzlaff et al. 2010). Despite their stable individual composition, Bechstein’s bat colonies are fission-fusion societies that regularly split into 2 – 6 temporary subgroups that use separate roosts for a few days before they remerge (Kerth and König 1999). As a consequence of their frequent roost switching, colony members must make daily group decisions over where to roost. We study group decision-making in wild Bechstein’s bat colonies using experiments in which we provided individuals with conflicting information about the suitability of potential roosts (Kerth et al. 2006). We also manipulated roost quality to increase asymmetries in individual preferences among colony members, for example between lactating females that prefer warmer roosts compared to non-reproductive females (Kerth et al. 2001). Permanent roost monitoring with automatic PIT-tag readers allowed us to record the bats’ nightly information transfer as well as the day-roosting behaviour (Kerth and Reckardt 2003). In one field experiment, we provided individuals with conflicting information about the suitability of new bat boxes placed inside the home ranges of two colonies. In another experiment, we provided members of one colony with conflicting information (disturbance versus no disturbance) about the suitability of their current day roost. The bats’ individual behaviour suggested that they considered both their own information and the behaviour of

Group decision-making in animal societies

255

other colony members when deciding when to switch a roost and where to roost next (Kerth et al. 2006). Conflicting interests in the heterogeneous colonies may have prevented the bats from relying entirely on social information, as this only makes sense if the needs of individuals are similar (Danchin et al. 2004). We also observed that most of the colony members inspected a novel roost and transferred information about its suitability days before it is used as a day roost for the first time. By relying on many inspecting individuals, a colony may ensure that the best communal roost is chosen. Bechstein’s bat colonies are small compared to those of eusocial insects and a large fraction of a bat colony may be required to make an optimal decision. Finally, the group decisions about communal day roosts reflected the information available to a majority of the bats roosting together but conflicting information led to increased fission in one of the two colonies. Our study suggests that fission-fusion societies allow individuals to avoid majority decisions that are not in their favour without foregoing grouping benefits that arise from communal roosting, such as social thermoregulation (Kerth et al. 2006).

Fig. 9.2 Part of a colony of Bechstein’s bats that daily make group decisions about communal day roosts. The bats are individually marked with coloured rings and subcutaneously implanted PIT-tags (Myotis bechsteinii; photo © by K. Weissman, naturefilm).

256

Gerald Kerth

tained the relatively high-ranking individuals, which received a foraging benefit at the experimental patches. The other subgroup comprised the lower-ranking individuals that would have experienced a consensus cost at the experimental food patches. During group fission, only the high-ranking animals that suffered no consensus costs visited the experimental patch. Finally, whenever the dominant male had been mate-guarding a receptive female, the group did not visit the experimental food patches. On those days this receptive female determined where the group moved and she did not visit the experimental food patch since she was relatively low in rank and would have suffered foraging costs there. The study by King et al. (2008) convincingly shows that unshared group decisions can emerge if a potential leader has both a strong incentive to lead and sufficient social influence to make the other group members follow even though this means they suffer consensus costs. That strong inter-individual conflicts can lead to temporary group fission instead of a consensus decision has also been shown in domestic pigeons (Columba livia f. domestica) that make group decisions about travel routes (Biro et al. 2006) and in Bechstein’s bats (Myotis bechsteinii) that make group decisions about communal roosts (Kerth et al. 2006; BOX 9.2). Biro and co-workers (2006) released single pigeons that carried small GPS receivers, which allowed the researchers to reconstruct the route of the individuals flying home. Each of the singly released pigeons followed an individual travel route that differed from that of other individuals but which was stable over time. In a second step, the researchers released the same individuals in pairs. They observed that the two birds chose an intermediate route, which represented a compromise between their original individual travel routes, whenever the original routes had been similar to each other. However, if the initial travel routes differed strongly between the two individuals, two different outcomes were possible. Either one bird gave up its own route and followed that of its partner or the two pigeons split up and each of them used its original travel route. This elegant study provides further evidence that the outcome of group decisions depends on the level of inter-individual conflict and may vary depending on the situation. Currently we do not know how larger pigeon flocks react to interindividual conflict. In Bechstein’s bats colonies, which comprise up to 40 females, colony members can also avoid group decisions that are against their individual interest by temporarily splitting into smaller subgroups that later merge again. Nevertheless, field experiments showed that even in situations with conflicting information about the suitability of a communal roost, Bechstein’s bats often do achieve a consensus, which reflects the information available to the majority of the colony members (Kerth et al. 2006; see

Group decision-making in animal societies

257

BOX 9.2 for more details). This study, as well as the studies on chacma baboons and domestic pigeons (Biro et al. 2006, King et al. 2008), suggests that group decision-making processes in fission-fusion societies may not be fundamentally different from those in animal societies that depend more strongly on reaching a consensus.

9.4 Open questions about group decision-making Despite the recent increase in studies on group decision-making in animals, we still lack the empirical data from a sufficient number of wild vertebrates to fully evaluate the assumptions and predictions of the models dealing with group decisions. Future empirical research must identify in what context different forms of group decisions (shared versus unshared; self-organised versus decision-making in a situation where one or several individuals have global overview) occur in different species in the wild, and then to make comparisons to other species. In particular, new empirical studies must consider the constraints (time, cognitive abilities) of group decision-making and measure the influence individuals have on group decisions in different contexts, and over longer time periods. To understand reasons for a potential asymmetry of individual influence, one also has to investigate the genetic and social relationships among group members, as well as their individual ‘personalities’ (Wolf et al. 2007, Harcourt et al. 2009, Bergmüller this volume). In nature, this will often require long-term studies, particularly in long-lived mammals (Kruuk and Hill 2008). Studies on wild animals are needed that vary the time available for a decision and then measure to what degree the resulting group decision are shared among group members in situations where individuals have different interests, analogous to what has been done for ants and honeybees in situations with no conflict of interest (Franks et al. 2003, Seeley et al. 2006). By manipulating the level of conflict in a group, its influence on the group decision processes can be tested (Biro et al. 2006, Kerth et al. 2006, King et al. 2008). An obvious experiment would be to vary group structure, for example by increasing the variation in age, relatedness, reproductive state, or body condition and then presenting the group with a situation that makes compromises between individuals difficult. In many animal societies it may be difficult to perform such manipulations, particularly in the field. Here, it will still be rewarding to study group decision-making by detailed observations of individual behaviour and describing the relationship among group members in their natural habitat (Fischhoff et al. 2007). At the level of the individual, one has to identify alternative decisions avail-

258

Gerald Kerth

able, as well as their costs and benefit in each situation. This remains a challenge for field studies on long-lived vertebrates, but King et al. (2008) showed that it is feasible. Another open question is the influence of kinship and cooperation on group decision-making processes. Kinship among group members may lower the consensus costs if an individual that refrains from taking its optimal action receives indirect benefits from helping a close relative to implement his or her optimal decision. Kin selection is not the only factor that may influence the formation of coalitions during group decisionmaking. Other forms of relevant cooperative behaviours include direct and indirect reciprocity (Couzin 2008, Conradt and List 2009). As long as subordinates are not able to form coalitions, a dominant individual does not have to fight against all subordinates together. In several cooperatively breeding species where only the dominant female breeds, she concentrates her aggression on specific individuals, and the other subordinates do not support the attacked group member (Clutton-Brock et al. 1998). Hence, the dominant individual could enforce despotic group decision-making. In addition, different interests of individual subordinates may prevent the formation of coalitions. Only in primates do subordinates form coalitions against dominant group members. Even in humans where coalition forming is common, alliances by opposing people are typically successfully suppressed in despotic societies (Harcourt 1992, Summers 2005). As a result, in the absence of kin selection or other forms of cooperation that may lower inter-individual conflict-levels, despotic decisions may be more common in heterogeneous animal societies than suggested by the initial model by Conradt and Roper (2003). Finally, it is unclear to what degree group decision-making processes are constrained by the cognitive abilities of the interacting individuals. Most of the theoretical models assume that group decisions in animals are based on simple behavioural rules, which do not involve high cognitive abilities (Couzin et al. 2005, Couzin 2008). This probably reflects many of the group decisions well where no conflict over the outcome exists, and group members only try to find the optimal way to reach and implement a decision. Travelling fish and social insects often make this kind of group decision. However, in situations with conflict among group members, the few empirical studies available on mammals suggest that complicated cognitive processes are involved in decision-making, for example in elephants that avoid potentially dangerous conspecifics (McComb et al. 2001). Since complexity of information processing also results from the social environment, the structure and size of a group should influence group decisionmaking processes. Mallon and co-workers (2001) proposed that in situations when group decisions are based on information gathering and trans-

Group decision-making in animal societies

259

fer among group members that share a common goal, small groups might place a greater cognitive burden on individuals than larger groups. This idea is in agreement with the model by Couzin and co-workers (2005), which predicts that in larger groups a smaller proportion of informed individuals is necessary for achieving an accurate decision. Mechanisms used to make group decisions that involve self-organisation do not require high cognitive abilities (Couzin 2008). Nevertheless, it should be highly rewarding to identify differences in group decision-making processes depending on the cognitive abilities and the social and ecological environment of a species, while correcting for phylogeny.

9.5 Synthesis and outlook Even though most models predict shared group decisions, several field studies show that in heterogeneous animal societies, specific individuals may have more influence than others. Reasons for this are: 1) their dominance status may give them more power in conflicts, 2) their personality makes them bolder leaders, 3) they may be favourite associates of other group members because they provide protection, 4) they may have the best knowledge because they are more experienced, 5) hold more information, or 6) have higher cognitive abilities. In all these cases, it may be beneficial for other individuals to follow their decision, even if this means accepting some (short-term) consensus costs (King et al. 2008). Depending on the social structure of a group and the situation in which a decision has to be made, individual influence on a group decision may be stable over several decisions or it may differ from decision to decision. Long-term leadership is the result of a strong asymmetry of individual influence over many decisions in different contexts. Several studies suggest that interactions between the different factors (conflict, distribution of information, constraints) influencing group decisions cause trade-offs, which can lead to flexible and sometimes suboptimal group decisions. For example, the number of individuals involved in the group decision-making process of house-hunting ants appears to be a compromise between accuracy and speed (Franks et al. 2003). Experiments with varying time-constraints confirmed that ant colonies were able to speed up the group decision-making by concentrating the process to a quorum involving fewer individuals, albeit at the cost of lower accuracy (Pratt et al. 2002, Franks et al. 2003). Trade-off between accuracy and speed is also observed in human group decisions where time constraints create a uniformity of preferences, lead to decisions with fewer individuals

260

Gerald Kerth

involved, and may prevent groups to achieve the best solution (Kerr and Tindale 2004). Furthermore, human groups are often not able to use the available information in the optimal way when a conflict among group members exists (Kerr and Tindale 2004). Research on group decision-making made tremendous progress in the last decade. However, many open questions remain (see Sect. 1.4) and empirical studies continue to be difficult and new approaches are required to successfully investigate the complexity of group decision-making in nature. More studies on group decision-making in wild animal societies that combine detailed observation of individual behaviour with carefully designed experiments are required to answer the remaining open questions. The available studies of this kind demonstrate that this can yield fascinating and unexpected results (McComb et al. 2001, Biro et al. 2006, Kerth et al. 2006, King et al. 2008). More sophisticated studies on group decisionmaking in animals will further also our understanding of the social structure and the stability of animal societies over time.

Acknowledgements This review was inspired by a workshop on communal decision-making, which Marta Manser and I organised in Zürich in September 2004 and which was funded by the Cogito Foundation. This book chapter benefited greatly from the comments of many people sharing their ideas on group decision-making with me over the last years, in particular Marta Manser and Tim Clutton-Brock. For comments on the manuscript I thank Roman Furrer, Stefan Greif, Marta Manser, Rachel Page and two anonymous referees. The German Research Foundation (DFG) funded me while I was writing this book chapter (KE476/3-1).

References Amé JM, Halloy J, Rivault C, Detrain C, Deneubourg JL (2006) Collegial decision making based on social amplification leads to optimal group formation. Proc Natl Acad Sci USA 103:5835-5840 Archie EA, Moss CJ, Alberts SC (2006) The ties that bind: genetic relatedness predicts the fission and fusion of social groups in wild African elephants. Proc R Soc Lond B 273:513-522

Group decision-making in animal societies

261

Aureli F, Schaffner CM, Boesch C, Breader SK, Call J, Chapman CA, Connor R, Di Fiore A, Dunbar RIM, Henzi SP, Holekamp K, Korstjens AH, Layton R, Lee P, Lehmann J, Manson JH, Ramos-Fernandez G, Strier KB, van Schaik CP, Boehm C, Izar P, Kreth G, Kummer H, Talebi M, Yamagiwa J (2008) Fission-fusion dynamics: new research frameworks. Curr Anthropol 49:627654 Ballerini M, Cabbibo N, Candelier R, Cavagna A, Cisbani E, Giardina I, Orlandi A, Parisi G, Procaccini A, Viale M, Zdravkovic V (2008) Empirical investigation of starling flocks: a benchmark study in collective animal behaviour. Anim Behav 76:201-215 Biro D, Sumpter DJT, Meade J, Guilford T (2006) From compromise to leadership in pigeon homing. Curr Biol 16:2123-2128 Black JM (1988) Preflight signaling in swans – a mechanism for group cohesion and flock formation. Ethology 79:143-157 Boinski S, Campbell AF (1995) Use of trill vocalizations to coordinate troop movement among white-faced capuchins – a second field-test. Behaviour 132:875-901 Bonabeau E, Theraulaz G, Deneubourg JL, Aron S, Camazine S (1997) Selforganization in social insects. Trends Ecol Evol 12:188-193 Byrne RW (2000) How monkeys find their way: leadership, coordination, and cognitive maps of African baboons. In: Boinski S, Garber PA (eds) On the Move: How and Why Animals Travel in Groups. University of Chicago Press, Chicago Camazine S, Deneubourg JL, Franks NR, Sneyd J, Theraulaz G, Bonabeau E (2001) Self-Organization in Biological Systems. Princeton University Press, Princeton Clutton-Brock TH, Brotherton PNM, Smith R, McIlrath GM, Kansky R, Gaynor D, O’Riain MJ, Skinner JD (1998) Infanticide and expulsion of females in a cooperative mammal. Proc R Soc Lond B 265:2291-2295 Conradt L (1998) Could asynchrony in activity between the sexes cause intersexual social segregation in ruminants? Proc R Soc Lond B 265:1359-1363 Conradt L, List C (2009) Group decisions in humans and animals: a survey. Philos Trans R Soc Lond B 364:719-742 Conradt L, Roper TJ (2003) Group decision-making in animals. Nature 421:155158 Conradt L, Roper TJ (2005) Consensus decision making in animals. Trends Ecol Evol 20:449-456 Conradt L, Roper TJ (2007) Democracy in animals: the evolution of shared group decisions. Proc R Soc Lond B 274:2317-2326 Conradt L, Roper TJ (2009) Conflicts of interest and the evolution of decision sharing. Philos Trans R Soc Lond B 364:807-819 Conradt L, Krause J, Couzin ID, Roper TJ (2009) ‘Leading according to need’ in self-organizing groups. Am Nat 173:304-312 Couzin ID (2008) Collective cognition in animal groups. Trends Cog Sci 13:36-43 Couzin ID, Franks NR (2003) Self-organized lane formation and optimized traffic flow in army ants. Proc R Soc Lond B 270:139-146

262

Gerald Kerth

Couzin ID, Krause J (2003) Self-organization and collective behavior in vertebrates. Adv Stud Behav 32:1-75 Couzin ID, Krause J, Franks NR, Levin SA (2005) Effective leadership and decision-making in animal groups on the move. Nature 433:513-516 Croft DP, James R, Krause J (2008) Exploring Animal Social Networks. Princeton University Press, Princeton Dall SRX, Boyd IL (2004) Evolution of mammals: lactation helps mothers to cope with unreliable food supplies. Proc R Soc Lond B 271:2049-2057 Danchin E, Giraldeau L-A, Valone TJ, Wagner RH (2004) Public information: from nosy neighbors to cultural evolution. Science 305:487-491 Deneubourg JL, Goss S (1989) Collective patterns and decision-making. Ethol Ecol Evol 1:295-311 Erhart EM, Overdorff DJ (1999) Female coordination of group travel in wild Propithecus and Eulemur. Int J Primatol 20:927-940 Fischhoff IR, Sundaresan SR, Cordingley J, Larkin HM, Sellier M-J, Rubenstein DI (2007) Social relationships and reproductive state influence leadership roles in movements of plains zebra, Equus burchellii. Anim Behav 73:825831 Franks NR, Richardson T (2006) Teaching in tandem-running ants. Nature 439:153 Franks NR, Pratt SC, Mallon EB, Britton NF, Sumpter DJT (2002) Information flow, opinion polling and collective intelligence in house-hunting social insects. Philos Trans R Soc Lond B 357:1567-1583 Franks NR, Dornhaus A, Fitzsimmons JP, Stevens M (2003) Speed versus accuracy in collective decision making. Proc R Soc Lond B 270:2457-2463 Franks NR, Dechaume-Moncharmont F-X, Hanmore E, Reynolds JK (2009) Speed versus accuracy in decision-making ants: expediting politics and policy implementation. Philos Trans R Soc Lond B 364:845-852 Furrer RD, Kunc HP, Manser MB (submitted) Variable leadership in the lowskew, cooperatively breeding banded mongoose is influenced by reproductive state Gompper ME (1996) Sociality and asociality in white-nosed coatis (Nasua narica): foraging costs and benefits. Behav Ecol 7:254-263 Harcourt AH (1992) Coalitions and alliances: are primates more complex than non-primates? In: Harcourt AH, de Waal FBM (eds) Coalitions and Alliances in Human and Non-Human Animals. Oxford University Press, New York Harcourt JL, Ang TZ, Sweetman G, Johnstone RA, Manica A (2009) Social feedback and the emergence of leaders and followers. Curr Biol 19:248-252 Jacobs A, Maumy M, Petit O (2008) The influence of social organisation on leadership in brown lemurs (Eulemur fulvus fulvus) in a controlled environment. Behav Process 79:111-113 Kerr NL, Tindale RS (2004) Group performance and decision making. Annu Rev Psychol 55:623-655 Kerth G (2008) Causes and consequences of sociality in bats. BioScience 58:737746

Group decision-making in animal societies

263

Kerth G, König B (1999) Fission, fusion and nonrandom associations in female Bechstein’s bats (Myotis bechsteinii). Behaviour 136:1187-1202 Kerth G, Morf L (2004) Behavioural and genetic data suggest that Bechstein’s bats predominantely mate outside the breeding habitat. Ethology 110:987-999 Kerth G, Reckardt K (2003) Information transfer about roosts in female Bechstein’s bats: an experimental field study. Proc R Soc Lond B 270:511515 Kerth G, Weissmann K, König B (2001) Day roost selection in female Bechstein’s bats (Myotis bechsteinii): a field experiment to determine the influence of roost temperature. Oecologia 126:1-9 Kerth G, Safi K, König B (2002) Mean colony relatedness is a poor predictor of colony structure and female philopatry in the communally breeding Bechstein’s bat (Myotis bechsteinii). Behav Ecol Sociobiol 52:203-210 Kerth G, Ebert C, Schmidtke C (2006) Group decision making in fission-fusion societies: evidence from two-field experiments in Bechstein’s bats. Proc R Soc Lond B 273:2785-2790 King AJ, Douglas CMS, Huchard E, Isaac NJB, Cowlishaw G (2008) Dominance and affiliation mediate despotism in a social primate. Curr Biol 18:1833-1838 Krause J, Ruxton GD (2002) Living in Groups. Oxford University Press, New York Krause J, Bumann D, Todt T (1992) Relationship between the position preference and nutritional state of individuals in schools of juvenile roach (Rutilus rutilus). Behav Ecol Sociobiol 30:177-180 Kruuk LEB, Hill WG (2008) Introduction. Evolutionary dynamics of wild populations: the use of long-term pedigree data. Proc R Soc Lond B 275:593-596 Kummer H (1968) Social Organisations of Hamadryas Baboons. A Field Study. Chicago University Press, Chicago Lachlan RF, Crooks L, Laland KN (1998) Who follows whom? Shoaling preferences and social learning of foraging information in guppies. Anim Behav 56:181-190 Laland KN (2004) Social learning strategies. Learn Behav 32:4-14 Laland KN, Williams K (1998) Social transmission of maladaptive information in the guppy. Behav Ecol 9:493- 499 Leca J-B, Gunst N, Thierry B; Petit O (2003) Distributed leadership in semifreeranging white-faced capuchin monkeys. Anim Behav 66:1045-1052 List C (2004) Democracy in animal groups: a political science perspective. Trends Ecol Evol 19:168-169 List C, Elsholtz C, Seeley TD (2009) Independence and interdependence in collective decision making: an agent-based model of nest-site choice by honeybee swarms. Philos Trans R Soc Lond B 364:755-762 Lusseau D, Conradt L (2009) The emergence of unshared consensus decisions in bottlenose dolphins. Behav Ecol Sociobiol 63:1067-1077 Mallon EB, Pratt SC, Franks NR (2001) Individual and collective decision-making during nest site selection by the ant Leptothorax albipennis. Behav Ecol Sociobiol 50:352-359

264

Gerald Kerth

McComb K, Moss C, Durant SM, Baker L, Sayialel S (2001). Matriarchs as repositories of social knowledge in African elephants. Science 292:491-494 Pratt SC, Mallon EB, Sumpter DJT, Franks NR (2002) Quorum sensing, recruitment, and collective decision-making during colony emigration by the ant Leptothorax albipennis. Behav Ecol Sociobiol 52:117-127 Pretzlaff I, Kerth G, Dausmann KH (2010) Communally breeding bats use physiological and behavioral adjustments to optimize daily energy expenditure. Naturwissenschaften, in press Prins HHT (1996) Ecology and Behaviour of the African Buffalo: Social Inequality and Decision Making. Chapman and Hall, London Rands SA, Cowlishaw G, Pettifor RA, Rowcliffe JM, Johnstone RA (2003) Spontaneous emergence of leaders and followers in foraging pairs. Nature 423:432434 Reckardt K, Kerth G (2007) Roost selection and roost switching of female Bechstein’s bats (Myotis bechsteinii) as a strategy of parasite avoidance. Oecologia 154:581-588 Reebs SG (2000) Can a minority of informed leaders determine the foraging movements of a fish shoal? Anim Behav 59:403-409 Reinhardt V (1983) Movement orders and leadership in a semi-wild cattle herd. Behaviour 83:251-264 Ruckstuhl KE, Neuhaus P (2002) Sexual segregation in ungulates: a comparative test of three hypotheses. Biol Rev 77:77-96 Safi K, Kerth G (2003) Secretions of the interaural gland contain information about individuality and colony membership in the Bechstein’s bat. Anim Behav 65:363-369 Safi K, König B, Kerth G (2007) Sex differences in population genetics, home range size and habitat use of the parti-colored bat (Vespertilio murinus, Linnaeus 1758) in Switzerland and their consequences for conservation. Biol Conserv 137:28-36 Schaller GB (1963) The Mountain Gorilla: Ecology and Behavior. Chicago: University of Chicago Press Seeley TD, Buhrman SC (1999) Group decision making in swarms of honey bees. Behav Ecol Sociobiol 45:19-31 Seeley TD, Buhrman SC (2001) Nest-site selection in honey bees: how well do swarms implement the ‘best-of-N’ decision rule? Behav Ecol Sociobiol 49:416-427 Seeley TD, Camazine S, Sneyd J (1991) Collective decision-making in honey bees: how colonies choose among nectar sources. Behav Ecol Sociobiol 28:277-290 Seeley TD, Visscher PK, Passino KM (2006) Group decision making in honey bee swarms. Am Sci 94:220-229 Sherman PW (1977) Nepotism and the evolution of alarm calls. Science 197:1246-1253 Simons AM (2004) Many wrongs: the advantage of group navigation. Trends Ecol Evol 19:453-455

Group decision-making in animal societies

265

Stewart KJ, Harcourt AH (1994) Gorillas’ vocalizations during rest periods: signals of impending departure? Behaviour 130:29-40 Stueckle S, Zinner D (2008) To follow or not to follow: decision making and leadership during the morning departure in chacma baboons. Anim Behav 75:1995-2004 Sueur C, Petit O (2008) Organization of group members at departure is driven by social structure in Macaca. Int J Primatol 29:1085-1098 Summers K (2005) The evolutionary ecology of despotism. Evol Hum Behav 26:106-135 Sumpter DJT, Pratt SC (2009) Quorum responses and consensus decision making. Philos Trans R Soc Lond B 364:743-753 Sumpter DJT, Krause J, Richard James R, Couzin ID, Ward AJW (2008) Consensus decision making by fish. Curr Biol 18:1773-1777 Sundaresan SR, Fischhoff IR, Dushoff J, Rubenstein DI (2007) Network metrics reveal differences in social organization between two fission-fusion species, Grevy’s zebra and onager. Oecologia 151:140-149 Trillmich J, Fichtel C, Kappeler PM (2004) Coordination of group movements in wild Verreaux’s sifakas (Propithecus verreauxi). Behaviour 141:1103-1120 Wallraff HG (1978) Social interrelations involved in migratory orientation of birds – possible contribution of field studies. Oikos 30:401-404 Ward AJW, Sumpter DJT, Couzin ID, Hart PJB, Krause J (2008) Quorum decision making facilitates information transfer in fish shoals. Proc Natl Acad Sci USA 105:6948-6953 Whitehead H (2008) Analysing Animal Societies: Quantitative Methods for Vertebrate Social Analysis. University of Chicago Press, Chicago Wolf M, van Doorn GS, Leimar O, Weissing FJ (2007) Life-history trade-offs favour the evolution of animal personalities. Nature 447:581-584

Chapter 10

Parental care: adjustments to conflict and cooperation FRITZ TRILLMICH

ABSTRACT In many species parental care is needed to rear offspring that survive to reproduce, a good measure of benefits in fitness terms. Such care may involve major costs to the individual. Balancing benefits and costs of care almost inevitably leads to tensions among individuals that provide and use the resource ‘parental care’. Thus parental care and with it parental investment (i.e. the ultimate costs of care) is enacted in a game of conflict and cooperation. Using mostly examples from mammals, I discuss Tinbergen’s four questions as they apply to parental care. Phylogenetic analyses of parental care are complicated by substantial intraspecific variability of this trait. Understanding the physiological and ontogenetic processes underlying parental care behaviour helps to understand how differences in the cost of care depend on the state of parents and their environment. A deeper insight into the strategies of conflict and cooperation in care can be derived from consideration of the behavioural mechanisms available to participants.

10.1 Introduction In this chapter, I will concentrate on mammalian examples of parental care without restricting the content to just mammals when other groups of organisms provide better examples. As we need a combined treatment of Tinbergen’s four questions (Tinbergen 1963) to understand any aspect of animal and human behaviour comprehensively, I highlight the close connection between phylogenetic, functional, mechanistic and ontogenetic aspects of parental care. Given that parental care requires a major effort from parents in time, energy, and risk taking, selection will act on individuals participating in the rearing of offspring in such a way as to minimize the

268

Fritz Trillmich

costs incurred. Thereby, conflict between parents about the extent of care performed by self versus a partner becomes likely. Nevertheless, cooperation in brood rearing can be observed quite frequently, as often two cooperating parents are more than twice as effective in rearing offspring than one might be (Maynard Smith 1977, Houston et al. 2005) and even unrelated individuals may help caring for infants (Hrdy 2009). Similarly, selection on parents differs from that on offspring concerning the amount of brood care expended on self versus siblings (whether in the same or a subsequent litter) and this is expected to lead to parent offspring conflict (Fig. 10.1; Trivers 1974, Mock and Parker 1997). In mammals, where females gestate and provide nourishment to offspring through lactation, there are many opportunities for males to forgo parental care. Therefore, active paternal care is the exception and needs a special explanation. It is found, for example, in marmoset males, who carry young (Tardif 1994), or in carnivores, who feed young by regurgitation (Asa and Valdespino 1998), in polygynous striped mice (Rhabdomys pumilio: Schubert et al. 2009a), and monogamous prairie voles (Microtus ochrogaster: Wang and Insel 1996). Even more surprising, unrelated individuals sometimes care for young as in cooperatively breeding banded mongoose (Mungos mungo: Bell 2008). Despite these examples, paternal care is rare and male-only care is almost impossible (except in humans through specialised technology like bottle feeding) and the probability of the evolution of male care is low due to physiological constraints, even though the question remains unresolved why males did not evolve lactation (with perhaps one exception in the Dayak fruit bat, Dyacopterus spadiceus: Francis et al. 1994). Therefore, in mammals, in contrast to birds, females alone care for the offspring in 90% of the species and only in 10% do males contribute substantially to rearing of offspring (CluttonBrock 1991).

10.2 Evolutionary preconditions How did this asymmetry in parental care duties arise? In the animal kingdom and beyond, females are defined as the sex with larger gametes, and in brood care, if it exists in a species, females usually expend more effort than males. These typical sex roles have often been explained as a consequence of anisogamy (Trivers 1972). The fact that females are characterised by higher initial investment in gametes supposedly causes them to continue to invest more since every single gamete of a female is more valuable than that of a male. This kind of argument is prone to the ‘Con-

Parental care: adjustments to conflict and cooperation

269

corde fallacy‘ (Dawkins and Carlisle 1976), i.e., it assumes that large past investment makes it advantageous to continue to invest. This logic represents a fallacy because it is the probability of future pay-offs that justifies investment, not the cost of past investment. The likelihood of getting offspring to the point of independence and eventual recruitment (their expected reproductive value (RV), defined as the sum of the number of offspring produced by an individual over its lifetime devalued by population growth rate; Fisher 1930) makes them valuable to a parent and justifies taking risk or otherwise expending effort on them. Another argument is that the initial greater investment by females induces male competition for females due to a skew in the operational sex ratio (OSR), i.e., the proportion of males in the pool of individuals ready for mating (Kvarnemo and Ahnesjö 1996). This fundamental sex difference, the argument goes, leads to male investment in competitive ability rather than in brood care and therefore leaves females ‘holding the baby’, i.e., expending a higher effort on brood care. This argument is often uncritically accepted. However, it overlooks the fact that a high OSR automatically leads to frequency dependent selection since every offspring of necessity has only one father and one mother (Queller 1997, Kokko and Jennions 2008). Such frequency dependent selection should favour increased parental care by whichever sex faces more serious competition for mates. This argument is based on the same logic as in the better-known case of sex ratio evolution, where the more frequent sex faces reduced chances of reproduction and should therefore be produced less, leading to a balanced sex ratio (Fisher 1930). Analogously, a male who cannot expect to compete successfully for access to females would do better by investing his resources into increased parental care rather than in fruitless attempts to compete for females. We have to assume quite strong sexual selection due to high variance in male mating success to override such a Fishercondition (Kokko and Jennions 2008). An important factor that helps to explain the conventional sex roles is female multiple mating. This female strategy reduces male paternity certainty, thereby making male investment in offspring that may not be their own less valuable. Kokko and Jennions (2008) point out that the adult sex ratio (ASR) in the population also plays a major role in determining the ESS parental care involvement of the sexes (ESS: Evolutionary Stable Strategy; a strategy which, if adopted by everyone in the population, cannot be invaded by an alternative strategy; for details see Maynard Smith 1982). As these factors (sexual selection, multiple paternity, ASR and OSR) interact with each other, the explanation of why females are more often the caring sex than males is not as easy as suggested by the simple anisogamy argument (Kokko and Jennions 2008).

270

Fritz Trillmich

BOX 10.1 Problems with defining parental care Brood care hardly needs a definition for mammals, as this taxon is characterised by its major brood care feature, the milk glands (‘mammae’). However, it is often harder to determine for other species what constitutes brood care as the decisive characteristics, a cost to the parent and benefits of parental behaviour to offspring, are not always obvious. For example, in many fish species male brood care may be incidental to the defense of a territory or some specific resource like a burrow in which females can lay their eggs. If such a shelter is essential for the attraction of mates, as for example in some blenniid fishes (Almada and Santos 1995), it is not at all clear whether the observed defense behaviour is primarily brood care or rather mating investment. Only a male with a territory will be able to attract females that may then lay eggs into his retreat. In such a case, brood care by the male will happen coincidentally to the attraction of mates. There is a benefit to the offspring through a reduced predation risk, but the ‘care’ provided may come without an additional, brood care-specific, cost to the male. The cost of defense here should rather be classified as a cost of mating effort. Similarly, female care may also be rather coincidental. In the snail cichlid (Lamprologus ocellatus), females permit the brood to use the snail shell she inhabits until she lays the next clutch. Only careful observation can determine whether shell defense is selfprotection or brood care. Snail cichlid females were found to signal to young when danger approaches and defended a larger area around the snail shell when fry began to leave the shell (Brandmann and Trillmich, unpubl. data). This demonstrates that females adjust behaviour to the presence of the brood and do not simply defend a shell for their own protection from nocturnal predators. Thus before calling a behaviour brood care, for many species it is necessary to prove the benefits of a given behaviour for the offspring and show that the actions of the parent that cause the benefit would not occur in the absence of eggs, larvae, or young.

10.3 Asking Tinbergen’s questions about parental care 10.3.1 Phylogenetic aspects Female mammals have taken on the main load of brood care through gestation and lactation. This has led to mainly maternal care in mammals with a few, taxonomically widely spaced exceptions. We find paternal care rarely in our closest relatives, the primates, but there mostly among twinning callitrichids and in humans. This rather wide separation suggests independent origins of paternal care within the primates (Snowdon 1995,

Parental care: adjustments to conflict and cooperation

271

Reynolds et al. 2002, Hrdy 2009), and separate origins are equally likely within the other orders and families. Caring fathers are found in canids, viverrids (for example meerkats, Suricata suricatta, and banded mongoose: Solomon and French 1997), hyaenids (aardwolf, Proteles cristatus: Koehler and Richardson 1990) and rodents (Clutton-Brock 1991). Within rodents, paternal care again has evolved several times independently, and even within the Muridae paternal care appears in several subfamilies: in the Murinae (striped mouse: Schradin et al. 2009), the Gerbillinae (Mongolian gerbil, Meriones unguiculatus: Elwood 1975, Weinandy and Gattermann 1999), the Arvicolinae (pine vole, Microtus pinetorum and prairie vole: Gruder-Adams and Getz 1985, Oliveras and Novak 1986), and in the Cricetidae (California mouse, Peromyscus californicus: Gubernick and Teferi 2000; Djungarian hamster, Phodopus campbelli: Wynne-Edwards and Lisk 1989). Finally, in the Caviidae, paternal care has recently been described for the yellow toothed cavy (Galea monasteriensis: Adrian et al. 2005). At the moment, this is the only mammalian species with precocial young, where paternal behaviour has been documented. Intraspecific variability in the occurrence of parental behaviour and its complex relationship to social and endocrinological state (Schradin 2008, Schradin et al. 2009) raises the question to which extent we can consider paternal behaviour to have evolved as a fixed, species typical pattern, rather than as a phenotypically plastic trait that is flexibly adjusted to the circumstances which a male finds himself in. The surprising flexibility in the expression of paternal care will be treated below, but here it highlights a more general problem of comparative analyses by demonstrating that it is hard to justify the assignment of a given species to only one parental care category. This is not only a problem in studying the phylogeny of parental care: For many behavioural traits adaptive phenotypic plasticity might have evolved rather than one fixed trait, and underlying hormonal mechanisms have proven remarkably flexible (West-Eberhard 2003). This makes it challenging to characterise species by ascribing just one trait in order to facilitate phylogenetic analyses. However, it is evident from the scattered phylogenetic distribution of male parental care in mammals that this trait and the behavioural plasticity enabling this option under specific ecological and social circumstances has evolved many times independently in different branches of the mammalian phylogeny, and thus appears to be a trait that can evolve rather easily. This malleability is not unexpected, given that all females express the trait and male mammals therefore also should carry the genetic endowment for parental care. However, its complete development has been suppressed in males of most species.

272

Fritz Trillmich

Another trait explaining the evolution of mammalian brood care concerns the developmental state of young at birth. Species can be classified as either altricial or precocial, a difference that has been described as a major grade of adaptation (Martin and MacLarnon 1985). However, these terms each comprise a wide variety of developmental stages at birth (Derrickson 1992). For example, precocial young, such as cetaceans and pinnipeds, may be sensorily, thermoregulatorily and locomotorily highly developed; yet they are completely dependent on maternal feeding. Other precocial species, such as some bovids and cervids (Knott et al. 2005) and many caviomorphs, can feed on solid food almost from birth and become independent very early (Künkele and Trillmich 1997). It has been argued that the difference between species producing precocial and altricial young might lie in differences in the resolution of parentoffspring conflict (Stockley and Parker 2002). Essentially, the argument proposed by these authors posits that with more competition for milk after birth, offspring gain advantages by growing faster so that they surpass their siblings in size and thereby gain in sibling competition. In retaliation, mothers may shorten the period of gestation to curtail the options of offspring for additional in utero growth, thereby producing altricial young which are less able to compete aggressively amongst each other. Under these circumstances, altriciality and precociality would represent the ESS solution to this parent-offspring conflict. However, variation in traits affected, in the degree of precociality, as well as its distribution across mammals (and similarly across birds: Starck and Ricklefs 1998) appears to suggest specific phylogenetic and ecological, rather than general parentoffspring conflict scenarios to underly this variation. In addition, some parameters necessary for the model to work (especially the number of offspring per teat) are out of the range of biologically likely values. With some notable exceptions (many caviomorphs), most mammals (particularly rodents) have more teats than offspring in a litter (Gilbert 1986), yet the model assumes up to 10 young per teat and needs strong sibling competition for access to the teats to produce the predicted effects. This assumption poorly fits the basic facts of mammalian biology. For example, guinea pigs and cavies produce many (1-10; mean 3.5) precocial young, despite the fact that these species have only two teats and therefore should be competing most strongly for maternal milk. Yet, they produce perhaps the most highly precocial young among mammals and competition among offspring is weak (Fey and Trillmich 2008).

Parental care: adjustments to conflict and cooperation

273

10.3.2 Functional aspects Parental care will only evolve when the benefits of such care surpass its costs. As parental care is the period of highest energy requirements for adult mammals (Drent and Daan 1980), it is tempting to measure benefits and costs in terms of energy expenditure. This aspect is certainly very important, yet it covers only part of the cost incurred by parents. Further costs include risks and the time taken to raise offspring to independence. Given limited life time, time may sometimes be even more limited than energy. Similarly, benefits can be measured as infant growth rate and size at independence (Hofer and East 2008), as well as additional aspects that influence eventual recruitment of young. However, all of these are proximate currencies and to determine parental investment (BOX 10.2) we need to measure costs and benefits in fitness currency, best perhaps as RV. As every member of a family is under different selection because of different relatedness to the other family members, this leads to conflicting benefits and costs of a given distribution of resources to family members. In this way, conflicts arise within the family (Trivers 1974, Mock and Parker 1997). There is conflict between the parents about the effort each one of them should expend, between parents and offspring about the amount of care provided overall, and among siblings about the share of care received by self versus its siblings, whether in the same or later broods (Fig. 10.1).

Mother

Sexual conflict: how much to invest in brood care

Father

Parent-offspring conflict: how much brood care to expend on each offspring

Offspring 1 Offspring 2 Offspring 3

Fig. 10.1 The structure of conflicts within families.

Sibling conflict: which proportion of care for self versus others

274

Fritz Trillmich

BOX 10.2 Definitions and measurement of parental input, parental effort and parental investment These terms are often used with different meanings. To avoid confusion, I provide some workable definitions. Take the example of a hyena mother. She may sometimes have to run 60 km to forage in order to bring milk home to her cubs. In effect, she stays away longer from her offspring, the lower the local food abundance (Hofer and East 1993, 2008). Thus, she works harder for her offspring, expends a higher maternal effort when resources are less abundant because migrating prey have moved away from her home territory. But the increased foraging expenditure of the hyena female (in terms of time and energy spent) does not translate into increased food intake for her cubs. Instead, due to the long absence of the mother, input per unit time is reduced when mothers have to work harder for food. As a consequence, offspring growth decreases. Thus parental input to the offspring is reduced, despite increased maternal effort. Whether or not this maternal effort turns out to be parental investment in the sense of Trivers (1972) is a different question that needs to be addressed by measuring future maternal reproductive success. Therefore, we need different measures to separate these three concepts. Parental input is measured as benefit to the offspring in terms of energy intake, resulting growth or safety provided through parental behaviour. Parental effort requires measurement of energetic or time expenditure of a parent for the benefit of the offspring. Parental investment is most difficult to measure because we ideally need a lifetime measure of ultimate costs to the parent, like reduced future reproductive success, while simultaneously demonstrating that this parental investment is of benefit to the fitness of offspring. Reductions of parental fitness can come about by reduced survival or fertility due to parental care. The causes behind these effects can be highly variable. Taking more risks during parental care activities might reduce survival through increased exposure to predation. Increased energy expenditure might lead to reduced immune defenses (Harshman and Zera 2007) or production of reactive oxygen compounds thereby damaging the metabolic machinery during brood care activities (Dowling and Simmons 2009). A long time spent in brood care (for example by late weaning), high expenditure of energy, specific nutrients, or body stores on a given brood may reduce fertility in the next reproductive cycle, thereby diminishing lifetime fertility.

10.3.2.1 Conflict between parents In mammals, conflict between the parents has largely been resolved in favour of males as > 90% of the species show female care only. An excellent review dealing with the many aspects of sexual conflict about brood care across a broad taxonomic spectrum has recently been provided by Wedell

Parental care: adjustments to conflict and cooperation

275

et al. (2006; see also Arnqvist and Rowe 2005). Suffice it to say that physiological constraints like internal fertilization (with the accompanying high paternity uncertainty) and a long gestation time make males more likely to search for further mates (ecological conditions permitting) and thereby potentially increase their reproductive success (Clutton-Brock and Vincent 1991). In contrast, females gain reproductive success mainly by optimizing the use of energetic resources for gestation and lactation. Even though this sounds highly plausible, it suffers from the explanatory problems pointed out above in Sect. 10.2. Parent-offspring conflict (POC) has been studied intensively in mammals. It is expected to be particularly acute in this group as lactation directly drains maternal reserves, while buffering the provisioning of young against environmental fluctuations in food supply (Dall and Boyd 2004). Thereby, female condition not only influences her ability to provide for young, but provisioning young also directly feeds back on her own physiology and potentially her survival. For a female, the decision about allocation to self-maintenance versus brood care is expected to be biased towards self-maintenance since an adult’s RV is generally higher than that of its offspring. That the behaviour of real animals closely follows this theoretical prediction was elegantly demonstrated in a study by Schubert et al. (2009b). Female house mice confronted with a decrease in the efficiency of foraging decreased overall food intake, thereby reducing energy available to themselves as well as to the offspring. Of these diminished resources they allocated proportionally more and more to self-maintenance and thereby increasingly reduced energy export to the offspring (Fig. 10.2).

Energy

Self-maintenance Milk energy output

Increasing foraging cost

Fig. 10.2 The change of allocation to self-maintenance and maternal care (here estimated as milk energy output) with increasing foraging costs. The sum of selfmaintenance and milk energy output represents net intake, which decreases with a decrease in food abundance (after Schubert et al. 2009b).

276

Fritz Trillmich

A similar effect was found by Trillmich and Kooyman (2001) for Galápagos fur seals (Arctocephalus galapagoensis), where females during a period of lower food availability even increased their own body condition, while reducing input to their offspring as evidenced by reduced growth rate of the latter. Some details about the metabolic machinery behind such changes in allocation are currently only known for insects and the model nematode, Caenorhabditis (Harshman and Zera 2007). Even when parents save expenditure on care, the rearing of offspring can still lead to considerable costs to the mother (Festa Bianchet et al. 1998). Rearing an offspring often reduces the mother’s fertility in the next reproductive cycle. This has been shown for monotocous species, for example, red deer (Clutton-Brock et al. 1983, 1989) and Galápagos fur seals and sea lions (Trillmich 1986, Trillmich and Wolf 2008). However, even in these cases it is unclear whether the reduction in fertility is a cost of rearing an offspring or represents a reproductive strategy flexibly adapted by a female in relation to resource abundance. Only experimental manipulation of parental effort can answer this question, but this is difficult to do with large free-living mammals. Interestingly, such costs of reproduction are less evident in small mammals, like ground squirrels (Neuhaus et al. 2004) and other rodents. In striking contrast to large mammals, these smaller species often prove more fertile when mating in a post-partum oestrus and thereby exposed to the double cost of simultaneous pregnancy and lactation, rather than after a period of rest in between litters (Fuchs 1982, Martinez-Gomez et al. 2004, Rehling and Trillmich 2008a). Moreover, a cost of reproduction often becomes evident only in young and old individuals, as for example in North American red squirrels, where prime-aged females showed no detectable cost of reproduction even under unfavourable ecological conditions (Descamps et al. 2009). When analysed across a population, a positive correlation is often found between reproductive expenditure and survival or fertility: The more fertile animals are found to survive better in contrast to the naïve expectation that costly reproduction should reduce future output. This comes about because animals in good condition or with better access to resources can allocate more resources to both functions, self-maintenance and reproductive effort, than those in poor condition – for example due to poor conditions during early ontogeny (Wells 2003, Naguib et al. 2006) – or with lower resources (see BOX 10.3). In striking contrast to the expectation that the cost of reproduction determines many life history trade-offs, reproductive males and females in the eusocial mole-rat, Cryptomys anselli, enjoy higher survival prospects than non-reproductives (Dammann and Burda 2006). This exceptional case parallels the unexpectedly long survival of queens of eusocial insects, which is thought to be an adaptation to low ex-

Parental care: adjustments to conflict and cooperation

277

trinsic mortality rates and the consequent high RV of queens (Keller and Genoud 1997). BOX 10.3 Individual Optimisation Hypothesis

Survival rate

When comparing parental care behaviour between individuals, we have to scale their effort against their abilities. It may seem obvious that a female providing more milk to her offspring is expending greater effort than another female with lower milk production. However, depending on female status (age, condition), higher effort of one female may ultimately be less costly to her, i.e., reduce her survival and future reproduction less, than a lower effort of another female of lower status (van Noordwijk and deJong 1986; see figure). This explains the often found positive correlation between parental effort and survival and/or fecundity of individuals. For example, individual red deer hinds of prime age were found to have the highest birth rates, greatest success in weaning offspring and lowest mortality rates, whereas younger and older individuals suffered increased mortality or reduced fertility after rearing an offspring (Clutton Brock et al. 1989). Similarly, common terns (Sterna hirundo) that prospected their future breeding colony at a young age were found to begin breeding early and had higher breeding success throughout their lives than late arrivals (Becker et al. 2008). Such observations are best explained by the hypothesis of ‘optimal investment’ (Högstedt 1980, Pettifor et al. 2001): individuals apparently can adjust parental effort in such a way that they optimise reproductive success relative to their own individual abilities and resources. These abilities may differ due to, for example, conditions experienced in early ontogeny (Naguib et al. 2006, Reid et al. 2006), current resource availability, or age.

Variance in allocation

Variance in resources

Parental effort

Variance in resource availability within a population combined with limited variance in resource allocation to reproduction versus own survival can easily suggest a positive correlation between reproductive effort and survival.

278

Fritz Trillmich

Given the potential for high costs of reproduction, parental care should be expended on own offspring. Of course, it is sometimes also used as a kind of mating effort allowing a male to approach or remain with an adult female which might mate later on with the caring male (Price 1990, Wagner et al. 1996, but see Nguyen et al. 2009). In some species care is performed by non-breeders that are tolerated by the parents (Allainé et al. 2000): Here juveniles are ‘paying for staying’ (Kokko et al. 2002) by providing a certain amount of care for younger offspring of their parents (or the breeding pair in residence). Finally, we will later see that even alloparental care (i.e. care by unrelated individuals) occurs in cooperatively breeding species. The mechanisms involved in insuring that primarily own young are provisioned often differ between altricial and precocial species. As long as altricial young are still too small to leave the nest, parents will readily accept and nurse foster young, even though they may be able to tell own from alien young (Beach and Jaynes 1956). Only when young become mobile and leave the nest, so that a nest site can no longer provide a safe cue to recognition, discrimination against foreign offspring sets in. In contrast to altricial species, where offspring discrimination happens relatively late during brood care, recognition becomes established rapidly in an imprinting-like process in some animals producing precocial offspring (Klopfer et al. 1964). Female sheep and goats need to learn the specific features of their offspring quickly, as they live in a social group, and indiscriminate nursing would likely result in the neglect of their own offspring. Therefore, females separate from the group before birth and within a few hours after parturition establish a highly specific bond with the newborn lamb that enables them to recognise their own and rebuff alien lambs (for the mechanisms involved see below). After establishing this specific bond, they immediately return to the safety of the herd (Lévy and Keller 2008). This or a similar mechanism seems to be involved in selectivity of parental care primarily in group-living precocial mammals because in other precocial species, like cavies (Cavia aperea), mothers readily accept foreign pups even if these differ in age (Rehling and Trillmich 2008b). It is unclear whether here, as in many fish, protection of own young from predation through a dilution effect (Fraser and Keenleyside 1995) was selectively more important than the limited cost of lactation (Künkele and Trillmich 1997). In contrast, altricial mouse lemurs (Microcebus murinus) nesting with one or two other females in the same tree hole recognise their own young and yet allo-nurse and adopt offspring of their partners should one of the communally nesting females die. Eberle and Kappeler (2006) were able to show that these females always were maternally related to each

Parental care: adjustments to conflict and cooperation

279

other, thereby explaining this unusual degree of cooperation by kin selection. 10.3.2.2 Competition and cooperation in parent-offspring conflict (POC) Once recognition is established, be it through a fixed site or individual recognition, parent and offspring communicate apparently in order to enable efficient allocation of brood care. Various calls and behaviours indicate hunger, the need for warmth or the offspring’s separation from parents. They evolved under strong selection to guarantee normal development of offspring (also by maintaining parental brood care motivation; see below 10.3.3). However, these same displays are used to compete for privileged access to parental (e.g. food) and nest resources (e.g. warmth). In this context, differing selection on parent(s) and offspring (Trivers 1974) raises the question of POC and of the honesty of such communication. The cases of deadly POC clearly make the point that POC can have serious fitness consequences. In brown bears (Ursus arctos), females are far more likely to abandon singleton offspring than twins. In this case the mother can gain fitness for herself by beginning another reproductive cycle earlier at a high cost to her offspring that is likely to grow slower or even die of starvation (Tait 1980, Dahle and Swenson 2003). Similarly, domestic pigs are known to abort small numbers of fetuses and cycle again, rather than rearing a suboptimally small litter (Taborsky 1985). In these cases offspring clearly lose fitness and mothers decide the POC in their own favour. These cases demonstrate that for some species the cost of time spent rearing offspring may be more important than the energetic costs involved. Cases of brood and litter size reduction are much more difficult to interpret as the fitness consequences of brood reduction may be beneficial to both, parents and offspring (except the poor victim; Mock and Parker 1997). Offspring begging may also involve shuffling for the best place to obtain food – be this the most productive teat (Drake et al. 2008) or the best position in the nest (Bautista et al. 2008). This is achieved by more or less aggressive interactions (Fraser and Thompson 1991, Hofer and East 2008, Trillmich and Wolf 2008) among young as well as by signalling, which make the offspring particularly obvious to the parent. In birds, parents may feed young according to their signalling intensity (Kilner 1997) and may also show favouritism for weaker offspring (Stamps et al. 1985). Differential feeding of young is less easy to achieve for a lactating female mammal as there are few options for differential provisioning during suckling. When feeding on solid food by regurgitation (in canids) or by passing food items on to begging young (meerkats), such favouritism can be achieved,

280

Fritz Trillmich

however, as easily as in bird parents feeding offspring in a nest. In cooperatively breeding species, like banded mongoose, pups may actually specialise on a particular individual (their ‘escort’) that is most likely to feed them, and escorts in good body condition are more likely to feed begging pups than less well-fed ones (Bell 2008). During begging episodes, signals, including vocal and olfactory signals as well as mechanical stimulation of the teats, may be of different quality. For example in pigs, stimulation by massage of the teats by many young is needed to achieve the let-down providing every offspring simultaneously with milk. Here, a clear cooperative aspect of begging becomes obvious (Drake et al. 2008). Offspring signals serve to communicate the presence of needy young, and only several young together may provide sufficient stimulation to maintain the brood care motivation of the sow in the long term, and to induce milk ejection in the shorter term. Indeed, very small litters may be given up by mothers due to suboptimal stimulation. It has previously been suggested that siblings may compete directly through begging by honestly advertising their need for a fixed level of provisioning (Godfray 1995), or that siblings may cooperate in their begging in order to jointly elevate the level of provisioning by adults (Johnstone 2004). The honesty of such begging signals has been discussed intensively. If increases in the intensity or frequency of begging signals augment parental provisioning one may assume that deception is likely to invade existing behavioural strategies, as observed in the communication between the cuckoo chick and its host (Kilner et al. 1999). Alternatively, offspring may compete for the best position and such scramble competition will lead to phenomena which are hard to separate form those predicted for honest signalling (Royle et al. 2004). While many studies have found results that appear to be consistent with honest signalling, others have found phenomena which appear to support the existence of scramble competition or a cooperative system. Furthermore, offspring may not only signal need but also quality (Mas et al. 2009). If so, parents should base their provisioning strategy as much on offspring quality (that is their RV) as on their need. The analysis is complicated even further, because offspring are quick to learn which solicitation behaviours are rewarded and thereby change begging behaviour in such a way that rewards are maximised (Kedar et al. 2000, Langmore et al. 2008). For example, in great tits needier chicks learn to approach the position from which the female feeds because she more likely feeds hungrier chicks (Kölliker and Richner 2004). Similarly, pups of cooperatively breeding meerkats and banded mongoose signal their need in a way that is strategically adjusted to the individual they are begging from (Bell 2008, Madden et al. 2009) and not directly related to

Parental care: adjustments to conflict and cooperation

281

their hunger level. The situation is complicated even further by the finding that, depending on the situation, these pups use more than one type of begging call (Kunc et al. 2007). In addition, siblings may negotiate about the next item brought by the parent while the parents are absent, as observed in barn owls (Roulin et al. 2009). These empirical findings complicate understanding parent-offspring interactions. The present models of begging capture only part of the communication process which simultaneously appears influenced by (1) competition and cooperation among offspring, (2) the variability in offspring quality, (3) their learning about the response of caregivers to the signals, (4) the ecological situation in which these signals are given, and (5) the need for a sufficient signalling level to maintain parental care motivation to weaning or fledging. The complexity of the situation seems best suited for analyses by negotiation models tailored to the species and the ecological situation (McNamara et al. 1999), rather than by general models that elucidate principles but do not fit any specific situation (Trillmich and Weissing 2006). Even in this type of modelling we should, however, not forget that a substantial part of offspring signalling may well function to maintain parental care motivation (see below), a factor that has been stressed for in utero cooperation of embryos in maintaining pregnancy (Gosling 1986) probably through sufficiently high secretion of chorionic gonadotropins (Haig 1993, Forbes 2002). Parents take a very active role in sibling competition as most clearly shown when mothers interfere in sibling conflict by defending a weaker sibling, separating fighting sibs (in hyenas: Hofer and East 2008, White 2008; in fur seals and sea lions: Trillmich and Wolf 2008), and by evolving counter-strategies that force offspring to quit fighting during feeding. In pigs, for example, the latter mechanism involves a maternal signal for a short let-down period during which fighting must cease, lest the individual will miss the chance to feed (Drake et al. 2008). Parental offering (supply) strategies have evolved which again reflect cooperation and conflict. On the one hand, mothers stimulate feeding and entice offspring to approach the milk offering teat. Such parental signals can be mechanical, olfactory, acoustic, or visual. For example, sows signal the likelihood of milk let-down through a change in grunting rate, thereby communicating to the piglets when sucking will be efficient (Weary and Fraser 1995, Drake et al. 2008). Female rabbits visit their young only once within 24 hours for a few minutes, apparently to avoid predation on the nest. To achieve rapid milk transfer, they signal the source of the milk (in a dark burrow) in a different modality, a strongly attracting pheromone secreted around the teats (Hudson and Distel 1983, Schaal et al. 2003). The amount of milk provided is then regulated through the demand of the off-

282

Fritz Trillmich

spring via a physiological feedback mechanism: The more young suck, the more milk is produced and ejected in the medium term. The discovery of such a mechanism has led to the formulation of the ‘restaurant hypothesis’ for pigs, where young massage the teat for prolonged periods after let down, i.e., when no more milk is flowing, presumably to signal their milk demand and maintain maternal milk production at a sufficient level (Jensen et al. 1998, Drake et al. 2008). In this way, demand will positively influence supply, contrary to a honest signalling scenario, which assumes that supply is independent of signalling intensity (Royle et al. 2004). In contrast, guinea pigs and cavies have evolved a maternal strategy that responds hardly at all to offspring demand (Laurien-Kehnen and Trillmich 2003, Rehling and Trillmich 2007, 2008a,b). In this species, mothers appear to set the agenda based on their own state. This is most clearly seen in a lack of response to increased pup needs. Pups given no independent option of food intake vigorously demanded more milk, but did not induce additional milk production (Laurien-Kehnen and Trillmich 2003). Weaning conflict. Much research has been done on the initiation of brood care and the internal (hormonal) and external (communication) signals that induce parents to provide care to newly hatched or newborn young. However, there is also the problem of when to end care. Trivers (1974) has pointed out that there should be conflict about the end of parental care, but we have little insight into the processes that underlie the end of parental care. Internal processes in parents may end parental care irrespective of further signalling by offspring, thus ensuring that offspring cannot manipulate parents into providing substantially more care than is optimal for their own fitness. In guinea pigs, mothers were found to wean offspring largely according to their own preset schedule (Rehling and Trillmich 2007, 2008a) and they did not react by extending the time to weaning when their own offspring were exchanged against younger pups (Rehling and Trillmich 2008b). However, older offspring transferred to mothers of younger pups used the opportunity provided and sucked for much longer than normal (Rehling and Trillmich 2008b), suggesting the existence of a mother-offspring conflict about the timing of weaning. However, as Mock and Parker (1997) have pointed out, not every behavioural squabble observed is indicative of a real evolutionary conflict between parent and offspring. Even though there is an evolutionary conflict between parent(s) and offspring, one of the major problems in proving such conflict is the lack of sufficiently accurate measurements of parental and offspring fitness to decide whether such a conflict actually exists in a given interaction, and if so, who wins. Determining the fitness consequences for all parties involved in the conflict requires measures of the fit-

Parental care: adjustments to conflict and cooperation

283

ness consequences of decisions by parents and offspring on the RV of all participants in the conflict, a tall order indeed. In addition, for example, in the case of hyena siblicide (Hofer and East 2008), at least for the victim no such decision exists, as it does not have a say in the matter. Even if the winning sibling outcompetes the sibling before its mother would have given up (Trillmich and Wolf 2008, White 2008), this does not mean that the demise of the victim does not turn out to be the best solution for the mother as well if she cannot provide enough milk for both of her offspring. So, while conflict on the behavioural level signals problems, it need not signal evolutionary conflict that may not exist in each and every case (Mock and Forbes 1992). Lack of conflict about the timing of weaning has also been documented in a few cases when offspring end parental care. For example, in northern fur seals (Callorhinus ursinus) females alternate between foraging at sea and visits to shore where they nurse the pup. In this species, pups leave the colony before mothers stop to return (Macy 1982). Apparently, it becomes important for offspring to get started in time with the southward migration. Moreover, energy provisioned by the mother to the pup towards the end of maternal care may not be sufficient to insure continued growth of the offspring which then perhaps has a better option in foraging for itself (Trillmich 1996). The mother will return one or two times after her pup has left, but then also begins its migration. Clearly, the offspring weans itself and no parent-offspring conflict exists at this point in time. The complex measurements needed make it difficult to judge whether offspring or parents win POC about provisioning. Presently, in most cases we do not have enough information on the fitness outcomes to judge whether an ESS is reached in this (signalling) game, and, if so, whether it serves best the fitness of the offspring, the parents, or both parties (see also Kilner and Hinde 2008). 10.3.3 Mechanisms and ontogenetic influences underlying parental care To understand constraints on, and potentials of, parental behaviour it is important to know the mechanisms underlying parental care. Clearly, unlimited phenotypic plasticity cannot be assumed (DeWitt et al. 1998) and knowledge of mechanisms can help to understand limitations to parental flexibility. Here, I first describe a few aspects of mechanisms working in male parental care and then go on to the much better known female side of mechanisms. Due to the enormous extent of that literature, I can only present a few aspects that suggest how the same function may depend on

284

Fritz Trillmich

different mechanisms and how understanding mechanisms feeds back on functional interpretations. 10.3.3.1 Male brood care Male mammals often, but not exclusively, get involved in paternal care when they invest in monogamous or stable polygynous relationships because they lack better alternative ways to increase reproductive success (see Wittenberger and Tilson 1980). Thus, the function of paternal care is evident where it occurs, but it is less clear how a male mammal is changed from a mate searching, deserting beast into a caring father. Thus, the proximate question arises how such a switch in behaviour is physiologically enabled. While in birds male care usually involves a reduction in testosterone levels and an increase in progesterone and prolactin (see Buntin 1996) this may not be the same in rodents, where females enter postpartum oestrus so that males should maintain regular testosterone levels to be able to copulate during the period of brood care. Indeed, in striking contrast to birds, in the California mouse (Peromyscus californicus) testosterone was found, perhaps in a threshold manner, to even increase paternal behaviour (Trainor and Marler 2001). While prolactin is of obvious importance in brood care behaviour of females, recent studies of paternal behaviour in mammals revealed quite variable involvement of prolactin in the proximate control of paternal care. Whereas fathers were found to have higher prolactin levels than nonfathers in a captive study of Djungarian hamsters (Wynne-Edwards 2001), field studies of striped mice and meerkats revealed a more variable relationship. In striped mice, where males may follow different reproductive strategies as philopatric (young) males, territorial (dominant) breeding males or roaming males, caring territorial males had higher prolactin levels than the equally caring allopaternal (i.e. the young males that were not the fathers of these pups) philopatric males. This finding suggests that male brood care behaviour is not directly proportional to prolactin levels and that parental care may be regulated differently in breeding males and alloparenting males (Schradin 2008). In meerkats, male helpers were found to display higher prolactin levels on days when they baby-sitted at the den than when they went out foraging, and prolactin levels declined over the course of the day as these baby sitters were fasting while guarding (Carlson et al. 2006a). However, during normal foraging, when male meerkats feed pups that beg for food, cortisol levels were more closely correlated with feeding rates than prolactin levels (Carlson et al. 2006b). These data demonstrate that similar brood care behaviours may be regulated differently in male and female mammals as well as between males following

Parental care: adjustments to conflict and cooperation

285

different strategies. Therefore, it is not warranted to conclude from similar behaviours to similar underlying regulatory networks, be they endocrinological or neuronal. 10.3.3.2 Maternal care The onset of maternal care is a complex process, in which indifference or aggressive defensiveness against the unknown stimuli presented by newborns needs to be replaced by attachment to offspring and maternal care for a brood. This crucial transition in behaviour has been analysed only in a few selected species: in altricial species I here concentrate on the paradigmatic rats and mice, and in precocial species on sheep and goats. These studies have demonstrated that different mechanisms can lead to similar functional outcomes, but also that mechanisms are finely tuned to social and ecological needs. Initially, most female mammals of altricial species avoid very young, still dependent conspecifics, and if they get too close they may attack or even wound or kill them. In males the equivalent response towards young (as found in lions or male house mice before mating) can be explained by sexual selection, but in virgin female rats it is a fear response that leads to avoidance or killing of young. The avoidance can be overcome by habituation to the stimulus by repeatedly exposing females to newborns. This process is called sensitisation and consists initially of a desensitisation to certain fear-evoking pup stimuli, many of them olfactory. Thereby, a female rat gets used to the proximity of pups and is increasingly exposed to pup stimuli, which then lead to the establishment of maternal behaviour. Virgin rats repeatedly exposed to newborns over a period of about a week indeed change their response towards infants from infanticide or avoidance to maternal care behaviour. They begin nest building and gather the pups in the nest where the female hovers or huddles over them in a position similar to nursing. This change in motivation is induced by the repeated exposure to the stimulus of newborns and sets into motion changes in neuronal circuits that permit maternal care behaviour (Numan and Insel 2003). Interestingly, young females near weaning (around day 24) do not yet show such a fear response that leads to avoidance of newborn pups, but will readily accept them and show maternal behaviour after much shorter sensitisation than adult nulliparous females. In some primates such as hanuman langurs (Semnopithecus entellus) virgin females will actually fight for access to newborns and are much more attentive to them than older females (Hrdy 1977). On the ultimate level, the difficulty non-reproducing females experience in accepting young are adaptive in that they protect such females from

286

Fritz Trillmich

wasting care on non-related young, but, at the proximate level, it also serves to highlight the fine-tuning needed to overcome the avoidance of and aggression against young that all of a sudden appear at parturition. This would apply in particular for cooperatively breeding species, where helpers must habituate and later care for suddenly appearing young. It also could explain why females in some of these species commit infanticide if another female gives birth before them. Under normal circumstances and in single breeders, the endocrine changes during pregnancy, in particular the rapid change in the ratio of progesterone to estrogen before birth, and the increase of prolactin during pregnancy lead to the immediate acceptance of offspring by repressing all fear responses to the novel stimuli presented by the newborns (Numan and Insel 2003). However, hormonal stimuli in the peripheral system are less required for the maintenance of maternal behaviour in rats. The exposure of the female to the sight of pups, and in particular the mechanical stimulation of the ventrum and the teats by the rooting young, serves to maintain maternal care (even though central oxytocin may play a role; Neumann 2008). This is the basis for the phenomenon that rat mothers will maintain lactation for much longer than a normal rearing period if her own pups are experimentally exchanged against younger pups (Wiesner and Sheard 1933, Pfister et al. 1986). Thus, maternal care here is maintained by the stimuli emitted by young pups and switched off as these stimuli are replaced by increasingly adult-liked stimuli (for example, the growing fur) during the normal course of offspring development. This mechanism ensures appropriate weaning under natural circumstances. Pfister et al. (1986) maintained maternal care beyond normal weaning by exchanging only some young in a litter against very young foreign pups. In such a situation older young, which were left in the nest, continued suckling for much longer than normal (about 20 – 24 days), up to more than 60 days. This shows that there is a conflict between mother and offspring about the timing of weaning which under natural circumstances mothers appear to win. Such a process does not operate in the precocial guinea pig, where exchange of young against younger offspring does not lead to a lengthening of the lactation period. In this species, the mother decides on weaning based on her own internal timer (Rehling and Trillmich 2007). Here again, young offered a chance to suckle for longer will use this option (Rehling and Trillmich 2008b). Originally, Gubernick and Alberts (1989) described that maintenance of paternal care motivation in the California mouse depended on the presence (at least olfactorily) of the mother, not on that of the pups. Thus, the male behaves as if it would make its decision to care depending on the presence of the female he mated with and less on the presence of known offspring. More recent studies showed, however, that even ‘virgin’ males showed in-

Parental care: adjustments to conflict and cooperation

287

terest in pup stimuli (de Jong et al. 2009), and some of them even behaved paternally whereas other males were initially infanticidal and would only behave paternally after birth of their own young (Gubernick et al. 1994). Here, as in rats, the onset and maintenance of maternal care does not go along with individual recognition of young, demonstrating that these processes are regulated independently. Only late in the rearing period when young become increasingly mobile does individual recognition between mother and offspring get established. Studies of sheep and goats revealed that the birth process itself plays a major role in the acceptance of young. In sheep, in contrast to rats, steroids are not sufficient to prime maternal responses. The vagino-cervical stimulation (VCS) happening during the birth process leads to a massive release of oxytocin, not only via the pituitary into the peripheral system, but also in certain brain regions, most importantly the PVN (paraventricular nucleus). Experiments showed that both, VCS and oxytocin release within the brain, are necessary to establish maternal responsiveness. This effect is furthered by the attraction to the smell of amniotic fluid briefly after birth, whereas amniotic fluid otherwise is strongly aversive. Attraction to amniotic fluid also seems to be induced by oxytocin that is released into the main olfactory bulb, where it modulates its activity and selectivity. Selectivity of the maternal response to the lamb thereby becomes established within a few hours after birth through an imprinting-like process that leads to some rearrangement of connectivity within the main olfactory bulb (Lévy and Keller 2008). Interestingly, maternal responsiveness to the young was shown to be independent of selectivity of the response to the own lamb (Lévy et al. 2004). Females respond strongly and indiscriminately to any lamb immediately after parturition. At this stage, maternal responsiveness is fully functioning, yet selectivity has still not become established. If the mother is separated from the lamb when it is one week old, maternal responsiveness fades within 36 – 72 hours. However, at this age the selective recognition of the lamb does not fade as fast as maternal responsiveness, demonstrating that the olfactory signature of the lamb has been consolidated elsewhere (mostly in cortical areas concerned with olfactory memory). Thus, VCS synchronises the onset of learning of specific olfactory cues with the onset of maternal responsiveness, thereby generating a smooth transition form a female that rejects lambs to one that accepts them, albeit only her own. The correct specificity of mother-offspring bonding is insured by the separation of the parturient ewe from the herd before birth of her pup. Acoustic and visual recognition cues are learned much slower by the mother than the olfactory signature of her lamb(s). Recognition by hearing

288

Fritz Trillmich

works within about 24 hours, but visual recognition appears to become established only after 3 weeks. These recognition memories reside in other brain regions than the one for the olfactory signature of the lamb. This is most strikingly demonstrated when observing that an anosmic female individually recognises her lamb acoustically from a distance, but nurses indiscriminately as the decision to allow access to the udder depends entirely on olfactory cues. Thus, there is no cross-check between these modalities. Functionally, this makes sense, as a nursing lamb usually cannot call, but offers the mother every opportunity for frequent olfactory checks, thereby establishing the lamb’s identity and insuring against allo-suckling (Lévy and Keller 2008). In rats, and similarly in many other mammalian mothers, pregnancy and lactation affect behaviour profoundly even beyond direct maternal care behaviour. Whereas pregnant females tend to behave rather cautiously, anxiously avoiding unknown stimuli and situations, lactating mothers are not as easily stressed as non-reproducing females and tend to be more riskprone (Slattery and Neumann 2008). Maternal aggression also becomes established around birth and appears to be maintained (in the rat) by pup stimulation to the teats and the ventrum. Importantly, these changes in behaviour seem adaptive during the brood care period as they allow the lactating female to accept more risk during foraging, if necessary to obtain sufficient energy for lactation. Furthermore, maternal aggression against conspecifics may act in protecting offspring against infanticidal males and competing females (Ebensperger 1998). Thus, pups, at least in rats, by stimulating her ventrum and her teats, prepare the mother for defense of her offspring and this example neatly exposes the sometimes unexpected links between mechanistic (proximate) and functional (ultimate) aspects of behaviour. We know less about the mechanisms determining the end of care (weaning and post-weaning care) than about those facilitating its onset, despite the fact that Trivers (1974), 35 years ago, using the example of weaning conflict, suggested that POC is to be expected given the genetic interests of the participants in the conflict. In a way then, this lack of understanding the processes and mechanisms producing the separation of mother and offspring may be considered a case of ‘arrested development’ (Mock and Forbes 1992) of research into this aspect of POC.

Parental care: adjustments to conflict and cooperation

289

10.4 Cooperative brood care Just because of all the emphasis on conflicts over brood care, it is worth stressing the cooperation necessary among family members: siblings benefit from joint stimulation of the brood care activity of the parents (Hussell 1988), grow better through maintenance of high temperature within a nest of altricial young (Bautista et al. 2008) and may benefit in an as yet little documented way from early experience in interactions with siblings (Hudson and Trillmich 2008), as amply demonstrated for humans (Sulloway 1996). However, in a few mammalian species, members of the extended family and non-related individuals actually cooperate in brood care. Even though kinship may play an important role in the evolution of such cooperative brood care (e.g., black-backed jackals Canis mesomelas: Moehlman 1979; elephants, Loxodonta africana: Lee 1987) ecological factors also prove highly influential (for primate examples see Hrdy 2009). This is particularly evident in the process of ‘group augmentation’ (Kokko et al. 2001), where the well-being of the group is vital to the individual for ecological reasons, largely independent of kinship. For example, in meerkats, banded mongoose (Bell 2008), and wild dogs (Lycaon pictus: Malcolm and Marten 1982, Creel et al. 1997), helpers, often not directly related, are needed to enable the rearing of offspring, to maintain the group in the face of serious predation pressure, and to enable the group to compete with its neighbours. For example, in wild dogs, only a group can successfully hunt larger prey, protect young from predators, and will succeed in rearing offspring to independence. Therefore, male helpers benefit by rearing offspring even if they are not the father of these young (Malcolm and Marten 1982). In perhaps the best investigated case, individual meerkat in larger groups profit from the activity of others that act as sentinels against predators, feed (related and unrelated) young, and may even forego foraging for 24 hours to protect young that are left behind at a den. Predation pressure here plays a major role in causing the advantages of larger groups. Large groups are much better able to rear young than small groups, despite the fact that competition for food within the group increases with group size (Brotherton et al. 2001). In addition, larger group size can confer advantages in competition with neighbouring groups. But, beware that all this cooperation within the group does not prevent individuals to compete fiercely for breeding opportunities: Generally, only the dominant female breeds successfully. She drives out subdominant females during the last three weeks of her pregnancy, thereby reducing their chances to conceive

290

Fritz Trillmich

and increasing the rate of abortion. If a subordinate nevertheless manages to breed, the dominant is likely to kill its offspring even though these may be her daughter’s young. But pregnant subordinates fight back and will kill the dominant’s offspring thereby becoming the main source of pup mortality for the dominant (Young and Clutton-Brock 2006). This intense and costly conflict is strong evidence for the absolute need for grouping to survive and breed as well as for the limits of kinship arguments to explain cooperation in allo-parental care. The pinnacle of such (somewhat forced) cooperation is reached in naked mole-rats (Heterocephalus glaber), which rear young in a eusocial group (for review of the bathyergid group see Faulkes and Bennett 2007). Here the dominant female breeder even changes her brain morphology upon becoming dominant (Holmes et al. 2007) and suppresses the breeding of other females in the group by bullying subordinates (Faulkes and Abbott 1997). Even brain evolution in humans has been suggested to be based on the evolution of cooperative breeding (Hrdy 2009). In our species, Hrdy postulates that the high reproductive rate (in comparison to other great apes) can only be achieved through cooperative care of the baby which releases the mother from the extensive need to carry her rather helpless offspring for years. This saves energetic expenditure and frees the mother to invest in further offspring. The sharing of costs by helpers includes the extended family, not least post-menopausal grandmothers. This need for cooperative care, according to Hrdy (2009) has selected for cooperative social interactions on the side of adults and, in the babies, for understanding the intentions of social companions and manipulating their brood care tendencies. The enormous amount of social knowledge needed to achieve this degree of social cooperation may thereby have selected for the development of our extreme brain size.

Acknowledgements I am grateful to Peter Kappeler, Doug Mock and an anonymous referee for many constructive suggestions and comments. My own research on this topic has greatly benefitted from longterm support by the DFG.

Parental care: adjustments to conflict and cooperation

291

References Adrian O, Brockmann I, Hohoff C, Sachser N (2005) Paternal behaviour in wild guinea pigs: a comparative study in three closely related species with different social and mating systems. J Zool Lond 265:97-105 Allainé D, Brondex F, Graziani L, Coulon J, Till-Bottraud I (2000) Male-biased sex ratio in litters of Alpine marmots supports the helper repayment hypothesis. Behav Ecol 11:507-514 Almada VC, Santos RS (1995) Parental care in the rocky intertidal: a case study of adaptation and exaptation in Mediterranean and Atlantic blennies. Rev Fish Biol Fish 5:23-37 Arnqvist G, Rowe L (2005) Sexual Conflict. Princeton University Press, Princeton Asa CS, Valdespino C (1998) Canid reproductive biology: an integration of proximate mechanisms and ultimate causes. Am Zool 38:251-259 Bautista A, García-Torres E, Martínez-Gómez M, Hudson R (2008) Do newborn domestic rabbits Oryctolagus cuniculus compete for thermally advantageous positions in the litter huddle? Behav Ecol Sociobiol 62:331-339 Beach FA, Jaynes J (1956) Studies of maternal retrieving in rats. I. Recognition of young. J Mammal 37:177-180 Becker PH, Dittmann T, Ludwigs J-D, Limmer B, Ludwig SC, Bauch C, Braasch A, Wendeln H (2008) Timing of initial arrival at the breeding site predicts age at first reproduction in a long-lived migratory bird. Proc Natl Acad Sci USA 105:12349-12352 Bell MBV (2008) Strategic adjustment of begging effort by banded mongoose pups. Proc R Soc Lond B 275:1313-1319 Brotherton PNM, Clutton-Brock TH, O’Riain MJ, Gaynor D, Sharpe L, Kansky R, McIlrath GM (2001) Offspring food allocation by parents and helpers in a cooperative mammal. Behav Ecol 12:590-599 Buntin JD (1996) Neural and hormonal control of parental behavior in birds. Adv Stud Behav 25:161-213 Carlson AA, Manser M, Young AJ, Russel AF, Jordan NR, McNeilly AS, Clutton-Brock TH (2006a) Cortisol levles are positively associated with pupfeeding rates in male meerkats. Proc R Soc Lond B 273:571-577 Carlson AA, Russel AF, Young AJ, Jordan NR, McNeilly AS, Parlow AF, Clutton-Brock TH (2006b) Elevated prolactin levles immediately precede decisions to babysit by meerkat helpers. Horm Behav 50:94-100 Clutton-Brock TH (1991) The Evolution of Parental Care. Princeton University Press, Princeton Clutton-Brock TH, Vincent ACJ (1991) Sexual selection and the potential reproductive rates of males and females. Nature 351:58-60 Clutton-Brock TH, Guiness FE, Albon SD (1983) The cost of reproduction to red deer hinds. J Anim Ecol 52:367-383 Clutton-Brock TH, Albon SD, Guinness FE (1989) Fitness costs of gestation and lactation in wild mammals. Nature 337:260-262

292

Fritz Trillmich

Creel S, Creel MN, Mills MGL, Monfort SL (1997) Rank and reproduction in cooperatively breeding African wild dogs: behavioral and endocrine correlates. Behav Ecol 8:298-306 Dahle B, Swenson JE (2003) Factors influencing length of maternal care in brown bears (Ursus arctos) and its effect on offspring. Behav Ecol Sociobiol 54:352358 Dall SRX, Boyd IL (2004) Evolution of mammals: lactation helps mothers to cope with unreliable food supplies. Proc R Soc Lond B 271:2049-2057 Dammann P, Burda H (2006) Sexual activity and reproduction delay ageing in a mammal. Curr Biol 16:R117-R118 Dawkins R, Carlisle TR (1976) Parental investment, mate desertion and a fallacy. Nature 262:131-133 de Jong TR, Chauke M, Harris BN, Saltzman W (2009) From here to paternity: neural correlates of the onset of paternal behavior in California mice (Peromyscus californicus). Horm Behav 56:220-231 Derrickson EM (1992) Comparative reproductive strategies of altricial and precocial eutherian mammals. Funct Ecol 6:57-65 Descamps S, Boutin S, McAdam AG, Berteaux D, Gaillard J-M (2009) Survival costs of reproduction vary with age in North American red squirrels. Proc R Soc Lond B 276:1129-1135 DeWitt TJ, Sih A, Wilson DS (1998) Costs and limits of phenotypic plasticity. Trends Ecol Evol 13:77-81 Dowling DK, Simmons LW (2009) Reactive oxygen species as universal constraints in life-history evolution. Proc R Soc Lond B 276:1737-1745 Drake A, Fraser D, Weary DM (2008) Parent-offspring resource allocation in domestic pigs. Behav Ecol Sociobiol 62:309-320 Drent RH, Daan S (1980) The prudent parent. Energetic adjustments in avian breeding. Ardea 68:225-252 Ebensperger LA (1998) Strategies and counterstrategies to infanticide in mammals. Biol Rev 73:321-346 Eberle M, Kappeler PM (2006) Family insurance: kin selection and cooperative breeding in a solitary primate (Microcebus murinus). Behav Ecol Sociobiol 60:582-588 Elwood RW (1975) Paternal and maternal behaviour in the Monoglian gerbil. Anim Behav 23:766-772 Faulkes CG, Abbott DH (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: Solomon NG, French JA (eds) Cooperative Breeding in Mammals. Cambridge University Press, Cambridge, pp 302-334 Faulkes CG, Bennett NC (2007) African mole-rats: social and ecological diversity. In: Wolff JO, Sherman PW (eds) Rodent Societies: An Ecological and Evolutionary Perspective. Chicago University Press, Chicago, pp 427-437 Festa-Bianchet M, Gaillard J-M, Jorgenson JT (1998) Mass- and densitydependent reproductive success and reproductive costs in a capital breeder. Am Nat 152:367-379

Parental care: adjustments to conflict and cooperation

293

Fey K, Trillmich F (2008) Sibling competition in guinea pigs (Cavia aperea f. porcellus): scrambling for mother’s teats is stressful. Behav Ecol Sociobiol 62:321-329 Fisher RA (1930) The Genetical Theory of Natural Selection. Oxford University Press, Oxford Forbes S (2002) Pregnancy sickness and embryo quality. Trends Ecol Evol 17:115-120 Francis CM, Anthony ELP, Brunton JA, Kunz TH (1994) Lactation in male fruit bats. Nature 367:691-692 Fraser SA, Keenleyside MHA (1995) Why are larger convict cichlid (Cichlasoma nigrofasciatum) fry sometimes adopted into broods of smaller fry? Behav Ecol Sociobiol 37:343-347 Fraser D, Thompson BK (1991) Armed sibling rivalry among suckling piglets. Behav Ecol Sociobiol 29:9-15 Fuchs S (1982) Optimality of parental investment: the influence of nursing on reproductive success of mother and female young house mice. Behav Ecol Sociobiol 10:39-51 Gilbert AN (1986) Mammary number and litter size in Rodentia – the one-half rule. Proc Natl Acad USA Sci 83:4828-4830 Godfray HCJ (1995) Evolutionary theory of parent-offspring conflict. Nature 376:133-138 Gosling LM (1986) Selective abortion of entire litters in the coypu: adaptive control of offspring production in relation to quality and sex. Am Nat 127:772795 Gruder-Adams S, Getz LL (1985) Comparison of the mating system and paternal behavior of Microtus ochrogaster and M. pennsylvanicus. J Mammal 66:165167 Gubernick DJ, Alberts JR (1989) Postpartum maintenance of paternal behaviour in the biparental California mouse, Peromyscus californicus. Anim Behav 37:656-664 Gubernick DJ, Teferi T (2000) Adaptive significance of male parental care in a monogamous mammal. Proc R Soc Lond B 267:147-150 Gubernick DJ, Schneider KA, Jeanotte LA (1994) Individual differences in the mechanisms underlying the onset and maintenance of paternal behavior and the inhibition of infanticide in the monogamous biparental California mouse, Peromyscus californicus. Behav Ecol Sociobiol 34:225-231 Haig D (1993) Genetic conflicts in human pregnancy. Q Rev Biol 68:495-532 Harshman LG, Zera AJ (2007) The cost of reproduction: the devil in the details. Trends Ecol Evol 22:80-86 Högstedt G (1980) Evolution of clutch size in birds: adaptive variation in relation to territory quality. Science 210:1148-1150 Hofer H, East ML (1993) The commuting system of Serengeti spotted hyaenas: how a predator copes with migratory prey. III. Attendance and maternal care. Anim Behav 46:575-589 Hofer H, East ML (2008) Siblicide in Serengeti spotted hyenas: a long-term study of maternal input and cub survival. Behav Ecol Sociobiol 62:341-351

294

Fritz Trillmich

Holmes MM, Rosen GJ, Jordan CL, de Vries GJ, Goldman BD, Forger NG (2007) Social control of brain morphology in a eusocial mammal. Proc Natl Acad Sci USA 104:10548-10552 Houston AI, Székely T, McNamara JM (2005) Conflicts between parents over care. Trends Ecol Evol 20:33-38 Hrdy SB (1977) The Langurs of Abu: Female and Male Strategies of Reproduction. Harvard University Press, Cambridge/MA Hrdy SB (2009) Mothers and Others: The Evolutionary Origins of Mutual Understanding. Harvard University Press, Cambridge/MA Hudson R, Distel H (1983) Nipple location by newborn rabbits: behavioural evidence for pheromonal guidance. Behaviour 85:260-274 Hudson R, Trillmich F (2008) Sibling competition and cooperation in mammals: challenges, developments and prospects. Behav Ecol Sociobiol 62:299-307 Hussel DJT (1988) Supply and demand in tree swallow broods: a model of parentoffspring food-provisioning interactions in birds. Am Nat 131:175-202 Jensen P, Gustafsson M, Augustsson H (1998) Teat massage after milk ingestion in domestic piglets: an example of honest begging? Anim Behav 55:779-786 Johnstone RA (2004) Begging and sibling competition: how should offspring respond to their rivals? Am Nat 163:388-406 Kedar H, Rodríguez-Gironés MA, Yedvab S, Winkler DW, Lotem A (2000) Experimental evidence for offspring learning in parent-offspring communication. Proc R Soc Lond B 267:1723-1727 Keller L, Genoud M (1997) Extraordinary lifespans in ants: a test of evolutionary theories of ageing. Nature 389:958-960 Kilner RM (1997) Mouth colour is a reliable signal of need in begging canary nestlings. Proc R Soc Lond B 264:963-968 Kilner RM, Hinde CA (2008) Information warfare and parent-offspring conflict. Adv Stud Behav 38:283-336 Kilner RM, Noblet DG, Davies NB (1999) Signals of need in parent-offspring communication and their exploitation by the common cuckoo. Nature 397:667-672 Klopfer PH, Adams DK, Klopfer MS (1964) Maternal ‘imprinting’ in goats. Proc Natl Acad Sci USA 52:911-914 Knott KK, Barboza PS, Bowyer RT (2005) Growth in arctic ungulates: postnatal development and organ maturation in Rangifer tarandus and Ovibos moschatus. J Mammal 86:121-130 Koehler CE, Richardson PRK (1990) Proteles cristatus. Mammal Species 363:1-6 Kölliker M, Richner H (2004) Navigation in a cup: chick positioning in great tit, Parus major, nests. Anim Behav 68:941-948 Kokko H, Jennions MD (2008) Parental investment, sexual selection and sex ratios. J Evol Biol 21:919-948 Kokko H, Johnstone RA, Clutton-Brock TH (2001) The evolution of cooperative breeding through group augmentation. Proc R Soc Lond B 268:187-196 Kokko H, Johnstone RA, Wright J (2002) The evolution of alloparental effort in cooperatively breeding groups: when should helpers pay to stay? Behav Ecol 13:291-300

Parental care: adjustments to conflict and cooperation

295

Künkele J, Trillmich F (1997) Are precocial young cheaper? Lactation energetics in the guinea pig. Physiol Zool 70:589-596 Kunc HP, Madden JR, Manser MB (2007) Begging signals in a mobile feeding system: the evolution of different call types. Am Nat 170:617-624 Kvarnemo C, Ahnesjö I (1996) The dynamics of operational sex ratios and competition for mates. Trends Ecol Evol 11:404-408 Langmore NE, Maurer G, Adcock GJ, Kilner RM (2008) Socially acquired hostspecific mimicry and the evolution of host races in Horsfield’s bronze cuckoo Chalcites byasalis. Evolution 62:1689-1699 Laurien-Kehnen C, Trillmich F (2003) Lactation performance of guinea pigs (Cavia porcellus) does not respond to experimental manipulation of pup demands. Behav Ecol Sociobiol 53:145-152 Lee PC (1987) Allomothering among African elephants. Anim Behav 35:278-291 Lévy F, Keller M (2008) Neurobiology of maternal behavior in sheep. Adv Stud Behav 38:399-437 Lévy F, Keller M, Poindron P (2004) Olfactory regulation of maternal behavior in mammals. Horm Behav 46:284-302 Macy SK (1982) Mother-pup interaction in the Northern fur seal. PhD Thesis, University of Washington, Seattle/WA Madden JR, Kunc HP, English S, Manser MB, Clutton-Brock TH (2009) Calling in the gap: competition or cooperation in littermates’ begging behaviour. Proc R Soc Lond B 276:1255-1262 Malcom JR, Marten K (1982) Natural selection and the communal rearing of pups in African wild dogs (Lycaon pictus). Behav Ecol Sociobiol 10:1-13 Martin RD, MacLarnon AM (1985) Gestation period, neonatal size and maternal investment in placental mammals. Nature 313:220-223 Martínez-Gómez M, Juárez M, Distel H, Hudson R (2004) Overlapping litters and reproductive performance in the domestic rabbit. Physiol Behav 82:629-636 Mas F, Haynes KF, Kölliker M (2009) A chemical signal of offspring quality affects maternal care in a social insect. Proc R Soc Lond B 276:2847-2853 Maynard Smith J (1977) Parental investment: a prospective analysis. Anim Behav 25:1-9 Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge McNamara JM, Gasson CE, Houston AI (1999) Incorporating rules for responding into evolutionary games. Nature 401:368-371 Mock DW, Forbes LS (1992) Parent-offspring conflict: a case of arrested development. Trends Ecol Evol 7:409-413 Mock DW, Parker GA (1997) The Evolution of Sibling Rivalry. Oxford University Press, Oxford Moehlman PD (1979) Jackal helpers and pup survival. Nature 277:382-383 Naguib M, Nemitz A, Gil D (2006) Maternal developmental stress reduces reproductive success of female offspring in zebra finches. Proc R Soc Lond B 273:1901-1905

296

Fritz Trillmich

Neuhaus P, Brussard DR, Murie JO, Dobson FS (2004) Age of primiparity and implications of early reproduction on life history in female Columbian ground squirrels. J Anim Ecol 73:36-43 Neumann ID (2008) Brain oxytocin: a key regulator of emotional and social behaviours in both females and males. J Neuroendocrinol 20:858-865 Nguyen N, van Horn RC, Alberts SC, Altmann J (2009) ‘Friendships’ between new mothers and adult males: adaptive benefits and determinants in wild baboons (Papio cynocephalus). Behav Ecol Sociobiol 63:1331-1344 Numan M, Insel TR (2003) The Neurobiology of Parental Behavior. Springer Verlag, New York Oliveras D, Novak M (1986) A comparison of paternal behaviour in the meadow vole Microtus pennsylvanicus, the pine vole M. pinetorum and the prairie vole M. ochrogaster. Anim Behav 34:519-526 Pettifor RA, Perrins CM, McCleery RH (2001) The individual optimization of fitness: variation in reproductive output, including clutch size, mean nestling mass and offspring recruitment, in manipulated broods of the great tits Parus major. J Anim Ecol 70:62-79 Pfister JF, Cramer CP, Blass EM (1986) Suckling in rats extended by continuous living with dams and their preweanling litters. Anim Behav 34:415-420 Price EC (1990) Infant carrying as a courtship strategy of breeding male cottontop tamarins. Anim Behav 40:784-786 Queller DC (1997) Why do females care more than males? Proc R Soc Lond B 264:1555-1557 Rehling A, Trillmich F (2007) Weaning in the guinea pig (Cavia aperea f. porcellus): who decides and by what measure? Behav Ecol Sociobiol 62:149-157 Rehling A, Trillmich F (2008a) Maternal effort is state dependent: energetic limitation or regulation? Ethology 114:318-326 Rehling A, Trillmich F (2008b) Changing supply and demand by cross-fostering: effects on the behaviour of pups and mothers in guinea pigs, Cavia aperea f. porcellus and cavies, Cavia aperea. Anim Behav 75:1455-1463 Reid JM, Bignal EM, Bignal S, McCracken DI, Monaghan P (2006) Spatial variation in demography and population growth rate: the importance of natal location. J Anim Ecol 75:1201-1211 Reynolds JD, Goodwin NB, Freckleton RP (2002) Evolutionary transitions in parental care and live bearing in vertebrates. Philos Trans R Soc Lond B 357:269-281 Roulin A, Dreiss A, Fioravanti C, Bize P (2009) Vocal sib-sib interactions: how siblings adjust signalling level to each other. Anim Behav 77:717-725 Royle NJ, Hartley IR, Parker GA (2004) Parental investment and family dynamics: interactions between theory and empirical tests. Pop Ecol 46:231-241 Schaal B, Coureaud G, Langlois D, Giniés C, Sémon E, Perrier G (2003) Chemical and behavioural characterization of the rabbit mammary pheromone. Nature 424:68-72 Schradin C (2008) Differences in prolactin levels between three alternative male reproductive tactics in striped mice (Rhabdomys pumilio). Proc R Soc Lond B 275:1047-1052

Parental care: adjustments to conflict and cooperation

297

Schradin C, Scantlebury M, Pillay N, König B (2009) Testosterone levels in dominant sociable males are lower than in solitary roamers: physiological differences between three male reproductive tactics in a sociably flexible mammal. Am Nat 173:376-388 Schubert M, Pillay N, Schradin C (2009a) Parental and alloparental care in a polygynous mammal. J Mammal 90:724-731 Schubert KA, de Vries G, Vaanholt LM, Meijer HAJ, Daan S, Verhulst S (2009b) Maternal energy allocation to offspring increases with environmental quality in house mice. Am Nat 173:831-840 Slattery DA, Neumann ID (2008) No stress please! Mechanisms of stress hyporesponsiveness of the maternal brain. J Physiol 586:377-385 Snowdon CT (1995) Infant care in cooperatively breeding species. Adv Stud Behav 25:643-689 Solomon NG, French JA (1997) Cooperative Breeding in Mammals. Cambridge University Press, Cambridge Stamps J, Clark A, Arrowood P, Kus B (1985) Parent-offspring conflict in budgerigars. Behaviour 94:1-40 Starck JM, Ricklefs RE (1998) Avian Growth and Development: Evolution within the Altricial-Precocial Spectrum. Oxford University Press, New York Stockley P, Parker GA (2002) Life history consequences of mammalian sibling rivalry. Proc Natl Acad Sci USA 99:12932-12937 Sulloway FJ (1996) Born to Rebel: Birth Order, Family Dynamics, and Creative Lives. Pantheon, New York Taborsky M (1985) On optimal parental care – commentary. Z Tierpsychol 70:331-336 Tait DEN (1980) Abandonment as a reproductive tactic – the example with grizzly bears. Am Nat 115:800-808 Tardif SD (1994) Relative energetic cost of infant care in small-bodied neotropical primates and its relation to infant-care patterns. Am J Primatol 34:133-143 Tinbergen N (1963) On aims and methods of ethology. Z Tierpsychol 20:410-433 Trainor BC, Marler CA (2001) Testosterone, paternal behavior, and aggression in the monogamous California mouse (Peromyscus californicus). Horm Behav 40:32-42 Trillmich F (1986) Maternal investment and sex-allocation in the Galápagos fur seal, Arctocephalus galapagoensis. Behav Ecol Sociobiol 19:157-164 Trillmich F (1996) Parental investment in pinnipeds. Adv Stud Behav 25:533-577 Trillmich F, Kooyman GL (2001) Field metabolic rate of lactating female Galápagos fur seals (Arctocephalus galapagoensis): the influence of offspring age and environment. Comp Biochem Physiol 129A:741-749 Trillmich F, Weissing FJ (2006) Lactation patterns of pinnipeds are not explained by optimization of maternal energy delivery rates. Behav Ecol Sociobiol 60:137-149 Trillmich F, Wolf JBW (2008) Parent-offspring and sibling conflict in the Galápagos fur seals and sea lions. Behav Ecol Sociobiol 62:363-375

298

Fritz Trillmich

Trivers RL (1972) Parental investment and sexual selection. In: Campbell B (ed) Sexual Selection and the Descent of Man, 1871-1971. Aldine, Chicago, pp 136-179 Trivers RL (1974) Parent-offspring conflict. Am Zool 11:249-264 van Noordwijk AJ, deJong G (1986) Acquisition and allocation of resources: their influence on variation in life history tactics. Am Nat 128:137-142 Wagner RH, Schug MD, Morton ES (1996) Confidence of paternity, actual paternity and parental effort by purple martins. Anim Behav 52:123-132 Wang Z, Insel TR (1996) Parental behavior in voles. Adv Stud Behav 25:361-384 Weary DM, Fraser D (1995) Calling by domestic piglets: reliable signals of need? Anim Behav 50:1047-1055 Wedell N, Kvarnemo C, Lessels CM, Treganza T (2006) Sexual conflict and life histories. Anim Behav 71:999-1011 Weinandy R, Gattermann R (1999) Parental care and time sharing in the Mongolian gerbil. Z Säugetierkd 64:169-175 Wells JCK (2003) Parent-offspring conflict theory, signalling of need, and weight gain in early life. Q Rev Biol 78:169-202 West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford University Press, Oxford White PA (2008) Maternal response to neonatal sibling conflict in the spotted hyena, Crocuta crocuta. Behav Ecol Sociobiol 62:353-361 Wiesner BP, Sheard NM (1933) Maternal Behaviour in the Rat. Oliver and Boyd, London Wittenberger JF, Tilson RL (1980) The evolution of monogamy: hypotheses and evidence. Annu Rev Ecol Syst 11:197-232 Wynne-Edwards KE, Lisk RD (1989) Differential effects of paternal presence on pup survival in two species of dwarf hamster (Phodopus sungorus and Phodopus campbelli). Physiol Behav 45:465-469 Wynne-Edwards KE (2001) Hormonal changes in mammalian fathers. Horm Behav 40:139-145 Young AJ, Clutton-Brock TH (2006) Infanticide by subordinates influences reproductive sharing in cooperatively breeding meerkats. Biol Lett 2:385-387

Part III Sex and reproduction

Chapter 11

The quantitative study of sexual and natural selection in the wild and in the laboratory WOLF BLANCKENHORN

ABSTRACT I discuss the continuum of approaches that exist when studying sexual or natural selection in the wild and the laboratory. These range from behavioural observations in the laboratory, via experimental manipulations of particular traits or environments, to phenomenological studies in nature. I focus on the study of body size and related life history traits, particularly drawing from our own studies on dung flies as examples. For any given species or phenomenon, ideally all types of studies should be integrated to obtain a complete picture of the evolution of particular traits in terms of mechanisms and function, proximate and ultimate explanations. I particularly advocate the use of standardised selection measures, which are well established in the literature but underused, and which I discuss in the chapter as practical guidance for the general reader. Utilisation of such measures even in experimental, laboratory studies of behavioural mechanisms, which is often possible but rarely done, would greatly facilitate any future meta-analyses of particular traits, species and evolutionary phenomena.

11.1 Introduction Whether studying behaviour, ecology or evolution, Darwin’s (1859, 1871) principle of natural selection is the central paradigm guiding our research. Natural selection is indeed the sole alternative mechanism explaining organic evolution besides the null hypothesis of random genetic drift. Therefore a thorough understanding of the process of natural selection is central to any research in organismic biology, including behaviour. All research in behaviour and ecology originates from observation and description. For example, before investigating the mechanisms of mate choice in the long-tailed widowbird, Andersson (1982) had to understand

302

Wolf Blanckenhorn

and describe the animals’ mating system. A hypothesis potentially explaining the long tail of males was then formulated based on sexual selection: for some reason females prefer to mate with long-tailed males. From this hypothesis predictions could be derived to be tested in an experiment: the mating success of males with artificially enlarged (shortened) tails should increase (decrease) relative to that of un- or mock-manipulated controls, independent of male quality. With such an experiment, sexual selection could be documented for one particular trait in this particular species. In fact, this experiment yielded not only a qualitative demonstration of sexual selection by female choice, but also a quantitative estimate of its strength: by how much did male mating success increase with tail length? However, Andersson (1982) did not calculate this, although he did in a later study on a related bird (Andersson 1989). This simple but clever experiment thus showed that natural selection can be quantified in nature. Why do we want to quantify natural selection? Evolutionary theory and also intuition tells us that the evolutionary change in a given trait X, e.g. tail length in the preceding example, i.e. the evolutionary response to selection R, is proportional to the heritability of the trait h2 and the strength of phenotypic selection S: R = h2S (Falconer 1989). Thus, if we can quantify the heritability of a trait (see Mousseau and Roff 1987) and selection on it, both of which are measurable entities, we can quantitatively predict the response to selection and ultimately the evolutionary change of a trait in a quantitative genetic framework. Although the theory of quantitative genetics was well elaborated by animal and plant breeders during the 20th century, including the methods for estimating genetic variation underlying quantitative traits (Falconer 1989, Lynch and Walsh 1998), it was not until the early 1980s that a handful of seminal papers by Lande and Arnold (1983) and Arnold and Wade (1984a,b; see also Manly 1985) outlined standardised statistical methods for estimating natural selection in nature. This has made possible quantitative comparisons among species, traits and environments at a new level, providing a database for powerful meta-analyses such as those by Endler (1986) or Kingsolver et al. (2001). After all, such meta-analyses of the literature are essential for ultimately evaluating the importance of natural selection in evolution. To give a hypothetical example: if pre-copulatory sexual selection by female choice has been demonstrated in 80 out of 100 study species, while male-male competition only occurs in 50 out of 100, and post-copulatory sperm competition only in 20 out of 100 cases, it can be concluded that pre-copulatory female choice is globally a more important selection mechanism than either male-male or sperm competition. Note that any single species study is only one data point at this level. It is not sufficiently appreciated that standardised selection measures often can

The quantitative study of sexual and natural selection

303

also be calculated for behavioural studies of sexual selection conducted in the field or even at semi-natural or artificial conditions in the laboratory. Unfortunately, this is rarely done. In this chapter I discuss the continuum of approaches that exist when studying sexual or natural selection in the wild and the laboratory. These range from behavioural observations, via experimental manipulations of particular traits or environments, to phenomenological studies in nature. Focus here is on the study of body size and related life history traits, such as growth rate, development time or survival, drawing on particular examples from our own studies on dung flies. I advocate and start with explaining the use of standardised selection measures in this context, as a practical guide for the unfamiliar reader. My major message is to urge researchers studying the behavioural mechanisms underlying sexual selection to calculate standardised selection coefficients whenever possible, as this typically cannot be easily done retrospectively from the data presented in a given paper. Such practice would tremendously broaden the database available for any systematic comparative meta-analyses of particular traits, species and evolutionary phenomena.

11.2 Defining natural selection and fitness In brief, natural selection with regard to a certain trait results when individuals expressing a superior phenotype enjoy greater fitness (defined below), leading to a change in the phenotypic distribution of the trait within a generation. If the trait has a heritable (genetic) component, the selected individuals pass on more genes into the next generation relative to conspecifics expressing an inferior phenotype, leading to changes in the trait distribution over generations, i.e. evolution. Any trait can be selected, for example colouration or body size (morphological traits), the intensity or duration of a mating display (a behavioural trait), immunocompetence (a physiological trait), or even particular alleles (a genetic trait). As selection acts on phenotypes, it merely requires trait and fitness variation, irrespective of the genetic background of the trait. For evolution to result by natural selection, the trait requires a heritable basis. Evolution by random genetic drift only requires (phenotypic) trait variation and trait heritability, fitness (say, mortality) being per definition random, whereas evolution by natural selection additionally requires trait-mediated fitness differences and an identifiable selective cause.

304

Wolf Blanckenhorn

Table 11.1 Hierarchy of fitness estimators, decreasing in precision from top to bottom but at the same time increasing in practicability. Hierarchy of fitness estimators Population (and species) fitness components  (expected) time to extinction  r: intrinsic rate of increase Individual fitness components  number of grand-offspring  number of offspring = lifetime reproductive success (R0 = ert) lx (survival)  longevity  winter survival  energy reserves before winter  body mass increase  food accumulated  foraging time

mx (fecundity)  offspring/lifetime  offspring/season  clutch size  energy investment/ clutch  time investment/clutch

mating success  matings per lifetime  number of mates/day  mating probability  mated/not mated once

Most generally, fitness can be defined as the relative contribution of a certain unit of selection (e.g. a species, an individual, a gene, etc.) to the subsequent generation. Strictly speaking, fitness is only defined for a particular class of individuals, but the definition can be easily extended to groups of individuals. There are a multitude of fitness estimators depending on the field, which can be arranged hierarchically in terms of quality, precision, and practicability (Endler 1986; Table 11.1). The (expected) time to extinction of a unit of selection (a species or individual) is arguably the best fitness estimator available, but it is impractical for most biologists except perhaps paleontologists. Ecologists use the intrinsic rate of increase r of a population, as defined by the Euler-Lotka equation, ∑ e-rx lx mx=1, but this is only practical for small, fast-growing organisms such as bacteria, protists or water fleas. Behavioural ecologists commonly consider the individual as the unit of selection, for which lifetime reproductive success is the ideal fitness measure. However, even this is unattainable for most organisms in the field except very large ones (Clutton-Brock 1988). More typically, people estimate at times rather distant sub-components (Table 11.1) of one of the three fitness components of individual reproductive success: survival (the likelihood of reaching a particular stage or age x: lx in the Euler-Lotka equation above), fecundity (i.e. number of offspring:

The quantitative study of sexual and natural selection

305

Selection Natural selection Artificial selection (breeding by humans) Sexual selection Intra-sexual selection (mate competition)

Non-sexual (natural) selection Inter-sexual selection (mate choice)

(b) Stabilising

Fecundity selection

(c) Disruptive

Fitness

Frequency

(a) Directional

Viability selection

Trait (e.g. body size)

Fig. 11.1 Two common classification schemes of natural selection based on the individual fitness components (top) and based on its outcome (bottom).

mx) and, primarily for males, mating success. It should be obvious that male (age-specific) fecundity mx is some product or sum of his mating success (e.g. his number of mates) and his mates’ offspring number. The three sub-components of individual fitness also give rise to a common classification of natural selection (Fig. 11.1, top), so there are viability (or survival), fecundity and sexual selection. Sexual selection can be further subdivided into intra-sexual selection, i.e. competition among members of the same sex for access to the other sex, and inter-sexual selection, largely mate choice of members of one sex by members of the other sex (Fig. 11.1, top). Other synonyms are in use, such as epigamic selection, but the ones presented in Fig. 11.1 are most commonly used. Also, natural selection is sometimes equated with non-sexual selection and opposed to sexual selection (cf. Darwin 1859, 1871; see e.g. Endler 1986). It is best to stick to the precise terms shown in Fig. 11.1. If you subscribe to

306

Wolf Blanckenhorn

the anthropocentric view that sets humans apart from nature, artificial selection can be defined separately (Fig. 11.1, top); if not, such plant or animal breeding orchestrated by humans is natural selection just as well. A second common classification is based on the evolutionary outcome of selection, differentiating between directional, stabilising (or balancing), and disruptive selection (Fig. 11.1, bottom). Again, there are synonyms such as positive (instead of directional) or diversifying (instead of stabilising) selection used by molecular biologists.

11.3 Measuring natural selection In principle, the quantitative assessment of selection is straightforward and based on regression techniques that most behavioural ecologists should be familiar with. Technical descriptions are available in Lande and Arnold (1983), Arnold and Wade (1984a,b), Brodie and Janzen (1996) and Brodie et al. (1995), as well as many evolution textbooks. Generally, any fitness measure (Table 11.1), the dependent variable (Y), is regressed upon one (univariate) or several (multivariate) metric trait(s) (Xi; e.g., body size, mating display intensity, colouration, etc.), the explanatory variables putatively affecting fitness. Selection coefficients can only be calculated if the traits Xi are continuous, ideally normally distributed, while fitness Y can be expressed in various ways. To directly compare the resulting selection coefficients among species, traits and fitness components, however, the Y and X variables must be standardised in a specific way as described below. Confusingly perhaps, there are various selection coefficients (which is the general, non-technical term) used in the literature. I define most of them below (and in BOX 11.1) and explain their relationships. Before starting, two statistical comments are necessary. First, although in principle the regression approach can be universally used, for statistical reasons the preferred method depends on the nature of the dependent (Y) variable. Thus, fitness can be measured in a binary way (dead or alive; mated or not), in which case binary logistic regression must be used. If fitness is instead expressed by a real or ordinal (countable) number (such as number of offspring produced), regular least-squares regression must be used. If only few outcomes are possible, such as when counting the number of mates when mating is a rare event (e.g. 0, 1, 2 or 3 mates), multinomial or loglinear approaches are needed. In modern statistical terms, however, these are all General Linear Models (GLMs) with different error distributions (binary, normal, multinomial, etc.). Second, unless purely descriptive, any statistical method has two elements to it: the estimate and the significance

The quantitative study of sexual and natural selection

307

test. Standard regression thus estimates the slope of Y on X, and the corresponding significance t-test tests whether the slope is different from zero (as the default), but potentially also differences from one or any other a priori prediction. Interestingly, evolutionary biologists and physiologists are often more interested in the estimates, whereas behavioural ecologists traditionally put more emphasis on significance testing, often not even reporting the corresponding estimates or presenting them in a useless form (e.g. unstandardised, scale-dependent regression coefficients instead of standardised (partial) correlation coefficients). It is important to keep in mind that these two statistical elements are equally important and always should come together. The standardised selection coefficients discussed below are the estimates, which come with a statistical test of significance. 11.3.1 Univariate linear selection gradients and differentials It is widely known that larger females of most ectothermic species produce more offspring (Blanckenhorn 2000). As the number of offspring is one of the central fitness measures (Table 11.1), this by itself already constitutes directional selection on (here female) body size. Linear regressions of clutch size on body size (here estimated by head width) for samples from three Swiss populations of the black scavenger or dung fly Sepsis cynipsea are plotted in Fig. 11.2 (left-hand side; data from Blanckenhorn et al. 1999a). If both the fitness (Y) variable and the trait (X) are transformed appropriately, these data yield standardised measures of the strength of fecundity selection for each population generally called selection gradients (Lande and Arnold 1983, Arnold and Wade 1984a,b; Fig. 11.2, right-hand side). The trait X, here head width, is normalised by deducting the population mean (average) from each individual value Xi and dividing by the population standard deviation (SD), to yield a so-called z-score: zi = (Xi – mean(X))/SD(X). This transforms the trait distribution into a standard normal distribution with mean = 0 and SD = 1, which most statistical programs will do automatically for any sample specified. Fitness Y, here estimated by clutch size and generally denoted by w, is transformed differently by dividing each individual value by the population mean to yield the relative fitness: w’ = wi / mean(w). While the slopes of the linear regressions of the raw data on the left in Fig. 11.2 denote the number of additional eggs produced per unit increase in head width for females of each population, the corresponding slopes on the right are dimensionless and denote the (standardised) intensity i of directional fecundity selection favouring larger female body size (because the relationship is positive), in standard deviation units.

308

Wolf Blanckenhorn

BOX 11.1 Calculating univariate linear and non-linear selection coefficients. Univariate selection gradient via regression. To calculate linear and nonlinear selection coefficients using the regression method (i.e. gradients) when only one predictor (x) variable is used (e.g. body size), first produce standardised z-scores for x by subtracting the sample mean from each individual value xi and dividing the difference by the standard deviation: z i  xi  x  / SD x . For binary pairing success or mortality (paired/unpaired; dead/alive), then calculate relative fitness w’ as absolute fitness w, i.e. pairing success or survival (1 or 0), divided by the mean fitness w of the entire, representative sample, which is akin to the operational sex ratio (mated males (= number of mated females)/all males) or the proportion of surviving individuals, respectively: w' wi / w . If necessary, estimate these proportions prior to sampling, take any (non-random) sample, and later adjust according to Blanckenhorn et al. (1999b). The procedure is analogous for any continuous fitness measure such as the number of offspring (clutch size; cf. Fig. 11.2). Use the univariate model of relative fitness on standardised trait size w'  c  1 z to estimate the univariate linear selection gradient  1 , and then use the model w'  c   '1 z   2 z ² to estimate the univariate non-linear (quadratic) selection gradient   2  2 , which quantifies the curvature with the linear component statistically controlled (Arnold and Wade 1984a,b). As a rule,  '1  1 . The figure below indicates that, for Bumpus’ (1899) data displayed below, there is no directional viability selection on body size (nonsignificant linear selection gradient  1 ) but instead stabilising selection (significant negative non-linear selection gradient   2  2 ). Univariate selection differential approach if fitness is binary. Calculate zscores as above. Then take only the mean z survive of the 21 surviving birds with absolute fitness 1 with its associated standard deviation SD( z survive ), from which the 95% confidence interval (CI) can be derived as CI = 1.96* SD( z survive )/ n  1 (see light grey underlain numbers in data set below). The resulting selection differential S is the same as 1 above.

Relative surviorship

2.5 2.0 1.5 1.0 0.5 0.0 -3

-2

-1

0

1

Z-score body size

2

3

The non-significant linear viability selection gradient (black line with 95% CI) indicating lack of directional selection for body size, and the significant nonlinear gradient indicating stabilising selection on body size (red; presumably via flight ability), for Bumpus‘ (1899) data on house sparrow (Passer domesticus) adult mortality after a severe storm.

The quantitative study of sexual and natural selection

309

Sample data set: Bumpus’ (1899) mortality of sparrows after a severe storm. Dead/alive Rel. survival 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Mean SD 95% CI N

0.429

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333 2.333

Body size

Z(body size)

2.987 3.051 3.061 3.100 3.105 3.116 3.129 3.163 3.165 3.169 3.183 3.186 3.186 3.192 3.201 3.201 3.220 3.230 3.232 3.234 3.246 3.248 3.260 3.298 3.300 3.307 3.363 3.379 3.101 3.126 3.143 3.150 3.161 3.161 3.170 3.192 3.193 3.202 3.205 3.208 3.211 3.212 3.222 3.229 3.246 3.252 3.276 3.284 3.289

-2.791 -1.959 -1.818 -1.301 -1.238 -1.100 -0.923 -0.472 -0.441 -0.392 -0.217 -0.172 -0.167 -0.086 0.022 0.028 0.284 0.405 0.439 0.464 0.625 0.645 0.805 1.303 1.340 1.432 2.165 2.374 -1.298 -0.966 -0.734 -0.642 -0.495 -0.495 -0.376 -0.088 -0.075 0.035 0.084 0.118 0.153 0.173 0.306 0.395 0.618 0.703 1.014 1.129 1.186

3.199 0.076 0.021 49

S = 0.036 0.674 0.296 21 Survivors

310

Wolf Blanckenhorn 100 80

2

Luzern

60

1

40

0,5

20 0.8

0.9

1

1. 1

60 40 20

100

0.8

0.9

1

1. 1

1.2

-3

-2

-1

0

1

2

3

2

3

2

3

Obfelden: S = +0.122

1,5 1 0,5 0

-4

-3

-2

-1

0

1

2 Fehraltorf: S = +0.118

Fehraltorf

80

1,5

60

1

40

0,5

20 0 0.7

-4

2

Obfelden

0 0.7

0

1.2

Relative clutch size

Clutch size

0 0.7 100 80

Luzern: S = +0.097

1,5

0.8

0.9

1

1. 1

Head width [mm]

1.2

0

-4

-3

-2

-1

0

1

Z-score head width

Fig. 11.2 Unstandardised relationship of female fecundity (clutch size) with body size (left) and the corresponding standardised fecundity selection differentials S (in standard deviation units; right) for three populations of the black scavenger fly Sepsis cynipsea (from Blanckenhorn et al. 1999a).

Of course, each slope comes with a standard error or confidence interval and a corresponding test of statistical significance. The slopes in Fig. 11.2 slightly differ between the populations, and if we enter population as a grouping factor in the statistical model, the interaction term between population and the trait X tests whether the fecundity selection intensities differ significantly between the populations or not (which is not the case in Fig. 11.2). Note that the significance test does not change due to the transformation performed, i.e. it yields the same result for the raw as for the transformed data, but the estimate crucially does. The data in Fig. 11.2 describe directional selection, because the relationship between fitness and the trait is monotonically increasing and linear, as in Fig. 11.1a, bottom left. Non-linear fitness-trait relationships are of course possible, and it is standard to model stabilising and disruptive selection (Figs. 11.1b, c, bottom) by negative and positive quadratic functions,

The quantitative study of sexual and natural selection

311

respectively. Thus, by adding a quadratic term X2 in the statistical model, as explained in BOX 11.1, we can additionally estimate and test for nonlinear selection. Note that the linear term in the model with the quadratic term will necessarily be different from the linear estimate in the model without the quadratic term. Thus the convention is that directional selection is first estimated using the simple linear model, whereafter non-linear selection is estimated using the model with both the linear and quadratic terms, only paying attention to the quadratic and ignoring the linear estimate (e.g. Fairbairn and Preziosi 1994). What if fitness is not continuous as in Fig. 11.2? If fitness is binary or multinomial, the statistical model for the significance test has to be modified, as stated above. Nevertheless, the same regression method can be applied to produce the estimate. An example is Bumpus’ (1899) famous data set on the mortality of sparrows after a severe storm in relation to their morphology (see BOX 11.1). Of 49 birds, 21 (given absolute binary fitness 1) survived and 28 died (given absolute fitness 0). Thus, assuming this is a random sample, the average fitness mean (w) = 21/49 = 0.429, which is also the proportion of surviving individuals. The two possible relative fitnesses of individuals are 0 or the inverse of mean fitness, 1/0.429 = 2.33. In this case it is immediately obvious that the standardised data produce a steeper relationship between fitness and the trait than the raw, absolute fitness (1/0) data (because 2.33 > 1). While logistic regression needs to be used to test for the statistical significance of the slope (using the raw or standardised data), regular least-squares regression on the standardised data can be used to estimate selection intensity. However, Janzen and Stern (1998) present an equivalent method allowing back-calculation of the selection intensity from the logistic slope estimate. Further and in general, randomisation techniques can, of course, also be used for significance testing (e.g. Manly 1985, Mitchell-Olds and Shaw 1987). Only if fitness is binary, as just discussed, is there a simpler, univariate method originally derived from classical quantitative genetics (Falconer 1989). Assuming the trait X (e.g. body size) is normally distributed and we have truncation selection (dead = not selected = 0 and alive = selected = 1, or mated and not mated, etc.), the selection differential is the mean trait value of the selected proportion of individuals (e.g. the largest ones) minus the mean trait value of the whole sample: S = mean(XSel) – mean(XPop). This value is yet unstandardised and in trait units, but by z-scoretransforming all data as above, standardisation is again achieved. As per definition the population mean(z) is zero, the selection differential is then equal to the mean standardised selected subpopulation mean(zSel), thus becoming identical to the selection gradient derived from regression as described above (see BOX 11.1).

312

Wolf Blanckenhorn 3.43

500 300

400

S = +0.32

300

100

200

Unpaired

200

Paired

Frequency

500

3.53

100

Selection differential

400

100

0

0

100 200 300 400

3.39

500 2.0

2.5

3.0

200 300 400 500

3.5

4.0

4.5

-4

Pairing status 0

-3

-2

-1

0

1

2

3

4

3

Selection gradient

Relative pairing success

1

2

1

S = +0.32

0 2.0

2.5

3.0

3.5

4.0

Male hind tibia length [mm]

4.5

-4

-3

-2

-1

0

1

2

3

Z-score male hind tibia length

4

Fig. 11.3 Unstandardised (left) and standardised (right) (a) frequency distribution of paired (top) and unpaired (bottom) male yellow dung flies (Scathophaga stercoraria) as a function of their body size in Fehraltorf, Switzerland, over 2 years; and (b) the corresponding relationship of absolute (paired = 1, unpaired = 0; left) and relative pairing status (ca. 31% of all males were paired) with body size (data from Jann et al. 2000). Both methods yield the same selection intensity of S = +0.320 (differential top; gradient bottom).

Figure 11.3 illustrates the two procedures to arrive at the same sexual selection intensity estimate for pairing success, using data from Jann et al.’s (2000) longitudinal study of yellow dung flies. I must re-emphasise that in case of binary fitness measures the selection intensity depends crucially on correct estimation of the proportion of mated (or live) vs. unmated (or dead) individuals in the population, because the slope of the regression in Fig. 11.3, bottom right, becomes steeper when this proportion (i.e. mean fitness) decreases. This should be intuitive, as fitness is a relative measure and a particular male should have higher fitness if he out-competes more other males. The sample taken thus needs to truly reflect this proportion (Arnold and Wade 1984a,b). Luckily, the selection estimate can be corrected if the sample does not reflect the true proportion (e.g. when collectors focus on copulating pairs because they are rare and more difficult to collect), but if an independent estimate of the true proportion of mated in-

The quantitative study of sexual and natural selection

313

dividuals is otherwise available (Blanckenhorn et al. 1999b). For the interested reader, Blanckenhorn et al. (1999b), but also e.g. Brodie and Janzen (1996) or Freeman and Herron’s (2007) Evolutionary Analysis textbook, derive the relationships between the different selection coefficients. 11.3.2 Multivariate linear and non-linear selection gradients Two extensions of the regression logic outlined above are obvious. First, selection on several traits Xi can be estimated at the same time as in standard multiple regression. Such multivariate approaches to selection are preferable to simple univariate studies, as quantitative partitioning of selection provides more detailed insights into the precise mechanism of selection. For example, is it larger body size or mass per se that confers an advantage in predator avoidance for a flying insect, or is it the size or shape of the wing relative to body size that does so? Second, the relationship between fitness and the trait may take any non-linear shape. In particular, stabilising and disruptive selection (Fig. 11.1b, c, bottom) can be modeled using negative and positive quadratic functions, respectively. I shall now describe these methods in turn. BOX 11.2 Calculating multivariate linear, non-linear and correlational selection coefficients. Produce standardised z-scores for all traits xi as described in BOX 11.1 by subtracting the sample mean from each value and dividing by the standard deviation: zi  xi  x  / SDx . Calculate relative fitness w’ as absolute fitness w divided by the mean fitness w of the entire, representative sample: w' wi / w . Use the model w'  c    multi,i zi for multivariate linear selection gradients, and the model w'  c    ' multi,i z i  0.5  multi,i z i2    i , j z i z j

for multivariate pair-wise correlational,  i, j , and non-linear (quadratic) selection gradients,  multi,i . While the linear coefficients as calculated in BOX 11.1 reflect the combined effects of direct and indirect selection on body size (Endler 1986), the multivariate coefficients calculated here estimate direct selection on each trait with selection on correlated traits statistically removed.

314

Wolf Blanckenhorn

By assessing several traits at once, selection on, for example, male body size and the duration of his mating display can be estimated simultaneously (BOX 11.2). As in any multiple regression, this yields independent selection coefficients for each trait. By doing so, correlational selection can additionally be estimated (Lande and Arnold 1983, Brodie et al. 1995). Intuitively, if two traits are correlated and both affect fitness, e.g. here if larger males also show longer mating displays, then selection on one trait (e.g. mating display duration by female choice) will produce an indirect, correlated change in the other trait (here body size) even if there is no direct selection on the second trait. This is what is being estimated by correlational selection when entering a product (or interaction) term between two Xi (BOX 11.2). Correlational effects are highly biologically relevant, particularly if selection on two traits that are positively correlated goes in opposite directions. For example, under drought conditions viability selection in the medium ground finch (Geospiza forti) favours deep but narrow beaks with which birds can crack the remaining large and strong seeds best. However, as is typical, both morphological traits are positively genetically correlated, so birds tend to have deeper and wider beaks (Grant and Grant 1995). It should be intuitive that direct selection for narrower beaks (yielding a negative directional selection gradient) will be counteracted by the correlated selection favouring wider beaks (positive correlational selection gradient) that occurs because deeper beaks are favoured by directional selection (positive direct selection gradient). As a result, the response to selection will be constrained until the genetic correlation between the two traits is broken down. Of course, this approach can be extended to incorporate more than two traits as well as quadratic terms estimating stabilising or disruptive selection, as described next. By entering a quadratic term in addition to a linear (and an interaction) term, stabilising and disruptive selection (Fig. 11.1b, c, bottom) can be estimated (BOX 11.2). These two types of selection are mutually exclusive for any given trait, a negative coefficient implying stabilising and a positive coefficient disruptive selection. Furthermore, linear and quadratic terms can both be (independently) significant, thus indicating a combination of linear and non-linear selection, for example asymptotic or accelerating selection (Fig. 11.4). It should be clear that any complex relationship between fitness and a trait is possible. Such relationships are best visualised by non-parametric regression (Phillips and Arnold 1989, Brodie et al. 1995, Blows 2007), e.g. using Schluter’s cubic spline software (http://www.zoology.ubc.ca/~schluter/software.html). Finally, it should be mentioned that historically there are other quantitative measures of selection in the literature that may be relevant in these contexts, such as e.g. the opportunity for selection (Crow 1958, Shuster

The quantitative study of sexual and natural selection Scathophaga stercoraria

315

Sepsis cynipsea

100

Clutch size

80 60 40 20 0

1.5

2

2.5

3

3.5 0.7 Hind tibia length

0.9

1.1

1.3

Fig. 11.4 Unstandardised relationship of female fecundity (clutch size) with body size showing accelerating (left, for the yellow dung fly Scathophaga stercoraria; from Jann et al. 2000) and asymptotic selection (right, for the black scavenger fly, Sepsis cynipsea; from Blanckenhorn et al. 1999a). In both cases the linear and non-linear (quadratic) terms were significant.

and Wade 2003). These are often mathematically related to the aforementioned coefficients but nevertheless different, so the user has to be careful not to confuse the various measures. For the interested reader, further extensions of the statistical approaches described here also exist (Blows 2007, Hunt et al. 2009).

11.4 Study types conducive to estimation of selection coefficients 11.4.1 Phenomenological selection studies in nature Besides the classic phenomenological field studies of selection, standardised selection coefficients can in principle be calculated for a wide variety of studies in behavioural ecology, and this would be useful if only for comparative purposes. These various study types can be loosely categorised as done in Table 11.2, which I shall now discuss, using primarily our own studies on dung flies with an emphasis on body size as examples, although any analogous studies on other taxa or traits would serve just as well. Note that we also did not supply selection coefficients in several of these studies even though we could have.

316

Wolf Blanckenhorn

Table 11.2 Types of behavioural studies principally conducive to calculating selection coefficients for comparative purposes. Study type Experimental unit/ replicate

Fitness component

Phenomenological

Pairing success Field

Non/ Population

Blanckenhorn et al. 1999a

Pairing success Field

None/ season; year

Jann et al. 2000

Pairing success Field

Tail length/ none

Andersson 1982

Survival

Field

None/ season; year

Grant and Grant 1995

Survival

Lab

Predation/ wing damage

Mühlhäuser and Blanckenhorn 2002

Pairing success Lab

Nutrition/ energy depletion

Blanckenhorn et al. 2008

Fecundity; longevity

Mating system/ laboratory line

Martin and Hosken 2003

Pairing success Lab

Sex ratio/ none

Blanckenhorn et al. 2000

Pairing success Lab

Light conditions/parasitism

Milinski and Bakker 1990

Pairing success Lab

None/ species

Abt and Reyer 1993

Pairing success Lab

None/ none

Blanckenhorn et al. 2000

Fecundity; longevity

Lab

None/ copulation number

Blanckenhorn et al. 2000

Fecundity; longevity; pairing success

Lab

Nutrition/ energy depletion

Blanckenhorn et al. 1995

Population

Population- Group based

Individualbased

Pair

Individual

Field/ Manipulation/ lab other factors

Lab

Example

Classic studies document selection as it occurs in the wild and are therefore at the heart of evolutionary ecology. As they are typically not manipu-

The quantitative study of sexual and natural selection

317

lated, they are phenomenological, meaning that fecundity, viability or sexual selection are assessed from field samples without regard to the underlying behavioural mechanisms, e.g. whether mating success is mediated by female choice or male-male competition. In practice a sample of mated and unmated males (e.g. Jann et al. 2000), or a number of live and dead birds (Bumpus 1899, Grant and Grant 1995) are collected, on which any measurements, often relating to morphology but also e.g. colouration, can be taken to investigate which of these traits influence fitness (i.e. here, pairing success or survival). Such studies have been reviewed by Endler (1986), Andersson (1994) and, most recently, by Kingsolver et al. (2001) and Cox and Calsbeek (2009). It must be emphasised that there are fewer such phenomenological selection studies than one might expect, and the existing studies are often incomplete in terms of fitness components and certainly not a random sample. A major reason for the latter is that many species do not occur in sufficiently high numbers in a spatio-temporally aggregated fashion to make such a study practical, as substantial sample sizes are typically required because of large natural variation (Arnold and Wade 1984a,b, Palmer 2000, Kingsolver et al. 2001). For example, if a species is very secretive and mates only rarely and/or for a very short time, sexual selection coefficients will be hard to come by because mating pairs cannot be found. Furthermore, not all fitness components are equally easy to assess in every species. While fecundity selection can be well studied in insects because they lay many eggs (e.g. Fig. 11.2), insect survival in the field is difficult to measure because they are so small (Blanckenhorn et al. 1999a). The reverse is true for larger mammals, which typically have few offspring but can be marked and tracked over long distances and time (Clutton-Brock 1988). Thus, field studies of selection in a given species rarely if ever assess lifetime reproductive success, arguably the best fitness estimate for an individual, but instead make do with more practical but also more distant fitness components lower down in the hierarchy (Table 11.1). Luckily, selection can be studied by a piece-meal or cross-sectional (as opposed to longitudinal) approach (Arnold and Wade 1984b; e.g. Blanckenhorn 2007), whereby the effects of a trait on fitness are investigated for a sample of individuals only at a particular life stage. Several such estimates at various stages, often called selection episodes, for various fitness components can ultimately be integrated, in the ideal case yielding a measure equivalent to lifetime reproductive success. For example, the effect of body size on pairing success would be one such sexual selection episode and, given pairing, whether larger males also mate with larger females, i.e. whether there is assortative mating, is the subsequent selection episode (e.g. Jann et al. 2000). (It should be obvious that both episodes influence

318

Wolf Blanckenhorn

male fitness, because male reproductive success depends on his number of mates and the number of offspring his mate produces with his sperm.) Because fitness is multiplicative, selection differentials are additive, and several selection episodes can so be strung together (Arnold and Wade 1984a,b; see e.g. Blanckenhorn 2007). Cross-sectional estimation of multiple selection episodes and fitness components makes a number of assumptions if it is to correctly reflect overall selection over the entire lifetime of an organism in a variable environment. Most of these assumptions relate to any particular field sample being truly random with regard to various potentially confounding variables. First, it assumes that any field sample correctly reflects the age structure of the population, thus including older and younger individuals of both sexes. Needless to say, the sexes should be assessed separately, although in the juvenile stage this is often not possible. Second, it assumes that there are no systematic effects of age on the fitness component estimated; otherwise early fecundity may systematically overestimate lifetime fecundity if fecundity diminishes with age. Third, selection depends strongly on the environmental conditions at the time. Therefore multiple samples at various times or environmental conditions at more than one place, or in several populations, are necessary to encompass the possibility of spatio-temporal variation in selection for a given species (Istock 1981). Assessment in more than one environment is particularly advised for any experimental estimation of selection in the field or laboratory. Fourth, selection at any life stage is contingent upon the probability of an individual reaching this life stage, so the magnitude of selection has to be adjusted for this probability (Blanckenhorn et al. 1999a, Blanckenhorn 2007). If these assumptions are not fulfilled, biased estimates of selection will result. In phenomenological selection studies, any natural or experimentally manipulated extraneous variable likely producing variation in selection can and should be entered in the statistical model as a factor or covariate to test for variation in selection intensity among treatments or populations via interaction terms, and to arrive at an average estimate of selection over several environments. For example, in Fig. 11.2, fecundity selection gradients, i.e. the slopes, do not differ among the three populations of black scavenger flies, so an average of the values presented for the three populations well represents fecundity selection of the species (Blanckenhorn et al. 1999a). As mentioned in the introduction, Andersson (1982) could have calculated sexual selection coefficients with regard to tail length in the long-tailed widowbird but did not do so. Note that because he manipulated tail length he produced three somewhat discrete tail-length categories (as opposed to continuous tail-length data on the X-axis; cf. Fig. 11.3), al-

The quantitative study of sexual and natural selection

319

Table 11.3 Available sex-specific mean ± 95% CI field estimates of linear sexual, fecundity or viability selection differentials or gradients on any morphological trait for a number of animal species (from Kingsolver et al. 2001 and Blanckenhorn 2007). There is only one available estimate for juvenile viability.

Fecundity or sexual selection

Adult viability selection

Females (N)

Males (N)

+0.127 ± 0.063 (13)

+0.255 ± 0.120 (22)

Differentials (unpaired data)

+0.079 ± 0.114 (8)

+0.333 ± 0.221 (16)

Gradients (unpaired data)

+0.149 ± 0.084 (9)

+0.256 ± 0.139 (9)

Invertebrates1 (paired data)

+0.063 ± 0.155 (6)

+0.239 ± 0.150 (6)

Vertebrates2 (paired data)

0.039 ± 0.113 (9)

+0.042 ± 0.073 (9)

All estimates and species

+0.072 ± 0.078 (9)

Not differentiated by sex

1

Allenomobius socius (Fedorka and Mousseau 2002), Aquarius remigis (Preziosi and Fairbairn 2000), Callosobruchus maculatus (Savalli and Fox 1999), Clibanarius dugeti (Harvey 1990), Gammarus pulex (Ward 1988), Plathhemis lydia (Koenig and Albano 1987), Scathophaga stercoraria (Jann et al. 2000, Blanckenhorn et al. 2003), Sepsis cynipsea (Blanckenhorn et al. 1999a, 2004), Stator limbatus (Savalli and Fox 1998); 2Crocidura russula (Boutellier and Perrin 2005), Carpodacus mexicanus (Badyaev et al. 2000), Geospiza conirostris (Grant 1985), Geospiza fortis (Price 1984), Niveoscineus microlepidates (Olsson et al. 2002), Parus major (Björklund and Lindén 1993)

though this still permits regression to calculate a selection gradient with regard to the number of mates obtained for the whole sample. Some patterns on phenomenological field selection studies emerge from the available data (Blanckenhorn 2000, 2005, 2007, Kingsolver et al. 2001, Cox and Calsbeek 2009). The vast majority of studies (ca. 80%) estimate selection with regard to morphological traits, generally reflecting body size. Selection data for physiological (e.g. Blanckenhorn et al. 2003, 2004) and behavioural traits are scarce in comparison, so I encourage all behavioural ecologists to generate such data. Even a common trait such as body size has been estimated for few species using several fitness components, and for fewer still (9 invertebrates and 6 vertebrates), simultaneous selec-

320

Wolf Blanckenhorn

tion estimates for both sexes exist (paired data in Table 11.3). Sexual or fecundity selection favouring larger male body size is generally stronger than (sexual or) fecundity selection favouring female body size (Table 11.3). Moreover, selection on body size rarely has a different direction in males and females (Cox and Calsbeek 2009). At the same time, adult and especially juvenile viability selection estimates are very rare and typically not different from zero, i.e. survival is largely independent of body size (Table 11.3). According to the differential equilibrium hypothesis of sexual size dimorphism (Blanckenhorn 2007), taken together these patterns would suggest that most species should have male-biased size dimorphism, because sexual selection on males exceeds fecundity selection on females and viability counter-selection is low and does not differ between the sexes. This is inconsistent with the data, as several of the species listed in Table 11.3 have female-biased size dimorphism, thus globally rejecting the hypothesis. 11.4.2 Population selection studies in the laboratory In nature, the scope for manipulations is limited and therefore selection mechanisms are difficult to assess. In such cases animals (such as Drosophila) may be kept and investigated in groups (or populations) in the laboratory in a manner resembling the field situation, although population sizes will necessarily be much smaller (Table 11.2). Thus, similar approaches as in the field can be used to generate selection coefficients. For example, Martin and Hosken (2003a,b) kept several replicate experimental evolution populations of Sepsis cynipsea in the laboratory either at naturally polyandrous conditions or under enforced monogamy, thus manipulating the mating system. Any estimates of, for example, body size dependent sexual, fecundity or viability selection generated with individuals of these populations would yield one selection coefficient (with its associated standard error) per laboratory population replicate. As Martin and Hosken (2003a,b) had three replicate populations per mating system, they could have ended up with 2 times 3 population estimates, thus documenting variation in selection between the mating systems (which was not done). Mühlhäuser and Blanckenhorn (2002) provided a second example. Predation is difficult to observe and manipulate in the field. A number (24) of S. cynipsea individuals of various sizes and two wing damage treatments (wings clipped and unclipped) were exposed to a predator (a yellow dung fly female) in each of 24 replicate containers (= populations; 12 with females and 12 with males). Although body size and behaviour were not measured in this experiment, given that the predator only ate a subset of

The quantitative study of sexual and natural selection

321

the flies each container would have yielded two size-dependent viability selection coefficients (one for the wing-damaged and one for the undamaged subset of flies). The whole experiment would have yielded 12 (males) + 12 (females) such pairs of replicate coefficients, allowing for a test of size-dependent predation with sex and wing-damage as additional factors. Teuschl et al. (2010) adopted a similar approach and calculated replicate selection coefficients. A final example is provided by Blanckenhorn et al. (2008), who simulated mating competition among 9 Scathophaga stercoraria males of different sizes on a dung pat in an enclosure in the laboratory, to which a limited number of females was added. Thus, each replicate container yielded one sexual selection coefficient (for pairing success) with regard to body size. As there was an additional nutrient treatment of the males, again difficult to perform in nature, this laboratory study assessed size-dependent mating success as a function of nutritional state at semi-natural conditions. Note that all the above studies produced replicate selection coefficients per container or population. Although each coefficient comes with its own standard error, this becomes largely irrelevant, as the group/container/population becomes the statistical unit of interest, and a mean selection coefficient with its associated standard error based on the number of group/container/population replicates is being generated (Table 11.2). 11.4.3 Individual-based selection studies in the laboratory In behavioural ecology it is rather common to perform choice experiments where (typically) a female is allowed to choose between a pair of males that may or may not interact, simulating female choice in nature (Table 11.2; e.g. Milinski and Bakker (1990) for sticklebacks; Abt and Reyer (1993) for frogs; Blanckenhorn et al. (2000) for S. cynipsea). Often these males differ in size, although differences in mating display, mating attempts, calling rates or parasitism are sometimes additionally tested. Here the group consists of two individuals only. Nevertheless, a primitive, twoindividual selection coefficient can be calculated for each pair of males, which is equal to the (standardised) deviation of the winner’s (body size, display, etc.) score from the mean of the two males. As the choice tests are replicated, several such coefficients will again yield a mean selection coefficient with respect to body size, behaviour, etc., in principle also allowing multivariate approaches as outlined above if sample sizes are high enough. The simplest situation conceivable consists of laboratory assessment of individuals (Table 11.2). If, for example, a sample of individuals are held singly or in groups and e.g. their fecundity or longevity in the laboratory

322

Wolf Blanckenhorn

GLOSSARY Evolution: A change in gene frequencies over generations affecting any trait Fitness: The relative contribution of a unit of selection (e.g. a species, an individual, a gene, etc.) to the subsequent generation. Best calculated as lifetime reproductive success when studying individuals (cf. Table 11.1) Fitness component: Any (incomplete) subcomponent of fitness only estimating part of fitness (cf. Table 11.1) Heritability h2: The proportion of the total phenotypic variation in a quantitative trait that can be attributed to genetic variation, typically within a given population Natural selection: Cf. Fig. 11.1 Quantitative genetics: A biological field of study of the genetics underlying continuous or quantitative traits, i.e. those that are affected by multiple genes. Traditionally a statistical (phenomenological) approach is used without necessarily identifying all underlying genes Phenomenological selection studies: Typically field studies of natural selection quantifying selection without necessary reference to the underlying mechanism. Sexual selection: Cf. Fig. 11.1 Selection coefficient: A generic term subsuming all terms below Selection differential: A quantitative measure of the strength of selection on a single trait calculated as the difference between the (selected) individuals leaving offspring into the next generation and the overall phenotypic population mean Selection episode: Selection during a single event, point in time or life history stage. Subcomponent of total, lifetime selection Selection gradient: A standardised measure of the strength of selection defined as the covariance between relative fitness and the standardised trait value (cf. Fig. 11.2). Typically estimated as the slope of a uni- or multivariate regression. Selection intensity: The standardised selection differential (see above), calculating by dividing the selection differential by the phenotypic standard deviation (SD) Selection balancing: Selection process whereby both extreme phenotypes have lower fitness (cf. Fig. 11.1, bottom) correlational: In bi- or multivariate selection studies estimating the interactive non-linear selection component between two traits (cf. BOX 11.1 and 11.2) directional: Selection process whereby one of the extreme phenotypes has higher fitness (cf. Fig. 11.1, bottom) disruptive: Selection process whereby both extreme phenotypes have higher fitness (cf. Fig. 11.1, bottom) diversifying: Same as balancing selection (used by molecular biologists)

The quantitative study of sexual and natural selection

323

Selection epigamic: Same as intra-sexual selection fecundity: Selection referring to the fitness component fecundity (number of offspring; cf. Fig. 11.1, top) intra-sexual: Type of sexual selection whereby individuals of the same sex (typically males) compete for access to the other sex. Also called male-male competition inter-sexual: Type of sexual selection whereby individuals of one sex (typically females) choose members of the other sex. Also called female choice non-linear: Selection estimating the non-linear, curvature component indicating stabilising, disruptive (also called quadratic; cf. Fig. 11.1, bottom), asymptotic (cf. Fig. 11.4), or correlational selection (cf. BOX 11.1 and 11.2). linear: Selection estimating the linear, directional component (cf. BOX 11.1 and 11.2) positive: Same as directional selection (as used by molecular biologists) stabilising: Same as balancing selection survival: Selection referring to the fitness component survival or survivorship (the inverse of mortality; cf. Fig. 11.1, top) viability: Same as survival selection

(under any treatment) is investigated, one can obtain fecundity or longevity selection data for the sample. This is how the data in Fig. 11.2 were generated, producing one fecundity or viability selection coefficient per population (Blanckenhorn et al. 1999a; see also Blanckenhorn et al. 1995, 2000 for analogous data sets). If females are supplied with a male (of a given size, performing particular behaviours), this further yields size- or behaviour-dependent pairing success data for males and females, as some females will mate and others not, from which again one sexual selection coefficient can be generated (Blanckenhorn et al. 1995, 2000). Note that in general calculating selection coefficients requires little extra effort besides standardisation, as the statistical models to test for effects are performed anyway.

11.5 Conclusions In this chapter, I described and discussed the available standardised quantitative methods to estimate natural (including sexual) selection in the field and the laboratory. In so doing, the advantages of investigating selection have been addressed, particularly with regard to its underlying behavioural

324

Wolf Blanckenhorn

mechanisms, with a continuum of laboratory and field approaches (Table 11.2). I hope that this treatise helps particularly the unfamiliar readers to familiarise themselves with such approaches, also in practice. My further hope is that these methods become more widely used, especially in the field of behaviour in the context of sexual selection. Although most of my examples treat body size and morphology, as does the available literature (Kingsolver et al. 2001), the methods can be easily transferred to any quantifiable behavioural or physiological trait. In this day and age of metaanalyses the scientific community is seriously discussing data banks for ecological data, similar to those for molecular data, and some do exist already (e.g. http://datadryad.org/repo/handle/10255/dryad.167). As behavioural and ecological data are inherently variable, idiosyncratic and often species and study specific, any available standardisation, such as selection coefficients, would make such an endeavour more feasible. One may hope to see more selection coefficients also in the behavioural literature.

References Abt G, Reyer H-U (1993) Mate choice and fitness in a hybrid frog – Rana esculenta females prefer Rana lessonae males over their own. Behav Ecol Sociobiol 32:221-228 Andersson M (1982) Female choice selects for extreme tail length in a widowbird. Nature 299:818-820 Andersson S (1989) Sexual selection and cues for female choice in leks of Jacksons’s widowbird. Behav Ecol Sociobiol 25:403-410 Andersson M (1994) Sexual Selection. Princeton University Press, Princeton Arnold SJ, Wade MJ (1984a) On the measurement of natural and sexual selection: theory. Evolution 38:709-719 Arnold SJ, Wade MJ (1984b) On the measurement of natural and sexual selection: applications. Evolution 38:720-734 Badyaev AV, Hill GE, Stoehr AM, Nolan PM, McGraw KJ (2000) The evolution of sexual size dimorphism in the house finch. II. Population divergence in relation to local selection. Evolution 54:2134-2144 Björklund M, Lindén M (1993) Sexual size dimorphism in the great tit (Parus major) in relation to history and current selection. J Evol Biol 6:397-415 Blanckenhorn WU (2000) The evolution of body size: what keeps organisms small? Q Rev Biol 75:385-407 Blanckenhorn WU (2005) Behavioral causes and consequences of sexual size dimorphism. Ethology 111:977-1016 Blanckenhorn WU (2007) Case studies of the differential-equilibrium hypothesis of sexual size dimorphism in dung flies. In: Fairbairn DJ, Blanckenhorn WU, Székely T (eds) Sex, Size and Gender Roles: Evolutionary Studies of Sexual Size Dimorphism. Oxford University Press, Oxford, pp. 106-114

The quantitative study of sexual and natural selection

325

Blanckenhorn WU, Preziosi RF, Fairbairn DJ (1995) Time and energy constraints and the evolution of sexual size dimorphism: to eat or to mate? Evol Ecol 9:369-381 Blanckenhorn WU, Morf C, Mühlhäuser C, Reusch T (1999a) Spatiotemporal variation in selection on body size in the dung fly Sepsis cynipsea. J Evol Biol 12:563-576 Blanckenhorn WU, Reuter M, Ward PI, Barbour AD (1999b) Correcting for sampling bias in quantitative measures of selection when fitness is discrete. Evolution 53:286-291 Blanckenhorn WU, Mühlhäuser C, Morf C, Reusch T, Reuter M (2000) Female choice, female reluctance to mate and sexual selection on body size in the dungfly Sepsis cynipsea. Ethology 106:577-593 Blanckenhorn WU, Kraushaar URS, Reim C (2003) Sexual selection on morphological and physiological traits and fluctuating asymmetry in the yellow dung fly. J Evol Biol 16:903-913 Blanckenhorn WU, Kraushaar URS, Teuschl Y, Reim C (2004) Sexual selection on morphological and physiological traits and fluctuating asymmetry in the black scavenger fly Sepsis cynipsea. J Evol Biol 17:629-641 Blanckenhorn WU, Birrer M, Meier CM, Reim C, Teuschl Y, Weibel D (2008) Size-dependent mating success at various nutritional states in the yellow dung fly. Ethology 114:752-759 Blows MW (2007) A tale of two matrices: multivariate approaches in evolutionary biology. J Evol Biol 20:1-8 Bouteiller-Reuter C, Perrin N (2005) Sex-specific selective pressures on body mass in the greater white-toothed shrew, Crocidura russula. J Evol Biol 18:290-300 Brodie ED III, Janzen FJ (1996) On the assignment of fitness values in statistical analyses of selection. Evolution 50:437-442 Brodie ED III, Moore AJ, Janzen FJ (1995) Visualizing and quantifying natural selection. Trends Ecol Evol 10:313-318 Bumpus HC (1899) The elimination of the unfit as illustrated by the introduced sparrow, Passer domesticus. Biol Lect Woods Hole Marine Biol Lab 6:209226 Clutton-Brock TH (1988) Reproductive Success: Studies of Individual Variation in Contrasting Breeding Systems. University of Chicago Press, Chicago Cox RM, Calsbeek R (2009) Sexually antagonistic selection, sexual dimorphism and the resolution of intralocus sexual conflict. Am Nat 173:176-187 Crow JF (1958) Some possibilities for measuring selection intensities in man. Hum Biol 30:1-13 Darwin C (1859) The Origin of Species by Means of Natural Selection. John Murray, London Darwin C (1871) The Descent of Man and Selection in Relation to Sex. John Murray, London Endler JA (1986) Natural Selection in the Wild. Princeton University Press, Princeton

326

Wolf Blanckenhorn

Fairbairn DJ, Preziosi RF (1994) Sexual selection and the evolution of allometry for sexual size dimorphism in the water strider, Aquarius remigis. Am Nat 144:101-118 Falconer DS (1989) Introduction to Quantitative Genetics. Longman, Harlow Fedorka KM, Mousseau TA (2002) Nuptial gifts and the evolution of male body size. Evolution 56:590-596 Freeman S, Herron JC (2007) Evolutionary Analysis. Prentice Hall, Upper Saddle River/NJ Grant BR (1985) Selection on bill characters in a population of Darwin’s finches: Geospiza conirostris on Isla Genovesa, Galapagos. Evolution 39:523-532 Grant PR, Grant BR (1995) Predicting microevolutionary responses to directional selection on heritable variation. Evolution 49:241-251 Harvey AW (1990) Sexual differences in contemporary selection acting on size in the hermit crab Clibanarius digueti. Am Nat 136:292-304 Hunt J, Breuker CJ, Sadowski JA, Moore AJ (2009) Male-male competition, female mate choice and their interaction: determining total sexual selection. J Evol Biol 22:13-26 Istock CA (1981) Natural selection and life history variation: theory plus lessons from a mosquito. In: Denno RF, Dingle H (eds) Insect Life History Patterns: Habitat and Geographic Variation. Springer, Heidelberg, pp 113-126 Jann P, Blanckenhorn WU, Ward PI (2000) Temporal and microspatial variation in the intensities of natural and sexual selection in the yellow dung fly Scathophaga stercoraria. J Evol Biol 13:927-938 Janzen FJ, Stern HS (1998) Logistic regression for empirical studies of multivariate selection. Evolution 52:1564-1571 Kingsolver JG, Hoekstra HE, Hoekstra JM, Berrigan D, Vignieri SN, Hill CE, Hoang A, Gibert P, Beerli P (2001) The strength of phenotypic selection in natural populations. Am Nat 157:245-261 Koenig WD, Albano SS (1987) Lifetime reproductive success, selection and the opportunity for selection in the white-tailed skimmer, Plathemis lydia (Odonata: Libellulidae). Evolution 41:22-36 Lande R, Arnold SJ (1983) The measurement of selection on correlated characters. Evolution 37:1210-1226 Lynch M, Walsh B (1998) Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland/MA Manly BFJ (1985) The Statistics of Natural Selection on Animal Populations. Chapman and Hall, London Martin OY, Hosken DJ (2003a) The evolution of reproductive isolation through sexual conflict. Nature 423:979-982 Martin OY, Hosken DJ (2003b) Costs and benefits of evolving under experimentally enforced polyandry or monogamy. Evolution 57:2765-2772 Milinski M, Bakker TCM (1990) Female sticklebacks use male coloration in mate choice and hence avoid parasitized males. Nature 344:330-333 Mitchell-Olds T, Shaw RG (1987) Regression analysis of natural selection: statistical inference and biological interpretation. Evolution 41:1149-1161

The quantitative study of sexual and natural selection

327

Mousseau TA, Roff DA (1987) Natural selection and the heritability of fitness components. Heredity 59:181-197 Mühlhäuser C, Blanckenhorn WU (2002) The costs of avoiding matings in the dung fly Sepsis cynipsea. Behav Ecol 13:359-365 Olsson M, Shine R, Wapstra E, Ujvari B, Madsen T (2002) Sexual dimorphism in lizard body shape: the roles of sexual selection and fecundity selection. Evolution 56:1538-1542 Palmer AR (2000) Quasireplication and the contract of error: lessons from sex ratios, heritabilities and fluctuating asymmetry. Annu Rev Ecol Syst 31:441-480 Phillips PC, Arnold SJ (1989) Visualizing multivariate selection. Evolution 43:1209-1222 Preziosi RF, Fairbairn DJ (2000) Lifetime selection on adult body size and components of body size in a waterstrider: opposing selection and maintenance of sexual size dimorphism. Evolution 54:558-566 Price TD (1984) Sexual selection on body size, territory and plumage variables in a population of Darwin’s Finches. Evolution 38:327-341 Savalli UM, Fox CW (1998) Sexual selection and the fitness consequences of male body size in the seed beetle Stator limbatus. Anim Behav 55:473-483 Savalli UM, Fox CW (1999) The effect of male size, age, and mating behavior on sexual selection in the seed beetle Callosobruchus maculatus. Ethol Ecol Evol 11:49-60 Shuster SM, Wade MJ (2003) Mating Systems and Strategies. Princeton University Press, Princeton Teuschl Y, Reim C, Blanckenhorn WU (2010) No size-dependent reproductive costs in black scavenger fly (Sepsis cynipsea) males. Behav Ecol 21:85-90 Ward PI (1988) Sexual selection, natural selection, and body size in Gammarus pulex (Amphipoda). Am Nat 131:348-359

Chapter 12

Mate choice and reproductive conflict in simultaneous hermaphrodites NILS ANTHES

ABSTRACT Simultaneous hermaphroditism defines sexual systems in which animals display male and female sex at the same time. Given that each individual now inherently expresses the ‘interests’ of both sexes, this form of gender expression can have profound consequences for the evolution of behavioural reproductive strategies. The outcome of reproductive interactions between two hermaphrodites is likely to vary with (i) intra-individual trade-offs between male and female reproduction, (ii) inter-individual courtship and mate choice between the sexes, (iii) inter-individual competition within the sexes, and (iv) competition with rival individuals in the social group. While earlier work has argued that precopulatory mate choice may only be weakly expressed in simultaneous hermaphrodites, the first section of the chapter documents that, instead, mate choice is prevalent and may include sophisticated mate discrimination based on traits such as body size, mating history, or relatedness. The second section illustrates putative conflicts that individuals may face during copulation, primarily focussing on the decision over mating roles that is central to understand hermaphrodite reproductive behaviour. Available evidence indicates that conditional reciprocity, where individuals accept matings in their disfavoured mating role in order to achieve access to their favoured role, are less widespread than initially though. The final section explores the idea that simultaneous hermaphroditism should enhance selection on postcopulatory mate choice and sperm discrimination mechanisms, including harmful male mating strategies. Support for this prediction remains limited to date, calling for much broader empirical quantifications of the fitness costs and benefits associated with hermaphrodite mating strategies.

330

Nils Anthes

12.1 Introduction Although reproduction represents a generally cooperative act between males and females, there are numerous components of sexual interactions over which the interests of two partners may actually diverge. Sexual conflict is known to occur over traits such as mating frequency, remating with novel mates, the number, size or sex ratio of offspring, and parental care (Arnqvist and Rowe 2005, see also Kempenaers and Schlicht this volume, Schneider and Fromhage this volume). The divergence in evolutionary trait optima underlying such conflicts traces back to anisogamy as the defining difference between the sexes, and may vary with traits such as gamete replenishment rates or the costs and benefits associated with mate search and parental care (Kokko and Jennions 2008). In a very simplistic view, it may appear that sexual conflict occurs because sexual partners (= males and females) inherently pursue different interests. Whilst this logic seems to fit when looking at prime model systems for behavioural

a

b

Courtship & competition

Competition

... & cooperation

Sexual conflict ...

& cooperation

n tio eti mp Co

Sexual conflict

p

Competition

Courtship & competition

Fig. 12.1 Reproductive interactions in separate sex animals (a) and hermaphrodites (b). Shared reproduction between males and females (a) can show components of both sexual conflict and sexual cooperation. The same applies to hermaphrodites (b), but there is additional scope for intrasexual competition between the two male functions (and the two female functions) of both mates. These interactions may further be affected by the intra-individual trade-offs in resource allocation.

Mate choice and reproductive conflict in simultaneous hermaphrodites

331

ecology in mammals, birds or insects, the analogy takes an interesting twist when applied to simultaneous hermaphrodites, where individuals express male and female sex at the same time (BOX 12.1; referred to as ‘hermaphrodites’ throughout). In these animals, mating partners do not differ in their sex (both are hermaphrodites), meaning that mates are more equal to each other than is the case in species with separate sexes. One may argue that this similarity between mates favours mutualistic reproduction, as perhaps indicated by the primarily bi-directional (reciprocal) sperm transfer exhibited by most hermaphrodites (Michiels 1998). Yet, a closer look at hermaphrodite mating reveals that individuals are involved in a rather complicated set of intraand inter-sexual interactions (Fig. 12.1).

Table 12.1 Intrinsic differences between separate sex animals and simultaneous hermaphrodites. This overview focuses on mechanisms that affect reproductive behaviour (modified from Michiels 1998). Mechanism / trait

Separate sex animals

Simultaneous hermaphrodites

Self fertilisation

Inexistent

Facultative, primarily at low densities

Sex-specific gene expression

Prevalent (e.g. via sex chromosomes)

Impossible, restricting sexually dimorphic trait optimisation and ‘masking’ of sexually antagonistic alleles.

Sexual roles during mating

Pre-defined

Preferences for one particular mating role can generate conflict over mating roles.

Paternal care / nuptial gifts

Widespread strategies that increase paternal reproductive success

Largely absent, presumably because investment in own female reproduction pays off more than investment in paternal offspring for which there is no fertilisation guarantee.

Sex allocation

Rare deviations from 1:1 sex ratio (e.g. under local mate competition); Sex ratio adjustments take effect only in next generation

Resource allocation to male and female reproduction can be adjusted ad hoc to environmental variation.

Male mate choice

Present

Perhaps more prevalent: ‘female function’ offers a profitable alternative investment strategy, making ‘sperm dumping’ maladaptive.

332

Nils Anthes

BOX 12.1 Hermaphroditism and its evolution Hermaphroditism defines the expression of male and female sex (i.e., sperm and egg production) in a single individual. Depending on whether both sexes are functional successively or synchronously, we distinguish sequential and simultaneous hermaphrodites, respectively. Sequential hermaphroditism is favoured in systems with substantial sex-differences in the size-fecundity relationship (Ghiselin 1969, Warner 1975). A classic example are haremic fish, where large males can monopolize multiple small females, favouring size-dependent sex change from female to male (Munday et al. 2006). Simultaneous hermaphroditism, being the focus of this chapter, occurs in all major animal clades except insects and vertebrates other than fish (Fig. 12.2; Ghiselin 1969, Michiels 1998, Jarne and Auld 2006). Phylogenetic analyses indicate that simultaneous hermaphroditism represents the ancestral state among animals (Iyer and Roughgarden 2008, but see Ghiselin 1969), followed by repeated transitions between simultaneous hermaphroditism and gonochorism (Eppley and Jesson 2008). Prevailing explanations for the adaptive significance of simultaneous hermaphroditism date back to Tomlinson (1966), Ghiselin (1969), and Charnov et al. (1976), who identified ‘low density’ as a prime condition favouring simultaneous hermaphroditism. Low densities can arise for a variety of reasons, including small or highly structured populations (as in internal parasites), restricted mobility, poor mate-searching capacity, or otherwise rare mate encounters (Puurtinen and Kaitala 2002). In all these cases, individuals benefit if every encounter with a conspecific individual allows outcrossing. Self-fertilisation remains an ‘emergency option’ if no mate is acquired within a reasonable period of time (Charlesworth and Charlesworth 1987, Jarne and Auld 2006), leading to so-called ‘delayed selfing’ (Tsitrone et al. 2003). Under a paradigm of adaptive evolution, one would expect animals to express the reproductive mode that maximises fitness under the current environmental and social conditions. While this type of adaptive flexibility is approached in broadcasting invertebrates with their simple reproductive structures, internal fertilisation and complex reproductive morphologies as present, for example, in gastropods seem to restrict flexibility. Such constraints may explain why whole clades maintain an apparently ‘maladaptive’ reproductive mode, as is the case with simultaneous hermaphroditism under high population densities (Mank et al. 2006, Michiels et al. 2009).

Mate choice and reproductive conflict in simultaneous hermaphrodites

333

Hermaphroditism predominant minority occasional / absent

Fig. 12.2 Schematic view of the distribution of simultaneous hermaphroditism across major animal clades. The tree represents a reduced version of the composite eukaryote phylogenetic tree in Eppley and Jesson (2008). Data modified from Michiels (1998).

334

Nils Anthes

BOX 12.2 Hermaphrodite ‘model systems’ Current knowledge about the evolutionary ecology of hermaphrodite reproductive behaviour has accumulated from a diverse range of organisms. Depending on the research context, a number of animal groups have proven particularly useful study systems, and I briefly allude to those receiving recent emphasis below (Fig. 12.3). Sex allocation theory for simultaneous hermaphrodites tackles the ultimate and proximate factors that influence the proportional division of reproductive resources between male and female function (Charnov 1982). It has stimulated observational and experimental work in a long list of organisms (reviewed in Schärer 2009). The perhaps most profound advances in recent years are due to flourishing research in a small intertidal flatworm, Macrostomum lignano. This species allows non-invasive quantification of sex allocation and provides access to a suite of molecular, developmental and histological techniques (Ladurner et al. 2005). Ongoing research investigates the genetic basis for sex allocation, its connection with mating behaviour (Janicke and Schärer 2009), and its evolutionary trajectories in response to social and environmental variation. The idea that simultaneous hermaphrodites often face conflicts over mating roles (Charnov 1979) has triggered extensive research on ‘egg trading’ in external fertilisers such as serranid fish Serranus and polychaete worms Ophryotrocha (recently Sella and Ramella 1999, Sella and Lorenzi 2000, Petersen 2006, Crowley and Hart 2007) and on ‘sperm trading’ in internal fertilisers such as flatworms, earthworms, and gastropods (recently Anthes et al. 2005, Jordaens et al. 2005, Koene and Ter Maat 2005, Facon et al. 2008). Studies on these and many other species have also tackled questions regarding mate choice criteria, costs and benefits of polyandry, pre- and postcopulatory sexual selection, and physiological partner manipulation (reviews: Leonard 1991, 2006; Baur 1998, Michiels 1998, Jordaens et al. 2007). The option for self-fertilisation is another characteristic specific to simultaneous hermaphrodites. The evolutionary significance of selfing, in addition to factors determining the position of species along the selfing-outcrossing continuum, have been studied primarily in facultatively selfing freshwater molluscs, in particular Biomphalaria glabrata, Physa acuta and Lymnaea spp. (e.g. Escobar et al. 2007, Tian-Bi et al. 2008). The future development of a few of the above-mentioned species into model systems for the evolutionary ecology of hermaphrodite reproduction will provide insights into the underlying mechanisms. At the same time, the field has always drawn from novel insights gained by looking at a great diversity of study systems. To this end, our knowledge is still incomplete for a range of taxonomic groups, calling for further work on currently underrepresented taxa with often intriguing reproductive strategies such as leeches, gastrotrichs, or arrow worms.

Mate choice and reproductive conflict in simultaneous hermaphrodites a)

b)

c)

d)

f)

335

e)

g)

Fig. 12.3 A gallery of classic hermaphroditic study systems: the ‘sex allocation model’ Macrostomum lignano (a, photo © Lukas Schärer), body-piercing earthworms Lumbricus terrestris (b, photo © Nico K. Michiels), the sperm trading sea slug Chelidonura hirundinina (c, photo © Nils Anthes), the penis-fencing flatworm Pseudoceros bifurcus (d, photo © Nico K. Michiels), the dart-shooting garden snail Cornu aspersum (Helix aspersa) (e, photo © Joris M. Koene and Ronald Chase), and two egg traders, the polychaet worm Ophryotrocha diadema (f, photo © Gabriella Sella), and the reef fish Serranus tabacarius (g, photo © Mary K. Hart).

The interplay between individual sex allocation, intra-sexual competition and reciprocal mate choice may in fact generate substantial sexual conflict and favour the evolution of harmful mating strategies even more than is known from separate sex organisms.

336

Nils Anthes

The effect of simultaneous hermaphroditism on the evolution and mechanisms of reproductive behaviour is thus anything but straightforward. In this chapter, I explore how hermaphroditism may lead to sexually antagonistic interactions, and how these can be solved or avoided (Table 12.1). The chapter develops along three basic elements of reproductive behaviour: pre-copulatory mate assessment and choice, sexual interactions during mating, and post-copulatory selection and manipulation. Throughout, I highlight the diversity in hermaphrodite reproductive strategies by referring to findings in a diverse array of study systems (BOX 12.2).

12.2 Pre-copulatory mate assessment and choice Charles Darwin argued that simultaneous hermaphrodites would not exhibit mate assessment or mate choice, primarily because sexual dimorphism cannot be expressed and because ‘lower’ invertebrates would lack the ‘mental powers’ to engage in ‘mutual rivalry’ (Darwin 1871:321). Indeed, frequent mating as seen in many hermaphroditic animals may convey the impression of indiscriminate sexual interactions. For example, Aplysia sea slugs tend to form large mating aggregations where multiple individuals form mating chains in which each individual copulates in both sexual roles with two different partners (Pennings 1991). Yet, frequent mating alone does not prove the absence of mate choice, which can involve very subtle cues to generate non-random mating. The sections below therefore explore the capacity of hermaphrodites to display precopulatory courtship as well as mate choice from both female and male perspectives.

12.2.1 Size-dependent mate choice When the costs of insemination are non-trivial and fecundity increases with size, animals are expected to preferentially inseminate partners of equal or larger body size because small partners will produce fewer (if any) eggs. Substantial variation in adult size, for example in systems with indeterminate growth (e.g. molluscs or flatworms), is thus considered to favour size-dependent mate choice (DeWitt 1996). The size-fecundity relationship, and thus the benefits of size-dependent mate choice, further increase when larger animals invest proportionally more resources into eggs (rather than sperm), as predicted by sex allocation theory (Klinkhamer et al. 1997, Angeloni et al. 2002, Cadet et al. 2004) and confirmed empiri-

Mate choice and reproductive conflict in simultaneous hermaphrodites

337

cally (Vizoso and Schärer 2007). Size-dependent mate choice may be manifested via three pathways, and I briefly review supporting and conflicting evidence below. First, size-dependent mate choice may affect the within-pair decision over mating roles. An agreement over mating roles is primarily necessary in species with unilateral matings, where one individual acts as male and the other as female only. Theory then predicts the larger partner to preferentially take the female role. Mating patterns that conform to this prediction have been found in sea hares Aplysia spp. (Otsuka et al. 1980, Angeloni and Bradbury 1999), freshwater snails Physa spp. (DeWitt 1996, Ohbayashi-Hodoki et al. 2004, but see Dillen et al. 2008) and Helisoma trivolis (Norton et al. 2008), the land snail Achatina fulica (Tomiyama 1996), and the sea slug Chelidonura sandrana (Sprenger et al. 2009). Yet, partner discrimination by body size is all but universal (e.g. SwitzerDunlap et al. 1984, Peters and Michiels 1996a, Koene et al. 2007), and may be absent in particular in species where substantial investment in mate searching makes mate rejection too costly to be advantageous (Baur 1992). Second, in species with bi-directional (= reciprocal) mating, a preference to inseminate larger individuals ultimately results in pair formation between similar-sized individuals. This generates a positive correlation between the body sizes of mating individuals across the population, a pattern known as size-assortative mating. Assortative mating is known from separate-sex animals (Crespi 1989, Rowe and Arnqvist 1996) but should be even more widespread in hermaphroditic populations (Ridley 1983): Here, any preference for larger partners exhibited via one sex function automatically acts reciprocally between partners, whereas both males and females must show a preference for larger partners for size-assortative mating to spread in separate-sex species. Contrary to this prediction, evidence for positive size-assortment in hermaphrodites remains ambiguous: Some studies found positive relationships between partner body sizes in field and laboratory populations of earthworms, limpets, and molluscs (Crozier 1917, Tomiyama 1996, Yusa 1996, Angeloni and Bradbury 1999, Michiels et al. 2001, Angeloni 2003, Angeloni et al. 2003, Monroy et al. 2005, Pal et al. 2006), but most of these correlations were rather weak. Several other studies failed to detect any mating patterns that are consistent with sizeassortative mating (reviewed in Chaine and Angeloni 2005, Koene et al. 2007). Hence, the prevalence and biological relevance of size-assortative mating in hermaphrodites appears lower than expected. Finally, the assessment of a partner’s body size may not affect pair formation per se, but rather lead to prudent male mating effort, where individuals donate larger ejaculates to more fecund partners (Wedell et al. 2002). This effect has been nicely demonstrated in the earthworm Eisenia

338

Nils Anthes

Copulation (mutual penis insertion)

‘Partner screening’

Positioning of gonopores ‘Sandwich posture’

Fig. 12.4 Mating pattern in the flatworm Dugesia gonocephala. Pre-copulatory partner screening and the flattened sandwich posture serve mutual size-assessment. Copulation only follows when partners closely match in size (modified from Vreys and Michiels 1997 with permission from the Royal Society of London).

andrei, where individuals approximately double their ejaculate size when inseminating a partner with twice their body mass (Velando et al. 2008). Similarly, several opisthobranch sea slugs prolong insemination, and thus likely transfer more sperm, when mating with large partners (Angeloni 2003, Michiels et al. 2003, Anthes et al. 2006a). As with the above two mechanisms, prudent male mating should only evolve if this discriminatory behaviour really pays off. For example, if mating rates are low and competition among rival sperm is weak, strategic sperm allocation may not be favoured; this could explain its absence in snails such as Arianta arbustorum and Succinea putris (Baur et al. 1998, Baminger et al. 2000, Jordaens et al. 2005, Dillen et al. 2008). The proximate mechanisms employed by hermaphrodites to assess partner body size have to date received little attention. However, it is known that the parasitic cestode Schistocephalus solidus actively searches large partners using tactile and/or chemical cues, a behaviour that should ultimately also result in size-assortative mating (Lüscher and Wedekind 2002). An intriguing tactile mechanism for size assessment and mate choice has been proposed for the flatworm Dugesia gonocephala (Fig. 12.4): during pre-copulatory contact, prospective mates completely stretch

Mate choice and reproductive conflict in simultaneous hermaphrodites

339

out their body while laying on top of each other (Vreys and Michiels 1997). Even though the receptive pathway remains unknown, this behaviour apparently serves to estimate the partner’s body surface area relative to own body size. Flattening is more likely followed by copulation when partners are similar in body size, whereas dissimilar pairings are interrupted more frequently. 12.2.2 Mate choice for reduced sperm competition Many hermaphrodites are capable of storing allosperm for prolonged periods (Baur 1998), such that sperm from multiple donors may compete for access to unfertilised eggs. The mating history of a putative partner can therefore become a relevant mate choice criterion, in particular from a ‘male’ perspective. A very illustrative example comes from the opisthobranch, Aeolidiella glauca, which deposits spermatophores on the partner’s skin and thus makes recent mating externally visible (Haase and Karlsson 2004). For several days, the presence of a spermatophore reliably signals to putative novel partners that their sperm will be exposed to competition with sperm stored from the previous copulation. In accordance with the expectation that individuals preferentially inseminate previously unmated partners, spermatophore-carrying slugs are rejected more frequently during pre-copulatory interactions than individuals that had the spermatophore experimentally removed (Haase and Karlsson 2004). Most hermaphrodites do not externally deposit spermatophores, but inseminate their partner. As a result, their mating history is not externally visible. Even among such systems, some studies have found individuals to more eagerly copulate with previously unmated partners (Michiels and Bakovski 2000, Anthes et al. 2006a). This indicates the existence of subtle cues that reveal a partner’s mating history, possibly involving tactile receptors in the male copulatory organ that detect the presence of a rival’s ejaculate in the female genital tract. Another option involves chemical cues that could signal previous contact with other mates (Koene and Ter Maat 2007). Clearly more research is needed to reveal the underlying mechanisms. Rather than completely rejecting a previously mated partner, animals may instead strategically adjust their male mating effort to the perceived intensity of sperm competition. Sperm competition models predict that ejaculate sizes should be small when inseminating a virgin partner, reach a maximum when inseminating a partner where the own ejaculate will compete with sperm from a single previous donor, and successively decrease thereafter (Wedell et al. 2002). Concordant with these models, earth-

340

Nils Anthes

worms, Eisenia fetida, approximately double their ejaculate size when inseminating a once mated partner compared to a virgin (Velando et al. 2008). This could be beneficial because mate encounters are frequent in this species and male resources therefore limited. No such adjustment of male mating effort occurs in the landsnail, Arianta arbustorum, where individuals donate sperm independent of partner mating history (Baur et al. 1998) or the actual risk of sperm competition (Locher and Baur 2000). A. arbustorum typically occurs at rather low population densities and only copulates a few times per season, providing a possible explanation for why male choosiness has apparently not evolved. An individual’s mating behaviour may vary also with own mating history. In the freshwater snail, Physa acuta, individuals that obtain a novel mating opportunity after sexual isolation preferentially assume the female mating role (Facon et al. 2007). This is consistent with theory that predicts a preference for the female mating role after periods of sexual isolation because of allosperm depletion. Receiving an ejaculate therefore ensures fertility and maximises immediate fitness benefits (Anthes et al. 2006b). Such ultimate factors may, however, be offset by proximate factors that more directly affect mating propensity. In two pulmonate snails, Lymnaea stagnalis and Succinea putris, it is the male (rather than female) motivation that increases after sexual isolation. In these species, sexual motivation is linked to the filling status of the seminal fluid producing prostate gland (Koene and Ter Maat 2005, Dillen et al. 2008). With the prostate gland depleted after mating, male mating drive remains low for several days, whereas copulations in the female role are still readily accepted despite full allosperm stores. The male motivational effect in these species may have an intriguing consequence for population operational sex ratios: Assuming that male copulations are typically followed by an unreceptive period of several days, whereas female mating motivation is maintained after sperm receipt, populations should on average contain more individuals that are receptive in their female function. The resultant female-biased operational sex ratio could limit access to ‘males’ for receptive ‘females’, possibly leading to sex role reversal as to date only known from separate-sex organisms (e.g. Berglund and Rosenqvist 2003). 12.2.3 Mate choice for relatedness and immune function Inbreeding often confers deleterious effects to offspring through the expression of recessive diseases or reduced genetic diversity at immunerelevant loci (such as MHC). Mate choice for the genetic quality of a partner should therefore be highly beneficial. A corresponding picture emerges

Mate choice and reproductive conflict in simultaneous hermaphrodites

341

from studies in freshwater snails Physa. Individuals from French P. acuta populations increasingly reject mating opportunities along the gradient from cross population, through within population to sib mating (Facon et al. 2006). American P. gyrina discriminate most strongly against outbreeding across populations (McCarthy 2004). Together, these studies indicate a preference for ‘intermediate outbreeding’ (McCarthy and Sih 2008), which provides a sensitive balance between the maintenance of genetic diversity and the disruption of beneficial genotypes. In both species, choosy behaviour regarding relatedness was only observed for matings in the female (but not the male) mating role. The female role is also the one that primarily suffers from the rather substantial inbreeding depression in this species (Jarne et al. 2000). Mate discrimination thus seems to associate with the role that pays the highest cost of ‘wrong choices’. The absence of inbreeding avoidance in other species could have one of several alternative reasons. First, the deleterious effects of sib-mating may be too weak to afford the costs of mate discrimination, and instead be offset by the inclusive fitness benefit that is accrued from incestuous mating (Kokko and Ots 2006). Data consistent with this idea have been reported from the landsnail Arianta arbustorum (Baur and Baur 1997). Second, deleterious alleles may have already been successfully purged, removing the genetic load of inbreeding. Inbreeding may then yield even higher fitness than outbreeding, leading to a preference for sib-matings as documented in the intestinal parasitic cestode Schistocephalus solidus (Schjørring and Jäger 2007). Finally, even in cases where inbreeding is costly, life-history strategies other than mate choice may sufficiently avoid such events. A concordant example comes from the freshwater flatworm Schmidtea polychroa, where efficient dispersal makes encounters between siblings so unlikely that mechanisms for kin recognition and selective mating seem to be obsolete (Peters and Michiels 1996b). In simultaneous hermaphrodites, mate choice for relatedness extends to the option of self-fertilisation as the far end of the ‘relatedness-continuum’ (Jarne and Charlesworth 1993; Jarne and Auld 2006). Selfing offers some intriguing options for the sperm recipient. For example, following mating with a ‘low quality’ partner (say, with an incompatible genetic make-up), the fitness costs of self-fertilisation may be lower than those of outcrossing. Even though not yet tested explicitly, work on the tapeworm Schistocephalus solidus indeed indicates that mate identity affects the degree of selfing, where selfing rates are higher after pairings that involved more conflict between partners (Lüscher and Milinski 2003). Selfing may thus not only be an emergency option when facing allosperm depletion, but can also be considered as a component of an individual’s reproductive strategy whenever inbreeding depression is not too severe. Such ‘strategic selfing’

342

Nils Anthes

would obviously not be in the interest of a previous sperm donor, generating a conflict load over selfing between partners (Bedhomme et al. 2009). Mate discrimination based on genetic traits similarly applies to specific heritable traits such as parasite resistance. Again, mate discrimination primarily relates to the sex function that takes the costs of wrong decisions. For example, genotypes of the snail Biomphalaria glabrata resistant or susceptible to infection with the castrating parasite Schistosoma mansoni mate indiscriminately in both mating roles with uninfected partners, but resistant genotypes refuse matings in the female role with infected partners (Webster et al. 2003, Webster and Gower 2006). Parasites that only weakly depress reproductive success or performance seem to have only limited impact on mating behaviour, as documented for Monocystisinfections of the earthworm Lumbricus terrestris (Field et al. 2003, Field and Michiels 2005). As with body size recognition, we currently have only vague ideas about the mechanisms involved in the detection of relatedness or parasite susceptibility. In the cestode Schistocephalus solidus, mate choice experiments that prevented physical contact between partners implicated chemical cues as informative agents indicating relatedness (Schjørring 2009). Similarly, chemical compounds in the mucous or shell have been suggested to be involved in the detection of resistance status in Biomphalaria glabrata (Webster et al. 2003). Facon et al. (2006) found no evidence that relatedness affected pre-mating behaviour and mucous-trail following in Physa acuta. This indicates that recognition is not associated with chemical cues in the mucous but may rather occur during penis intromission or via other contact signals.

12.3 Conflicts over mating roles and conditional reciprocity While the sexual roles during mating are inherently determined in copulations between males and females, there are three possible combinations for simultaneous hermaphrodites: act in both sex roles, act male only (with the partner acting female), and vice versa. Conflict over mating roles defines situations where two prospective mates disagree over the choice of mating roles for subsequent copulation (Wethington and Dillon 1996, Anthes et al. 2006b). It occurs whenever two individuals prefer incompatible mating roles, for example if both intend to mate in the female but not in the male role (Michiels 1998, Fig. 12.5). Several outcomes of this conflict of interest are conceivable. First, one individual accepts (or is forced into) the less

Insist on ♀ role

Unilateral No conflict

Unilateral No conflict

Conflict over mating roles

Accept ♂ and ♀ role

Unilateral No conflict

Reciprocal No conflict

Unilateral No conflict

Insist on ♂ role

Hermaphrodite 2

Mate choice and reproductive conflict in simultaneous hermaphrodites

Conflict over mating roles

Unilateral No conflict

Unilateral No conflict

Insist on ♂ role

Accept ♂ and ♀ role

Insist on ♀ role

343

Hermaphrodite 1

Fig. 12.5 Scheme illustrating the occurrence of conflicts over mating roles between two prospective hermaphroditic mating partners.

preferred role, resulting in unilateral matings. Second, no mating takes place because no consent is achieved during pre-copulatory ‘negotiation’. Finally, the dilemma is solved if both partners accept mating in the less preferred role and this also guarantees access to their preferred role. Such ‘cooperative’ behaviour would result in balanced reciprocal mating. The last scenario, where individuals reciprocate in the less preferred role in order to obtain access to the preferred role, has been coined trading, or conditional reciprocity (Leonard and Lukowiak 1984). The currency used for trading can be either type of gamete. Egg trading has been experimentally assessed in externally fertilising polychaete worms of the genus Ophryotrocha (Sella 1985, Sella et al. 1997; Fig. 12.3) but to date received most attention in serranid sea basses (details in Fischer 1984, Petersen 1995, 2006; Fig. 12.3). In the final hours before dusk, these hermaphroditic fish engage in alternated bouts of spawning. A first individual releases a small parcel of eggs into the water to be fertilised by its male acting partner. Thereafter, roles are changed repeatedly until both individuals have released and fertilised a balanced number of egg parcels. This behaviour resembles a tit-for-tat game of strategy, where each individual only offers costly eggs when the partner follows suit. The alternative strategy would be very risky: An individual releasing all its eggs in a single go is prone to cheating. Its partner may just fertilise these eggs and leave without offering an own batch of eggs, rather seeking another ‘cheap’ male mating opportunity. Hence, repeated alternation of sex roles assures that

344

Nils Anthes

each partner takes a fair share in both male and female reproduction. Interestingly, egg trading strategies vary strikingly even between closely related species. Evolutionary stable strategy models indicate that this variation may be driven by differences in local conditions, such as mating group size, population density, and predation risk (Crowley and Hart 2007). In sperm trading, it is the male gametes that are exchanged conditionally between partners. Sperm trading is primarily expected in internal fertilisers (Leonard and Lukowiak 1984) where fertilisation is under female control. This makes the ‘fertilisation game’ much more subtle than in egg traders: Only sperm exchange is directly connected to mating and thus verifiable. Eggs, in contrast, are not transferred or offered, such that a male-acting individual has no honest information regarding fertilisation success. Longterm sperm storage, multiple mating, and the presence of sperm digestion organs in many hermaphrodites make male fertilisation success very unpredictable. Hence, even though hermaphrodites are expected to primarily mate in order to donate rather than to receive sperm (Charnov 1979, but see opposite prediction by Leonard 1991), sperm donors should remain prudent about their sperm allocation. By donating sperm only to partners that return this investment, individuals at least (i) ensure own fertility, (ii) obtain a potential ‘repayment’ for their risky ejaculate investment (via resources available for digestion), and (iii) acquire possibly informative signals about the partner’s quality or condition (Landolfa 2002, Anthes et al. 2005). Following the establishment of the concept of sperm trading based on observations in the sea slug Navanax inermis (Leonard and Lukowiak 1984), many follow-up studies found mating patterns that were consistent with this mechanism (reviewed in Anthes et al. 2006b). However, while many reciprocally mating hermaphrodites indeed appear like engaging in a balanced sexual trade, pure behavioural patterns may also occur as a simple by-product of the partners’ mutual willingness to donate and receive sperm and are thus not by themselves sufficient criteria to define conditional reciprocity (Anthes et al. 2006b). For example in the snail, Physa actua, non-random alternation of mating roles is the result of the individual strategy to alternate sex roles between successive matings. Reciprocation then often occurs with novel partners, and does not involve the conditional nature of sex role alternation between two mates that is required for trading (Facon et al. 2007, 2008). Likewise, in the freshwater snail, Lymnaea stagnalis, serial reciprocity with the same partner only occurs when snails had been sexually isolated and thus display elevated male mating motivation (Koene and Ter Maat 2005). This indicates that serial reciprocity is a product of sexual motivation, not of conditionality.

Mate choice and reproductive conflict in simultaneous hermaphrodites a)

345

b)

(ii)

(i)

(iii)

g

T-focal T-cheater

(iv)

C-focal C-partner

p 57 T-pairs

51 C-pairs

Fig. 12.6 Mating cycle (a) and experimental treatment (b) for testing sperm trading in the sea slug Chelidonura hirundinina. Initial body contact (i) is followed by pre-copulatory entwining (ii) and unilateral or reciprocal penis insertion (p) (iii). This cycle is multiply repeated until one partner deserts (iv). In the experiment (b), the sperm groove (violet) transporting ejaculate parcels from the gonopore (g) to the penis was cauterised in imposed ‘cheaters’. All other individuals received a sham-treatment (red dots). The study measured how focal slugs responded to not receiving sperm in the treatment group (T) relative to the control group (C). Reprinted from Anthes et al. (2005) with permission from Elsevier.

Given that sperm trading should ultimately serve to balance the amount of sperm exchanged between partners, between-partner correlations in ejaculate sizes have been considered a second indicator for this mating strategy. A corresponding pattern has been found in the freshwater planarian, Dugesia gonocephala, which closely balances sperm amounts during mutual penis intromission and preparation of the spermatophore (Vreys and Michiels 1998). Interestingly, in the closely related Schmidtea polychroa, sperm amounts are not balanced between partners (Michiels and Bakovski 2000), but instead vary with the size of selfsperm stores of the sperm donor (Michiels and Streng 1998), again indicating individually optimised strategies rather than a conditional ‘agreement’ between partners. Similarly, studies in several molluscs failed to find correlations in ejaculate sizes between partners (Baur et al. 1998, Jordaens et al. 2005), indicating that trading of sperm amounts may in fact be rather rare among simultaneous hermaphrodites. Conditionality as the central component of trading (Leonard and Lukowiak 1984) requires that attempts to cheat (by avoiding the less preferred

346

Nils Anthes

role) are punished by the cheated individual. Recent manipulative studies could confirm that sperm exchange can indeed be conditional in internal fertilisers. In the sea slug Chelidonura hirundinina (Fig. 12.3), multiple bouts of reciprocal and unilateral sperm transfer highly synchronize mating outcomes between partners (Putz et al. 2008) and are suggestive of sperm trading. In an experiment disabling sperm transfer in a subset of individuals to generate ‘cheaters’ in the male mating role (Fig. 12.6), sham-treated focal slugs interrupted matings significantly earlier and after fewer sex role alternations when mated to such cheaters than in matings with shamtreated control partners (Anthes et al. 2005). This shows that in this species sperm donation is indeed conditional upon sperm receipt, and matings are interrupted if the partner does not reciprocate. Interestingly, sperm trading is not universal among closely related species. A congener, Chelidonura sandrana, showed no indications for conditional sperm exchange when exposed to experimental cheaters (Anthes and Michiels 2005). Combined with the above overview on studies failing to provide support for trading, the currently available datasets suggest that trading occurs only in a limited number of systems, rather than being a prevalent mating strategy among simultaneous hermaphrodites as initially expected (Axelrod and Hamilton 1981, Leonard and Lukowiak 1984, Michiels 1998).

12.4 Male harm and post-copulatory partner manipulation 12.4.1 Male harm Male harm occurs when male-male competition gives rise to traits that provide a male fertilisation advantage but also impose collateral or intentional damage upon the female mating partner. Even though widespread among separate sex animals (Johnstone and Keller 2000, Arnqvist and Rowe 2005, Lessells 2006), recent models predict that male harm even more likely evolves in simultaneous hermaphrodites and takes more extreme forms (Michiels and Koene 2006, Preece et al. 2009). The underlying reason is that – according to these models – hermaphrodites can accept matings that are costly for the female function as long as this fitness loss is at least outweighed by the expected paternal success through mating in the male function. Clear evidence for male harm in hermaphrodites is still scarce, and many studies did not detect relevant fitness costs of mating behaviours that look rather damaging at first sight (e.g. dart-shooting: Chase and Vaga 2006 and further details below). Nevertheless, support for manipulative male mating strategies with damaging side-effects for the partner has re-

Mate choice and reproductive conflict in simultaneous hermaphrodites

347

cently begun to emerge. One example is traumatic sperm injection, which often inflicts substantial damage to the female function. This insemination strategy is rare among separate sex organisms (Siva-Jothy 2006), but widespread among hermaphroditic sea slugs, flatworms, or leeches. For example, the externally applied ejaculates of polyclad flatworms (Pseudobiceros bedfordi, Pseudoceros bifurcus, Fig. 12.3) contain aggressive substances that dissolve up to 50% of a receiver’s body tissue while sperm actively travel towards the fertilisation pouches (Michiels and Newman 1998, Arnqvist and Rowe 2005). Evidence for fitness costs of hypodermic injections comes from the sacoglossan sea slug, Alderia willowi. Here, individuals that receive sperm via hypodermic injection produce 30-50% fewer eggs than non-mated individuals that lay unfertilised egg masses (Smolensky et al. 2009). 12.4.2 Physiological partner manipulation A second category of partner manipulation concerns the transfer of manipulative allohormones (Koene 2005), referring to bioactive substances that interfere with postcopulatory processes such as fertilisation, remating, or resource allocation to female function (Michiels 1998, Bedhomme et al. 2009, Schärer and Janicke 2009). Manipulative ejaculate components such as the seminal sex peptides in Drosophila (Chapman et al. 1995, Fricke et al. 2009) are well documented for separate sex organisms. Simultaneous hermaphrodites may have particularly easy access to manipulations of the partner’s reproductive physiology: The genes for male and female regulatory substances are present and expressed in each individual. Hence, active compounds that affect the partner’s reproductive system are inherently available and could easily be applied during mating (Koene 2005). The broad role that such mechanisms may play in hermaphrodite mating systems is indicated by the application of gland secretions during mating in diverse systems, including opisthobranch sea slugs (Anthes and Michiels 2007), land slugs (Reise et al. 2007), land snails, and earthworms. The latter two groups provide the best studied examples to date: The garden snail Cornu aspersum is well-known for its dart-shooting behaviour (Fig. 12.3), where sexual partners initiate copulation by shooting a calcareous dart into the partner’s body. Establishing the dart’s function has attracted a series of experimental studies (Koene and Chase 1998a, Koene and Chase 1998b, Landolfa et al. 2001, Rogers and Chase 2001, Evanno et al. 2005). Only recently, Chase and Blanchard (2006) provided compelling evidence that it is indeed the bioactive compounds delivered with the dart’s mucous that increase a dart-shooters paternity. Interestingly, this effect cannot be gen-

348

Nils Anthes

eralised to all dart-shooting snails, with darts having no effect on sperm delivery in Arianta arbustorum (Baminger et al. 2000). Lumbricus terrestris earthworms follow a similar strategy: During reciprocal copulation and sperm transfer (Fig. 12.3), ~40 copulatory setae pierce into the partner’s body and inject secretions of the setal glands (Grove 1925, Koene et al. 2002). Experimental injections have shown that these products increase sperm uptake and delay remating, both of which are in the interest of the sperm donor but may conflict with the interest of the sperm receiver (Koene et al. 2005). Proteomic analysis of the secretions suggest a role for ubiquitin in the manipulative process (König et al. 2006), but its exact functionality remains to be established. 12.4.3 Costs and benefits of multiple mating Despite the potential costs of matings in the female role, many hermaphrodites do not seem particularly reluctant to accept female matings (Mulvey and Vrijenhoek 1981, Baur 1994, Angeloni et al. 2003, Pongratz and Michiels 2003). This implies that multiple mating not only imposes costs on the female-acting partner, but also confers some kind of direct or indirect benefit (Arnqvist and Nilsson 2000). Some pulmonate land snails such as Arianta arbustorum possess complex sperm storage organs with multiple tubules that could serve differential storage and use of sperm that were received from multiple donors (Bojat and Haase 2002). Such control over fertilisation would indeed benefit female interests, but whether or not controlled fertilisation occurs and is beneficial to the female function awaits experimental confirmation (Chen and Baur 1993, reviewed in Baur 2007). In the sea slug Chelidonura sandrana, egg production decreases steadily with female mating rate (Sprenger et al. 2008b). These mating costs are at least partly offset through polyandry: Multiply mated mothers lay larger eggs that hatch larger veliger larvae than monandrous mothers (Sprenger et al. 2008a). This effect was strongest when experimental slugs were exposed to mating rates matching those experienced in the field (Sprenger et al. 2008a), indicating that in this species it is the female function that maintains mating rates near its optimum. Interestingly, the benefits of polyandry in this species do not depend on male identity: Additive genetic sire (= ‘good gene’) effects were absent, indicating that female benefits do not accrue from the genetic quality of male partners (Sprenger et al. 2010). Instead, variation in offspring traits was largely governed by maternal effects, where mothers more or less actively elevate investment per offspring after multiple mating. These slugs therefore possibly follow a bet-hedging strategy, where the production of genetically diverse offspring could re-

Mate choice and reproductive conflict in simultaneous hermaphrodites

349

duce mortality under the unpredictable environmental conditions that larvae will encounter after settlement (Sprenger et al. 2010).

12.5 Perspectives Our current knowledge on the way sexual selection affects the evolution of hermaphrodite reproductive behaviour primarily suffers from a paucity of explicit quantifications of the sex-specific (lifetime) costs and benefits of the diverse mating strategies. It therefore often remains hard to evaluate whether a given behaviour is an expression of sexual mutualism or sexual antagonism. Closing this gap will be a major goal for the near future. Moreover, given that descriptions of hermaphrodite reproductive behaviour remain often based on anecdotal reports, more extensive experimental research on the mechanisms of mate assessment, mate choice, and cryptic post-copulatory mechanisms will substantially complement our picture of sexual selection in these groups. These goals should greatly benefit from ongoing attempts to establish selected hermaphroditic species as model systems for studying the genetics underlying the variation in reproductive strategies.

Ackowledgements I wish to thank Peter Kappeler, Hanna Kokko, Nico K. Michiels, Dennis Sprenger, and one anonymous referee for constructive comments to this chapter, Stuart Pengelley for linguistic corrections, and Mary K. Hart, Joris M. Koene, Nico K. Michiels, Lukas Schärer, and Gabriella Sella for providing the photos of their study pets displayed in Fig. 12.3.

References Angeloni L (2003) Sexual selection in a simultaneous hermaphrodite with hypodermic insemination: body size, allocation to sexual roles and paternity. Anim Behav 66:417-426 Angeloni L, Bradbury JW (1999) Body size influences mating strategies in a simultaneously hermaphroditic sea slug, Aplysia vaccaria. Ethol Ecol Evol 11:187-195 Angeloni L, Bradbury JW, Charnov EL (2002) Body size and sex allocation in simultaneously hermaphroditic animals. Behav Ecol 13:419-426

350

Nils Anthes

Angeloni L, Bradbury JW, Burton RS (2003) Multiple mating, paternity, and body size in a simultaneous hermaphrodite, Aplysia californica. Behav Ecol 14:554-560 Anthes N, Michiels NK (2005) Do ‘sperm trading’ simultaneous hermaphrodites always trade sperm? Behav Ecol 16:188-195 Anthes N, Michiels NK (2007) Precopulatory stabbing, hypodermic injections and unilateral matings in a hermaphroditic sea slug. Biol Lett 3:121-124 Anthes N, Putz A, Michiels NK (2005) Gender trading in a hermaphrodite. Curr Biol 15:R792-R793 Anthes N, Putz A, Michiels NK (2006a) Hermaphrodite sex role preferences: the role of partner body size, mating history and female fitness in the sea slug Chelidonura sandrana. Behav Ecol Sociobiol 60:359-367 Anthes N, Putz A, Michiels NK (2006b) Sex role preferences, gender conflict and sperm trading in simultaneous hermaphrodites: a new framework. Anim Behav 72:1-12 Arnqvist G, Nilsson T (2000) The evolution of polyandry: multiple mating and female fitness in insects. Anim Behav 60:145-164 Arnqvist G, Rowe L (2005) Sexual Conflict. Princeton University Press, Princeton Axelrod R, Hamilton WD (1981) The evolution of cooperation. Science 211:13901396 Baminger H, Locher R, Baur B (2000) Incidence of dart shooting, sperm delivery, and sperm storage in natural populations of the simultaneously hermaphroditic land snail Arianta arbustorum. Can J Zool 78:1767-1774 Baur B (1992) Random mating by size in the simultaneously hermaphroditic land snail Arianta arbustorum: experiments and an explanation. Anim Behav 43:511-518 Baur B (1994) Multiple paternity and individual variation in sperm precedence in the simultaneously hermaphroditic land snail Arianta arbustorum. Behav Ecol Sociobiol 35:413-421 Baur B (1998) Sperm competition in molluscs. In: Birkhead TR, Møller AP (eds) Sperm Competition and Sexual Selection. Academic Press, London, pp 255305 Baur B (2007) Reproductive biology and mating conflict in the simultaneously hermaphroditic land snail Arianta arbustorum. Am Malacol Bull 23:157-172 Baur B, Baur A (1997) Random mating with respect to relatedness in the simultaneously hermaphroditic land snail Arianta arbustorum. Invert Biol 116:294298 Baur B, Locher R, Baur A (1998) Sperm allocation in the simultaneously hermaphroditic land snail Arianta arbustorum. Anim Behav 56:839-845 Bedhomme S, Bernasconi G, Koene JM, Lankinen A, Arathi HS, Michiels NK, Anthes N (2009) How does breeding system variation modulate sexual antagonism? Biol Lett 5:717-720 Berglund A, Rosenqvist G (2003) Sex role reversal in pipefish. Adv Stud Behav 32:131-167 Bojat NC, Haase M (2002) Sperm storage in the simultaneously hermaphroditic land snail Arianta arbustorum. J Zool Lond 258:497-503

Mate choice and reproductive conflict in simultaneous hermaphrodites

351

Cadet C, Metz JAJ, Klinkhamer PGL (2004) Size and the not-so-single sex: disentangling the effects of size and budget on sex allocation in hermaphrodites. Am Nat 164:779-792 Chaine A, Angeloni L (2005) Size-dependent mating and gender choice in a simultaneous hermaphrodite, Bulla gouldiana. Behav Ecol Sociobiol 59:58-68 Chapman T, Liddle LF, Kalb JM, Wolfner MF, Partridge L (1995) Cost of mating in Drosophila melanogaster females is mediated by male accessory gland products. Nature 373:241-244 Charlesworth D, Charlesworth B (1987) Inbreeding depression and its evolutionary consequences. Annu Rev Ecol Syst 18:237-268 Charnov EL (1979) Simultaneous hermaphroditism and sexual selection. Proc Natl Acad Sci USA 76:2480-2484 Charnov EL (1982) The Theory of Sex Allocation. Princeton University Press, Princeton Charnov EL, Bull JJ, Maynard-Smith J (1976) Why be an hermaphrodite? Nature 263:125-126 Chase R, Blanchard KC (2006) The snail’s love-dart delivers mucus to increase paternity. Proc R Soc Lond B 273:1471-1475 Chase R, Vaga K (2006) Independence, not conflict, characterizes dart-shooting and sperm exchange in a hermaphroditic snail. Behav Ecol Sociobiol 59:732739 Chen X, Baur B (1993) The effect of multiple mating on female reproductive success in the simultaneously hermaphroditic land snail Arianta arbustorum. Can J Zool 71:2431-2436 Crespi BJ (1989) Causes of assortative mating in arthropods. Anim Behav 38:9801000 Crowley PH, Hart MK (2007) Evolutionary stability of egg trading and parceling in simultaneous hermaphrodites: the chalk bass revisited. J Theor Biol 246:420-429 Crozier WJ (1917) Evidence of assortive mating in a nudibranch. Proc Natl Acad Sci USA 3:519-522 Darwin C (1871) The Descent of Man and Selection in Relation to Sex. John Muray, London DeWitt TJ (1996) Gender contests in a simultaneous hermaphrodite snail: a sizeadvantage model for behaviour. Anim Behav 51:345-351 Dillen L, Jordaens K, Dieleman W, Backeljau T (2008) Effects of isolation and body size on the mating behaviour of the hermaphroditic land snail Succinea putris. Anim Behav 75:1401-1411 Eppley SM, Jesson LK (2008) Moving to mate: the evolution of separate and combined sexes in multicellular organisms. J Evol Biol 21:727-736 Escobar JS, Epinat G, Sarda V, David P (2007) No correlation between inbreeding depression and delayed selfing in the freshwater snail Physa acuta. Evolution 61:2655-2670 Evanno G, Madec L, Arnaud J-F (2005) Multiple paternity and postcopulatory sexual selection in a hermaphrodite: what influences sperm precedence in the garden snail Helix aspersa? Mol Ecol 14:805-812

352

Nils Anthes

Facon B, Ravigné V, Goudet J (2006) Experimental evidence of inbreeding avoidance in the hermaphroditic snail Physa acuta. Evol Ecol 20:395-406 Facon B, Ravigné V, Sauteur L, Goudet J (2007) Effect of mating history on gender preference in the hermaphroditic snail Physa acuta. Anim Behav 74:14551461 Facon B, Ravigné V, Goudet J (2008) Gender-role alternation in the simultaneously hermaphroditic freshwater snail Physa acuta: not with the same partner. Behav Ecol Sociobiol 62:713-720 Field SG, Michiels NK (2005) Parasitism and growth in the earthworm Lumbricus terrestris: fitness costs of the gregarine parasite Monocystis sp. Parasitology 130:397-403 Field SG, Schirp HJ, Michiels NK (2003) The influence of Monocystis sp infection on growth and mating behaviour of the earthworm Lumbricus terrestris. Can J Zool 81:1161-1167 Fischer EA (1984) Egg trading in the chalk bass, Serranus tortugarum, a simultaneous hermaphrodite. Z Tierpsychol 66:143-151 Fricke C, Wigby S, Hobbs R, Chapman T (2009) The benefits of male ejaculate sex peptide transfer in Drosophila melanogaster. J Evol Biol 22:275-286 Gascoigne T (1993) Stylets, styles and other cuticular extensions of the male duct in the order Sacoglossa (Gastropoda). Boll Malacol 29:139-146 Ghiselin MT (1969) The evolution of hermaphroditism among animals. Q Rev Biol 44:189-208 Grove AJ (1925) On the reproductive processes of the earthworm, Lumbricus terrestris. Q J Microsc Sci 69:245-290 Haase M, Karlsson A (2004) Mate choice in a hermaphrodite: you won’t score with a spermatophore. Anim Behav 67:287-291 Iyer P, Roughgarden J (2008) Dioecy as a specialization promoting sperm delivery. Evol Ecol Res 10:867-892 Janicke T, Schärer L (2009) Determinants of mating and sperm-transfer success in a simultaneous hermaphrodite. J Evol Biol 22:405-415 Jarne P, Auld JR (2006) Animals mix it up too: the distribution of self-fertilization among hermaphroditic animals. Evolution 60:1816-1824 Jarne P, Charlesworth D (1993) The evolution of the selfing rate in functionally hermaphrodite plants and animals. Ann Rev Ecol Syst 24:441-466 Jarne P, Perdieu M-A, Pernot A-F, Delay B, David P (2000) The influence of selffertilization and grouping on fitness attributes in the freshwater snail Physa acuta: population and individual inbreeding depression. J Evol Biol 13:645655 Johnstone RA, Keller L (2000) How males can gain by harming their mates: sexual conflict, seminal toxins, and the cost of mating. Am Nat 156:368-377 Jordaens K, Pinceel J, Backeljau T (2005) Mate choice in the hermaphroditic land snail Succinea putris (Stylommatophora: Succineidae). Anim Behav 70:329337 Jordaens K, Dillen L, Backeljau T (2007) Effects of mating, breeding system and parasites on reproduction in hermaphrodites: pulmonate gastropods (Mollusca). Anim Biol 57:137-195

Mate choice and reproductive conflict in simultaneous hermaphrodites

353

Klinkhamer PGL, de Jong TJ, Metz H (1997) Sex and size in cosexual plants. Trends Ecol Evol 12:260-265 Koene JM (2005) Allohormones and sensory traps: a fundamental difference between hermaphrodites and gonochorists? Invert Reprod Dev 48:101-107 Koene JM, Chase R (1998a) Changes in the reproductive system of the snail Helix aspersa caused by mucus from the love dart. J Exp Biol 201:2313-2319 Koene JM, Chase R (1998b) The love dart of Helix aspersa Müller is not a gift of calcium. J Mollus Stud 64:75-80 Koene JM, Ter Maat A (2005) Sex role alternation in the simultaneous hermaphroditic pond snail Lymnaea stagnalis is determined by the availability of seminal fluid. Anim Behav 69:845-850 Koene JM, Ter Maat A (2007) Coolidge effect in pond snails: male motivation in a simultaneous hermaphrodite. BMC Evol Biol 7:212 Koene JM, Sundermann G, Michiels NK (2002) On the function of body piercing during copulation in earthworms. Invert Reprod Dev 41:35-40 Koene JM, Pförtner T, Michiels NK (2005) Piercing the partner’s skin influences sperm uptake in the earthworm Lumbricus terrestris. Behav Ecol Sociobiol 59:243-249 Koene JM, Montagne-Wajer K, Ter Maat A (2007) Aspects of body size and mate choice in the simultaneously hermaphroditic pond snail Lymnaea stagnalis. Anim Biol 57:247-259 König S, Mehlich A-M, Bullesbach J, Michiels NK (2006) Allohormones in Lumbricus terrestris? Mass spectrometry of the setal gland product indicates possible role of ubiquitin. Invert Reprod Dev 49:103-112 Kokko H, Jennions MD (2008) Parental investment, sexual selection and sex ratios. J Evol Biol 21:919-948 Kokko H, Ots I (2006) When not to avoid inbreeding. Evolution 60:467-475 Ladurner P, Schärer L, Salvenmoser W, Rieger RM (2005) A new model organism among the lower Bilateria and the use of digital microscopy in taxonomy of meiobenthic Platyhelminthes: Macrostomum lignano, n. sp. (Rhabditophora, Macrostomorpha). J Zool Syst Evol Res 43:114-126 Landolfa MA (2002) On the adaptive function of gamete trading in the black hamlet Hypoplectrus nigricans. Evol Ecol Res 4:1191-1199 Landolfa MA, Green DM, Chase R (2001) Dart shooting influences paternal reproductive success in the snail Helix aspersa. Behav Ecol 12:773-777 Leonard JL (1991) Sexual conflict and the mating systems of simultaneously hermaphroditic gastropods. Am Malacol Bull 9:45-58 Leonard JL (2006) Sexual selection: lessons from hermaphrodite mating systems. Integr Comp Biol 46:349-367 Leonard JL, Lukowiak K (1984) Male-female conflict in a simultaneous hermaphrodite resolved by sperm trading. Am Nat 124:282-286 Lessells CM (2006) The evolutionary outcome of sexual conflict. Philos Trans R Soc Lond B 361:301-317 Lüscher A, Milinski M (2003) Simultaneous hermaphrodites reproducing in pairs self-fertilize some of their eggs: an experimental test of predictions of mixedmating and Hermaphrodite’s Dilemma theory. J Evol Biol 16:1030-1037

354

Nils Anthes

Lüscher A, Wedekind C (2002) Size-dependent discrimination of mating partners in the simultaneous hermaphroditic cestode Schistocephalus solidus. Behav Ecol 13:254-259 Mank JE, Promislow DEL, Avise JC (2006) Evolution of alternative sexdetermining mechanisms in teleost fishes. Biol J Linn Soc 87:83-93 McCarthy TM (2004) Effects of pair-type and isolation time on mating interactions of a freshwater snail, Physa gyrina (Say, 1821). Am Malacol Bull 19:4755 McCarthy TM, Sih A (2008) Relatedness of mates influences mating behaviour and reproductive success of the hermaphroditic freshwater snail Physa gyrina. Evol Ecol Res 10:77-94 Michiels NK (1998) Mating conflicts and sperm competition in simultaneous hermaphrodites. In: Birkhead TR, Møller AP (eds) Sperm Competition and Sexual Selection. Academic Press, London, pp 219-254 Michiels NK, Bakovski B (2000) Sperm trading in a hermaphroditic flatworm: reluctant fathers and sexy mothers. Anim Behav 59:319-325 Michiels NK, Koene JM (2006) Sexual selection favours harmful mating in hermaphrodites more than in gonochorists. Integr Comp Biol 46:473-480 Michiels NK, Newman LJ (1998) Sex and violence in hermaphrodites. Nature 391:647 Michiels NK, Streng A (1998) Sperm exchange in a simultaneous hermaphrodite. Behav Ecol Sociobiol 42:171-178 Michiels NK, Hohner A, Vorndran IC (2001) Precopulatory mate assessment in relation to body size in the earthworm Lumbricus terrestris: avoidance of dangerous liaisons? Behav Ecol 12:612-618 Michiels NK, Raven-Yoo-Heufes A, Kleine Brockmann K (2003) Sperm trading and sex roles in the hermaphroditic opisthobranch sea slug Navanax inermis: eager females or opportunistic males? Biol J Linn Soc 78:105-116 Michiels NK, Crowley PH, Anthes N (2009) Accessory male investment can undermine the evolutionary stability of simultaneous hermaphroditism. Biol Lett 5:709-712 Monroy F, Aira M, Velando A, Domínguez J (2005) Size-assortative mating in the earthworm Eisenia fetida (Oligochaeta, Lumbricidae). J Ethol 23:69-70 Mulvey M, Vrijenhoek RC (1981) Multiple paternity in the hermaphroditic snail, Biomphalaria obstructa. J Hered 72:308-312 Munday PL, Buston PM, Warner RR (2006) Diversity and flexibility of sexchange strategies in animals. Trends Ecol Evol 21:89-95 Norton CG, Johnson AF, Mueller RL (2008) Relative size influences gender role in the freshwater hermaphroditic snail, Helisoma trivolvis. Behav Ecol 19:1122-1127 Ohbayashi-Hodoki K, Ishihama F, Shimada M (2004) Body size-dependent gender role in a simultaneous hermaphrodite freshwater snail, Physa acuta. Behav Ecol 15:976-981 Otsuka C, Rouger Y, Tobach E (1980) A possible relationship between size and reproductive behaviour in a population of Aplysia punctata (Cuvier, 1803). Veliger 23:159-163

Mate choice and reproductive conflict in simultaneous hermaphrodites

355

Pal P, Erlandsson J, Sköld M (2006) Size-assortative mating and non-reciprocal copulation in a hermaphroditic intertidal limpet: test of the mate availability hypothesis. Mar Biol 148:1273-1282 Pennings SC (1991) Reproductive behavior of Aplysia californica Cooper: diel patterns, sexual roles and mating aggregations. J Exp Mar Biol Ecol 149:249266 Peters A, Michiels NK (1996a) Do simultaneous hermaphrodites choose their mates? Effects of body size in a planarian flatworm. Freshwater Biol 36:623630 Peters A, Michiels NK (1996b) Evidence for lack of inbreeding avoidance by selective mating in a simultaneous hermaphrodite. Invert Biol 115:99-103 Petersen CW (1995) Reproductive behavior, egg trading, and correlates of male mating success in the simultaneous hermaphrodite, Serranus tabacarius. Environm Biol Fishes 43:351-361 Petersen CW (2006) Sexual selection and reproductive success in hermaphroditic seabasses. Integr Comp Biol 46:439-448 Pongratz N, Michiels NK (2003) High multiple paternity and low last-male sperm precedence in a hermaphroditic planarian flatworm: consequences for reciprocity patterns. Mol Ecol 12:1425-1433 Preece T, Mao Y, Garrahan JP, Davison A (2009) Harmful mating tactics in hermaphrodites. Am Nat 173:632-639 Putz A, Michiels NK, Anthes N (2008) Mating behaviour of the sperm trading hermaphroditic sea slug Chelidonura hirundinina: repeated sex role alternation balances reciprocity. Ethology 114:85-94 Puurtinen M, Kaitala V (2002) Mate-search efficiency can determine the evolution of separate sexes and the stability of hermaphroditism in animals. Am Nat 160:645-660 Reise H, Visser S, Hutchinson JMC (2007) Mating behaviour in the terrestrial slug Deroceras gorgonium: is extreme morphology associated with extreme behaviour? Anim Biol 57:197-215 Ridley M (1983) The Explanation of Organic Diversity. Oxford University Press, Oxford Rogers DW, Chase R (2001) Dart receipt promotes sperm storage in the garden snail Helix aspersa. Behav Ecol Sociobiol 50:122-127 Rowe L, Arnqvist G (1996) Analysis of the causal components of assortative mating in water striders. Behav Ecol Sociobiol 38:279-286 Schärer L (2009) Tests of sex allocation theory in simultaneously hermaphroditic animals. Evolution 63:1377-1405 Schärer L, Janicke T (2009) Sex allocation and sexual conflict in simultaneously hermaphroditic animals. Biol Lett 5:705-708 Schjørring S (2009) Sex allocation and mate choice of selfed and outcrossed Schistocephalus solidus (Cestoda). Behav Ecol 20:644-650 Schjørring S, Jäger I (2007) Incestuous mate preference by a simultaneous hermaphrodite with strong inbreeding depression. Evolution 61:423-430

356

Nils Anthes

Schmitt V, Anthes N, Michiels NK (2007) Mating behaviour in the sea slug Elysia timida (Opisthobranchia, Sacoglossa): hypodermic injection, sperm transfer and balanced reciprocity. Front Zool 4:17 Sella G (1985) Reciprocal egg trading and brood care in a hermaphroditic polychaete worm. Anim Behav 33:938-944 Sella G, Lorenzi MC (2000) Partner fidelity and egg reciprocation in the simultaneously hermaphroditic polychaete worm Ophryotrocha diadema. Behav Ecol 11:260-264 Sella G, Ramella L (1999) Sexual conflict and mating systems in the dorvilleid genus Ophryotrocha and the dinophilid genus Dinophilus. Hydrobiologia 402:203-213 Sella G, Premoli MC, Turri F (1997) Egg trading in the simultaneously hermaphroditic polychaete worm Ophryotrocha gracilis (Huth). Behav Ecol 8:83-86 Siva-Jothy MT (2006) Trauma, disease and collateral damage: conflict in cimicids. Philos Trans R Soc Lond B 361:269-275 Smolensky N, Romero MR, Krug PJ (2009) Evidence for costs of mating and selffertilization in a simultaneous hermaphrodite with hypodermic insemination, the opisthobranch Alderia willowi. Biol Bull 216:188-199 Sprenger D, Anthes N, Michiels NK (2008a) Multiple mating affects offspring size in the opisthobranch Chelidonura sandrana. Mar Biol 153:891-897 Sprenger D, Faber J, Michiels NK, Anthes N (2008b) Natural female mating rate maximizes hatchling size in a marine invertebrate. J Anim Ecol 77:696-701 Sprenger D, Lange R, Michiels NK, Anthes N (2009) The role of body size in early mating behavior in a simultaneous hermaphrodite, Chelidonura sandrana. Behav Ecol Sociobiol 63:953-958 Sprenger D, Lange R, Michiels N, Anthes N (2010) Sources of phenotypic variance in egg and larval traits in a marine invertebrate. Evol Ecol, doi: 10.1007/ s10682-10009-19300-x Switzer-Dunlap M, Meyers-Schulte K, Gardner EA (1984) The effect of size, age, and recent egg laying on copulatory choice of the hermaphroditic mollusc Aplysia juliana. Int J Invert Repr Dev 7:217-225 Tian-Bi Y-NT, N’Goran EK, N’Guetta S-P, Matthys B, Sangare A, Jarne P (2008) Prior selfing and the selfing syndrome in animals: an experimental approach in the freshwater snail Biomphalaria pfeifferi. Gen Res 90:61-72 Tomiyama K (1996) Mate-choice criteria in a protandrous simultaneously hermaphroditic land snail Achatina fulica (Férussac) (Stylommatophora:Achatinidae). J Moll Stud 62:101-111 Tomlinson J (1966) The advantages of hermaphroditism and parthenogenesis. J Theor Biol 11:54-58 Trowbridge CD (1995) Hypodermic insemination, oviposition, and embryonic development of a pool-dwelling ascoglossan (= sacoglossan) opisthobranch: Ercolania felina (Hutton, 1882) on New Zealand shores. Veliger 38:203-211 Tsitrone A, Duperron S, David P (2003) Delayed selfing as an optimal mating strategy in preferentially outcrossing species: theoretical analysis of the optimal age at first reproduction in relation to mate availability. Am Nat 162:318331

Mate choice and reproductive conflict in simultaneous hermaphrodites

357

Velando A, Eiroa J, Domínguez J (2008) Brainless but not clueless: earthworms boost their ejaculates when they detect fecund non-virgin partners. Proc R Soc Lond B 275:1067-1072 Vizoso DB, Schärer L (2007) Resource-dependent sex-allocation in a simultaneous hermaphrodite. J Evol Biol 20:1046-1055 Vreys C, Michiels NK (1997) Flatworms flatten to size up each other. Proc R Soc Lond B 264:1559-1564 Vreys C, Michiels NK (1998) Sperm trading by volume in a hermaphroditic flatworm with mutual penis intromission. Anim Behav 56:777-785 Warner RR (1975) The adaptive significance of sequential hermaphroditism in animals. Am Nat 109:61-82 Webster JP, Gower CM (2006) Mate choice, frequency dependence, and the maintenance of resistance to parasitism in a simultaneous hermaphrodite. Integr Comp Biol 46:407-418 Webster JP, Hoffman JI, Berdoy M (2003) Parasite infection, host resistance and mate choice: battle of the genders in a simultaneous hermaphrodite. Proc R Soc Lond B 270:1481-1485 Wedell N, Gage MJG, Parker GA (2002) Sperm competition, male prudence and sperm-limited females. Trends Ecol Evol 17:313-320 Wethington AR, Dillon RT Jr (1996) Gender choice and gender conflict in a nonreciprocally mating simultaneous hermaphrodite, the freshwater snail, Physa. Anim Behav 51:1107-1118 Yusa Y (1996) The effects of body size on mating features in a field population of the hermaphroditic sea hare Aplysia kurodai Baba, 1937 (Gastropoda: Opisthobranchia). J Moll Stud 62:381-386

Chapter 13

Extra-pair behaviour BART KEMPENAERS AND EMMI SCHLICHT

ABSTRACT In many socially monogamous species, males and females pursue copulations with individuals other than their social mate. The outcome of this behaviour is that broods often contain offspring of mixed paternity. Here, we first show how the frequency of extra-pair paternity varies among species and among populations of the same species, and we discuss how this variation can be explained. We then examine potential costs and benefits of extra-pair behaviour for males and females. Extra-pair behaviour in both sexes might have evolved because it allows males to produce more offspring. Female extra-pair behaviour could also be favoured by selection because it increases female fitness, for example through increased offspring quality. In the third part of this chapter, we consider some of the evolutionary consequences of the occurrence of extra-pair behaviour. In particular, we evaluate how extra-pair behaviour influences the strength of sexual selection.

13.1 Introduction In many sexually reproducing organisms, a single well-timed copulation would in theory be sufficient to fertilise all the eggs in a clutch. However, in reality males and females often copulate much more frequently and they also often mate with more than one partner during a single bout of reproduction. Although such multiple mating or promiscuity comes with some obvious costs, such as increased risk of predation or increased risk of contracting a sexually transmitted disease, it turns out to be extremely common. Hence, we need an explanation for why multiple mating has evolved. In this chapter, we discuss a special case of promiscuity, which can occur in all species where individuals form pairs for breeding (BOX 13.1).

360

Bart Kempenaers and Emmi Schlicht

BOX 13.1 Pair-living and parenting patterns Pair-living or social monogamy is uncommon and unevenly distributed across the animal kingdom, but it does occur in a wide range of taxa among invertebrates (Lorenzi and Sella 2000, Lombardo et al. 2004, Baeza 2008, Beltran and Boissier 2009) and vertebrates (Sefc et al. 2008, Cohas and Allainé 2009, While et al. 2009a). Birds are unusual in that most avian species are socially monogamous (Bennett and Owens 2002). We define extra-pair behaviour as a special case of promiscuity where copulations occur with more mates than are included in the social mating system. How do behavioural ecologists know which individuals form a social pair? The male and female of a pair have social interactions that are linked to reproduction but go beyond what is required for the act of fertilisation. These interactions lead to an association of pair members in space and time, in particular during the female’s fertile period, that is not observed between unpaired individuals (Westneat et al. 1990). For example, in many birds members of a social pair typically engage in behaviours such as prolonged courtship, mate guarding, territory defence, nest building, incubation of the eggs, feeding or protecting the young, and so on. In birds, parenting patterns often reflect the social mating system (Bennett and Owens 2002) and in the majority of species social monogamy goes hand in hand with biparental care. A male of the socially monogamous American robin (Turdus migratorius) feeding the offspring in his nest. Whether he sired those offspring is another issue. Photo © Bruce E. Lyon

Biparental care may be especially common in birds because eggs and offspring need considerable parental investment to survive (Thomson et al. 1998, Tullberg et al. 2002), and male birds can contribute substantially to all parental activities except egg production. Furthermore, a high mortality cost of caring as compared to competing for mates will lead to selection favouring egalitarian sex roles and hence biparental care (Kokko and Jennions 2008). This comes about because different sex roles can cause sex-specific mortality, which may feedback on the availability of mates. Assume that sex roles have diverged a little bit so that females tend to care more than males

Extra-pair behaviour

361

and males are investing more in competing. Further sex role divergence may now be prevented if caring is the more costly activity. This is because females will become rarer as they suffer the higher risk of mortality due to caring, and the mating pool will become more strongly biased towards males. Hence, it will take deserting males longer to find another mate, and despite the higher mortality cost for caring than for competing, caring may become the better option.

An extra-pair mating can be defined as a copulation with an individual that is not the social partner, and it can lead to extra-pair paternity or mixed paternity within a brood or litter. The study of promiscuity and extra-pair behaviour is interesting, not only in itself, but also because promiscuity has several interesting evolutionary consequences. For example, it can help explain variation in the morphology of reproductive organs, in sperm production, sperm size, and sperm motility, in copulation behaviour, or in patterns of parental care. It may also lead to selection that favours behaviours such as mate guarding or male aggression. Finally, extra-pair behaviour can affect the fitness of an individual and it can dramatically increase the intensity of sexual selection, particularly in socially monogamous species. In this chapter we focus exclusively on birds. In this group of vertebrates, a large variety of mating patterns has been described, but social monogamy with biparental care is by far the most common mating system (BOX 13.1). However, parentage analyses using DNA fingerprinting techniques revealed that social monogamy often goes hand in hand with multiple mating and extra-pair paternity. Studies on birds have played a prominent role in our understanding of the evolutionary causes and consequences of extra-pair mating, partly because bird behaviour is relatively easy to study in the field. Many of the issues discussed here not only apply to birds and are more generally relevant in the context of understanding promiscuous behaviour.

13.2 The occurrence of extra-pair behaviour 13.2.1 Frequency of extra-pair paternity The frequency of extra-pair paternity, as revealed by studies using molecular techniques, varies dramatically among species (Griffith et al. 2002). In some species, extra-pair paternity seems to be absent or extremely rare. For example, in the socially monogamous Lanyu scops owl (Otus elegans

362

Bart Kempenaers and Emmi Schlicht

Y

X

W

A

B

1/7

1

C 1

D

E

4/4

0/2

1

2

1

1/4

0/5

3

1

3

1

5/5

0/7

5/6

5/6

1

0/1

1

4

5

1

1/4

1

3/6

2/6

3/4

2

6/6

1

floater # 1 4

4/5

5

6

7

6/6

1/5

2/5

5 1

2 1

floater # 2

2/5

floater # 3 3/5

2

ST-DI TG-CI

Fig. 13.1 Patterns of extra-pair paternity in a population of tree swallows (Tachycineta bicolor). Each square represents a nestbox. The numbers inside each box refer to the number of extra-pair young and the total number of offspring in the brood. Arrows indicate extra-pair males that sired offspring in a particular nest and the number next to the arrow shows how many offspring they sired. The encircled birds on the right cared for a brood elsewhere, whereas the other three birds were unpaired ‘floaters’ without their own brood. Nestboxes marked with the same colour belonged to one of four socially polygynous males. Reprinted with permission from Kempenaers et al. 2001).

botelensis) only one extra-pair offspring was detected among a total of 200 genotyped offspring from 108 families (Hsu et al. 2006). In other species, extra-pair males sire the majority of offspring. This is for example the case in the socially monogamous tree swallow (Tachycineta bicolor), where 51% of 229 offspring from 49 broods were extra-pair, and extra-pair paternity was found in 75% of all broods (Kempenaers et al. 1999; see Fig.

Extra-pair behaviour

363

13.1). The record holder is found in Australia: in the superb fairy-wren (Malurus cyaneus), a cooperatively breeding bird, extra-group paternity was detected in almost all broods (95% of 40, Mulder et al. 1994), and 61% of 1895 genotyped offspring were sired by males outside the female’s social unit (Double and Cockburn 2003). Species also vary in the distribution of extra-pair young among broods and in the number of fathers within a brood. In many cases, broods show mixed paternity, with some offspring fathered by the social male, and some by an extra-pair male. However, broods that contain exclusively extra-pair young are found in many species (Fig. 13.1). In the superb fairy-wren, this occurred in no less than 48% of the broods (Mulder et al. 1994). Females may also copulate with several extra-pair males, so that each offspring can have a different father (e.g. in the tree swallow; Whittingham et al. 2006, Dunn et al. 2009). There are two important points to be made regarding the frequency of extra-pair paternity. First, one should be aware that multiple paternity in a brood can occur for reasons other than extra-pair mating. These are: (1) rapid switching of social mates (e.g. Mee et al. 2004), (2) sequential polyandry and sperm storage (e.g. Oring et al. 1992), and (3) cooperative breeding with more than one reproductively active male (e.g. Haydock et al. 2001). The second point is that the frequency of extra-pair paternity does not necessarily reflect the frequency of extra-pair behaviour (Brommer et al. 2007, Griffith 2007). Thus, it is possible that extra-pair copulations are rather common, but rarely lead to extra-pair paternity. For example, a study on northern fulmars (Fulmarus glacialis) found no extra-pair paternity although 2.4% of copulations were extra-pair (Hunter et al. 1992). Extensive behavioural observations suggest that the social males may have sired all the offspring because they always obtained the majority of copulations and the last copulation before egg laying. Because observing copulation behaviour is time-consuming, and because extra-pair copulations may be harder to observe than within-pair copulations, it is difficult to link behaviour with paternity. Experimental work with bluethroats (Luscinia svecica) provides further insight. Fossøy et al. (2006) employed a simple device (see photo) to prevent social males from transferring sperm during copulation. Females mated to such males should lay infertile eggs, unless they performed extra-pair copulations. Indeed, Fossøy et al. (2006) showed that most of the experimental pairs had extra-pair offspring in their brood. Because the level of extra-pair paternity in natural broods is much lower, the results suggest that many females that appear truly monogamous are in fact promiscuous. Note that the opposite is also possible: a single, well-timed extra-pair copulation may sire a disproportionate number of offspring (Michl et al. 2002).

364

Bart Kempenaers and Emmi Schlicht

A male bluethroat (Luscinia svecica) with a rubber tube attached around the cloaca to prevent sperm transfer during copulation. Photo © Jan T. Lifjeld (reprinted with kind permission from Springer Science and Business Media; Fossøy et al. 2006).

Extra-pair paternity is not only the outcome of copulation behaviour, but also of post-copulatory processes such as sperm competition and cryptic female choice. The importance of sperm competition was demonstrated in a study on mallards (Anas platyrhynchos): when females were artificially inseminated with equal numbers of sperm from a brother and an unrelated male, differences in fertilisation success could be explained by betweenejaculate differences in average sperm swimming speed and motility, but not by male relatedness to the female (Denk et al. 2005). Evidence for female sperm selection comes from a study on red junglefowl (Gallus gallus): the number of sperm that were counted on the perivitelline membrane of eggs, i.e. the site of fertilisation, was lower when a female was inseminated by a brother compared to an unrelated male (Pizzari et al. 2004). Nonetheless, we still lack a general understanding of whether and how processes such as sperm selection and sperm competition, and factors such as the timing of copulations affect the success of extra-pair copulations.

Extra-pair behaviour

365

13.2.2 Determinants of variation in levels of extra-pair paternity Which factors can explain the observed variation in the frequency of extrapair paternity that occurs between species, populations and individuals? The frequency of extra-pair paternity in a population is the outcome of processes that play at the level of individuals with potentially strong conflicts of interest between males and females (Petrie and Kempenaers 1998, Westneat and Stewart 2003). Thus, to understand variation in extra-pair paternity, we need to consider the costs and benefits for individuals engaging in extra-pair matings (see below). Here, we briefly consider four groups of relevant factors and illustrate their complex interactions with some examples. (1) Species-specific factors and life-history. Comparative analyses have shown that variation in the frequency of extra-pair paternity can be explained by variation in general life-history characteristics of the species (Arnold and Owens 2002). Extra-pair paternity is most frequent in species with high adult mortality (Fig. 13.2) and high annual fecundity, in species where males provide less care (relative to the female), or in species where the effect of male care on offspring fitness is limited. In contrast, extra-pair paternity is generally rare or absent in species that are long-lived and show a low annual fecundity, and in species where male care is more important. Life-history, paternal care, and extra-pair behaviour may be directly linked through the evolutionary response of male behaviour to an increased frequency of cuckoldry. The more common extra-pair paternity becomes, the less it will pay males to care for offspring (BOX 13.2). Extra-pair behaviour and extensive paternal care may then become mutually exclusive and male care in the absence of extra-pair paternity may be the only stable equilibrium for species with slow life-histories. Yet, how exactly slow life-history prevents the invasion of cuckoldry remains unclear. Maybe it is the potential costs of extra-pair matings for females, imposed by the males’ facultative response to their mate’s infidelity (BOX 13.2). Male retaliation in terms of reduced care or divorce after perceived loss of paternity presumably has a stronger negative impact on female fitness in slow life-history species and might thus prevent females from pursuing extra-pair matings. This is also referred to as the ‘constrained female hypothesis’ (Mulder et al. 1994, Gowaty 1996). Alternatively, the association between life-history characteristics and the frequency of extra-pair paternity may be caused by a yet unknown factor.

366

Bart Kempenaers and Emmi Schlicht

BOX 13.2 Multiple paternity and paternal care It is characteristic for birds that males allocate their care to clutches in which other males may have fertilised some or all of the eggs. There is no empirical evidence that males can identify the extra-pair young and theory suggests that the evolution of kin recognition is unlikely in this context (Westneat et al. 1995, Kempenaers and Sheldon 1996, Pagel 1997; but see Johnstone 1997). Female extra-pair behaviour will then not only cause a direct reduction in a male’s fitness, but will also lead to males spending part of their parental investment on unrelated young (Trivers 1972). Should males with reduced paternity – and hence reduced benefits from caring – reduce their care? Intuitively the answer seems ‘yes’, but in fact it depends on the behavioural options available to the male (Grafen 1980). Even when the payoff from paternal care is reduced, it may remain larger than the payoff from alternative behaviours, for instance when chances of survival or remating are remote (Clutton-Brock 1984, Kempenaers and Sheldon 1997). Thus, reduced care is only expected if reallocation to alternative forms of investment (i.e. ultimately to future broods) leads to greater fitness. In general, changes in the level of extra-pair paternity can cause an evolutionary response in the care behaviour of all males or a facultative response in the care of individual males (Westneat and Sherman 1993). On an evolutionary timescale, a population-wide increase in female promiscuity reduces benefits from paternal care for males and increases mating opportunities outside the pair. The first effect alone will not cause selection for reduced male care because males will on average have similar paternity in future broods (Maynard Smith 1978, Grafen 1980, Westneat and Sherman 1993), but due to the second effect – an increased probability of future mating – reduced paternity in a species should lead to a reduction in paternal care (Queller 1997, Houston and McNamara 2002). On the individual level, a male that can assess his paternity in a brood should adjust his level of paternal care so as to maximise his fitness. Depending on the circumstances, he may achieve this by ignoring paternity loss, by reducing his care, or even by increasing his care (Kempenaers and Sheldon 1997, Whittingham and Dunn 2001, Holen and Johnstone 2007). It is not possible to identify a universally optimal male response to paternity loss. This is because (a) the effect of paternal care on the success of the brood may vary among species (Whittingham et al. 1992, Houston 1995), (b) not all males are equal, so that males differ in how their care decisions affect their prospects of survival and future matings (Westneat and Sherman 1993), and (c) the opportunity costs of male care relate to the care decisions of other individuals (Houston and McNamara 2002). For example, whether other males decide to desert or care will influence a male’s extra-pair mating opportunities. It is thus a matter of some intricacy to determine how an individual male should optimally respond when he perceives paternity loss.

Extra-pair behaviour

367

100

Rate of EFP (% of broods)

80

60

40

20

0 0

10

20

30

40

50

60

70

80

Rate of annual adult mortality (%)

Fig. 13.2 The relationship between the frequency of extra-pair paternity and adult mortality rate. Each dot represents one species. Reprinted with permission from Arnold and Owens (2002).

(2) Population-specific factors and genetic diversity. If females mate with extra-pair males to increase the genetic quality of their offspring through ‘good genes’ or ‘compatible genes’, levels of extra-pair paternity may vary depending on the availability of males with corresponding alleles (Kempenaers 2007, Lindstedt et al. 2007). One could then predict a positive relationship between levels of extra-pair paternity and population-wide variation in male genetic quality or genomewide diversity (Petrie et al. 1998). Indirect evidence comes from the observation that levels of extra-pair paternity are generally lower on island populations (Griffith 2000), which may also be less genetically diverse. For example, an island population of house sparrows (Passer domesticus) had both lower levels of extra-pair paternity and reduced genetic diversity compared to several mainland populations (Ockendon et al. 2009). When mainland birds with higher genetic diversity were released on the island, the level of extra-pair paternity strongly increased, as predicted. However, contrary to the idea that females would benefit from mating with immigrant males by maximising the genetic diversity of their offspring, all extra-pair fathers were island males.

368

Bart Kempenaers and Emmi Schlicht

(3) Population-specific factors and ecology. Levels of extra-pair paternity in a population may also vary depending on ecological factors that influence the likelihood of meeting extra-pair partners. For example, the availability of copulation partners may be larger in high-density populations, or in populations with high breeding synchrony (Stutchbury and Morton 1995). Note, however, that high breeding synchrony may also lead to lower levels of extra-pair paternity, if males face a tradeoff between mate guarding and pursuing extra-pair copulations (Westneat et al. 1990, van Dongen and Mulder 2009). Similarly, one could argue that higher levels of extra-pair paternity are expected in low-density populations, if this forces females to settle with nonpreferred males and, hence, to pursue extra-pair copulations more often. Although effects of breeding density and synchrony have been extensively discussed and tested, they rarely seem important (e.g. Kempenaers 1997, Stewart et al. 2006, Lindstedt et al. 2007, Rowe and Weatherhead 2007, but see the experimental study by Charmantier and Perret 2004). Another example is the potential effect of habitat characteristics on extra-pair paternity. Dense vegetation may allow females as well as extra-pair males to hide from guarding males and to sneak extra-pair copulations, whereas open habitats make it easier for territory owners to observe intruding males and may also increase the probability of male retaliation, and thus increase the costs of extra-pair paternity for females (Valera et al. 2003). (4) Individual-specific factors. Only few studies have addressed whether loss or gain of paternity is repeatable for individual males, females or pairs (e.g. Dietrich et al. 2004). At the level of the individual, the question thus remains whether some individuals are more likely to engage in extra-pair behaviour than others (everything else being equal), and what factors play a role. Males or females may intrinsically differ in a variety of ways that influence their propensity to show extra-pair behaviour. For example, individuals may differ in their genetic predisposition to be promiscuous (Forstmeier 2007), or individual extrapair behaviour may depend on maternal effects (e.g. hormone levels in the egg) or on the rearing environment. Males may also differ in quality, as reflected in age, condition, or expression of sexual ornaments, which may influence their capacity to gain extra-pair matings or to avoid losing paternity. One of the more robust findings in studies on extra-pair paternity is that older males are more likely to sire extrapair offspring (e.g. Delhey et al. 2003, Bitton et al. 2007, Bouwman et al. 2007, Schmoll et al. 2007a). Similarly, females may differ in quality, condition, size, or experience (age), which may influence their ability to cope with reduced male investment, and hence the potential

Extra-pair behaviour

369

costs of engaging in extra-pair matings. Alternatively, larger or stronger females might be better at avoiding costly extra-pair copulations and, hence, are less likely to have extra-pair offspring (Rosivall et al. 2009). In reality, it will often be difficult to disentangle the factors that play a role in creating a particular level of extra-pair paternity in a population, unless experiments are conducted. The following examples illustrate how difficult it is to unequivocally interpret results even in carefully designed studies. In the coal tit (Parus ater) Dietrich et al. (2004) showed that the level of extra-pair paternity increased significantly from first to second broods (overall and within pairs). There are several, not always mutually exclusive explanations (see also discussion in Dietrich et al. 2004). (a) Second broods may be less valuable to males so they would invest less in paternity protection strategies. (b) Males whose mates did not lay a second clutch may have invested heavily in extra-pair behaviour causing high paternity loss for males whose mates did produce second broods. (c) Young in second broods may face harsher conditions, so that effects of good or compatible alleles on offspring fitness are more important (Schmoll et al. 2005, Kempenaers 2007), and selection would favour females who engage more often in extra-pair behaviour. (d) Females may invest more in extra-pair behaviour because the environmental conditions have improved so that they are less dependent on male care. (e) Females may have increased opportunities to engage in extra-pair copulations for the second brood because the eggs of the second brood are fertilised when the young of the first brood still require care and males have to trade off mate-guarding with offspring care. The second example concerns studies on the relationship between weather conditions and the level of extra-pair paternity. In a Norwegian population of bluethroats, low temperatures during the peak fertile period resulted in lower levels of extra-pair paternity (Johnsen and Lifjeld 2003), presumably because individuals invested more in self-maintenance during a spell of bad weather. Moreover, it may be harder for males or females to detect suitable extra-pair partners in adverse weather conditions. Finally, when weather conditions are likely to stay bad during the nestling phase, male care may become more important, and this may cause a shift in female behaviour towards less extra-pair behaviour. Similar reasoning can explain why rainfall during the fertile period caused a decrease in levels of extra-pair paternity in the reed bunting (Emberiza schoeniclus: Bouwman and Komdeur 2006). However, in this species, low temperatures led to increased levels of extra-pair paternity, which the authors explained by reduced mate guarding because of investment in self-maintenance when temperatures are low. What can we learn from such conflicting results?

370

Bart Kempenaers and Emmi Schlicht

Clearly, the level of extra-pair paternity is hard to predict, even when information about variation in very specific ecological parameters is available. This is because the level of extra-pair paternity is a population parameter that emerges from behavioural interactions between individuals, and we are still far from understanding the complex dynamics causing population level changes. However, we can focus on particular behaviours or behavioural interactions and use the above results to predict the effect of e.g. weather conditions on mate guarding in male reed buntings or on solicitation of extra-pair copulations in female bluethroats. Ultimately, seemingly contradictory results may not be inconsistent after all. Weather conditions likely affect general activity patterns and may do so more in one sex than in the other or more for particular activities depending on the specific ecological situation and on the species. The final example is an experimental study on song sparrows (Melospiza melodia: MacDougall Shackleton et al. 2006) where food and predator presence was manipulated. Extra-pair paternity was more frequent when the environment was more challenging (no extra food, predators present), and females with extra-pair offspring were more stressed than faithful females. This contradicts the hypothesis that females that are more in need of male help are less likely to engage in extra-pair behaviour. Instead, general activity patterns may again have changed in response to the experimental manipulation, leading also to changes in the exposition to and opportunities for extra-pair mating attempts. For example, hungry females might move around more and, hence, meet more extra-pair males, and females might be less likely to conspicuously resist copulation attempts if predators are around. Arnold and Owens’ (2002) comparative analysis showed that 55% of the interspecific variation in the frequency of extra-pair paternity in birds is explained by the taxonomic classification (family and order). They thus suggested a hierarchical explanation for the variation in extra-pair paternity. Variation among major avian lineages can best be explained by variation in the potential costs and evolutionary consequences of extra-pair behaviour, as explained above. Variation among closely related species, or among populations of the same species, may be better explained by the opportunities to engage in extra-pair copulations and the benefits that can be gained from this behaviour. It would be interesting to find out whether this hierarchical explanation can be extended to other vertebrates. The available data are too scant to allow formal analysis, but indicate that there may be as much variation in levels of extra-pair paternity in other taxa as observed in birds (Sefc et al. 2008, Cohas and Allainé 2009, While et al. 2009a).

Extra-pair behaviour

371

13.3 Evolutionary causes of extra-pair behaviour When the existence of extra-pair behaviour is the result of natural and sexual selection, the fitness benefits an individual derives from acquiring additional mates must be higher than the costs associated with it, at least for some individuals. Costs are not only those that come about directly from performing the behaviour, but also include opportunity costs. These arise from missing the opportunity to perform an alternative behaviour. Extrapair behaviour will trade-off with behaviours such as self-maintenance, mate guarding or parental care. An individual that invests in extra-pair behaviour can only do so at the cost of investing less into such other behaviours, thereby foregoing other fitness benefits. Generally, the reproductive success of male birds is mainly limited by the access to females, because males can produce more gametes than they can obtain eggs to fertilise. Therefore, males should always copulate with additional females, unless there are substantial costs associated with this behaviour. In contrast, female birds usually cannot increase their offspring number by mating with more males and it is therefore not so clear whether they should mate multiply, even when this behaviour is not costly. Due to this initial difference between the sexes, conflicts of interest between the members of a socially monogamous pair (and other individuals) may arise over the level of promiscuity. The observed extra-pair behaviour will be the evolutionary outcome shaped by these conflicts of interest. Here we discuss the costs and benefits of engaging in extra-pair behaviour for males and females separately. 13.3.1 Benefits and costs of extra-pair matings for females 13.3.1.1 Female benefits of extra-pair behaviour Selection can act on females to mate multiply in two ways (Jennions and Petrie 2000). (1) Selection may favour females who increase their number of mates independent of the quality of their social mate. (2) Selection may favour females who improve on the quality of their social mate by ensuring fertilisation by extra-pair males of higher quality. In both cases the fitness increase for multiply mated females can be brought about directly via an increase in female fecundity, in female survival, or in the number or non-genetic quality of the offspring that can be raised (direct benefits). Alternatively, fitness of multiply mated females can increase indirectly through genetic effects on the fitness of offspring. Most likely direct and indirect benefits often occur jointly in the same species (Jennions and Petrie 2000).

372

Bart Kempenaers and Emmi Schlicht

A. Direct benefits Material benefits. The fitness of a female can be enhanced directly via extra-pair behaviour when she gains material benefits in the form of additional male investment such as courtship feeding, nutritional resources in the ejaculate, offspring care, access to additional resources etc. In this context, it is important to note that female multiple mating will reduce the fitness benefit of parental investment for males (BOX 13.2). Also, distributing paternal care among several nests would lead to a reduction in efficiency, especially in territorial species. Selection will therefore favour males who focus their care on the brood(s) where their expected paternity is highest (Souzou and Houston 1994, Iwasa and Harada 1998). This makes it unlikely that females can gain additional male investment such as offspring care from extra-pair mates. Furthermore, when females only seek extra-pair copulations to obtain additional investment, a simple way in which males could reduce multiple mating of their female is to provide her with as much care as possible (cf. Hunter and Davis 1998). Taken together this should concentrate material benefits within the social pair (Akçay and Roughgarden 2007) and inhibit selection for female extra-pair behaviour solely through benefits such as paternal care. In line with this, material benefits from extra-pair copulations are generally thought to be unimportant (e.g. Birkhead and Møller 1992:201-203). Nevertheless, a few studies have found evidence for such benefits and we discuss those cases in more detail below. In the great grey shrike (Lanius excubitor), a socially monogamous species with biparental care, males give food to females immediately before copulation. Tryjanowski and Hromada (2005) showed that extra-pair females are provided with prey items of high energy value that are costly to hunt (Fig. 13.3). Why do males not allocate this investment to their social mate? The benefit for males from this investment is twofold: it not only functions as parental effort but also as mating effort (Møller and Thornhill 1998). Forms of male investment that occur after fertilisation, such as incubation or feeding young, primarily serve to increase offspring fitness and cannot influence mating success anymore (although they might sometimes influence copulation opportunities for future broods, e.g. Freeman-Gallant 1996). In contrast, providing females with courtship gifts may create extrapair opportunities that are otherwise absent. Additional access to fertile females may also explain why male redwinged blackbirds (Agelaius phoeniceus) allow extra-pair mates to forage on their territory (Gray 1997a). Material benefits in the form of foraging opportunities could further be facilitated because males have to defend a territory for their social nest(s) in any case, and the cost of allowing other

Extra-pair behaviour

373

14

Number of prey items

12 10 8 6 4 2 0

Birds

Voles

Lizards

Insects

Without gift

Fig. 13.3 Number of courtship gifts offered by male great grey shrikes (Lanius excubitor) to females in the context of within-pair copulations (open bars) or extrapair copulations (filled bars). Handling time of hunting shrikes is longer for vertebrate prey (birds, voles, lizards) than for insects. Successful copulation attempts are associated with gifts of higher energy content. Reprinted with permission from Tryjanowski and Hromada (2005).

females to forage in their territory may be minimal. Finally, it has been observed that unpaired males in some species provided material benefits to females when they copulated with them (Hunter and Davis 1998, Blomqvist et al. 2005). In this case, the care-providing extra-pair males do not have their own brood to care for. Such a situation shows some resemblance to a cooperative breeding system where several males share in siring offspring and in providing care. Similarly, males may start caring for an ‘extra-pair brood’ after their ‘social brood’ failed (Kempenaers 1993). Insurance benefits. An increase in the number of mates can also lead to higher fecundity because it ensures fertilisation when some males are unable to fertilise a female’s eggs (Sheldon 1994). Even if a male produces functional sperm, he may be unable to fertilise the ova of his social mate due to temporary sperm depletion (Wetton and Parkin 1991) or due to a failure to overcome female barriers against polyspermy (Morrow et al. 2002). There is considerable variation among bird species in hatching success, but no positive relationship with the rate of extra-pair paternity has been found (Morrow et al. 2002). However, such a relationship is not nec-

374

Bart Kempenaers and Emmi Schlicht

essarily expected because the present state may represent an evolutionary response to past selection against infertility (Sheldon 1994), so that species with high levels of extra-pair paternity are now best insured against infertility (Lifjeld 1994). Fertility benefits are difficult to study both for theoretical and practical reasons. First, with observational data it is almost impossible to exclude the hypothesis that females obtain indirect benefits from extra-pair behaviour. Whenever there is some heritable component to an extra-pair male’s ability to fertilise a female’s eggs that her social mate failed to fertilise, genetic benefits are inevitable and it is difficult to quantify the evolutionary importance of one benefit relative to the other (Sheldon 1994, Griffith et al. 2002, Griffith 2007). Second, it may be problematic to ascertain whether eggs have failed due to lack of fertilisation or due to early embryo death (e.g. Kempenaers et al. 1996, Friedl and Klump 2005, but see Birkhead et al. 2008). Temporary sperm depletion is perhaps most likely in polygynous species. Indeed, the strongest evidence for fertility benefits comes from two socially polygynous species (Gray 1997b, Friedl and Klump 2005), whereas evidence from studies of socially monogamous species is relatively weak (Wetton and Parkin 1991, Wagner 1992, Lifjeld 1994, Krokene et al. 1998). Extra-pair behaviour may not only function as an insurance against male infertility but also against mate loss through death of the partner, if it enables multiply mating females to find a new male more quickly (Petrie and Kempenaers 1998). This hypothesis has not been explicitly studied, but interspecific variation in the rate of extra-pair paternity is positively associated with variation in mortality rates (see above, Arnold and Owens 2002). However, such a pattern can easily be explained in other ways because mortality will fundamentally influence the costs and benefits of infidelity in several respects (Mauck et al. 1999, Wink and Dyrcz 1999, Arnold and Owens 2002, Jeschke and Kokko 2008). Insurance benefits could also come into play as a result of genetic pleiotropy when the same genes control the frequency of within- and extrapair copulations (positive genetic correlation; Arnqvist and Kirkpatrick 2005, Forstmeier 2007). It is likely that there is an optimal frequency of within-pair copulations, for example to avoid infertility of eggs, to secure male care, or to strengthen the pair bond. Hence, selection may favour female extra-pair behaviour indirectly because only unfaithful females copulate frequently enough with their social partner to avoid infertility or mate loss. A related benefit of female extra-pair behaviour could lie in mate sampling (Heg et al. 1993) and pair formation (Colwell and Oring 1989). Females in suboptimal pair bonds could assess available males via extra-pair

Extra-pair behaviour

375

behaviour and initiate new social relationships for the following season. In socially monogamous bird species a high divorce rate was positively associated with a high rate of extra-pair paternity (Cézilly and Nager 1995). Again, it is easy to come up with alternative explanations for such an association, because the divorce rate of a species also emerges as a trait from complex interactions between individuals (Cézilly and Nager 1995, Catry and Furness 1997). Diversity benefits. Multiple paternity broods will show a higher genetic diversity. This may directly increase the success of the brood, if it leads to positive interactions among nestlings (Yasui 1998; for tests and discussion see Schmoll et al. 2007b and Dunn et al. 2009). For example, genetic diversity may hinder the transfer of infectious diseases among offspring (Jennions and Petrie 2000). Here it is considered a direct and not an indirect benefit because it is independent of offspring genomic quality itself. It has to be kept in mind, however, that the presence of half-siblings will lead to an increase in sibling-competition as an ecological (Boncoraglio and Saino 2008) or evolutionary (Briskie et al. 1994) response to decreased relatedness among nest-mates. This might be beneficial in some cases when it eliminates genetically inferior competitors quickly (Jennions and Petrie 2000), but in general is expected to be against the parents’ interest (Birkhead and Møller 1992:204). B. Indirect benefits The genetic quality of an offspring will influence its fitness and thus indirectly maternal fitness. Females could therefore benefit from extra-pair behaviour by optimising the genetic quality of their offspring. How females should behave will depend on the information they can obtain about expected fitness returns. It may sometimes be impossible to maximize offspring genetic quality, because the optimal male genotype is uncertain or not readily identifiable. In such a situation genetic bet hedging may be adaptive (Yasui 1998; for tests and discussion see e.g. Schmoll et al. 2007b). Multiple mating produces offspring of variable genetic quality, whereas broods sired by only one male contain young that are more or less of the same (superior or inferior) genetic quality. The average genetic quality of a brood with multiple sires is therefore less likely to be extreme. Hence, increasing genetic diversity by mating multiply causes a reduction in fitness variance between broods. In the long run, this reduction in between-brood variance in offspring quality is associated with increased fitness. This is because over

376

Bart Kempenaers and Emmi Schlicht

evolutionary time, above-average success with one mate cannot compensate similar below-average success with another (the geometric mean is more relevant for long-term fitness than the arithmetic mean; Gillespie 1977, Philippi and Seger 1989). However, it has been convincingly argued that fitness benefits from genetic bet-hedging are usually limited and unlikely to be important in maintaining female promiscuity, except in special cases (Yasui 2001). When females are able to determine the optimal genetic constellation for their offspring, their choice of a paternal genotype should be influenced by two effects: allelic quality and allelic compatibility. The former refers to the absolute, additive effects certain paternal alleles (‘good genes’) have on the fitness of the offspring genotype (Jennions and Petrie 2000, Neff and Pitcher 2005). These will be alleles that increase viability or attractiveness (mating success) and they confer their fitness effects relatively independent of the genetic background in which they happen to occur. Genetic compatibility refers to the effects of interactions within a genome between alleles at the same locus (e.g. heterozygote advantage) or at different loci (epistasis). In the context of mate choice, genetic compatibility issues might arise when the fitness effects of the alleles from the maternal haplotype and other maternal genetic elements depend in a non-additive way on the paternal haplotype they are combined with. This means that the quality of offspring genotypes does not only depend on the sum of the effects of the maternal and paternal alleles, but also on their interaction (Zeh and Zeh 1996, Brown 1997, Neff and Pitcher 2005). Female choice for allelic quality and compatibility are not mutually exclusive (Jennions and Petrie 2000) and it is expected that females trade off one benefit against the other when they are negatively associated (Neff and Pitcher 2005). When female preferences for genetically superior or compatible males drive female extra-pair behaviour, extra-pair copulations are expected only when females can improve on the social mate’s genotype. Thus, females that secured the genetically optimal male as social mate should not engage in extra-pair copulations for genetic benefits. The good genes and the compatible genes hypothesis make different predictions about the congruence among females in their extra-pair mate choice: additive effects lead to agreement about who is the ‘top-quality’ male, whereas non-additive effects lead to a different ranking of males for different females (Zeh and Zeh 1996, Brown 1997, Neff and Pitcher 2005). However, both hypotheses predict that extra-pair young should have higher quality genomes than within-pair young of the same brood (maternal half-sib comparisons; Griffith et al. 2002, Arnqvist and Kirkpatrick 2005, Akçay and Roughgarden 2007). A recent meta-analysis found only small and presumably evolutionary irrelevant effects of paternity status on offspring fitness (Arnqvist and

Extra-pair behaviour

377

Kirkpatrick 2005; see also Akçay and Roughgarden 2007). This suggests that selection on females to pursue extra-pair copulations for genetic benefits is absent. However, the estimates on which this meta-analysis is based can be criticized for three reasons. (1) Estimates of offspring fitness that can be obtained from data in the field and that are reported in studies on extra-pair paternity do not necessarily represent an important component of total fitness, because they often do not give information on residual reproduction or offspring reproduction (Eliassen and Kokko 2008). (2) Differences in genetic quality among offspring within a brood need not transfer into easily measurable differences in fitness when genetic effects explain only a small amount of the variation in offspring fitness (Møller and Alatalo 1999). Yet, they might constitute the most important fitness consequence of a female’s mating decision, which is then shaped by these minor differences because a small fitness benefit can still be selected for over evolutionary time. (3) Parental effects may differ among the young in a brood (Price 1998). This could confound the genetic differences (Griffith et al. 2002) and conceal them when parents invest more into within-pair offspring. Such compensatory allocation (Bluhm and Gowaty 2004) can be very difficult to control for because it can occur already at the stage of egg production (Saino et al. 2002, Bolund et al. 2009; see also Magrath et al. 2009). There is evidence for both good genes and genetic compatibility benefits from studies on a variety of species (e.g. Hasselquist et al. 1996, Blomqvist et al. 2002, Forstmeier et al. 2002, Eimes et al. 2005, Fossøy et al. 2008, Kawano et al. 2009). For example, in a study on the blue tit (Cyanistes caeruleus) extra-pair offspring were more heterozygous than withinpair offspring from the same nest (Fig. 13.4a). Heterozygosity is associated with reproductive success and sexual attractiveness in this species presumably because it reduces the number of recessive deleterious alleles expressed or increases the diversity of gene products synthesized (Foerster et al. 2003, García-Navas et al. 2009). Interestingly, the difference in heterozygosity was only present when extra-pair sires were not close neighbours (Fig. 13.4a). This is because females were generally less related to more distantly breeding males (Foerster et al. 2006). Hence, females that copulated with extra-pair males breeding further away obtained compatibility benefits (Foerster et al. 2003). Further, it could be shown that only close neighbours, but not non-neighbours, were older and larger than the social male they cuckolded (Fig. 13.4b). This suggests that females may have obtained good genes benefits from mating with neighbouring extra-pair males. Thus, in this population the details are more complicated because different females may obtain different genetic

0.10

NS

P<0.10 P<0.001

0.08 0.06 0.04 0.02 0

101

15

44

58

-0.02 -0.04

c

-0.06

b  Tarsus length (EP - WP )

P<0.05

 Heterozygosity (EPY - WPY)

 Heterozygosity (EPY - WPY)

a

Bart Kempenaers and Emmi Schlicht

P<0.01

NS

0.2 0.1 0

73 31

-0.1 -0.2 All extra- Close Local non- Non-local males pair males neighb. neighb.

0.10

P<0.05

P<0.05

NS

79

62

23

P<0.001

P<0.01

NS

47

39

10

All extrapair males

Local males

Non-local males

0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06

d  PHA-response (EPY - WPY)

378

0.2

0.1

0

Fig. 13.4 Evidence for genetic benefits of extra-pair behaviour in blue tits (Cyanistes caeruleus; panel a and b on the left) and bluethroats (Luscinia svecica; panel c and d on the right). Top panels (a, c): in both species extra-pair young (EPY) were more heterozygous than their within-pair half-sibs (WPY). In the blue tit this difference was found only when the extra-pair sires bred far away from the focal female (non-neighbours and non-locals in panel a). Bottom panels (b, d): in the blue tit extra-pair sires (EP♂) that were close neighbours to the female were older (not shown) and larger than the males they cuckolded (WP♂; panel b). In the bluethroat, extra-pair offspring showed a stronger immune response (PHA test) than their within-pair half-sibs (panel d). The two effects in the bluethroat (panel c and d) were independent from each other and both were present only when the extra-pair sire was a local male. Left panels: adapted with permission from Foerster et al. (2003). Right panels: reprinted with permission from Føssoy et al. (2008).

benefits from their extra-pair behaviour. Similarly, in bluethroats, extrapair offspring were more heterozygous and showed a higher cell-mediated immune response than their within-pair maternal half-sibs (Fig. 13.4c, d).

Extra-pair behaviour

379

However, both these effects were independent of each other (Fossøy et al. 2008), again suggesting that females may obtain different genetic benefits from engaging in extra-pair copulations. The previous example illustrates that the type of selection on female extra-pair behaviour is likely to vary not only between species and populations but also between individuals (Petrie and Kempenaers 1998, Westneat and Stewart 2003). Hence, measuring an average effect size by drawing together measurements across many studies on different species might not be very meaningful (Eliassen and Kokko 2008). Complex patterns can also arise when individual females differ in the benefits they derive from extrapair behaviour depending on their condition (Friedl and Klump 2005, Dreiss et al. 2008), or on environmental conditions (e.g. Johnsen and Lifjeld 2003, Schmoll et al. 2005, Bouwman and Komdeur 2006, Garvin et al. 2006, O’Brien and Dawson 2007). As discussed above, intricate relationships are probably the rule rather than the exception. Direct and indirect benefits or different types of indirect benefits may often occur simultaneously. Furthermore, whenever extra-pair copulations occur, indirect benefits will necessarily be invoked because all females possess pre- or postcopulatory mechanisms that make fertilisations more probable for certain paternal haplotypes. The question is thus ‘What is the importance of female genetic benefits for maintaining extra-pair behaviour relative to other female benefits, to female costs, and to male costs and benefits?’ Another genetic benefit of female extra-pair behaviour may result when alleles causing an increase in extra-pair behaviour have opposite effects on male and female fitness (intralocus sexual conflict). If the genetic basis underlying promiscuity is identical for both sexes (i.e. if there is a positive genetic correlation between male and female extra-pair behaviour), strong selection favouring male extra-pair behaviour will result in increased extra-pair behaviour of both sexes (Halliday and Arnold 1987). The evolutionary stable level of promiscuity could then be above the level optimal for females because it reflects a compromise between the costs to females and the benefits to males. However, there would also be positive selection acting on females to express such a high level of extra-pair behaviour because without this the male descendants of their line would be unsuccessful as extra-pair sires. Hence, in this case the genetic benefit for females is the increased extra-pair behaviour of their sons and grandsons. Benefits of extra-pair copulations are then rather concealed and females may even appear to suffer a cost, although the behaviour increases their overall fitness.

380

Bart Kempenaers and Emmi Schlicht

13.3.1.2 Female costs of extra-pair behaviour There are at least four types of costs associated with engaging in extra-pair copulations (listed in Westneat et al. 1990). First, as with any behaviour, the costs of extra-pair mating to females are associated with the time and energy needed to express the behaviour, i.e. with the availability of additional mates. Second, copulations with additional mates may increase the risk of predation or of infection with parasites or sexually transmitted diseases. Third, a male may react to potential infidelity of his mate with mate guarding, frequent copulation, physical punishment, reduced care or divorce which may all be costly for females. This group of costs is thus a manifestation of sexual conflict between females and their social mates. Fourth, producing broods containing half- instead of full-siblings tends to increase sibling competition, which in most cases will be costly to parents. Should males reduce their investment in parental care when their paternity is lower? And do they? It remains a challenge both to predict how males should respond (BOX 13.2) and to test whether they respond (BOX 13.3). To date there is no generally accepted study design that allows a decisive evaluation of the hypothesis of facultative male responses to reduced (certainty of) paternity (Sheldon 2002; Fig. 13.5). The strongest experimental setup to address this question has been developed in a study on bluegill sunfish (Lepomis macrochirus, Neff 2003). It makes use of the fact that males that tend a nest can assess their relatedness to the fry (recently hatched young) via chemical cues, but not to the unhatched eggs (Neff and Sherman 2003, 2005). By experimentally reducing perceived paternity during the egg stage and then revealing the true paternity during the fry stage, Neff has been able to induce an increase in paternal care from the egg to the fry stage that was not found in untreated controls. Furthermore, when true paternity was experimentally lowered without providing the male with a cue of cuckoldry at the egg stage, a decrease in paternal care occurred. Thus, when males are able to assess true paternity reliably at a well-defined point in time, manipulations of perceived and true paternity become much more meaningful, especially because the response was shown in a randomly selected sample of males and in opposite directions between the two experiments. Reduced male care is often considered the most important cost of extrapair matings to females (e.g. Birkhead and Møller 1992). Across species the rate of extra-pair paternity and the level of paternal care are inversely associated (Møller 2000, Møller and Cuervo 2000, Arnold and Owens 2002). A high male contribution to brood care may have constrained the evolution of extra-pair behaviour, so that divergence in parental care patterns between major avian lineages has led to different extra-pair paternity

Extra-pair behaviour

381

BOX 13.3 Studying male facultative responses In species with biparental care, selection can favour males that adjust their level of care to their (perceived) share of paternity (BOX 13.2). Many studies have attempted to empirically test such facultative male responses using correlational or experimental approaches (Wright 1998, Whittingham and Dunn 2001, Sheldon 2002). Unfortunately, correlations between levels of male care and share of paternity are hard to interpret (Kempenaers and Sheldon 1997, Wright 1998). The main issue is that young of unfaithful females may receive less paternal care than those of faithful females even without the existence of facultative male responses. This is because a third factor, such as territory quality or male quality, condition and age can cause a decrease both in male investment and in within-pair success. Most of these factors can be excluded by comparing successive breeding attempts of the same pairs in the same season, but even here alternative explanations are possible (Sheldon 2002; Fig. 13.5). Experimental approaches are more promising to test for a facultative male response (Kempenaers and Sheldon 1997, Sheldon 2002). However, here the difficulty is to ensure that the experimental treatment is effective in manipulating a male’s perception of his paternity (Wright 1998). Studies that manipulated male certainty of paternity or perceived paternity (reviewed by Whittingham and Dunn 2001 and Sheldon 2002) provide some support (Arnqvist and Kirkpatrick 2005), though not unequivocal evidence (Sheldon 2002, Griffith 2007), for the existence of facultative male responses. Yet, this does not mean that males that had their confidence of paternity experimentally reduced and invested less in the brood in response would ever have suffered paternity loss or that all males would have responded in this way. It is possible and often even expected that the males with the lowest risk of losing paternity suffer the highest opportunity costs of care and are selected to react strongest to paternity loss, and vice versa (Eliassen and Kokko 2008). The experimental studies can be seen as evidence for the existence of strong facultative responses at least in some males. These males could be mated to females that remain faithful, maybe also because of the risk of losing male care otherwise. Nevertheless, the females that are most strongly selected to have extra-pair young could well be the females whose social mates cannot provide a high amount of care and are selected to ignore paternity loss, because these males cannot improve their fitness otherwise. Thus, even if females that engage in extra-pair behaviour receive only a small amount of paternal care – as indicated by some observational studies – this does not imply that extra-pair behaviour comes with the cost of reduced male care for females that engage in it.

Difference male provisioning rate

382

Bart Kempenaers and Emmi Schlicht

a

0.3

b

2.0

0.6

1.0

0.1

0.4

0.5

0.2

0.0

0.0

0.0

-0.1

-0.5

-0.2

-1.0

-0.4

-0.2

-0.6

-1.5

-0.3 -1

-0.5

0

0.5

1

Difference proportion EPP

c

0.8

1.5

0.2

-1

-0.5

0

0.5

1

Difference proportion EPP

-0.8 -1

-0.5

0

0.5

1

Difference proportion EPP

Fig. 13.5 Paternity and paternal effort in three studies of the reed bunting (Emberiza schoeniclus) using the same study design. To exclude confounding factors, the relationship between the change in the level of extra-pair paternity and the change in male provisioning rate (feeds per hour per nestling) between two sequential broods of the same pair is considered. (a) UK population (Dixon et al. 1994). (b) Dutch population (Bouwman et al. 2005). (c) Swiss population (Suter et al. 2009). In (a) and (c) the negative correlation is significant, whereas in (b) there is no significant association (note that in (a) changes in provisioning are rather small in absolute terms). A significant relationship can be seen as evidence that individual males adjust their parental effort to their level of paternity in the nest. However, a similar pattern could arise when males trade off somatic effort (investment in survival) against mating and parental effort combined (Magrath and Komdeur 2003). This trade-off may vary for individual males over the course of a season. Males that invest more into survival early in the season will have lower paternity (low mating effort) in the first brood and care less (low parental effort) for this brood, whereas males that invest more into survival later in the season will have lower paternity and care less in the second brood. (a) Adapted from Dixon et al. (1994) with permission, (b) reprinted with permission from Bouwman et al. (2005), (c) reprinted with permission from Suter et al. (2009).

rates (Arnold and Owens 2002, Griffith et al. 2002; but see Møller 2000 and Møller and Cuervo 2000 for alternative explanations). Could the importance of paternal care also constrain female extra-pair behaviour at the population or the individual level? A recent meta-analysis emphasized that the cost of reduced male care, measured for individual females, appears to be high, especially when compared to the benefits in the form of good or compatible genes (Arnqvist and Kirkpatrick 2005). However, the cost measurements on which this conclusion is based may not reflect responses of males to female extra-pair behaviour (BOX 13.3; Eliassen and Kokko 2008). Leaving this aside, the estimates of these costs (Arnqvist and Kirkpatrick 2005) are negatively correlated with the rate of extra-pair paternity for this selection of species (Albrecht et al. 2006), supporting the

Extra-pair behaviour

383

idea that at the species level females that suffer the highest costs of extrapair copulations are the least likely to engage in this behaviour. The costs of loss of male care can be reduced experimentally, for example by providing females with additional food. One would then predict that such females are more likely to engage in extra-pair behaviour, but studies that used this approach produced mixed results (Hoi-Leitner et al. 1999, Václav et al. 2003). Hence, the evolutionary relevance of the cost of reduced male care in shaping patterns of female extra-pair behaviour probably varies between species. 13.3.2 Benefits and costs of extra-pair matings for males 13.3.2.1 Male benefits of extra-pair behaviour In general, males may gain the same benefits from extra-pair behaviour as females. However, unlike females, males should always accept additional copulations when these can be acquired cheaply (Trivers 1972). Hence, selection will strongly favour males that increase their number of mating partners, independently of the quality of their social mate. In contrast, selection to improve the quality of the offspring through extra-pair copulations with high-quality females will be relatively unimportant. Thus, male benefits from extra-pair behaviour are mostly direct benefits, namely acquiring additional female investment in the form of eggs. Males can also gain benefits from extra-pair paternity that do not apply to females, i.e. benefits that arise from distributing offspring among several nests or from exploiting the care of other males. This requires that the female does not place the extra-pair eggs in the nest of the sire (quasiparasitism), a behaviour that appears to be rare in birds (Griffith et al. 2004). For example, by siring young in multiple broods, males could bethedge against the effects of total offspring loss through nest predation (Webster et al. 2007), but such benefits are generally thought to be small (Bulmer 1984, Hopper et al. 2003; but see Pöysä and Pesonen 2007). Furthermore, if extra-pair young are of higher quality than their within-pair half-sibs (see above), they might experience reduced sib competition compared to the level of competition they would experience among their full siblings (Holen and Johnstone 2007). However, as outlined above, the decreased relatedness among young of a mixed paternity brood can also cause an increase in sibling competition (Boncoraglio and Saino 2008), which would work against any such benefits to the extra-pair male.

384

Bart Kempenaers and Emmi Schlicht

13.3.2.2 Male costs of extra-pair behaviour For males the cost of extra-pair copulation itself is probably small. Whenever copulations can be acquired cheaply, males only pay the costs associated with performing the copulation and producing viable sperm. As for females, the most important costs of copulating are an increased risk of predation and the risk of infection with parasites or with a sexually transmitted disease. In contrast to females, the cost associated with the availability of additional mates is most likely a major factor for males. Fertile females will always be a limited resource because males should accept all extra-pair copulations that can be acquired cheaply. When competition for females is high, the opportunity costs associated with attempts to obtain extra-pair copulations are high, and this can lead to selection favouring male investment in other behaviours (cf. Kokko and Jennions 2008). Specifically, extra-pair behaviour may trade off with securing paternity in the nest of the social mate, because females may be free to pursue extra-pair copulations when their social mate is absent. This will lead to males that invest into paternity protection behaviours, such as frequent copulation or mate guarding, rather than courting additional females (Westneat et al. 1990, Westneat and Stewart 2003). Pursuing extra-pair copulations may also trade off against providing care for the offspring in the nest of the social mate (Westneat et al. 1990), especially when material investment is needed to acquire extra-pair mates. Thus, males may refrain from extra-pair copulations because caring for offspring has a higher fitness pay-off. Note that such a trade-off can also explain the relationship found across species between the level of paternal care and the rate of extra-pair paternity discussed earlier (Møller 2000, Møller and Cuervo 2000). Independent of the availability of mates and the opportunity costs of extra-pair behaviour there can be a fixed cost for every additional mate a male acquires when material benefits are important for maintaining female extra-pair behaviour. 13.3.3 Is extra-pair behaviour male or female driven? The question whether extra-pair behaviour is male or female driven has received much attention in the recent literature (e.g. Westneat and Stewart 2003, Arnqvist and Kirkpatrick 2005, Akçay and Roughgarden 2007, Eliassen and Kokko 2008). We use the term ‘male (or female) driven’ in the sense of evolutionary changes in extra-pair behaviour resulting from selection on males (or females). The question is tightly linked to the net-effect of extra-pair behaviour on male and female fitness. If the costs of extra-

Extra-pair behaviour

385

pair behaviour for females outweigh the benefits, infidelity is selected against in females and will only occur as a result of selection on males to pursue extra-pair copulations. Note, however, that extra-pair behaviour can also be male driven when females gain a net benefit from cuckoldry, as long as selection for infidelity is much stronger in males than in females, which seems likely. Then, evolutionary changes in extra-pair behaviour would occur chiefly through selection on males, even though selection favouring the behaviour acts on both sexes. Furthermore, as discussed above, male and female extra-pair behaviour may be genetically correlated, so that expression of extra-pair behaviour in females may be non-adaptive and a consequence of strong selection on male behaviour. Then females appear to suffer a net cost, but extra-pair behaviour is not the result of male pursuit alone. It is not easy to determine whether the extra-pair behaviour observed in a population is the result of selection on males, females or both. The answer to this question may appear obvious if judged by the behaviour of females. For example, in the New Zealand stitchbird (or Hihi, Notiomystis cincta) groups of extra-pair males perform prolonged chases of fertile females which alarm call and apparently try to evade the males. When a male can successfully get hold of a female, the female is pinned with her back to the ground during copulation (Castro et al. 1996). Observations such as these have been made in a variety of species and have been referred to as ‘forced copulations’ (McKinney et al. 1983, Birkhead and Møller 1992:104-106, Westneat and Stewart 2003). They strongly suggest that in many instances females actively avoid extra-pair copulations, reflecting a sexual conflict. However, resistance against copulation attempts does not necessarily equal female avoidance of costly copulations, because females may ensure copulations with high quality males by allowing only the most persistent males to copulate (resistance as a ploy: Cox and Le Boeuf 1977, Westneat et al. 1990). Furthermore, females are often not equally resistant towards all males (female manipulation hypothesis: Westneat and Stewart 2003) and even solicit copulations from some while fleeing from others. In the blue tit and in the superb fairy-wren, fertile females conduct forays to the territories of other males which are the sires of some of the offspring in their brood (Kempenaers et al. 1992, Double and Cockburn 2000). In the fairy-wrens these forays occur very early in the morning when it is still dark and are the first movement of females after leaving their roost site. They consist of rapid flights directly to the centre of the visited territory and back (Double and Cockburn 2000). Upon arrival at the male’s territory, females were observed to perch near the male, and to closely approach the male after he started a courtship display until copula-

386

Bart Kempenaers and Emmi Schlicht

tion took place (Cockburn et al. 2009). Thus, females of some species apparently actively pursue extra-pair copulations, which – at least for an adaptationist – is hard to reconcile with the idea that extra-pair copulations come at a net cost for females. Similarly, in the Adélie penguin (Pygoscelis adeliae) females have been observed to join extra-pair males at their nest site and solicit courtship and copulation (Hunter and Davis 1998). After copulation however, these females picked up a stone from the extra-pair male’s nest which they then brought to their own nest. Adélie Penguins’ nests consist of a pile of stones and these stones are a limited resource in their breeding area. Females therefore probably trade nesting material for copulations and it is impossible to determine whether they leave their nest site for extra-pair copulations or in search of stones. This illustrates the general problem to determine the function of female forays (Westneat and Stewart 2003). When the copulation itself is not observed – which can be exceedingly difficult – it is impossible to rule out that females visit the territories of extra-pair males in search of nesting material or food and copulations are the result of male coercion upon encounter. In sum, studying female behaviour can give strong indications in one or the other direction, but it does not allow a conclusive decision about whether extra-pair behaviour comes at a net cost or benefit to females, because all forms of female behaviour are open to alternative explanations (Westneat et al. 1990, Westneat and Stewart 2003). Nevertheless, in most bird species it seems that some female cooperation is necessary for successful copulation (Gowaty and Buschhaus 1998). Also, females can control paternity to a certain degree after copulation, at least in some species (Birkhead and Møller 1993, Pizzari et al. 2004). When the female has evolved the ability to prevent fertilisation by sperm from particular males and extra-pair fertilisations still occur, this suggests a net benefit of extra-pair paternity to females. However, because such control is never perfect, and because males will be selected to exploit the ‘gaps’ (Rice and Holland 1997), extra-pair offspring could be the result of fertilisations escaping cryptic female choice. Interestingly, a cost-benefit analysis of male extra-pair behaviour has rarely been made, probably because it is assumed that the benefits from siring additional young cared for by other males are overwhelming. However, a study on great tits (Parus major) that examined the net effect of extra-pair behaviour for individual males found that males that gained extrapair young also had a higher than average probability to lose paternity in their nest (Fig. 13.6a; Lubjuhn 2005). It is not necessarily true that paternity loss could have been avoided if those males would not have engaged in extra-pair copulations. Yet, if there is indeed a trade-off, males pursuing extra-pair copulations might not do better than those that guard their mate

60

P = 0.004

5

50

N = 42 pairwise comparisons

4

40 30 20 10

387

b

Number of EPY

a

Proportion of broods with EPY (%)

Extra-pair behaviour

P = 0.003

3 2 1 0

0

Genetic sires ( N = 42)

Population mean (N = 233)

Losses

Gains

Fig. 13.6 (a) Comparison of the proportion of broods containing extra-pair young in the great tit (Parus major). The proportion observed among the broods of 42 males identified as extra-pair sires is higher than the proportion expected based on the population mean. (b) Comparison of paternity loss (in own brood) and paternity gain (extra-pair offspring sired) for the same 42 males. Shown are mean and standard deviation.

(i.e. alternative reproductive strategies with equal pay-off; see Taborsky and Brockman this volume). In the great tit study this was not the case: the number of young gained in extra-pair nests more than compensated for the number of young lost in the social nest (Fig. 13.6b). Up to now we have treated the male and female perspective separately to simplify the line of argument. However, it is important to realise that in reality these two viewpoints are merged. Every extra-pair copulation involves the reproductive interests of at least the copulating male and female, and the female’s social partner. Therefore, male and female costs and benefits inescapably come into play simultaneously and the behaviour of all ‘players’ determines the adaptive value of the realised outcome. In the superb fairy-wrens discussed above, paternity analysis showed that satellite males were successful in securing on average one third of extragroup fertilisations by mimicking the behaviour of attractive males in a sort of reproductive parasitism (Cockburn et al. 2009). Thus, despite strong behavioural evidence for female ‘control’, males are obviously often able to evade this control. In the superb fairy-wren males do so inconspicuously, but in other species they resort to more overt behaviour such as harassing and coercing females. The selective effect in both sexes drives evolutionary changes in mating behaviour and both sexes will evolve adaptations to maximise their fitness payoff from mating interactions. The

388

Bart Kempenaers and Emmi Schlicht

often very different costs and benefits from the male and female perspective outlined above emphasise the sexual conflict that is inherent in the evolution of extra-pair behaviour (Petrie and Kempenaers 1998, Snook 2001, Chapman et al. 2003, Westneat and Stewart 2003, Wedell et al. 2006).

13.4 Evolutionary consequences of extra-pair mating 13.4.1 Overview The occurrence of extra-pair paternity has a range of important evolutionary consequences. We first provide a brief general overview of these consequences, and then discuss one of them in detail. (1) Selection on male and female copulation behaviour. Selection will favour males who develop behavioural strategies that allow them to maximise the chances to obtain extra-pair copulations. This could lead to selection on male display behaviour, or on sneaky behaviour in search of fertile females. Similarly, females may develop behavioural strategies that increase the probability to obtain extra-pair copulations from favoured males at peak fertility and that avoid copulations with unattractive males (see Westneat and Stewart 2003 for a review). (2) Selection on male parental behaviour. See BOXes 13.2 and 13.3. (3) Evolution of paternity protection behaviour. If males risk losing paternity either because other males intrude on their territory or because females leave the territory to visit other males, selection may favour paternity protection behaviour such as territory defence, close mate guarding, and frequent copulation during the period when their mates are fertile (see Birkhead and Møller 1992). (4) Selection on sperm production and sperm characteristics. Female birds can store sperm in storage tubules for a period of at least a week (Birkhead and Møller 1992:60-64). This means that when females copulate with multiple males, sperm from different males will compete for the fertilisation of the egg(s), a process known as sperm competition. This will lead to selection on male traits that improve their success in sperm competition, such as increased sperm production, increased sperm swimming speed and increased sperm longevity. (5) Selection on sperm selection mechanisms. If females benefit from having their offspring sired by a particular male, mechanisms may have evolved that increase the likelihood that the ‘right’ sperm will fertilise the egg. This cryptic female choice is expected to be more

Extra-pair behaviour

389

important in species where females have less behavioural control over copulations, for example species where males have an intromittent organ and can force copulations on females (see Eberhard 2009). (6) Selection on offspring behaviour. If offspring in a brood are half-sibs rather than full-sibs, increased competition is expected, which may manifest itself in the loudness of begging or in aggression among offspring ( see e.g. Royle et al. 1999, Kilner and Hinde 2008). (7) Effects on the intensity of sexual selection. See next section. 13.4.2 Extra-pair paternity and sexual selection 13.4.2.1 Variance in mating success is the basis for sexual selection Sexual selection is selection acting on variation among individuals in their ability to obtain access to mating partners. The extent of this variation often differs considerably between males and females, with males commonly showing higher variance. Because variation provides the substrate for selection (BOX 13.4), greater variance in mating and reproductive success generally causes males to experience stronger sexual selection than females. This is known as Bateman’s principle. Sex differences in the strength of sexual selection are the primary cause of many of the sex differences in behaviour and morphology that can be found in organisms. They can explain the evolution of conspicuous traits, such as ornamentation, weaponry or display behaviour in one sex and they may account for the particular expression of sex roles and mating systems in different species (Andersson 1994). Pronounced sexual dimorphism is thought to be a sign of a strong sex difference in the strength of sexual selection (Dale et al. 2007). In a strictly monogamous species there is no variation in mate number and variation in offspring numbers is equal for males and females. As a consequence, selection pressures on males and females are identical and there is little scope for sexual conflict. A priori, the evolution of sexual dimorphism and sex-specific roles is suppressed. In a strictly monogamous species one would therefore – at least theoretically – expect both sexes to exhibit similar levels of intrasexual competition for mates and intersexual mate choice. Because of the similarity in the intensity of sexual selection one would also expect that males and females are endowed with ornaments and armaments to the same degree. However, in many apparently monogamous bird species males and females differ substantially in characteristics such as size, plumage colour, ornaments, song, etc. Prior to the discovery of extra-pair paternity this pre-

390

Bart Kempenaers and Emmi Schlicht

BOX 13.4 Fitness variance and the strength of selection Fitness variance is caused by variation between individuals in reproductive success, which in turn can arise from variance in mating success. When fitness is completely heritable, directional selection causes a change in fitness from one generation to the next with a rate that is equal to the variance in fitness before selection (Fisher 1930). The traits that cause variation in fitness will change at a rate depending on their contribution to heritable fitness, but their rate of change can never exceed that of fitness itself. Hence, fitness variance places an upper bound on the change in any phenotypic trait from one generation to the next and can be used to measure the maximum strength of selection (Crow 1958, O’Donald 1970). For standardisation, variance in relative fitness is compared, which is equivalent to dividing variance in (absolute) fitness by mean (absolute) fitness squared. This estimate is now independent of the particular study system and is called the opportunity for selection I (Arnold and Wade 1984). Because they are standardised, measurements of I are frequently used to compare the strength of selection between populations or species. However, this can be criticised for three main reasons. Measurements of I are (1) expected to increase with mean fitness values (Downhower et al. 1987, Ruzzante et al. 1996), and (2) expected to increase with sample size (Ruzzante et al. 1996). Furthermore, (3) measurements of I may not correlate with the strength of selection when the importance of random variance differs between the compared groups (Sutherland 1985a,b). The (standardised) variance in fitness is only one of many possibilities to estimate inequality of fitness among individuals (Kokko et al. 1999, Fairbairn and Wilby 2001). Several alternative indices have been developed, all of which behave differently with respect to changes in inequality (Kokko et al. 1999). Unfortunately, all of them suffer to some degree from the problems described above. The great advantage of estimating inequality by the opportunity for selection I is the close connection to selection theory, allowing a direct interpretation with regard to the evolutionary process (Jones et al. 2002, 2004).

sented a conundrum (Mock 1985, Møller and Birkhead 1994). Sexual dimorphism in monogamous species can sometimes be explained by a biased sex ratio (causing many males to obtain no mates) or by competition related to mate quality instead of mate number (Darwin 1871:221-222). However, such explanations should in principle apply equally to both sexes. It thus remains difficult to see why female choice and male-male competition should be more important than male choice and competition among females. Extra-pair paternity allows a relaxation of the constraints imposed on sexual selection by strict monogamy. Could extra-pair pater-

Extra-pair behaviour

391

nity cause sex differences in sexual selection in socially monogamous birds and account for the observed sexual dimorphism? 13.4.2.2 Extra-pair paternity and sexual selection: measurements Realized and apparent variance in mating success. Several studies have attempted to assess the role of extra-pair paternity for the sex difference in the strength of sexual selection by comparing standardised variances (BOX 13.4) in two measurements of male reproductive success: (1) the reproductive success a male would have had without extra-pair paternity, i.e. the number of young raised with the social mate (apparent reproductive success), and (2) the number of young a male actually sired in his and other males’ nests (realised reproductive success). When extra-pair paternity contributes to the sex difference in sexual selection, variance in realised reproductive success should be larger than variance in apparent reproductive success (Gibbs et al. 1990, Mock and Fujioka 1990). This is because an increase in sexual selection on males through extra-pair paternity is only expected when some males can gain paternity at the cost of others (instead of males merely exchanging paternity). In order to affect sexual selection, extra-pair paternity should thus increase the variation among males in their reproductive success, that is, the ratio Irealised / Iapparent should be larger than 1. The published studies that report this ratio have produced mixed results (Tables in Møller and Ninni 1998, Freeman-Gallant et al. 2005, Whittingham and Dunn 2005, and Albrecht et al. 2007). The interpretation is difficult because unassigned extra-pair sires can artificially inflate standardised variances of realised reproductive success (FreemanGallant et al. 2005). Comparisons of Irealised / Iapparent ratios between studies may therefore be of limited value. Fitness components. An alternative approach to estimate the effect of extra-pair paternity on the strength of sexual selection is based on variance in realised male reproductive success. This variance can be partitioned into a component due to extra-pair success, a component due to within-pair success, and a component due to the covariance between the two (Webster et al. 1995). Further partitioning allows accounting for the relative importance of mate number, mate quality as reflected by investment in reproduction (e.g. clutch size), and fertilisation success (as well as their covariances) in producing each component. Figure 13.7 illustrates this method of variance partitioning. Based on this example, the effect of extra-pair paternity can now be assessed in three different ways.

392

Bart Kempenaers and Emmi Schlicht

Variance in WP success 41%

Variance in total ♂ reproductive success

Variance in No. of WP ♀

0%

Variance in WP ♀ quality

6%

Variance in WP fertilisation success 36% Covariance and remainder terms

-1%

Variance in No. of EP ♀

33%

Variance in EP ♀ quality

1%

Variance in EP fertilisation success

2%

Covariance and remainder terms

9%

Covariance (WP, EP) (2x) 13 %

Variance in EP success 46%

Fig. 13.7 Estimates of components of variance in reproductive success in male eastern kingbirds (Tyrannus tyrannus). Variance in total (T) male reproductive success (number of young) is partitioned into components due to within-pair (WP) success, extra-pair (EP) success, and the covariance between the two. Var (T) = Var (WP) + Var (EP) + 2Cov (WP, EP); see Webster et al. (1995). Variance in within- and extra-pair success is further partitioned into components due to mate number, mate quality (clutch size), and fertilisation success (proportion of clutch sired). Shown are the percentages of the total variance explained by each component. Terms indicative of the importance of extra-pair paternity for variance in male reproductive success are highlighted. Data from Table 4 in Dolan et al. (2007).

(1) Variance in extra-pair success explains almost half (46%) of the total variance in male reproductive success. This shows that extra-pair paternity constitutes an important path through which sexual selection can act in this species. (2) The covariance between extra-pair and within-pair success is positive. This indicates that males that are successful as extra-pair sires also tend to be successful with their social mate. Thus, increased extra-pair success is not generally cancelled out by decreased paternity in the own nest. Rather, extra-pair paternity shifts offspring towards successful sires, thereby increasing the variance in male reproductive success and the opportunity for (sexual) selection on males. (3) Variance in within-pair fertilisation success is also high (36% of total variance). This is variance among males in their ability to secure paternity in their own nest. When paternity loss constitutes an important

Extra-pair behaviour

393

part of the total variance in reproductive success, as is here the case, much of the observed opportunity for selection is mediated by extrapair paternity. Many studies have found at least one of these three indications for an effect of extra-pair paternity on the opportunity for selection (e.g. Webster et al. 1995, Weatherhead and Boag 1997, Kleven et al. 2006, Albrecht et al. 2007). In other studies evidence is equivocal or weak (Webster et al. 2001, Freeman-Gallant et al. 2005, Whittingham and Dunn 2005). Again, differences between populations and species are probably important and deserve further study. However, there are two caveats that should be kept in mind. First, there is no general agreement on the size of variance proportions needed to infer an important role of extra-pair paternity (e.g. compare Webster et al. 2001 with Whittingham and Dunn 2005). Second, in some studies estimates are highly inconsistent between years (Freeman-Gallant et al. 2005, Webster et al. 2007; see also Weatherhead and Boag 1997, Kleven et al. 2006). This indicates that estimates may not be representative of long-term selection pressures either because annual fluctuations are highly influential or because of the low level of confidence associated with one-year estimates of fitness components (cf. Griffith et al. 2002). The rationale of using variance-based estimates is that they represent opportunity measurements. Thus these approaches rest on the assumption that the strength of sexual selection on males can be compared by calculating the opportunity for selection I (BOX 13.4). It is important to recall that the opportunity for selection is not a measurement of sexual selection itself. I only provides an upper limit to the strength of directional selection on any trait and thus also limits sexual selection. Sexual dimorphism however is thought to rest on a sex difference in the strength of sexual selection. The variance-based estimates presented above are not necessarily strongly associated with the effect of extra-pair paternity on this sex difference. For example, much of the variation in male reproductive success may stem from effects outside sexual selection and females may also experience strong sexual selection. Bateman gradients. An alternative method to estimate the strength of sexual selection, the Bateman gradient, directly links mating and reproductive success (BOX 13.5). Including this measurement into analyses of sexual selection may add to the picture and provide more reliable information on the strength of sexual selection than variance-based estimates alone (BOX 13.5; Jones et al. 2004, 2005, Mills et al. 2007; see also Bjork and Pitnick 2006).

394

Bart Kempenaers and Emmi Schlicht

BOX 13.5 The Bateman gradient The strength of selection on a particular trait depends on how fitness is affected by the trait in question. This can be formally estimated by regressing (relative) fitness on the phenotypic trait value and illustrated by plotting fitness against the trait value. Fitness can also covary with a trait due to selection on correlated traits (indirect selection). To remove these fitness effects the partial regression of relative fitness on the trait in question is used, which then estimates only direct selection on a trait. The selection gradient β is the corresponding partial regression coefficient (Lande 1979). In the context of sexual selection, we are interested in the trait ‘ability to obtain mates’. The selection gradient for this trait is called the Bateman gradient (Bateman 1948, Andersson and Iwasa 1996). It measures the association between the number of mates (phenotypic trait) and the number of offspring produced (fitness value), that is, it shows the influence of additional mates on reproductive success and is thus a direct measure of the strength of sexual selection. Furthermore, a combination of a positive male and a negative female Bateman gradient may be an indicator of sexual conflict over mating events. A negative Bateman gradient in females can result when additional matings are harmful to them (Levitan 2008). In general, selection gradients provide an estimate of evolutionary change for given fitness and error variances. For example, different Bateman gradients (values of β) for males and females imply a sex difference in the strength of sexual selection. This is true as long as the larger gradient is not associated with large error variance or with smaller selection opportunity (small fitness variance), which generally applies for comparisons between the sexes. However, these assumptions need not hold for other comparisons. Furthermore, the size of β is specific to the particular situation in which it is measured (Arnold and Wade 1984, Andersson 1994:91-94). Therefore it is problematic to use β for comparisons of the strength of selection between traits, populations, and species. The opportunity for selection (BOX 13.4) provides a standardised and comparable measure of the maximal strength of selection I. However, comparing estimates of I instead of β ignores some of the information contained in the Bateman gradient, which might be biologically meaningful (Jones et al. 2004, 2005, Mills et al. 2007).

In the context of extra-pair mating, Bateman gradients have potential to reveal the role of extra-pair paternity for creating a sex difference in sexual selection. In an otherwise monogamous system, extra-pair mating can create variation between individuals in the number of mates (copulation partners) within a season. We can then estimate the Bateman gradient for both sexes, reflecting the presence of variation in mate number in both sexes. We can further estimate the difference in the male and female Bateman

Extra-pair behaviour

395

Average number of young sired

6 5 4 3 2 1 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Average number of mates per season

Fig. 13.8 Bateman gradient illustrating the relationship between number of mates and reproductive success in the splendid fairy-wren (Malurus splendens). Male reproductive success is strongly correlated with mate number (F1,159 = 261.7, R2 = 0.62, P < 0.0001). This is a cooperatively breeding socially monogamous species where auxiliary males associate with a breeding pair and help raising the offspring. Auxiliary males routinely sire young in their own group. Therefore only the reproductive success of breeding males is shown here. Breeding males can obtain multiple mates only through extra-pair copulations. Lifetime reproductive success was recorded, but success per season is shown to remove the effect of differences in lifespan. Reprinted with permission from Webster et al. (2007).

gradient, which corresponds to the sex difference in sexual selection caused by extra-pair paternity. Only few studies on birds have calculated Bateman gradients (Woolfenden et al. 2002, Webster et al. 2007, Krakauer 2008, Balenger et al. 2009). In a study of the splendid fairy-wren (Malurus splendens), a strong positive effect of extra-pair mates on male reproductive success has been found (Fig. 13.8; Webster et al. 2007). This is further corroborated by variance partitioning, which showed that a high proportion of the total variance in male reproductive success (> 30%) is caused by variation in extrapair success. Thus, in this study the Bateman gradient suggests that extrapair paternity is an important source of sexual selection on males and might contribute to the sex difference in sexual selection. This could be the ultimate cause of the pronounced plumage dimorphism of splendid fairywrens: the body of breeding males is covered in different shades of lus-

396

Bart Kempenaers and Emmi Schlicht

trous blue while females are mostly brown. It could also be the reason why male splendid fairy-wrens exhibit rather extreme reproductive morphology (Tuttle et al. 1996) with large testes, large cloacal protuberances and huge numbers of stored sperm (up to 8.3 billion). Indeed, these are common characteristics of all Australian members of the fairy-wrens (genus Malrus), with the exception of the purple-crowned fairy-wren (M. coronatus), incidentally the only species with little sexual dimorphism and low levels of extra-pair paternity (Kingma et al. 2009). 13.4.2.3 Extra-pair paternity and sexual selection: correlates and consequences Exaggeration of male traits such as plumage brightness, tail length or testes size is thought to be a sign of intense sexual selection on males. The estimates of sexual selection intensity from the studies mentioned above indicate that extra-pair paternity can sometimes help explain the presence of such traits in socially monogamous species. Extra-pair paternity probably has a stronger influence on sexual selection in species where it is common. Based on this logic, comparative studies have tested for a correlation between the frequency of extra-pair paternity and indicators of sexual selection on males. Across species, high levels of extra-pair paternity are associated with relatively larger testes and brighter plumage in males (Møller and Birkhead 1994, Møller and Briskie 1995), as well as pronounced sexual dimorphism in plumage coloration (Møller and Birkhead 1994, Owens and Hartley 1998, Dunn et al. 2001), and in wing and tail length (Dunn et al. 2001). In many species sexual selection thus appears to operate through the path opened by extra-pair mating. Still, evolutionary effects of extra-pair paternity are less important than effects of the social mating system (i.e. dimorphism in polygynous and lekking species is often even more pronounced; Dunn et al. 2001). This is not surprising because social monogamy will tend to equalize selection on males and females as long as it is not completely annulled by extra-pair paternity, whereas in polygynous species many males may not obtain a single mate. The fact that higher levels of extra-pair paternity can cause an increase in sexual selection on males beyond that observed in species with less infidelity has two important consequences. First, sexual selection on males via extra-pair copulations goes hand in hand with a potential for genetic benefits from extra-pair behaviour to females. Sexual selection on males through extra-pair mating is selection on heritable variation among males in their ability to obtain extra-pair paternity. As a result, the genetic constitution of extra-pair young will differ from that of within-pair young in fitness-relevant traits, at least for male offspring. Second, the effect of extra-

Extra-pair behaviour

397

pair paternity on the strength of sexual selection on males provides indirect information about the net effect of extra-pair behaviour on male fitness. The estimates of the strength of selection mentioned above can be used to examine the balance between costs and benefits of male infidelity. When (a) additional mates lead to additional offspring (positive Bateman gradient), (b) extra-pair success positively covaries with within-pair success (analysis of fitness components), and (c) extra-pair paternity increases variance in male reproductive success (higher Irealised / Iapparent ratios), extrapair fertilisations increase the success of certain males at the expense of others. Thus, an increase in the strength of sexual selection on males through extra-pair paternity is indicative of a net benefit of male extra-pair behaviour. In summary, several studies have attempted to quantify the effect of extra-pair paternity on sexual selection by estimating selection strength or by comparing indicators of sexual selection. Some studies have found evidence for an important role of extra-pair paternity (e.g. Owens and Hartley 1998, Dolan et al. 2007), but some of the more recent studies caution against an overestimation (Dunn et al. 2001, Freeman-Gallant et al. 2005, Whittingham and Dunn 2005). Independent of quantification, the effect of extra-pair paternity on sexual selection in the context of social monogamy is of a fundamental nature. It releases the constraint of equal mate and offspring numbers for males and females. Thus, extra-pair paternity generates greater plasticity in avian mating systems, which sets the stage for sexual conflict and for a far wider field of evolutionary processes than possible under the restrictions of genetic monogamy.

13.5 Outlook In the sections above, we provided an overview of what is currently known about the causes and consequences of extra-pair behaviour in birds. As usual in science, new insights generate new questions and many older questions are still not fully answered. We will finish this chapter by sketching some of the problems that await to be tackled and some approaches that might be useful.  As discussed in some detail benefits of extra-pair behaviour to females remain controversial. There is a need for research in two directions here. First, studies that compare fitness-relevant traits of maternal halfsibs should separate genetic and non-genetic (e.g. hatch order) effects. Such studies should also examine the influence of environmental fac-

398

Bart Kempenaers and Emmi Schlicht

tors on the relative performance of extra-pair young, for instance by comparing broods under various levels of environmental stress. This is necessary because fitness differences between maternal half-sibs may only reveal themselves under more stressful environmental conditions. Second, experiments should be performed whereby the fitness of females that are free to pursue extra-pair copulations is compared with the fitness of females that are forced to be monogamous. A comparable experimental design has been successfully used in a variety of other taxa to address the fitness benefits of polyandry for females (e.g. Tregenza and Wedell 1998, Fisher et al. 2006, Firman and Simmons 2008). Transferring it to birds in the wild – to obtain the most relevant fitness measures – is going to be quite a challenge, however.  So far, most studies have focused on male quality and differences among males in explaining variation in paternity loss or gain. It appears interesting to extend this type of inquiry to females. How does female quality affect female extra-pair behaviour? Do females change their mating behaviour depending on age, or on their condition? Along similar lines, do females change their mating behaviour depending on variation in environmental stress?  Many of the cost-benefit considerations we described above assume total behavioural flexibility of individuals. That is, we assume that an individual bird is able to freely adapt its behaviour so as to optimize its fitness in every feasible situation it might be confronted with during its lifetime. This is undoubtedly an overly adaptionist view. In fact, behaviour might not be this plastic (see Kappeler and Kraus this volume). Research on (avian) personality shows that an individual’s behaviour may be correlated over a wide range of different contexts, such as escape situations, object handling, exploration, social dominance or aggressiveness. This means that expressing a particular behaviour comes with a whole suite of other behaviours, which may or may not be adaptive in the individual’s environment (Bergmüller this volume). In mating system research, the existence of a ‘sexual personality’ has not been explicitly considered, but first attempts have been made at linking variation in personality traits with variation in extra-pair behaviour (van Oers et al. 2008, While et al. 2009b). Indeed, it might well be that an individual’s propensity to form a strong or loose pair bond, to invest more in parental care or in courtship, to show a low or high sex drive, etc. are correlated characteristics that lead to variation in extra-pair behaviour. Newly developed methods from quantitative genetics now make it feasible to unravel the underlying genetic basis of these traits.

Extra-pair behaviour

399

 Many behavioural ecologists that study extra-pair mating still focus exclusively on an individual’s behaviour before and up to copulation. However, much insight can be gained from studying post-copulatory processes, including behaviour (e.g. sperm ejection; Pizzari and Birkhead 2000). In the arena of avian sperm competition and cryptic female choice, where it is likely that outcomes are determined by intrinsic individual characteristics and that restricted flexibility is the rule rather than the exception, there is still much scope for discovery.  Above, we mentioned the ongoing debate on whether extra-pair behaviour is male or female driven and emphasised that the male and female side have to be considered simultaneously. One interesting challenge here is to pinpoint the antagonistic and synergistic evolutionary dynamics, i.e. to answer the question when is it cooperation, when conflict? The discovery of extra-pair behaviour in birds and the study of its causes and consequences would not have been possible without molecular techniques that allow reliable paternity assessment. However, studying paternity is not the same as studying behaviour, and the former has become easier than the latter. It is remarkable how little we still understand about how the pattern of paternity in a brood relates to the copulation behaviour of the female. The difficulties of studying behaviour in the wild are clear, but even in captive birds, few researchers have undertaken in-depth investigations of the link between copulation behaviour and paternity. Many of the issues discussed above would be much better understood if more was known about the behaviour. The challenge thus has changed from developing new molecular tools to developing new methods to study behaviour in the field in much more detail (as for example in Cockburn et al. 2009). The study of extra-pair paternity in birds has revealed a plasticity in mating systems formerly unknown and produced much insight into the multitude of factors underlying these dynamics. Because birds have been studied for so long and in so much detail they offer the unique opportunity to tackle problems that go deep into the biology of the species and require much background knowledge. However, birds represent only a fraction of animal diversity. The findings on avian extra-pair paternity should be an incentive for students of other animal taxa to extend our understanding beyond birds.

400

Bart Kempenaers and Emmi Schlicht

Acknowledgements We are grateful to Pernilla Borgström, Wolfgang Forstmeier, Peter Kappeler, Holger Schielzeth, Lotte Schlicht and David Westneat for valuable comments on this chapter.

References Akçay E, Roughgarden J (2007) Extra-pair paternity in birds: review of the genetic benefits. Evol Ecol Res 9:855-868 Albrecht T, Kreisinger J, Piálek J (2006) The strength of direct selection against female promiscuity is associated with rates of extrapair fertilizations in socially monogamous songbirds. Am Nat 167:739-744 Albrecht T, Schnitzer J, Kreisinger J, Exnerová A, Bryja J, Munclinger P (2007) Extrapair paternity and the opportunity for sexual selection in long-distant migratory passerines. Behav Ecol 18:477-486 Andersson M (1994) Sexual selection. Princeton University Press, Princeton Andersson M, Iwasa Y (1996) Sexual Selection. Trends Ecol Evol 11:53-58 Arnold KE, Owens IPF (2002) Extra-pair paternity and egg dumping in birds: life history, parental care and the risk of retaliation. Proc R Soc Lond B 269:12631269 Arnold SJ, Wade MJ (1984) On the measurement of natural and sexual selection: theory. Evolution 38:709-719 Arnqvist G, Kirkpatrick M (2005) The evolution of infidelity in socially monogamous passerines: the strength of direct and indirect selection on extrapair copulation behavior in females. Am Nat 165:26-37 Baeza JA (2008) Social monogamy in the shrimp Pontonia margarita, a symbiont of Pinctada mazatlanica, off the Pacific coast of Panama. Mar Biol 153:387395 Balenger SL, Johnson LS, Mays Jr. HL, Masters BS (2009) Extra-pair paternity in the socially monogamous mountain bluebird Sialia currucoides and its effect on the potential for sexual selection. J Avian Biol 40:173-180 Bateman AJ (1948) Intra-sexual selection in Drosophila. Heredity 2:349-368 Beltran S, Boissier J (2009) Are schistosomes socially and genetically monogamous? Parasitol Res 104:481-483 Bennett PM, Owens IPF (2002) Evolutionary Ecology of Birds: Life Histories, Mating Systems and Extinction. Oxford University Press, Oxford Berglund A, Rosenqvist G, Svensson I (1988) Multiple matings and paternal brood care in the pipefish Syngnathus typhle. Oikos 51:184-188 Birkhead TR, Møller AP (1992) Sperm Competition in Birds: Evolutionary Causes and Consequences. Academic Press, London Birkhead TR, Møller AP (1993) Female control of paternity. Trends Ecol Evol 8:100-104

Extra-pair behaviour

401

Birkhead TR, Hall J, Schut E, Hemmings N (2008) Unhatched eggs: methods for discriminating between infertility and early embryo mortality. Ibis 150:508517 Bitton P-P, O’Brien EL, Dawson RD (2007) Plumage brightness and age predict extrapair fertilization success of male tree swallows, Tachycineta bicolor. Anim Behav 74:1777-1784 Bjork A, Pitnick S (2006) Intensity of sexual selection along the anisogamyisogamy continuum. Nature 441:742-745 Blomqvist D, Andersson M, Küpper C, Cuthill IC, Kis J, Lanctot RB, Sandercock BK, Székely T, Wallander J, Kempenaers B (2002) Genetic similarity between mates and extra-pair parentage in three species of shorebirds. Nature 419:613-615 Blomqvist D, Fessl B, Hoi H, Kleindorfer S (2005) High frequency of extra-pair fertilisations in the moustached warbler, a songbird with a variable breeding system. Behaviour 142:1133-1148 Bluhm CK, Gowaty PA (2004) Reproductive compensation for offspring viability deficits by female mallards, Anas platyrhynchos. Anim Behav 68:985-992 Bolund E, Schielzeth H, Forstmeier W (2009) Compensatory investment in zebra finches: females lay larger eggs when paired to sexually unattractive males. Proc R Soc Lond B 276:707-715 Boncoraglio G, Saino N (2008) Barn swallow chicks beg more loudly when broodmates are unrelated. J Evol Biol 21:256-262 Bouwman KM, Komdeur J (2006) Weather conditions affect levels of extra-pair paternity in the reed bunting Emberiza schoeniclus. J Avian Biol 37:238-244 Bouwman KM, Lessells CM, Komdeur J (2005) Male reed buntings do not adjust parental effort in relation to extrapair paternity. Behav Ecol 16:499-506 Bouwman KM, van Dijk RE, Wijmenga JJ, Komdeur J (2007) Older male reed buntings are more successful at gaining extrapair fertilizations. Anim Behav 73:15-27 Briskie JV, Naugler CT, Leech SM (1994) Begging intensity of nestling birds varies with sibling relatedness. Proc R Soc Lond B 258:73-78 Brommer JE, Korsten P, Bouwman KM, Berg ML, Komdeur J (2007) Is extrapair mating random? On the probability distribution of extrapair young in avian broods. Behav Ecol 18:895-904 Brown JL (1997) A theory of mate choice based on heterozygosity. Behav Ecol 8:60-65 Bulmer MG (1984) Risk avoidance and nesting strategies. J Theor Biol 106:529535 Castro I, Minot EO, Fordham RA, Birkhead TR (1996) Polygynandry, face-toface copulation and sperm competition in the hihi Notiomystis cincta (Aves: Meliphagidae). Ibis 138:765-771 Catry P, Furness RW (1997) Territorial intrusions and copulation behaviour in the great skua Catharacta skua. Anim Behav 54:1265-1272 Cézilly F, Nager R (1995) Comparative evidence for a positive association between divorce and extra-pair paternity in birds. Proc R Soc Lond B 262:7-12

402

Bart Kempenaers and Emmi Schlicht

Chapman T, Arnqvist G, Bangham J, Rowe L (2003) Sexual conflict. Trends Ecol Evol 18:41-47 Charmantier A, Perret P (2004) Manipulation of nest-box density affects extrapair paternity in a population of blue tits (Parus caeruleus). Behav Ecol Sociobiol 56:360-365 Clutton-Brock TH (1984) Reproductive effort and terminal investment in iteroparous animals. Am Nat 123:212-229 Cockburn A, Dalziell AH, Blackmore CJ, Double MC, Kokko H, Osmond HL, Beck NR, Head ML, Wells K (2009) Superb fairy-wren males aggregate into hidden leks to solicit extragroup fertilizations before dawn. Behav Ecol 20:501-510 Cohas A, Allainé D (2009) Social structure influences extra-pair paternity in socially monogamous mammals. Biol Lett 5:313-316 Colwell MA, Oring LW (1989) Extra-pair mating in the spotted sandpiper: a female mate acquisition tactic. Anim Behav 38:675-684 Cox CR, Le Boeuf BJ (1977) Female incitation of male competition: a mechanism in sexual selection. Am Nat 111:317-335 Crow JF (1958) Some possibilities for measuring selection intensities in man. Human Biol 30:1-13 Dale J, Dunn PO, Figuerola J, Lislevand T, Székely T, Whittingham LA (2007) Sexual selection explains Rensch’s rule of allometry for sexual size dimorphism. Proc R Soc Lond B 274:2971-2979 Darwin C (1871) The Descent of Man and Selection in Relation to Sex. Prometheus Books, New York, reprinted 1998 Delhey K, Johnsen A, Peters A, Andersson S, Kempenaers B (2003) Paternity analysis reveals opposing selection pressures on crown coloration in the blue tit (Parus caeruleus). Proc R Soc Lond B 270:2057-2063 Denk AG, Holzmann A, Peters A, Vermeirssen ELM, Kempenaers B (2005) Paternity in mallards: effects of sperm quality and female sperm selection for inbreeding avoidance. Behav Ecol 16:825-833 Dietrich V, Schmoll T, Winkel W, Epplen JT, Lubjuhn T (2004) Pair identity – an important factor concerning variation in extra-pair paternity in the coal tit (Parus ater). Behaviour 141:817-835 Dixon A, Ross D, O’Malley SLC, Burke T (1994) Paternal investment inversely related to degree of extra-pair paternity in the reed bunting. Nature 371:698700 Dolan AC, Murphy MT, Redmond LJ, Sexton K, Duffield D (2007) Extrapair paternity and the opportunity for sexual selection in a socially monogamous passerine. Behav Ecol 18:985-993 Double MC, Cockburn A (2000) Pre-dawn infidelity: females control extra-pair mating in superb fairy-wrens. Proc R Soc Lond B 267:465-470 Double MC, Cockburn A (2003) Subordinate superb fairy-wrens (Malurus cyaneus) parasitize the reproductive success of attractive dominant males. Proc R Soc Lond B 270:379-384

Extra-pair behaviour

403

Downhower JF, Blumer LS, Brown L (1987) Opportunity for selection: an appropriate measure for evaluating variation in the potential for selection? Evolution 41:1395-1400 Dreiss AN, Silva N, Richard M, Moyen F, Théry M, Møller AP, Danchin E (2008) Condition-dependent genetic benefits of extrapair fertilization in female blue tits Cyanistes caeruleus. J Evol Biol 21:1814-1822 Dunn PO, Whittingham LA, Pitcher TE (2001) Mating systems, sperm competition, and the evolution of sexual dimorphism in birds. Evolution 55:161-175 Dunn PO, Lifjeld JT, Whittingham LA (2009) Multiple paternity and offspring quality in tree swallows. Behav Ecol Sociobiol 63:911-922 Eberhard WG (2009) Postcopulatory sexual selection: Darwin’s omission and its consequences. Proc Natl Acad Sci USA 106:10025-10032 Eimes JA, Parker PG, Brown JL, Brown ER (2005) Extrapair fertilization and genetic similarity of social mates in the Mexican jay. Behav Ecol 16:456-460 Eliassen S, Kokko H (2008) Current analyses do not resolve whether extra-pair paternity is male or female driven. Behav Ecol Sociobiol 62:1795-1804 Fairbairn DJ, Wilby AE (2001) Inequality of opportunity: measuring the potential for sexual selection. Evol Ecol Res 3:667-686 Firman RC, Simmons LW (2008) Polyandry, sperm competition, and reproductive success in mice. Behav Ecol 19:695-702 Fisher RA (1930) The Genetical Theory of Natural Selection. Dover Publications, New York, reprinted 1958 Fisher DO, Double MC, Blomberg SP, Jennions MD, Cockburn A (2006) Postmating sexual selection increases lifetime fitness of polyandrous females in the wild. Nature 444:89-92 Foerster K, Delhey K, Johnsen A, Lifjeld JT, Kempenaers B (2003) Females increase offspring heterozygosity and fitness through extra-pair matings. Nature 425:714-717 Foerster K, Valcu M, Johnsen A, Kempenaers B (2006) A spatial genetic structure and effects of relatedness on mate choice in a wild bird population. Mol Ecol 15:4555-4567 Forstmeier W (2007) Do individual females differ intrinsically in their propensity to engage in extra-pair copulations? PLoS ONE 2:e952, doi:10.1371/journal.pone.0000952 Forstmeier W, Kempenaers B, Meyer A, Leisler B (2002) A novel song parameter correlates with extra-pair paternity and reflects male longevity. Proc R Soc Lond B 269:1479-1485 Fossøy F, Johnsen A, Lifjeld JT (2006) Evidence of obligate female promiscuity in a socially monogamous passerine. Behav Ecol Sociobiol 60:255-259 Fossøy F, Johnsen A, Lifjeld JT (2008) Multiple genetic benefits of female promiscuity in a socially monogamous passerine. Evolution 62:145-156 Freeman-Gallant CR (1996) DNA fingerprinting reveals female preference for male parental care in savannah sparrows. Proc R Soc Lond B 263:157-160

404

Bart Kempenaers and Emmi Schlicht

Freeman-Gallant CR, Wheelwright NT, Meiklejohn KE, States SL, Sollecito SV (2005) Little effect of extrapair paternity on the opportunity for sexual selection in savannah sparrows (Passerculus sandwichensis). Evolution 59:422430 Friedl TWP, Klump GM (2005) Extrapair fertilizations in red bishops (Euplectes orix): do females follow conditional extrapair strategies? Auk 122:57-70 García-Navas V, Ortego J, Sanz JJ (2009) Heterozygosity-based assortative mating in blue tits (Cyanistes caeruleus): implications for the evolution of mate choice. Proc R Soc Lond B 276:2931-2940 Garvin JC, Abroe B, Pedersen MC, Dunn PO, Whittingham LA (2006) Immune response of nestling warblers varies with extra-pair paternity and temperature. Mol Ecol 15:3833-3840 Gibbs HL, Weatherhead PJ, Boag PT, White BN, Tabak LM, Hoysak DJ (1990) Realized reproductive success of polygynous red-winged blackbirds revealed by DNA markers. Science 250:1394-1397 Gillespie JH (1977) Natural selection for variances in offspring numbers: a new evolutionary principle. Am Nat 111:1010-1014 Gowaty PA (1996) Battles of the sexes and origins of monogamy. In: Black J (ed) Partnerships in Birds: The Study of Monogamy. Oxford University Press, Oxford, pp 21-52 Gowaty PA, Buschhaus N (1998) Ultimate causation of aggressive and forced copulation in birds: female resistance, the CODE hypothesis, and social monogamy. Am Zool 38:207-225 Grafen A (1980) Opportunity cost, benefit and degree of relatedness. Anim Behav 28:967-968 Gray EM (1997a) Female red-winged blackbirds accrue material benefits from copulating with extra-pair males. Anim Behav 53:625-639 Gray EM (1997b) Do female red-winged blackbirds benefit genetically from seeking extra-pair copulations? Anim Behav 53:605-623 Griffith SC (2000) High fidelity on islands: a comparative study of extrapair paternity in passerine birds. Behav Ecol 11:265-273 Griffith SC (2007) The evolution of infidelity in socially monogamous passerines: neglected components of direct and indirect selection. Am Nat 169:274-281 Griffith SC, Owens IPF, Thuman KA (2002) Extra pair paternity in birds: a review of interspecific variation and adaptive function. Mol Ecol 11:2195-2212 Griffith SC, Lyon BE, Montgomerie R (2004) Quasi-parasitism in birds. Behav Ecol Sociobiol 56:191-200 Halliday T, Arnold SJ (1987) Multiple mating by females: a perspective from quantitative genetics. Anim Behav 35:939-941 Hasselquist D, Bensch S, von Schantz T (1996) Correlation between male song repertoire, extra-pair paternity and offspring survival in the great reed warbler. Nature 381:229-232 Haydock J, Koenig WD, Stanback MT (2001) Shared parentage and incest avoidance in the cooperatively breeding acorn woodpecker. Mol Ecol 10:15151525

Extra-pair behaviour

405

Heg D, Ens BJ, Burke T, Jenkins L, Kruijt JP (1993) Why does the typically monogamous oystercatcher (Haematopus ostralegus) engage in extra-pair copulations? Behaviour 126:247-289 Hoi-Leitner M, Hoi H, Romero-Pujante M, Valera F (1999) Female extra-pair behaviour and environmental quality in the serin (Serinus serinus): a test of the ‘constrained female hypothesis’. Proc R Soc Lond B 266:1021-1026 Holen ØH, Johnstone RA (2007) Parental investment with a superior alien in the brood. J Evol Biol 20:2165-2172 Hopper KR, Rosenheim JA, Prout T, Oppenheim SJ (2003) Within-generation bet hedging: a seductive explanation? Oikos 101:219-222 Houston AI (1995) Parental effort and paternity. Anim Behav 50:1635-1644 Houston AI, McNamara JM (2002) A self-consistent approach to paternity and parental effort. Philos Trans R Soc Lond B 357:351-362 Hsu Y-C, Li S-H, Lin Y-S, Philippart MT, Severinghaus LL (2006) High frequency of extra-pair copulation with low level of extra-pair fertilization in the Lanyu scops owl Otus elegans botelensis. J Avian Biol 37:36-40 Hunter FM, Davis LS (1998) Female Adélie penguins acquire nest material from extrapair males after engaging in extrapair copulations. Auk 115:526-528 Hunter FM, Burke T, Watts SE (1992) Frequent copulation as a method of paternity assurance in the northern fulmar. Anim Behav 44:149-156 Iwasa Y, Harada Y (1998) Female mate preference to maximize paternal care. II. Female competition leads to monogamy. Am Nat 151:367-382 Jennions MD, Petrie M (2000) Why do females mate multiply? A review of the genetic benefits. Biol Rev 75:21-64 Jeschke JM, Kokko H (2008) Mortality and other determinants of bird divorce rate. Behav Ecol Sociobiol 63:1-9 Johnsen A, Lifjeld JT (2003) Ecological constraints on extra-pair paternity in the bluethroat. Oecologia 136:476-483 Johnstone RA (1997) Recognition and the evolution of distinctive signatures: when does it pay to reveal identity? Proc R Soc Lond B 264:1547-1553 Jones AG, Arguello JR, Arnold SJ (2002) Validation of Bateman’s principles: a genetic study of sexual selection and mating patterns in the rough-skinned newt. Proc R Soc Lond B 269:2533-2539 Jones AG, Arguello JR, Arnold SJ (2004) Molecular parentage analysis in experimental newt populations: the response of mating system measures to variation in the operational sex ratio. Am Nat 164:444-456 Jones AG, Rosenqvist G, Berglund A, Avise JC (2005) The measurement of sexual selection using Bateman’s principles: an experimental test in the sex-rolereversed pipefish Syngnathus typhle. Int Comp Biol 45:874-884 Kawano KM, Yamaguchi N, Kasuya E, Yahara T (2009) Extra-pair mate choice in the female great tit Parus major: good males or compatible males. J Ethol 27:349-359 Kempenaers B (1993) A case of polyandry in the blue tit: female extra-pair behaviour results in extra male help. Ornis Scand 24:246-249 Kempenaers B (1997) Does reproductive synchrony limit male opportunities or enhance female choice for extra-pair paternity? Behaviour 134:551-562

406

Bart Kempenaers and Emmi Schlicht

Kempenaers B (2007) Mate choice and genetic quality: a review of the heterozygosity theory. Adv Stud Behav 37:189-278 Kempenaers B, Sheldon BC (1996) Why do male birds not discriminate between their own and extra-pair offspring? Anim Behav 51:1165-1173 Kempenaers B, Sheldon BC (1997) Studying paternity and paternal care: pitfalls and problems. Anim Behav 53:423-427 Kempenaers B, Verheyen GR, van den Broeck M, Burke T, van Broeckhoven C, Dhondt AA (1992) Extra-pair paternity results from female preference for high-quality males in the blue tit. Nature 357:494-496 Kempenaers B, Adriaensen F, van Noordwijk AJ, Dhondt AA (1996) Genetic similarity inbreeding and hatching failure in blue tits: are unhatched eggs infertile? Proc R Soc Lond B 263:179-185 Kempenaers B, Congdon B, Boag P, Robertson RJ (1999) Extrapair paternity and egg hatchability in tree swallows: evidence for the genetic compatibility hypothesis? Behav Ecol 10:304-311 Kempenaers B, Everding S, Bishop C, Boag P, Robertson RJ (2001) Extra-pair paternity and the reproductive role of male floaters in the tree swallow (Tachycineta bicolor). Behav Ecol Sociobiol 49:251-259 Kilner RM, Hinde CA (2008) Information warfare and parent-offspring conflict. Adv Stud Behav 38:283-336 Kingma SA, Hall ML, Segelbacher G, Peters A (2009) Radical loss of an extreme extra-pair mating system. BMC Ecol 9:15 Kleven O, Jacobsen F, Izadnegahdar R, Robertson RJ, Lifjeld JT (2006) Male tail streamer length predicts fertilization success in the North American barn swallow (Hirundo rustica erythrogaster). Behav Ecol Sociobiol 59:412-418 Kokko H, Jennions MD (2008) Parental investment, sexual selection and sex ratios. J Evol Biol 21:919-948 Kokko H, Mackenzie A, Reynolds JD, Lindström J, Sutherland WJ (1999) Measures of inequality are not equal. Am Nat 154:358-382 Krakauer AH (2008) Sexual selection and the genetic mating system of wild turkeys. Condor 110:1-12 Krokene C, Rigstad K, Dale M, Lifjeld JT (1998) The function of extrapair paternity in blue tits and great tits: good genes or fertility insurance? Behav Ecol 9:649-656 Lande R (1979) Quantitative genetic analysis of multivariate evolution applied to brain: body size allometry. Evolution 33:402-416 Levitan DR (2008) Gamete traits influence the variance in reproductive success, the intensity of sexual selection, and the outcome of sexual conflict among congeneric sea urchins. Evolution 62:1305-1316 Lifjeld JT (1994) Do female house sparrows copulate with extra-pair mates to enhance their fertility? J Avian Biol 25:75-76 Lindstedt ER, Oh KP, Badyaev AV (2007) Ecological social and genetic contingency of extrapair behavior in a socially monogamous bird. J Avian Biol 38:214-223 Lombardo R, Christy JH, Rissanen R (2004) Social monogamy in a simultaneous hermaphrodite. Integr Comp Biol 44:594

Extra-pair behaviour

407

Lorenzi MC, Sella G (2000) Is individual recognition involved in the maintenance of pairbonds in Ophryotrocha diadema (Dorvilleidae Polychaeta)? Ethol Ecol Evol 12:197-202 Lubjuhn T (2005) Consequences of being unfaithful – costs and benefits of extra pair copulations in birds. Vogelwarte 43:3-13 MacDougall Shackleton E, Clinchy M, Zanette L, Neff B, Wingfield JC, Boonstra R (2006) Environmental quality physiological stress and female mating strategies in song sparrows. J Ornithol 147:11-12 Magrath MJL, Komdeur J (2003) Is male care compromised by additional mating opportunity? Trends Ecol Evol 18:424-430 Magrath MJL, Vedder O, van der Velde M, Komdeur J (2009) Maternal effects contribute to the superior performance of extra-pair offspring. Curr Biol 19:792-797 Mauck RA, Marschall EA, Parker PG (1999) Adult survival and imperfect assessment of parentage: effects on male parenting decisions. Am Nat 154:99109 Maynard Smith J (1978) The Evolution of Sex. Cambridge University Press, Cambridge McKinney F, Derrickson SR, Mineau P (1983) Forced copulation in waterfowl. Behaviour 86:250-294 Mee A, Whitfield DP, Thompson DBA, Burke T (2004) Extrapair paternity in the common sandpiper Actitis hypoleucos, revealed by DNA fingerprinting. Anim Behav 67:333-342 Michl G, Török J, Griffith SC, Sheldon BC (2002) Experimental analysis of sperm competition mechanisms in a wild bird population. Proc Natl Acad Sci USA 99:5466-5470 Mills SC, Grapputo A, Koskela E, Mappes T (2007) Quantitative measure of sexual selection with respect to the operational sex ratio: a comparison of selection indices. Proc R Soc Lond B 274:143-150 Mock DW (1985) An introduction to the neglected mating system. Ornithol Monogr 37:1-10 Mock DW, Fujioka M (1990) Monogamy and long-term pair bonding in vertebrates. Trends Ecol Evol 5:39-43 Møller AP (2000) Male parental care, female reproductive success, and extrapair paternity. Behav Ecol 11:161-168 Møller AP, Alatalo RV (1999) Good-genes effects in sexual selection. Proc R Soc Lond B 266:85-91 Møller AP, Birkhead TR (1994) The evolution of plumage brightness in birds is related to extrapair paternity. Evolution 48:1089-1100 Møller AP, Briskie JV (1995) Extra-pair paternity, sperm competition and the evolution of testis size in birds. Behav Ecol Sociobiol 36:357-365 Møller AP, Cuervo JJ (2000) The evolution of paternity and paternal care in birds. Behav Ecol 11:472-485 Møller AP, Ninni P (1998) Sperm competition and sexual selection: a metaanalysis of paternity studies of birds. Behav Ecol Sociobiol 43:345-358

408

Bart Kempenaers and Emmi Schlicht

Møller AP, Thornhill R (1998) Male parental care, differential parental investment by females and sexual selection. Anim Behav 55:1507-1515 Morrow EH, Arnqvist G, Pitcher TE (2002) The evolution of infertility: does hatching rate in birds coevolve with female polyandry? J Evol Biol 15:702709 Mulder RA, Dunn PO, Cockburn A, Lazenby-Cohen KA, Howell MJ (1994) Helpers liberate female fairy-wrens from constraints on extra-pair mate choice. Proc R Soc Lond B 255:223-229 Neff BD (2003) Decisions about parental care in response to perceived paternity. Nature 422:716-719 Neff BD, Pitcher TE (2005) Genetic quality and sexual selection: an integrated framework for good genes and compatible genes. Mol Ecol 14:19-38 Neff BD, Sherman PW (2003) Nestling recognition via direct cues by parental male bluegill sunfish (Lepomis macrochirus). Anim Cogn 6:87-92 Neff BD, Sherman PW (2005) In vitro fertilization reveals offspring recognition via self-referencing in a fish with paternal care and cuckoldry. Ethology 111:425-438 O’Brien EL, Dawson RD (2007) Context-dependent genetic benefits of extra-pair mate choice in a socially monogamous passerine. Behav Ecol Sociobiol 61:775-782 Ockendon N, Griffith SC, Burke T (2009) Extrapair paternity in an insular population of house sparrows after the experimental introduction of individuals from the mainland. Behav Ecol 20:305-312 O’Donald P (1970) Change of fitness by selection for a quantitative character. Theor Popul Biol 1:219-232 Oring LW, Fleischer RC, Reed JM, Marsden KE (1992) Cuckoldry through stored sperm in the sequentially polyandrous spotted sandpiper. Nature 359:631-633 Owens IPF, Hartley IR (1998) Sexual dimorphism in birds: why are there so many different forms of dimorphism? Proc R Soc Lond B 265:397-407 Pagel M (1997) Desperately concealing father: a theory of parent-infant resemblance. Anim Behav 53:973-981 Petrie M, Kempenaers B (1998) Extra-pair paternity in birds: explaining variation between species and populations. Trends Ecol Evol 13:52-58 Petrie M, Doums C, Møller AP (1998) The degree of extra-pair paternity increases with genetic variability. Proc Natl Acad Sci USA 95:9390-9395 Philippi T, Seger J (1989) Hedging one’s evolutionary bets, revisited. Trends Ecol Evol 4:41-44 Pizzari T, Birkhead TR (2000) Female feral fowl eject sperm of subdominant males. Nature 405:787-789 Pizzari T, Løvlie H, Cornwallis CK (2004) Sex-specific, counteracting responses to inbreeding in a bird. Proc R Soc Lond B 271:2115-2121 Pöysä H, Pesonen M (2007) Nest predation and the evolution of conspecific brood parasitism: from risk spreading to risk assessment. Am Nat 169:94-104 Price T (1998) Maternal and paternal effects in birds. In: Mousseau TA, Fox CW (eds) Maternal Effects as Adaptations. Oxford University Press, Oxford, pp 202-226

Extra-pair behaviour

409

Queller DC (1997) Why do females care more than males? Proc R Soc Lond B 264:1555-1557 Rice WR, Holland B (1997) The enemies within: intergenomic conflict, interlocus contest evolution (ICE), and the intraspecific Red Queen. Behav Ecol Sociobiol 41:1-10 Rosivall B, Szöllõsi E, Hasselquist D, Török J (2009) Effects of extrapair paternity and sex on nestling growth and condition in the collared flycatcher, Ficedula albicollis. Anim Behav 77:611-617 Rowe KMC, Weatherhead PJ (2007) Social and ecological factors affecting paternity allocation in American robins with overlapping broods. Behav Ecol Sociobiol 61:1283-1291 Royle NJ, Hartley IR, Owens IPF, Parker GA (1999) Sibling competition and the evolution of growth rates in birds. Proc R Soc Lond B 266:923-932 Ruzzante DE, Hamilton DC, Kramer DL, Grant JWA (1996) Scaling of the variance and the quantification of resource monopolization. Behav Ecol 7:199207 Saino N, Bertacche V, Ferrari RP, Martinelli R, Møller AP, Stradi R (2002) Carotenoid concentration in barn swallow eggs is influenced by laying order, maternal infection and paternal ornamentation. Proc R Soc Lond B 269:17291733 Schmoll T, Dietrich V, Winkel W, Epplen JT, Schurr F, Lubjuhn T (2005) Paternal genetic effects on offspring fitness are context dependent within the extrapair mating system of a socially monogamous passerine. Evolution 59:645657 Schmoll T, Mund V, Dietrich-Bischoff V, Winkel W, Lubjuhn T (2007a) Male age predicts extrapair and total fertilization success in the socially monogamous coal tit. Behav Ecol 18:1073-1081 Schmoll T, Schurr FM, Winkel W, Epplen JT, Lubjuhn T (2007b) Polyandry in coal tits Parus ater: fitness consequences of putting eggs into multiple genetic baskets. J Evol Biol 20:1115-1125 Sefc KM, Mattersdorfer K, Sturmbauer C, Koblmüller S (2008) High frequency of multiple paternity in broods of a socially monogamous cichlid fish with biparental nest defence. Mol Ecol 17:2531-2543 Sheldon BC (1994) Male phenotype, fertility, and the pursuit of extra-pair copulations by female birds. Proc R Soc Lond B 257:25-30 Sheldon BC (2002) Relating paternity to paternal care. Philos Trans R Soc Lond B 357:341-350 Snook RR (2001) Sexual selection: conflict, kindness and chicanery. Curr Biol 11:R337-R341 Sozou PD, Houston AI (1994) Parental effort in a mating system involving two males and two females. J Theor Biol 171:251-266 Stewart IRK, Hanschu RD, Burke T, Westneat DF (2006) Tests of ecological phenotypic, and genetic correlates of extra-pair paternity in the house sparrow. Condor 108:399-413 Stutchbury BJ, Morton ES (1995) The effect of breeding synchrony on extra-pair mating systems in songbirds. Behaviour 132:675-690

410

Bart Kempenaers and Emmi Schlicht

Suter SM, Bielańska J, Röthlin-Spillmann S, Strambini L, Meyer DR (2009) The cost of infidelity to female reed buntings. Behav Ecol 20:601-608 Sutherland WJ (1985a) Measures of sexual selection. Oxf Surv Evol Biol 2:90101 Sutherland WJ (1985b) Chance can produce a sex difference in variance in mating success and explain Bateman’s data. Anim Behav 33:1349-1352 Thomson DL, Monaghan P, Furness RW (1998) The demands of incubation and avian clutch size. Biol Rev 73:293-304 Tregenza T, Wedell N (1998) Benefits of multiple mates in the cricket Gryllus bimaculatus. Evolution 52:1726-1730 Trivers RL (1972) Parental investment and sexual selection. In: Campbell B (ed) Sexual Selection and the Descent of Man, 1871-1971. Aldine, Chicago, pp 136-179 Tryjanowski P, Hromada M (2005) Do males of the great grey shrike Lanius excubitor, trade food for extrapair copulations? Anim Behav 69:529-533 Tullberg BS, Ah-King M, Temrin H (2002) Phylogenetic reconstruction of parental-care systems in the ancestors of birds. Philos Trans R Soc Lond B 357:251-257 Tuttle EM, Pruett-Jones S, Webster MS (1996) Cloacal protuberances and extreme sperm production in Australian fairy-wrens. Proc R Soc Lond B 263:13591364 Václav R, Hoi H, Blomqvist D (2003) Food supplementation affects extrapair paternity in house sparrows (Passer domesticus). Behav Ecol 14:730-735 Valera F, Hoi H, Krištín A (2003) Male shrikes punish unfaithful females. Behav Ecol 14:403-408 van Dongen WFD, Mulder RA (2009) Multiple ornamentation, female breeding synchrony, and extra-pair mating success of golden whistlers (Pachycephala pectoralis). J Ornithol 150:607-620 van Oers K, Drent PJ, Dingemanse NJ, Kempenaers B (2008) Personality is associated with extrapair paternity in great tits, Parus major. Anim Behav 76:555563 Wagner RH (1992) The pursuit of extra-pair copulations by monogamous female razorbills: how do females benefit? Behav Ecol Sociobiol 29:455-464 Weatherhead PJ, Boag PT (1997) Genetic estimates of annual and lifetime reproductive success in male red-winged blackbirds. Ecology 78:884-896 Webster MS, Pruett-Jones S, Westneat DF, Arnold SJ (1995) Measuring the effects of pairing success, extra-pair copulations and mate quality on the opportunity for sexual selection. Evolution 49:1147-1157 Webster MS, Chuang-Dobbs HC, Holmes RT (2001) Microsatellite identification of extrapair sires in a socially monogamous warbler. Behav Ecol 12:439-446 Webster MS, Tarvin KA, Tuttle EM, Pruett-Jones S (2007) Promiscuity drives sexual selection in a socially monogamous bird. Evolution 61:2205-2211 Wedell N, Kvarnemo C, Lessells CM, Tregenza T (2006) Sexual conflict and life histories. Anim Behav 71:999-1011 Westneat DF, Sherman PW (1993) Parentage and the evolution of parental behavior. Behav Ecol 4:66-77

Extra-pair behaviour

411

Westneat DF, Stewart IRK (2003) Extra-pair paternity in birds: causes correlates and conflict. Annu Rev Ecol Evol Syst 34:365-396 Westneat DF, Sherman PW, Morton ML (1990) The ecology and evolution of extra-pair copulations in birds. Curr Ornithol 7:331-369 Westneat DF, Clark AB, Rambo KC (1995) Within-brood patterns of paternity and paternal behavior in red-winged blackbirds. Behav Ecol Sociobiol 37:349-356 Wetton JH, Parkin DT (1991) An association between fertility and cuckoldry in the house sparrow, Passer domesticus. Proc R Soc Lond B 245:227-233 While GM, Uller T, Wapstra E (2009a) Family conflict and the evolution of sociality in reptiles. Behav Ecol 20:245-250 While GM, Sinn DL, Wapstra E (2009b) Female aggression predicts mode of paternity acquisition in a social lizard. Proc R Soc Lond B 276:2021-2029 Whittingham LA, Dunn PO (2001) Male parental care and paternity in birds. Curr Ornithol 16:257-298 Whittingham LA, Dunn PO (2005) Effects of extra-pair and within-pair reproductive success on the opportunity for selection in birds. Behav Ecol 16:138-144 Whittingham LA, Taylor PD, Robertson RJ (1992) Confidence of paternity and male parental care. Am Nat 139:1115-1125 Whittingham LA, Dunn PO, Stapleton MK (2006) Repeatability of extra-pair mating in tree swallows. Mol Ecol 15:841-849 Wink M, Dyrcz A (1999) Mating systems in birds: a review of molecular studies. Acta Ornithol 34:91-109 Woolfenden BE, Gibbs HL, Sealy SG (2002) High opportunity for sexual selection in both sexes of an obligate brood parasitic bird, the brown-headed cowbird (Molothrus ater). Behav Ecol Sociobiol 52:417-425 Wright J (1998) Paternity and paternal care. In: Birkhead TR, Møller AP (eds) Sperm Competition and Sexual Selection. Academic Press, London, pp 117145 Yasui Y (1998) The ‘genetic benefits’ of female multiple mating reconsidered. Trends Ecol Evol 13:246-250 Yasui Y (2001) Female multiple mating as a genetic bet-hedging strategy when mate choice criteria are unreliable. Ecol Res 16:605-616 Zeh JA, Zeh DW (1996) The evolution of polyandry. I. Intragenomic conflict and genetic incompatibility. Proc R Soc Lond B 263:1711-1717

Chapter 14

Extreme polyandry in social Hymenoptera: evolutionary causes and consequences for colony organisation F. BERNHARD KRAUS AND ROBIN F.A. MORITZ

ABSTRACT Polyandry, the multiple mating of females with more than one male, is a behavioural pattern found throughout the animal kingdom. In eusocial Hymenoptera (the bees, wasps and ants) polyandry is present in many taxa, and also the species with the highest degrees of polyandry known so far are found in this group. Polyandry has evolved multiple times and the genera with polyandrous species provide excellent test systems to study the evolution of multiple mating and its consequences for natural selection. Polyandrous colonies are composed of many paternal subfamilies (= patrilines), creating conflict potential on the one hand, and a template for variability among the colony members on the other hand.

14.1 Evolution of polyandry Because polyandry appears to be a maladaptive behaviour at first sight, it has gained the attention of many researchers in past decades (e.g. Zeh and Zeh 1996, 2001, Jennions and Petrie 2000). Polyandry is obviously a costly behaviour, since it requires time and energy on behalf of the female and it also poses the risk of predation or the transmission of venereal diseases. Moreover, in most species mating with a single mate provides a female with sufficient sperm to produce offspring. Thus, given the disadvantages of polyandrous behaviour, the question arises why females should be polyandrous at all? In addition to these disadvantages mentioned above, polyandry in social Hymenoptera is directly correlated with an effect at the level of the colony. The female offspring of a singly mated queen has a high average related-

414

F. Bernhard Kraus and Robin F.A. Moritz

ness of 75% among full sisters (= supersisters), which results from the haplo-diploid sex determination of the Hymenoptera. This close relatedness amongst the workers of a colony is one of the basic pillars for the evolution of sociality via kin selection in inclusive fitness theory (Hamilton 1964). Each additional mating of a queen reduces the average relatedness amongst the workers of a colony, hence distorting the indirect fitness benefits from male haploidy. Kin selection theory predicts that such a reduction of average relatedness is likely to trigger conflicts among the members of different patrilines. For example in the honeybee Apis mellifera, intracolonial selection takes place in emergency queen rearing (Moritz et al. 1996, 2005) and laying workers in queenless colonies engage in pheromonal competition for becoming a reproductive worker (Moritz et al. 2000). Whereas conflict is common in queenless colonies, it is the exception in queenright colonies. This is probably due to the fact that polyandry in social Hymenoptera is a derived trait and has only secondarily evolved in lineages with worker castes which already had lost their reproductive totipotency (Hughes et al. 2008a). In addition, polyandry and decreased intracolonial relatedness can also lead to a reduction of conflict potential. A well documented behaviour which is correlated with a low relatedness in social insect colonies is policing, where workers eat the unfertilised eggs of other workers to prevent selfish reproduction (Wenseleers and Ratnieks 2006). Polyandry also reduces the queen worker conflict over sex investment ratios, with both castes preferring an even resource allocation to either sex (Moritz 1985, Starr 1985). A large number of adaptive explanations, theories and hypotheses have been put forward to explain the evolution of this mating behaviour and its consequences for colony organisation. Some of these hypotheses consider the queen as the main target of natural selection, while other hypotheses more strongly focus on the effects and consequences of polyandrous matings at the colony level (Korb this volume). Empirical tests of these hypotheses are notoriously difficult and have only been conducted for a few selected species. So far, no single hypothesis convincingly succeeded as stand-alone explanation for the evolution of polyandry in social Hymenoptera. In fact, it seems likely that depending on the social organisation of a species, and on the specific environmental conditions, different selective pressures may act simultaneously at various levels of selection, which in the end concert the evolution of polyandry (Crozier and Fjerdingstad 2001, Kronauer et al. 2007). In the following, a selection of hypotheses will be presented, which so far have gained most support and appear to be the most plausible hypotheses to explain the evolution of polyandry. We here want to provide the reader with a basis of the theoretical and experimental work dealing with the behavioural phenomenon of polyandry in social

Extreme polyandry in social Hymenoptera

415

Hymenoptera with special emphasis on the honeybee (Apis mellifera), whose mating system is characterised by extreme polyandry. Thus, we purposely omit an overly detailed account of each and every possible evolutionary pathway to polyandry, a topic well covered in Crozier and Fjerdingstad (2001). 14.1.1 Hypotheses for the evolution of polyandry 14.1.1.1 Sperm competition hypothesis Starr (1985) suggested that multiple mating can evolve through postcopulatory sperm competition. For example, sperm of fitter males may move faster than others and are more likely to fertilise an egg than slower ones. If sperm fertilisation success is correlated with the quality of a male and the overall individual fitness of the resulting queen’s offspring, mating with multiple males would enhance the probability of a female to obtain offspring from high fitness males. This logic only functions for post-copulatory within-female selection among the sperms of the various males. Hence, the queen would produce higher fitness offspring as a result of post-copulatory competition amongst sperm. If sperm competition plays a role in social Hymenoptera, then species with extreme polyandry like the honeybees (genus Apis) or several army ants (e.g. Eciton, Dorylus) are expected to be most likely affected by sperm competition. However, even in a well studied species like A. mellifera, we are not aware of any study showing a correlation between sperm quality and male fitness. In contrast, studies addressing this issue empirically report on random sperm use of the honeybee queen, with no indications for sperm competition (Moritz 1986, Page 1986, Schlüns et al. 2004). Another version of the sperm competition hypothesis has focused on the ‘winning sperm’ concept, which is exclusively based on the correlation between male fitness and sperm competiveness, but not on the fitness of the females (Keller and Reeve 1995). Even though males in Hymenoptera have no sons, because males are parthenogenetically produced from unfertilised eggs, they have grandsons. The ‘winning sperm’ hypothesis predicts that any such trait that leads to higher fertilisation success will be transferred from the male to its grandsons and, hence, solely operates on male selection. Thus, by mating with multiple males, the queen would enhance the probability of having grandsons whose sperms are more competitive in fertilising eggs. Unfortunately also in the case of the winning sperm hypothesis empirical evidence is still lacking. This is not surprising, since it requires three physical generations to measure any effect transferred from

416

F. Bernhard Kraus and Robin F.A. Moritz

a male, to a queen, to an offspring queen, and finally to a grandson male. Not a particularly appealing task for most students of social insect biology! A more fundamental theoretical problem of the sperm competition theory stems from its assumption of genetic variance for sperm fitness. Any gene promoting sperm fitness should quickly go towards fixation in the population, hence terminating a selective advantage of the fittest sperm. The sperm competition concept is only an evolutionary stable strategy (ESS) if there is a trade-off for sperm fitness, causing stable equilibria with other fitness-relevant life history traits (Anderson and Simmons 2006). 14.1.1.2 Sperm limitation hypothesis The sperm limitation hypothesis (Cole 1983) might be considered as the most straightforward and intuitively also the most appealing explanation of polyandry in social Hymenoptera, especially in the cases of extreme polyandry (more than 10 matings), as is found in army ants and honeybees. In contrast to the sperm competition hypothesis, which is about sperm quality, it focuses on the importance of the sperm quantity transferred to the female. The basic concept holds that the number of sperm transferred by a single male to the queen is insufficient to produce tens of thousands, in some cases even millions of fertilised eggs over many years. Since the queens of large-colony species are often very long-lived (Keller and Genoud 1997), they have to mate with multiple males to obtain a sufficient amount of sperm. This hypothesis is supported by a significant positive correlation between colony size and the number of matings found in ant species (Cole 1983, Boomsma and Ratnieks 1996, Schmid-Hempel and Crozier 1999). Comparing various subspecies of the highly polyandrous honeybee the number of matings of a queen was negatively correlated with the amount of sperm per drone of a given subspecies (Kraus et al. 2004). Also in army ants, which establish the largest known monogynous colonies with up to 20 million workers (Dorylus: Gotwald 1995), queens are extremely polyandrous with up to 32 observed matings (Kronauer et al. 2004). Similar results were found for other army ants from the genus Eciton which have colony sizes up to 1 million workers and where the number of matings per queen ranged between 10 and 20 (Denny et al. 2004, Kronauer et al. 2006, Jaffé et al. 2007) In spite of these data, there are also various arguments against the relevance of sperm limitation for the evolution of polyandry. The most prominent one is that larger males, which are capable of transferring more sperm to the queen, should evolve because of a strong selective pressure towards more sperm per male (Crozier and Page 1985). However, male size and

Extreme polyandry in social Hymenoptera

417

sperm number will clearly have Bauplan limitations. Males need to fly in order to find queens. Moreover, optimal male size may also be driven by male-male competition rather than sufficient sperm transfer alone, resulting in male sizes optimising overall mating success rather than plain sperm number (Davidson 1982). For example there may be trade-offs for flight ability versus testes size and the size of the vesiculae seminales to store fertile sperm. Another argument against the sperm limitation hypothesis is that honeybee queens expel large amounts of sperm after insemination (Crozier and Fjerdingstad 2001). A single A. mellifera drone produces sufficient sperm to inseminate a queen, however 96% of the semen is expelled after each mating flight and the honeybee queen conducts repeated nuptial flight until she has obtained a sufficient sperm load (Schlüns et al. 2005). This generous and apparently inefficient sperm handling seems to contradict the sperm limitation hypothesis. However, large colony size species are derived from small colony size species, and there can be various evolutionary routes for queens to gain sufficient sperm. One route could be through evolving more efficient sperm storage mechanisms to manage the large semen quantities produced by the males. The other equally plausible evolutionary pathway could be through evolving a different mating behaviour, and maintaining the ancestral semen handling apparatus, which can only handle small amounts of semen, sufficient for a small but not for a large colony. Once an appropriate mating behaviour has evolved to allow for the collection of large semen volumes, it may be very difficult to modify the semen handling organs through evolutionary adaptations. After copulation the sperm of the mating drones is compressed into both lateral oviducts of the queen. These are balloon-like inflated with semen when the queens return from mating flights. When releasing the sperm from the lateral oviducts, they pass by the very narrow spermathecal canal, which poses a narrow bottleneck for sperm to enter the spermatheca. Clearly it will be very difficult at best for Breslau’s semen pump (Ruttner and Koeniger 1971) to transport all sperm from a single mating flight into the spermatheca. 14.1.1.3 Transfer of benefits hypothesis The idea behind this hypothesis is that females mate with multiple males to obtain or to accumulate fitness enhancing resources provided by the males for mating. Such benefits might either be transferred as nuptial gifts before mating to increase the male’s chances to mate with a given female (Vahed 1998), or they might be directly transferred with sperm during mating. However, so far no nuptial gifts have been observed in Hymenoptera. Moreover, there is no evidence for the transmission of nutritional acces-

418

F. Bernhard Kraus and Robin F.A. Moritz

sory substances with the sperm so far. In fact, quite the opposite has been found in the monandrous bumblebee, Bombus terrestris. Here the accessory material transferred with the sperm in form of a mating plug contains substances which reduce receptivity for further matings (Baer et al. 2001). Also the honeybee queen expels almost all of the accessory material provided by the drones. In sum, even though this hypothesis has not been explicitly falsified and the significance of trace compounds for queen fitness has not been excluded, it seems an unlikely explanation for the evolution of polyandry in social Hymenoptera in light of the available empirical data. 14.1.1.4 Sex locus load hypothesis The sex locus load hypothesis is linked to the haplo-diploid complimentary sex determination mechanism of the Hymenoptera. Individuals which are diploid and heterozygous at the sex-determining locus (csd) develop into females, whereas haploid individuals develop into males (Beye et al. 2003). The crucial point in terms of colony fitness is that individuals which are diploid but homozygous at the sex determining locus also develop into males which are sterile, dysfunctional or sometimes even nonviable (van Wilgenburg et al. 2006). Producing such dysfunctional males can pose a considerable genetic load for a colony and selection should favour any mechanism reducing the chance of their production. Depending on the number of functional csd alleles in a population, singly mated queens have a certain risk to fail completely in colony foundation when mating with a male which has the very same sex allele than she has. By mating with multiple males she ensures to have a sufficient number of alternative sex alleles, which allow for the production of viable worker offspring, even though with each additional mating she also increases the chance to encounter a male with the same sex allele. So, on average, the risk of producing a diploid male is the same irrespective of the mating strategy. Although, this mechanism is less important for independent colony founding species, it can be an important factor for perennial colonies with colony fission (Kronauer et al. 2007). At the colony level, multiple mating reduces variance among colonies. Every colony will have a similar, but very small genetic load, which is probably negligible for colony viability. Under single mating, the variance among colonies is high, with some colonies failing completely, whereas others are not different from those with the negligible load under polyandry. Hence, whereas all colonies have an equally high fitness under polyandry, some will have a dramatically reduced fitness under monandry. The selective advantage of polyandry under the sex locus load hypothesis lies in the fitness of the colonial phenotype.

Extreme polyandry in social Hymenoptera

419

14.1.1.5 Genetic compatibility hypothesis The basis for the genetic compatibility hypothesis is the assumption that the compatibility of the female’s genome with that of the male is highly variable from male to male. In principle, it is the more general version of the sex locus load hypothesis because it is based on many loci interacting with each other in a complex way. By mating with many males, the female increases the probability to mate with males with optimally matching genomes, allowing for the production of high fitness offspring (Zeh and Zeh 1996, 1997). Thus females mate with many males because they cannot identify the ‘optimal male’ before mating and lack pre-copulatory female choice. The queen just averages the genetic compatibility by mating multiply and mainly avoids the possibility to mate with a single low compatibility male. Natural selection should therefore favour queens that are able to choose compatible sperm by some post-copulatory selection mechanism to avoid non- or less viable offspring. So far, however, we are not aware of any evidence that queens can choose the sperm they utilise for fertilisation of their eggs in social insects. For honeybee queens it has been indirectly shown that such a sperm selection mechanism does not exist. The most detrimental combination of sperm and egg is when both sex alleles are identical (see above). Hence, if a queen could choose a particularly viable combination, we should see deviations from random sperm selection in fertilised eggs. Unfortunately for many beekeepers this does not happen (Woyke 1965). In brother – sister matings diploid drones occur exactly at the expected frequencies predicted by the Mendelian laws, which are based on random recombination. Since there is no sperm selection for this trait (viable or non-viable), it seems farfetched to assume sperm recognition mechanism in the queen or egg to enhance viability and fitness in other non-lethal traits. So, similar to the sex locus load, also here fitness advantages might result from preventing high intracolonial frequencies of a low compatibility male offspring. Support for this idea stems from a study by Schwander and Keller (2008), reporting that in the harvester ant Pogonomyrmex rugosus incompatibilities affect caste determination and only queens which mate multiply can ensure the production of both worker and queen castes. 14.1.1.6 Sex allocation manipulation hypothesis There is an inherent sex ratio conflict between a singly mated queen and her workers in social Hymenoptera colonies. Whereas the queens prefer a 1:1 investment ratio, workers prefer a 1:3 ratio (Trivers and Hare 1976). One way for the queen to overcome this conflict is to mate multiply and to

420

F. Bernhard Kraus and Robin F.A. Moritz

shift the optimal worker sex ratio closer to her own (Moritz 1985, Starr 1985). By mating multiply, the queen is in a position to actively reduce intracolonial relatedness independent of the workers (Fig. 14.1). This behaviour reduces the conflict potential for sex allocation between queen and workers in the colony, since the average relatedness of a worker to nestmate workers decreases with every additional mating of the queen. As a consequence, the investment sex ratio (if worker controlled) is expected to be more biased towards the production of males and thus towards the sex ratio optimal for the queen (Moritz 1985, Ratnieks and Boomsma 1997). In fact, worker policing, with workers preventing other workers from egg laying, becomes a highly adaptive trait under a polyandrous mating system (Ratnieks 1988). Worker policing also corroborates the reproductive dominant position of the queen, hence again resolving conflict in the colony. 14.1.1.7 Genetic polyethism hypothesis The genetic polyethism hypothesis implies that the increase of the withincolony genetic variability, caused by multiple mating of the queen, leads to a more efficient division of labour (Page et al. 1989). If the behaviour of the workers and their thresholds to perform a certain task have a genetic basis, the increased within-colony variability will inevitably lead to an increase in polyethism. That behaviour, and also caste determination, has a genetic basic and has indeed been shown in many species of social insects. In the honeybee, for example, patrilines differ in their thresholds to perform given tasks (Robinson and Page 1995) and in ant species like Acromyrmex leafcutter ants (Hughes et al. 2003) and Eciton army ants, there is a genetic basis for caste determination (Jaffé et al. 2007). The question, however, is whether this genetic polyethism or caste determination is an ultimate factor causing polyandry and leading to a fitness advantage of multiply mated colonies over single mated colonies, or rather a proximate phenomenon reflecting genetic diversity in the colony. 14.1.1.8 Increased parasite resistance hypothesis This hypothesis also focuses on the increased genetic diversity within the colony, suggesting that colonies headed by multiple mated queens show a higher resistance towards pathogens and parasites. This mechanism is not based on queens selecting specific males (Mays and Hill 2004), but the increased number of different genotypes present in a multiple mated colony is assumed to confer improved overall colony (‘herd’) immunity (SchmidHempel 1994). The concept is that some of the subfamiles in the colony

Extreme polyandry in social Hymenoptera

421

are resistant to specific pathogens. Even if a specific pathogen genotype is perfectly adapted to one specific host genotype, it will not be able to overcome the immune defences of other host genotypes in the colony as easily, resulting in improved colony resistance. For this hypothesis there is direct evidence from the bumblebee Bombus terrestris. Even though B. terrestris is a species with singly mated queens, Baer and Schmid-Hempel (1999) artificially inseminated queens with the sperm from several males and showed that indeed the resulting higher intra-colonial diversity was positively correlated with a higher resistance towards pathogens. However, B. terrestris has not evolved polyandry and thus it seems that the costs associated with polyandry exceed the benefits in this particular species. Recently, Mattila and Seeley (2007) showed that honeybee colonies with a multiply mated queen had an enhanced resistance against various diseases in comparison to singly mated queens. This was the first empirical evidence showing that a lack of intra-colonial genotypic variance did have an effect on the colony phenotype. Indeed resistance to pathogens and parasites is thought to be a major driver of evolution and genetic heterogeneity has been considered repeatedly to be an efficient mechanism providing group resistance. 14.1.2 Polygyny versus polyandry Three of the above described hypotheses are explicitly focused on the increased genetic diversity caused by the multiple mating of queens. They assume that this increased genetic diversity per se leads to positive effects at the colony level. Since a high intra-colonial genetic diversity is also present in polygynous species, one would expect that polyandry should be more frequent in monogynous species. Indeed, polyandrous behaviour has so far been mainly described for monogynous species and there is a negative relationship between polyandry and polygyny (Keller and Reeve 1994, Hughes et al. 2008b; but see Schmid-Hempel and Crozier 1999). Kronauer and Boomsma (2007) show that within the army ants a polygynous species was monandrous whereas the monogynous one was highly polyandrous. However, polygyny is the much more efficient mode to enhance genetic diversity in the colony and various highly polygynous ant species reach intra-colonial relatedness values close to zero (Helanterä et al. 2009) which is obviously impossible in monogynous colonies. This may actually weaken the arguments of those hypotheses that are exclusively aiming at genetic diversity as a major evolutionary driver of polyandry. Polygyny is a possible evolutionary pathway and obviously much more efficient in enhancing intracolonial genotypic variability. Including polygyny as a factor

422

F. Bernhard Kraus and Robin F.A. Moritz

for low intracolonial relatedness into the discussion of the evolution of low intracolonial relatedness does not resolve whether sperm limitation or genetic diversity is the prime driver of evolution. 14.1.3 Evolutionary pathways 14.1.3.1 Transitions from monandry to extreme polyandry Even though the above hypotheses provide a theoretical framework to explain the evolution from a single mating to a few matings (oligoandry), most of them fall short in explaining the evolution of extreme polyandry, like it is found in the honeybees and army ants. The main problem for all hypotheses which are based on an increased within-colony genetic variability due to multiple mating is that they assume a linear increase in fitness with each additional mating. However, the increase in genetic diversity, which is directly correlated with the decrease of average relatedness in the colony, is not linearly correlated with the number of matings (Fig. 14.1). While there is a sharp initial increase of intra-colonial genetic diversity from one to two matings, there is nearly no additional gain after the sixth mating. Thus, any hypothesis based on genetic variability fails to explain mating numbers above six or seven, unless we can plausibly explain extremely nonlinear fitness responses when further increasing the number of matings. Hence, the problem is not to explain the evolutionary transition from one to two matings (as in Ratnieks and Boomsma 1995). The problem is to explain the adaptive significance of a transition from say 25 to 26 matings (or 99 to 100 in the case of the giant Asian honeybee A. dorsata). One approach to overcome this problem is to assume that queens mate with an excess of males in a nonadaptive fashion (Tarpy and Page 2000). This, however, would imply zero mating costs for these additional matings that do not contribute to enhanced queen fitness. In the case of army ants, low costs of mating might indeed exist, since the queens mate in the safety of their colonies, thus avoiding mating risk by predation (Kronauer et al. 2007). However, predation is only one and probably only a minor risk. The risk of transmitting veneral diseases increases linearly with every additional mating, as has recently been shown for the highly polyandrous honeybee (Yue et al. 2007). Moreover, mating costs are often far from zero. Honeybee queens fly to drone congregation areas, and thus the mating flight costs are less correlated with the number of matings than with the number of mating flights. At least for queens under apicultural conditions the mating risk can be substantial and exceeds 15-30% if weather conditions are poor (Ruttner 1980, Moritz 1985). Moreover, Schlüns et al. (2005) showed that queens with an insufficient

Extreme polyandry in social Hymenoptera

423

Relatedness / genetic diversity

0.8

0.6 Relatedness Genetic diversity 0.4

0.2

1

4

7

10

13

Number of matings

Fig. 14.1 Correlation between the average degree of relatedness and the genetic diversity within a colony. After an initial sharp decline of the average relatedness from 0.75 within the first six matings, relatedness approaches the lower limit of 0.25.

number of matings and low sperm numbers in the spermatheca embarked on additional mating flights, hence substantially increasing the mating costs for additional matings. In conclusion, the assumption that matings come for free for queens of social Hymenoptera is not well supported in honeybees and doubtful in other social insect species. Only few of the genetic variance hypotheses remain plausible when taking the non-linear decrease of relatedness and the increasing costs for mating into account. Given that the costs of matings increase linearly, the benefits need to grow exponentially to compensate for the declining gains in genetic variance as the number of matings increase. Here, most of the above hypotheses fail because they do not include non-linear fitness benefits or insufficient steep fitness gains as a result of extreme polyandry. One attempt was made by Fuchs and Moritz (1999) who argued that ‘specialist’ workers need to be rare in a colony. If they become too frequent, there is a cost at the colony level, i.e., if there are too many plumbers on a construction site and insufficient masons there will be no building at the end of the day. The large multitude of males will automatically guarantee that all genotypes are rare and hence local intra-colonial processes can activate the specialists whenever needed. But then how many specialists are needed? Is genetic variance really essential, if intracolonial environmental variance alone can already provide a sufficiently wide spectrum of castes and specialists even in singly mated species as shown for army ants (Kronauer et al. 2007)? For the time being, we see limitations for the genetic variance

424

F. Bernhard Kraus and Robin F.A. Moritz

hypothesis to plausibly explain the evolution of extreme polyandry because they largely ignore the costs of mating. Depending on fitness functions to be verified, pathogen resistance, the sex allele model and the rare specialist model hypotheses remain plausible. 14.1.3.2 Back to square one? In light of the empirical evidence, we feel that the sperm limitation hypothesis has received insufficient appreciation in the field when discussing the evolution of extreme polyandry. The consequences of a queen depleting her sperm store are obviously most dramatic for the colonial phenotype. It does not require delicate, fine-tuned experiments to measure some barely significant fractions of fitness reduction to see that a colony with a queen without semen has a fitness disadvantage, compared to one with a queen with sufficient semen. In fact, we know a large suite of behavioural and physiological adaptations in many social insects to regulate the process of queen supersedure. This is generally a very delicate and fragile phase in the life history of a colony. Just arguing that there might be evolutionary pathways to overcome sperm limitation by enhancing the sperm produced by a male or changing the queen’s Bauplan so that it can handle more semen does not tell us anything about the actual evolutionary pathway. Evolution seeks solutions that work and not necessarily optimal designs envisaged by theory. Once a suboptimal solution reaches a sufficiently high fitness peak, it will be difficult to change highly complex biological units such as insect societies to optimal solutions requiring very radical changes in anatomy, behaviour and life history. Moreover we should keep in mind that with the multitude and diversity of social insects systems and the independently evolved polyandry in various social Hymenoptera, different evolutionary pathways may have led to extreme polyandry. It is clearly a highly derived trait, evolved in response to the requirements of large and highly complex societies.

14.2 The consequences of polyandry for colony organisation 14.2.1 Genetic variance as an intra-colonial template Irrespective of the multitude of evolutionary theories, models and concepts for the ultimate reasons of the extreme polyandry observed in social Hymenoptera, the consequences of polyandry for colony structure are very clear. Polyandry inevitably enhances intracolonial genetic diversity and re-

Extreme polyandry in social Hymenoptera

425

duces average intracolonial relatedness. This sets the stage for intracolonial conflict (see Heinze this volume) and reduces all relatedness based indirect fitness benefits resulting from male haploidy (see Korb this volume). The other inevitable consequence is an enhanced phenotypic variance among the colony members, which relates to all phenotypic traits which show genetic variance. Hence, many morphological traits, but also inescapably a wide range of behavioural traits are affected, unless we assume that genes cannot interfere with phenotypes and question behavioural genetics as a scientific discipline as such. Since we know that genes control behaviour in social insects (e.g. Linksvayer et al. 2009) and we do know that there is genetic variance for behaviour among individuals, it should be impossible not to find genetic variance for behaviour including sophisticated traits such as learning, task specialisation and reproduction within a group of several thousands of half sib individuals. 14.2.2 Genetic variance for reproduction The trait ‘reproduction’ has been a long standing research topic in social insect research because of the dramatic and fundamental differences between the reproductive female caste (queen) and the sterile caste (worker). Even when combining otherwise solitary individuals, hierarchies can be established in a self-organised fashion (Fewell and Page 1999). For a long time it seemed ‘obvious’ that genetic variance resulting from multiple mating of queens was not suitable to explain the diversity of form and function among the members of a colony. In fact, the equal chance for all genotypes in a colony to become the queen is a very fundamental assumption and the basis for inclusive fitness theory. As soon as there is a bias for a specific ‘royal’ allele, enhancing the probability to be selected as a queen, this should immediately go towards fixation in the population, again restoring equal chance for all. Since there is only a single queen in a monogynous colony, this seemed to rule out any genetic variance controlling the reproductive caste. Nevertheless, first claims about a genetic mechanism for caste determination came from the stingless bee Melipona quadrifasciata (Kerr 1950). A pair of two Mendelian segregating loci was supposed to be involved in queen determination. Kerr’s claim resulted in a long lasting controversy with the Darchens (Darchen and Delage-Darchen 1971, 1977), who advocated an exclusively trophic mode of queen determination in M. quadrifasciata. It became clear, however, that the picture was neither only nutrition nor only genetic determination, but, as so often in biological systems, a mixture of both (Campos et al. 1979). In subsequent decades the link be-

426

F. Bernhard Kraus and Robin F.A. Moritz

tween genotype and insect phenotype was found in a variety of social insects, including wasps, ants and bees (Calderone and Page 1988, 1991, Frumhoff and Baker 1988, Kolmes et al. 1989, Robinson and Page 1989, Stuart and Page 1991, Oldroyd et al. 1992, Snyder 1993, Keller and Ross 1995, Ross et al. 1996, O’Donnell 1998). With the rise of molecular ecology and the availability of microsatellite DNA analyses in social insects it became quickly clear that genetic variance might have much deeper consequences for colonial organisation. Microsatellite DNA analyses allow for precise paternity assignment, and, hence, it was possible to test which subfamily members were in which caste and engaged in which task in the colony. Today it seems clear that genotype-environment interactions are the major drivers for caste determination in social insects. The high intracolonial variability appears to be the key to regulate task organisation among the workers (Robinson and Page 1989, Bonabeau et al. 1996, Fewell and Page 2000). In particular the threshold response model (Page and Mitchell 1998; BOX 14.1) is most suited to explain how even subtle genetic differences among the workers of the various patrilines might result in large behavioural differences in the colony. 14.2.3 Caste determination and genetic variance Knowing genes that interfere with social behaviour is one aspect, but in the context of polyandry it is the variability among different genotypes which becomes important. How does genetic control of castes vary among workers of different subfamilies in the colony if it varies at all? Task specialisation among morphologically similar individuals may be seen as the result of behavioural genetic variance. In ants, however, task specialisation often coincides with morphological differentiation among workers. Distinct worker castes perform distinct tasks in the colony (Oster and Wilson 1978). Although the colony composition in such multi-worker-caste species looks qualitatively different compared to the single worker caste, the occurrence of multiple physical worker castes does not pose a conceptually new problem for the stimulus threshold model. It just changes from a behavioural level to a developmental level: from ethology to ontogeny. Larvae are fed based on the attention they receive from the feeding workers. Individuals who are fed large amounts of food should grow faster and larger then those receiving less food. As soon as there is genetic variance among larvae for attracting feeding workers, there will be variance for growth among those individuals and eventually differently sized adults will develop.

Extreme polyandry in social Hymenoptera

427

BOX 14.1 The threshold response model Task specialisation of honeybees under the threshold response model is based on individual worker decisions to participate in a certain task based on its individual response to an external stimulus releasing the behaviour to engage in the task. The effect such individual response thresholds can have on colony organisation are best illustrated by a hypothetical example. Let us assume we have a honeybee colony with six workers at the flight entrance that are intensively engaged in guarding. They will do this for a while until something attracts the attention of one of them which quits doing its task (i.e. it got hungry and went inside the colony to feed on honey before it got distracted by a dancing forager to be recruited to nectar foraging). As a consequence, there are only five workers left at the flight entrance to do the guarding. If workers now have an intrinsic threshold for the number of workers they want to see at the flight entrance guarding, any worker coming along which feels this number should be five or less is happy and goes by to do nothing or engage in another task. A worker who feels there should be six guarding workers will, however, join the group. If a worker came by which had a response threshold of seven workers to engage in guarding, she would also join the group. Let us now assume that there is genetic variance for this response threshold to join the guards. In that case workers from different patrilines are expected to have different thresholds for joining the guard force. The workers from the patriline with the highest threshold will eventually accumulate at the flight entrance. All other workers will rarely engage in guarding because there are always a sufficient number of guards at the flight entrance. We do not know if most of them are actually puzzled by the high density of guards and assume that their nest mates are simply wasting their time. But then bees are not humans. In any case, response thresholds are not constraint to social insects and are also an important driver of day to day family life when it comes to who is to switch on the vacuum cleaner in response to dust stimuli in the house, resulting in task specialisation of the family members and sometimes hot debates.

Since morphological worker castes are only known in ants, this has been exclusively studied in polyandrous ant species. Various studies on monogynous ant species with polyandrous queens have demonstrated a genetic effect on worker differentiation into either small or large workers (Hughes et al. 2003, Rheindt et al. 2005, Jaffé et al. 2007). The most extreme case is that of the army ant Eciton burchelli, where queens show extensive multiple mating with 10-25 males (Denny et al. 2004, Kronauer et al. 2006) and colonies have strongly differentiated morphological worker castes (Franks 1985). Jaffé et al. (2007) found that the four worker castes,

428

F. Bernhard Kraus and Robin F.A. Moritz

Head width [mm]

3

2

1

majors submajors medias minors 5

10

15

Patrilines

Fig. 14.2 Ranked worker head width variation among the most frequent patrilines of a E. burchellii colony (data from Jaffé et al. 2007). The overall caste representation among patrilines was significantly skewed (p < 0.01) among the four worker castes (minors, medias, submajors and majors). Mean values, standard errors and standard deviations are represented by horizontal bars, grey boxes and whiskers, respectively.

minors, medias, submajors and majors were significantly biased among the various patrilines (Fig. 14.2). Many patrilines produced exclusively individuals within two or three castes and did not contribute to others. 14.2.4 The honeybee as a model system Since the honeybee Apis mellifera was the first social insect for which microsatellite DNA loci were developed (Estoup et al. 1993, 1995) it is here where most information is available. An interesting property is the extreme degree of polyandry in the honeybee (Palmer and Oldroyd 2000). Multiple mating in A. mellifera can exceed 50 and in A. dorsata, the Giant Asian honeybee, even 100 drones per queen (Moritz et al. 1995, Palmer and Oldroyd 2000, Wattanachaiyingcharoen 2003), which makes the honeybee a superb system for research on the impact of polyandry on social behaviour. Although not all drones are equally represented in the sperm store of

Extreme polyandry in social Hymenoptera

429

a queen’s spermatheca, the effective number of males per queen is still much higher than in other social Hymenoptera. The sequence of mating behaviour is best described for the Western honeybee Apis mellifera. Virgin queens conduct multiple nuptial flights to so-called drone congregation areas (Alber et al. 1955). In flight they mate with a variable number of males representing a random sample of the drone population. Honeybee queens conduct repeated flights until they have mated with an adequate number of drones to obtain a sufficient spermathecal filling (Woyke 1964). It appears that the number of matings is an important signal used by the queen to initiate oviposition (Schlüns et al. 2005; but see Tarpy and Page 2000). Once the queen has commenced egg laying, she will not mate again, and successively depletes the semen stored in the spermatheca to fertilise her eggs destined to produce female offspring. Depending on the origin of the sperm, the resulting female will belong to one of the various subfamilies represented in the stored semen. In principle this mating strategy will generate a rather unpredictable genotypic composition of the colony. The unpredictability of sperm use is further enhanced because the sperm from each sire is not equally represented in the spermatheca, and the distribution of sperm in the spermatheca is not completely homogeneous (Haberl and Tautz 1998). The genotypic composition of the honeybee colony is therefore mainly characterised by two major factors: its complexity and its unpredictability within the limits set by the mating system. Stochastic dynamics will hence be an essential issue in colony composition, setting the stage for a large variance for any genetically determined behavioural trait which may potentially affect colony efficiency. 14.2.4.1 Gene effects and worker task specialisation A variety of experiments have shown that honeybee workers vary in their response to task stimuli. Page et al. (1998) tested the response of individual pollen and nectar foragers towards sucrose in a learning paradigm using the proboscis extension reflex. Bees expose their proboscis (= tongue) in a reflex if a droplet of sugar water touches their antennae (unconditional stimulus). If an odour is paired with the sugar reward, the bee will learn to associate odour with reward and expose its proboscis if only the odour is presented (conditional stimulus). Since not all bees show the same learning performace, this assay allows for the quantification of stimuli thresholds and revealed the connection between behaviour, stimulus and genotype (Page et al. 1998, Scheiner et al. 2001). With increasing knowledge of the genome, many genes have been identified that interfere with the control and regulation of social behaviour and

430

F. Bernhard Kraus and Robin F.A. Moritz

reproduction. The connection between reproduction and task specialisation of workers was proposed by West-Eberhard (1987, 1996) as the Reproductive Ground Plan Hypothesis (RGPH) in which worker reproductive anatomy and physiology is correlated with biases in non-reproductive social behaviour. The basic idea is that fertile solitary females, once they are mated, need proteins for egg production and subsequently also must provide their brood with even more protein. Hence the activation of ovaries and the brood care phase is expected to be associated with increased pollen foraging. Since vitellogenin is the major yolk protein and an indicator of fertility, Amdam and Omholt (2003) developed a model integrating this concept into the vitellogenin physiology in honeybees. Experimental work confirmed the concept and worker bees with more ovarioles and a bias toward pollen foraging had increased vitellogenin titres. This association between the reproductive physiology and the foraging behaviour seems to have persisted through the evolutionary process towards sociality, because their genetic bases are largely congruent. The effects of vitellogenin on foraging behaviour was verified by Vg knockdown (Nelson et al. 2007) and recently also the correlation with ovary size, as predicted by the RGPH, became evident. Wang et al. (2009) showed that the activity of two candidate genes for behaviour, PDK1 and HR46, have direct genetic relationships to ovary size. This shows how selection may have acted on gene networks that affect reproductive resource allocation and behaviour to facilitate the evolution of social foraging in honey bees. In the end, there is a highly sophisticated regulatory network governing honeybee phenotypic plasticity of reproduction and task specialisation (for a review see Page and Amdam 2007) evolved from the genetic tool box of its solitary ancestors. 14.2.4.2 Selection in the honeybee colony Whereas task specialisation for non-reproductive traits can be plausibly explained within a framework of inclusive fitness theory, it is much more difficult to give reasons for genetic variance of traits that determine reproductive caste. Nevertheless, a series of studies in honeybees and ants has shown that there is also considerable genetic variance among the subfamilies for the determination of the queen caste. Studies of honeybees revealed that after experimentally removing the old queen, emergency queens did not represent a random sample of the available brood. Surprisingly, the rare patrilines in the colony were overrepresented in the queen sample (Fig. 14.3; Moritz et al. 2005). Of course such a system is only evolutionary stable if there is a trade-off for becoming a queen. The empirical data suggested a possible trade-off

Extreme polyandry in social Hymenoptera

431

0.7

Worker Queen

Mean subfamily frequency

0.6 0.5 0.4 0.3 0.2 0.1 0

0 - 20

21 - 40

41 - 60

61 - 80

81 - 100

Subfamiliy quintiles

Fig.14.3 Frequencies of the 20th percentiles of subfamily frequencies of workers and queens from an A. mellifera colony (data from Moritz et al. 2005). Standard errors are indicated by whiskers. The subfamily distributions were significantly different among the queen and worker samples (p < 0.01).

between drone and queen fitness, i.e. offspring of drones poorly represented in the colony might become highly attractive in the queen selection phase. A simple population genetic model based on a single locus with two alleles, one making larvae attractive for queen rearing (r) and the other one not (+), showed that a wide parameter space allowed for stable equilibria between r and + in an otherwise classical Hardy-Weinberg population. The model is based on the fitness of a male (Wm) carrying the r allele is reduced by the selection coefficient s (Wm = 1 − s). Under the extreme assumption that only those female larvae with the largest number of r alleles (either one or two) are raised to queens, the equilibrium frequency can be derived as p = (1/s) − 1 (for 0 < p < 1)

(1)

which results in stable equilibria for 0.5 < s < 1. Other conditions for the rearing of queen larvae allow for many other possible equilibria conditions and expand the fitness range in which both the r and + allele can coexist. Alternatively, the trade-off for individual female reproductive success could be at the colony level. This is particularly relevant if worker repro-

432

F. Bernhard Kraus and Robin F.A. Moritz

duction is considered. Laying workers typically do not engage in other tasks but egg laying (for examples in ants, see also Heinze this volume and Korb this volume). A special case is the Cape honeybee, A. mellifera capensis, which is probably the best studied example of laying worker activity in insect colonies. Laying workers of this subspecies parthenogenetically produce female offspring (Onions 1912), which cause very specific relatedness patterns in the colony (Neumann and Moritz 2002). Laying worker offspring are almost clonally produced (Moritz and Haberl 1994) and hence have a relatedness of r ~ 1. Inclusive fitness theory predicts high conflict in these colonies and indeed there are many indications of physical and pheromonal competition between queens and laying workers (Moritz et al. 2003) and among laying workers (Simon et al. 2005). In case of the Cape honeybee, the reproductive worker trait should spread in the population under a wide parameter space (Moritz 1985, Greeff 1996). In particular, the risk of virgin queens on mating flights may be an important factor in natural selection because thelytokous laying workers are in principle able to replace the queen. The genetic mode of this behaviour could be pinned down to a single locus (thelytoky) located on chromosome 13 (Lattorff et al. 2005). Interestingly, this locus did not only control the mode of parthenogenesis, but also the activation of ovaries, and the queen pheromone production in the mandibular glands. Lattorff et al. (2007) could map the gene to a region of 11.4 cM including 15 open reading frames, two of which coded for transcription factors known to control many different traits at the same time. The control of the trait by a single major locus enhances the value of the population genetic models that were based on a single biallelic gene coding for worker reproduction. The single locus system of the Cape honeybee offers a simple model system to understand genetic control and selection of social behaviour and reproduction at the individual, the colony and the population level. If thelytoky affords laying workers such a large individual fitness what are the costs for thelytokous worker reproduction? What is the potential trade-off that causes laying thelytokous workers not to be the rule but rather the exception in social insects and remain confide to the small region around the Cape of Good Hope? The individual fitness benefits seem to be overwhelming. Laying workers can produce queens and workers without any relatedness loss. Colonies can requeen themselves from worker-laid offspring. The most likely trade-off for the rise of laying worker alleles in the population may be at the colony level. Hillesheim et al. (1989) could show that colonies composed of ‘selfish’ laying workers had an extremely reduced capacity for brood rearing. Combs with eggs introduced into the

Extreme polyandry in social Hymenoptera

433

colony are typically reared to the adult stage in honeybee colonies composed of regular ‘altruistic’ workers. Egg combs introduced into colonies with ‘selfish’ workers were not reared to adult bees. The workers removed all the eggs to replace them with their own eggs. These eggs were then again removed by other laying workers in the colony. It thus becomes clear that the colonial composition is essential for both individual and colony fitness.

14.3 Conclusion Even after more than two decades of research, the evolution of extreme polyandry in insect societies remains a controversially discussed phenomenon. This is not necessarily surprising because of the manifold evolutionary pathways to large colony units that arose independently in the various genera. Its consequences reach deep into the heart of colony organisation governing conflict and cooperation. The study of proximate and ultimate factors driving polyandry will therefore remain a rewarding research field in behavioural ecology. It is not difficult to foresee that with the next generation DNA technology we will be able to address the interactions among the members of a colony at the transcriptome level, eventually closing the gap between the gene in the individual worker and the complex phenotype of a colony. The emerging synergies between bioinformatics, genomics and classical behavioural ecology will eventually result in understanding of how a colony works: the foremost goal of any student of social insect biology.

References Alber M, Jordan M, Ruttner F, Ruttner H (1955) Von der Paarung der Honigbiene. Z Bienenforsch 3:1-28 Amdam GV, Omholt SW (2003) The hive bee to forager transition in honeybee colonies: the double repressor hypothesis. J Theor Biol 223:451-64 Andersson M, Simmons LW (2006) Sexual selection and mate choice. Trends Ecol Evol 21:296-302 Baer B, Schmid-Hempel P (1999) Experimental variation in polyandry affects parasite loads and fitness in a bumble-bee. Nature 397:151-154 Baer B, Morgan ED, Schmid-Hempel P (2001) A nonspecific fatty acid within the bumblebee mating plug prevents females from remating. Proc Natl Acad Sci USA 98:3926-3928

434

F. Bernhard Kraus and Robin F.A. Moritz

Beye M, Hasselmann M, Fondrk MK, Page RE Jr, Omholt SW (2003) The gene csd is the primary signal for sexual development in the honeybee and encodes an SR-type protein. Cell 114:419-429 Bonabeau E, Theraulaz G, Deneubourg J-L (1996) Quantitative study of the fixed threshold model for the regulation of division of labour in insect societies. Proc R Soc Lond B 263:1565-1569 Boomsma JJ, Ratnieks FLW (1996) Paternity in eusocial Hymenoptera. Philos Trans R Soc Lond B 351:947-975 Calderone NW, Page RE Jr (1988) Genotypic variability in age polyethism and task specialization in the honey bee, Apis mellifera (Hymenoptera: Apidae). Behav Ecol Sociobiol 22:17-25 Calderone NW, Page RE Jr (1991) Evolutionary genetics of division-of-labour in colonies of the honeybee (Apis mellifera). Am Nat 138:69-92 Campos LAO, Kerr WE, da Silva DLN (1979) Sex determination in bees. VIII. Relative action of genes xa and xb on sex determination in Melipona bees. Rev Bras Genet 2:267-280 Cole BJ (1983) Multiple mating and the evolution of social behaviour in the Hymenoptera. Behav Ecol Sociobiol 12:191-201 Crozier RH, Fjerdingstad EJ (2001) Polyandry in social Hymenoptera – disunity in diversity? Ann Zool Fenn 38:267-285 Crozier RH, Page RE Jr (1985) On being the right size – male contributions and multiple mating in social Hymenoptera. Behav Ecol Sociobiol 18:105-115 Darchen R, Delage-Darchen B (1971) Le déterminisme des castes chez les Trigones (Hyménoptères Apidés). Insect Soc 18:121-134 Darchen R, Delage-Darchen B (1977) Sur le déterminisme des castes chez les Meliponinés (Hymenoptères: Apidés). Bull Biol Fr Belg 11:91-109 Davidson DW (1982) Sexual selection in harvester ants (Hymenoptera: Formicidae: Pogonomyrmex) Behav Ecol Sociobiol 10:245-250 Denny AJ, Franks NR, Powell S, Edwards KJ (2004) Exceptionally high levels of multiple mating in an army ant. Naturwissenschaften 91:396-399 Estoup A, Solignac M, Harry M, Cornuet J-M (1993) Characterization of (GT)n and (CT)n microsatellites in two insect species: Apis mellifera and Bombus terrestris. Nuc Acids Res 21:1427-1431 Estoup A, Garnery L, Solignac M, Cornuet J-M (1995) Microsatellite variation in honey bee (Apis mellifera L.) populations: hierarchical genetic structure and test of the infinite allele and stepwise mutation models. Genetics 140:679-695 Fewell JH, Page RE Jr (1999) The emergence of division of labour in forced associations of normally solitary ant queens. Evol Ecol Res 1:537-548 Fewell JH, Page RE Jr (2000) Colony-level selection effects on individual and colony foraging task performance in honeybees, Apis mellifera L. Behav Ecol Sociobiol 48:173-181 Franks NR (1985) Reproduction, foraging efficiency and worker polymorphism in army ants. In: Hölldobler B, Lindauer M (eds) Experimental Behavioral Ecology and Sociobiology: in Memoriam Karl von Frisch, 1886-1982, Vol 31. Sinauer Associates, Sunderland/MA, pp 91-107

Extreme polyandry in social Hymenoptera

435

Frumhoff PС, Baker J (1988) A genetic component to division of labour within honey bee colonies. Nature 333:358-361 Fuchs S, Moritz RFA (1999) Evolution of extreme polyandry in the honeybee Apis mellifera L. Behav Ecol Sociobiol 9:269-275 Gotwald WH (1995) Army Ants: The Biology of Social Predation. Cornell University Press, Ithaca Greeff JM (1996) Effects of thelytokous worker reproduction on kin-selection and conflict in the Cape honeybee, Apis mellifera capensis. Philos Trans R Soc Lond B 351:617-625 Haberl M, Tautz D (1998) Sperm usage in honey bees. Behav Ecol Sociobiol 42:247-255 Hamilton WD (1964) The genetical evolution of social behaviour. J Theor Biol 7:1-16 Helanterä H, Strassmann JE, Carrillo J, Queller DC (2009) Unicolonial ants: where do they come from, what are they and where are they going? Trends Ecol Evol 24:341-349 Hillesheim E, Koeniger N, Moritz RFA (1989) Colony performance in honeybees (Apis mellifera capensis Esch) depends on the proportion of subordinate and dominant workers. Behav Ecol Sociobiol 24:291-296 Hughes WOH, Sumner S, Van Borm S, Boomsma JJ (2003) Worker caste polymorphism has a genetic basis in Acromyrmex leaf-cutting ants. Proc Natl Acad Sci USA 100:9394-9397 Hughes WOH, Oldroyd BP, Beekman M, Ratnieks FLW (2008a) Ancestral monogamy shows kin selection is key to the evolution of eusociality. Science 320:1213-1216 Hughes WOH, Ratnieks FLW, Oldroyd BP (2008b) Multiple paternity or multiple queens: two routes to greater intracolonial genetic diversity in the eusocial Hymenoptera. J Evol Biol 21:1090-1095 Jaffé R, Kronauer DJC, Kraus FB, Boomsma JJ, Moritz RFA (2007) Worker caste determination in the army ant Eciton burchellii. Biol Lett 3:513-516 Jennions MD, Petrie M (2000) Why do females mate multiply? A review of the genetic benefits. Biol Rev 75:21-64 Keller L, Genoud M (1997) Extraordinary lifespans in ants: a test of evolutionary theories of ageing. Nature 389:958-960 Keller L, Reeve HK (1994) Genetic variability, queen number, and polyandry in social Hymenoptera. Evolution 48:694-704 Keller L, Reeve HK (1995) Why do females mate with multiple males? The sexually selected sperm hypothesis. Adv Stud Behav 24:291-315 Keller L, Ross KG (1995) Gene by environment interaction – effects of a singlegene and social environment on reproductive phenotypes of fire ant queens. Funct Ecol 9:667-676 Kerr WE (1950) Genetic determination of castes in the genus Melipona. Genetics 35:143-152 Kolmes SA, Winston ML, Fergusson LA (1989) The division of labor among worker honey bees – the effects of multiple patrilines. J Kansas Entomol Soc 62:80-95

436

F. Bernhard Kraus and Robin F.A. Moritz

Kraus FB, Neumann P, van Praagh J, Moritz RFA (2004) Sperm limitation and the evolution of extreme polyandry in honeybees (Apis mellifera L.). Behav Ecol Sociobiol 55:494-501 Kronauer DJC, Boomsma JJ (2007) Multiple queens means fewer mates. Curr Biol 17:R753-R755 Kronauer DJC, Schöning C, Pedersen JS, Boomsma JJ, Gadau J (2004) Extreme queen-mating frequency and colony fission in African army ants. Mol Ecol 13:2381-2388 Kronauer DJC, Berghoff SM, Powell S, Denny AJ, Edwards KJ, Franks NR, Boomsma JJ (2006) A reassessment of the mating system characteristics of the army ant Eciton burchellii. Naturwissenschaften 93:402-406 Kronauer DJC, Johnson RA, Boomsma JJB (2007) The evolution of multiple mating in army ants. Evolution 61:413-422 Lattorff HMG, Moritz RFA, Fuchs S (2005) A single locus determines thelytokous parthenogenesis of laying honeybee workers (Apis mellifera capensis). Heredity 94:533-537 Lattorff HMG, Moritz RFA, Crewe RM, Solignac M (2007) Control of reproductive dominance by the thelytoky gene in honeybees. Biol Lett 3:292-295 Linksvayer TA, Fondrk MK, Page RE Jr (2009) Honeybee social regulatory networks are shaped by colony-level selection. Am Nat 173:E99-E107 Mattila HR, Seeley TD (2007) Genetic diversity in honey bee colonies enhances productivity and fitness. Science 317:362-364 Mays HL Jr, Hill GE (2004) Choosing mates: good genes versus genes that are a good fit. Trends Ecol Evol 19:554-559 Moritz RFA (1985) The effects of multiple mating on the worker-queen conflict in Apis mellifera L. Behav Ecol Sociobiol 16:375-377 Moritz RFA (1986) Intracolonial worker relationship and sperm competition in the honeybee (Apis mellifera L). Experientia 42:445-448 Moritz RFA, Haberl M (1994) Lack of meiotic recombination in thelytokous parthenogenesis of laying workers of Apis mellifera capensis (the Cape honeybee). Heredity 73:98-102 Moritz RFA, Kryger P, Koeniger G, Koeniger N, Estoup A, Tingek S (1995) High degree of polyandry in Apis dorsata queens detected by DNA microsatellite variability. Behav Ecol Sociobiol 37:357-363 Moritz RFA, Kryger P, Allsopp MH (1996) Competition for royalty in bees. Nature 384:31 Moritz RFA, Simon UE, Crewe RM (2000) Pheromonal contest between honeybee workers (Apis mellifera capensis). Naturwissenschaften 87:395-397 Moritz RFA, Pflugfelder J, Crewe RM (2003) Lethal fighting between honeybee queens and parasitic workers (Apis mellifera). Naturwissenschaften 90:378381 Moritz RFA, Lattorff HMG, Neumann P, Kraus FB, Radloff SE, Hepburn HR (2005) Rare royal families in honeybees, Apis mellifera. Naturwissenschaften 92:488-491

Extreme polyandry in social Hymenoptera

437

Nelson CM, Ihle KE, Fondrk MK, Page RE Jr, Amdam GV (2007) The gene vitellogenin has multiple coordinating effects on social organization. Plos Biol 5: 5:e62, doi: 10.1371/journal.pbio.0050062 Neumann P, Moritz RFA (2002) The Cape honeybee phenomenon: the sympatric evolution of a social parasite in real time? Behav Ecol and Sociobiol 52:271281 O’Donnell S (1998) Dominance and polyethism in the eusocial wasp Mischocyttarus mastigophorus (Hymenoptera: Vespidae). Behav Ecol Sociobiol 43:327331 Oldroyd BP, Rinderer TE, Harbo JR, Buco SM (1992) Effects of intracolonial genetic diversity on honey-bee (Hymenoptera, Apidae) colony performance. Ann Entomol Soc Am 85:335-343 Onions GW (1912) South African ‘fertile worker bees’. Agr J Union S Afr 1:720728 Oster GF, Wilson EO (1978) Caste and Ecology in the Social Insects. Princeton University Press, Princeton Page RE Jr (1986) Sperm utilization in social insects. Annu Rev Entomol 31:297320 Page RE Jr, Amdam GV (2007) The making of a social insect: developmental architectures of social design. Bioessays 29:334-343 Page RE Jr, Mitchell SD (1998) Self-organization and the evolution of division of labor. Apidologie 29:171-190 Page RE Jr, Robinson GE, Fondrk MK (1989) Genetic specialists, kin recognition and nepotism. Nature 338:576-579 Page RE Jr, Erber J, Fondrk MK (1998) The effect of genotype on response thresholds to sucrose and foraging behavior of honey bees (Apis mellifera L.). J Comp Physiol A 182:489-500 Palmer KA, Oldroyd BP (2000) Evolution of multiple mating in the genus Apis. Apidologie 31:235-248 Ratnieks FLW (1988) Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. Am Nat 132:217-236 Ratnieks FLW, Boomsma JJ (1995) Facultative sex allocation by workers and the evolution of polyandry by queens in social Hymenoptera. Am Nat 145:969993 Ratnieks FLW, Boomsma JJ (1997) On the robustness of split sex ratio predictions in social Hymenoptera. J Theor Biol 185:423-439 Rheindt FE, Strehl CP, Gadau J (2005) A genetic component in the determination of worker polymorphism in the Florida harvester ant Pogonomyrmex badius. Insectes Soc 52:163-168 Robinson GE, Page RE Jr (1989) Genetic determination of nectar foraging, pollen foraging, and nest-site scouting in honey bee colonies. Behav Ecol Sociobiol 24:317-323 Robinson GE, Page RE Jr (1995) Genotypic constraints on plasticity for corpse removal in honey bee colonies. Anim Behav 49:867-876 Ross KG, Vargo EL, Keller L (1996) Simple genetic basis for important social traits in the fire ant Solenopsis invicta. Evolution 50:2387-2399

438

F. Bernhard Kraus and Robin F.A. Moritz

Ruttner H (1980) Haltung der Königinnen während der Paarungszeit. In: Ruttner F (ed) Königinnenzucht. Apimondia, Bukarest, pp 225-267 Ruttner F, Koeniger G (1971) Die Füllung der Spermatheka der Bienenkönigin. Z Vgl Physiol 72:411-422 Scheiner R, Page RE Jr, Erber J (2001) The effects of genotype, foraging role, and sucrose responsiveness on the tactile learning performance of honey bees (Apis mellifera L.). Neurobiol Learn Mem 76:138-150 Schlüns H, Koeniger G, Koeniger N, Moritz RFA (2004) Sperm utilization pattern in the honeybee (Apis mellifera). Behav Ecol Sociobiol 56:458-463 Schlüns H, Moritz RFA, Neumann P, Kryger P, Koeniger G (2005) Multiple nuptial flights, sperm transfer and the evolution of extreme polyandry in honeybee queens. Anim Behav 70:125-131 Schmid-Hempel P (1994) Infection and colony variability in social insects. Philos Trans R Soc Lond B 346:313-321 Schmid-Hempel P, Crozier RH (1999) Polyandry versus polygyny versus parasites. Philos Trans R Soc Lond B 354:507-515 Schwander T, Keller L (2008) Genetic compatibility affects queen and worker caste determination. Science 322:552 Simon UE, Moritz RFA, Crewe RM (2005) Reproductive dominance among honeybee workers in experimental groups of Apis mellifera capensis. Apidologie 36:413-419 Snyder LE (1993) Non-random behavioral interactions among genetic subgroups in a polygynous ant. Anim Behav 46:431-439 Starr CK (1985) Sperm competition, kinship and sociality in the aculeate Hymenoptera. In: Smith RL (ed) Sperm Competition and the Evolution of Animal Mating Systems. Academic Press, New York, pp 428-464 Stuart RJ, Page RE Jr (1991) Genetic component to division of labor among workers of a leptothoracine ant. Naturwissenschaften 78:375-377 Tarpy DR, Page RE Jr (2000) No behavioral control over mating frequency in queen honey bees (Apis mellifera L.): implications for the evolution of extreme polyandry. Am Nat 155:820-827 Trivers RL, Hare H (1976) Haplodiploidy and theevolution of the social insects. Science 191:249-263 Vahed K (1998) The function of nuptial feeding in insects: review of empirical studies. Biol Rev 73:43-78 van Wilgenburg E, Driessen G, Beukeboom L (2006) Single locus complementary sex determination in Hymenoptera: an ‘unintelligent’ design? Front Zool 3:1 Wang Y, Amdam GV, Rueppell O, Wallrichs MA, Fondrk MK, Kaftanoglu O, Page RE Jr (2009) PDK1 and HR46 gene homologs tie social behavior to ovary signals. PloS ONE 4:e4899, doi:10.1371/journal.pone.0004899 Wattanachaiyingcharoen W, Oldroyd BP, Wongsiri S, Palmer K, Paar R (2003) A scientific note on the mating frequency of Apis dorsata. Apidologie 34:85-86 Wenseleers T, Ratnieks FLW (2006) Comparative analysis of worker reproduction and policing in eusocial Hymenoptera supports relatedness theory. Am Nat 168:E163-E179

Extreme polyandry in social Hymenoptera

439

West-Eberhard MJ (1987) Flexible strategy and social evolution. In: Ito Y, Brown JL, Kikkawa J (eds) Animal Societies, Theories and Facts. Japan Scientific Societies, Tokyo, pp 35-51 West-Eberhard MJ (1996) Wasp societies as microcosms for the study of development and evolution. In: Turillazzi S, West-Eberhard MJ (eds) Natural History and Evolution of Paper Wasps. Oxford University, New York, pp 290317 Woyke J (1964) Causes of repeated mating flights by queen honeybees. J Apic Res 3:17-23 Woyke J (1965) Genetic proof of the origin of diploid drones from fertilised eggs of the honeybee. J Apic Res 4:7-11 Yue C, Schröder M, Gisder S, Genersch E (2007) Vertical-transmission routes for deformed wing virus of honeybees (Apis mellifera). J Gen Virol 88:2329-2336 Zeh JA, Zeh DW (1996) The evolution of polyandry. I. Intragenomic conflict and genetic incompatibility. Proc R Soc Lond B 263:1711-1717 Zeh JA, Zeh DW (1997) The evolution of polyandry. II. Post-copulatory defences against genetic incompatibility. Proc R Soc Lond B 264:69-75 Zeh JA, Zeh DW (2001) Reproductive mode and the genetic benefits of polyandry. Anim Behav 61:1051-1063

Chapter 15

Monogynous mating strategies in spiders JUTTA SCHNEIDER AND LUTZ FROMHAGE

ABSTRACT To understand the evolution of mating systems, we need to consider why reproductive strategies differ between species in the way they do. For example, males in most mammals appear entirely specialised on mating with many females, whereas males in many birds also invest in offspring quality by providing paternal care. Alternatively or additionally, males may invest in enhancing their paternity share in any mating they get. Taken to the extreme, this may result in males investing maximally in gaining paternity with one or two females only, in the absence of paternal care. Mating systems with monogynous (or bigynous) males and polyandrous females are taxonomically widespread but relatively rare overall. In spiders, however, low male mating frequencies have evolved several times independently and are associated with remarkable adaptations that include sexual cannibalism and genital damage. In the first part of this chapter, we describe the mating strategies, and the associated costs and benefits of sexual cannibalism and genital damage, of selected spider species that represent independent evolutionary origins of these traits. We then introduce models that investigate the evolution of mating systems with low male mating frequencies. The models predict that a male-biased sex ratio is required for monogyny to evolve. We outline how such a sex ratio bias may have arisen in concert with female-biased sexual size dimorphism. Finally, we discuss monogyny in the light of sexual conflict theory.

15.1 Sex roles and the evolution of mating systems Accounts of sexual selection often start by citing the classical prediction that, in order to maximise reproductive success, males should compete for mating with as many females as possible, whereas females should attempt to mate only with the most suitable males (Bateman 1948, Trivers 1972).

442

Jutta Schneider and Lutz Fromhage

These so-called conventional sex roles are indeed very common, and have received much attention by researchers in the past. However, to understand the principles of the evolution of mating strategies, it is necessary to acknowledge that these sex roles are but one of several possible outcomes. The tendency of males to strive for mating with many females is sometimes portrayed as the outcome of a self-reinforcing process that began with the evolution of anisogamy: because sperm are designed to compete for fertilisations, males are deemed to be predisposed to compete for matings. A particularly clear summary of this view was formulated by Cronin (2006): ‘If you specialise in competing, you gain most selective advantage by putting more into competing; and the same for caring [for offspring]. And so the divergence widened over evolutionary time, with natural selection proliferating and amplifying the differences, down the generations, in every sexually reproducing species that has ever existed.’ It has recently been emphasised, however, that narratives of this kind are incomplete to the point of being misleading (Kokko and Jennions 2008). There is no logical necessity by which competitiveness in one sex, e.g. males, should beget even greater competitiveness. On the contrary: provided that alternative behavioural options to increase fitness (e.g. parental care) exist, strong mating competition can act as an evolutionary disincentive for joining the competition, leading to more egalitarian sex roles (Yamamura and Tsuji 1993, Webb et al. 1999, Fromhage et al. 2007b). In other words, if males face a choice between caring for an existing set of offspring or deserting them to search for an additional receptive female, the latter option becomes less rewarding as receptive females become rare due to intense competition. To explain why males in many species nevertheless provide less care for offspring than do females, one must invoke additional factors, such as uncertainty of paternity and non-random variance in male mating success (as mediated by female choice or male-male competition, Queller 1997, Kokko and Jennions 2008). Paternity uncertainty is relevant here because it dilutes any benefits that will reach a caring male’s genetic offspring; variance in mating success is relevant because very successful males, whose reproductive rate is higher than that of the average male and female, have more to lose than females by spending time with their offspring (Queller 1997, Kokko and Jennions 2008). Whether male care can evolve in any particular species depends on the combined effects of all of the above factors. Thus, males can enhance their reproductive success by multiple mating, as observed in the majority of mammals and/or by caring for offspring, as observed in many bird and fish species (see e.g., Kempenaers and Schlicht this volume). Alternatively or additionally, males may also invest in en-

Monogynous mating strategies in spiders

443

hancing the paternity share obtained in any mating they get. Here we focus on the relatively small number of species in which males specialise entirely on this latter aspect, investing heavily in paternity enhancement with only one or a few females. We suggest that studying such exceptions can be particularly instructive, because any general rule can be said to be truly understood only to the degree that its exceptions are understood (see also Anthes this volume). In the following sections, we will review the mating systems of spiders and other taxa with low male mating frequencies, outline theoretical explanations, and discuss these ideas in a context of sexual conflict and life history evolution.

15.2 Spiders as a model system for studying the evolution of mating systems Spiders are ideal for studying the evolution of mating strategies because they include notable examples of specialisation on paternity maximisation, and more generally because they exhibit large variation in reproductive strategies between and within species. Although spider males do not provide paternal care (one notable exception among the Opilionidae: Mora 1990), males in a relatively large number of species will mate with only one or at most two females in their lifetime. A recent surge of interest in this topic has opened up new ways of understanding the evolution of behavioural and morphological spider traits that have puzzled researchers in the past (see below). For a more general perspective on spider mating behaviour, including on species with high male mating frequencies, see Elgar (1998), Huber (2005) and Gaskett (2007). Low male mating frequencies have evolved independently in several taxa among spiders and are associated with remarkable mating strategies. Such strategies include self-sacrifice, as can be found in some Latrodectus species, e.g. in the form of the copulatory somersault or spontaneous death, and genital mutilation (BOX 15.1-3). The latter refers to damage of the male intromittent organs (pedipalps) that occurs during copulation (Fig.15.1), usually rendering the used pedipalp functionally sterile. In some species whole pedipalps are ectomised (Kuntner 2005, Knoflach and van Harten 2006). Male mating frequency can also be limited by sexual cannibalism without obvious complicity on the part of the male. Whether there is selection on males to avoid this fate depends critically on the timing of cannibalism (Elgar 1992, Elgar and Schneider 2004). Males have paired pedipalps and in several species can make two insertions, one with each pedipalp (‘one-shot’ genitalia), usually filling one of the female’s paired sperm storage organs at a time.

444

Jutta Schneider and Lutz Fromhage

BOX 15.1 Family Araneidae, Genus Argiope [distribution global; 82 species; no species phylogeny] Natural history All species show extreme sexual size dimorphism; species in temperate climates have a very pronounced seasonal reproductive phase; the mating season can be shorter than 2 weeks. Egg sacs hatch synchronously after overwintering. Argiope bruennichi The protandrous males are attracted via airborne chemical cues to the webs of mature females. No reliable sex ratio data are available, but an experimental field study found that virgin females are visited by an average of 3.4 males during the brief (5 days) peak of the mating season, 70% of which go on to mate. Because males are subject to high mortality, particularly through sexual cannibalism, the initial male bias becomes reversed as the season progresses (K.F. Schulte, G. Uhl and J.M. Schneider unpubl. data). Copulations are preceded by a brief, energetic courtship display by the male, to which females respond by assuming a characteristic mating position. Females never reject or attack a courting male, but stereotypically attack him immediately after the onset of copulation; only about 20-30% of males escape such an attack (Schneider et al. 2006). Males use only one of their paired pedipalps during each copulatory bout. The tip of the sperm transferring structure (embolus) breaks off during sperm transfer and remains inside the genital duct (Uhl et al. 2007). Here the fragment functions as a mating plug that drastically reduces copulation duration of a subsequent male, and thereby the amount of sperm transferred. Thus, the paternity share of any male attempting to copulate into the same genital opening later is significantly reduced (Nessler et al. 2007). It is important to note that a male killed after his first insertion leaves the female’s second genital opening unused. If a second male mates into this unused side, he can expect to share paternity equally with the first male (Schneider and Lesmono 2009). Under such conditions there is no first male priority. However, because second males are unable to detect which side is unused, they face a 50% risk of using the plugged side instead. Hence, first males enjoy an overall advantage, which may explain why males prefer to mate with virgin females. The attractiveness of females decreases rapidly after mating (K.F. Schulte, G. Uhl and J.M. Schneider unpubl. data). To survive a copulation, A. bruennichi males must jump off the female almost immediately after inserting a pedipalp; if they do not do so within ca. 10 seconds, their survival chance is virtually zero (Schneider et al. 2006). On the other hand, if a male does not attempt to escape, he can copulate for a bit longer while already held in the female’s fangs. The rate of sperm transfer is such that a pedipalp is about one third empty after 10 s, and completely emp-

Monogynous mating strategies in spiders

445

ty after 35 s (Schneider et al. 2006). Males can choose between two mating tactics: a high variance tactic that offers high (but uncertain) potential gains, and a low variance tactic that offers more predictable gains. The high variance tactic involves a maximum of two copulations: one that lasts for ca. 10 s and is ended by a lucky escape at the last possible moment, and a second one that is ended by sexual cannibalism. Depending on whether the escape succeeds, this tactic can lead to a relatively long or to a very short total copulation duration. In contrast, a male using the low variance tactic does not attempt to escape and is therefore able to extend his copulation for several seconds beyond the female’s attack, resulting in an intermediate total copulation duration. Interestingly, the presence or absence of female pheromones in the environment of maturing males influences the choice of mating tactic: males were more likely to use the low variance strategy if they matured in the presence of female pheromones (Nessler et al. 2009). The reason for this behaviour is unclear, as it contradicts the intuition that finding another female should be easier under high female density. One possible explanation is that female density will often be correlated with male density, and high male density reduces the gains from further mate search. Similarities and differences between species Species differ in the frequency of sexual cannibalism and in the rates of genital damage and plugging. In A. aurantia, males always attempt two copulations with the same female and spontaneously die (without any attack by the female) during the second copulation (Foellmer and Fairbairn 2003); A. keyserlingi males only mate once with a female, guard her for some time, and then leave to search for another female (Herberstein et al. 2005). In the same species, virgin males prefer virgin females over mated females; a preference that is not present in mated males (Gaskett et al. 2004). a

b

c

d e2

e1

Photos: a) female and male Argiope bruennichi during courtship; the female is in the mating position; b) male paired genitalia (pedipalps); the tips of the black structures (emboli) can break off during copulation (red circle); c, d) macerated epigyne with tips of male genitalia stuck in female genital openings (e1 and e2) (c – dorsal view; d – ventral view), sp – paired spermathecae

446

Jutta Schneider and Lutz Fromhage

BOX 15.2 Family Nephilidae, Genus Nephila [tropical and subtropical distribution; 15 species; phylogeny described] Natural history All species show extreme sexual size dimorphism. Reproduction is synchronised in the sub-tropics and continuous in the tropics. Seasonality affects upper limits of female size (Higgins 2000). Nephila fenestrata Males of this species are generally monogynous. Often several males reside on the large, relatively permanent webs of females (tertiary sex ratio: 1.4) and compete for copulations (Fromhage et al. 2007a). Unlike in Argiope, males can copulate with each of their pedipalps in succession without getting off the female. Each pedipalp can be used only once. Females generally attempt to terminate copulation aggressively, but cannot easily reach the male while he keeps his small body pressed at hers during copulation. Up to 40% of males are eventually caught and cannibalised. However, males can greatly reduce this risk by mating with a female while she is feeding (Fromhage et al. 2007a). After mating, stuck pedipalp fragments in the female genital openings make it difficult for subsequent suitors to copulate (Fromhage and Schneider 2006). However, such mating plugs offer only limited protection, so that surviving males can benefit from post-copulatory mate-guarding. Because this requires male survival, sexual cannibalism in N. fenestrata is against the male’s interest, even if he is functionally sterile after mating (Fromhage and Schneider 2005b). Similarities and differences between species N. fenestrata is probably the only Nephila species in which male genital fragments are effective mating plugs. In the other seven species with common male genital damage, the frequent occurrence of multiple fragments per genital opening suggests the absence of a plug function (Kuntner et al. 2007, 2008a,b). In one case, this conjecture has been supported by experimental data: male N. plumipes that mate into a previously used genital opening transfer similar amounts of sperm, and achieve similar paternity, than males mating into an unused opening of a previously mated female. Perhaps genital damage in this species occurs because the hooked tip of the embolic conductor firmly attaches to the female, and then breaks off in order to facilitate escape from her attack. In this species, males have little time for mating and have to wait until the female captures a prey item. Males follow the hunting female and can only copulate while the female is holding the prey with her chelicerae. In the remaining Nephila species, genital damage is uncommon. In at least two species (N. edulis, N. clavipes) males can use each pedipalp more than once. There are four genera in the family Nephilidae and the phylogeny is very suggestive regarding the evolution of monogyny. Whereas the basic clade,

Monogynous mating strategies in spiders

447

Clitaetra, has large males, no genital damage and no sexual cannibalism, the remaining three genera, Herennia, Nephilengys and Nephila all contain species with extreme sexual size dimorphism and genital damage (Kuntner et al. 2008). Broken-off fragments of male genitalia are large in Herennia and are effective plugs. H. multipuncta males ectomise their entire pedipalp after use. Palpless males (eunuchs) survive and remain on the female’s web, probably also to defend the female against other males. Getting rid of useless appendages may improve agility and/or signal readiness to invest maximally in fighting (gloves-off hypothesis, Kuntner et al. 2009). A comparative analysis revealed that female genitalia co-vary in their degree of complexity with male pedipalps. Sexual selection and antagonistic coevolution are the suggested forces behind this pattern; alternative explanations in the context of natural selection are not supported (Kuntner et al. 2009). simple genitalia multiple plugs possible complex genitalia, mostly single plugs

Nephila

cruentata

borbonica

Nephilengys

Herennia

no palpal removal

episinoides perroti clathrata simoni irenae thisbe

no plugs known

multipuncta etruscilla deelemanae tone agnarssoni gagamba papuana

polyandry slit epigynum simple genitalia, simple palp

papuana

chambered epigynum

malabarensis

fenestrata

clavipes sexpunctata constricta pilipes sumptuosa inaurata ardentipes turneri komaci antipodiana senegalensis plumipes edulis clavata

(unknown in N. papuana)

embolic plugging of epigynum palpal removal / eunuchs

Clitaetra

extreme sexual size dimorphism chambered epigynum complex palp

a

b

Pictures: a) female and male Nephila fenestrata during mating; b) phylogenetic tree of the Nephilidae (with permission from Kuntner et al. 2008)

448

Jutta Schneider and Lutz Fromhage

BOX 15.3 Family Theridiidae, Genera  Latrodectus [distribution global; 31 species; described phylogeny]  Tidarren [sub-/tropical distribution; <10 species; no species phylogeny] Natural history All Latrodectus species show extreme sexual size dimorphism. Egg-sacs are produced throughout the summer and hatch successively, so that many interactions are between spiders of different ages. Latrodectus hasselti Copulating males somersault their tiny abdomen onto the fangs of the female, thus actively inviting cannibalism. This self-sacrifice provides the male with a paternity advantage under sperm competition (Andrade 1996, Snow and Andrade 2004). Males protect their paternity by leaving the tip of their sperm-transferring embolus deep inside the convoluted insemination duct of the female, where it blocks the entrance to the spermatheca (Snow et al. 2006). It is crucial, however, that the sclerite is accurately placed at the entrance of the spermatheca; otherwise the plug is ineffective. In L. hasselti, 49% of the males that initiate mating survive to inseminate and plug the female’s second spermatheca. Males somersault during each copulation and, to defer death despite being fed on, they retract vital internal organs to a protected position at the front end of their body (Andrade et al. 2005). This tactic confers a substantial advantage: an experimental treatment simulating cannibalism killed 89% of control males within one hour, but only 22% of males with retracted internal organs. Males preferentially associate with virgin females, as expected given the first male priority found in this species. In the presence of pheromones emitted by virgin females, males shorten their penultimate instar and mature sooner, at the expense of body size and condition (Kasumovic and Andrade 2006). Males mate after several hours of elaborate courtship and females reward this effort by reducing the probability of premature cannibalism for long-courting males (Stoltz et al. 2009). Similarities and differences between species Copulatory somersault is optional in L. geometricus, where it occurs more often with virgin than mated females (Segoli et al. 2008). Although male monogyny is the rule in this genus, there appears to be variation in the degree to which this reflects male strategies rather than external constraints. Survival of roving males has been measured in two studies and was found to be around 20% in L. hasselti and L. pallidus; up to six males were found in female webs of both species (Andrade 2003, Segoli et al. 2006). Another interesting genus in this family is Tidarren, in which males always ectomise one of their pedipalps at maturation, and then set out to fill only one spermatheca of a single female (Knoflach and van Harten 2006). Males of T. argo always die during copulation, using their entire bodies as temporary mating plugs (Knoflach and van Harten 2001).

Monogynous mating strategies in spiders

(a)

449

sp

sp

cd

cd

(b)

2

1

a

b

c

Photos: a) female consuming male Latrodectus hasselti (photo © M.C.B. Andrade 2003) b) position of sclerites in female insemination duct (with permission from Snow et al. 2006); c) Tidarren cuneolatum: sexual cannibalism of single-palp male (after Knoflach and van Harten 2001, photo © B. Knoflach-Thaler); sp = spermatheca, cd = copulatory duct, 1 and 2 mark sclerites

Cannibalism before the first insertion obviously reduces a male’s reproductive success to zero; this outcome is infrequent in the species in question (see BOX 15.1-3). Getting cannibalised after the first insertion reduces a male’s maximum mating success by half, whereas death after the second insertion leaves his total mating frequency unchanged. The most prominent spider families exhibiting these patterns are the Araneidae, Theriididae and Nephilidae. In BOX 15.1-3 we describe some examples from these families, in which males appear specialised to maximise paternity with only one or at most two females. The outcome of sexual interactions between males and females depends on many factors, including their evolutionary interests and their capacity to manipulate or physically coerce each other (Arnqvist and Rowe 2005, Hosken et al. 2009). Spiders are exceptional in that females are usually larger than males, and thus appear physically able to dominate and control interactions with males. Moreover, in contrast to many insect species, male spiders do not possess clasping mechanisms (with one exception, see Eberhard and Huber in press) suited to constrain females during copulation, and there are very few reports of sexual coercion of females by males (e.g. Becker et al. 2005, Rezac 2009). However, male spiders can sometimes circumvent female control by more subtle means, e.g. by initiating mating while the female is moulting or feeding (Schneider and Lubin 1998). But even if a female loses control over mating itself, there remains a possibility that she retains some control over fertilisation. This is because in the Entelegynae (the most diverse spider group, comprising >75% of all

450

Jutta Schneider and Lutz Fromhage

EC

EC

a

E

b

Nephila fenestrata pedipalp intact (a) and mutilated (b) EC = embolic conductor E = embolus

Fig. 15.1 Male genitalia from N. fenestrata before (a) and after use (b) for sperm transfer (with permission from Kuntner et al. 2008).

species, Platnick 2009), females possess paired sperm storage organs (spermathecae) that are often filled by different males, potentially allowing the female to use the ejaculate from one or the other male selectively (see also Hellriegel and Ward 1998). For example, in an experimental study on Argiope lobata, once-mated females were assigned to receive a second ejaculate, either from a brother or a non-brother, to fill the as yet unused spermatheca. Subsequent sperm counts revealed that fewer sperm from the second male were stored if he was a brother, as would be expected if females avoid inbreeding (Welke and Schneider 2009). Several studies have shown that female spiders play an active role in determining how often and for how long males copulate. For example, in the black widow Latrodectus hasselti, the monogynous males can achieve a maximum of two copulations, but the female sometimes kills them after only one copulation. Because this ends a male’s life prematurely, it has been termed premature sexual cannibalism (Stoltz et al. 2009). Premature sexual cannibalism is less likely after a male has shown elaborate courtship (Stoltz et al. 2009). The authors interpret this result such that females punish poorly courting males by preventing their second copulation. Similarly, in A. bruennichi, courtship positively influences paternity, although it is not clear how this relationship arises at a mechanistic level (Schneider and Lesmono 2009). In another species, A. keyserlingi, females end copulations by abruptly attacking the male, varying the timing of these attacks such that relatively smaller males, if they happen to mate after a larger predecessor, can copulate for longer and thus obtain a greater share of paternity (Elgar et al. 2000). Although these examples suggest that females prefer certain sires over others, no study has as yet linked such preferences to female lifetime reproductive success, or the success of the resulting offspring. It therefore remains a pressing question whether (and if so, how) spider females benefit from being choosy.

Monogynous mating strategies in spiders

451

Males too are to a certain degree flexible in their reproductive decisions (Nessler et al. 2009). For example, A. bruennichi skip courtship, proceeding directly to copulation instead, if a rival male is simultaneously present on the same (virgin) female’s web. On the other hand, they often reject non-virgin females and they can influence the probability of premature sexual cannibalism (BOX 15.1). In L. hasselti, males accelerate maturation in the presence of mature females (see below for more detail) and are adapted to survive for a certain time even after the onset of cannibalism (see BOX 15.3).

15.3 Theory Two facts in need of explanation have emerged from the previous section: spider males can be limited in their mating frequency (1) behaviourally, by sexual cannibalism i.e., and (2) morphologically, i.e., by the structure of their genitalia. How could such seemingly disadvantageous traits evolve and be maintained? A simple, but short-sighted answer is available if we consider each of these traits in isolation while taking the other as given: given that functionally sterile males have no reproductive future anyway, cannibalism seems cost-free and thus requires no special explanation. Similarly, if males are cannibalised after mating anyway, functional sterility seems cost-free. Explaining one paradoxical trait by another leaves open, however, why either trait evolved in the first place. This kind of reasoning offers no plausible account of the historical sequence of evolutionary events, and thus represents a so-called ‘sequence fallacy’ (Simmons and Parker 1989). To avoid this fallacy, one may argue that sexual cannibalism evolved as a female foraging strategy with no adaptive significance from the male’s perspective (Newman and Elgar 1991). This view appears plausible for species where males occasionally get cannibalised despite attempting to escape this fate. It cannot account, however, for lack of escape behaviour or even active male complicity (see BOX 15.3) in species where sexual cannibalism is common. Alternatively, sexual cannibalism and/or male sterility can be explained in terms of their fitness consequences for males. A useful concept in this context is the distinction between ‘paternal investment’ and ‘mating effort’ (Fig. 15.2). Paternal investment is defined as an action performed by a male that increases a female’s reproductive success; a mating effort is an action that confers a paternity advantage over rival males (Simmons and Parker 1989).

452

Jutta Schneider and Lutz Fromhage males

female

a

x

b

x

c

x

eggs

Fig. 15.2 Paternal investment and mating effort. a) Baseline case with neither paternal investment nor mating effort, where two males mate with the same female and each sire half of her offspring (indicated by corresponding egg colours). b) By performing a ‘paternal investment’, the yellow male causes the female to produce more eggs. Paternity over these additional eggs is shared equally among males. c) By performing a ‘mating effort’, the yellow male increases his paternity at the cost of the red male; the total number of eggs remains constant.

It has been suggested that sexual cannibalism may be a form of paternal investment, where males are selected to be cannibalised because they thereby provide a nutritious meal to the female (Buskirk et al. 1984). If this extra energy enables females to produce more (or better) offspring that are at least in part fathered by the cannibalised male, then his net fitness may increase as a result (Fig. 15.2b). Exploring this idea with a mathematical model, Buskirk et al. (1984) predicted that self-sacrifice should evolve under conditions where eating a male strongly enhances female fecundity, and where male mortality during mate search is high. The latter effect arises because a male who is unlikely to encounter more than one female in his lifetime should attempt to make the most of any opportunity he gets. The idea of sexual cannibalism as a form of paternal investment is not supported by empirical findings, however: perhaps because of the much smaller size of males compared to females in highly cannibalistic spiders, no nutritional benefits of mate consumption have been found (reviewed in Andrade and Kasumovic 2005). However, there are some other (spider and

Monogynous mating strategies in spiders

453

insect) species in which such benefits exist (reviewed in Wilder et al. 2009). Alternatively, sexual cannibalism and/or male sterility may represent an example of mating effort (Fig. 15.2c), with males limiting themselves to mating with one or a few females because by doing so they can achieve a higher paternity share (as first suggested by Elgar 1992, for ceratopogonid midges). In the extreme case where selection favours paternity maximisation with a single female, any means to achieve this goal is available for evolution to work on, including suicidal mating and the modification of genitalia for one-way use. According to this view, sexual cannibalism and post-mating sterility may be manifestations of the same general phenomenon. But why should selection favour male monogamy (monogyny) in the first place? The adaptive value of such a strategy has been investigated using theoretical models that compared the fitness of monogynous males (assumed to have a paternity advantage) with males capable of multiple mating (Andrade 2003, Fromhage et al. 2005). A central insight from the latter model was that the evolution of monogyny requires a male-biased effective sex ratio (ESR), defined as the ratio between males and females that mate at least once. Specifically, monogyny was found to be favoured by selection if a monogynists’ average paternity share was greater than 1/ESR (Fromhage et al. 2005), implying that a lower paternity share requires a correspondingly greater ESR bias for monogyny to evolve. The intuition behind this result is illustrated in Fig. 15.3. Another factor found to be of importance in this context is male mortality during mate search, which has two conflicting effects: on the one hand, high search mortality can favour monogynists by rendering the alternative of continued searching unprofitable. On the other hand, by reducing the number of rival males, it can make paternity protection unnecessary. Which of these effects prevails depends on other factors, such as the sex ratio at maturation and the frequency of monogynists in the population. Three possible evolutionary outcomes emerge from this model: pure monogyny, pure polygyny, and a mixture of both. Empirical evidence linking monogyny with a male-biased sex ratio is provided in Sects. 4 and 5. In contrast to Fromhage et al.’s (2005) assumption that monogyny and polygyny are immediate alternatives, males in several spider species are constrained to use each of their pedipalps only once. This limits males to perform either two copulations with one female (monogyny), or one copulation with each of two females (bigyny). Incorporating this constraint, Fromhage et al. (2008) analysed a model in which monogynous males could protect their paternity either by out-competing rivals in sperm competition, and/or by reducing the probability of female remating. As in the

454

Jutta Schneider and Lutz Fromhage

a

b

Fig. 15.3 Sex ratio and the evolution of monogyny. a) In a population where males (drawn around the periphery) outnumber females (drawn in a circle representing the total amount of eggs produced), a monogynous mutant male (arrow) monopolises paternity over the eggs of one female (shaded area). b) This male has a relative advantage because excess males reduce the average paternity per male.

previous model, possible evolutionary outcomes of this model included monogyny, bigyny, and a mixture of both.

15.4 Sex ratio and sexual size dimorphism Theory suggests that a male-biased sex ratio is a main driving force in the evolution of monogyny. Census data consistent with this prediction are available for several of the species in question, with estimates ranging between 1.5 and 10 maturing males for each maturing female (Christenson et al. 1985, Christenson 1989, Miyashita 1993, Andrade 1996, Foellmer and Fairbairn 2005a, Fromhage et al. 2007a). Here the number of ten males per female is an upper estimate not accounting for female mortality; it is based on the observed median number of two males per female (Andrade 1996) in a species where four out of five males die during mate search without ever reaching a female (Andrade 2003). Moreover, a comparative study across orb-web spiders (Araneoidae) found a correlation between monogyny and the accumulation of multiple males in the female

Monogynous mating strategies in spiders

455

web, the latter being a qualitative indicator of a male-biased sex ratio (Miller 2007). But which factors generate a male-biased sex ratio in the first place? Frequency-dependent selection is generally expected to maintain the primary sex ratio near unity (Fisher 1930), and no exceptions to this prediction are known from the spider species in question. In contrast, there is no theoretical reason why the sex ratio at maturation (the tertiary sex ratio) should necessarily be at unity, and a bias in the tertiary ratio can arise from sex-specific growth strategies. Indeed, Miller (2007) found a correlation between extreme sexual size dimorphism and the accumulation of multiple males in the female web. Spider males generally mature after fewer instars and at a much smaller body size than females (Higgins 2002). This difference can result from males growing at a slower rate and/or for a shorter time than females. Slow growth increases survival by reducing risks associated with foraging; a short growth period reduces the opportunity for mortality to strike before maturation. Males in some species can reduce their development time to a fraction of that of females (Vollrath and Parker 1997, Higgins 2000). In some seasonally reproducing species, however, males are forced to keep their development in approximate synchrony with females, or else they would not be able to find a mate (Higgins 2002). Early male maturation (protandry), in combination with progressive maturation of females, can result in seasonal biases of sex ratios, with early maturing females experiencing an excess of mating partners whereas others remain unmated (for a similar case in butterflies, see Calabrese et al. 2008). Similar processes could act in space, with some females never being reached by males, whereas others are found by many (Kasumovic et al. 2008). Rather than being just a side-effect of a particular growth strategy, protandry can also directly be favoured by selection (Morbey and Ydenberg 2001). This will be the case if early maturing males have better access to virgin females, and virgin females are particularly valuable mates, e.g., because there is a first-male advantage. The above argument establishes a link between several of the characteristic features of the mating strategies discussed here, namely sexual size dimorphism, protandry, sex-ratio bias, and first male priority, all of which may have influenced each other during the course of evolution. Studies on within-species patterns provide insights on how selection on body size and the timing of maturation may have contributed to the observed life-history strategies of males. A laboratory study on L. hasselti demonstrated that developmental plasticity in the timing of male maturation is an adaptive response to female availability: males mature sooner and at a lower body condition in the presence of female sex pheromones that signal

456

Jutta Schneider and Lutz Fromhage

availability of mating partners (Kasumovic and Andrade 2006). This flexible response allows males to track unpredictable changes in mate availability, which are likely to arise by chance in species where females mature throughout the season. If females of different developmental stages occur in clusters, as is the case in Nephila species, early maturing females may trigger male maturation and then face an excess of males, whereas late maturing females have fewer or perhaps even no mating partners (Kasumovic and Andrade 2006). Indeed, a field study on N. plumipes revealed differences in adult male body size between clusters, and found that these differences were correlated with the number of adult females present per cluster (Kasumovic et al. 2008). Although the causation of this pattern is unclear, these findings are consistent with males adjusting their developmental strategy to local demographic variation, and to the associated relative advantages of protandry versus size-related competitive ability. Thus far we have focussed our discussion on selection pressures on males. To explain sexual size dimorphism, however, it is necessary to consider selection pressures in both sexes. In principle, both sexes should benefit from having a high chance of reaching maturity, and might thus be selected to mature at a small size. However, this benefit is traded-off against sex-specific alternative benefits. In female spiders, large body size is strongly associated with high fecundity (Head 1995, Blanckenhorn 2000), whereas benefits of larger male size are less clear (Schneider et al. 2000). In N. clavipes and N. plumipes, relatively larger males are more likely to hold the hub position near the female, which may provide advantages in male-male competition (Vollrath 1980, Elgar and Fahey 1996). However, in a study on N. edulis, holding the hub position did not predict paternity (Elgar and Jones 2008). In a field study on Argiope aurantia, Foellmer and Fairbairn (2005a) found selection on large body size through male-male competition, but no consistent selection patterns on male body size during the roving phase (Foellmer and Fairbairn 2005b). Sometimes small male size can even be advantageous: in N. fenestrata, small males were better at sneaking up to a female guarded by a rival (Fromhage and Schneider 2005a). Taken together, these findings suggest that growing to a large size tends to be more advantageous for female than for male spiders.

15.5 Parallel cases in other taxa Mating systems involving monogyny are not restricted to spiders (Hosken et al. 2009). Other examples include social insects, such as certain ants, stingless bees and honey bees (Boomsma et al. 2005), whose mating beha-

Monogynous mating strategies in spiders

457

viour is no less dramatic than in spiders. For example, in the queenless ant Dinoponera quadriceps, the copulating male remains linked to the female until she severs the end of his abdomen, leaving parts of his genitalia attached to hers (Monnin and Peeters 1998). No female action is required to bring about the male honey bee’s death, when ‘as part of the mating, the distal portion of the male genitalia virtually explodes with a loud pop into the sting chamber of the female and breaks off there’ (Michener 1974:363). These species are characterised by a highly male-biased sex ratio among reproductive individuals, as is typical of social insects whose reproduction involves colony fission (Bulmer 1983). In ceratopogonid biting midges, the male is killed and sucked out by the female while mating is in progress, leaving his broken-off genitalia attached to the female when he is finally discarded (Downes 1978). In a number of other taxa, monogynous males survive mating but are permanently attached to the female’s body surface. Examples include molluscs, crustaceans, representatives of the phyla Cycliophora and Echiura, and even one vertebrate species, the angler fish (Vollrath 1998), all of which are characterised by female-biased size dimorphism and malebiased sex ratios. Describing such a case in barnacles, Darwin (1854:23) wrote that ‘These males are minute, often exceedingly minute, and consequently generally more than one is attached to a single female; and I have seen as many as fourteen adhering on one female!’ Among parasitic flatworms of the family Schistosomatidae, male-biased sex ratios predict a species’ tendency to be monogamous, with males guarding females in a specialised groove running the length of the male’s body (Morand and Muller-Graf 2000). Whereas the above examples concern species in which males mate with only one female in their lifetime, a similar logic may apply to sequential monogamy in longer-lived species. A comparative study on mammals found a positive association between monogamy and females holding nonoverlapping territories, but no effect of paternal care (Komers and Brotherton 1997). This suggests that monogamy in mammals often reflects a male mate-guarding strategy, perhaps driven by a shortage of reproductive females as might result from some females’ failure to obtain a territory. It is worth noting that monogamy is the exception in mammals, and that the more commonly observed polygyny in this group has been suggested to be driven by a female-biased adult sex ratio (Kokko and Jennions 2008).

458

Jutta Schneider and Lutz Fromhage

15.6 Conflicts between and within the sexes Males in any mating system can benefit from monopolising paternity over their mates’ offspring; but they may also benefit from undermining their rivals’ monopolisation attempts. Additionally, females may benefit from mating with more than one male per reproductive cycle (Arnqvist and Nilsson 2000), resulting in conflicts of interests within and between the sexes over female remating. The result can be a coevolutionary arms race in which repeated male adaptations for monopolisation provoke counteradaptations in both males and females (Chapman et al. 2003). What are the implications of monogyny in this context? One might expect that males who concentrate entirely on monopolising one female should be particularly good at this, thus potentially enforcing monogamy on the female as well. However, because counter-adaptations in either sex may undermine monopolisation, shared paternity need not be less common in species with monogynous males compared to other species. If monogynists are unable to monopolise females, there are two possible outcomes: either monogyny is nevertheless associated with sufficiently large paternity shares to be stable (see Sect. 3), or else it is destabilised and is replaced by polygyny. Male adaptations associated with monogyny include ‘one-shot’ genitalia, life-long mate-guarding, and suicidal mating. Female adaptations may include premature sexual cannibalism as a possible mechanism to avoid monopolisation. These traits may occur alternatively or simultaneously. Which traits evolve in any particular species is likely to depend, among other things, on the sperm priority pattern, i.e., the dependence of paternity on mating sequence. For example, if the female genital morphology is such that the first male to mate receives most paternity (first male priority; cf. Eberhard et al. 1993, Uhl 2002), males may benefit from guarding females that will soon moult to maturity. If the opposite pattern holds (last male priority) males can benefit more from guarding females after copulation until oviposition. Since this requires male survival after copulation, suicidal mating tactics may evolve less easily under last male priority. In conclusion, the diverse male and female behaviours associated with monogyny in various taxa are likely to depend on morphological details.

15.7 Outlook The existing literature on spider mating strategies has largely focussed on males. There are few reliable data on female mating frequencies and on

Monogynous mating strategies in spiders

459

any benefits of polyandry in species with monogynous males. Although data on the number of male visitors per female are available for many species, in most cases we do not know how many of these males are also mating partners. Measuring the costs and benefits of different mating rates for females will be a crucial step forward. Empirical and theoretical investigations of mating systems with monogyny have provided exciting insights, and have placed apparently paradoxical traits in a larger evolutionary framework. The described cases illustrate the fact that the general assertion of males maximising fitness through seeking high numbers of mating partners is not universally true, even if males do not provide paternal investment. To understand the species-specific dynamics that lead to the described mating strategies, costs and benefits of different mating rates of females need to be characterised in several model systems. Given the factors we have highlighted, future studies should particularly target the selection pressures and ecological conditions that initially favoured the evolution of male-biased adult sex ratios and sexual size dimorphism.

15.8 References Andrade MCB (1996) Sexual selection for male sacrifice in the Australian redback spider. Science 271:70-72 Andrade MCB (2003) Risky mate search and male self-sacrifice in redback spiders. Behav Ecol 14:531-538 Andrade MCB, Kasumovic MM (2005) Terminal investment strategies and male mate choice: extreme tests of Bateman. Integr Comp Biol 45:838-847 Andrade MCB, Gu L, Stoltz JA (2005) Novel male trait prolongs survival in suicidal mating. Biol Lett 1:276-279 Arnqvist G, Nilsson T (2000) The evolution of polyandry: multiple mating and female fitness in insects. Anim Behav 60:145-164 Arnqvist G, Rowe L (2005) Sexual Conflict. Princeton University Press, Princeton Bateman AJ (1948) Intra-sexual selection in Drosophila. Heredity 2:349-368 Becker E, Riechert S, Singer F (2005) Male induction of female quiescence/catalepsis during courtship in the spider, Agelenopsis aperta. Behaviour 142:57-70 Blanckenhorn WU (2000) The evolution of body size: what keeps organisms small? Q Rev Biol 75:385-407 Boomsma JJ, Baer B, Heinze J (2005) The evolution of male traits in social insects. Annu Rev Entomol 50:395-420 Bulmer MG (1983) Sex ratio theory in social insects with swarming. J Theor Biol 100:329-339

460

Jutta Schneider and Lutz Fromhage

Buskirk RE, Frohlich C, Ross KG (1984) The natural selection of sexual cannibalism. Am Nat 123:612-625 Calabrese JM, Ries L, Matter SF, Debinski DM, Auckland JN, Roland J, Fagan WF (2008) Reproductive asynchrony in natural butterfly populations and its consequences for female matelessness. J Anim Ecol 77:746-756 Chapman T, Arnqvist G, Bangham J, Rowe L (2003) Sexual conflict. Trends Ecol Evol 18:41-47 Christenson TE (1989) Sperm depletion in the orb-weaving spider Nephila clavipes (Araneae, Araneidae). J Arachnol 17:115-118 Christenson TE, Brown SG, Wenzl PA, Hill EM, Goist KC (1985) Mating behaviour of the golden-orb-weaving spider, Nephila clavipes: I. Female receptivity and male courtship. J Comp Psychol 99:160-166 Cronin H (2006) The battle of the sexes revisited. In: Grafen A, Ridley M (eds) Richard Dawkins: How a Scientist Changed the Way We Think. Reflections by Scientists, Writers, and Philosophers. Oxford University Press, Oxford, pp 14-26 Darwin C (1854) A Monograph on the Sub-Class Cirripedia, with Figures of All the Species, Vol 2. The Ray Society, London Downes JA (1978) The feeding and mating in the insectivorous Ceratopogoninae (Diptera). Mem Entomol Soc Can 104:1-62 Eberhard WG, Huber BA (in press) Spider genitalia: precise maneuvers with a numb structure in a complex lock. In: Leonard J, Córdoba-Aguilar A (eds) Evolution of Primary Sexual Characters in Animals. Oxford University Press, Oxford Eberhard WG, Guzmangomez S, Catley KM (1993) Correlation between spermathecal morphology and mating systems in spiders. Biol J Linn Soc Lond 50:197-209 Elgar MA (1992) Sexual cannibalism in spiders and other invertebrates. In: Elgar MA, Crespi BJ (eds) Cannibalism: Ecology and Evolution among Diverse Taxa. Oxford University Press, Oxford, pp 129-156 Elgar MA (1998) Sperm competition and sexual selection in spiders and other arachnids. In: Birkhead TR, Møller AP (eds) Sperm Competition and Sexual Selection. Academic Press, San Diego, pp 307-339 Elgar MA, Fahey BF (1996) Sexual cannibalism, competition, and size dimorphism in the orb-weaving spider Nephila plumipes Latreille (Araneae: Araneoidea). Behav Ecol 7:195-198 Elgar MA, Jones TM (2008) Size-dependent mating strategies and the risk of cannibalism. Biol J Linn Soc Lond 94:355-363 Elgar MA, Schneider JM (2004) Evolutionary significance of sexual cannibalism. Adv Stud Behav 34:135-163 Elgar MA, Schneider JM, Herberstein ME (2000) Female control of paternity in the sexually cannibalistic spider Argiope keyserlingi. Proc R Soc Lond B 267:2439-2443 Fisher RA (1930) The Genetical Theory of Natural Selection. Clarendon Press, Oxford

Monogynous mating strategies in spiders

461

Foellmer MW, Fairbairn DJ (2003) Spontaneous male death during copulation in an orb-weaving spider. Proc R Soc Lond B 270:S183-S185 Foellmer MW, Fairbairn DJ (2005a) Competing dwarf males: sexual selection in an orb-weaving spider. J Evol Biol 18:629-641 Foellmer MW, Fairbairn DJ (2005b) Selection on male size, leg length and condition during mate search in a sexually highly dimorphic orb-weaving spider. Oecologia 142:653-662 Fromhage L, Schneider JM (2005a) Safer sex with feeding females: sexual conflict in a cannibalistic spider. Behav Ecol 16:377-382 Fromhage L, Schneider JM (2005b) Virgin doves and mated hawks: contest behaviour in a spider. Anim Behav 70:1099-1104 Fromhage L, Schneider JM (2006) Emasculation to plug up females: the significance of pedipalp damage in Nephila fenestrata. Behav Ecol 17:353-357 Fromhage L, Elgar MA, Schneider JM (2005) Faithful without care: the evolution of monogyny. Evolution 59:1400-1405 Fromhage L, Jacobs K, Schneider JM (2007a) Monogynous mating behaviour and its ecological basis in the golden orb spider Nephila fenestrata. Ethology 113:813-820 Fromhage L, McNamara JM, Houston AI (2007b) Stability and value of male care for offspring – is it worth only half the trouble? Biol Lett 3:234-236 Fromhage L, Houston AI, McNamara JM (2008) A model for the evolutionary maintenance of monogyny in spiders. J Theor Biol 250:524-531 Gaskett AC (2007) Spider sex pheromones: emission, reception, structures, and functions. Biol Rev 82:26-48 Gaskett AC, Herberstein ME, Downes BJ, Elgar MA (2004) Changes in male mate choice in a sexually cannibalistic orb-web spider (Araneae: Araneidae). Behaviour 141:1197-1210 Head G (1995) Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (Class Araneae). Evolution 49:776-781 Hellriegel B, Ward PI (1998) Complex female reproductive tract morphology: Its possible use in postcopulatory female choice. J Theor Biol 190:179-186 Herberstein ME, Barry KL, Turoczy MA, Wills E, Youssef C, Elgar MA (2005) Post-copulation mate guarding in the sexually cannibalistic St Andrew’s Cross spider (Araneae Araneidae). Ethol Ecol Evol 17:17-26 Higgins L (2000) The interaction of season length and development time alters size at maturity. Oecologia 122:51-59 Higgins L (2002) Female gigantism in a New Guinea population of the spider Nephila maculata. Oikos 99:377-385 Hosken DJ, Stockley P, Tregenza T, Wedell N (2009) Monogamy and the battle of the sexes. Annu Rev Entomol 54:361-378 Huber BA (2005) Sexual selection research on spiders: progress and biases. Biol Rev 80:363-385 Kasumovic MM, Andrade MCB (2006) Male development tracks rapidly shifting sexual versus natural selection pressures. Curr Biol 16:R242-R243

462

Jutta Schneider and Lutz Fromhage

Kasumovic MM, Bruce MJ, Andrade MCB, Herberstein ME (2008) Spatial and temporal demographic variation drives within-season fluctuations in sexual selection. Evolution 62:2316-2325 Knoflach B, van Harten A (2001) Tidarren argo sp. nov. (Araneae: Theridiidae) and its exceptional copulatory behaviour: emasculation, male palpal organ as a mating plug and sexual cannibalism. J Zool Lond 254:449-459 Knoflach B, van Harten A (2006) The one-palped spider genera Tidarren and Echinotheridion in the Old World (Araneae, Theridiidae), with comparative remarks on Tidarren from America. J Nat Hist 40:1483-1616 Kokko H, Jennions MD (2008) Parental investment, sexual selection and sex ratios. J Evol Biol 21:919-948 Komers PE, Brotherton PNM (1997) Female space use is the best predictor of monogamy in mammals. Proc R Soc Lond B 264:1261-1270 Kuntner M (2005) A revision of Herennia (Araneae, Nephilidae, Nephilinae), the Australasian ‘coin spiders’. Invertebr Syst 19:391-436 Kuntner M, Coddington JA, Hormiga G (2008) Phylogeny of extant nephilid orbweaving spiders (Araneae, Nephilidae): testing morphological and ethological homologies. Cladistics 24:147-217 Kuntner M, Coddington JA, Schneider JM (2009) Intersexual arms race? Genital coevolution in nephilid spiders (Araneae, Nephilidae). Evolution 63:14511463 Michener CD (1974) The Social Behavior of the Bees: A Comparative Study. Harvard University Press, Cambridge/MA Miller JA (2007) Repeated evolution of male sacrifice behavior in spiders correlated with genital mutilation. Evolution 61:1301-1315 Miyashita T (1993) Male male competition and mating success in the orb-web spider, Nephila clavata, with reference to temporal factors. Ecol Res 8:93-102 Mora G (1990) Paternal care in a neotropical harvestman, Zygopachylus albomarginis (Arachnida, Opiliones, Gonyleptidae). Anim Behav 39:582-593 Morand S, Muller-Graf CDM (2000) Muscles or testes? Comparative evidence for sexual competition among dioecious blood parasites (Schistosomatidae) of vertebrates. Parasitology 120:45-56 Morbey YE, Ydenberg RC (2001) Protandrous arrival timing to breeding areas: a review. Ecol Lett 4:663-673 Nessler SH, Uhl G, Schneider JM (2007) Genital damage in the orb-web spider Argiope bruennichi (Araneae: Araneidae) increases paternity success. Behav Ecol 18:174-181 Nessler SH, Uhl G, Schneider JM (2009) Scent of a woman – the effect of female presence on sexual cannibalism in an orb-weaving spider (Araneae: Araneidae). Ethology 115:633-640 Newman JA, Elgar MA (1991) Sexual cannibalism in orb-weaving spiders: an economic model. Am Nat 138:1372-1395 Platnick NI (2009) The world spider catalog, version 9.5. American Museum of Natural History, online at http://research.amnh.org/entomology/spiders/catalog/index.html

Monogynous mating strategies in spiders

463

Queller DC (1997) Why do females care more than males? Proc R Soc Lond B 264:1555-1557 Rezac M (2009) The spider Harpactea sadistica: co-evolution of traumatic insemination and complex female genital morphology in spiders. Proc R Soc Lond B 276:2697-2701 Schneider JM, Lesmono K (2009) Courtship raises male fertilization success through post-mating sexual selection in a spider. Proc R Soc Lond B 276: 3105-3111 Schneider JM, Lubin Y (1998) Intersexual conflict in spiders. Oikos 83:496-506 Schneider JM, Herberstein ME, De Crespigny FC, Ramamurthy S, Elgar MA (2000) Sperm competition and small size advantage for males of the golden orb-web spider Nephila edulis. J Evol Biol 13:939-946 Schneider JM, Gilberg S, Fromhage L, Uhl G (2006) Sexual conflict over copulation duration in a cannibalistic spider. Anim Behav 71:781-788 Segoli M, Harari AR, Lubin Y (2006) Limited mating opportunities and male monogamy: a field study of white widow spiders, Latrodectus pallidus (Theridiidae). Anim Behav 72:635-642 Segoli M, Arieli R, Sierwald P, Harari AR, Lubin Y (2008) Sexual cannibalism in the brown widow spider (Latrodectus geometricus). Ethology 114:279-286 Simmons LW, Parker GA (1989) Nuptial feeding in insects: mating effort versus paternal investment. Ethology 81:332-343 Snow LSE, Andrade MCB (2004) Pattern of sperm transfer in redback spiders: implications for sperm competition and male sacrifice. Behav Ecol 15:785792 Snow LSE, Abdel-Mesih A, Andrade MCB (2006) Broken copulatory organs are low-cost adaptations to sperm competition in redback spiders. Ethology 112:379-389 Stoltz JA, Elias DO, Andrade MCB (2009) Male courtship effort determines female response to competing rivals in redback spiders. Anim Behav 77:79-85 Trivers RL (1972) Parental investment and sexual selection. In: Campbell B (ed) Sexual Selection and the Descent of Man, 1871-1971. Aldine, Chicago, pp 136-179 Uhl G (2002) Female genital morphology and sperm priority patterns in spiders (Araneae). In: Toft S, Scharff N (eds) European Arachnology 2000. Aarhus University Press, Aarhus, pp 145-156 Uhl G, Nessler SH, Schneider J (2007) Copulatory mechanism in a sexually cannibalistic spider with genital mutilation (Araneae: Araneidae: Argiope bruennichi). Zoology 110:398-408 Vollrath F (1980) Male body size and fitness in the web-building spider Nephila clavipes. Z Tierpsychol 53:61-78 Vollrath F (1998) Dwarf males. Trends Ecol Evol 13:159-163 Vollrath F, Parker GA (1997) Giant female or dwarf male spiders? Reply. Nature 385:688-688 Webb JN, Houston AI, McNamara JM, Szekely T (1999) Multiple patterns of parental care. Anim Behav 58:983-993

464

Jutta Schneider and Lutz Fromhage

Welke K, Schneider JM (2009) Inbreeding avoidance through cryptic female choice in the cannibalistic orb-web spider Argiope lobata. Behav Ecol, doi:10.1093/beheco/arp097 Wilder SM, Rypstra AL, Elgar MA (2009) The importance of ecological and phylogenetic conditions for the occurrence and frequency of sexual cannibalism. Annu Rev Ecol Evol Syst 40:21-39 Yamamura N, Tsuji N (1993) Parental care as a game. J Evol Biol 6:103-127

Chapter 16

Mating systems, social behaviour and hormones WOLFGANG GOYMANN AND HERIBERT HOFER

ABSTRACT There has been an extensive debate in behavioural biology to what extent the function and adaptive value of behavioural traits require, or are limited by, specific physiological processes triggered by hormones. The observation that testis size predicts mating systems suggested that testosterone may be an important proximate factor involved in the regulation of mating systems. Indeed, there is evidence that seasonal patterns of circulating testosterone are related to mating systems. Testosterone manipulations affect mating decisions in some, but not all species. In addition, the peptide hormones oxytocin and arginine vasopressin play a major role in pair-bond formation and the expression of different mating systems in mammals. Exciting recent studies indicated that changes in complex social behaviours may be based on the differential expression pattern of just one hormone receptor gene. Rather than social status itself, the process by which social status is achieved and maintained determines the physiological and psychosocial effects associated with a given social status. Current knowledge suggests that the relationship between behaviour and hormones can be quite complex. It also shows that the study of function and adaptive value of behavioural traits and hormonal processes is most successful if the links between the two are explicitly recognised.

16.1 Introduction In this chapter we explore the link between physiological mechanisms, as exemplified by hormonal processes, and behavioural traits associated with mating systems and social behaviour. Such links were examined ever since comparative studies demonstrated that mating systems may be linked to morphological traits, such as testes size (Clutton-Brock and Harvey 1977), and because gamete production is under control of hormonal processes.

466

Wolfgang Goymann and Heribert Hofer

For instance, in ring-tailed lemur (Lemur catta) groups, females synchronise their estrus, males compete intensely for access to females, and males have the largest relative testis size of any strepsirrhine primate (Kappeler 1997). The naïve view of this link suggests that hormones are the building blocks from which behaviour is generated, or in other words, that there is a simple chain of causality from hormones to behaviour, and that there is a simple dose-response relationship between the concentration of a specific hormone and the intensity of a behavioural response of the organism. There has been substantial progress in this field over the past 30 years, much of which is associated with the development of improved analytical tools to assess hormone concentrations, particularly minimally-invasive or non-invasive procedures. The methodological progress facilitated the study of hormone concentrations without apparently influencing the behaviour of the animals in free-ranging populations, expanded the behavioural contexts in which hormones could be studied, and extended the range of species in which such studies could be performed. This has led to many empirical studies in a variety of vertebrate species, and carefully set-up laboratory experiments in rather fewer species, mostly birds and rodents. In this chapter we focus on the conceptual advances resulting from this work and use some well researched examples to illustrate how and why in the context of mating systems and social behaviour the joint investigation of hormonal processes and behavioural traits has shed new light on the evolution of both. A particular focus will be on testosterone and mating systems, but the relationship of other hormones and behaviour will also be explored.

16.2 Behaviour and hormones Before we delve deeper into hormone-behaviour relationships, it is important to clarify the four perspectives that we could use to study behaviour and hormones. These were established by Nikolaas Tinbergen based on previous work by Huxley (1916). We will illustrate Tinbergen’s (1963) four kinds of problems of behavioural biology by addressing the question ‘Why do male songbirds in temperate zones sing?’ 16.2.1 Tinbergen‘s four perspectives The first problem focuses on the proximate aspects of behaviour, the proximate causation or mechanism: the processes inside the animal that are required to perform the behaviour, and the environmental stimuli that elicit

Mating systems, social behaviour and hormones

467

the behaviour. Applied to our example, a study from this perspective would argue that male birds sing because an increase in day length induces growth of testes, resulting in elevated levels of testosterone. These elevated levels induce singing behaviour (Catchpole and Slater 2008). The second problem deals with the ontogeny of a behaviour: when and how does the behaviour emerge in the course of an individual’s development and which internal and external factors control the expression of behavioural changes? In our example, a study from this perspective would record changes in hormone concentrations during ontogeny and the occurrence of male song. Unfortunately, this question has been very difficult to tackle. In many songbirds, activational effects of testosterone seem to suffice to induce song, as demonstrated by experiments in which adult females received testosterone and then started to sing just like males. Other species, such as the zebra finch (Taeniopygia guttata), are more complicated: adult females treated with testosterone do not sing. Although giving estradiol to female nestlings and testosterone when adult does induce singing, direct genetic differentiation rather than organisational effects of hormones are likely to play the major role in the sexual differentiation of song in zebra finches (reviewed in Adkins-Regan 2005). The third problem focuses on ultimate aspects, i.e. the function of a behaviour: what is the impact of the behaviour on an individual’s chance of survival and reproduction? For our example, a study from this perspective would record the possible functions of the song of male birds and conclude that male passerines sing to attract mates and defend a territory against rivals (Krebs 1977, Krebs et al. 1978, Catchpole and Slater 2008). The final problem deals with how the behaviour evolved during the evolutionary history of a species. ‘Song’ as such is not specific to songbirds, but songbirds (as well as parrots and hummingbirds) are special because they learn their song. Hence, the evolutionary history suggests that songbirds sing learned song, because they originated from a single ancestral species in which song was learned. From this example it becomes clear that hormones immediately appeal from a perspective of proximate causation and ontogeny, because they have organisational effects during the development of an organism and activational effects throughout its life (Phoenix et al. 1959, Sachser and Kaiser this volume). So, why should hormones be interesting for anyone studying the function of behaviour? The function of many behaviours is intimately related to the life histories of individuals, and behavioural ecologists are interested in life histories because life history traits are highly relevant for the Darwinian fitness of an individual – its ability to survive and reproduce. Briefly, life history refers to how long an animal lives, how long it takes to reach reproductive maturity, how often and in which inter-

468

Wolfgang Goymann and Heribert Hofer

vals it reproduces, how many offspring it has, etc. Many life history traits are regulated by hormones as hormones regulate traits central to growth, metabolism and reproduction. In addition, they organise transitions between life history states and may act as the physiological mediators of functional trade-offs. Testosterone, for example, acts as a signal to increase the amount of territorial and mate attraction behaviour, such as singing and territorial defence in birds. All other things being equal, the energy and time that is allocated towards these activities needs to be taken away from other activities, such as foraging, parental care or – somewhat more controversially – the immune system (Folstad and Karter 1992, Roberts et al. 2004). In addition, such territorial behaviour may make the individual more conspicuous, thereby increasing the risk of predation. Thus, from a functional perspective, hormones are interesting because they mediate traits that are relevant for survival and reproduction (Hau 2007). 16.2.2 Hormones influence behaviour and vice versa Long before behavioural endocrinology was founded as a scientific discipline, people ‘knew’ that hormones influence behaviour – for instance farmers, who, for thousands of years, removed testicles to turn uncooperative and aggressive bulls into docile oxen. The year 1849 is now considered the official founding date of behavioural endocrinology as a scientific discipline. In that year, German scientist Arnold A. Berthold published his studies in which he re-implanted and transplanted testes in castrated cockerels, showing that typical male behaviours such as crowing, mounting and fighting are restored by a secretory blood-borne product from the testes (Berthold 1849), a substance and hormone now known as testosterone. This hormone is the principal steroid secreted by the vertebrate testes, except for teleost fish where it is 11-keto-testosterone. The secretion of testosterone is regulated by the central nervous system (Fig. 16.1). Testosterone plays a pivotal role in the regulation of morphological structures and behaviour of male vertebrates during reproduction. It modulates sexual behaviour as well as behaviour related to dominance and resource holding potential in a reproductive context, such as breeding season territoriality, vocalisations, and mate guarding. While Berthold conducted the first experiment in behavioural endocrinology, the principal founder of the discipline was Frank A. Beach, who named and defined the field (Beach 1948). Since these early days, countless studies have convincingly demonstrated that behaviour is influenced by hormones. The reverse idea that behaviour may also influence hormones became testable with the development of the radioimmunoassay by

Mating systems, social behaviour and hormones

Environmental information

469

Aggression, sexual behaviour

Brain Hypothalamus

Negative feedback

GnRH Anterior pituitary

Negative feedback

LH Testis

Testosterone

Fig. 16.1 Basic scheme of the hypothalamic-pituitary-gonadal axis (HPG-axis): environmental information (i.e. a change in day length, the presence of receptive females or competitors) induces the release of gonadotropin-releasing hormone (GnRH) from the hypothalamus. GnRH then induces the release of luteinising hormone (LH) from the anterior pituitary. LH then triggers the release of testosterone, which exerts negative feed-back on the hypothalamus and pituitary and affects sexual and aggressive behaviour (redrawn after Wingfield 1994).

Nobel Prize winner Rosalyn Yalow in the 1950s. For example, in a series of studies, Irwin S. Bernstein and colleagues discovered that fighting between males and access to females influenced testosterone concentration in male primates (Bernstein et al. 1983). A major step in understanding the relationship between behaviour and hormones was achieved through formulation of the Challenge Hypothesis by John Wingfield and colleagues (Wingfield et al. 1990). This hypothesis represented one of the first formalised attempts to explain the huge variation in plasma testosterone levels (spanning two orders of magnitude) as a function of differences in seasonality, male-male aggression and access to females, and it related this variation to the social mating system. While originally formulated for birds, the Challenge Hypothesis has since been highly influential for studies of all other vertebrate groups (see Sect. 16.3.1).

470

Wolfgang Goymann and Heribert Hofer

16.2.3 Hormones do not cause behaviour The realisation that hormones influence behaviour does not mean that hormones ‘cause’ behaviour. Rather, hormones change the likelihood of a behaviour occurring but only if the context is right. Consider the following example: Dusky-footed woodrats (Neotoma fuscipes) are solitary rodents residing in dwellings made of branches and twigs in oak woodland and stream habitats along the Pacific coast of the USA. The dwellings are crucial for survival by providing shelter from predators and weather extremes. Also, the dwellings are repositories for collected food. Male woodrats are aggressive towards other males and exclude them from their territories. To find out whether testosterone is involved in the degree of aggression in males, Caldwell et al. (1984) brought woodrats into the lab and tested them in a neutral arena. Surprisingly, castrated males fought as intensively as intact males and were equally successful in dominating encounters. Thus, aggression appeared to be independent of testosterone. However, when the researchers replaced the neutral arena with a more naturalistic setting and woodrats were allowed to construct and defend individual ‘houses’, the situation changed. Now, intact males had a clear advantage: testosterone did have an effect on aggression, but this effect depended on the presence of a defendable resource, i.e. the appropriate environmental context (Glickman and Caldwell 1994). This example nicely illustrates that hormones do not cause behaviour, but change the likelihood of a behaviour occurring if the context is right.

16.3 Mating systems and hormones Reproductive relationships between individuals can be described as mating systems. In the behavioural literature such relationships have been classified mainly in terms of the number of behavioural mates per male or per female. Although the recent literature has developed a much more sophisticated classification of mating systems (Shuster and Wade 2003), for the purpose of this contribution it is useful to recall the traditional broad classification of mating systems (e.g. Emlen and Oring 1977), provided we distinguish between social relationships and genetic parentage success. Social mating systems include: (1) Monogamy, in which one male and one female express a partner preference that leads to the formation and maintenance of either a temporary or a permanent pair-bond. Neither sex monopolises additional members of the opposite sex. (2) Polygyny, in which one male controls or gains access to several females. (3) Polyandry,

Mating systems, social behaviour and hormones

471

in which one female controls or gains access to several males. In the latter two cases, partner preference, pair-bond formation and maintenance are relaxed or at least biased towards one sex. Finally, there is (4) promiscuity, in which neither females nor males control access to members of the opposite sex and each individual may mate with multiple partners. In promiscuous species, pair-bond formation is absent or we know too little about them to be able to find out about mating preferences. After the establishment of genetic parentage analysis in the 1980s, it soon became clear that social mating systems are built upon genetic mating systems with varying degrees of extra-pair fertilisations (Kempenaers and Schlicht this volume). As a consequence, many more females produce offspring with several males than previously suspected. In other words, genetic mating systems describe who produces how many offspring with whom and with how many others (Shuster and Wade 2003). Genetic mating systems are also appropriate to explain alternative mating tactics, such as satellite or sneaker males, which attempt to fertilise females that visit territories of territorial males (Taborsky and Brockman this volume). For an example of proximate regulation of such alternative mating tactics see BOX 16.1. 16.3.1 Androgen responsiveness and mating systems: the Challenge Hypothesis The Challenge Hypothesis is central for an understanding of the relationship between hormones and mating systems. The development of the Challenge Hypothesis was triggered by the observation that testosterone levels of male birds may show large differences between species and huge seasonal fluctuations within species. Seasonally breeding birds with a high degree of male-male competition have high plasma androgen concentrations during periods when they establish territories, and/or when females are receptive. In contrast, high concentrations of circulating androgens are virtually absent in species that do not compete for territories or mates. The basic assumption of the Challenge Hypothesis is that elevations of circulating androgens above the breeding season baseline are not related to basal reproductive physiology, but are associated with temporal variations in aggressive and sexual behaviour. Indeed, there are studies that have demonstrated rapid effects of social interactions on plasma concentrations of testosterone in a wide array of vertebrate taxa, such as fish, amphibians, reptiles, birds and mammals, including humans (e.g. Wingfield and Wada 1989, Emerson and Hess 1996, Hirschenhauser et al. 2004, Oyegbile and Marler 2005, Archer 2006).

472

Wolfgang Goymann and Heribert Hofer

BOX 16.1 Alternative mating strategies and hormones in lizards The interplay between the functional significance of morphological or behavioural traits and of hormonal processes has been repeatedly explored in studies of male alternative mating tactics and hormones (Oliveira 2005, 2008, Miles et al. 2007). In this context, lizards provide well researched and instructive examples. Male side-blotched lizards (Uta stansburiana) are frequently found in the deserts of western North America and display one of the most spectacular examples of alternative mating strategies, where different mating strategies are linked with changes in morphological appearance, territorial defence and access to females, and testosterone concentrations. Upon sexual maturation, males develop bright throat colours. The three discrete colour morphs (orange, blue and yellow) are related to different mating tactics. Males with an orange throat are highly aggressive, defend large territories that overlap with the territories of several females, and have high levels of testosterone. Males with a blue throat have intermediate levels of testosterone, are less aggressive and defend small territories overlapping with those of one to few females. Males with a yellow throat express low levels of testosterone and do not defend territories but mimic females and sneak copulations mainly within the territories of orange males (Sinervo et al. 2000a). The three colour morphs play the lizard version of the rock-paper-scissor game (rock beats scissor, scissor beats paper, and paper beats rock). In the lizard version of this game, the highly aggressive and highly polygynous orange-throated males dominate the more monogamous and mate-guarding blue-throated males. The orange males in turn are bested by the yellowthroated sneakers, whose strategy is again bested by the mate-guarding strategy of the blue-throated males (Sinervo and Lively 1996). Each strategy in this game has weaknesses and strengths, which keeps the game running. Throat colour has a genetic basis (Zamudio and Sinervo 2000, Sinervo et al. 2000b, 2001, Sinervo et al. 2006) and its expression may be influenced by hormone levels during development. Apart from such organisational effects during ontogeny, hormones also have activational effects in adult lizards. Testosterone manipulation experiments suggest that testosterone is involved in the proximate regulation of mating strategies. If blue-throated and yellow-throated males are implanted with testosterone during adulthood, they become more active and acquire territory space from neighbouring sham-implanted males. The territories they defend after experimental treatment are almost as large as those of orange-throated males and overlap with territories of many females (DeNardo and Sinervo 1994, Sinervo et al. 2000a). Another study on a closely related lizard, the tree lizard Urosaurus ornatus focused on organisational effects of hormones during development. Similar to the side-blotched lizard, tree lizards come in several colour morphs.

Mating systems, social behaviour and hormones

473

Males with an orange-blue dewlap are aggressive and defend territories, whereas males with an orange dewlap are non-aggressive and do not establish territories. Males with high levels of testosterone and progesterone during ontogeny develop into the orange-blue morph, whereas males with low levels of testosterone and progesterone during development turn into the orange morph (Moore et al. 1998). Once they are adult, orange-blue males monopolise the territories of several females and orange males follow a sneaker strategy. However, testosterone concentrations do not differ between orange-blue and orange morphs in tree lizard males when they are adult (Moore 1991). Hence, the major hormonal effects on mating strategies seem to occur at different times during the life history of tree lizards and side-blotched lizards. Organisational effects of hormones during ontogenetic development are likely to shape mating tactics of tree lizards and these effects cannot be manipulated during adulthood (a fixed effect sensu Moore 1991). In contrast, hormonal effects in side-blotched lizards appear to be both organisational during development and activational during adulthood, thereby providing a measure of plasticity for controlling mating tactics in adult males.

Fig. 16.2 Three morphs of side-blotched lizards in the hand of Barry Sinervo: an orange homozygote, a blue homozygote, and a blue-yellow heterozygote. Photograph courtesy of Suzanne Mills and Barry Sinervo

474

Wolfgang Goymann and Heribert Hofer

C

Androgen concentration

C

Regulated (facultative)

Level C B

B

Regulated (seasonal)

Level B A

Level A

A

Homeostatic

Time of year

Fig. 16.3 The 3-level model of androgen secretion in seasonally breeding birds. Level A (black) represents the non-breeding androgen baseline required for feedback regulation of GnRH and gonadotropin release. Level B (blue) represents the androgen baseline during breeding induced by environmental cues such as the increase in day length. Level B is sufficient for spermatogenesis and for the expression of secondary sexual characters and reproductive behaviours. Level C (red) represents the physiological testosterone maximum that is assumed to be triggered during interactions with other males or receptive females. The increase from level A to level B occurs seasonally at the onset of the breeding season. The increase from level B to level C is facultative, i.e. only triggered by social stimulation or challenge during the breeding season. Redrawn from Wingfield et al. (1990).

In their formulation of the Challenge Hypothesis, Wingfield et al. (1990) postulated three (idealised) levels at which testosterone or other androgens are present in peripheral circulation (Fig. 16.3): First, there is a constitutive homoeostatic ‘Level A’ which represents the basal secretory activity of the Leydig cells during the non-breeding season. This level is presumed to maintain feedback regulation of both gonadotropin-releasing hormone (GnRH) from the hypothalamus and gonadotropin release from the pituitary gland (Fig. 16.1). Second, there is a regulated (periodic) breeding season baseline ‘Level B’, which represents the constitutive secretory activity stimulated by environmental cues such as day length (Fig. 16.3). Level B is considered sufficient for spermatogenesis, the development of secondary sexual characters and accessory organs, and the expression of reproductive behaviours. Levels A and B can be considered as the basic levels on which hormones influence behaviour, morphology and physiology. Finally, there is a maximum response ‘Level C’ that is reached through social stimulation from competing males or via interactions with receptive females. Thus, Level C is induced by behaviour feeding back on the secretion of the hormone. The increase of testosterone to

Mating systems, social behaviour and hormones

475

Androgen responsiveness

Monogamous Polyandrous Polygynous

Paternal care

Male aggression

Fig. 16.4 The relationship between androgen responsiveness and mating systems in birds, as originally formulated in the Challenge Hypothesis: species in which males provide substantial amounts of parental care (dark green) show a high androgen responsiveness. This means that they express low levels of testosterone (Level B) most of the time, but are capable of quickly increasing their testosterone concentrations to maximum (Level C) during periods of male-male interactions. In contrast, males that do not participate in parental care (yellow) but frequently interact with competitors show a low androgen responsiveness: they express high levels of testosterone (close to Level C) throughout the breeding season. Polygynous males that provide paternal care and experience intermediate levels of malemale conflicts show an intermediate pattern (light green). Drawing modified after Wingfield et al. (1990).

Level C can be short or long in duration, and small or great in magnitude. In contrast to the increase from Level A to Level B, which periodically occurs at the onset of the breeding season, the increase from Level B to Level C is considered facultative and may be expressed throughout all phases of the breeding life-cycle stage (Fig. 16.3). The three levels of testosterone release represent the first important cornerstone of the Challenge Hypothesis. The second important ingredient is the observation that high levels of testosterone interfere with male parental care or result in other costs that should be avoided. But what do these idealised levels A, B, and C of testosterone and the interference of testosterone with paternal care imply for the relationship between testosterone and mating systems? The Challenge Hypothesis states that temporal patterns of plasma testosterone are the result of a trade-off between the degree to which male parental care is necessary for reproductive success and the necessity or benefit of expressing aggressive behaviour (Fig. 16.4).

476

Wolfgang Goymann and Heribert Hofer

BOX 16.2 Sex role reversal, classical polyandry, and hormones The term ‘sex role reversal’ implies that there are ‘conventional’ and ‘reversed’ sex roles. This nomenclature is influenced by our perception and our evolutionary history as mammals – if you asked a teleost fish, its perception of different sex roles may be much more ambivalent than ours, and if you asked one of the numerous hermaphrodite species that have both male and female gonads (Anthes this volume), it might find it hard to understand the concept. The scientific concept of sex roles (in species with separate sexes) refers to the overall bias in the animal kingdom that females are more likely to provide parental care whereas males are more likely to compete for access to females and mating opportunities. Thus, sex roles may be considered as ‘reversed’ if females compete more intensely for access to males and males provide more parental care than females (Kokko and Jennions 2008). An example are species with a mating system termed classical polyandry, where females compete more intensely over mates than males over females, and a female mates with more than one male. In addition, each male raises the offspring with little or no help from the female (Andersson 2005). Classical polyandry is rare but occurs for instance in insects, fish, amphibians and birds. This mating system offers not only a unique possibility to test hypotheses about sexual selection and breeding systems, but also a good model to study the hormonal control of sex role reversed behaviour. Testosterone is involved in the modulation of aggression in a reproductive context in many male vertebrates. A naïve assumption would be that a reversal in sex roles comes with a reversal in hormone concentrations that regulate sex specific traits, i.e. females should show higher testosterone levels than males or males higher estradiol levels than females. Although initial results suggested that this might be the case (Höhn and Cheng 1967), all recent studies (summarised by Mayer et al. 1993, Eens and Pinxten 2000, Goymann and Wingfield 2004b), including the pioneering work by Rissman and Wingfield (1984) and Fivizzani and Oring (1986), did not find unusual sex steroid patterns in females and males of sex role reversed species. Alternatively, other hormones may modulate female aggression, or the ability of target tissues to receive a hormonal signal (i.e. the amount and density of receptors for the respective hormone) may be particularly high, or may increase during the breeding period. Without increasing peripheral hormone concentration, females of sex role reversed species might simply increase sensitivity of target cells by increasing the number of neurons and/or the number of receptors per neuron sensitive to testosterone. Recent studies of African black coucals (Centropus grillii, Fig. 16.5a, b), a classically polyandrous cuckoo species, suggested that both mechanisms may be important. Female black coucals are the more competitive sex, singing and defending territories, and are about 70% larger than males. They lay clutches for up to four males, each of which attends his own nest within the territory of one

Mating systems, social behaviour and hormones

477

female. Males rarely sing and do not defend territories, but build the nest, incubate the eggs and feed the young without help from the female (Vernon 1971, Goymann et al. 2004a, 2005). Female black coucals have lower testosterone levels than males. But females express high levels of androgen receptor mRNA in the Nucleus taeniae, a brain region that is involved in the regulation of social behaviour in birds. Hence, this nucleus appears to be more sensitive to testosterone in females than in males (Voigt and Goymann 2007). During the mating period, when females sing more often and territorial conflicts are intense, their testosterone levels are increased. Thus, the expression of mating and agonistic behaviour in females may be related to circulating testosterone levels, yet absolute concentrations may still be below those of males. Other hormones may play a role, too. When exposed to a simulated territorial intrusion (done by placing a stuffed female into a female’s territory and playing back conspecific song, Fig. 16.5b), females showed a decrease in progesterone levels. And when progesterone concentrations were experimentally elevated, females became less aggressive (Goymann et al. 2008). Because progesterone apparently suppresses territorial aggression, a decrease in progesterone concentrations during territorial interactions might be a mechanism to sustain high levels or persistence of aggression in female black coucals. Progesterone had a similar suppressive effect on aggression induced by estradiol and testosterone in female rats (Albert et al. 1992). Next to nothing is known about hormone patterns and manipulations of hormones during the ontogeny of any sex role reversed species, when many of the characteristics of sex role reversed species may be shaped. a

b

Fig. 16.5 a) Female black coucal singing on a perch. b) Female black coucal attacking a stuffed ‘dummy’. Owing to the vigour of the attack the dummy was dislocated from its perch. Photo © Wolfgang Goymann

478

Wolfgang Goymann and Heribert Hofer

The balance between costs and benefits of the effects of permanently high levels of testosterone is assumed to differ between monogamous and polygynous species. Socially monogamous species with a high degree of male parental care are predicted to show an increase in androgens from Level B to Level C (the androgen responsiveness) only during periods of territory establishment, acute male-male challenges, or when females are fertile, so that paternal care is not compromised. In other words: these species express a high androgen responsiveness because they show a strong rise in testosterone when challenged (Fig. 16.4). The same is true for classically polyandrous species in which males provide exclusive parental care (see also BOX 16.2). In contrast, androgen levels in polygynous species with little or no paternal care should be close to the maximum Level C throughout the breeding season because males interact continuously and intensely with each other and with receptive females. Hence, in such a competitive environment the benefits of permanently high levels of testosterone may outweigh associated costs. Thus, polygynous species without paternal care express a low androgen responsiveness, because their testosterone levels are high throughout the breeding season. Polygynous species in which males contribute to parental duties at the nest, however, show an androgen responsiveness intermediate between those two extremes. Interspecific comparisons of seasonal androgen patterns in birds and fish confirmed the existence of a relationship between mating system and androgen responsiveness (Hirschenhauser et al. 2003, Hirschenhauser and Oliveira 2006). The data suggest that socially monogamous species show a larger difference between Level B and Level C testosterone concentrations than polygynous species. This difference seems to be based mainly on lower Level B values in socially monogamous than polygynous species rather than on variation of Level C testosterone concentrations (but see below). In these comparative studies the influence of paternal care on androgen responsiveness persisted only in passerine birds: in passerines different levels of androgen responsiveness probably evolved in response to changes in the male’s paternal contribution during the incubation phase (Hirschenhauser et al. 2003). Thus, it is possible that costs unrelated to paternal care associated with high levels of testosterone drive this relationship between mating system and androgen responsiveness. One direct and potentially costly effect of testosterone may be that permanently high levels may enhance the likelihood of escalating conflict behaviour. This may lead to inappropriate expression of aggressive or risktaking behaviour, and, hence, increase the risk of injury or predation. Thus, if permanent high levels of testosterone do not pose an immediate advantage, it may be better to keep them near the breeding baseline and to elevate testosterone only when mating opportunities arise.

Mating systems, social behaviour and hormones

479

Recent findings – again in birds – suggest that some refinements of the Challenge Hypothesis or its interpretation may be necessary in the future. For example, testosterone does not universally suppress paternal behaviour in passerine birds. According to the Essential Paternal Care Hypothesis, males of some passerine bird species become insensitive to the suppressive effect of testosterone on paternal behaviour (Lynn 2008). Behavioural insensitivity should mainly occur in species in which offspring survival is severely hampered should the male stop feeding the young. Another issue is related to a methodological discrepancy between the predictions and the supporting data of the Challenge Hypothesis. The predictions of the Challenge Hypothesis, namely the increase from Level B to Level C, relate to a situational increase of testosterone during agonistic interactions between males. In other words, it predicts a behavioural effect of social challenges on testosterone secretion. But the support for the Challenge Hypothesis is mainly based on seasonal patterns of testosterone: Level C concentrations of testosterone are determined during the periods of territory establishment or mate-guarding and Level B concentrations are determined during the rest of the breeding season, assuming that the higher levels during territory establishment and mate-guarding are the result of these social interactions. During experimental inductions of situational male-male interactions, more than half of the bird species investigated so far did not show the expected increase in testosterone (reviewed in Goymann 2009). This discrepancy led to the distinction (Goymann et al. 2007) between the seasonal androgen response (Rseason, based on seasonal profiles of Level B and Level C testosterone concentrations) and the androgen responsiveness to male-male interactions (Rmale-male, based on testosterone concentrations measured during experimental inductions of territorial conflicts between males compared to those during control situations). Data based on Rseason broadly support the predictions of the Challenge Hypothesis (Wingfield et al. 1990, 2000, Hirschenhauser et al. 2003), but data based on situational Rmale-male are ambiguous (Goymann et al. 2007, Goymann 2009). Future work has to show whether the lack of androgen responsiveness to male-male challenges (Rmale-male) in many species and the discrepancy between Rseason and Rmale-male is related to specific ecological factors, which then need to be incorporated into refinements of the Challenge Hypothesis. Alternatively, the seasonal androgen response (Rseason) may not be entirely caused by social stimulation from competing males or receptive females. If so, additional, potentially intrinsic factors need to be considered in a modified hypothesis. In summary, there is substantial evidence that the androgen responsiveness based on seasonal testosterone profiles (Rseason) is related to mating systems in birds: males of socially monogamous species are more likely to

480

Wolfgang Goymann and Heribert Hofer

show large androgen responses, whereas males of polygynous species are more likely to express small androgen responses. The disparity between males of monogamous and polygynous species is most likely based on differences in the benefits and costs of the effect of persistently high levels of testosterone. This view gains further support when looking at maximal levels of testosterone, rather than the dynamics of testosterone or androgen responsiveness. A good predictor of maximum testosterone levels was the rate of extra-pair mating behaviour: bird species with higher rates of extrapair paternity expressed higher peak levels of testosterone (Garamszegi et al. 2005). This fits with another observation in a meta-analysis of all vertebrate groups, that – independent of mating system – high testosterone levels were strongly associated with high frequencies of sexual behaviour, with the largest effect observed in vertebrates without paternal care, in particular mammals (Hirschenhauser and Oliveira 2006). In contrast to androgen responsiveness, maximum testosterone levels are not related to the social mating system in birds (Goymann et al. 2004b, Garamszegi et al. 2008). Rather, length of breeding season and latitude predict testosterone concentrations: when breeding seasons are short and latitude is high, maximum testosterone levels are high (Goymann et al. 2004b, Garamszegi et al. 2008). Because birds at high latitudes tend to have short breeding seasons, latitude may be considered a proxy for the length of the breeding season. For other vertebrate groups comparable analyses are currently unavailable. Of interest are also the known exceptions to the rules proposed by the Challenge Hypothesis. For instance, in cooperatively breeding mammals, such as the dwarf mongoose (Helogale parvula) and the African wild dog (Lycaon pictus), androgen levels did not, as predicted, increase during the breeding season in either dominant or subordinate males, despite increased male-male aggression among dwarf mongooses (Creel et al. 1993, 1997). In African wild dogs, rates of aggression among male pack members declined during mating periods, even though some contests escalated and resulted in wounding (Creel and Creel 2002). Both species are characterised by reproductive suppression, a factor that played no role in the original formulation of the Challenge Hypothesis (Wingfield et al. 1990). 16.3.2 Experimental manipulations: is testosterone a proximate factor controlling individual mating tactics? The previous section considered comparative data regarding testosterone and mating systems, suggesting that there is a link between the two. Is there also evidence that testosterone is a factor involved in the proximate

Mating systems, social behaviour and hormones

481

regulation of male mating tactics? One way to investigate this is to experimentally manipulate testosterone concentrations in individual males and see whether this affects their mating tactics. Most experimental investigations of this kind have been done with birds. Testosterone implants increased the likelihood of males becoming polygynous or showing extra-pair behaviour in some bird species (Wingfield 1984, Moss et al. 1994, Raouf et al. 1997, De Ridder et al. 2000, Davis 2002, Reed et al. 2006), but not others (Silverin 1980, Beletsky et al. 1995, Stoehr and Hill 2000, Cordero et al. 2003, Foerster and Kempenaers 2004). Studies in lizards demonstrated that hormones may be involved in the proximate control of mating strategies, but species may differ with respect to the life history stage at which hormones influence mating tactics (BOX 16.1). Studies that manipulate mating tactics using hormone implants and then measure reproductive success face a particular problem. Long-term treatment with external testosterone may lead to a shut-down of testicular testosterone and sperm production. This pharmacological effect creates a potential problem for all studies that use testosterone treatment to manipulate mating behaviour and measure potential fitness consequences of this treatment such as the number of extra- and within-pair young. Any loss of paternity by testosterone-treated males compared to controls is likely to be an artefact. Hence, testosterone treatment is an efficient means to manipulate and study extra-pair behaviour, but effects on genetic paternity should be interpreted with caution. A frequent question in the discussion of the proximate control of mating tactics is: if high levels of testosterone support extra-pair behaviour or increase behavioural traits that induce polygyny, why do not all males boost their testosterone levels to maximise their reproductive success? There are several reasons why this is not the case. From a male perspective, reproductive success can be maximised by either emphasising pre-mating investment (such as sperm production or courtship behaviour) or post-mating investment (such as parental care). There is no a priori reason to assume that males should automatically put more emphasis on the former (Kokko and Jennions 2008, Schneider and Fromhage this volume). Also, high levels of testosterone may produce costs, such as an increased risk of injury or mortality, an increase in energy expenditure or a reduction in fat stores (Wingfield et al. 2001). There is also the issue of the potentially suppressive effect of testosterone on the immune system, a controversial topic (see Roberts et al. 2004), since there is currently more robust evidence that the immune system suppresses testosterone rather than vice versa (Boonekamp et al. 2008). Hence, there are

482

Wolfgang Goymann and Heribert Hofer

several reasons why males might want to keep levels of this anabolic hormone low, unless the benefits outweigh the associated costs. The female perspective is an important but largely neglected part in the discussion of proximate control factors of mating strategies: it takes two to tango and it takes even more to be polygamous. High levels of testosterone may increase the propensity of a male to seek additional mates. However, if male and female interests diverge (e.g. in the context of sexual conflict, Gowaty 1996), then male success will not improve with treatment if it is not in the interest of the female. There is a large body of literature that discusses factors that might influence the decision of females to become the secondary mate of a male. Such factors include male quality, territory quality or availability of unpaired males (Orians 1969, Krebs and Davis 1997). Currently, we know very little about the physiology of these females. Potential hormonal factors that may lead to or prevent the formation of a pair-bond between a female and a male are discussed in the next section. 16.3.3 The role of oxytocin and arginine vasopressin in pairbonding and mating systems: voles as a model In recent years, voles have become important model systems to understand the role of peptide hormones in the neuroendocrine regulation of mating decisions. The prairie vole (Microtus ochrogaster) is a socially monogamous species in North America. A female and a male form a lifetime pairbond and maintain a common nest and territory that they defend against other voles. They live in communal family groups consisting of the breeder pair and their offspring, which often function as ‘helpers at the nest’ (Trillmich this volume). In contrast, montane voles (Microtus montanus) and meadow voles (Microtus pennsylvanicus) are much less social: males and females have separate territories and nests, and only meet for mating. In these two species, males and females are highly promiscuous and parental care is only provided by the female (Carter et al. 1997). The most important modulators of pair-bonding and social behaviour in voles were discovered to be the two hormones oxytocin and arginine vasopressin. These hormones belong to an ancient peptide hormone family present in all vertebrates. In mammals, oxytocin released from the pituitary gland plays an important role during parturition and lactation as it stimulates contraction of the uterus and initiates the milk let-down from mammary glands (Bentley 1998). Pituitary arginine vasopressin reduces urinary water loss by increasing osmotic re-absorption of water from kidney tubules (Bentley 1998). Apart from their role as hormones acting in the pe-

Mating systems, social behaviour and hormones

483

riphery, both hormones also act as neuromodulators within the brain. Central oxytocin and arginine vasopressin play an important role in the regulation of affiliative and aggressive behaviours. For instance, central oxytocin is an essential component for mother-infant bonding, grooming and sexual behaviour. Central arginine vasopressin has been implicated in male social behaviours, including courtship and pair-bonding but also aggression and territorial behaviour (Nelson 2005). Typically, socially monogamous prairie voles form a pair-bond as a consequence of intense mating during a 24 hour period. Because mating results in central oxytocin release in mammals, it is likely that this mating activity stimulates oxytocin release in prairie voles and facilitates pairbond formation of female prairie voles. Indeed, oxytocin injections into the ventricle of the brain of unmated female prairie voles facilitated pairbonding (Insel and Hulihan 1995). In male prairie voles, arginine vasopressin rather than oxytocin mediated this effect: central arginine vasopressin is released during mating and its central injection facilitated pairbonding in unmated males (Winslow et al. 1993). The prairie vole brain expresses high levels of oxytocin receptors and arginine vasopressin receptors (of the V1aR subtype) in brain regions that are involved in the dopamine reward and reinforcement circuits, brain areas that play a role in addiction (Insel and Shapiro 1992, Insel et al. 1994). The release of oxytocin and arginine vasopressin upon mating in prairie voles activates the dopaminergic reward pathway in the brain and makes mating with a particular partner rewarding and presumably hedonic in monogamous prairie voles. Such a reinforcement effect may become associated with the identity of the mate, resulting in a conditioned partner preference (which does not prevent them from seeking extra-pair matings, Ophir et al. 2008). Blocking the dopamine receptors in these reward areas prevented the formation of partner preferences (Young and Wang 2004). In contrast to prairie voles, promiscuous montane voles do not form partner preferences and have few oxytocin and arginine vasopressin receptors in these brain reward regions. Thus, although sex is likely to be rewarding for montane voles, too, the reward stimulus is unlikely to be associated with the identity of a particular partner, thus preventing the formation of a partner preference. The comparison of the brains of several vole species suggests that the distribution of oxytocin and arginine vasopressin receptors differ between monogamous and promiscuous species: monogamous pine voles (Microtus pinetorum) are similar to monogamous prairie voles, whereas promiscuous meadow voles resemble promiscuous montane voles (Insel and Shapiro 1992, Insel et al. 1994).

484

Wolfgang Goymann and Heribert Hofer

BOX 16.3 Social behaviour, hormones and genitals in spotted hyenas Spotted hyenas (Crocuta crocuta) display female dominance, masculinisation of female genitalia (virilisation) and sibling rivalry that is sometimes fatal (siblicide). The joint appearance of these unusual traits has been repeatedly cited as evidence for fundamental links between androgens, the masculinisation of female genitalia and aggressive behaviour (Gowaty 1997, Hrdy 1999). Spotted hyenas are therefore an interesting model to discuss the evolution of sex-specific aggressive behaviour and its hormonal foundations. The prevailing assumption was that it must be elevated levels of androgens that produce large, hyperaggressive females and that only such females can outcompete males in feeding situations and acquire absolute dominance over males (Gould 1981, Glickman et al. 1993, Frank et al. 1995, Frank 1997, Hrdy 1999). Female spotted hyenas have an enlarged clitoris that is erectile and resembles a penis (Matthews 1939, Fig. 16.6 and 16.7). This female ‘pseudo-penis’ would have evolved as a by-product of selection for androgenised, aggressive females because virilisation of female genitalia is assumed to be controlled by androgens (Gould 1981, Hamilton et al. 1986, Frank et al. 1991, Frank 1997). Neonatal siblicide in spotted hyenas would then be a by-product of selection for androgenised, aggressive females (Frank et al. 1991). In the past 20 years, substantial progress has been made on these issues. Importantly, adult females have testosterone (T) and 5α-dihydrotestosterone (DHT) concentrations of an order of magnitude lower than those of males sampled before they dispersed or reproductively active males who had immigrated into a new clan (Goymann et al. 2001a). Female androstenedione (AE) concentrations were similar for females and post-dispersal males. In mammals where females are not virilised, AE levels are sometimes [brown hyena (Parahyaena brunnea), Racey and Skinner 1979] but not always (domestic dogs, humans: Feder 1985, Cashdan 1995, Feldman and Nelson 1996) lower in females than males. Total levels of circulating androgens (T, DHT, AE) in adult female spotted hyenas were therefore substantially lower than those of post-dispersal males. As androgen levels in spotted hyenas conform to the typical mammalian pattern, it is unlikely that female dominance is a consequence of selection for androgen-facilitated social dominance. This conclusion is consistent with another argument. Owing to specific enzymatic activity of the placenta, spotted hyena fetuses experience high levels of androgens (Lindeque and Skinner 1982, Yalcinkaya et al. 1992, Licht et al. 1992, 1998). These levels were not only supposed to be responsible for the virilisation of female genitalia (disproved in the meantime, see Drea et al. 1998, Glickman et al. 2006) but also for ‘enhanced’ female aggressiveness during adulthood (Glickman et al. 1992, 1993, 2006, Frank 1997). If so, androgens should lead to ‘enhanced’ aggressiveness in both

Mating systems, social behaviour and hormones

485

adult males and females (East et al. 1993), as females and males experience the same prenatal maternal environment, and circulating androgen levels are much higher in male than female fetuses (Browne et al. 2006). However, rates of aggression among post-dispersal males are low and these males typically submit to females (Frank 1986, East et al. 1993, East and Hofer 2001). Adult pre-dispersal males had significantly lower testosterone levels than post-dispersal males (Frank et al. 1985, Holekamp and Smale 1998, van Jaarsveld and Skinner 1991, Goymann et al. 2001a), yet pre-dispersal males dominate post-dispersal males during feeding and in social encounters, win all agonistic interactions with them, and show higher hourly rates of aggression (Holekamp and Smale 1998). This is in contrast to many other species (Holekamp and Smale 1998) where winners of contests have higher testosterone levels than losers, and more aggressive individuals have higher levels than less aggressive conspecifics. These data are consistent with the hypothesis that, in males, testosterone is involved in aggression related to reproduction, whereas aggression in a non-reproductive context may be independent of circulating testosterone (Schwabl and Kriner 1991, Wingfield et al. 1994, 1997, Canoine and Gwinner 2002). If androgen levels in spotted hyenas conform to the typical mammalian pattern, which evolutionary forces are responsible for female virilisation and dominance? Female virilisation could be a by-product of selection for precocial aggressive cubs of both sexes (East et al. 1993, Hofer and East 1995). In contrast to the hypothesis of androgen-facilitated social dominance (Gould 1981, Frank 1986, Hamilton et al. 1986, Glickman et al. 1993), this hypothesis neither assumes a different mechanism for the action of androgens during fetal development in females and males, nor enhanced aggressiveness in adult females. It is also consistent with current theories of facultative siblicide (Mock and Parker 1997, Hofer and East 2008). Female dominance is most likely a function of matrilineal associations, coalitions between females, and the lack of aggressiveness in males (East et al. 1993). Possible selection for active submission by post-dispersal males to females (East et al. 1993) and lack of aggression among post-dispersal males deserve further attention.

In a sophisticated experiment, viral vectors were used to over-express the Avpr1a gene, the gene that encodes the arginine vasopressin receptor V1aR, in the reward circuit of male promiscuous meadow voles. Unlike control males, these transgenic males showed an increased partner preference towards a female they were cohabitated with. If dopamine receptors were blocked in the reward areas of these transgenic animals, the formation of a partner preference, as expected, was prevented (Lim et al. 2004). What kind of mutation could have led to the different expression of arginine vasopressin receptors in the brains of monogamous and promiscu-

486

Wolfgang Goymann and Heribert Hofer

Fig. 16.6 Spotted hyena mating with the male’s penis clearly visible. Photo © Marion East and Heribert Hofer

Fig. 16.7 A young female spotted hyena with erected ‘pseudo-penis’. Photo © Marion East and Heribert Hofer

Mating systems, social behaviour and hormones

487

ous voles? The Avpr1a gene is highly homologous in prairie and montane voles. Upstream of the transcription start site, the prairie vole gene contains a highly repetitive sequence, a microsatellite. In montane voles this microsatellite sequence is much shorter. Transgenic mice with the prairie vole Avpr1a gene that included the long microsatellite sequence expressed the gene in similar brain regions as in prairie voles and, unlike control voles, showed a strong partner preference (Young et al. 1999). Thus, the change in the microsatellite region of the Avpr1a gene has been hypothesised to be the molecular event which induced a change in the expression of the arginine vasopressin receptor in the reward circuit of the vole brain. This change resulted in the biological potential to develop a conditioned partner preference. The vole example raises the exciting possibility that a relatively minor mutation in a single gene can induce major changes in social behaviour such as pair-bonding (Young and Wang 2004) and form the mechanistic basis for the transformation of a promiscuous vole into a monogamous one or vice versa. If this change in social behaviour is accompanied by an increase in Darwinian fitness, it has the potential to spread rapidly in a population.

16.4 Group-living, social status and hormones Animals can be roughly classified as living a solitary life, in pairs or in groups. Group-living has major advantages, such as cooperation and social support, but also disadvantages that arise from social conflict and competition between group members (Alexander 1974, Bertram 1978, Walters and Seyfarth 1987, East and Hofer in press, Kotrschal et al. this volume). Animal groups often form dominance hierarchies, which may result in inequalities with respect to access to resources and mating partners. This differential access can drastically affect quality of life. It has often been assumed that success in agonistic interactions and the achievement of high dominance status is a function of morphological and physiological characteristics (BOX 16.3). Recent studies demonstrated that, particularly in complex mammalian societies, there are many routes to high social status, and a large body size or high testosterone concentration may be of little relevance (BOX 16.3 and BOX 16.4). Social status and quality of life are often studied in the context of stress. Affiliative behaviour can reduce social stress in group-living animals (e.g. Hennessy et al. 2009, Kotrschal et al. this volume) and agonistic behaviour can increase social stress (e.g. Sapolsky 2002, Kotrschal et al. this volume). Psychosocial stressors within a group may activate the endocrine

488

Wolfgang Goymann and Heribert Hofer

BOX 16.4 Determinants of social status and hormones Social status refers to the position of an individual in a society. Traditionally, social status is scored as a rank by recording an individual’s submissive response to group members during pair-wise (dyadic) interactions. It is often assumed that large animals and/or those with high testosterone concentrations achieve high status. In many societies, however, status depends on other factors. As few studies record these factors along with body size or peripheral testosterone concentrations, it is unclear how these factors interact in the establishment of social status. Social context. Rank may depend on the social context of an interaction. In the domestic cat (Felis catus) males dominate females in most social contexts, yet females dominate males during feeding (Bonanni et al. 2007). Here, female dominance probably results from pay-off asymmetries in relation to the value of food and the cost of losing. Males are likely to suffer fitness costs by winning contests over food if females invest in the male’s offspring (Bonanni et al. 2007). This argument may also explain female dominance in lemurs (Young et al. 1990) and is consistent with the finding that in such species, testosterone concentrations of subordinate males are an order of magnitude higher than in females (Goymann et al. 2001a). Social queues. In some societies, social status or access to breeding territories or mates is determined by the sequence in which animals join a group or locate a receptive female: The first arrival has the highest rank, later arrivals successively lower ranks, and animals advance in status only when higherranking animals drop out. As a queue is not based on any individual trait, it is an arbitrary convention and in theory inherently unstable because it is open to cheating (Maynard Smith 1982). Yet the longest observed social queue of more than 20 immigrant males in spotted hyenas is also the most stable (East and Hofer 2001). Queues are widespread (Schwagmeyer and Parker 1990, Voigt and Streich 2003) as males often queue for territories on leks (Kokko et al. 1998, Kokko and Johnstone 1999) and subordinates of both sexes queue for breeding positions in societies with reproductive suppression. Recruitment of support. Contestants may recruit support from other group members in short-term coalitions or long-term alliances. The winner of a contest is likely to be the contestant who recruits a larger number of supporters (Bygott et al. 1979, Maynard Smith 1982, Hofer and East 1993). Many social mammals have a matrilineal group structure where females stay within their natal group and males disperse (Greenwood 1980). Matrilineal structures favour the recruitment of support amongst females (Harcourt and de Waal 1992, Silk et al. 2003), and permit coalitions of subordinate females to successfully challenge otherwise dominant males (Smuts 1986).

Mating systems, social behaviour and hormones

489

Rank inheritance. In some primates, carnivores and rodents, offspring acquire a social status close to that of a parent, a phenomenon termed rank inheritance (Hausfater et al. 1982). Females provide the environment in which offspring develop and therefore have a considerable effect on the expression of offspring phenotype. These ‘maternal effects’ are mediated through interconnected pathways and can have genetic, physiological/hormonal and behavioural components (Bernardo 1996, Rossiter 1996). In birds, crossfostering experiments disentangled maternal effects on the emergence of offspring phenotype (Griffith et al. 1999, Verboven et al. 2003). In mammals, a recent study in the spotted hyena used natural cases of adoption by surrogate mothers to shed light on the mechanisms that lead to rank inheritance (East et al. 2009). At adulthood, daughters usually obtain a social status similar to and below that of their mother (Holekamp and Smale 1991, Engh et al. 2000), as do non-reproductive sons before they emigrate or become reproductively active (Smale et al. 1993, East and Hofer 2001, Hofer and East 2003). When reaching adulthood, adopted offspring did not gain a social status similar to that held by their genetic mother on that date as predicted by the genetic pathway (Moore 1993, Moore et al. 1997). Adopted offspring also did not obtain a rank at adulthood close to that of their genetic mother at the end of gestation as predicted by the endocrine pathway (Dloniak et al. 2006). Adopted cubs obtained a rank at adulthood close to and below that of their surrogate mother (East et al. 2009), consistent with the idea that the emergence of offspring social status is predominantly influenced by maternal behavioural support (Hausfater et al. 1982, Horrocks and Hunte 1983, Hofer and East 2003). Knowledge of third-party relationships. Individuals in some species monitor social interactions to gain information about the relative competitive ability of conspecifics. If an animal recognises the relationship between two other animals it is said to recognise ‘third-party relationships’. It was initially suggested for primates (Tomasello and Call 1997) and has been debated ever since (e.g. Range and Noë 2005). Clear evidence for it comes from play-back experiments when female by-standers looked towards the mother whose infant’s distress screams were experimentally broadcast (Cheney and Seyfarth 1980, 1999, Bergman et al. 2003).

stress response including the hypothalamic pituitary adrenocortical (HPA) axis (Fig. 16.8). The stressor induces the release of corticotropin releasing hormone (CRH) from the hypothalamus, leading to a release of adrenocorticotropic hormone (ACTH) from the pituitary, which stimulates the production of corticosteroids in the cortex of the adrenal gland. Corticosteroids are known as ‘stress hormones’ and function to mobilise energy and tissue nitrogen, increase cardiovascular tone and regulate the immune system (von Holst 1998, Sapolsky 2002). Short-term elevations of corticos-

490

Wolfgang Goymann and Heribert Hofer

Environmental information

1

Brain Hypothalamus CRF

Anterior pituitary

3

ACTH

Gonad

4

2

Adrenal

CBG

Total plasma glucocorticoids

Active glucocorticoids Gluconeogenesis mating success

Pathological states immunosuppression

Fig. 16.8 The basic components of the adrenocortical response system (the HPAaxis). Adrenocorticotropic hormone (ACTH) is secreted by the anterior pituitary, which in turn stimulates the production of glucocorticoids in the adrenals. The biologically active fraction of glucocorticoids is the fraction of total plasma glucocorticoids that remains after most glucocorticoids are bound to corticosteroid binding globulins (CBG). Negative feedback of glucocorticoids on the anterior pituitary regulates the rate of production of ACTH. The actions of glucocorticoids may contribute fitness benefits (gluconeogenesis) and fitness costs (pathological changes in organs and immunosuppression). Comparisons between species suggest that natural selection has tuned the system in different ways in at least four components. (1) The hypothalamus may change sensitivity to environmental stimuli. (2) Negative feedback is either present or switched off. (3) Adrenal activity and gonadal activity may be negatively coupled or switched off. (4) Gonadal activity may either increase or decrease the rate of production of CBGs. Green arrows: positive effects; red arrows: negative effects; blue arrows: feedback switched off. Drawing modified after Hofer and East (1998).

teroids lead to adaptive behavioural and physiological processes (Landys et al. 2006), including the initiation of the emergency life history stage (Wingfield et al. 1998). Short-term elevations of corticosteroids suppress non-essential life processes to redirect efforts towards survival and recovery (Landys et al. 2006), but chronic elevations of corticosteroids may cause reproductive failure and disease (Sapolsky 2002). The precise feed-

Mating systems, social behaviour and hormones

491

back mechanisms are not necessarily uniform and may vary across species (Fig. 16.8), revealing a remarkable fine-tuning of hormonal processes to different reproductive life histories (Hofer and East 1998). Traditionally, subordinate animals have been considered to be more physiologically stressed, to show a chronic expression of higher levels of corticosteroids, and, thus, to be more likely to develop stress-related pathologies than dominants (reviewed in von Holst 1998). This view has been challenged by arguing that in free-living animals, dominants experience more psychosocial stress and often express higher levels of corticosteroids (Creel 2001). When looking at social rank and relating it to psychosocial stress, it is important to consider that rank as a descriptor of social status does not specify the process by which such a status is obtained and maintained. Thus, behavioural strategies related to status acquisition and maintenance, rather than social status per se may determine the degree of social stress (von Holst 1998, Goymann and Wingfield 2004a, Sapolsky 2005). For instance, there are social groups in which dominance rank is inherited or dominance is achieved via queuing conventions (BOX 16.4). In such cases, dominants are rarely challenged by subordinates, and there is little need for dominants to threaten subordinates. In many cooperativelybreeding birds and mammals, a single dominant pair breeds and other adult group members assist in the rearing of offspring. Reproductive suppression is here a natural condition experienced by most members of a population at some stage of their life cycle. In such species, specific neuroendocrine mechanisms are the prime cause of reproductive suppression, whereas the general adrenocortical response plays no role or only a minor one (Faulkes and Abbott 1997, Hofer and East 1998). Hence, as long as there are no acute challenges, rank-related patterns of corticosteroid levels are unlikely to be found in such societies (Goymann and Wingfield 2004a, Sapolsky 2005). Also, affiliative relationships or other coping outlets for subordinates can buffer rank-induced psychosocial stress. In contrast, if dominance is acquired or maintained via low-level or overt aggression and subordinates are frequently threatened by dominants, then a rank-related pattern of acute and chronic elevations of corticosteroid levels in subordinates may emerge (e.g. Goymann et al. 2001b). This is more likely in animal societies with strong group cohesion than in fission-fusion societies (Goymann and Wingfield 2004a, Sapolsky 2005). Alternatively, if rank hierarchies are non-linear and/or if dominants are frequently challenged by subordinates, the opposite pattern may emerge – dominants may experience chronic elevation of corticosteroid levels (Creel 2001). Thus, it may not be the dominant or subordinate position itself, but the process by which rank is acquired and maintained which determines the

492

Wolfgang Goymann and Heribert Hofer

physiological and psychosocial effects associated with a given social status (Sapolsky 1992, Abbott et al. 2003, Goymann and Wingfield 2004a).

References Abbott DH, Keverne EB, Bercovitch FB, Shively CA, Mendoza SP, Saltzman W, Snowdon CT, Ziegler TE, Banjevic M, Garland T, Sapolsky RM (2003) Are subordinates always stressed? A comparative analysis of rank differences in cortisol levels among primates. Horm Behav 43:67-82 Adkins-Regan E (2005) Hormones and Animal Social Behavior: Monographs in Behavior and Ecology. Princeton University Press, Princeton Albert DJ, Jonik RH, Walsh ML (1992) Interaction of estradiol, testosterone, and progesterone in the modulation of hormone-dependent aggression in the female rat. Physiol Behav 52:773-779 Alexander RD (1974) The evolution of social behavior. Annu Rev Ecol Syst 5:325-383 Andersson M (2005) Evolution of classical polyandry: three steps to female emancipation. Ethology 111:1-23 Archer J (2006) Testosterone and human aggression: an evaluation of the challenge hypothesis. Neurosci Biobehav Rev 30:319-345 Beach FA (1948) Hormones and Behavior. Paul B Hoeber, New York Beletsky LD, Gori DF, Freeman S, Wingfield JC (1995) Testosterone and polygyny in birds. Curr Ornithol 12:1-41 Bentley PJ (1998) Comparative Vertebrate Endocrinology. Cambridge University Press, Cambridge Bergman TJ, Beehner CJ, Cheney DL, Seyfarth RM (2003) Hierarchial classification by rank and kinship in baboons. Science 302:1234-1236 Bernardo J (1996) Maternal effects in animal ecology. Am Zool 36:83-105 Bernstein IS, Gordon TP, Rose RM (1983) The interaction of hormones, behaviour, and social context in nonhuman primates. In: Svare BB (ed) Hormones and Aggressive Behavior. Plenum Press, New York, pp 535-561 Berthold AA (1849) Transplantation der Hoden. Arch Anat Physiol Wiss Med 16:42-46 Bertram BCR (1978) Living in groups: predators and prey. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 64-96 Bonanni R, Cafazzo S, Fantini C, Pontier D, Natoli E (2007) Feeding-order in an urban feral domestic cat colony: relationship to dominance rank, sex and age. Anim Behav 74:1369-1379 Boonekamp JJ, Ros AHF, Verhulst S (2008) Immune activation suppresses plasma testosterone level: a meta-analysis. Biol Lett 4:741-744

Mating systems, social behaviour and hormones

493

Browne P, Place NJ, Vidal JD, Moore IT, Cunha GR, Glickman SE, Conley AJ (2006) Endocrine differentiation of fetal ovaries and testes of the spotted hyena (Crocuta crocuta): timing of androgen-independent versus androgendriven genital development. Reproduction 132:649-659 Bygott JD, Bertram BCR, Handy JP (1979) Male lions in large coalitions gain reproductive advantages. Nature 282:839-841 Caldwell GS, Glickman SE, Smith ER (1984) Seasonal aggression independent of seasonal testosterone in wood rats. Proc Natl Acad Sci USA 81:5255-5257 Canoine V, Gwinner E (2002) Seasonal differences in the hormonal control of territorial aggression in free-living European stonechats. Horm Behav 41:1-8 Carter CS, DeVries AC, Taymans SE, Roberts RL, Williams JR, Getz LL (1997) Peptides, steroids, and pair bonding. Ann NY Acad Sci 807:260-272 Cashdan E (1995) Hormones sex and status in women. Horm Behav 29:354-366 Catchpole CK, Slater PJB (2008) Bird Song. Biological Themes and Variations. Cambridge University Press, Cambridge Cheney DL, Seyfarth RM (1980) Vocal recognition in free-ranging vervet monkeys. Anim Behav 28:362-367 Cheney DL, Seyfarth RM (1999) Recognition of other individuals’ social relationships by female baboons. Anim Behav 58:67-75 Clutton-Brock TH, Harvey PH (1977) Primate ecology and social organization. J Zool Lond 183:1-39. Cordero PJ, Veiga JP, Moreno J, Parkin DT (2003) Extra-pair paternity in the facultatively polygynous spotless starling, Sturnus unicolor. Behav Ecol Sociobiol 54:1-6 Creel S (2001) Social dominance and stress hormones. Trends Ecol Evol 16:491497 Creel S, Creel NM (2002) The African Wild Dog: Behavior, Ecology and Conservation. Princeton University Press, Princeton Creel S, Wildt DE, Monfort SL (1993) Aggression, reproduction, and androgens in wild dwarf mongooses: a test of the challenge hypothesis. Am Nat 141:816825 Creel S, Creel NM, Mills MGL, Monfort SL (1997) Rank and reproduction in cooperatively breeding African wild dogs: behavioural and endocrine correlates. Behav Ecol 8:298-306 Davis ES (2002) Male reproductive tactics in the mallard, Anas platyrhynchos: social and hormonal mechanisms. Behav Ecol Sociobiol 52:224-231 DeNardo DF, Sinervo B (1994) Effects of steroid hormone interaction on activity and home-range size of male lizards. Horm Behav 28:273-287 De Ridder E, Pinxten R, Eens M (2000) Experimental evidence of a testosteroneinduced shift from paternal to mating behaviour in a facultatively polygynous songbird. Behav Ecol Sociobiol 49:24-30 Dloniak SM, French JA, Holekamp KE (2006) Rank related maternal effects of androgens on behaviour in wild spotted hyaenas. Nature 440:1190-1193

494

Wolfgang Goymann and Heribert Hofer

Drea CM, Weldele ML, Forger NG, Coscia EM, Frank LG, Licht P, Glickman SE (1998) Androgens and masculinization of genitalia in the spotted hyena (Crocuta crocuta). 2. Effects of prenatal anti-androgens. J Reprod Fertil 113:117127 East ML, Hofer H (2001) Male spotted hyenas Crocuta crocuta queue for status in social groups dominated by females. Behav Ecol 12:558-568 East ML, Hofer H (in press) Social environments, social tactics and their fitness consequences in complex mammalian societies. In: Székely T, Komdeur J, Moore AJ (eds) Social Behaviour: Genes, Ecology and Evolution. Cambridge University Press, Cambridge East ML, Hofer H, Wickler W (1993) The erect ‘penis’ is a flag of submission in a female-dominated society: greetings in Serengeti spotted hyaenas. Behav Ecol Sociobiol 33:355-370 East ML, Höner OP, Wachter B, Wilhelm K, Burke T, Hofer H (2009) Maternal effects on offspring social status in spotted hyenas. Behav Ecol 20:478-483 Eens M, Pinxten R (2000) Sex-role reversal in vertebrates: behavioural and endocrinological accounts. Behav Processes 51:135-147 Emerson SB, Hess DL (1996) The role of androgens in opportunistic breeding, tropical frogs. Gen Comp Endocrinol 103:220-230 Emlen ST, Oring LW (1977) Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223 Engh AL, Esch K, Smale L, Holekamp KE (2000) Mechanisms of maternal ‘rank inheritance’ in the spotted hyaena, Crocuta crocuta. Anim Behav 60:323-332 Faulkes CG, Abbott DH (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: Solomon NG, French JA (eds) Cooperative Breeding in Mammals. Cambridge University Press, Cambridge, pp 302-334 Feder HH (1985) Peripheral plasma levels of gonadal steroids in adult male and in adult, non-pregnant female mammals. In: Adler NT, Pfaff D, Goy RW (eds) Handbook of Behavioural Neurobiology, Vol 7. Plenum Press, New York, pp 299-370 Feldman EC, Nelson RW (1996) Canine and Feline Endocrinology and Reproduction. WB Saunders, Philadelphia Fivizzani AJ, Oring LW (1986) Plasma steroid hormones in relation to behavioural sex role reversal in the spotted sandpiper, Actitis macularia. Biol Reprod 35:1195-1201 Foerster K, Kempenaers B (2004) Experimentally elevated plasma levels of testosterone do not increase male reproductive success in blue tits. Behav Ecol Sociobiol 56:482-490 Folstad I, Karter AJ (1992) Parasites, bright males, and the immunocompetence handicap. Am Nat 139:603-622 Frank LG (1986) Social organisation of the spotted hyaena (Crocuta crocuta). II. Dominance and reproduction. Anim Behav 34:1510-1527 Frank LG (1997) Evolution of genital masculinization: why do female hyenas have such a large ‘penis’? Trends Ecol Evol 12:58-62

Mating systems, social behaviour and hormones

495

Frank LG, Davidson JM, Smith ER (1985) Androgen levels in the spotted hyena, Crocuta crocuta: the influence of social factors. J Zool Lond 206:525-531 Frank LG, Glickman SE, Licht P (1991) Fatal sibling aggression, precocial development, and androgens in neonatal spotted hyenas. Science 252:702-704 Frank LG, Weldele ML, Glickman SE (1995) Masculinization costs in hyaenas. Nature 377:584-585 Garamszegi LZ, Eens M, Hurtrez-Boussès S, Møller AP (2005) Testosterone, testes size, and mating success in birds: a comparative study. Horm Behav 47:389-409 Garamszegi LZ, Hirschenhauser K, Bókony V, Eens M, Hurtrez-Boussès S, Møller AP, Oliveira RF, Wingfield JC (2008) Latitudinal distribution, migration, and testosterone levels in birds. Am Nat 172:533-546 Glickman SE, Caldwell GS (1994) Studying natural behaviour in artificial environments: the problem of ‘salient elements’. In: Gibbons EF Jr, Wyers EJ, Waters E, Menzel EW Jr (eds) Naturalistic Environments in Captivity for Animal Behaviour Research. SUNY Press, Albany/NY, pp 197-216 Glickman SE, Frank LG, Pavgi S, Licht P (1992) Hormonal correlates of ‘masculinisation’ in female spotted hyenas (Crocuta crocuta). 1. Infancy to sexual maturity. J Reprod Fertil 95:451-462 Glickman SE, Frank LG, Holekamp KE, Smale L, Licht P (1993) Costs and benefits of ‘androgenization’ in the female spotted hyena: the natural selection of physiological mechanisms. Perspect Ethol 10:87-117 Glickman SE, Cunha GR, Drea CM, Conley AJ, Place NJ (2006) Mammalian sexual differentiation: lessons from the spotted hyena. Trends Endocrinol Metab 17:349-356 Gould SJ (1981) Hyena myths and realities. Nat Hist 90:16-24 Gowaty PA (1996) Battle of the sexes and origins of monogamy. In: Black JM (ed) Partnerships in Birds. Oxford University Press, Oxford, pp 21-52 Gowaty PA (1997) Feminism and Evolutionary Biology: Boundaries, Intersections and Frontiers. Chapman and Hall, New York Goymann W (2009) Social modulation of androgens in male birds. Gen Comp Endocrinol 163:149-157 Goymann W, Wingfield JC (2004a) Allostatic load, social status and stress hormones: the costs of social status matter. Anim Behav 67:591-602 Goymann W, Wingfield JC (2004b) Competing females and caring males: sex steroids in African black coucals, Centropus grillii. Anim Behav 68:733-740 Goymann W, East ML, Hofer H (2001a) Androgens and the role of female ‘hyperaggressiveness’ in spotted hyenas (Crocuta crocuta). Horm Behav 39:8392 Goymann W, East ML, Wachter B, Höner O, Möstl E, Van’t Hof TJ, Hofer H (2001b) Social, state-dependent and environmental modulation of faecal corticosteroid levels in free-ranging female spotted hyenas. Proc R Soc Lond B 268:2453-2459 Goymann W, Wittenzellner A, Wingfield JC (2004a) Competing females and caring males. Polyandry and sex-role reversal in African black coucals, Centropus grillii. Ethology 110:807-823

496

Wolfgang Goymann and Heribert Hofer

Goymann W, Moore IT, Scheuerlein A, Hirschenhauser K, Grafen A, Wingfield JC (2004b) Testosterone in tropical birds: effects of environmental and social factors. Am Nat 164:327-334 Goymann W, Kempenaers B, Wingfield J (2005) Breeding biology, sexually dimorphic development and nestling testosterone concentrations of the classically polyandrous African black coucal, Centropus grillii. J Ornithol 146:314324 Goymann W, Landys MM, Wingfield JC (2007) Distinguishing seasonal androgen responses from male-male androgen responsiveness: revisiting the challenge hypothesis. Horm Behav 51:463-476 Goymann W, Wittenzellner A, Schwabl I, Makomba M (2008) Progesterone modulates aggression in sex-role reversed African black coucals. Proc R Soc Lond B 275:1053-1060 Greenwood PJ (1980) Mating systems, philopatry and dispersal in birds and mammals. Anim Behav 28:1140-1162 Griffith SG, Owens IPF, Burke T (1999) Environmental determination of a sexually selected trait. Nature 400:358-360 Hamilton WJ, Tilson RL, Frank LG (1986) Sexual monomorphism in spotted hyenas, Crocuta crocuta. Ethology 71:63-73 Harcourt AH, de Waal FBM (1992) Cooperation and conflict: from ants to anthropoids. In: Harcourt AH, de Waal FBM (eds) Coalitions and Alliances in Humans and Other Animals. Oxford University Press, Oxford, pp 493-510 Hau M (2007) Regulation of male traits by testosterone: implications for the evolution of vertebrate life histories. BioEssays 29:133-144 Hausfater G, Altmann J, Altmann S (1982) Long-term consistency of dominance relations among female baboons (Papio cynocephalus). Science 217:752-755 Hennessy MB, Kaiser S, Sachser N (2009) Social buffering of the stress response: diversity, mechanisms, and functions. Front Neuroendocrinol 30: 470-482 Hirschenhauser K, Oliveira RF (2006) Social modulation of androgens in male vertebrates: meta-analyses of the challenge hypothesis. Anim Behav 71:265277 Hirschenhauser K, Winkler H, Oliveira RF (2003) Comparative analysis of male androgen responsiveness to social environment in birds: the effects of mating system and paternal incubation. Horm Behav 43:508-519 Hirschenhauser K, Taborsky M, Oliveira T, Canàrio AVM, Oliveira RF (2004) A test of the ‘challenge hypothesis’ in cichlid fish: simulated partner and territory intruder experiments. Anim Behav 68:741-750 Höhn EO, Cheng SC (1967) Gonadal hormones in Wilson’s phalarope (Steganopus tricolor) and other birds in relation to plumage and sex behavior. Gen Comp Endocrinol 8:1-11 Hofer H, East ML (1993) The commuting system of Serengeti spotted hyaenas: how a predator copes with migratory prey. II. Intrusion pressure and commuters’ space use. Anim Behav 46:559-574 Hofer H, East ML (1995) Virilised sexual genitalia as adaptations of female spotted hyaenas. Rev Suisse Zool 102:895-906

Mating systems, social behaviour and hormones

497

Hofer H, East ML (1998) Biological conservation and stress. Adv Stud Behav 27:405-525 Hofer H, East ML (2003) Behavioural processes and costs of co-existence in female spotted hyenas: a life history perspective. Evol Ecol 17:315-331 Hofer H, East ML (2008) Siblicide in Serengeti spotted hyenas: a long-term study of maternal input and cub survival. Behav Ecol Sociobiol 62:341-351 Holekamp KE, Smale L (1991) Dominance acquisition during mammalian social development: the ‘inheritance’ of maternal rank. Am Zool 31:306-317 Holekamp KE, Smale L (1998) Dispersal status influences hormones and behavior in the male spotted hyena. Horm Behav 33:205-216 Horrocks J, Hunte W (1983) Maternal rank and offspring rank in vervet monkeys: an appraisal of the mechanisms of rank acquisition. Anim Behav 31:772-782 Hrdy SB (1999) Mother Nature: Natural Selection and the Female of the Species. Random House, New York Huxley JS (1916) Bird-watching and biological science: some observations on the study of courtship in birds. Auk 33:142-161, 256-270 Insel TR, Hulihan TJ (1995) A gender-specific mechanism for pair bonding: oxytocin and partner preference formation in monogamous voles. Behav Neurosci 109:782-789 Insel TR, Shapiro LE (1992) Oxytocin receptor distribution reflects social organization in monogamous and polygamous voles. Proc Natl Acad Sci USA 89:5981-5985 Insel TR, Wang ZX, Ferris CF (1994) Patterns of brain vasopressin receptor distribution associated with social organization in microtine rodents. J Neurosci 14:5381-5392 Kappeler PM (1997) Intrasexual selection and testis size in strepsirhine primates. Behav Ecol 8:10-19 Kokko H, Jennions MD (2008) Parental investment, sexual selection and sex ratios. J Evol Biol 21:919-948 Kokko H, Johnstone RA (1999) Social queuing in animal societies: a dynamic model of reproductive skew. Proc R Soc Lond B 266:571-578 Kokko H, Lindström J, Alatalo RV, Rintamäki PT (1998) Queuing for territory positions in the lekking black grouse (Tetrao tetrix). Behav Ecol 9:376-383 Krebs JR (1977) Song and territory in the great tit Parus major. In: Stonehouse B, Perrins CM (eds) Evolutionary Ecology. Macmillan, London, pp 47-62 Krebs JR, Davies NB (1997) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford Krebs JR, Ashcroft R, Webber M (1978) Song repertoires and territory defence in the great tit. Nature 271:539-542 Landys MM, Ramenofsky M, Wingfield JC (2006) Actions of glucocorticoids at a seasonal baseline as compared to stress-related levels in the regulation of periodic life processes. Gen Comp Endocrinol 148:132-149 Licht P, Frank LG, Pavgi S, Yalcinkaya TM, Siiteri PK, Glickman SE (1992) Hormonal correlates of ‘masculinization’ in female spotted hyaenas (Crocuta crocuta). 2. Maternal and fetal steroids. J Reprod Fertil 95:463-474

498

Wolfgang Goymann and Heribert Hofer

Licht P, Hayes T, Tsai P, Cunha G, Kim H, Golbus M, Hayward S, Martin MC, Jaffe RB, Glickman SE (1998) Androgens and masculinization of genitalia in the spotted hyena (Crocuta crocuta). 1. Urogenital morphology and placental androgen production during fetal life. J Reprod Fertil 113:105-116 Lim MM, Wang Z, Olazábal DE, Ren X, Terwilliger EF, Young LJ (2004) Enhanced partner preference in a promiscuous species by manipulating the expression of a single gene. Nature 429:754-757 Lindeque M, Skinner JD (1982) Aseasonal breeding in the spotted hyaena (Crocuta crocuta, Erxleben) in southern Africa. Afr J Ecol 20:217-278 Lynn SE (2008) Behavioral insensitivity to testosterone: why and how does testosterone alter paternal and aggressive behavior in some avian species but not others? Gen Comp Endocrinol 157:233-240 Matthews LH (1939) Reproduction in the spotted hyaena (Crocuta crocuta, Erxleben). Philos Trans R Soc Lond B 230:1-78 Mayer I, Rosenqvist G, Borg B, Ahnesjö I, Berglund A, Schulz RW (1993) Plasma levels of sex steroids in three species of pipefish (Syngnathidae). Can J Zool 71:1903-1907 Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge Miles DB, Sinervo B, Hazard LC, Svensson EI, Costa D (2007) Relating endocrinology, physiology and behaviour using species with alternative mating strategies. Funct Ecol 21:653-665 Mock DW, Parker GA (1997) The Evolution of Sibling Rivalry. Oxford University Press, Oxford Moore MC (1991) Application of organization-activation theory to alternative male reproductive strategies: a review. Horm Behav 25:154-179 Moore MC, Hews DK, Knapp R (1998) Hormonal control and evolution of alternative male phenotypes: generalizations of models for sexual differentiation. Am Zool 38:133-151 Moss R, Parr R, Lambin X (1994) Effects of testosterone on breeding density, breeding success and survival of red grouse. Proc R Soc Lond B 258:175-180 Nelson RJ (2005) An Introduction to Behavioral Endocrinology. Sinauer Associates, Sunderland/MA Noë R, van Schaik CP, van Hooff JARAM (1991) The market effect: an explanation for pay-off asymmetries among collaborating animals. Ethology 87:97118 Oliveira RF (2005) Neuroendocrine mechanisms of alternative reproductive tactics in fish. Fish Physiol 24:297-357 Oliveira RF, Taborsky M, Brockmann HJ (2008) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge Ophir AG, Phelps SM, Sorin AB, Wolff JO (2008) Social but not genetic monogamy is associated with greater breeding success in prairie voles. Anim Behav 75:1143-1154 Orians GH (1969) On the evolution of mating systems in birds and mammals. Am Nat 103:589-603

Mating systems, social behaviour and hormones

499

Oyegbile TO, Marler CA (2005) Winning fights elevates testosterone levels in California mice and enhances future ability to win fights. Horm Behav 48:259-267 Phoenix CH, Goy RW, Gerall AA, Young WC (1959) Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology 65:369-382 Racey PA, Skinner JD (1979) Endocrine aspects of sexual mimicry in spotted hyaenas Crocuta crocuta. J Zool Lond 187:315-326 Range F, Noë R (2005) Can simple rules account for the pattern of triadic interactions in juvenile and adult female sooty mangabeys? Anim Behav 69:445-452 Raouf SA, Parker PG, Ketterson ED, Nolan V Jr, Ziegenfus C (1997) Testosterone affects reproductive success by influencing extra-pair fertilizations in male dark-eyed juncos (Aves: Junco hyemalis). Proc R Soc Lond B 264:1599-1603 Reed WL, Clark ME, Parker PG, Raouf SA, Arguedas N, Monk DS, Snajdr E, Nolan V Jr, Ketterson ED (2006) Physiological effects on demography: a long-term experimental study of testosterone’s effects on fitness. Am Nat 167:667-683 Rissman EF, Wingfield JC (1984) Hormonal correlates of polyandry in the spotted sandpiper, Actitis macularia. Gen Comp Endocrinol 56:401-405 Roberts ML, Buchanan KL, Evans MR (2004) Testing the immunocompetence handicap hypothesis: a review of the evidence. Anim Behav 68:227-239 Rossiter MC (1996) Incidence and consequences of inherited environmental effects. Annu Rev Ecol Syst 27:451-476 Sapolsky RM (1992) Neuroendocrinology of the stress-response. In: Becker JB, Breedlove SM, Crews D, McCarthy MM (eds) Behavioural Endocrinology. MIT Press, Cambridge/MA, pp 287-324 Sapolsky RM (2002) Endocrinology of the stress-response. In: Becker JB, Breedlove SM, Crews D, McCarthy MM (eds) Behavioural Endocrinology. MIT Press, Cambridge MA, pp 409-450 Sapolsky RM (2005) The influence of social hierarchy on primate health. Science 308:648-652 Schwabl H, Kriner E (1991) Territorial aggression and song of male European robins (Erithacus rubecula) in autumn and spring: effects of antiandrogen treatment. Horm Behav 25:180-194 Schwagmeyer PL, Parker GA (1990) Male mate choice as predicted by sperm competition in thirteen-lined ground squirrels. Nature 348:62-64 Shuster SM, Wade MJ (2003) Mating Systems and Strategies. Princeton University Press, Princeton Silk JB, Alberts SC, Altmann J (2003) Social bonds of female baboons enhance infant survival. Science 302:1231-1234 Silverin B (1980) Effects of long-acting testosterone treatment on free-living pied flycatchers, Ficedula hypoleuca, during the breeding period. Anim Behav 28:906-912 Sinervo B, Lively CM (1996) The rock-paper-scissors game and the evolution of alternative male strategies. Nature 380:240-243

500

Wolfgang Goymann and Heribert Hofer

Sinervo B, Miles DB, Frankino WA, Klukowski M, DeNardo DF (2000a) Testosterone, endurance, and Darwinian fitness: natural and sexual selection on the physiological bases of alternative male behaviors in side-blotched lizards. Horm Behav 38:222-233 Sinervo B, Svensson E, Comendant T (2000b) Density cycles and an offspring quantity and quality game driven by natural selection. Nature 406:985-988 Sinervo B, Bleay C, Adamopoulou C (2001) Social causes of correlational selection and the resolution of a heritable throat color polymorphism in a lizard. Evolution 55:2040-2052 Sinervo B, Chaine A, Clobert J, Calsbeek R, Hazard L, Lancaster L, McAdam AG, Alonzo S, Corrigan G, Hochberg ME (2006) Self-recognition, color signals, and cycles of greenbeard mutualism and altruism. Proc Natl Acad Sci USA 103:7372-7377 Smale L, Frank LG, Holekamp KE (1993) Ontogeny of dominance in free-living spotted hyaenas: juvenile rank relations with adult females and immigrant males. Anim Behav 46:467-477 Smuts BB (1986) Gender, aggression, and influence. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker TT (eds) Primate Societies. University of Chicago Press, Chicago, pp 400-412 Stoehr AM, Hill GE (2000) Testosterone and the allocation of reproductive effort in male house finches (Carpodacus mexicanus). Behav Ecol Sociobiol 48:407-411 Tinbergen N (1963) On aims and methods of ethology. Z Tierpsychol 20:410-433 Tomasello M, Call J (1997) Primate Cognition. Oxford University Press, Oxford van Jaarsveld AS, Skinner JD (1991) Plasma androgens in spotted hyaenas (Crocuta crocuta): influence of social and reproductive development. J Reprod Fertil 93:195-201 Verboven N, Monaghan P, Evans DM, Schwabl H, Evans N, Whitelaw C, Nager RG (2003) Maternal condition, yolk androgens and offspring performance: a supplemental feeding experiment in the lesser black-backed gull (Larus fuscus). Proc R Soc Lond B 270:2223-2232 Vernon CJ (1971) Notes on the biology of the black coucal. Ostrich 42:242-258 Voigt C, Goymann W (2007) Sex-role reversal is reflected in the brain of African black coucals (Centropus grillii). Dev Neurobiol 67:1560-1573 Voigt CC, Streich WJ (2003) Queuing for harem access in colonies of the greater sac-winged bat. Anim Behav 65:149-156 von Holst D (1998) The concept of stress and its relevance for animal behaviour. Adv Stud Behav 27:1-131 Walters JR, Seyfarth RM (1987) Conflict and cooperation. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker TT (eds) Primate Societies. University of Chicago Press, Chicago, pp 306-317 Wingfield JC (1984) Androgens and mating systems: testosterone-induced polygyny in normally monogamous birds. Auk 101:665-671 Wingfield JC (1994) Hormone-behaviour interactions and mating systems in male and female birds. In: Short RV, Balaban E (eds) The Differences Between the Sexes. Cambridge University Press, Cambridge, pp 303-330

Mating systems, social behaviour and hormones

501

Wingfield JC, Wada M (1989) Changes in plasma levels of testosterone during male-male interactions in the song sparrow, Melospiza melodia: time course and specifity of response. J Comp Physiol A 166:189-194 Wingfield JC, Hegner RE, Dufty AM Jr, Ball GF (1990) The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. Am Nat 136:829-846 Wingfield JC, Whaling CS, Marler P (1994) Communication in vertebrate aggression and reproduction: the role of hormones. In: Knobil E, Neill JD (eds) The Physiology of Reproduction, 2nd edn. Raven Press, New York, pp 303-342 Wingfield JC, Jacobs J, Hillgarth N (1997) Ecological constraints and the evolution of hormone-behavior interrelationships. Ann NY Acad Sci 807:22-41 Wingfield JC, Maney DL, Breuner CW, Jacobs JD, Lynn S, Ramenofsky M, Richardson RD (1998) Ecological bases of hormone-behaviour interactions – the emergency life-history stage. Am Zool 38:191-206 Wingfield JC, Jacobs JD, Tramontin AD, Perfito N, Meddle S, Maney DL, Soma K (2000) Toward an ecological basis of hormone-behaviour interactions in reproduction of birds. In: Wallen K, Schneider J (eds) Reproduction in Context. MIT Press, Cambridge/MA, pp 85-128 Wingfield JC, Lynn SE, Soma KK (2001) Avoiding the ‘costs’ of testosterone: ecological bases of hormone-behavior interactions. Brain Behav Evol 57:239251 Winslow JT, Hastings N, Carter CS, Harbaugh CR, Insel TR (1993) A role for central vasopressin in pair bonding in monogamous prairie voles. Nature 365:545-548 Yalcinkaya TM, Siiteri PK, Vigne J-L, Licht P, Pavgi S, Frank LG, Glickman SE (1993) A mechanism for virilization of female spotted hyenas in utero. Science 260:1929-1931 Young LJ, Wang Z (2004) The neurobiology of pair bonding. Nat Neurosci 7:1048-1054 Young AL, Richard AF, Aiello LC (1990) Female dominance and maternal investment in strepsirhine primates. Am Nat 135:473-488 Young LJ, Nilsen R, Waymire KG, MacGregor GR, Insel TR (1999) Increased affiliative response to vasopressin in mice expressing the V1a receptor from a monogamous vole. Nature 400:766-768 Zamudio KR, Sinervo B (2000) Polygyny, mate-guarding, and posthumous fertilization as alternative male mating strategies. Proc Natl Acad Sci USA 97:14427-14432

Part IV Behavioural variation

Chapter 17

The social modulation of behavioural development NORBERT SACHSER AND SYLVIA KAISER

ABSTRACT Individual differences in behaviour develop during ontogeny. They can be due to genetic or environmental factors as well as to their interactions. In this chapter we first address the role of genetic polymorphisms and geneby-environment interactions for the emergence of such variation. Then, we discuss the role of the social environment for the modulation of behavioural development in mammals during different phases of life. The social environment in which the pregnant female lives is of major importance for foetal brain development and, as a consequence, for offspring behaviour later in life. The effects are likely to be mediated by maternal hormones. During the early postnatal phase, contact with social companions provides a secure base for the infants, buffering their stress responses in challenging situations. Moreover, variation in maternal care can bring about distinct differences in offspring behavioural profiles, and specific behavioural traits can even be transmitted across generations by epigenetic inheritance. Finally, behavioural strategies and stress responsiveness can be canalised by social experiences during adolescence as well. From an evolutionary perspective, the hypothesis arises that social modulation of behavioural development is a mechanism of rapid adaptation, adjusting the individual efficiently to prevailing environmental conditions.

17.1 Introduction Members of the same species may vary conspicuously in their individual behavioural profiles, including social and sexual behaviours, cognitive abilities, as well as emotional and stress responses. From an evolutionary point of view, to understand such variation is of major importance because it is generally related to differences in reproductive success (e.g. Alcock

506

Norbert Sachser and Sylvia Kaiser

2005). In addition, individual behavioural profiles can have major consequences for quality of life and susceptibility to disease (Broom 2001; Korte et al. 2005). Hence, the analysis of variation in behavioural profiles is a focus not only in behavioural ecology, but also in animal welfare, biopsychological, and biomedical research. Individual differences develop during ontogeny, that is, the time from the moment the egg is fertilised until death. They can be due to genetic and/or environmental factors as well as to their interactions (Gross and Hen 2004). Concerning environmental factors, early postnatal influences were traditionally considered as most important for the individual’s behavioural development and adult phenotype (Immelmann et al. 1981), probably because socialisation and learning processes are thought to play a major role during this time. It has become evident, however, that environmental conditions during the prenatal phase (de Weerth et al. 2005) and during adolescence (Spear 2000) can shape brain development and, as a consequence, behaviour in adulthood as well. In this chapter we summarise some relevant findings with a focus on mammals, referring to humans when appropriate. Concerning the environment, we focus on social factors. In a first step, we address the significance of genes, environment, and their interactions for the emergence of individual differences in behaviour. In a second step, we focus on the modulation of behavioural development by conspecifics. We present findings for different phases of life – prenatal phase, early postnatal phase, and, briefly, adolescence -, discuss mechanisms that mediate the effects of social influences on behaviour and ask whether this social modulation represents an adaptive mechanism that helps the individual to adjust to current or future environmental conditions.

17.2 The role of genes and environment in behavioural development Even minor changes in specific alleles can lead to distinct inter-individual differences in aspects of behaviour, ranging from circadian rhythms (Vitaterna et al. 1994) to sexual (Nelson et al. 1995) and maternal behaviour (Brown et al. 1996), pair-bonding (Young et al. 1999), as well as learning and memory (Giese et al. 1998). Furthermore, the inactivation of the monoamine oxidase-A gene correlates with higher aggressiveness in both mice (Cases et al. 1995) and men (Brunner et al. 1993). Finally, in humans, a variable length polymorphism in the serotonin transporter gene promoter region has been described. Individuals carrying at least one copy of the short variant display higher levels of neuroticism and harm avoid-

The social modulation of behavioural development

507

ance as well as higher trait anxiety than homozygous carriers of the long variant (Lesch et al. 1996, Schinka et al. 2004). Consistently, mice lacking the serotonin transporter gene completely or in part – closely modelling the profile of the short allele combination in humans (Holmes et al. 2003a) – show higher degrees of anxiety-like behaviour (Holmes et al. 2003b). The identification of such genes is a major step in the analysis of neuronal and hormonal mechanisms underlying behaviour on a molecular level. It is important to emphasise, however, that behavioural differences between individuals can usually not be explained satisfactorily by corresponding variation in these genes. Instead, they result from the interplay of genes with environmental influences during behavioural development. This insight is not new. More than 50 years ago, Cooper and Zubek (1958) tested the learning ability of rats that were artificially selected for good or poor performance in a standard maze. However, when the ‘genetically’ dull rats were reared in an enriched environment and the ‘genetically’ bright rats had lived under restricted conditions, the latter performed significantly worse and the former significantly better. Thus, the genetic predisposition was overridden by the environmental conditions. Some decades later, genetically modified animals provide a promising new tool to study the interplay between genes and environment during development. For example, transgenic mice expressing exon 1 of the human huntingtin gene develop a neurodegenerative syndrome that closely models Huntington’s disease in humans. The exposure of these mice to a stimulating, enriched environment from early age on, however, helps to prevent the loss of cerebral volume and delays the onset of motor disorders indicative for the disease (van Dellen et al. 2000). Comparable findings emerged from studies on the effects of environmental enrichment on Alzheimer’s disease (AD). Transgenic mice carrying a genetic disposition (one allele of a human double mutated form of the amyloid precursor protein) to develop AD-like pathology show symptoms characteristic for human patients: impaired learning and memory (Chisthi et al. 2001, Janus et al. 2004, Görtz et al. 2008), stereotypic behaviour (Ambrée et al. 2006), and elevated levels of stress hormones (Touma et al. 2004). However, when the transgenic animals are kept in large semi-naturalistic enclosures from birth on, an exciting result is found: they cannot be differentiated from animals that do not carry the transgene by their behavioural profiles and basal levels of stress, neither at early ages, nor later in life when AD-like protein plaques are found in their brains (Lewejohann et al. 2009; see Fig. 17.1). Thus, genetic predispositions for specific traits can be modulated significantly by environmental influences. Individuals can differ significantly in their responsiveness towards environmental stimuli. Evolutionary modelling confirms that responsive and

508

Norbert Sachser and Sylvia Kaiser

Fig. 17.1 ‘Alzheimer-mice’ living in a semi-naturalistic enclosure. Up to 30 individually colour-marked (see bottom right) adult wildtype and transgenic mice used the spacious indoor enclosure with an effective surface area of 6.6 m². The transgenic mice carried one allele of a human double-mutated form of the amyloid precursor protein predisposing them for ‘Alzheimer-like’ pathology. However, when the transgenic animals were kept under these conditions from birth on, they could not be differentiated behaviourally and in basal levels of stress from wildtype mice that do not bear the transgene, neither at early ages nor later in life when ‘Alzheimer-like’ protein plaques appear in their brains (Lewejohann et al. 2009).

unresponsive personalities can indeed evolve (Wolf et al. 2008; Bergmüller this volume). There is increasing evidence that variation in behavioural profiles may result from specific genotype and gene-by-environment interactions, that is, some genotypes are more responsive to their environment than others.

The social modulation of behavioural development

509

BOX 17.1 Living in a dangerous world: modulation of behavioural profile by environment and serotonin transporter genotype Genetically engineered animals, such as knockout mice in which one or more genes have been turned off, provide a promising tool to study the role of single genes and their interaction with the environment for the development of individual variation in behaviour. The serotonin-transporter (5-HTT) is a key element in the regulation of the serotonergic system, and allelic variation in the (5-HTT) gene has been associated with traits of negative emotionality (Lesch et. al. 1996). Moreover, a central role of the 5-HTT in the control of social behaviour is suggested (Canli and Lesch 2007). The generation of mice with a targeted disruption of the 5-HTT gene allows examination of the consequences of diminished or absent function of the 5HTT. Using this animal model, the interaction between 5-HTT genotype and a dangerous environment during early phases of life and its impact on the development of behavioural profile in adulthood was studied (Heiming et al. 2009; see Fig. 17.2). During pregnancy and lactation dams were repeatedly exposed to olfactory cues of unfamiliar adult males by introducing male soiled bedding to their home cage (unfamiliar male bedding = UMB). Control females were treated with neutral bedding (NB). Behavioural data indicate that the UMB dams realised the relevance of the odour cues, which simulated the presence of potentially infanticidal males (vom Saal and Howard 1982, Elwood and Kennedy 1991). Regarding the behavioural profile of the offspring, varying in genotype (5-HTT+/+, 5-HTT+/–, 5-HTT–/–) there was a significant main effect of treatment, i.e. UMB offspring showed increased levels of anxiety-like behaviour and reduced exploratory locomotion compared to NB offspring. These behavioural alterations are likely to represent adaptive maternal effects: when living in a dangerous habitat it is beneficial to behave carefully and unobtrusively. There was also a main effect of genotype with 5-HTT–/– offspring showing higher levels of anxiety-like behaviour and lower levels of exploration than 5-HTT+/– and 5-HTT+/+ mice. Finally, concerning anxiety-like behaviour, the effects of a threatening environment in early life were most pronounced in 5-HTT–/– mice. In conclusion, if mothers live in a dangerous world during pregnancy and lactation, their offspring behavioural profile will, in principle, be shaped in an adaptive way. This process is modulated, however, by 5-HTT genotype. Subsequently it was studied whether anxiety-like behaviour and exploration can be modified by social experiences later in life (Jansen et al. 2010). Using the same mouse model, adult males of all three genotypes were provided with either the experience of being a winner or a loser in a residentintruder paradigm. Repeated social experience, irrespective of being a winner or loser, elevated levels of anxiety-like behaviour and decreased exploration. In losers, a distinct effect of genotype occurred, with 5-HTT–/– males

510

Norbert Sachser and Sylvia Kaiser

showing more anxiety-like behaviour and less exploration than the two other genotypes. In winners, no genotype-dependent variation was found. In conclusion, these data show that behavioural profiles can still be modulated by serotonin transporter genotype and social experiences in adulthood.

X 5-HTT +/–

X 5-HTT +/–

Treatment with neutral bedding during pregnancy and lactation ‘safe’ environment

5-HTT +/+

5-HTT +/–

Treatment with unfamiliar male bedding during pregnancy and lactation ‘dangerous’ environment

5-HTT –/–

5-HTT +/+

Dark light test *

250

*

*

200 150 100 50 0

5-HTT +/+ +/– –/–

Neutral bedding

5-HTT +/–

5-HTT –/–

Open field test

5-HTT +/+ +/– –/–

Unfamiliar male bedding

40

Path length (m)

Latency to enter light compartment [s]

300

5-HTT +/–

5-HTT +/–

*** ***

** ***

30 20 10 0

5-HTT +/+ +/– –/–

Neutral bedding

5-HTT +/+ +/– –/–

Unfamiliar male bedding

Fig. 17.2 The effects of a dangerous environment during early phases of life and serotonin transporter (5-HTT) genotype on anxiety-like and exploratory behaviour. Top: Experimental design: heterozygous (5-HTT+/–) females (red hair ribbon) were mated with heterozygous males. During pregnancy and lactation dams were repeatedly exposed to olfactory cues of unfamiliar adult males by introducing small amounts of male soiled bedding to their home cage (‘dangerous environment’). Control females were treated with neutral bedding (‘safe environment’). Offspring varying in genotype (5-HTT+/+, +/–, –/–) were examined for their anxiety-like and exploratory behaviour in adulthood. Bottom: Dark light test (DL): the DL test apparatus consists of a dark and a light compartment. The latency to enter the light compartment is a good measure for anxiety-like behaviour (the longer the latencies the higher the levels of anxiety). There was a significant main effect of genotype (p = 0.009) and treatment (p = 0.017). Open field test: the path length travelled in the unfamiliar open field indicates exploratory behaviour. There was a significant main effect of genotype (p < 0.0001) and treatment (p = 0.010). * p < 0.05, ** p < 0.01, *** p < 0.001 (Heiming et al. 2009).

The social modulation of behavioural development

511

For example, humans with one or two copies of the short allele of the serotonin transporter gene polymorphism exhibit more often depression and suicidal tendencies in relation to stressful life events, including health and relationship stressors, than individuals homozygous for the long allele (Caspi et al. 2003). An orthologue of the human gene polymorphism is present in rhesus monkeys (Macaca mulatta). In line with the studies in humans, macaques carrying the short allele are not only more anxious, but also seem to be more vulnerable to develop behavioural pathology in the face of chronic adversity (Barr et al. 2004, Spinelli et al. 2007). Finally, mice lacking the serotonin transporter totally or in part are affected more strongly by low maternal care (Carola et al. 2008), a dangerous environment during the prenatal and early postnatal phase (Heiming et al. 2009), and social defeat in adulthood (Jansen et al. 2010) than individuals with a functional serotonin transporter gene (see BOX 17.1 and Fig. 17.2). Thus, in humans, monkeys, and rodents, an individual’s response to environmental influences can be moderated by the specific genetic makeup. Future research might identify gene-by-environment interactions as a major source of inter-individual variation in behaviour (Belsky et al. 2009).

17.3 Effects of the social environment on offspring development 17.3.1 Prenatal phase A comprehensive understanding of behavioural development must include the prenatal phase because many parts of the central nervous system involved in the control of behaviour are organised during this phase of life. For example, in many species sex differences in behaviour are based on the organising effects of sex steroids during sensitive phases long before birth (see for example the classical study of Phoenix et al. 1959). One might argue that, in mammals, environmental influences during the prenatal phase are of minor importance for foetal development because the foetus is well protected within the mother. However, numerous studies have shown that particularly stressors acting upon the mother during pregnancy can have distinct and long-lasting effects on the offspring. A variety of functions and systems can be influenced, including behaviour, brain, reproduction, the neuroendocrine, autonomic, as well as the immune system (Ward 1972, Anderson et al. 1985, Herrenkohl 1979, Klein and Rager 1995, Maccari et al. 1995, Weinstock et al. 1998, for a review see de Kloet et al. 2005).

512

Norbert Sachser and Sylvia Kaiser

In most experimental studies on the effects of prenatal stress, pregnant females were subjected to non-social stressors (e.g. immobilisation, noise combined with flashing light, illumination-heat stress, handling). From these studies, important general insights have been derived regarding the prenatal programming of brain, behaviour, and physiological responses. Still, from a behavioural biologist’s point of view a critical comment can be made: such artificial stressors typically do not occur in the individual’s natural environment and one might doubt whether they represent ecologically relevant stimuli (Kaiser and Sachser 2005). In their natural habitat animals have to cope with a variety of stressors that depend on their ecological niche. They have to adjust to the physical environment (e.g. inclement weather) and the biotic world that surrounds them (e.g. predators, food shortage). A major part of an individual’s biotic environment consists of other members of the same species, which can be defined as its ‘social world’. In fact, a majority of human and animal daily expectations, motivations, and behaviours are directed toward encounters with conspecifics. On the one hand, this social world can sustain good welfare and health (e.g. through the effects of social support, see Sachser et al. 1998, Hennessy et al. 2009). On the other hand, it can result in severe stress, eventually leading to disease and even death (e.g. in the case of social defeat, social instability, or crowding; von Holst 1998). Thus, the social environment represents a most influential factor, which, during pregnancy, can be crucial for the development of the offspring as well (Kaiser and Sachser 2005, 2009). The most comprehensive insights regarding prenatal social influences on offspring behavioural development in non-human mammals are derived from studies in guinea pigs (review: Kaiser and Sachser 2005; see Fig. 17.3 and BOX 17.2). For example, daughters whose mothers have lived in an unstable world during pregnancy show conspicuous behavioural masculinisation later in life (Sachser and Kaiser 1996), as well as increased testosterone concentrations (Kaiser and Sachser 1998), and a male-typical expression pattern of sex hormone receptors in parts of the limbic system (Kaiser et al. 2003a) when compared to daughters whose mothers have lived in a stable social environment during this time. Sons whose mothers have lived in an unstable social environment during pregnancy show behavioural infantilisation, delayed development of the adrenocortical system, and down-regulation of androgen receptor expression in the limbic system compared to sons whose mothers have lived in a stable world (Kaiser et al. 2003b).

The social modulation of behavioural development Male-typical courtship behaviour

a

513

Testosterone

b 1,0

***

**

3,0

In ng/ml plasma

Frequencies/h

0,8

2,0

1,0

0,6

0,4

0,2 0,0

0,0

Daughters whose mothers have lived in a stable social environment Daughters whose mothers have lived in an unstable social environment

Distribution of androgen receptors (medial preoptic area) c

d

Daughter whose mother has lived in a stable social invironment

Daughter whose mother has lived in an unstable social environment

Fig. 17.3 The effects of the social environment during pregnancy on behaviour, endocrine system, and distribution of androgen receptors in daughters whose mothers have either lived in a stable or unstable social environment during this time. (a) Frequencies of male-typical courtship behaviour (rumba); (b) plasma testosterone concentrations; data in (a) and (b) are given as medians, 10%, 25%, 75%, and 90% quartiles. (c), (d) give an example of androgen receptor distribution, samples differ p < 0.001; ** p < 0.01, *** p < 0.001. Redrawn after Kaiser et al. (2003 a).

514

Norbert Sachser and Sylvia Kaiser

BOX 17.2 Of masculinised daughters and infantilised sons: effects of the social environment during pregnancy on offspring behaviour The social environment in which a female lives during pregnancy has major consequences for offspring behavioural profiles in adulthood. This insight derives in particular from studies in guinea pigs (for a review see Kaiser and Sachser 2005). Daughters and sons whose mothers have lived in a stable social environment during pregnancy and lactation show sex-specific differences regarding behaviour, endocrine systems, and brain development compared to those whose mothers have lived in an unstable social environment during this phase of life. A series of experiments revealed that the behavioural differences are exclusively due to the social instability during pregnancy, while the period of lactation does not seem to be of importance for this phenomenon (Sachser and Kaiser 1996). The stable social environment was achieved by keeping group composition (one male, five females) constant; in the unstable social environment situation two females from different groups were exchanged every third day. Daughters whose mothers have lived in an unstable social environment show a most conspicuous behavioural masculinisation later in life (see Fig. 17.3a): they display significant higher amounts of male-typical courtship and play behaviour than daughters whose mothers have lived in a stable social environment (Sachser and Kaiser 1996). Behavioural masculinisation is accompanied by increased testosterone concentrations (Fig. 17.3b; Kaiser and Sachser 1998) and a male-typical expression pattern of androgen receptors in specific areas of the limbic system, including the medial preoptic area (Fig. 17.3c; Kaiser et al. 2003a). Since this brain region plays an essential role in the control of masculine behaviours (Gréco et al. 1998), it seems likely that changes in these expression patterns are causally related to the daughters’ behavioural masculinisation. In male offspring, completely different effects were found. When mothers have lived in an unstable social environment, their sons displayed behavioural patterns that are usually shown only by very young guinea pigs (e.g., resting with body contact). These patterns are not only shown more frequently but also up to an older age compared to males whose mothers have lived in a stable social environment. In addition, their courtship behaviour was frequently integrated into play behaviour. This behavioural profile seems to be – at least in part – different from a ‘demasculinisation’, that is, the loss or attenuation of male-typical traits, as well as from a ‘feminisation’, defined as the expression of female-typical traits. We, therefore, introduced the term ‘behavioural infantilisation’ (Kaiser and Sachser 2001). This syndrome coincides with delayed development of the hypothalamic-pituitaryadrenocortical (HPA) system (Kaiser and Sachser 2001), which is typically regarded as the body’s primary stress responsive neuroendocrine system. Finally, a down-regulation of androgen receptor expression was found in the

The social modulation of behavioural development

515

medial preoptic area and the Nucleus arcuatus of the hypothalamus if mothers have lived in an unstable social environment during pregnancy (Kaiser et al. 2003b). We also studied whether these behavioural consequences of prenatal social influences are adaptive or whether they are simply the (pathological) consequences of physiological constraints. We showed that in reproductively challenging situations in adulthood the infantilised males change their behavioural pattern and can no longer be differentiated behaviourally from noninfantilised males. Thus, no indications for a behavioural disorder exist (Kemme et al. 2007, 2008). Currently the central question is: what is the benefit for sons who show less pronounced expression of male-typical traits and/or a delay in development? (And what is the benefit of being a masculinised daughter?). So far, there are no empirical data to answer these questions, but a reasonable hypothesis has been derived that is in accordance with data on the social life of guinea pigs under natural and laboratory conditions (Sachser 1986, 1998, Asher et al. 2004, 2008; see also BOX 17.3): masculinised daughters and infantilised sons might be better adapted to highdensity populations, while the opposite behavioural profiles may benefit individuals living under low densities (Kaiser and Sachser 2005). The analysis of reproductive data indicated that adaptive maternal effects might indeed exist: we compared litter sex ratios of female guinea pigs, exposed experimentally to a stable or unstable social environment. Under unstable social conditions, the sex ratio was significantly more biased towards daughters. This finding was consistent among four independent experiments. Life span can be dramatically reduced under conditions of social instability. Hence, mothers in such conditions should bias their investment towards the sex that reaches sexual maturity first, which is the female sex in this species. Thus, shifting the offspring sex ratio towards more daughters under conditions of social instability may represent a maternal strategy to maximise future reproductive success (Kemme et al. 2009).

Besides guinea pigs, studies of prenatal social influences on offspring development have been conducted only in a few other species (mice, rats, squirrel monkeys). Although a variety of different social stressors have been applied in these experiments (crowding, social confrontation, changing group membership), a common characteristic of all approaches is the induction of social instability. In general, the number of interactions with conspecifics increases under such conditions and the predictability and controllability of social encounters decline dramatically. Interestingly, stress research shows that situations of uncertainty or unpredictability are a major source of stress responses (von Holst 1998, Sachser 2001). When all experimental studies in different species are compared, some general conclusions can be drawn (Kaiser and Sachser 2005, 2009): the offspring is af-

516

Norbert Sachser and Sylvia Kaiser

fected by prenatal social stress in a sex-specific and sex-reversed way. Daughters show signs of masculinisation, including their behaviour, endocrine state, and brain development. In sons, a less pronounced expression of male-typical traits (described in terms of demasculinisation, feminisation, infantilisation) and/or a delay in development occurs. Concerning the neuroendocrine pathways mediating the effects of prenatal social stress on offspring behaviour, the general mechanism seems to be clear: environmental stimuli act on the female during pregnancy. The maternal organism responds with hormonal changes, which in turn influence the embryonic/foetal endocrine state. Hereby, foetal brain development is affected with major consequence for offspring behaviour later in life. However, the detailed mechanisms are not yet understood (for details see Kaiser and Sachser 2005), primarily because prenatal social stress affects the offspring in a sex-reversed way. While sons are characterised by less pronounced male-typical traits, daughters are behaviourally masculinised. The mechanisms underlying these differences are not known, however. In addition, different hypotheses exist concerning the specific neuroendocrine pathways that mediate the effects of prenatal social stress on behaviour. With respect to sons, for example, some researchers argue that social stressors activate the maternal hypothalamo-pituitary-adrenocortical (HPA) axis during pregnancy, resulting in elevated plasma glucocorticoid concentrations (Weinstock et al. 1998). These in turn affect foetal brain development, for example by eliminating or attenuating the testosterone surge in the male foetus (Harvey and Chevins 1984). As a consequence, less pronounced male-typical traits are expressed. There are some problems, however, with this explanation. For instance, placental 11 betahydroxysteroid dehydrogenase is known to act as a physiological foetoplacental ‘barrier’ to endogenous glucocorticoids, preventing maternal cortisol to cross the placental barrier freely (Seckl et al. 2000). Others have suggested an alternative pathway that may not involve the HPA system (Lieberman 1963, Kaiser et al. 2003c). They hypothesise that social stress during pregnancy induces a distinct activation of the maternal sympatheticadrenomedullary system. Hereby, maternal androgen secretion is suppressed, leading to lower serum levels of sexual steroids in the foetus. A significant impact on foetal brain development results, with downregulated androgen receptors in specific parts of the limbic system (Kaiser et al. 2003b). In summary, the present data suggest that different neuroendocrine pathways may cause the behavioural infantilisation in prenatally stressed sons (viz. masculinisation in daughters; Kaiser and Sachser 2005). Prenatal influences on behaviour have not only been studied in mammals, but also in birds (for a review, see Groothuis et al. 2005). In both

The social modulation of behavioural development

517

taxonomic groups comparable mechanisms are discussed in which maternal steroids play a decisive role: in birds, mothers deposit androgens in the yolk of their eggs (Schwabl 1996, 2001, Eising et al. 2003), whereas in mammals, mothers affect foetal steroid levels via the placenta. However, investigations of prenatal influences on behaviour in mammals and birds differ in some important aspects: in mammals, this subject is mainly studied in the context of prenatal stress research, predominantly by researchers with a psychological or biomedical background. Notably, they often regard the characteristic traits of individuals who were exposed to environmental stressors during pregnancy as deviation from the norm or even as pathological. In contrast, studies on prenatal influences on behaviour in birds are mainly conducted by biologists with a primary interest in evolutionary questions. These researchers do not use the term ‘prenatal stress’ but talk about ‘maternal effects’ (e.g., Birkhead et al. 2000), i.e., female control over offspring development. From this point of view, mothers try to maximise their own Darwinian fitness by adjusting their offspring efficiently to current or future environmental conditions. Hence, variation in offspring traits depending on the amount of maternal androgen deposited into egg yolk is regarded as a mechanism to fine-tune offspring growth, physiological state, competitiveness, etc. in order to pre-adapt them to the environment in which the mother lives (review: Groothuis et al. 2005). In evolutionary biology, maternal effects have received much attention (Mousseau and Fox 1998, Qvarnström and Price 2001). Because they control and modulate the variance and quality of the next generation, they directly modify intensity and direction of selection and, consequently, evolutionary change. Results of two recent studies in wild spotted hyenas and guinea pigs suggest that prenatal androgen exposure can adaptively influence offspring phenotype in mammals as well (Dloniak et al. 2006, Kemme et al. 2007). We can speak of an adaptive phenotype if such traits result in enhanced measures of fitness parameters such as social dominance and eventually reproductive success. Noteworthily, similar arguments for the adaptive value of prenatal stress effects in humans have been put forward. For example, Bateson et al. (2004) hypothesised that a period of starvation during pregnancy tells the developing foetus that food is probably going to be scarce in the future. Babies of such mothers often show small body weight and correspondingly modified metabolism. These traits are not necessarily pathological, inasmuch as they help the baby to cope with environments of low food availability. (For further hypotheses on the adaptive value of prenatal stress effects in humans see also Kaiser and Sachser 2009).

518

Norbert Sachser and Sylvia Kaiser

17.3.2 Early postnatal phase More than 50 years ago, experiments with rhesus monkeys by Harlow and co-workers made it clear for the first time that social relationships during the early postnatal phase might be crucial for the development of behavioural profiles. Monkeys raised in isolation are fearful and depressed in new situations, while they are hyper-aggressive in response to unfamiliar members of the same species and cannot build ‘normal’ social relationships (Harlow and Harlow 1962). A review on early life stress and the development of aggression in primate and rodent models confirmed that chronic deprivation of early maternal care as well as chronic deprivation of early physical interactions with peers is a profound risk factor for the development of inappropriate aggressive behaviours (Veenema 2009). Thus, the majority of mammals probably require at least some socialisation in order to be able to communicate and interact successfully with members of the same species (Sachser 2001). In many mammals close contact with conspecifics during early life provides the young with a sense of security. Only if this secure base is available, they can explore their environment in a non-anxious and nonstressful way (Bowlby 1972, Gandelmann 1992, Carter et al. 2005). In most species, particularly the mother functions as a secure base for the infant. Depending on the species’ social organisation, however, this role can also be played by the father, the extended family or the entire social group. Accordingly, companions with close, positive social relationships to the young are more likely to buffer their hormonal stress responses than conspecifics that are unfamiliar or to which the infants are less positively disposed. This insight has been demonstrated impressively in studies on social support (Hennessy et al. 2009). In species in which infants exhibit evidence of a true attachment to the mother, the presence of the mother effectively buffers the infant’s hormonal stress responses to stressors such as novelty, i.e., the increase of glucocorticoids is reduced (Smothermann et al. 1979, Hennessy et al. 1982, 2006a; see Fig. 17.4a). Moreover, this ability is typically selective, in that the mother usually is more effective than other adult females or siblings (Hennessy and Ritchey 1987, Sachser et al. 1998, Graves and Hennessy 2000). In titi monkeys (Callicebus molloch), as in some other monogamous species, the father plays an active role in care-taking, actually carrying the infant more than the mother, while the infant in turn appears more attached to the father than to the mother (Mendoza and Mason 1986). In this species, it is the father that is clearly most effective at buffering plasma cortisol response to novelty (Hoffman et al. 1995). In the monogamous Muenster yellow-toothed cavy (Galea monasteriensis), infants have

The social modulation of behavioural development

519

Cortisol stress response to novelty

Increases in ng/ml plasma

600

a

Preweaning

*

600

b

Late adolescence

600

500

500

400

400

400

300

300

300

200

200

200

100

100

100

0

0

0

500

Alone

c

Adulthood

*

With preferred social companion

Fig. 17.4 Cortisol stress responses to novelty in preweaning, late adolescent, and adult male guinea pigs living in large mixed-sex colonies. The animals were placed into an unfamiliar enclosure either alone or together with the preferred social companion in their colony. Blood samples were taken immediately before and 1 h after the beginning of the experiments. Increases in cortisol concentrations are given as means + SEM. The preferred social companion was (a) for infants always the mother; (b) for late adolescent and (c) adult males always an adult female. * p < 0.05. Redrawn after Hennessy et al. (2006b).

lower cortisol responses in a novel environment when with their mother than when with an unfamiliar female. In contrast, for the closely related harem-forming wild cavy such a difference between the mother and another female was not found. This variation in infant stress responses is reflected by patterns of interactions of infants and adults: in the harem-living species, interactions with mothers and unrelated females are largely similar in the novel environment. In the pair-living species, only interaction with the mother is characterised by physical closeness and socio-positive behaviour (Hennessy et al. 2006a). As a general rule, variation in the buffering of the stress response by specific individuals follows the pattern of attachments (Hennessy et al. 2009). Social experiences during early rearing appear capable of affecting the ability to benefit from social buffering in later life. In rhesus monkeys housed in long-standing pairs, the presence of the cage mate reduced the plasma cortisol response during novelty exposure, but only in monkeys that had been reared by their mothers (Winslow et al. 2003). Variation in the quality of care provided by biological mothers also appears to influence buffering effects. A percentage of rhesus monkey mothers exhibit

520

Norbert Sachser and Sylvia Kaiser

abusive behaviour toward their infants (Maestripieri et al. 2005). When young rhesus monkeys were exposed to a novel environment with their biological mothers, those with abusive mothers showed a pronounced hormonal stress response, whereas those with non-abusive mothers did not (Hennessy et al. 2009). In addition to these effects, a different line of studies suggests that variation in the quality of relationship between social companions and infant during early phases of life as, for instance, indicated by the frequency and intensity of maternal care, might be a major source of inter-individual differences in behaviour and stress responsiveness in adulthood. This insight is primarily derived from experiments in rats (Meaney 2001). Variation in maternal behaviour, particularly in licking and grooming, is a stable, individual characteristic of rat dams (Champagne et al. 2003). In their groundbreaking work, Michael Meaney and colleagues showed that such variation in maternal care alters the expression of genes that regulate behavioural and hormonal stress responses as well as hippocampal synaptic development of the offspring. These effects form the basis for the development of stable, individual differences in stress reactivity, certain forms of cognition, and anxiety-related behaviour (Liu et al. 1997, Meaney 2001). For example, as adults, the offspring of mothers that exhibited more licking and grooming of pups during the first 10 days of life have a reduced acute corticosterone stress response and decreased levels of hypothalamic corticotrophin-releasing hormone messenger RNA (Fig. 17.5). On a behavioural level, offspring that received high amounts of maternal care show a decreased startle response, increased open field exploration, and shorter latencies to eat food provided in a novel environment (Meaney 2001). In summary, these and further data (Francis and Meaney 1999) suggest that maternal behaviour can program offspring stress responsiveness and aspects of behavioural phenotype. Rat dams displaying low degrees of maternal care are more fearful and show increased stress responsiveness, compared to mothers that show high degrees of maternal care (Francis et al. 2000). These individual differences are apparently transmitted across generations: fearful mothers have more stress-reactive offspring, and daughters of high-licking and -grooming mothers show higher frequencies of maternal care than offspring of low licking and –grooming mothers (Francis et al. 1999). A series of crossfostering experiments suggested a non-genomic transmission of such individual differences and identified variation in maternal care as the potential mediator (Francis et al. 1999, Meaney 2001). The behavioural and stress physiological traits that are transmitted across generations are combined with stable alterations in gene expression (Champagne and Curley 2009). Since changes in DNA base sequences probably do not occur, this finding

The social modulation of behavioural development a

521

Maternal care 200

Licking/grooming

*

*

150

0.5

*

0.4 100

0.3 0.2

50

0.1 0

Low

High

0

Fig. 17.5 Maternal care and offspring stress responses in adulthood. (a) Maternal care in low and high licking/grooming rat dams. (b) Corticosterone concentrations 20 minutes after start of restraint stress in offspring of high and low licking/grooming mothers. (c) Levels of CRH mRNA in the Nucleus paraventricularis in the adult offspring of high and low licking/grooming mothers. (CRH = corticotropin releasing hormone, a neuropeptide significantly involved in the regulation of the corticosterone stress response). All values are given as means + SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. From Liu et al. (1997). Reprinted with permission from AAAS.

represents an example of epigenetic inheritance (Jablonka and Raz 2009). Epigenetic modification of steroid hormone receptor gene promoters, including their methylation, has indeed been identified as the molecular mechanism (Champagne et al. 2006, Weaver et al. 2004). To summarise, in rats, early life experience results in stable epigenetic programming with long-term consequences for specific aspects of behaviour and stress responsiveness. It remains poorly known whether these findings can be generalised to other species and other social behavioural patterns. In humans and primates, for example, behavioural development as well as the transmission of behavioural profiles from mother to offspring depends on social learning, and there are only some suggestions that this transmission involves epigenetic alterations (Champagne and Curley 2005). On the other hand, there are some indications that other social experiences like social defeat can also induce changes in the epigenetic regulation of genes and, hence, changes in gene expression (Tsankova et al. 2006). Such findings suggest that this mechanism may be more universal. Programming offspring stress responsiveness and behavioural phenotype by maternal behaviour is regarded as an adaptive pattern of develop-

522

Norbert Sachser and Sylvia Kaiser

ment. It is argued that conditions such as food availability, predation, social instability, and social status directly influence the emotional state of the mother and, thus, maternal care. The effects of these environmental challenges on the development of the pups are then mediated by alterations in maternal care. Variations in maternal care can thus serve to transduce an environmental signal to the pups. Under conditions of environmental adversity, maternal anxiety will increase. The resulting decrease in maternal care will program increased behavioural and endocrine reactivity to stress in the offspring. Under conditions of adversity, these responses are adaptive, in that they promote detection of potential threat, avoidance learning, and metabolic/cardiovascular responses that are beneficial under the increased demands of the stressor. Because offspring usually inhabit niches similar to those of their parents, the transmission of these traits from parent to offspring could be adaptive (Meaney 2001). However, from a functional point of view, experiments are needed to decide unequivocally whether the consequences of variation in maternal care indeed represent adaptations. So far, it is not clear, for example, whether offspring from adverse conditions cope better with adversity and attain higher reproductive success under conditions in which they have been reared than offspring that have lived in a generous, non-challenging world, and vice versa. 17.3.3 Adolescence The gradual transition from infancy to adulthood is called adolescence. It encompasses the attainment of sexual maturity and changes in neural circuitry and endocrine systems as well as in behaviour (Primus and Kellog 1990, Delville et al. 1998, Spear 2000, Romeo et al. 2002, Hayward, 2003, Sisk and Foster 2004, Sisk and Zehr 2005, Shen et al. 2007). During adolescence, the social environment changes dramatically, since the focus on interactions with parents shifts towards interactions with peers. In this context, the adolescent acquires knowledge of social rules and behaviour patterns that prove useful for life with conspecifics (Sachser 1993, Sachser et al. 1994). In addition, emigration from the natal group occurs often at this developmental stage (Baker 1978). Adolescence is furthermore characterised by a distinctive increase in novelty and sensation seeking and risk taking behaviours in humans (Trimpop 1999) as well as in other mammals (Macrí et al. 2002, Laviola et al. 2003). Hence, adolescence is regarded as an important, dangerous, and potentially stressful life stage (Spear 2000, Yurgelun-Todd 2007). There is some evidence for increased physiological stress responsiveness in animal and human adolescents (Spear 2000). However, evidence

The social modulation of behavioural development

523

for a reduced stress response during adolescence has been reported as well: periadolescent mice do not exhibit the strong rise in corticosterone concentrations on forced exposure to novelty that is typical for adult mice (Adriani and Laviola 2000). The responsiveness of the noradrenergic system also seems to be blunted around periadolescence (Choi and Kellog 1996). Moreover, a heretofore undescribed period of stress (cortisol) response suppression was discovered in maturing male guinea pigs, which does not exist during infancy and full adulthood (Hennessy et al. 2006b; see Fig. 17.4). This period coincides with the time when males begin to display social skills that allow them to negotiate potentially dangerous interactions with other males (Sachser et al. 1994, 1998). Thus, a change in stress responsiveness might somehow facilitate the acquisition or expression of these behavioural skills. Alternatively, reduced cortisol responsiveness to novelty might promote dispersal (Laviola et al. 2003). Reduced stress responsiveness during adolescence is not found under all environmental conditions, but depends on social experiences made during this phase of life (Kaiser et al. 2007, Lürzel et al. 2010): colony-housed male guinea pigs in late adolescence displayed a significantly lower cortisol stress response in a novel environment than equally aged males kept only with a female since weaning. These two groups of males differed also significantly in their behavioural profiles: males living in large mixed-sex colonies throughout adolescence were characterised by low aggressiveness, whereas males housed in heterosexual pairs from early adolescence on, were highly aggressive (Sachser et al. 1994, 1998, Hennessy et al. 2006b). Hence, the behavioural and physiological changes and canalisations that occur during adolescence may not reflect the realisation of fixed genetic programmes, but rather an adaptive modulation by social experiences (see BOX 17.3 and Fig. 17.6). This hypothesis supports the assumption that adolescence may not only be a time of high risks, but also an opportunity for adaptation (Steinberg 2005). Concerning underlying mechanisms the question arises: how can a suppression of stress (cortisol) responsiveness during adolescence be organised in males living in large heterosexual groups, but not in males living in pairs? In rodents, sex differences in stress responsiveness are well known. These differences emerge during puberty and are dependent on gonadal steroids, with males generally displaying lower activity of the HPA axis than females (Viau and Meaney 1996, Seale et al. 2004). In addition to these activational effects, evidence for an organisational influence of testosterone on the HPA axis of male rats has been reported, causing decreased base values as well as reduced stress hormone secretion in response to a stressor (McCormick et al. 1998, Seale et al. 2004). In these studies testosterone levels and androgen uptake were manipulated during

524

Norbert Sachser and Sylvia Kaiser

BOX 17.3 Queuing or fighting? Shaping of behavioural strategies during adolescence There is increasing evidence that individual behavioural strategies can be shaped significantly by social experiences during adolescence. This insight was demonstrated for the first time in a series of experiments with guinea pigs (Sachser et al. 1994, 1998; Hennessy et al. 2006b). In a first step, it was shown that living under different social housing conditions from early adolescence through full adulthood has major consequences for behaviour and stress responses. These studies were conducted with males that were born in large mixed-sex colonies. During early adolescence they were transferred either to another colony (COL), or they were housed with a single female (PAIR). At the time of transfer, they were already weaned, but had not yet reached sexual maturity. If, during full adulthood, two COL were placed into an unfamiliar enclosure in the presence of an unfamiliar female, they quickly established stable dominance relationships without displaying overt aggression, and no cortisol stress response occurred either in the winner or the loser. A completely different picture emerged if two PAIR were confronted in the same way. In such situations, high levels of aggressive behaviour were displayed and escalated fighting was frequent. Pronounced hormonal stress responses were found particularly in losers and many experiments had to be stopped to avoid injuries (Fig. 17.6; Sachser and Lick 1991). In a second step, the influence of age on the ability to adapt to unfamiliar conspecifics was analysed in PAIR and COL (Sachser 1993, Sachser et al. 1994). COL showed no distinct indications of stress, irrespective of whether the confrontations took place before, around, or after reaching sexual maturity. In contrast, PAIR adapted without problems only before reaching sexual maturity. When PAIR were confronted around the time of sexual maturity, social stress responses increased markedly and confrontations between older animals had to be terminated to avoid health problems. Thus, adolescence is a time during which PAIR became increasingly incompatible with unfamiliar males. During behavioural development PAIR were never involved in agonistic interactions because fighting and threat displays do not occur between the sexes in guinea pigs. In contrast, during adolescence, COL were increasingly involved in agonistic encounters with older, dominant males (Sachser and Pröve 1988). It seems likely that in this way they learned crucial social rules, that is, to assess the fighting ability of an opponent and to submit to higher ranking males. PAIR, in contrast, cannot learn these rules in the absence of other males. As a consequence, high levels of aggression occur when unfamiliar adult conspecifics meet in the presence of a female. Interestingly, the experience of only 50 minutes of agonistic encounters during adolescence is sufficient to shift the behaviour from a PAIR pattern to a COL pattern (Sachser et al. 1994). In summary, these data suggest a causal

The social modulation of behavioural development

525

relationship between social experiences during adolescence, aggressive behaviour as adults, and the degree of social stress in chronic encounters. The crucial role of social housing conditions has also been shown in a study taking a different approach (Sachser and Renninger 1993). Males reared individually, in pairs, or in colonies were introduced singly into unfamiliar colonies of conspecifics. COL easily adjusted to the new social situation, integrated into the social network of the established colonies and even gained higher social positions than in their native colonies. In the new colonies, changes could not be determined in either their body weights or in the activities of the organism’s stress axes. In contrast, individually or pairreared males were frequently involved in threat displays and fighting. As a consequence, they responded to the new situation with substantial decreases in body weight and with extreme increases in stress hormone concentrations. Originally, the behavioural profile of COL was regarded as a paradigm of ‘successful’ socialisation during adolescence, since these males were able to arrange with unfamiliar conspecifics in a low-aggressive and non-stressful way (Sachser et al. 1994, 1998). More recently, it was argued that the different behavioural strategies of COL and PAIR may represent adjustments to different social situations (Hennessy et al. 2006b). Indeed, the social organisation of guinea pigs is highly flexible and density-dependent male reproductive strategies exist (Sachser 1986). At low densities, the most promising strategy is to fight for dominance and to defend the mating partner. The high aggressiveness of PAIR might represent an adaptation to this situation. In the ancestor of the guinea pig, the wild cavy, a comparable situation exists in the natural habitat at low population densities (Asher et al. 2004). In contrast, at higher densities, a queuing strategy yields the highest reproductive success. Accordingly, the low aggressiveness of COL might promote this strategy, which is also found at high natural densities (Asher et al. 2008). Thus, social experiences during adolescence shape behavioural strategies that may indeed represent adaptations to the current social conditions (Hennessy et al. 2006b).

the perinatal period, but taken together with the knowledge that during adolescence massive (re-)organisation of neural circuits and behaviours takes place (Sisk and Zehr 2005, Uhlhaas 2009), it would stand to reason that the function of the HPA axis might be organized during this phase of life as well. In mammals, group-living males have higher testosterone levels than pair-living or solitary males (Sachser 1990) due to a testosterone surge that occurs during adolescence under group-living conditions (Sachser and Pröve 1988). This increase in testosterone levels coincides with increasing numbers of social interactions with adult conspecifics, namely agonistic interactions with other males and courtship and sexual behaviour oriented toward females, and is in concordance with the ‘challenge hypothesis‘

526

Norbert Sachser and Sylvia Kaiser Males raised with a female

Cortisol in ng/ml plasma

1000

*** ***

800

** *

Males raised in colonies 400 200

600 0 400

I

Initial value

200 0

I

1

1

6

Days after the beginning of the confrontation

6

Initial value Days after the beginning of the confrontation

Winner

Loser

Control

Fig. 17.6 Consequences of rearing conditions during adolescence for cortisol stress responses in male guinea pigs. Two males that were either raised in different large mixed-sex colonies or in heterosexual pairs are placed into an unfamiliar enclosure in the presence of an unfamiliar female. The bars represent means + SEM of controls and confronted animals 20 h before (initial values) and at different times after the beginning of the confrontations. * p < 0.05, ** p < 0.01, *** p < 0.001. For reference and original data see Sachser et al. (1998).

(Wingfield et al. 1990, Goymann et al. 2007, see also Goymann and Hofer this volume). Combining our knowledge about social interactions and testosterone with insights about the organisation of stress responsiveness by testosterone, an exciting hypothesis emerges: the frequency and intensity of social interactions during adolescence modulate the level of testosterone secretion, which in turn organises stress responsiveness and, thereby, facilitates adaptation to the social conditions experienced during this phase of life. It seems promising to test this hypothesis on the developmental canalisation of behavioural profiles and stress responsiveness during adolescence in order to obtain deeper insights into the interactions between behavioural mechanisms and functions.

17.4 Conclusions Individual differences in behaviour emerge during ontogeny. Concerning sources of variation, genetic polymorphisms play an important role, and even minor changes in specific alleles can have distinct effects. However, behavioural profiles are not determined by genetic information. Instead, they result from the interplay of genetic and environmental factors. Thus, it is timely to study in which way specific genotypes and genotype-by-

The social modulation of behavioural development

527

environment interactions contribute to individual differences in emotion, cognition, and behaviour. Concerning the role of the environment, the organism seems to be most susceptible to external influences during early development, i.e. the prenatal and early postnatal phase, when synaptic connections are elaborated and refined, and when brain circuits are highly plastic (Champagne and Curley 2005). Accordingly, the social environment in which a female lives during pregnancy can have major consequences for embryonic/foetal development and behaviour later in life. The effects are likely to be mediated by maternal hormones, although the detailed mechanisms are not yet understood. During the early postnatal phase social interactions between infants and conspecifics are most important. Lack of social contact can result in severe behavioural disturbance. The presence of social companions, in contrast, functions as a secure base, buffering the infant’s stress response in challenging situations. Notably, variation in maternal care can be a major source for offspring inter-individual differences in behaviour, and specific behavioural traits can be transmitted across generations by epigenetic inheritance. It must be emphasised, however, that extensive future research is needed to confirm the generality of the latter conclusions. It was not until recently that adolescence was identified as a decisive phase during which behavioural profile and endocrine stress responses can be shaped as well. There is some evidence that the behavioural and physiological changes and canalisations that occur during this phase of life are not the result of fixed genetic programs, but rather modulations by social experience. Up to now this hypothesis is based on only a rather limited number of studies in a small number of species. From an evolutionary point of view the question arises whether the social modulation of behavioural development is adaptive, or whether it simply represents the consequences of physiological constraints. In this context results achieved for the prenatal and early postnatal phases can be discussed logically within the concept of adaptive maternal effects, i.e. female control over offspring development, in order to adjust them efficiently to the prevailing or future environmental conditions. Whether the behavioural consequences of social modulation during these early phases of life indeed represent adaptations cannot be decided for mammals yet because empirical data, including measures of fitness, are largely lacking. At least some of the programming that takes place during the prenatal and early postnatal phase is potentially reversible later in life (Weaver et al. 2006, Kemme 2007, 2008). Such plasticity is highly adaptive because offspring might find themselves in situations for which they were not programmed during early phases of life, for example, if environmental conditions changed in an unpredictable way. It is tempting to speculate that dur-

528

Norbert Sachser and Sylvia Kaiser

ing adolescence the early programming might be reassessed and, if required, readjusted to the current environmental conditions. If so, the social modulation of behavioural development from the prenatal phase through adolescence could emerge as an effective mechanism of rapid adaptation.

Acknowledgements We thank Rebecca Heiming, Friederike Jansen, Lars Lewejohann, and Stephanie Lürzel for critical comments on the manuscript. Our studies on the social modulation of behavioural development in guinea pigs and mice were supported continuously by the German Research Foundation (DFG) and are currently supplied by grants to Sylvia Kaiser (FOR 1232: Ka 1546/6-1; Ka 1564/7-1) and Norbert Sachser (Sa 389/10-1; SFB-TRR 58, project A1; FOR 1232: Sa 389/11-1).

References Adriani W, Laviola G (2000) A unique hormonal and behavioral hyporesponsivity to both forced novelty and D-amphetamine in periadolescent mice. Neuropharmacol 39:334-346 Alcock J (2005) Animal Behavior: An Evolutionary Approach. Sinauer Associates, Sunderland Ambrée O, Touma C, Görtz N, Keyvani K, Paulus W, Palme R, Sachser N (2006) Activity changes and marked stereotypic behavior precede Abeta pathology in TgCRND8 Alzheimer mice. Neurobiol Aging 27:955-964 Anderson DK, Rhees RW, Fleming DE (1985) Effects of prenatal stress on differentiation of the sexually dimorphic nucleus of the preoptic area (SDN-POA) of the rat brain. Brain Res 332:113-118 Asher M, Spienelli de Oliveira E, Sachser N (2004) Social system and spatial organization of wild guinea pigs (Cavia aperea) in a natural low density population. J Mammal 85:788-796 Asher M, Lippmann T, Epplen JT, Kraus C, Trillmich F, Sachser N (2008) Large males dominate: ecology, social organization, and mating system of wild cavies, the ancestors of the guinea pig. Behav Ecol Sociobiol 62:1509-1521 Baker RR (1978) The Evolutionary Ecology of Animal Migration. Hodder and Stoughton, London Barr CS, Newman TK, Schwandt M, Shannon C, Dvoskin RL, Lindell SG, Taubman J, Thompson B, Champoux M, Lesch KP, Goldman D, Suomi SJ, Higley JD (2004) Sexual dichotomy of an interaction between early adversity and the serotonin transporter gene promoter variant in rhesus macaques. Proc Natl Acad Sci USA 101:12358-12363

The social modulation of behavioural development

529

Bateson P, Barker D, Clutton-Brock T, Deb D, D´Udine B, Foley RA, Gluckman P, Godfrey K, Kikwood T, Lahr MM, McNamara J, Metcalfe NB, Monaghan P, Spencer HG, Sultan SE (2004) Developmental plasticity and human health. Nature 430:419-421 Belsky J, Jonassaint C, Pluess M, Stanton M, Brummett B, Williams R (2009) Vulnerability genes or plasticity genes? Mol Psychiatry 14:746-754 Birkhead TR, Schwabl H, Burke T (2000) Testosterone and maternal effectsintegrating mechanisms and function. Trends Ecol Evol 15:86-87 Bowlby J (1972) Attachment. Penguin, Middlesex England Broom DM (2001) Coping with challenge: welfare in animals including humans. Dahlem Workshop Report 87. Dahlem University Press, Berlin Brown JR, Ye H, Bronson RT, Dikkes P, Greenberg ME (1996) A defect in nurturing in mice lacking the immediate early gene fosB. Cell 86:297-309 Brunner HG, Nelen M, Breakefield XO, Ropers HH, van Oost BA (1993) Abnormal behavior associated with a point mutation in the structural gene for monoamine oxidase A. Science 262:578-580 Canli T, Lesch KP (2007) Long story short: the serotonin transporter in emotion regulation and social cognition. Nat Neurosci 10:1103-1109 Carola V, Frazzetto G, Pascucci T, Audero E, Puglisi-Allegra S, Cabib S, Lesch KP, Gross C (2008) Identifying molecular substrates in a mouse model of the serotonin transporter x environment risk factor for anxiety and depression. Biol Psychiatry 63:840-846 Carter CS, Ahnert L, Grossmann KE, Hrdy SB, Lamb ME, Porges SW, Sachser N (2005) Attachment and Bonding: A New Synthesis. MIT Press, Cambridge/MA Cases O, Seif I, Grimsby J, Gaspar P, Chen K, Pournin S, Müller U, Aguet M, Babinet C, Shih JC, de Maeyer E (1995) Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA. Science 268:1763-1766 Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, McClay J, Mill J, Martin J, Braithwaite A, Poulton R (2003) Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301:386-389 Champagne FA, Curley JP (2005) How social experiences influence the brain. Curr Opin Neurobiol 15:704-709 Champagne FA, Curley JP (2009) Epigenetic mechanisms mediating the longterm effects of maternal care on development. Neurosci Biobehav Rev 33:593-600 Champagne FA, Francis DD, Mar A, Meaney MJ (2003) Variations in maternal care in the rat as a mediating influence for the effects of environment on development. Physiol Behav 79:359-371 Champagne FA, Weaver IC, Diorio J, Dymov S, Szyf M, Meaney MJ (2006) Maternal care associated with methylation of the estrogen receptor-alpha 1b promoter and estrogen receptor-alpha expression in the medial preoptic area of female offspring. Endocrinology 147:2909-2915

530

Norbert Sachser and Sylvia Kaiser

Chishti MA, Yang D-S, Janus C, Phinney AL, Horne P, Pearson J, Strome R, Zuker N, Loukides J, French J, Turner S, Lozza G, Grilli M, Kunicki S, Morissette C, Paquette J, Gervais F, Bergeron C, Fraser PE, Carlson GA, St. Gerorg-Hyslop S, Westaway D (2001) Early onset amyloid deposition and cognitive deficits in transgenic mice expressing a double mutant form of amyloid precursor protein 695. J Biol Chem 276:21562-21570 Choi S, Kellogg CK (1996) Adolescent development influences functional responsiveness of noradrenergic projections to the hypothalamus in male rats. Dev Brain Res 94:144-151 Cooper RM, Zubek JP (1958) Effects of enriched and restricted early environment on the learning ability of bright and dull rats. Can J Psychol 12:159-164 de Kloet ER, Sibug RM, Helmerhorst FM, Schmidt M (2005) Stress, genes and the mechanism of programming the brain for later life. Neurosci Biobehav Rev 29:271-281 Delville Y, Melloni RH Jr, Ferris CF (1998) Behavioral and neurobiological consequences of social subjugation during puberty in golden hamsters. J Neurosci 18:2667-2672 de Weerth C, Buitelaar JK, Mulder EJH (2005) Prenatal programming of behaviour, physiology and cognition. Neurosci Biobehav Rev 29:207-208 Dloniak SM, French JA, Holekamp KE (2006) Rank-related maternal effects of androgens on behaviour in wild spotted hyaenas. Nature 440:1190-1193 Eising CM, Visser GH, Müller W, Groothuis TG (2003) Steroids for free? No metabolic costs of elevated maternal androgen levels in the black-headed gull. J Exp Biol 206:3211-3218 Elwood RW, Kennedy HF (1991) Selectivity in paternal and infanticidal responses by male mice: effects of relatedness, location, and previous sexual partners. Behav Neural Biol 56:129-147 Francis DD, Meaney MJ (1999) Maternal care and the development of stress responses. Curr Opin Neurobiol 9:128-134 Francis DD, Diorio J, Liu D, Meaney MJ (1999) Nongenomic transmission across generations in maternal behavior and stress responses in the rat. Science 286:1155-1158 Francis DD, Champagne FC, Meaney MJ (2000) Variations in maternal behaviour are associated with differences in oxytocin receptor levels in the rat. J Neuroendocrinol 12:1145-1148 Gandelmann R (1992) Psychobiology of behavioral development. Oxford University Press, New York Giese KP, Fedorov NB, Filipkowski RK, Sivla AJ (1998) Autophosphorylation at Thr286 of the α calcium-calmodulin kinase II in LTP and learning. Science 279:870-873 Görtz N, Lewejohann L, Tomm M, Ambrée O, Keyvani K, Paulus W, Sachser N (2008) Effects of environmental enrichment on exploration, anxiety, and memory in female TgCRND8 Alzheimer mice. Behav Brain Res 191:43-48 Goymann W, Landys MM, Wingfield JC (2007) Distinguishing seasonal androgen responses from male-male andogen responsiveness: revisiting the challenge hypothesis. Horm Behav 51:463-476

The social modulation of behavioural development

531

Graves FC, Hennessy MB (2000) Comparison of the effects of the mother and an unfamiliar adult female on cortisol and behavioral responses of pre- and postweaning guinea pigs. Dev Psychobiol 36:91-100 Gréco B, Edwards DA, Michael RP, Clancy AN (1998) Androgen receptors and estrogen receptors are colocalized in male rat hypothalamic and limbic neurons that express fos immunoreactivity induced by mating. Neuroendocrinology 67:18-28 Groothuis TGG, Müller W, von Engelhardt N, Carere C, Eising C (2005) Maternal hormones as a tool to adjust offspring phenotype in avian species. Neurosci Biobehav Rev 29:329-352 Gross C, Hen R (2004) The developmental origins of anxiety. Nat Rev Neurosci 5:545-552 Harlow HF, Harlow MK (1962) Social deprivation in monkeys. Sci Am 207:136146 Harvey PW, Chevins PFD (1984) Crowding or ACTH treatment of pregnant mice affects adult copulatory behavior of male offspring. Horm Behav 18:101-110 Hayward C (2003) Gender Differences at Puberty. Cambridge University Press, Cambridge Heiming RS, Jansen F, Lewejohann L, Kaiser S, Schmitt A, Lesch KP, Sachser N (2009) Living in a dangerous world: the shaping of behavioral profile by early environment and 5-HTT genotype. Front Behav Neurosci: 3:1-11 Hennessy MB, Ritchey RL (1987) Hormonal and behavioral attachment responses in infant guinea pigs. Dev Psychobiol 20:613-625 Hennessy MB, Mendoza SP, Kaplan JN (1982) Behavior and plasma cortisol following brief peer separation in juvenile squirrel monkeys. Am J Primatol 3:143-151 Hennessy MB, Neisen G, Bullinger KL, Kaiser S, Sachser N (2006a) Social organization predicts nature of infant-adult interactions in two species of wild guinea pigs (Cavia aperea and Galea monasteriensis). J Comp Psychol 120:12-18 Hennessy MB, Hornschuh G, Kaiser S, Sachser N (2006b) Cortisol responses and social buffering: a study throughout the life span. Horm Behav 49:383-390 Hennessy MB, Kaiser S, Sachser N (2009) Social buffering of the stress response: diversity, mechanisms, and functions. Front Neuroendocrinol 30:470-482 Herrenkohl LR (1979) Prenatal stress reduces fertility and fecundity in female offspring. Science 206:1097-1099 Hoffman KA, Mendoza SB, Hennessy MB, Mason, WA (1995) Responses of infant titi monkeys, Callicebus moloch, to removal of one or both parents: evidence for paternal attachment. Dev Psychobiol 28:399-407 Holmes A, Li Q, Murphy DL, Gold E, Crawley JN (2003a). Abnormal anxietyrelated behavior in serotonin transporter null mutant mice: the influence of genetic background. Genes Brain Behav 2:365-380 Holmes A, Murphy DL, Crawley JN (2003b) Abnormal behavioral phenotypes of serotonin transporter knockout mice: parallels with human anxiety and depression. Biol Psychiatry 54:953-959

532

Norbert Sachser and Sylvia Kaiser

Immelmann K, Barlow G, Petrinovich L, Main M (1981) Behavioral Development: The Bielefeld Interdisciplinary Project. Cambridge University Press, Cambridge Jablonka E, Raz G (2009) Transgenerational epigenetic inheritance: prevalence, mechanisms, and implications for the study of heredity and evolution. Q Rev Biol 84:131-176 Jansen F, Heiming RS, Lewejohann L, Touma C, Palme R, Schmitt A, Lesch KP, Sachser N (2010) Modulation of behavioral profile and stress response by 5HTT genotype and social experience in adulthood. Behav Brain Res 207:2129 Janus C, Welzl H, Hanna A, Lovasic L, Lane N, St George-Hyslop P, Westaway D (2004) Impaired conditioned taste aversion learning in APP transgenic mice. Neurobiol Aging 25:1213-1219 Kaiser S, Sachser N (1998) The social environment during pregnancy and lactation affects the female offsprings’ endocrine status and behaviour in guinea pigs. Physiol Behav 63:361-366 Kaiser S, Sachser N (2001) Social stress during pregnancy and lactation affects in guinea pigs the male offsprings’ endocrine status and infantilizes their behaviour. Psychoneuroendocrinology 26:503-519 Kaiser S, Sachser N (2005) The effects of prenatal social stress on behaviour: mechanisms and function. Neurosci Biobehav Rev 29:283-294 Kaiser S, Sachser N (2009) Effects of prenatal social stress on offspring development: pathology or adaptation? Curr Dir Psychol Sci 18:118-121 Kaiser S, Kruijver FPM, Swaab DF, Sachser N (2003a) Early social stress in female guinea pigs induces a masculinization of adult behavior and corresponding changes in brain and neuroendocrine function. Behav Brain Res 144:199210 Kaiser S, Kruijver FPM, Straub RH, Sachser N, Swaab DF (2003b) Early social stress in male guinea pigs changes social behaviour, and autonomic and neuroendocrine functions. J Neuroendocrinol 15:761-769 Kaiser S, Heemann K, Straub RH, Sachser N (2003c) The social environment affects behaviour and androgens, but not cortisol in pregnant female guinea pigs. Psychoneuroendocrinology 28:67-83 Kaiser S, Haderthauer S, Sachser N, Hennessy MB (2007) Social housing conditions around puberty determine later changes in plasma cortisol levels and behavior. Physiol Behav 90:405-411 Kemme K, Kaiser S, Sachser N (2007) Prenatal maternal programming determines testosterone response during social challenge. Horm Behav 51:387-394 Kemme K, Kaiser S, Sachser N (2008) Prenatal stress does not impair coping with challenge later in life. Physiol Behav 93:68-75 Kemme K, Kaiser S, von Engelhardt N, Wewers D, Groothuis T, Sachser N (2009) An unstable social environment affects sex ratio in guinea pigs: an adaptive maternal effect? Behaviour 146:1513-1529 Klein SL, Rager DR (1995) Prenatal stress alters immune function in the offspring of rats. Dev Psychobiol 28:321-336

The social modulation of behavioural development

533

Korte SM, Koolhaas JM, Wingfield JC, McEwen BS (2005) The Darwinian concept of stress: benefits of allostasis and costs of allostatic load and the tradeoffs in health and disease. Neurosci Biobehav Rev 29:3-38 Laviola G, Macrì S, Morley-Fletcher S, Adriani W (2003) Risk-taking behavior in adolescent mice: psychobiological determinants and early epigenetic influence. Neurosci Biobehav Rev 27:19-31 Lesch K-P, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S, Benjamin J, Müller CR, Hamer DH, Murphy DL (1996) Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 274:1527-1531 Lewejohann L, Reefmann N, Widmann P, Ambrée O, Herring A, Keyvani K, Paulus W, Sachser N (2009) Transgenic Alzheimer mice in a semi-naturalistic environment: more plaques, yet not compromised in daily life. Behav Brain Res 201:99-102 Lieberman MW (1963) Early developmental stress and later behaviour. Science 141:824-825 Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, Freedman A, Sharma S, Pearson D, Plotsky PM, Meaney MJ (1997) Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal response to stress. Science 277:1659-1662 Lürzel S, Kaiser S, Sachser N (2010) Social interaction, testosterone, and stress responsiveness during adolescence. Physiol Behav 99:40-46 Maccari S, Piazza PV, Kabbaj M, Barbazanges A, Simon H, Le Moal M (1995) Adoption reverses the long-term impairment in glucocorticoid feedback induced by prenatal stress. J Neurosci 15:110-116 Macrí S, Adriani W., Chiarotti F, Laviola G (2002) Risk taking during exploration of a plus-maze is greater in adolescent than in juvenile or adult mice. Anim Behav 64:541-546 Maestripieri D, Lindell SG, Ayala A, Gold PW, Higley JD (2005) Neurobiological characteristics of rhesus macaque abusive mothers and their relation to social and maternal behaviour. Neuosci Biobehav Rev 29:51-57 McCormick CM, Furey BF, Child M, Sawyer MJ, Donohue SM (1998) Neonatal sex hormones have ‘organizational’ effects on the hypothalamic-pituitaryadrenal axis of male rats. Dev Brain Res 105:295-307 Meaney MJ (2001) Maternal care, gene expression, and the transmission of individual differences in stres reactivity across generations. Annu Rev Neurosci 24:1161-1192 Mendoza SP, Mason WA (1986) Parental division of labour and differentiation of attachments in a monogamous primate (Callicebus moloch). Anim Behav 34:1336-1347 Mousseau TA, Fox CW (1998) The adaptive significance of maternal effects. Trends Ecol Evol 13:403-407 Nelson RJ, Demas GE, Huang PL, Fishman MC, Dawson VL, Dawson TM, Snyder SH (1995) Behavioural abnormalities in male mice lacking neuronal nitric oxide synthase. Nature 378:383-386

534

Norbert Sachser and Sylvia Kaiser

Phoenix CH, Goy RW, Gerall AA, Young WC (1959) Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology 65:369-382 Primus RJ, Kellogg CK (1990) Gonadal hormones during puberty organize environment-related social interaction in the male rat. Horm Behav 24:311-323 Qvarnström A, Price TD (2001) Maternal effects: paternal effects and sexual selection. Trends Ecol Evol 16:95-100 Romeo RD, Richardson HN, Sisk CL (2002) Puberty and the maturation of the male brain and sexual behavior: recasting a behavioral potential. Neurosci Biobehav Rev 26:381-391 Sachser N (1986) Different forms of social organization at high and low population densities in guinea pigs. Behaviour 97:253-272 Sachser N (1990) Social organization, social status, behavioural strategies and endocrine responses in male guinea pigs. In: Balthazart J (ed) Hormones, Brain and Behaviour in Vertebrates. 2: Behavioural Activation in Males and Females – Social Interaction and Reproductive Endocrinology. Karger, Basel, pp 176-187 Sachser N (1993) The ability to arrange with conspecifics depends on social experiences around puberty. Physiol Behav 53:539-544 Sachser N (1998) Of domestic and wild guinea pigs: studies in sociophysiology, domestication, and social evolution. Naturwissenschaften 85:307-317 Sachser N (2001) What is important to achieve good welfare in animals? In: Broom DM (ed) Coping with Challenge: Welfare in Animals Including Humans. Dahlem Workshop Report 87. Dahlem University Press, Berlin, pp 3148 Sachser N, Kaiser S (1996) Prenatal social stress masculinizes the females´ behaviour in guinea pigs. Physiol Behav 60:589-94 Sachser N, Lick C (1991) Social experience, behavior, and stress in guinea pigs. Physiol Behav 50:83-90 Sachser N, Pröve E (1988) Plasma-testosterone development in colony and individually housed male guinea pigs. Ethology 79:62-70 Sachser N, Renninger SV (1993) Coping with social conflict: the role of rearing conditions in guinea pigs. Ethol Ecol Evol 5:65-74 Sachser N, Lick C, Stanzel K (1994) The environment, hormones and aggressive behaviour – a five-year-study in guinea pigs. Psychoneuroendocrinology 19:697-707 Sachser N, Dürschlag M, Hirzel D (1998) Social relationships and the management of stress. Psychoneuroendocrinology 23:891-904 Schinka JA, Busch RM, Robichaux-Keene N (2004) A meta-analysis of the association between the serotonin transporter gene polymorphism (5-HTTLPR) and trait anxiety. Mol Psychiatry 9:197-202 Schwabl H (1996) Maternal testosterone in the avian egg enhances postnatal growth. Comp Biochem Physiol A 114:271-276 Schwabl H (2001) Sex differences in avain yolk hormone levels. Nature 412:498

The social modulation of behavioural development

535

Seale JV, Wood SA, Atkinson HC, Harbuz MS, Ligthman SL (2004) Gonadal steroid replacement reverses gonadectomy-induced changes in the corticosterone pulse profile and stress-induced hypothalamic-pituitary-adrenal axis activity of male and female rats. J Neuroendocrinol 16:989-998 Seckl JR, Cleasby M, Nyirenda MJ (2000) Glucocorticoids, 11betahydroxysteroid dehydrogenase, and fetal programming. Kidney Int 57:14121417 Shen H, Gong QH, Aoki C, Yuan M, Ruderman Y, Dattilo M, Williams K, Smith SS (2007) Reversal of neurosteroid effects at alpha4beta2delta GABAA receptors triggers anxiety at puberty. Nat Neurosci 10:469-477 Sisk CL, Foster DL (2004) The neural basis of puberty and adolescence. Nat Neurosci 7:1040-1047 Sisk CL, Zehr JL (2005) Pubertal hormones organize the adolescent brain and behavior. Front Neuroendocrinol 26:163-174 Smotherman WP, Hunt LE, McGinnis LM, Levine S (1979) Mother-infant separation in group living rhesus macaques: a hormonal analysis. Dev Psychobiol 12:211-217 Spear LP (2000) The adolescent brain and age-related behavioral manifestations. Neurosci Biobehav Rev 24:417-463 Spinelli S, Schwandt ML, Lindell SG, Newman TK, Heilig M, Suomi SJ, Higley JD, Goldman D, Barr CS (2007) Association between the recombinant human serotonin transporter linked promoter region polymorphism and behavior in rhesus macaques during a separation paradigm. Dev Psychopathol 19:977-987 Steinberg L (2005) Cognitive and affective development in adolescence. Trends Cogn Sci 9:69-74 Touma C, Palme R, Sachser N (2004) Analyzing corticosterone metabolites in fecal samples of mice: a noninvasive technique to monitor stress hormones. Horm Behav 45:10-22 Trimpop RM, Kerr JH, Kirkcaldy B (1999) Comparing personality constructs of risk-taking behavior. Pers Individ Dif 26:237-254 Tsankova NM, Berton O, Renthal W, Kumar A, Neve RL, Nestler EJ (2006) Sustained hippocampal chromatin regulation in a mouse model of depression and antidepressant action. Nat Neurosci 9:519-525 Uhlhaas PJ, Roux F, Singer W, Haenschel C, Sireteanu R, Rodriguez E (2009) The development of neural synchrony reflects late maturation and restructuring of functional networks in humans. Proc Natl Acad Sci USA 106:98669871 van Dellen A, Blakemore C, Decaon R, York D, Hannan AJ (2000) Delaying the onset of Huntington´s in mice. Nature 404:721-722 Veenema AH (2009) Early life stress, the development of aggression and neuroendocrine and neurobiological correlates: What can we learn from animal models? Front Neuroendocrinol 30:497-518 Viau V, Meaney MJ (1996) The inhibitory effect of testosterone on hypothalamicpituitary-adrenal responses to stress is mediated by the medial preoptic area. J Neurosci 16:1866-1876

536

Norbert Sachser and Sylvia Kaiser

Vitaterna MH, King DP, Chang AM, Kornhauser JM, Lowrey PL, McDonald JD, Dove WF, Pinto LH, Turek FW, Takahashi JS (1994) Mutagensis and mapping of a mouse gene clock, essential for circadian behavior. Science 264:719725 vom Saal FS, Howard LS (1982) The regulation of infanticide and parental behaviour: implications for reproductive success in male mice. Science 215:12701272 von Holst D (1998) The concept of stress and its relevance for animal behavior. Adv Stud Behav 27:1-131 Ward IL (1972) Prenatal stress feminizes and demasculinizes the behavior of males. Science 175:82-84 Weaver ICG, Cervoni N, Champagne FA, D´Alessio AC, Sharma S, Seckl JR, Dymov S, Szyf M, Meaney MJ (2004) Epigenetic programming by maternal behavior. Nat Neurosci 7:847-854 Weinstock M, Poltyrev T, Schorer-Apelbaum, D, Men D, McCarty R (1998) Effect of prenatal stress on plasma corticosterone and catecholamines in response to footshock in rats. Physiol Behav 64:439-444 Wingfield JC, Hegner RE, Dufty AM Jr, Ball GF (1990) The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. Am Nat 136:829-846 Winslow JT, Noble PL, Lyons CK, Sterk SM, Insel TR (2003) Rearing effects on cerebrospinal fluid oxytocin concentration and social buffering in rhesus monkeys. Neuropsychopharmacol 28:910-918 Wolf M, van Doorn GS, Weissing FJ (2008) Evolutionary emergence of responsive and unresponsive personalities. Proc Natl Acad Sci USA 105:1582515830 Young LJ, Nilsen R, Waymire KG, MacGregor GR, Insel TR (1999) Increased affiliative response to vasopressin in mice expressing the V1a receptor from a monogamous vole. Nature 400:766-768 Yurgelun-Todd D (2007) Emotional and cognitive changes during adolescence. Curr Opin Neurobiol 17:251-257

Chapter 18

Alternative reproductive tactics and life history phenotypes MICHAEL TABORSKY AND H. JANE BROCKMANN

ABSTRACT Alternative reproductive tactics (ARTs) coexisting within a population are found in many organisms. Their existence has been an enduring puzzle in evolutionary biology. Why should selection produce distinctly different alternatives to reach the same goal? How can such alternative solutions coexist in a population? What determines their evolutionary stability? Here we outline ultimate and proximate mechanisms responsible for the origin, coexistence and stability of ARTs. We argue that behavioural and reproductive polymorphisms often reflect different allocation decisions in response to trade-offs in reproduction or life-history optima that may involve heritable threshold responses to environmental variation. Alternative tactics may either be fixed for life or plastic, with simultaneous or sequential switches between tactics. General principles include disruptive selection, negative frequency dependence, density dependence, and an interaction between genetic and environmental components to generate alternative tactics. ARTs are found often where individuals invest heavily in reproduction in a way that can be circumvented and exploited by competitors, which reflects disruptive selection on reproductive investment. This often coincides with consistent size variation between individuals pursuing bourgeois and parasitic tactics.

18.1 Introduction Alternative reproductive tactics refer to alternative ways to obtain fertilisations or, more generally, to reproduce (Taborsky et al. 2008). They are part of a much broader category of alternative phenotypes that include alternative life-history tactics, sex allocation, mimicry, polyphenism, polyethism and social insect castes, all of which are characterised by bimodal or mul-

538

Michael Taborsky and H. Jane Brockmann

timodal trait distributions (Lloyd 1987, Brockmann 2001, Brockmann and Taborsky 2008). ARTs are characterised by a discontinuous distribution of reproductive traits between individuals of the same sex. These distinct traits may include behavioural, morphological and physiological differences that form a stable and regular pattern within a population. For example, within a population of bluegill sunfish, Lepomis macrochirus, there are three kinds of males, ‘bourgeois’ males that construct and guard nests, and two types of ‘parasitic’ males that do not guard but exploit the effort of others (Gross 1982; BOX 18.1). In some scarab beetles (Onthophagus sp.), large males develop horns and guard the entrance to a tunnel containing a female, while small, hornless males sneak towards the guarded female for mating by digging a side tunnel (Emlen 1997). Female Sphecid digger wasps may either build a nest or usurp a nest of a conspecific female to deposit and raise their brood (Brockmann and Dawkins 1979). Males of some damselflies, katydids, wasps and bees, swordtails and guppies, frogs and toads, lizards, ruffs, ungulates and pinnipeds may either court females or intercept them on the mating ground to obtain copulations (see reviews in Oliveira et al. 2008a). Such variation within populations is a puzzle that demands an evolutionary explanation. What are the selective pressures that maintain such a pattern at stable frequencies across generations? Such variation is also puzzling from a developmental and physiological point of view. What, for example, are the hormonal, neural and metabolic mechanisms that can cause males to follow one of two different life-history trajectories? In this chapter we address the maintenance of variation from both, a proximate and an ultimate perspective through the study of ARTs. ARTs are particularly common when there is investment to be exploited by same-sex competitors (Brockmann and Dawkins 1979, Dominey 1984, Waltz and Wolf 1984, Taborsky 1994, 2001, Tallamy 2005). In principle, this is possible in both sexes, but because of the higher investment of females (the burden of anisogamy), ARTs evolve more often in the male sex. According to our current understanding, ARTs evolve most commonly when there is fitness to be gained by pursuing different reproductive tactics and when the intermediate expressions of a reproductive trait are selected against. In the bluegill sunfish, for example (BOX 18.1), males of intermediate size are selected against (disruptive selection) because they are too small to compete with the large, bourgeois and territorial male tactic and too large to gain access to females through the surreptitious, parasitic tactic (Gross 1982), which is a general pattern in fishes and other taxa (Taborsky 1997, 2008, Wolff 2008). In this case the frequency of each morph is thought to be set by frequency-dependent selection. ARTs also evolve

Alternative reproductive tactics and life history phenotypes

539

BOX 18.1 Bluegill sunfish: ARTs fixed for life and sequential ARTs A. ♀ 0

1

2

3

Parental ♂

4

Bourgeois tactic

B. ♂ 0

1

2

3

4

5

Parental

6

7

8

9

10

11

♂ ♀

Parasitic tactic

Female mimic Sneaker

Schematic representation of life histories and reproductive behaviour of females and alternative types of males in bluegill sunfish. Numbers denote age in years, shaded arrows indicate maturity (Picture by Brian Neff with permission).

In the bluegill sunfish, Lepomis macrochirus, males develop into either a bourgeois or parasitic morph. These tactics are fixed for life. Males pursuing the parasitic pathway mature sexually when still small, at about 2 years of age, and remain parasitic throughout life (Gross 1982). In one population, they grow much slower than males pursuing the bourgeois pathway during the years 2-4 of their lives (Taborsky 1994). Burgeois males mature only at an age of about 7 years, when they start to construct nests in densely packed colonies, and court and spawn sequentially with multiple females. They care for the developing eggs and fry in their nests; hence, they have been termed ‘parentals’ (Gross 1982). Parasitic males start with a sneaking tactic at an age of 2-3 years, hiding near nests of parentals from where they dart into the nest to steal fertilisations during female spawning. Once parasitic males reach the size of mature females at an age of 4-5 years, they mimic females to mislead the nest owner and participate when a female is spawning in his nest (Gross 1982, Neff and Gross 2001). At another population, both, satellites behaving as female mimics and bourgeois males reproduce at an average age of 6 years (Dominey 1980), underscoring the fact that parasitic and bourgeois phenotypes reflect fixed lifetime tactics. In contrast, sneaker behaviour is performed as a transient tactic early in the reproductive lives of satellite males. Offspring from parasitic males grow quicker (Neff 2004), but it is yet unknown whether and to what extent the bourgeois and parasitic tactics are heritable (Neff 2008). An ontogenetic switch mechanism involving a size threshold reaction norm seems likely.

540

Michael Taborsky and H. Jane Brockmann

when there are different reproductive niches that favour divergent tactics and specialisation for exploiting each niche. Such alternative tactics might be created in males, for example, if there were a polymorphism in female preference traits (Henson and Warner 1997, Alonzo and Warner 2000, Alonzo and Sinervo 2001, Morris et al. 2003, Neff 2008) or if some males exploited females at aggregated emergence sites whereas other males exploited widely dispersed females (Brockmann 2008). Disruptive selection would operate in these cases as well because intermediate phenotypes would not be as effective as specialised ones when exploiting the available options (Skúlason and Smith 1995, Smith and Skúlason 1996, Shuster and Wade 2003, Brockmann and Taborsky 2008). Here the frequency of morphs depends on the reproductive potential in each niche (Zera and Rankin 1989, Denno 1994, Langellotto and Denno 2001). Many species show continuous variation in size or sexually selected traits but ARTs evolve only when disruptive selection favours discrete phenotypes by selection against intermediates (Danforth and Desjardins 1999). Investment in gaining privileged access to mates or fertilisable gametes usually bears costs (e.g. Taborsky et al. 1987, Simmons et al. 1992, Plaistow and Tsubaki 2000, Wagner 2005). These costs may involve (i) the production and display of conspicuous signals suited to attract mates and repel rivals, which may also lure predators and competitors (Andersson 1994); (ii) the construction of costly structures for mate attraction, defence, or brood care (Hansell 2005, Schädelin and Taborsky 2009); or (iii) parental investment to protect, provision, and raise offspring (Clutton-Brock 1991). Individuals using surreptitious and parasitic tactics can omit these costs and in many cases exploit the investment of bourgeois conspecifics to gain access to mates or fertilisable gametes (reviewed in Taborsky 1994, Oliveira et al. 2008a). Often they use secretive or quick behavioural tactics that cannot be easily overcome by the exploited bourgeois individuals (Gross 1982, Correa et al. 2003; reviewed by Taborsky 1994, 2008 for fish and Westneat and Stewart 2003 for birds). Alternatively, individuals using exploitive reproductive tactics may take by force the resources needed for mating or brood care from bourgeois individuals (van den Berghe 1988, Sinervo and Lively 1996, Mboko and Kohda 1999). In a few species, competitive individuals defend areas that females frequent, such as leks, which otherwise lack obvious resources; other males intercept females as they arrive at the breeding ground (Thornhill and Alcock 1983, Zamudio and Chan 2008). Also, when some females are widely dispersed and others are clumped, there are opportunities for some males to gain fitness by searching and others by staying near the aggregation and fighting over females (Brockmann 2008). In principle, when acquiring mates is costly, limited resources are allocated using evolved decision rules that maximise repro-

Alternative reproductive tactics and life history phenotypes

541

ductive success in the face of inevitable trade-offs. ARTs evolve when those allocation rules involve mutually exclusive adaptations. Among females, individuals pursuing parasitic tactics often benefit from exploiting the maternal effort of conspecifics, for example by using nests built by other females (e.g. in digger wasps, Sphex ichneumoneus: Brockmann and Dawkins 1979, Brockmann et al. 1979, Field 1994) or by dumping eggs in a host female’s nest (or mouth) that will be cared for by the host (i.e. intraspecific brood parasitism in insects: Brockmann 1993, Zink 2003, Tallamy 2005; fish: Ribbink 1977, Yanagisawa 1985, Kellog et al. 1998; birds: Yom-Tov 1980, 2001, Petrie and Møller 1991). Thereby parasitic females save the effort required to prepare breeding sites and care for their brood (Sandell and Diemer 1999), or they can raise their productivity at the expense of their competitors (Tallamy and Horton 1990, Ahlund and Andersson 2001, Zink 2003). Female ARTs may also arise as a way to avoid sexual conflict (Alonzo 2008). In Ischnura ramburi and some other damselflies, females avoid long and costly mating interactions with males either by mimicking males or by avoiding males through cryptic appearance and behaviour (Sirot et al. 2003). As with male ARTs, distinct alternative tactics in females involve trade-offs and disruptive selection against the intermediates. ARTs are usually thought of as competitive but they may involve cooperative behaviour between competitors (Packer 1977, Harcourt and de Waal 1992, Taborsky 1994, 2001, Watts 1998, Feh 1999, Connor et al. 2000, Awata et al. 2005). Competing individuals may cooperate or ‘trade’ with resource holders by paying for access to reproductive options by mutualism or reciprocity (Taborsky 1985, Taborsky et al. 1987, Martin and Taborsky 1997, Oliveira et al. 2002, Dickinson 2004, Heg et al. 2009; BOX 18.2). These relationships are usually characterised by asymmetries in the resource-holding potential of individuals pursuing alternative tactics. The mechanisms regulating and stabilising such cooperative associations between reproductive competitors are a challenge for evolutionary theory and require experimental scrutiny (Vehrencamp 1983, Keller and Reeve 1994, Johnstone 2000, Skubic et al. 2004, Bergmüller et al. 2005, Stiver et al. 2005, Taborsky 2009). The evolution of ARTs is influenced by interactions between the sexes as well as competitive interactions within a sex (Alonzo 2008). For example, parasitic tactics often appear to circumvent not only the attacks of territorial males but also female choice (Taborsky 1994). When this is the case, females and bourgeois males have similar interests in preventing extra-pair mating whereas females and sneaker males have conflicting interests (Neff 2008). On the other hand, multiple mating may be advantageous to females, such as when the presence of sneaker males ensures that

542

Michael Taborsky and H. Jane Brockmann

BOX 18.2 Cooperation between reproductive competitors – a case of trading care with a share in reproduction In the cooperatively breeding cichlid Neolamprologus pulcher, both male and female subordinates help to raise the broods of dominant breeders of a group. They jointly defend a territory, where they dig out and maintain shelters to hide and breed (Taborsky and Limberger 1981, Taborsky 1984). Mature helpers may attempt to participate in reproduction (Taborsky 1985, Dierkes et al. 1999, Heg et al. 2006, 2008), but the reproductive tactics are sex-specific. Male subordinates attempt to steal fertilisations by instantaneous darts into the breeding shelter during spawning (Dierkes et al. 1999, Mitchell et al. 2009a). In contrast, subordinate females may produce clutches when they have separate shelters available in the territory (Heg and Hamilton 2008, Heg et al. 2009), posing differential costs for male and female dominant breeders (Taborsky 1985). Male breeders and helpers compete for fertilisations, i.e., each egg fertilised by a male subordinate is lost for the dominant male breeder, because male breeders and large helpers are rarely related to each other (Dierkes et al. 2005; Figure below). In contrast, a female breeder is not loosing offspring when a subordinate female produces an own clutch within the territory (Heg et al. 2008), although there may be competition between females and their offspring for resources and consequently mutual egg cannibalism (Heg and Hamilton 2008), which raises the conflict between male and female dominants (Mitchell et al. 2009a). Both, male and female dominants profit from the presence and effort of subordinates in the group by increased productivity (Taborsky 1984) and offspring survival (Brouwer et al. 2005). to female breeders to male breeders

Genetic relatedness

0.5 0.4 0.3 0.2 0.1 0.0 -0.1 0.0

1.0

2.0

3.0

4.0

5.0

6.0

Body length [SL cm]

Relatedness of subordinate members of N. pulcher groups to male and female dominant breeders (means + SEM and regression lines from a GLMM), as a function of the subordinate’s body size; reproductive maturity is reached at about 3.5 cm. Subordinates are hardly related to dominant male breeders, due to the high natural exchange rate of the latter (Dierkes et al. 2005).

Alternative reproductive tactics and life history phenotypes

543

Experiments revealed that helpers pay to stay in the territory of dominant breeders (Taborsky 1985, Balshine-Earn et al. 1998, Bergmüller and Taborsky 2005, Bergmüller et al. 2005, Bruintjes and Taborsky 2008), which serves to increase their survival chances due to the protection they get (Taborsky 1984, Heg et al. 2004a), but is also a reproductive tactic (Taborsky 1985, Dierkes et al. 1999, 2008, Skubic et al. 2004, Heg et al. 2009). Especially male subordinates gain from producing offspring in the dominant’s territory (Dierkes et al. 1999), and they are more successful if other male helpers are present (Heg et al. 2008). However, apparently due to the costs this entails for male breeders, the latter suppress the growth of large male helpers (Heg et al. 2004b), and to some degree also their gonadal development (Fitzpatrick et al. 2006). This is not the case for females, where subordinates do not entail noticeable fitness costs to dominants and where growth is not suppressed (Hamilton and Heg 2008, Heg et al. 2008). Hence in this species, reproductive competitors join forces to raise offspring, but they rival for their production, especially in the male sex. Careful examination of empirical evidence revealed that conventional reproductive skew models cannot account for the complexity of evolutionary mechanisms involved in reproductive participation among group members of this species (Taborsky 2009). Instead, a dynamic programming approach using body size as a state variable might provide a better model for the decisions of male subordinates to reproduce in the group or not (Skubic et al. 2004). Males of N. pulcher show sequential ARTs. They start to reproduce early, usually when still staying in their natal group (Stiver et al. 2004). However, their reproductive effort in testes and behaviour is much lower than when they have obtained a dominant breeding position later in life (Fitzpatrick et al. 2006, Mitchell et al. 2009b). Their early reproductive activity reflects an opportunistic tactic while growing up in a safe territory, in dependence of the costs of expulsion they risk by their participation in reproduction. Once they manage to take over a territory as dominant breeder, they will do so (Balshine-Earn et al. 1998, Stiver et al. 2006) and raise their reproductive effort accordingly (Fitzpatrick et al. 2006).

all the female’s eggs will be fertilised (Jennions and Petrie 2000). If this is the case then selection will favour collaboration between females and sneakers and conflict with bourgeois males (Reichard et al. 2007). In general, female behaviour alters the benefits and costs of male phenotypes and in this way shapes the evolution of male ARTs (Henson and Warner 1997, Alonzo 2008). Alternative tactics may be performed by individuals either simultaneously (the choice of tactic depending on circumstances), in succession during different life stages (sequential tactics), or ARTs may be fixed for the entire life of an animal (Taborsky 1998, Brockmann 2001, Taborsky et al. 2008; Fig. 18.1). Simultaneous and sequential ARTs result from flexible or

544

Michael Taborsky and H. Jane Brockmann

Fixed

ART

Simultaneous Plastic Fixed sequence Sequential Immature Burgeois phenotype

Reversible sequence

Parasitic phenotype

Fig. 18.1 Alternative reproductive tactics can be fixed over the lifetime of an animal or plastic. Plastic tactics may be performed at the same time interval (simultaneous ARTs) or in a fixed or reversible sequence (sequential ARTs). Examples are given in the text. (From Taborsky et al. 2008)

plastic responses to conditions (BOX 18.3). Such phenotypic plasticity is favoured if individual fitness is correlated with conditions varying either with regard to the physical or social environment of an animal, or its own physical condition (West-Eberhard 2003). If conditions change predictably with ontogeny or with size, which applies in particular to organisms with indeterminate growth, a sequential expression of tactics may be favoured (Warner et al. 1975, Magnhagen 1992, Dierkes et al. 1999, Alonzo et al. 2000, Utami et al. 2002). If conditions change either rarely during a lifetime or if the change occurs unpredictably, then fixed tactics may result (Shuster and Wade 2003). Highly unpredictable conditions, for example regarding partner availability, competition and tactic-specific risk, may favour also simultaneous ARTs, which are common in fishes (Taborsky 1994), anurans (Zamudio and Chan 2008), and birds (Westneat and Stewart 2003). The existence (and coexistence) of fixed and flexible tactics is also influenced by tactic specific success and the costs of plasticity (Brockmann 2001, Plaistow et al. 2004).

Alternative reproductive tactics and life history phenotypes

545

BOX 18.3 Definitions of terms ‘Tactic’ vs. ‘Strategy’: In evolutionary game theory models (Maynard Smith 1982), ‘strategy’ denotes a particular life history pattern or ‘genetically based program’ (Gross 1996), and tactic classifies the application of rules that are part of a strategy (i.e. the phenotype; Shuster and Wade 2003). However, with empirical data our potential for inference is usually limited to the phenotype level. Information about underlying genotypes is virtually always missing. Moreover, the difference made between phenotypic traits produced by same or different genotypes ignores the fact that virtually all phenotypic traits are the product of genotypic and environmental influence (Scheiner 1993). Hence, in reality the borders between the concepts underlying the terms ‘strategy’ and ‘tactic’ are vague and flexible, and the underlying mechanisms are usually unknown. Therefore, an operational use of terms is preferable to one encumbered with functional implications: ‘tactic’ and ‘strategy’ are synonymous terms, but in the context of ARTs ‘tactic’ is a less equivocal expression because researchers mainly deal with phenotypes, and because of the connotations of the term strategy as outlined. ‘Fixed’ vs. ‘Plastic’ vs. ‘Conditional’ Tactics: ‘Fixed tactics’ refer to cases where individuals retain a specific phenotype throughout life. In contrast, ‘plastic tactics’ allow for a flexible response to the environment (Fig. 18.1). If alternative phenotypes can be expressed at the same time, the choice of tactic is conditional on instantaneous circumstances – hence the term ‘conditional tactics’. If tactic expression is sequential, it does usually not depend on present conditions, but on conditions during earlier stages of life, or on the passing of a criterion (e.g., a threshold body size). Therefore, these tactics are not ‘conditional’ in the strict sense of the term, even though the switch points may be influenced by conditions (especially when the sequence is reversible). If these tactics are denoted ‘conditional’, fixed tactics should also be thus termed, at least sometimes, as tactic choice may also depend on some threshold criterion relating to conditions early in ontogeny. Therefore, the term ‘conditional tactic’ is somewhat equivocal, because in the broad sense of the term it refers to virtually all cases of ARTs, whereas in the strict sense it is reserved for cases where different phenotypes can be expressed by an individual at a time, i.e., to plastic ARTs with simultaneous tactic flexibility (see Fig. 18.1). ‘Bourgeois’ vs. ‘Parasitic’ Tactics: The term ‘bourgeois’ tactic refers to individuals investing in privileged access to mates, by behavioural (e.g., defence, courtship, nest building, brood care), physiological (e.g., pheromones), or morphological means (e.g., secondary sexual characters). In contrast, ‘parasitic’ tactic denotes individuals exploiting the investment of bourgeois conspecifics. In general discussions

546

Michael Taborsky and H. Jane Brockmann

of the function of alternative tactics, these terms are preferable to the more descriptive terms often used in particular case studies (e.g., guarders, primary males, parentals, or sneakers, streakers, satellites, and cuckolders; Taborsky 1997). It is important to note, however, that not all alternative reproductive tactics fall into this dichotomy of bourgeois ‘investor’ and parasitic ‘exploiter’, as mutualistic or reciprocal relationships between reproductive competitors are also possible (cf. BOX 18.2).

18.2 Evolutionary mechanisms 18.2.1 Threshold responses and selection on switch points Above we argued that alternative phenotypes reflect different allocation decisions in response to trade-offs in reproduction or life-history optima. Different decisions often involve threshold responses to environmental variation that show heritable variation and are the product of selection like other heritable traits (Roff 1996, Tomkins and Brown 2004). Alternative phenotypes may result from threshold responses of one genotype to environmental changes (reaction norm), two or more genotypes reacting differently to an environmental threshold value (genetic polymorphism), or from a genotype expressing different trait values or tactics depending on their own phenotypic transitions during a lifetime (condition-dependent or status-dependent switch, such as when tactics change with size or age; Dempster and Lerner 1950, Falconer and Mackay 1996, Shuster 2008). A polygenic model that assumes genetic variation in reaction norm switch points (and the switch points are defined by the environment at which a phenotype switches from one alternative to another) reveals that the evolutionary outcome depends primarily on two parameters: (i) the probability density function of the environmental variation acting on the switch points that are responsible for the production of the alternative phenotypes, and (ii) the magnitude of the fitness trade-offs of the phenotypes across this environmental variation (Hazel et al. 1990). Other factors may also influence the evolution of alternative phenotypes, including the effects of other switch points (more than one threshold mechanism may be involved in the production of phenotypic polymorphisms; Rowland and Emlen 2009) and genetic, developmental and physiological trade-offs or coupling that cause selection on one morph to act also on the other (Tomkins et al. 2005, Tomkins and Moczek 2009; cf. Falconer 1952). As with sexual dimorphism, the action of selection on one alternative phenotype (male) can af-

Alternative reproductive tactics and life history phenotypes

547

fect the phenotype of the other (female) because of genetic correlations (antagonistic pleiotropy) or linkage disequilibrium (Roff 1990). This will favour the evolution of mechanisms that break up the genetic correlations between alternative phenotypes and favour phenotype-limited traits (Via and Lande 1985). Once this has occurred, unique adaptations for each of the alternative phenotypes can evolve (Crnokrak and Roff 1998, Brockmann and Taborsky 2008). Behavioural polymorphisms based on heritable threshold responses have been described in several species. For example in scarab beetles, males with alternative morphologies (large or small horns relative to body size) adopt different mating tactics (Emlen 1997, Moczek and Emlen 2000): large males with large horns guard tunnels dug by females for egg deposition and thereby attempt to monopolise access to the female(s) inside. Males not passing a threshold body size develop only rudimentary horns and do not guard tunnels. Instead, they sneak into the tunnels past guarding males or dig side tunnels to bypass them. In the beetle Onthophagus acuminatus, the switch point for male dimorphism was found to be heritable by conducting artificial selection experiments. Emlen (1996) bred those males that had the shortest horn lengths relative to their body size for the ‘low’ line and those with the longest horns relative to body size for the ‘high’ line. All males were mated to random females. The result in both lines was a shift in the body size at which the animals switched from developing small horns to large horns (Fig. 18.2). Heritable, size-dependent, life-history switch points have been found also in Atlantic salmon, suggesting discontinuous reaction norms for age and size at maturity (AubinHorth and Dodson 2004, Piché et al. 2008). Salmon males either swim to the sea to grow large (1-10 kg) and return to their home stream for breeding, typically at an age of 4-8 years (anadromous males); or they reproduce at an age of 1-4 years at a very small size (10-150 g; parr males), without having left their home stream. On the spawning ground, anadromous males search for, court and fight for mates, whereas parr males sneak under the egg-laying female or dart in to shed sperm, thereby parasitising the effort of bourgeois males. It has been assumed that the success of salmon male tactics is strongly size-dependent, with bourgeois males benefitting from large and parasitic males from small body size (Gross 1984, 1985). However, experimental results revealed that size was not important for the reproductive success of anadromous males within the size range tested, and parr males benefitted from large instead of small size (Thomaz et al. 1997, Jones and Hutchings 2002). This suggests that, although both male types gain from being larger, the slopes of body size effects on fitness differ between the two male morphs (Taborsky 1999; Fig. 18.3).

548

Michael Taborsky and H. Jane Brockmann

1

a)

b)

c)

Guarding male Sneaking male

1.0

Female

0.5

0.0

2.5

3.0

3.5

4.0

Horn length

a) Residual horn length

Prothorax width

c) Down lines

4

b) Down lines

Egg Brood

Prothorax width [mm]

Up lines

0.2

Response

Horn length [mm]

1.5

3

Dung

2

Residual horn length

0.1 0.0 -0.1 -0.2 0.0

d) Up lines

Up lines Down lines 0.5

1.0

1.5

Cumulative selection differential

Horn length

Residual horn length

e)

Residual horn length

f)

Horn length [mm]

5 1.2 0.8 0.4 0.0 2.0

Prothorax width

Prothorax width

Up line, gen. 7 Down line, gen. 7

2.5

3.0

3.5

Prothorax width [mm]

4.0

Fig. 18.2 Artificial selection for male ARTs in Onthophagus acuminatus. 1. a) Allometric relationship between body size and horn length. The frequency distribution of body sizes is unimodal, whereas horn length representing male ARTs shows a bimodal distribution; b) hornless and c) horned male phenotypes. 2. Sketch of male tactics in scarab beetles, with bourgeois males guarding tunnels with females and parasitic males sneaking past them to fertilise females. 3. Scheme of the artificial selection procedure for alternative body size/horn length phenotypes. a) Residual horn lengths were calculated for all males as the difference between their horn length and the horn length expected for their body size; the curve shows the nonlinear regression between body size and horn length. b) Males with the largest and smallest residual horn lengths were selected. c) – f) This resulted in respective shifts in the body size/horn length allometry. 4) Linear regressions of mean response to selection for residual horn length. 5) Final horn length/body size distributions for lines selected for long (open circles) and short (closed circles) horns for seven generations. (From Emlen 1996, 2008)

Alternative reproductive tactics and life history phenotypes

Fitness

B

P

Size

549

Fig. 18.3 Selection may cause limited size distributions of bourgeois (B) and parasitic (P) males when the slopes of their fitness functions differ. Even if increasing body size affects fitness positively in both, bourgeois and parasitic tactics, selection may favour parasitic reproductive behaviour in small males (size range marked grey) and bourgeois behaviour in large males (to the right of the intersection). This pattern is probably prevalent. (From Taborsky 1999)

Threshold switches between morphs have been proposed also as explanations for many other traits, for instance environmentally sensitive sex determination (Kraak and Pen 2002), wing polymorphisms in crickets (Zera and Huang 1999), seasonal polyphenisms and partial bivoltinism (Dingle and Winchell 1997, Nijhout 1999), the expression of insect castes (Nijhout and Wheeler 1982), and the age polyethism in honeybees (Whitfield et al. 2003). 18.2.2 Discrete phenotypes and conditional tactics As described above, alternative phenotypes may reflect either life-long individual specialisations or flexible responses to intraspecific competition. The evolutionary mechanisms involved in these alternative scenarios differ substantially. Fixed alternative tactics are characterised by equal average lifetime fitness if they reflect an evolutionarily stable state (Maynard Smith 1982), whereas tactics adopted in response to variation in individual condition, status or environment do not necessarily generate equivalent fitness in the alternative phenotypes (Lively 1986, Repka and Gross 1995, Gross 1996, Tomkins and Hazel 2007). Populations frequently consist of a mixture of conditional and pure tactics (Plaistow et al. 2004). Conditional tactics (cf. BOX 18.3) have been found in a wide range of taxa and in many different functional contexts, being quite likely the most common form of discrete variation within species (West-Eberhard 2003, Oliveira et al. 2008a). Conditional tactics in response to environmental variation, where ‘environment’ also includes competitors and social conditions, have been studied in some acarid mites, where ‘fighter’ and ‘scrambler’ males compete for access to females (Radwan 1993, 1995, Radwan et al. 2002).

550

Michael Taborsky and H. Jane Brockmann

Experimental studies revealed considerable heritability of morph expression due to an adaptive response of the threshold reaction norm in Sancassania berlesei (Tomkins et al. 2004, Unrug et al. 2004). Essential preconditions for conditional tactics to be evolutionarily stable are that (i) there is discrete environmental variation, (ii) environmental cues are reliable (i.e., they allow predictions about the environment that are better than random), and (iii) alternative phenotypes must have different fitness optima in different environments (Lively 1986). In the reproductive context such environments might include different opportunities to access partners, such as the size-dependent access options of males in the bee Centris pallida (Alcock et al. 1977). Here, large males patrol emergence sites to locate females that they defend aggressively, whereas small males wait for females hovering near emergence sites. As is consistent with our understanding of conditional tactic choice, the hovering tactic pursued by small males provides considerably less fitness rewards than the patrolling and fighting tactic adopted by large males, which Dawkins (1980) referred to as pursuing the ‘best of a bad job’. The polygenic environmental threshold model for the evolution and maintenance of conditional tactics developed by Hazel and collaborators suggests that for the conditional tactic ‘if small hover, if large patrol and fight’ to be maintained, there must be a size-dependent fitness trade-off between these phenotypes, such that there is some body size above which ‘patrolling’ provides higher fitness rewards than ‘hovering’, and below which it is the other way round (Hazel et al. 1990, 2004). For such a trait to evolve, it is important that there is heritable genetic variation for the tactic switch point (i.e., the response to the environmental cue) to be subject to natural selection (Dempster and Lerner 1950, Hazel et al. 1990, Tomkins and Hazel 2007; cf. Shuster and Wade 2003). This example is, however, complicated by the fact that offspring size is set by maternal investment tactics: large males are the product of females providing large amounts of food and small males come from reduced female investment. This means that what is played out as a ‘best of a bad job’ tactic among males may reflect alternative maternal investment tactics (Danforth and Neff 1992, Alonzo 2008, Brockmann 2008). In the Mediterranean wrasse Symphodus ocellatus, nest-building, gaudy males attract females to spawn in their nests where they fan and guard the eggs until hatching. Smaller, inconspicuous males adopt two alternative tactics, either opportunistically darting into the nest during spawning (sneakers) or joining a bourgeois male to defend the nest against other reproductive parasites while attempting to fertilise eggs from a privileged position (satellites; Taborsky et al. 1987; Fig. 18.4). Analyses of growth patterns revealed that there are three different life-history pathways in

Alternative reproductive tactics and life history phenotypes

2

0.8

2-year-old

Satellite

Nesting male

0.7

1-year-old

Sneaker Satellite Non-reproductive

3 0.6

Nesting males

Nonreproductives

One-year-old satellites

Two-year-old satellites

Sneakers

Two-year-old females

One-year-old females

0.5

Weight loss or gain [% per day]

Otolith growth in the first year [mm]

1

551

0.5

0.0

-0.5

Territorial Satellites Sneakers Nonmales reproductive males

Fig. 18.4 Three different life history pathways in males of the Mediterranean wrasse Symphodus ocellatus are revealed by otolith growth patterns. 1) In the first year of life growth varies significantly between age and behaviour types (means ± SD). Sneaker males and 2-year-old satellites had grown less in their first year than 1-year-old satellites, non-reproductive males and nesting males. 2) The three male life-history pathways 1- to 2-year-olds are: sneaker – satellite, satellite – nesting male, and non-reproductive – nesting male. 3) Reproductive males lose on average 0.44% of their body weight per day and ~17% per season due to their behavioural and gonadal effort. Males not participating in reproduction put on weight during the same period, demonstrating a high growth potential during that season. The large and similar reproductive costs of bourgeois and parasitic males can be regarded as an adaptation to sperm competition. (From Taborsky et al. 1987, Taborsky 1998, and Alonzo et al. 2000)

males of this species (Alonzo et al. 2000): one-year-old males may reproduce as sneakers or satellites, or they refrain from reproduction in order to grow (Taborsky 1998; Fig. 18.4); two-year-old males reproduce as satellites (after behaving as sneakers in their first year) or nest males (after behaving as satellites or non-reproductives in their first year). Hence, one life-history pathway includes parasitic and bourgeois behaviour, another one two types of parasitic behaviour and the third one delayed reproduction by the bourgeois tactic. The growth during the first year of life, i.e. be-

552

Michael Taborsky and H. Jane Brockmann

fore their first reproductive season, apparently determines which trajecotory males adopt (Alonzo et al. 2000). This illustrates the potential importance of the date of birth in a seasonal environment (‘birthdate effect’, Taborsky 1998); if growth conditions differ strongly between reproductive and non-reproductive seasons, individuals born late in the season will be exposed to favourable growth conditions before their first winter when compared with early born individuals. In the next season, late-born will be much smaller than early-born individuals, which affects their reproductive opportunities and, hence, tactic choice. Birthdate might be an important trigger for tactic expression also in salmonids (Thorpe 1986). Conditional alternative tactics are not confined to the reproductive context. Other functional conditions in which they have been demonstrated include size-dependent trophic polyphenisms (Frankino and Pfennig 2001), seasonal polyphenism (Shapiro 1976), sex ratio determination (CluttonBrock et al. 1986, Tomkins et al. 2001), predator induced defence polymorphisms (Lively 1986), and dispersal polymorphism (Zera and Rankin 1989, Denno 1994, Denno et al. 1996, Zhao and Zera 2002). In these cases the cues used to switch conditional tactics include environmental cues, such as food availability, food quality, local population density, rainfall, chemical cues from predators, and day length, or they may involve internal cues about the individual’s internal state such as growth rate, size, condition or dominance. Regardless of the proximate cause, these cues are used because they provide individuals with information about the fitness prospects when following different alternative tactics. 18.2.3 Frequency dependence When individuals exhibit alternative reproductive tactics, their success depends on the frequency with which their own tactic is pursued and on the frequency with which alternative tactics are adopted in the population. The evolutionary mechanism responsible for tactic frequencies in a population is negative frequency dependence: the success of a tactic depends on its proportion in the population. In the simplest case, selection will produce an equilibrium distribution of alternative tactics in the population that is stable over evolutionary time. For this situation game theory models assume that evolutionary stability is reached either by equilibrium proportions of pure behavioural strategies (e.g., 70% bourgeois and 30% parasitic individuals), or by equilibrium proportions of behaviour (e.g., on average 70% bourgeois and 30% parasitic behaviour) shown by all individuals of the population (mixed strategies). Both pure and mixed strategies are referred to as ‘evolutionary stable strategies’ (ESS) if the condition of a stable fre-

Alternative reproductive tactics and life history phenotypes

553

quency distribution is reached by negative frequency-dependent selection (Maynard Smith 1982, Parker 1984). Negative frequency dependence affects the distribution of tactics regardless of whether they are fixed for life or dependent on conditions (Brockmann and Taborsky 2008). However, plastic responses to present conditions allow a much quicker response to deviations from equilibrium (i.e., within a generation), whereas with fixed tactics the population is returned to an equilibrium point by natural selection (i.e., between generations). Due to this retarded dynamic, fixed alternative tactics are expected to occur mainly if crucial environmental conditions change rarely in relation to the average lifetime of organisms, or if change is unpredictable and, hence, information about the best tactic is not available or too costly to obtain (Brockmann 2001, Shuster and Wade 2003). These predictions have not yet been scrutinised, however. Three fixed alternative tactics are exhibited by male marine isopods (Paracerceis sculpta: Shuster 1992), and there is evidence that these tactics have a genetic basis and lead to equal reproductive success (Shuster and Wade 1991, Shuster and Sassaman 1997). In livebearing swordtail fish (Xiphophorus nigrensis) two size-dependent male alternative tactics have been described (Zimmerer and Kallman 1989): large males courting females before copulation, and small males reproducing by surreptitious, forced copulations, at least when large competitors are present (Ryan and Causey 1989). Male size classes are genetically based as size and age at maturation are determined by a series of sex-linked alleles, with little growth occurring after maturation (Kallman 1989, Zimmerer and Kallman 1989; maturation may also depend on social status, however; Sohn 1977). The fitness of male swordtails pursuing alternative tactics seems to be balanced due to a higher probability to survive until reproduction for small males and a greater copulation success by large males (Ryan et al. 1992). A somewhat similar balance between a greater reproductive rate of bourgeois males and a longer reproductive lifespan of parasitic males resulting in equivalent lifetime fitness estimates has been observed in the damselfly Mnais costalis (Tsubaki et al. 1997). The lek mating system of ruffs (Philomachus pugnax) is characterised by dark, bourgeois males displaying to attract females and light satellite males staying close by, interfering and copulating with females that approach the court (Hogan-Warburg 1966, van Rhijn 1973). A captive rearing study showed that differential morph development is genetically controlled and consistent with a single-locus, two-allele autosomal genetic polymorphism (Lank et al. 1995). Courting males seem to benefit from the presence of satellites due to some female mating preference for male couples (Hugie and Lank 1997, Widemo 1998). Although it has been assumed that the lower reproductive success of satellite males might be compen-

554

Michael Taborsky and H. Jane Brockmann

sated by a longer reproductive lifespan like in the damselfly described above (Widemo 1998), it is yet unclear whether morph frequencies are stable over time due to negative frequency dependence. In addition, female mimics appear to be a third, very rare morph reproducing parasitically on leks (Jukema and Piersma 2006). Sometimes selection does not seem to produce an equilibrium distribution of alternative tactics in a population, but morph frequencies oscillate over time. In males of side-blotched lizards (Uta stansburiana), for example, there are three genetically determined throat colour morphs (orange, blue and yellow), each of them linked to a different mating tactic. Blue throated males are monogamous and guard their mates, orange males are polygynous, defend large territories and dominate blue neighbours siring some of their females’ offspring, and yellow males adopt a surreptitious parasitic tactic, siring offspring of females defended by orange males by secretive copulations (Zamudio and Sinervo 2000). An ESS-model suggested a frequency-dependent cycle of male morphs with an oscillation period of six years, reflecting a rock-paper-scissors dynamic (Sinervo and Lively 1996). Laboratory breeding experiments (Sinervo et al. 2001) and gene mapping (Sinervo et al. 2006) have suggested a single genetic factor of major effect controlling this male polymorphism. Manipulations of tactic frequencies in six populations revealed that the respective common phenotypes lost fitness in comparison to their antagonists (Bleay et al. 2007), which is a crucial prediction of the balancing function of negative frequency dependence. Oscillations of morph frequencies around an equilibrium point by negative frequency dependence are probably widespread also in systems with two alternative morphs, as exemplified in the feeding polymorphism of scale-eating Lake Tanganyika fishes with asymmetrical mouths (Perissodus microlepis: Hori 1993, Takahashi and Hori 1994). Negative frequency dependence will balance morph frequencies whenever fitness lines of alternative tactics cross, with selection pushing frequency deviations back to the point of intersection (Gadgil 1972, Waltz 1982, Brockmann and Taborsky 2008). This will lead to stable equilibrium conditions or oscillations of various phase length and amplitude around the intersection point. Negative frequency dependence is a straightforward and captivating mechanism that has been demonstrated to work by observational and manipulative studies (e.g., Hori 1993, Bleay et al. 2007), but experimental proof that alternative tactics are balanced by negative frequency-dependent selection is still lacking. Dimorphic or polymorphic trait distributions in natural populations do not necessarily imply simple genetic or even Mendelian causation and equilibrium morph frequencies (Pienaar and Greeff 2003). Fitness lines do not always cross, for instance when conditional tactics are employed by individuals diverging in quality due to

Alternative reproductive tactics and life history phenotypes

555

variation in developmental constraints (Schlichting and Pigliucci 1998), which may result in morph frequencies that are not balanced by negative frequency-dependent selection (Taborsky et al. 2008). 18.2.4 Density dependence When same or different phenotypes are competing for the same resources or mates, then the economic defendability of the resource or mate depends on the density of competitors in a population (Emlen and Oring 1977). This affects the occurrence, coexistence and frequency of alternative reproductive tactics. At high densities the bourgeois (i.e. investing) tactic may become much less rewarding due to more intense resource competition. Dragonfly (Nannophya pygmaea) males, for example, are more likely to switch from a territorial to a satellite tactic at high population densities (Tsubaki and Ono 1986). Male field crickets (Gryllus integer) switch from calling to searching at high densities (Cade and Cade 1992), whereas male white-footed mice (Peromyscus leucopus) change from territoriality to wandering at low female densities (Wolff and Cicirello 1990). Similarly, in fallow deer (Dama dama), male reproductive tactics vary mainly with male and female densities (Langbein and Thirgood 1989). Density dependence often interacts with frequency dependence in the evolution of alternative phenotypes (Bleay et al. 2007), especially as frequency dependence may be much stronger at high than at low densities (Eadie and Fryxell 1992, Lucas and Howard 1995, 2008). In addition to population density, operational sex ratio may also influence tactic choice (Forchhammer and Boomsma 1998, Zamudio and Chan 2008) and the response to frequencydependent selection (Andrés et al. 2002). The interaction between density dependence and frequency dependence may be influenced by other ecological factors such as predation or parasitism risk (Walker and Cade 2003) and female choice (Rios-Cardenas et al. 2007). If alternative phenotypes are using different resources or habitats (Halama and Reznick 2001), density within phenotypes but not frequency dependence between them will affect their occurrence. For example when animals dwell in a patchwork of interspersed niches (or temporally varying habitats), selection may favour multiple phenotypes that are specialised for exploiting resources in each niche (Brockmann and Taborsky 2008). Intermediate phenotypes will not be as effective as specialists in exploiting the resource or at finding mates, resulting in disruptive selection for morphs specialised for each habitat. In this situation, the success of one phenotype does not depend on the other, as each morph simply exploits the habitat to which it is adapted. The two morphs should occur at frequencies

556

Michael Taborsky and H. Jane Brockmann

that match the resource availability in the different niches (Seger and Brockmann 1987), i.e., the success of alternative phenotypes is density dependent (Denno et al. 1985) but not frequency dependent. Similarly, alternative reproductive phenotypes in one sex may be favoured by a polymorphism in preference traits of the other sex (Henson and Warner 1997, Alonzo and Sinervo 2001, Morris et al. 2003), which would mean that members of a phenotype compete with one another but not with members of the other phenotype and would be subject to density but not frequency dependence. It should be considered also that density effects may cause frequency oscillations between alternative phenotypes in ways that are similar to negative frequency dependence, although it may need additional frequency dependence effects to reach equilibrium conditions (Sinervo et al. 2000). Population density and resource abundance may influence morph frequencies of males also through maternal allocation decisions, as occurs for example in Dawson’s burrowing bees (Amegilla dawsoni), where females invest in either large or small males that then pursue alternative mating tactics based on their size (Alcock 1999, Tomkins et al. 2001). In this situation, an understanding of the factors responsible for male morph frequencies requires information about the factors influencing maternal allocation tactics and not just the relative success of male tactics (Brockmann 2008). Density effects may vary between populations with different selection histories, as demonstrated by responses to experimental variation of population size on distinct male morphs in acarid mites (Tomkins et al. 2004). In Sancassania berlesei the decision of males to turn into fighters or non-fighters depends on social conditions and food during development (Radwan 1995, Radwan et al. 2002). Small and low density populations contain larger proportions of the fighting male morph. In a captive breeding experiment, most males originating from three natural populations could be manipulated to become fighters when kept alone, but a pronounced response to increasing density by emergence of a much smaller number of fighters was only shown by two of the three populations (Tomkins et al. 2004). When disruptive selection occurs and different traits are favoured under different conditions, selection favours the evolution of a threshold switching mechanism between alternatives (Nijhout 2003).

Alternative reproductive tactics and life history phenotypes

557

18.3 Proximate mechanisms 18.3.1 Genetic and environmental components As with any other biological trait, ARTs are a product of genes and the environment and of interactions between these causal factors; it would be naïve to assume that ARTs are either ‘genetically’ or ‘environmentally’ determined (Caro and Bateson 1986). Dimorphic traits usually seem to be threshold traits (Roff 1996) influenced by quantitative trait loci: morph expression depends on whether a ‘liability’ value is above or below a threshold (Falconer and Mackay 1996). This has been demonstrated for the expression of different male morphs in mites, in which the threshold reaction norm was shifted in artificial selection experiments (Unrug et al. 2004). Threshold traits frequently operate during development, and alternative pathways may change abruptly, e.g. at a particular body size, producing different phenotypes on either side of the threshold (Emlen and Nijhout 2000, Nijhout 2003, Lee 2005). As developmental thresholds have a genetic basis, trait expression is both conditional and heritable, allowing alternative phenotypes to evolve largely independently from one another. This greatly increases the scope for the evolution of alternative tactics (West-Eberhard 1989, 2003). Since thresholds or developmental switch points involved in the determination of tactics have a genetic basis they will be subject to selection and adaptive evolution (Tomkins et al. 2004). It is important to bear in mind that even if divergent conditional responses have different genetic bases and result in different fitness outcomes for alternative tactics they may still persist in a population (Hazel et al. 1990; see Sect. 18.2.2). 18.3.1.1 Plastic tactics When alternative phenotype expression remains flexible throughout life, the decision to choose one or the other option depends on cues specific to the current situation. Stickleback males, for example, display bright nuptial colours, defend territories, build nests and attract females, but if their nest is already full of eggs and a female spawns nearby in a neighbour’s nest, this male turns drab mimicking a female-like colour pattern and sneaks into the neighbor’s nest to steal fertilisations (van den Assem 1967, Rowland 1979). Stickleback males in general seem to switch between tactics instantaneously, adjusting to the respective situation in response to a number of factors that include the state of their own brood, the distance to neighbours and their courtship success, among others. Similar opportunistic reactions to present conditions are shown, for example, in male frogs

558

Michael Taborsky and H. Jane Brockmann

and toads that switch from calling to satellite behaviour in response to their neighbourhood on the pond (Perrill et al. 1982, Arak 1988; reviewed in Zamudio and Chan 2008), and in birds switching between bourgeois and parasitic mating tactics in response to their social environment (Birkhead and Møller 1992, Westneat and Stewart 2003). When tactic expression is sequential, such as in numerous fish species where males exhibit parasitic tactics when small, but turn into bourgeois males when large (Taborsky 1994, 2008), the decision to behave in one or the other way will again depend on momentary conditions, which includes the individual’s relative resource holding potential as a function of body size. This ontogenetic switch has a strong causal relationship to indeterminate growth as exhibited by most fishes (Taborsky 1999), which generates age-related size variation among reproductive competitors. Similar patterns have been observed for example in frogs (Howard 1984, Forester and Lykens 1986) and marine iguanas (Wikelski et al. 1996). A similar mechanism of size-dependent tactic choice can be observed in species with ontogenetic sex change. Here, the switch is not from one to another reproductive tactic within a sex, but from one sex to another (Warner et al. 1975, Shapiro 1987, Ross 1990). The direction of switch depends on environmental conditions determining which sex benefits more from large size or dominant status (Warner et al. 1975, Muñoz and Warner 2004). In contrast, in sequential intrasexual choice of alternative reproductive tactics the direction of change is always from parasitic to bourgeois, because the latter tactic inevitably benefits more from large size. A sequential choice of tactics is not confined to species with indeterminate growth, however, even though size and age dependence may still be important in species halting growth at maturity (livebearing fishes: Constantz 1975, Farr 1980; mammals: Le Boeuf 1974, Wirtz 1982, Gosling and Petrie 1990, Pemberton et al. 2004). In horseshoe crabs for example age-related condition and not size affects tactic switches (Brockmann and Penn 1992, Brockmann 2003). The important point is that selection favours mechanisms (such as status dependence of sex change; Rodgers et al. 2007) that switch the individual from one tactic to another at the point that maximises lifetime fitness. 18.3.1.2 Fixed tactics: genotype polymorphism When the expression of ARTs is fixed for life, alternative tactics may be determined either by polymorphic genotypes originating from major gene or polygenic effects, or by some developmental mechanism. Genetic polymorphisms responsible for the expression of fixed alternative reproductive tactics have been described for the marine isopod Paracerceis sculpta, where male tactics are apparently controlled mainly by the Men-

Alternative reproductive tactics and life history phenotypes

559

delian segregation of three alleles at one autosomal locus of major effect (Shuster and Wade 1991, Shuster and Sassaman 1997). The involved alleles show directional dominance and interact with alleles at other loci, setting off developmental cascades that determine the three alternative morphological and behavioural phenotypes. Interactions between alleles of this and another autosomal locus, primary sex-determining factors and possibly unknown extrachromosomal factors, affect sex determination in this species, which can distort sex ratios within groups and influence male tactic frequencies (Shuster et al. 2001). In plumage polymorphic male ruffs that show either a bourgeois courtship tactic or parasitic satellite behaviour, offspring phenotype is strongly influenced by paternal inheritance. Common garden breeding experiments provided evidence for a single locus, two-allele autosomal genetic polymorphism, with the best fit to the pedigree data obtained with a satellite allele dominance model (Lank et al. 1995). Variance in environmental factors was minimised in this experiment, so a potential modifying effect by environmental or social conditions may occur. Similarly, males of the damselfly Mnais costalis differing in wing colouration pursue either a bourgeois tactic by defending potential oviposition sites or a parasitic tactic where females are opportunistically pursued, which is most successful when accomplished on bourgeois male territories (Tsubaki et al. 1997). A captive rearing experiment suggested that this male polymorphism is genetically controlled by a single-locus, two-allele autosomal polymorphism (Tsubaki 2003). In males of the swordtail Xiphophorus nigrensis that show determinate growth, a simple Mendelian polymorphism linked to the male sex chromosome is largely responsible for different body sizes (Kallman 1984, 1989). A locus (‘pituitary locus’) with three alleles triggers the timing of activation of the hypothalamic-pituitary-gonadal axis during ontogeny and hence the secretion of androgens, resulting in three male size classes, and body size heritabilities > 90% (Ryan and Wagner 1987, Kallman 1989). The males of the smallest size class perform primarily forced copulations, which is different from the courtship based mating tactic preferably used by large males (Ryan and Causey 1989). Intermediate males perform either one or the other tactic, depending on their size. Tactic choice was shown to be genetically determined also independently of body size effects (Zimmerer and Kallman 1989; this study apparently used the sister species, X. multilineatus; cf. Ryan et al. 1992). Hence, in this system genotype affects the choice of mating tactic through body size and through size-independent effects. It should be noted, though, that there is overlap both in body size and mating behaviour between the three male types (Ryan and Causey 1989, Zimmerer and Kallman 1989), suggesting more complex genetic and environmental causation of male mating behaviour than is hitherto under-

560

Michael Taborsky and H. Jane Brockmann

stood. In another male-polymorphic poeciliid fish, Limia perugiae, also three male size morphs have been described, which coincides with similar specialisation between courting and forced copulation tactics as observed in Xiphophorus males. Here, breeding experiments suggested that male morphs are determined by Y-chromosome linked genes for small and large size interacting with an autosomal recessive repressor responsible for the development of intermediate males (Erbelding-Denk et al. 1994). The dominant allele of the recessive repressor might activate the Y-genes for large or small size, respectively, resulting in males that only attempt forced copulations (small males) or both, forced matings and copulations induced by courtship (intermediate and large males). 18.3.1.3 Fixed tactics: developmental threshold mechanisms Tactics that are fixed for the adult life of organisms are often generated by developmental mechanisms where the decision on which life-history trajectory to follow depends on environmental or social conditions. For example in a number of species including salmon (Thorpe and Morgan 1980, Thorpe et al. 1998, Hutchings and Jones 1998), the rate of growth during the first year results in a permanent condition-dependent developmental switch to the parasitic tactic. The switch may also depend on passing a threshold size by a particular life stage, such as the end of the larval period (Emlen 1994, Moczek and Nijhout 2002, Thériault and Dodson 2003). Such threshold-dependent switch mechanisms are probably a much more widespread phenomenon than the rather rigidly determined polymorphisms just described (Brockmann 2001; reviewed in Emlen 2008). In seasonal environments, condition-dependent switches may be influenced by the time of birth of an individual in relation to environmental oscillations (the ‘birthdate effect’, Taborsky 1998; see above). In temperate marine fishes, for example, the time available to grow before the first winter may determine the choice of reproductive tactic in the subsequent reproductive season, because late-born males may be too small to compete with early born or older rivals on equal terms (Alonzo et al. 2000, Oliveira et al. 2001, 2002). Gene-environment interactions are of prime importance for the expression of alternative tactics, as illustrated by reproductive patterns of male salmon (Heath et al. 1994, Aubin-Horth et al. 2005a). Size thresholds responsible for tactic expression may differ between individuals within or between populations so that accelerated early development does not influence the choice of tactic uniformly across all genotypes, and the thresholds can be affected by environmental conditions and are subject to selection (Heath et al. 1994, Aubin-Horth and Dodson 2004, Baum et al. 2004, Piché et al. 2008). Potential cues other than size or growth rate during de-

Alternative reproductive tactics and life history phenotypes

561

velopment that may affect phenotype expression include temperature, humidity, photoperiod, light conditions, food components and population density; in effect any external or internal cue that provides information to the individual about potential fitness effects of choosing one or another phenotype expression or life history trajectory (Dunbar 1982; reviewed in Levins 1968, Moran 1992, West-Eberhard 2003, Emlen 2008). 18.3.2 Physiological regulation Neural circuits may strongly differ between individuals showing alternative behaviours, or biochemical switching of existing circuits may be caused by neuromodulators. These neural mechanisms interact closely with hormonal mechanisms, as neurons synthesise peptides regulating hormone production and, in turn, structural organisation of neural circuits can be influenced by organisational effects of hormones. Biochemical switches, in contrast, can be driven by activational effects of hormones on regulatory pathways of behaviour. Structural reorganisation and biochemical switching are the major mechanism underlying behavioural plasticity (Zupanc and Lamprecht 2000). It has been argued that these regulatory processes might differ systematically between fixed and flexible alternative tactics. Flexible tactics requiring rapid and transient changes in neural activity are mediated by biochemical switches involving activational effects of hormones, whereas fixed and sequential tactics are mediated by structural (re)organisation of neural networks (Oliveira et al. 2008b). Threshold responses are usually associated with hormonal differences (e.g., differences in rates of synthesis or degradation of hormones, timing of secretion or receptor expression, changes in binding affinities or in the number of receptors expressed; Emlen 2008) and they have been found to be associated with changes in gene expression in the brain that affect behaviour (Whitfield et al. 2003). In vertebrates, the forebrain’s preoptic area (POA) together with the anterior hypothalamus is a region of the brain that is highly conserved across classes, synthesising a wide range of neuropeptides and concentrating sexual steroids. This region has important control functions of reproductive and social behaviour patterns by its connections to the somatic and visceral motor systems and the pituitary gland. Peptides synthesised by POA neurons (most vertebrates) or axons reaching the pituitary directly (teleosts) trigger the activity of secretory cells in the anterior pituitary that release peptidergic hormones targeting organs throughout the body via the circulatory system. In addition, POA neurons synthesise and release arginine vasotocin- (AVT; in teleosts) or vasopressin- (AVP; in mammals)-like pep-

562

Michael Taborsky and H. Jane Brockmann

tides into the posterior pituitary, which again interfaces with the circulatory system throughout the body. Among other functions, AVT/AVP has facilitating effects on courtship behaviour. The POA is also partly responsible for the production of gonadotropin- (GnRH; in teleosts) or luteinising- (LHRH; in mammals)-releasing hormones, that via the anterior pituitary influence gonad development and the production of sexual steroids. In species with ARTs, the number of neuropeptide-containing neurons within the POA vary systematically with developmental trajectories and alternative tactics (Bass and Grober 2001, Rhen and Crews 2002, Bass and Forlano 2008). In male teleosts, for example, GnRH dimorphisms are associated with differences in relative gonad size and reproductive tactics, with the morph with a larger gonad mass/body mass ratio showing larger or more GnRH-releasing POA neurons (Bass and Forlano 2008; see Bass and Grober 2009 for review of neuroendocrine mechanisms of reproductive plasticity in fish). When aiming to understand the mediating effect of hormones on the expression of alternative reproductive tactics, it is important to distinguish between their activational effects, which are transient and affect individuals throughout life, and organisational effects, which are long-lasting and typically act during a critical period of development early in ontogeny (Arnold and Breedlove 1985). Moore (1991, Moore et al. 1998) proposed that fixed reproductive tactics should be mainly affected by the organisational role of hormones, whereas flexible phenotypes would rather depend on their activational function (the ‘relative plasticity hypothesis’). The two predictions derived from this hypothesis received only limited support, however: (1) In species with ARTs remaining flexible throughout life, the levels of hormones affecting phenotype expression should differ between morphs, whereas in species with fixed ARTs hormone profiles should be similar between alternative morphs. Regarding androgens, this prediction is consistent with roughly 80% of studied cases with ARTs in mammals, but with less then 50% of cases in birds and only 60 to 70% of cases in other vertebrate classes (see Oliveira et al. 2008b for review). The comparative patterns suggests that this hypothesis might reflect hormonal mechanisms of phenotype expression adequately only in connection with genetic sex determination and male heterogamety, where the expression of secondary sexual characters is androgen dependent (i.e., the condition in eutherian mammals). (2) Hormone manipulations during adulthood should hence affect only the expression of flexible ARTs (activational effect), whereas in species with fixed ARTs such manipulations should be effective only when applied during early development (organisational effect). Twelve experimental studies have again produced equivocal results (reviewed in Oliveira et al. 2008b). In tree lizards (Urosaurus ornatus), the

Alternative reproductive tactics and life history phenotypes

563

only species with fixed ARTs tested to date by a manipulation during early ontogeny, castration or hormonal implants in males significantly affected morph expression when applied at hatching (Hews et al. 1994). In contrast, six of eleven studies where hormone levels were manipulated during adulthood did not confirm the second prediction of the relative plasticity hypothesis. In marine iguanas, Wikelski and collaborators (2005) tested whether bourgeois males may be transformed into the parasitic phenotype when implanting them with an androgen blocker in combination with an aromatase inhibitor to obstruct the effects of testosterone on bourgeois male behaviour. Even though the treated males reduced their territorial behaviour and attracted fewer females they did not switch to parasitic behaviour. Marine iguanas show sequential ARTs with males developing from parasitic towards bourgeois behaviour with increasing age and size, and the described experiments provide no clue that this switch is solely triggered by the action of testosterone. This and other experimental and correlative evidence suggests that the effects of hormones on the expression of ARTs is not as simple as originally assumed. A ‘one-hormone-one-tactic’ relationship seems unrealistic; rather, different hormones modulate neural pathways underlying behaviour patterns in concert, involving interaction effects and feedback from the social environment (Oliveira 2004). A twoway type of interaction between hormones and behaviour is consistent also with the observation that regarding androgens, for example, their influence on morphological differentiation is much greater than on behavioural traits (this applies to intrasexually polymorphic species; Oliveira et al. 2008b). New insight into the regulatory function of factors responsible for the expression of male tactics comes from a study of gene expression in alternative male types of Atlantic salmon using microarray technology (AubinHorth et al. 2005b). This study revealed differential expression of 15% of the 2917 genes tested between reproductive parasites and bourgeois (anadromous) males when tested at the same age. Most of the upregulated genes in sneakers are associated with reproduction (e.g., the production of gonadotropins, growth hormone and prolactin), neural plasticity and neural signalling, whereas the upregulated genes of anadromous males are mainly associated with somatic growth and maturation. Interestingly, these results suggest a greater role for neural plasticity in the parasitic males, which was unexpected. Understanding the threshold mechanisms involved in the development and expression of ARTs is particularly fruitful also for a comprehension of underlying evolutionary patterns. In horned beetles with guarding and sneaking males, experiments revealed that the juvenile hormone provides the important signal. This hormone varies with larval diet, and male larvae

564

Michael Taborsky and H. Jane Brockmann

with levels above a threshold (which are the small animals) generate a pulse of ecdysone that acts as a secondary signal (Emlen and Nijhout 1999, 2001). This signal prevents significant horn growth in females and small males, perhaps by affecting the sensitivity of horn cells to juvenile hormone during a sensitive period. The physiological mechanisms regulating the divergent behaviour of the resulting male morphs, however, are hitherto less well understood. Importantly, genetic changes in downstream regulatory pathways involving hormones and neural mechanisms may cause changes of tactics, in morphology, physiology and behaviour alike, that are expressed more or less independently from the alternative tactics (Emlen 2008). This insight illustrates basic mechanisms how genes can create divergent phenotypes among individuals of the same species and sex relatively independently from each other (West-Eberhard 2003).

18.4 General questions and future prospects Which patterns emerge when comparing the occurrence of ARTs in different major taxa (BOX 18.4) and which general principles can be uncovered by the study of alternative phenotypes? Research on alternative tactics can help to understand a range of mechanisms and principles that are of general importance in biology. This includes the role of trade-offs for the evolution of allocation decisions, the costs and benefits of flexibility, the origin of complex suites of characters, the action of disruptive selection, the evolution of decision rules, the functionality of threshold mechanisms and developmental switches, the causes and consequences of individual (behavioural, physiological, morphological) variation among conspecifics, the evolution of cooperation as a means to cope with competition, specialisation and adaptations to sperm competition (at gametic, anatomical and behavioural levels), and the coexistence of several evolutionarily stable strategies within a population. ARTs are an important component of biological diversity and may be a source of sympatric speciation (GarciaVazquez et al. 2002, West-Eberhard 2003). All reproductive traits are variable but ARTs are different because the variation is significantly discrete, i.e. bimodal or multimodal, rather than continuous. Most ARTs are characterised by suites of co-varying and coadapted traits tightly associated with each alternative phenotype. This means that to understand the evolution and expression of ARTs we need to understand how discrete variation comes about and why it is favoured by selection. We have argued that discrete variation usually arises through threshold mechanisms, i.e. the animal switches from one reproductive

Alternative reproductive tactics and life history phenotypes

565

phenotype to another when specific conditions are reached (Emlen 2008). These threshold mechanisms have been well studied both physiologically and developmentally (Emlen 1996, Roff 1998, Nijhout 1999, Hartfelder and Emlen 2005). There is a growing literature demonstrating the heritability of thresholds and an ability to respond quickly to selection (Emlen 1996, Zera and Zhang 1995, Shuster and Wade 2003). Threshold mechanisms are favoured by selection when individuals possessing intermediate values of a continuously varying trait are selected against, i.e. under disruptive selection (Danforth and Desjardin 1999). Some alternative phenotypes are fixed during development and some are flexible over the life of the animal. Selection favours flexible alternative tactics when information exists (e.g. from the individual, environment) on the relationship between phenotype and fitness (Brockmann 2001, Shuster 2008). Selection against intermediates occurs because extreme phenotypes have higher fitness (Brockmann 2008, Brockmann and Taborsky 2008). This may be because of trade-offs (physical, behavioural, physiological, developmental or genetic; see Zera and Harshman 2001) or genetic correlations, or because of different reproductive niches for different phenotypes within a population. Selection for extreme phenotypes is particularly common under intense sexual selection when animals pay high costs for specialised structures that enhance performance in male-male competition or female choice. Any male that can gain even a few fertilisations without paying those extreme costs may have higher than average fitness, since under conditions of high variance in fitness most individuals produce no offspring (Shuster and Wade 2003). While this broad outline for the evolution of ARTs seems clear, we have little ability to predict the patterns, frequency and occurrence of alternative phenotypes. Evolutionary outcomes are very difficult to understand without the aid of models (Brockmann 2001, Wolf et al. 2007). To predict complex interactions between an individual’s physiological condition, prevailing environmental conditions, time constraints, and frequency- and density-dependent factors operating within a population demands a modelling approach (Lucas and Howard 2008). Models allow us to examine assumptions and to specify the factors that we think are influencing the evolution of tactics in a particular system. Dynamic models (Lucas et al. 1996, Skubic et al. 2004) or multi-level dynamic games (Alonzo and Warner 2000) and other kinds of models (e.g. Sinervo et al. 2000, Alonzo and Sinervo 2001) provide subtle and sometimes counterintuitive outcomes that result from complex and unexpected interactions among variables and parameters. For example, in a model of flexible caller/satellite tactics in chorusing frogs (Lucas et al. 1996), the effects of costly singing, predation risk at choruses, variation in female arrival rates, frequency-dependent effects

566

Michael Taborsky and H. Jane Brockmann

BOX 18.4 Emerging patterns: why do ARTs differ between major taxa? When comparing ARTs among major taxa, their prevalence and forms appear to differ substantially. Male ARTs are widespread and variable in fish, for example, which may relate to four conditions (Taborsky 2008): (1) The fertilisation mechanism: the vast majority of fish taxa show external fertilisation, which has two important consequences. Firstly, it is difficult for males to monopolise access to partners or fertilisable eggs. Secondly, external fertilisation selects for large numbers of sperm, which is a precondition for male success in sperm competition. External fertilisation also allows for very diverse spawning patterns (e.g., in the water column, on the ground or on/in a substrate), which in turn increases the variability of ARTs. (2) Indeterminate growth: the vast majority of fishes do not stop to grow after maturation, which often causes enormous intrasexual size differences within species (Taborsky 1999). In a sample of eight species of fish with ARTs from which data of male size were available, comprising salmon, sunfish, cichlids, wrasses and blennies, the largest males of a species were on average 18 times heavier than their smallest male conspecifics (Taborsky 1999). In contrast, the intrasexual size dimorphism of animals with determinate growth is much smaller. In 490 passeriform bird species, for example, the largest males were on average only 1.19 times heavier than their smallest male conspecifics (Taborsky 1999). As a consequence, no specialised reproductive parasites are apparently found in passerine birds. (3) Parental roles: paternal investment is particularly widespread in fish (Blumer 1979), which allows male competitors to exploit such investment by the performance of ARTs. In addition, fish show an unprecedented diversity of parental care patterns (Blumer 1982), which relates to a great variability in mating patterns and, consequently ARTs (Taborsky 1994). (4) Variable sex determination: in fishes we find gonochorism, simultaneous and sequential hermaphroditism, and the latter with either males (protandry) or females (protogyny) preceding each other (Shapiro 1987). Gonochorism and sex change may even coexist within species (Robertson et al. 1982). This variability of sex determination and differentiation mechanism might also relate to the evolution of alternative mating tactics in fish, as suggested by the frequent existence of ARTs in simultaneous and sequential hermaphrodites (see Taborsky 2008 for review). In insects, ARTs can be found at all different steps of (male) reproductive behaviour, which include locating a mate, getting access and copulating with her, and showing postcopulatory behaviour serving to raise the fertilisation chances (Brockmann 2008). ARTs are particularly common at the mate searching stage, especially when females mate in diverse locations. In species with maternal care, investing females are often exploited by other females by facultative, intraspecific brood parasitism (Tallamy 2005). The prevalence varies greatly between insect orders, with Hymenoptera and Coleoptera spearheading others, whereas in the also well-studied Hemiptera and

Alternative reproductive tactics and life history phenotypes

567

Lepidoptera ARTs are apparently rare. This variation may again relate to parental care, which is more common in the first two orders (Trumbo 1996), providing opportunities for exploitation by male and female parasitism. In amphibians, ARTs are mainly confined to mate acquisition in males (Zamudio and Chan 2008), for example with bourgeois caller and parasitic satellite tactics in frogs and toads. In salamanders, male ARTs involve paedomorphosis (the retention of larval characters in mature adults), female mimicry and sexual interference. Female amphibians hardly show ARTs. In contrast, intraspecific female reproductive parasitism (egg dumping) is widespread in birds, where high parental investment makes exploitation worthwhile (Krüger 2008). In male birds, opportunistic simultaneous ARTs are very widespread, with males that invest in resource defence, mate acquisition and parental care also pursuing a parasitic tactic to fertilise neighbouring females (Westneat and Stewart 2003). In mammals, harem defence and resource defence polygyny in pinnipeds and ungulates is frequently accompanied by alternative male tactics exploiting the investment of bourgeois males (Wolff 2008). A pattern with territorial bourgeois males and extra-pair fertilisations by satellite, neighbouring or wandering males is also widespread in rodents. In contrast, sexual coercion as an alternative male tactic is most widespread in primates (Smuts and Smuts 1993), where sneak matings by subordinate males with females guarded by dominants also abound (Setchell 2008). Female ARTs are very rare in mammals. Probably the most obvious general pattern is that ARTs abound where individuals invest heavily in reproduction in a way that can be circumvented and exploited by same-sex competitors. As soon as selection favours investment, there is also potential to select for low-cost tactics to attain the same goal by alternative means (Taborsky 1997, 2001). This reflects disruptive selection on reproductive investment (e.g., Schütz et al. 2010). A second generality may be the important effect of body size variation among samesex competitors, which may have different causes in the evolutionary ecology of different species, including indeterminate growth, variable growth conditions during ontogeny, and diverging feeding specialisations of juveniles. In species with environmentally determined size variation between competitors, a specialisation should result in bourgeois tactics if large and parasitic tactics if small. This expectation is apparently often met. When comparing gross patterns between taxa the caveat is, however, that our knowledge depends largely on the accessibility of information, conspicuousness of traits, research approach, ideas and hypotheses of researchers, and underlying research traditions. Therefore, the differences we conceive between major taxa must be treated with caution – in the end they might be more apparent than real.

568

Michael Taborsky and H. Jane Brockmann

since satellites parasitise the singing behaviour of callers (only callers attract females) and a limited breeding season combined to produce some unexpected results: periods of intense calling followed by periods with no animals present. Such fluctuations in numbers at choruses had been observed but were always assumed to be due to some unknown environmental factors. This model also demonstrated that a variable that had never been measured, female arrival rate, was a crucial parameter in determining the behaviour of the model and the frequency of tactics. Thus, models can identify new important factors; they can evaluate the importance of variables through sensitivity analyses; and they can even suggest new tactics that were not previously suspected. Even some of the best-studied ART systems at the empirical level have never been modelled and our understanding of the factors that shape their evolution has not been tested. We do not know, for example, whether the effects of frequency and density-dependence are sufficient to maintain alternative tactics at observed frequencies in most systems or even what factors might be creating frequency-dependent effects (e.g. female choice traits, interactions among males, heterozygote advantage). To create a model of an ART system, the factors affecting fitness must be hypothesised, formalised and their functional relationships with other variables specified. Even simple models would take us a long way toward incrementing our understanding of ARTs. In addition, few alternative phenotype systems have been critically tested at the empirical level. Experimental testing requires manipulating the frequencies of tactics to see whether fitness changes in the predicted directions (Bleay et al. 2007). Alternatively, increasing or decreasing the fitness of one of the tactics should show whether the frequencies of tactics change in the predicted directions or whether selection differentials are affected correspondingly (Lande and Arnold 1983). While such experiments are not easy to carry out, they will reveal whether our picture of the evolution and maintenance of alternative tactics is realistic. ARTs, along with other sorts of alternative tactics, provide one important source of variation upon which evolution acts (West-Eberhard 2003). A well studied discontinuous distribution of reproductive traits is male and female function. Animals may either pursue a fixed tactic by adopting one of the two sexes for a lifetime (gonochorism); develop both sexual functions simultaneously and use them interchangeably depending on circumstances (simultaneous hermaphroditism); or switch from one sexual function to the other either once or repeatedly (sequential hermaphroditism with fixed or reversible sequence) depending on age or condition (Anthes this volume). Hence the same scheme that we used to classify ARTs also applies to sex allocation (Fig. 18.1). There is a large literature on sex allo-

Alternative reproductive tactics and life history phenotypes

569

cation (Charnov 1982), both at the ultimate and proximate levels, and incorporating this aspect of ARTs here was beyond the scope of this chapter. However, it is important to consider that the mechanisms underlying systematic variation in reproductive tactics between sexes and within a sex are similar (Brockmann 2001). Studies of populations with different frequencies of alternative phenotypes can provide insight into the factors that influence phenotypic evolution (Bailey et al. 2008, Pizzo et al. 2008). Alternative phenotypes also provide a unique opportunity to study the underlying genetic architecture of adaptations. For example, the parasitic tactic is often a mosaic of both male and female phenotypes (e.g. sexual mimicry), which evolves through the decoupling of traits that usually occur together (Brockmann et al. 2008). ARTs allow an opportunity to study the nature of correlated traits and the mechanisms by which they may be decoupled (Gonçalves et al. 2008). Further, genes that are differently expressed in alternative phenotypes make it possible to identify the genes and regulatory mechanisms that are consistently up- or down-regulated between alternative phenotypes (Aubin-Horth et al. 2005b, Toth et al. 2007, Gonçalves et al. 2008, Renn et al. 2008, St-Cyr and Aubin-Horth 2009). The study of alternative phenotypes provides a special opportunity for studying the evolution of the mechanisms that underlie adaptation.

Acknowledgements We thank Peter Kappeler for the invitation to contribute to this book and Matthew Grober, Peter Kappeler and an anonymous referee for constructive comments on an earlier version of this manuscript. Michael Taborsky acknowledges support of the Swiss National Science Foundation.

References Ahlund M, Andersson M (2001) Female ducks can double their reproduction. Nature 414:600-601 Alcock J (1999) The nesting behavior of Dawson’s burrowing bee, Amegilla dawsoni (Hymenoptera: Anthophorini), and the production of offspring of different sizes. J Insect Behav 12:363-384 Alcock J, Jones CE, Buchmann SL (1977) Male mating strategies in the bee Centris pallida Fox (Anthophoridae: Hymenoptera). Am Nat 111:145-155

570

Michael Taborsky and H. Jane Brockmann

Alonzo SH (2008) Conflict between the sexes and alternative reproductive tactics within a sex. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 435-450 Alonzo SH, Sinervo B (2001) Mate choice games, context-dependent good genes, and genetic cycles in the side-blotched lizard, Uta stansburiana. Behav Ecol Sociobiol 49:176-186 Alonzo SH, Warner RR (2000) Female choice, conflict between the sexes and the evolution of male alternative reproductive behaviours. Evol Ecol Res 2:149170 Alonzo SH, Taborsky M, Wirtz P (2000) Male alternative reproductive behaviors in a Mediterranean wrasse, Symphodus ocellatus: evidence from otoliths for multiple life-history pathways. Evol Ecol Res 2:997-1007 Andersson M (1994) Sexual Selection. Princeton University Press, Princeton Andrés JA, Sánchez-Guillén RA, Rivera AC (2002) Evolution of female colour polymorphism in damselflies: testing the hypotheses. Anim Behav 63:677685 Arak A (1988) Callers and satellites in the natterjack toad: evolutionarily stable decision rules. Anim Behav 36:416-432 Arnold AP, Breedlove SM (1985) Organizational and activational effects of sex steroids on brain and behavior: a reanalysis. Horm Behav 19:469-498 Aubin-Horth N, Dodson JJ (2004) Influence of individual body size and variable thresholds on the incidence of a sneaker male reproductive tactic in Atlantic salmon. Evolution 58:136-144 Aubin-Horth N, Letcher BH, Hofmann HA (2005a) Interaction of rearing environment and reproductive tactic on gene expression profiles in Atlantic salmon. J Hered 96:261-278 Aubin-Horth N, Landry CR, Letcher BH, Hofmann HA (2005b) Alternative life histories shape brain gene expression profiles in males of the same population. Proc R Soc Lond B 272:1655-1662 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. Behav Ecol Sociobiol 58:506-516 Bailey NW, McNabb JR, Zuk M (2008) Preexisting behavior facilitated the loss of a sexual signal in the field cricket Teleogryllus oceanicus. Behav Ecol 19:202207 Balshine-Earn S, Neat FC, 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. Behav Ecol 9:432-438 Bass AH, Forlano PM (2008) Neuroendocrine mechanisms of alternative reproductive tactics: the chemical language of reproductive and social plasticity. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 109-131

Alternative reproductive tactics and life history phenotypes

571

Bass AH, Grober MS (2001) Social and neural modulation of sexual plasticity in teleost fish. Brain Behav Evol 57:293-300 Bass AH, Grober MS (2009) Reproductive plasticity in fish: evolutionary lability in the patterning of neuroendocrine and behavioral traits underlying divergent sexual phenotypes. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, Rubin RT (eds) Hormones, Brain and Behavior. Vol 1. Academic Press, San Diego, pp 579-609 Baum D, Laughton R, Armstrong JD, Metcalfe NB (2004) Altitudinal variation in the relationship between growth and maturation rate in salmon parr. J Anim Ecol 73:253-260 Bergmüller R, Taborsky M (2005) Experimental manipulation of helping in a cooperative breeder: helpers ‘pay to stay’ by pre-emptive appeasement. Anim Behav 69:19-28 Bergmüller R, Heg D, Taborsky M (2005) Helpers in a cooperatively breeding cichlid stay and pay or disperse and breed, depending on ecological constraints. Proc R Soc Lond B 272:325-331 Birkhead TR, Møller AP (1992) Sperm Competition in Birds: Evolutionary Causes and Consequences. Academic Press, London Bleay C, Comendant T, Sinervo B (2007) An experimental test of frequencydependent selection on male mating strategy in the field. Proc R Soc Lond B 274:2019-2025 Blumer LS (1979) Male parental care in the bony fishes Q Rev Biol 54:149-161 Blumer LS (1982) A bibliography and categorization of bony fishes exhibiting parental care. Zool J Linn Soc 76:1-22 Brockmann HJ (1993) Parasitizing conspecifics: comparisons between Hymenoptera and birds. Trends Ecol Evol 8:2-4 Brockmann HJ (2001) The evolution of alternative strategies and tactics. Adv Stud Behav 30:1-51 Brockmann HJ (2003) Male competition and satellite behavior. In: Shuster CN, Barlow RB, Brockmann HJ (eds) The American Horseshoe Crab. Harvard University Press, Cambridge/MA, pp 50-82 Brockmann HJ (2008) Alternative reproductive tactis in insects. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 177-223 Brockmann HJ, Dawkins R (1979) Joint nesting in a digger wasp as an evolutionarity stable preadoption to social life. Behavior 71:203-244 Brockmann HJ, Penn D (1992) Male mating tactics in the horseshoe crab, Limulus polyphemus. Anim Behav 44:653-665 Brockmann HJ, Taborsky M (2008) Alternative reproductive tactics and the evolution of alternative allocation phenotypes. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 25-51 Brockmann HJ, Grafen A, Dawkins R (1979) Evolutionarily stable nesting strategy in a digger wasp. J Theor Biol 77:473-496

572

Michael Taborsky and H. Jane Brockmann

Brockmann HJ, Oliveira RF, Taborsky M (2008) Integrating mechanisms and function: prospects for future research. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 471-489 Brouwer L, Heg D, Taborsky M (2005) Experimental evidence for helper effects in a cooperatively breeding cichlid. Behav Ecol 16:667-673 Bruintjes R, Taborsky M (2008) Helpers in a cooperative breeder pay a high price to stay: effects of demand, helper size and sex. Anim Behav 75:1843-1850 Cade WH, Cade ES (1992) Male mating success, calling and searching behavior at high and low densities in the field cricket, Gryllus integer. Anim Behav 43:49-56 Caro TM, Bateson P (1986) Organization and ontogeny of alternative tactics. Anim Behav 34:1483-1499 Charnov EL (1982) The Theory of Sex Allocation. Princeton University Press, Princeton Clutton-Brock TH (1991) The Evolution of Parental Care. Princeton University Press, Princeton Clutton-Brock TH, Albon SD, Guinness FE (1986) Great expectations: dominance, breeding success and offspring sex ratios in red deer. Anim Behav 34:460-471 Connor RC, Read AJ, Wrangham R (2000) Male reproductive strategies and social bonds. In: Mann J, Connor RC, Tyack PL, Whitehead H (eds) Cetacean Societies: Field Studies of Dolphins and Whales. University of Chicago Press, Chicago, pp 247-269 Constantz GD (1975). Behavioural ecology of mating in the male Gila topminnow, Poeciliopsis occidentalis. Ecology 56:966-973 Correa C, Baeza JA, Hinojosa IA, Thiel M (2003) Male dominance hierarchy and mating tactics in the rock shrimp Rhynchocinetes typus (Decapoda: Caridea). J Crustacean Biol 23:33-45 Crnokrak P, Roff DA (1998) The genetic basis of the trade-off between calling and wing morph in males of the cricket Gryllus firmus. Evolution 52:11111118 Danforth BN, Desjardins CA (1999) Male dimorphism in Perdita portalis (Hymenoptera, Andrenidae) has arisen from preexisting allometric patterns. Insectes Soc 46:18-28 Danforth BN, Neff JL (1992) Male polymorphism and polyethism in Perdita texana (Hymenoptera, Andrenidae). Ann Entomol Soc Am 85:616-626 Dawkins R (1980) Good strategy or evolutionarily stable strategy? In: Barlow GW, Silverberg J (eds) Sociobiology: Beyond Nature/Nurture? Reports, definitions, and debate. Westview Press, Boulder, pp 331-367 Dempster ER, Lerner IM (1950) Heritability of threshold characters. Genetics 35:212-236 Denno RF (1994) The evolution of dispersal polymorphisms in insects – the influence of habitats, host plants and mates. Res Popul Ecol 36:127-135

Alternative reproductive tactics and life history phenotypes

573

Denno RF, Douglass LW, Jacobs D (1985) Crowding and host plant nutrition: environmental determinants of wing form in Prokelisia marginata. Ecology 66:1588-1596 Denno RF, Roderick GK, Peterson MA, Huberty AF, Döbel HG, Eubanks MD, Losey JE, Langellotto GA (1996) Habitat persistence underlies intraspecific variation in the dispersal strategies of planthoppers. Ecol Monogr 66:389-408 Dickinson JL (2004) A test of the importance of direct and indirect fitness benefits for helping decisions in western bluebirds. Behav Ecol 15:233-238 Dierkes P, Taborsky M, Kohler U (1999) Reproductive parasitism of broodcare helpers in a cooperatively breeding fish. Behav Ecol 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. Ecol Lett 8:968-975 Dierkes P, Taborsky M, Achmann R (2008) Multiple paternity in the cooperatively breeding fish Neolamprologus pulcher. Behav Ecol Sociobiol 62:15811589 Dingle H, Winchell R (1997) Juvenile hormone as a mediator of plasticity in insect life histories. Arch Insect Biochem Physiol 35:359-373 Dominey WJ (1980) Female mimicry in male bluegill sunfish – a genetic polymorphism? Nature 284:546-548 Dominey WJ (1984) Alternative mating tactics and evolutionarily stable strategies. Am Zool 24:385-396 Dunbar RIM (1982) Intraspecific variations in mating strategy. In: Bateson PPG, Klopfer PH (eds) Perspectives in Ethology. Plenum Press, New York, pp 385431 Dunning JB Jr (1992) CRC Handbook of Avian Body Masses. CRC Press, Boca Raton Eadie JM, Fryxell JM (1992) Density dependence, frequency dependence, and alternative nesting strategies in goldeneyes. Am Nat 140:621-641 Emlen DJ (1994) Environmental control of horn length dimorphism in the beetle Onthophagus acuminatus (Coleoptera, Scarabaeidae). Proc R Soc Lond B 256:131-136 Emlen DJ (1996) Artificial selection on horn length body size allometry in the horned beetle Onthophagus acuminatus (Coleoptera: Scarabaeidae). Evolution 50:1219-1230 Emlen DJ (1997) Alternative reproductive tactics and male-dimorphism in the horned beetle Onthophagus acuminatus (Coleoptera: Scarabaeidae). Behav Ecol Sociobiol 41:335-341 Emlen DJ (2008) The roles of genes and the environment in the expression and evolution of alternative tactics. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 85-108 Emlen DJ, Nijhout HF (1999) Hormonal control of male horn length dimorphism in the dung beetle Onthophagus taurus (Coleoptera: Scarabaeidae). J Insect Physiol 45:45-53

574

Michael Taborsky and H. Jane Brockmann

Emlen DJ, Nijhout HF (2000) The development and evolution of exaggerated morphologies in insects. Annu Rev Entomol 45:661-708 Emlen DJ, Nijhout HF (2001) Hormonal control of male horn length dimorphism in Onthophagus taurus (Coleoptera: Scarabaeidae): a second critical period of sensitivity to juvenile hormone. J Insect Physiol 47:1045-1054 Emlen ST, Oring LW (1977) Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223 Erbelding-Denk C, Schröder JH, Schartl M, Nanda I, Schmid M, Epplen JT (1994) Male polymorphism in Limia perugiae (Pisces: Poeciliidae). Behav Gen 24:95-101 Falconer DS (1952) The problem of environment and selection. Am Nat 86:293298 Falconer DS, Mackay TFC (1996) Introduction to Quantitative Genetics. Longman, Harlow Farr JA (1980) Social behavior patterns as determinants of reproductive success in the guppy, Poecilia reticulata Peters (Pisces, Poeciliidae) – an experimental study of the effects of intermale competition, female choice, and sexual selection. Behaviour 74:38-91 Feh C (1999) Alliances and reproductive success in Camargue stallions. Anim Behav 57:705-713 Field J (1994) Selection of host nests by intraspecific nest-parasitic digger wasps. Anim Behav 48:113-118 Fitzpatrick JL, Desjardins JK, Stiver KA, Montgomerie R, Balshine S (2006) Male reproductive suppression in the cooperatively breeding fish Neolamprologus pulcher. Behav Ecol 17:25-33 Forchhammer MC, Boomsma JJ (1998) Optimal mating strategies in nonterritorial ungulates: a general model tested on muskoxen. Behav Ecol 9:136-143 Forester DC, Lykens DV (1986) Significance of satellite males in a population of spring peepers (Hyla crucifer). Copeia 1986:719-724 Frankino WA, Pfennig DW (2001) Condition-dependent expression of trophic polyphenism: effects of individual size and competitive ability. Evol Ecol Res 3:939-951 Gadgil M (1972) Male dimorphism as a consequence of sexual selection. Am Nat 106:574-580 Garcia-Vazquez E, Moran P, Perez J, Martinez JL, Izquierdo JI, de Gaudemar B, Beall E (2002) Interspecific barriers between salmonids when hybridisation is due to sneak mating. Heredity 89:288-292 Gonçalves D, Teles M, Alpedrinha J, Oliveira RF (2008) Brain and gonadal aromatase activity and steroid hormone levels in female and polymorphic males of the peacock blenny Salaria pavo. Horm Behav 54:717-725 Gosling LM, Petrie M (1990) Lekking in topi: a consequence of satellite behaviour by small males at hotspots. Anim Behav 40:272-287 Gross MR (1982) Sneakers, satellites and parentals – polymorphic mating strategies in North-American sunfishes. Z Tierpsychol 60:1-26

Alternative reproductive tactics and life history phenotypes

575

Gross MR (1984) Sunfish, salmon, and the evolution of alternative reproductive strategies and tactics in fishes. In: Potts GW, Wootton RJ (eds) Fish Reproduction: Strategies and Tactics. Academic Press, London, pp 55-75 Gross MR (1985) Disruptive selection for alternative life histories in salmon. Nature 313:47-48 Gross MR (1996) Alternative reproductive strategies and tactics: diversity within sexes. Trends Ecol Evol 11:92-98 Halama KJ, Reznick DN (2001) Adaptation, optimality, and the meaning of phenotypic variation in natural populations. In: Orzack SH, Sober E (eds) Adaptationism and Optimality. Cambridge University Press, Cambridge, pp 242-272 Hamilton IM, Heg D (2008) Sex differences in the effect of social status on the growth of subordinates in a co-operatively breeding cichlid. J Fish Biol 72:1079-1088 Hansell M (2005) Animal Architecture. Oxford University Press, Oxford Harcourt AH, de Waal FBM (1992) Coalitions and Alliances in Humans and Other Animals. Oxford University Press, Oxford Hartfelder K, Emlen DJ (2005) Endocrine control of insect polyphenism. In: Gilbert LI, Iatrou K, Gill SS (eds) Comprehensive Molecular Insect Science. Vol 3: Endocrinology. Elsevier, Boston, pp 651-703 Hazel WN, Smock R, Johnson MD (1990) A polygenic model for the evolution and maintenance of conditional strategies. Proc R Soc Lond B 242:181-187 Hazel W, Smock R, Lively CM (2004) The ecological genetics of conditional strategies. Am Nat 163:888-900 Heath DD, Devlin RH, Heath JW, Iwama GK (1994) Genetic, environmental and interaction effects on the incidence of jacking in Oncorhynchus tshawytscha (Chinook salmon). Heredity 72:146-154 Heg D, Hamilton IM (2008) Tug-of-war over reproduction in a cooperatively breeding cichlid. Behav Ecol Sociobiol 62:1249-1257 Heg D, Bachar Z, Brouwer L, Taborsky M (2004a) Predation risk is an ecological constraint for helper dispersal in a cooperatively breeding cichlid. Proc R Soc Lond B 271:2367-2374 Heg D, Bender N, Hamilton IM (2004b) Strategic growth decisions in helper cichlids. Proc R Soc Lond B 271:S505-S508 Heg D, Bergmüller R, Bonfils D, Otti O, Bachar Z, Burri R, Heckel G, Taborsky M (2006) Cichlids do not adjust reproductive skew to the availability of independent breeding options. Behav Ecol 17:419-429 Heg D, Jutzeler E, Bonfils D, Mitchell JS (2008) Group composition affects male reproductive partitioning in a cooperatively breeding cichlid. Mol Ecol 17:4359-4370 Heg D, Jutzeler E, Mitchell JS, Hamilton IM (2009) Helpful female subordinate cichlids are more likely to reproduce. PLoS ONE 4:e5458, doi:10.1371/journal.pone.0005458 Henson SA, Warner RR (1997) Male and female alternative reproductive behaviors in fishes: a new approach using intersexual dynamics. Annu Rev Ecol Syst 28:571-592

576

Michael Taborsky and H. Jane Brockmann

Hews DK, Knapp R, Moore MC (1994) Early exposure to androgens affects adult expression of alternative male types in tree lizards. Horm Behav 28:96-115 Hogan-Warburg AJ (1966) Social behavior of the ruff, Philomachus pugnax (L.). Ardea 54:109-229 Hori M (1993) Frequency-dependent natural selection in the handedness of scaleeating cichlid fish. Science 260:216-219 Howard RD (1984) Alternative mating behaviors of young male bullfrogs. Am Zool 24:397-406 Hugie DM, Lank DB (1997) The resident’s dilemma: a female choice model for the evolution of alternative mating strategies in lekking male ruffs (Philomachus pugnax). Behav Ecol 8:218-225 Hutchings JA, Jones MEB (1998) Life history variation and growth rate thresholds for maturity in Atlantic salmon, Salmo salar. Can J Fish Aquat Sci 55:22-47 Jennions MD, Petrie M (2000) Why do females mate multiply? A review of the genetic benefits. Biol Rev 75:21-64 Johnstone RA (2000) Models of reproductive skew: a review and synthesis. Ethology 106:5-26 Jones MW, Hutchings JA (2002) Individual variation in Atlantic salmon fertilization success: implications for effective population size. Ecol Appl 12:184-193 Jukema J, Piersma T (2006) Permanent female mimics in a lekking shorebird. Biol Lett 2:161-164 Kallman KD (1984) A new look at sex determination in poeciliid fishes. In: Turner BJ (ed) Evolutionary Genetics of Fishes. Plenum Press, New York, pp 95-171 Kallman KD (1989) Genetic control of size at maturity in Xiphophorus. In: Meffe GK, Snelson FF Jr (eds) Ecology and Evolution of Livebearing Fishes (Poeciliidae). Prentice-Hall, Englewood Cliffs/NJ, pp 163-184 Keller L, Reeve HK (1994) Partitioning of reproduction in animal societies. Trends Ecol Evol 9:98-102 Kellog KA, Markert JA, Stauffer JR, Kocher TD (1998) Intraspecific brood mixing and reduced polyandry in a maternal mouth-brooding cichlid. Behav Ecol 9:309-312 Kraak SBM, Pen I (2002) Sex-determining mechanisms in vertebrates. In: Hardy ICW (ed) Sex Ratios: Concepts and Research Methods. Cambridge University Press, Cambridge, pp 158-177 Krüger O (2008) Alternative reproductive tactics in birds. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 343-355 Lande R, Arnold SJ (1983) The measurement of selection on correlated characters. Evolution 37:1210-1226 Langbein J, Thirgood SJ (1989) Variation in mating systems of fallow deer (Dama dama) in relation to ecology. Ethology 83:195-214 Langellotto GA, Denno RF (2001) Benefits of dispersal in patchy environments: mate location by males of a wing-dimorphic insect. Ecology 82:1870-1878

Alternative reproductive tactics and life history phenotypes

577

Lank DB, Smith CM, Hanotte O, Burke T, Cooke F (1995) Genetic polymorphism for alternative mating behaviour in lekking male ruff Philomachus pugnax. Nature 378:59-62 Le Boeuf BJ (1974) Male-male competition and reproductive success in elephant seals. Am Zool 14:163-176 Lee JSF (2005) Alternative reproductive tactics and status-dependent selection. Behav Ecol 16:566-570 Levins R (1968) Evolution in Changing Environments: Some Theoretical Explorations. Princeton University Press, Princeton Lively CM (1986) Canalization versus developmental conversion in a spatially variable environment. Am Nat 128:561-572 Lloyd DG (1987) Parallels between sexual strategies and other allocation strategies. In: Stearns SC (ed) The Evolution of Sex and Its Consequences. Birkhäuser Verlag, Basel, pp 263-281 Lucas JR, Howard RD (1995) On alternative reproductive tactics in anurans: dynamic games with density and frequency dependence. Am Nat 146:365-397 Lucas JR, Howard RD (2008) Modeling alternative mating tactics as dynamic games. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 63-82 Lucas JR, Howard RD, Palmer JG (1996) Callers and satellites: chorus behaviour in anurans as a stochastic dynamic game. Anim Behav 51:501-518 Magnhagen C (1992) Alternative reproductive behaviour in the common goby Pomatoschistus microps: an ontogenetic gradient? Anim Behav 44:182-184 Martin E, Taborsky M (1997) Alternative male mating tactics in a cichlid, Pelvicachromis pulcher: a comparison of reproductive effort and success. Behav Ecol Sociobiol 41:311-319 Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge Mboko SK, Kohda M (1999) Piracy mating by large males in a monogamous substrate-breeding cichlid in Lake Tanganyika. J Ethol 17:51-55 Mitchell JS, Jutzeler E, Heg D, Taborsky M (2009a) Gender differences in the costs that subordinate group members impose on dominant males in a cooperative breeder. Ethology 115:1162-1174 Mitchell JS, Jutzeler E, Heg D, Taborsky M (2009b) Dominant members of cooperatively-breeding groups adjust their behaviour in response to the sexes of subordinates. Behaviour 146:1665-1686 Moczek AP, Emlen DJ (2000) Male horn dimorphism in the scarab beetle, Onthophagus taurus: do alternative reproductive tactics favour alternative phenotypes? Anim Behav 59:459-466 Moczek AP, Nijhout HF (2002) Developmental mechanisms of threshold evolution in a polyphenic beetle. Evol Dev 4:252-264 Moore MC (1991) Application of organization-activation theory to alternative male reproductive strategies: a review. Horm Behav 25:154-179

578

Michael Taborsky and H. Jane Brockmann

Moore MC, Hews DK, Knapp R (1998) Hormonal control and evolution of alternative male phenotypes: generalizations of models for sexual differentiation. Am Zool 38:133-151 Moran NA (1992) The evolutionary maintenance of alternative phenotypes. Am Nat 139:971-989 Morris MR, Nicoletto PF, Hesselman E (2003) A polymorphism in female preference for a polymorphic male trait in the swordtail fish Xiphophorus cortezi. Anim Behav 65:45-52 Muñoz RC, Warner RR (2004) Testing a new version of the size-advantage hypothesis for sex change: sperm competition and size-skew effects in the bucktooth parrotfish, Sparisoma radians. Behav Ecol 15:129-136 Neff BD (2004) Increased performance of offspring sired by parasitic males in bluegill sunfish. Behav Ecol 15:327-331 Neff BD (2008) Alternative mating tactics and mate choice for good genes or good care. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 421-434 Neff BD, Gross MR (2001) Dynamic adjustment of parental care in response to perceived paternity. Proc R Soc Lond B 268:1559-1565 Nijhout HF (1999) Hormonal control in larval development and evolution: insects. In: Hall BK, Wake MH (eds) The Origin and Evolution of Larval Forms. Academic Press, New York, pp 217-254 Nijhout HF (2003) Development and evolution of adaptive polyphenisms. Evol Dev 5:9-18 Nijhout HF, Wheeler DE (1982) Juvenile hormone and the physiological basis of insect polymorphisms. Q Rev Biol 57:109-133 Oliveira RF (2004) Social modulation of androgens in vertebrates: mechanisms and function. Adv Stud Behav 34:165-239 Oliveira RF, Gonçalves EJ, Santos RS (2001) Gonadal investment of young males in two blenniid fishes with alternative mating tactics. J Fish Biol 59:459-462 Oliveira RF, Carvalho N, Miranda J, Gonçalves EJ, Grober M, Santos RS (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 Oliveira RF, Taborsky M, Brockmann HJ (2008a) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge Oliveira RF, Canário AVM, Ros AFH (2008b) Hormones and alternative reproductive tactics in vertebrates. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 132-174 Packer C (1977) Reciprocal altruism in Papio anubis. Nature 265:441-443 Parker GA (1984) Evolutionarily stable strategies. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 3061

Alternative reproductive tactics and life history phenotypes

579

Pemberton JM, Coltman DW, Smith JA, Bancroft DR (2004) Mating patterns and male breeding success. In: Clutton-Brock TH, Pemberton J (eds) Soay Sheep: Dynamics and Selection in an Island Population. Cambridge University Press, Cambridge, pp 166-189 Perrill SA, Gerhardt HC, Daniel RE (1982) Mating strategy shifts in male green treefrogs (Hyla cinerea): an experimental study. Anim Behav 30:43-48 Petrie M, Møller AP (1991) Laying eggs in other’s nests: intraspecific brood parasitism in birds. Trends Ecol Evol 6:315-320 Piché J, Hutchings JA, Blanchard W (2008) Genetic variation in threshold reaction norms for alternative reproductive tactics in male Atlantic salmon, Salmo salar. Proc R Soc Lond B 275:1571-1575 Pienaar J, Greeff JM (2003) Different male morphs of Otitesella pseudoserrata fig wasps have equal fitness but are not determined by different alleles. Ecol Lett 6:286-289 Pizzo A, Roggero A, Palestrini C, Moczek AP, Rolandoa A (2008) Rapid shape divergences between natural and introduced populations of a horned beetle partly mirror divergences between species. Evol Dev 10:166-175 Plaistow SJ, Tsubaki Y (2000) A selective trade-off for territoriality and nonterritoriality in the polymorphic damselfly Mnais costalis. Proc R Soc Lond B 267:969-975 Plaistow SJ, Johnstone RA, Colegrave N, Spencer M (2004) Evolution of alternative mating tactics: conditional versus mixed strategies. Behav Ecol 15:534542 Radwan J (1993) The adaptive significance of male polymorphism in the acarid mite Caloglyphus berlesei. Behav Ecol Sociobiol 33:201-208 Radwan J (1995) Male morph determination in two species of acarid mites. Heredity 74:669-673 Radwan J, Unrug J, Tomkins JL (2002) Status-dependence and morphological trade-offs in the expression of a sexually selected character in the mite, Sancassania berlesei. J Evol Biol 15:744-752 Reichard M, Le Comber SC, Smith C (2007) Sneaking from a female perspective. Anim Behav 74:679-688 Renn SCP, Aubin-Horth N, Hofmann HA (2008) Fish and chips: functional genomics of social plasticity in an African cichlid fish. J Exp Biol 211:30413056 Repka J, Gross MR (1995) The evolutionarily stable strategy under individual condition and tactic frequency. J Theor Biol 176:27-31 Rhen T, Crews D (2002) Variation in reproductive behaviour within a sex: neural systems and endocrine activation. J Neuroendocrinol 14:517-531 Ribbink AJ (1977) Cuckoo among Lake Malawi chichlid fish. Nature 267:243244 Rios-Cardenas O, Tudor MS, Morris MR (2007) Female preference variation has implications for the maintenance of an alternative mating strategy in a swordtail fish. Anim Behav 74:633-640

580

Michael Taborsky and H. Jane Brockmann

Robertson DR, Reinboth R, Bruce RW (1982) Gonochorism, protogyneous sexchange and spawning in three sparisomatinine parrotfishes from the western Indian Ocean. Bull Mar Sci 32:868-879 Rodgers EW, Earley RL, Grober MS (2007) Social status determines sexual phenotype in the bi-directional sex changing bluebanded goby Lythrypnus dalli. J Fish Biol 70:1660-1668 Roff DA (1990) Antagonistic pleiotropy and the evolution of wing dimorphism in the sand cricket, Gryllus firmus. Heredity 65:169-177 Roff DA (1996) The evolution of threshold traits in animals. Q Rev Biol 71:3-35 Roff DA (1998) Evolution of threshold traits: the balance between directional selection, drift and mutation. Heredity 80:25-32 Ross RM (1990) The evolution of sex-change mechanisms in fishes. Environ Biol Fishes 29:81-93 Rowland WJ (1979) Stealing fertilizations in the fourspine stickleback, Apeltes quadracus. Am Nat 114:602-604 Rowland JM, Emlen DJ (2009) Two thresholds, three male forms result in facultative male trimorphism in beetles. Science 323:773-776 Ryan MJ, Causey BA (1989) Alternative mating behavior in the swordtails Xiphophorus nigrensis and Xiphophorus pygmaeus (Pisces, Poeciliidae). Behav Ecol Sociobiol 24:341-348 Ryan MJ, Wagner WE (1987) Asymmetries in mating preferences between species: female swordtails prefer heterospecific males. Science 236:595-597 Ryan MJ, Pease CM, Morris MR (1992) A genetic polymorphism in the swordtail, Xiphophorus nigrensis: testing the prediction of equal fitnesses. Am Nat 139:21-31 Sandell MI, Diemer M (1999) Intraspecific brood parasitism: a strategy for floating females in the European starling. Anim Behav 57:197-202 Schaedelin FC, Taborsky M (2009) Extended phenotypes as signals. Biol Rev 84:293-313 Scheiner SM (1993) Genetics and evolution of phenotypic plasticity. Annu Rev Ecol Syst 24:35-68 Schlichting CD, Pigliucci M (1998) Phenotypic Evolution: A Reaction Norm Perspective. Sinauer Associates, Sunderland/MA Schütz D, Pachler G, Ripmeester E, Goffinet O, Taborsky M (2010) Reproductive investment of giants and dwarfs: specialized tactics in a cichlid fish with alternative male morphs. Funct Ecol 24:131-140 Seger J, Brockmann HJ (1987) What is bet-hedging? In: Harvey PH, Partridge L (eds) Oxford Surveys of Evolutionary Biology. Oxford University Press, Oxford, pp 182-211 Setchell JM (2008) Alternative reproductive tactics in primates. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 373-398 Shapiro AM (1976) Seasonal polyphenism. Evol Biol 9:259-333 Shapiro DY (1987) Differentiation and evolution of sex change in fishes. Bioscience 37:490-497

Alternative reproductive tactics and life history phenotypes

581

Shuster SM (1992) The reproductive behavior of α-, β-, and γ -male morphs in Paracerceis sculpta, a marine isopod crustacean. Behaviour 121:231-257 Shuster SM (2008) The expression of crustacean mating strategies. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 224-250 Shuster SM, Sassaman C (1997) Genetic interaction between male mating strategy and sex ratio in a marine isopod. Nature 388:373-377 Shuster SM, Wade MJ (1991) Equal mating success among male reproductive strategies in a marine isopod. Nature 350:608-610 Shuster SM, Wade MJ (2003) Mating Systems and Strategies. Princeton University Press, Princeton Shuster SM, Zinser G, Keim P, Ballard JWO, Sassaman C (2001) The influence of genetic and extrachromosomal factors on population sex ratio in the marine isopod, Paracerceis sculpta. In: Kensley BF, Brusca RC (eds) Crustacean Issues 13: Isopod Systematics and Evolution. Balkema, Brookfield, pp 313-326 Simmons LW, Teale RJ, Maier M, Standish RJ, Bailey WJ, Withers PC (1992) Some costs of reproduction for male bush-crickets, Requena verticalis (Orthoptera, Tettigoniidae) – allocating resources to mate attraction and nuptial feeding. Behav Ecol Sociobiol 31:57-62 Sinervo B, Lively CM (1996) The rock-paper-scissors game and the evolution of alternative male strategies. Nature 380:240-243 Sinervo B, Svensson E, Comendant T (2000) Density cycles and an offspring quantity and quality game driven by natural selection. Nature 406:985-988 Sinervo B, Bleay C, Adamopoulou C (2001) Social causes of correlational selection and the resolution of a heritable throat color polymorphism in a lizard. Evolution 55:2040-2052 Sinervo B, Chaine A, Clobert J, Calsbeek R, Hazard L, Lancaster L, McAdam AG, Alonzo S, Corrigan G, Hochberg ME (2006) Self-recognition, color signals, and cycles of greenbeard mutualism and altruism. Proc Natl Acad Sci USA 103:7372-7377 Sirot LK, Brockmann HJ, Marinis C, Muschett G (2003) Maintenance of a female-limited polymorphism in Ischnura ramburi (Zygoptera: Coenagrionidae). Anim Behav 66:763-775 Skubic E, Taborsky M, McNamara JM, Houston AI (2004) When to parasitize? A dynamic optimization model of reproductive strategies in a cooperative breeder. J Theor Biol 227:487-501 Skúlason S, Smith TB (1995) Resource polymorphisms in vertebrates. Trends Ecol Evol 10:366-370 Smith TB, Skúlason S (1996) Evolutionary significance of resource polymorphisms in fish, amphibians and birds. Annu Rev Ecol Syst 27:111-134 Smuts BB, Smuts RW (1993) Male aggression and sexual coercion of females in nonhuman primates and other mammals: evidence and theoretical implications. Adv Stud Behav 22:1-63 Sohn JJ (1977) Socially induced inhibition of genetically determined maturation in platyfish, Xiphophorus maculatus. Science 195:199-201

582

Michael Taborsky and H. Jane Brockmann

St-Cyr S, Aubin-Horth N (2009) Integrative and genomics approaches to uncover the mechanistic bases of fish behavior and its diversity. Comp Biochem Physiol A 152:9-21 Stiver KA, 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. J Fish Biol 65:91-105 Stiver KA, Dierkes P, Taborsky M, Gibbs HL, Balshine S (2005) Relatedness and helping in fish: examining the theoretical predictions. Proc R Soc Lond B 272:1593-1599 Stiver KA, Fitzpatrick J, Desjardins JK, Balshine S (2006) Sex differences in rates of territory joining and inheritance in a cooperatively breeding cichlid fish. Anim Behav 71:449-456 Taborsky M (1984) Broodcare helpers in the cichlid fish Lamprologus brichardi – their costs and benefits. Anim Behav 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 (1994) Sneakers, satellites, and helpers: parasitic and cooperative behavior in fish reproduction. Adv Stud Behav 23:1-100 Taborsky M (1997) Bourgeois and parasitic tactics: do we need collective, functional terms for alternative reproductive behaviours? Behav Ecol Sociobiol 41:361-362 Taborsky M (1998) Sperm competition in fish: bourgeois males and parasitic spawning. Trends Ecol Evol 13:222-227 Taborsky M (1999) Conflict or cooperation: what determines optimal solutions to competition in fish reproduction? In: Almada VC, Oliveira RF, Gonçalves EJ (eds) Behaviour and Conservation of Littoral Fishes. Instituto Superior de Psicologia Aplicada, Lisboa, pp 301-349 Taborsky M (2001) The evolution of parasitic and cooperative reproductive behaviors in fishes. J Heredity 92:100-110 Taborsky M (2008) Alternative reproductive tactics in fish. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 251-299 Taborsky M (2009) Reproductive skew in cooperative fish groups: virtue and limitations of alternative modeling approaches. In: Hager R, Jones CB (eds) Reproductive Skew in Vertebrates: Proximate and Ultimate Causes. Cambridge University Press, Cambridge, pp 265-304 Taborsky M, Limberger D (1981) Helpers in fish. Behav Ecol Sociobiol 8:143145 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 RF, Brockmann HJ (2008) The evolution of alternative reproductive tactics: concepts and questions. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 1-21

Alternative reproductive tactics and life history phenotypes

583

Takahashi S, Hori M (1994) Unstable evolutionarily stable strategy and oscillation: a model of lateral asymmetry in scale-eating cichlids. Am Nat 144:10011020 Tallamy DW (2005) Egg dumping in insects. Annu Rev Entomol 50:347-370 Tallamy DW, Horton LA (1990) Costs and benefits of the eggdumping alternative in Gargaphia lace bugs (Hemiptera: Tingidae). Anim Behav 39:352-359 Thériault V, Dodson JJ (2003) Body size and the adoption of a migratory tactic in brook charr. J Fish Biol 63:1144-1159 Thomaz D, Beall E, Burke T (1997) Alternative reproductive tactics in Atlantic salmon: factors affecting mature parr success. Proc R Soc Lond B 264:219226 Thornhill R, Alcock J (1983) The Evolution of Insect Mating Systems. Harvard University Press, Cambridge/MA Thorpe JE (1986) Age at first maturity in Atlantic Salmon, Salmo salar: freshwater period influences and conflicts with smelting. In: Meerburg DJ (ed) Salmonid Age at Maturity. Can Spec Publ Fish Aquat Sci 89:7-14 Thorpe JE, Morgan RIG (1980) Growth rate and smolting rate of progeny of male Atlantic salmon parr (Salmo salar L.). J Fish Biol 17:451-459 Thorpe JE, Mangel M, Metcalfe NB, Huntingford FA (1998) Modelling the proximate basis of salmonid life-history variation, with application to Atlantic salmon, Salmo salar L. Evol Ecol 12:581-599 Tomkins JL, Brown GS (2004) Population density drives the local evolution of a threshold dimorphism. Nature 431:1099-1103 Tomkins JL, Hazel W (2007) The status of the conditional evolutionarily stable strategy. Trends Ecol Evol 22:522-528 Tomkins JL, Moczek AP (2009) Patterns of threshold evolution in polyphenic insects under different developmental models. Evolution 63:459-468 Tomkins JL, Simmons LW, Alcock J (2001) Brood-provisioning strategies in Dawson’s burrowing bee, Amegilla dawsoni (Hymenoptera: Anthophorini). Behav Ecol Sociobiol 50:81-89 Tomkins JL, LeBas NR, Unrug J, Radwan J (2004) Testing the status-dependent ESS model: population variation in fighter expression in the mite Sancassania berlesei. J Evol Biol 17:1377-1388 Tomkins JL, Kotiaho JS, LeBas NR (2005) Phenotypic plasticity in the developmental integration of morphological trade-offs and secondary sexual trait compensation. Proc R Soc Lond B 272:543-551 Toth AL, Varala K, Newman TC, Miguez FE, Hutchison SK, Willoughby DA, Simons JF, Egholm M, Hunt JH, Hudson ME, Robinson GE (2007) Wasp gene expression supports an evolutionary link between maternal behavior and eusociality. Science 318:441-444 Trumbo ST (1996) Parental care in invertebrates. Adv Stud Behav 25:3-51 Tsubaki Y (2003) The genetic polymorphism linked to mate-securing strategies in the male damselfly Mnais costalis Selys (Odonata: Calopterygidae). Popul Ecol 45:263-266

584

Michael Taborsky and H. Jane Brockmann

Tsubaki Y, Ono T (1986) Competition for territorial sites and alternative mating tactics in the dragonfly, Nannophya pygmaea Rambur (Odonata, Libellulidae). Behaviour 97:234-252 Tsubaki Y, Hooper RE, Siva-Jothy MT (1997) Differences in adult and reproductive lifespan in the two male forms of Mnais pruinosa costalis Selys (Odonata: Calopterygidae). Res Popul Ecol 39:149-155 Unrug J, Tomkins JL, Radwan J (2004) Alternative phenotypes and sexual selection: can dichotomous handicaps honestly signal quality? Proc R Soc Lond B 271:1401-1406 Utami SS, Goossens B, Bruford MW, de Ruiter JR, van Hooff JARAM (2002) Male bimaturism and reproductive success in Sumatran orang-utans. Behav Ecol 13:643-652 van den Assem J (1967) Territory in the three-spined stickleback Gasterosteus aculeatus L.: an experimental study in intra-specific competition. Behaviour (suppl) 16:1-164 van den Berghe EP (1988) Piracy: a new alternative male reproductive tactic. Nature 334:697-698 van Rhijn JG (1973). Behavioural dimorphism in male ruffs, Philomachus pugnax (L.). Behaviour 47:153-229 Vehrencamp SL (1983) A model for the evolution of despotic versus egalitarian societies. Anim Behav 31:667-682 Via S, Lande R (1985) Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39:505-522 Wagner WE (2005) Male field crickets that provide reproductive benefits to females incur higher costs. Ecol Entomol 30:350-357 Walker SE, Cade WH (2003) A simulation model of the effects of frequency dependence, density dependence and parasitoid flies on the fitness of male field crickets. Ecol Modell 169:119-130 Waltz EC (1982) Alternative mating tactics and the law of diminishing returns – the satellite threshold model. Behav Ecol Sociobiol 10:75-83 Waltz EC, Wolf LL (1984) By Jove – why do alternative mating tactics assume so many different forms? Am Zool 24:333-343 Warner RR, Robertson DR, Leigh EG Jr (1975) Sex change and sexual selection. Science 190:633-638 Watts DP (1998) Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale National Park, Uganda. Behav Ecol Sociobiol 44:43-55 West-Eberhard MJ (1989) Phenotypic plasticity and the origins of diversity. Annu Rev Ecol Syst 20:249-278 West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford University Press, Oxford Westneat DF, Stewart IRK (2003) Extra-pair paternity in birds: causes, correlates, and conflict. Annu Rev Ecol Evol Syst 34:365-396 Whitfield CW, Cziko A-M, Robinson GE (2003) Gene expression profiles in the brain predict behavior in individual honey bees. Science 302:296-299 Widemo F (1998) Alternative reproductive strategies in the ruff, Philomachus pugnax: a mixed ESS? Anim Behav 56:329-336

Alternative reproductive tactics and life history phenotypes

585

Wikelski M, Carbone C, Trillmich F (1996) Lekking in marine iguanas: female grouping and male reproductive strategies. Anim Behav 52:581-596 Wikelski M, Steiger SS, Gall B, Nelson KN (2005) Sex, drugs and mating role: testosterone-induced phenotype-switching in Galapagos marine iguanas. Behav Ecol 16:260-268 Wirtz P (1982) Territory holders, satellite males and bachelor males in a high density population of waterbuck (Kobus ellipsiprymnus) and their associations with conspecifics. Z Tierpsychol 58:277-300 Wolf M, van Doorn GS, Leimar O, Weissing FJ (2007) Life-history trade-offs favour the evolution of animal personalities. Nature 447:581-584 Wolff JO (2008) Alternative reproductive tactics in nonprimate male mammals. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 356-372 Wolff JO, Cicirello DM (1990) Mobility versus territoriality: alternative reproductive strategies in white-footed mice. Anim Behav 39:1222-1224 Yanagisawa Y (1985) Parental strategy of the cichlid fish Perissodus microlepis, with particular reference to intraspecific brood ‘farming out’. Env Biol Fish 12:241-249 Yom-Tov Y (2001) An updated list and some comments on the occurrence of intraspecific nest parasitism in birds. Ibis 143:133-143 Yom-Tov Y (1980) Intraspecific nest parasitism in birds. Biol Rev 55:93-108 Zamudio KR, Chan LM (2008) Alternative reproductive tactics in amphibians. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 300-331 Zamudio KR, Sinervo E (2000) Polygyny, mate-guarding, and posthumous fertilization as alternative male mating strategies. Proc Natl Acad Sci USA 97:14427-14432 Zera AJ, Harshman LG (2001) The physiology of life history trade-offs in animals. Annu Rev Ecol Syst 32:95-126 Zera AJ, Huang Y (1999) Evolutionary endocrinology of juvenile hormone esterase: functional relationship with wing polymorphism in the cricket, Gryllus firmus. Evolution 53:837-847 Zera AJ, Rankin MA (1989) Wing dimorphism in Gryllus rubens: genetic basis of morph determination and fertility differences between morphs. Oecologia 80:249-255 Zera AJ, Zhang C (1995) Evolutionary endocrinology of juvenile hormone esterase in Gryllus assimilis: direct and correlated responses to selection. Genetics 141:1125-1134 Zhao Z, Zera AJ (2002) Differential lipid biosynthesis underlies a tradeoff between reproduction and flight capability in a wing-polymorphic cricket. Proc Natl Acad Sci USA 99:16829-16834 Zimmerer EJ, Kallman KD (1989) Genetic basis for alterntive reproductive tactics in the pygmy swordtail, Xiphophorus nigrensis. Evolution 43:1298-1307

586

Michael Taborsky and H. Jane Brockmann

Zink AG (2003) Intraspecific brood parasitism as a conditional reproductive tactic in the treehopper Publilia concava. Behav Ecol Sociobiol 54:406-415 Zupanc GKH, Lamprecht J (2000) Towards a cellular understanding of motivation: structural reorganization and biochemical switching as key mechanisms of behavioral plasticity. Ethology 106:467-477

Chapter 19

Animal personality and behavioural syndromes RALPH BERGMÜLLER

ABSTRACT Behaviours are considered to be among the most flexible traits in animals and often reflect conditional responses upon the behaviours of others, or to changing environmental conditions. However, in contrast to the presumed advantages of behavioural flexibility, individuals within the same species or populations often consistently differ in their behaviour, that is, some individuals are consistently more aggressive, more explorative, or shyer than others. This phenomenon has been termed ‘animal personality’ or a ‘behavioural syndrome’. This chapter gives an overview about the current state of research in the field of animal personality by summarising the evolutionary concepts that have been put forward to explain the three questions arising from the phenomenon, i.e. why individuals are consistent in their behaviour, why there are individual differences in behaviour and why behavioural traits are sometimes correlated among each other. It has been shown that animal personality is heritable and entails fitness consequences, which demonstrates that it is subject to evolutionary processes. Moreover, as consistent individual differences in behaviour can result from developmental processes, research on animal personality integrates proximate and functional questions about animal behaviour. Therefore, animal personality research provides an integrative approach to understand animal behaviour by taking into account the causes and potential effects of intrinsic individual differences in behaviour.

19.1 Introduction Behaviours are among the most flexible traits in animals and largely reflect conditional responses upon the behaviours of others or to changing environmental conditions and are therefore sometimes assumed to show largely unlimited plasticity (Sih et al. 2004a). However, when observing animals

588

Ralph Bergmüller

for some time, it often seems that individuals differ in their behaviour, that is, some individuals appear to be more aggressive, more explorative, or shyer than others, suggesting some limitations to plasticity. Indeed, a growing number of systematic studies provided evidence showing that individuals in many species consistently differ in their behaviour, sometimes even in whole suites of behavioural traits that can be important in different functional contexts and reflect inherent individual differences in behavioural type (Gosling and John 1999, Sih et al. 2004a, Dingemanse and Réale 2005, Réale et al. 2007). In fish (Huntingford 1976, Bell 2005, Bergmüller and Taborsky 2007, Dingemanse et al. 2007), birds (Verbeek et al. 1994) and rodents (Koolhaas et al. 1999), individuals that are more aggressive towards conspecifics have also been found to be bolder when exploring novel environments or when encountering predators. This phenomenon has been termed ‘animal personality’ (Gosling and John 1999, Drent et al. 2003) or a ‘behavioural syndrome’ (Sih et al. 2004a). Consistent individual differences in behaviour across time and context have been discovered in many species, ranging from insects to humans, suggesting that the phenomenon is widespread (Koolhaas et al. 1999, Sih et al. 2004b). Recent reviews have also highlighted the parallels between personality in humans and other animals (Wilson et al. 1994, Gosling 2001). But why do these differences between individuals exist? Why do behaviours in different functional contexts sometimes correlate among each other? These questions about animal behaviour are equally interesting from an ultimate (evolutionary) and from a proximate (mechanistic) perspective. Behavioural ecologists traditionally aim at understanding why animals behave the way they do by assuming that behaviours are adaptations. The underlying idea is that behaviours result from long-term selection pressures that have adjusted the responses of animals to specific situations so that they show, on average, adaptive responses to these situations (Kappeler and Kraus this volume). For instance, species or populations that live under high predation risk should have evolved different behavioural responses to this challenge, compared to populations facing low predation risk. As a consequence, animals in a population are expected to evolve towards a mean optimal phenotype to cope with a particular challenge, i.e. the trait should produce the highest fitness (Krebs and Davies 1997). However, assuming such a mean optimal phenotype has diverted attention away from variation around the mean, which therefore has been regarded as some sort of ‘noise’, merely reflecting stochastic effects, and, thus, not relevant in evolutionary terms. In contrast, evolutionary personality research is interested precisely in this variation. The question is whether the observed variation might itself

Animal personality and behavioural syndromes

589

be a result of evolution and whether searching only for the optimal phenotype may not always provide a suitable answer for why animals behave in a certain way (Wilson 1998). For instance, within the same population, different individuals might be differently affected or exposed to predation, resulting in individual differences in tendencies in behaviours that are used to cope with this challenge. Alternatively, different individuals might use different strategies to cope with the same ecological or social challenges. Hence, similarly to research on alternative strategies (Schneider and Fromhage this volume, Taborsky and Brockman this volume), sexual selection (Kempenaers and Schlicht this volume) and cooperation (Bshary, Heinze, and Kerth this volume), research on animal personality emphasises the fact that there is often not only one solution to grow up, survive and reproduce. This switch in perspective raises the question why different types of individuals often co-exist in a population, whether there is heritable variation in these traits and whether they have fitness consequences. Other interesting questions include how individual differences develop during ontogeny and which mechanisms are involved in the expression of different behavioural phenotypes. In this chapter, I will introduce and discuss the current scientific state of affairs in animal personality research in regard to these issues. 19.1.1 The phenomenon of personality 19.1.1.1 Concepts and definitions The term ‘animal personality’ describes a particular aspect of the behavioural phenotype (BOX 19.1). While personality (the individual type) is an individual-level characteristic (Sih et al. 2004a, Bell 2007; Fig. 19.1), the term behavioural syndrome describes a characteristic of a population (see BOX 19.1). Other terms that are used in the same context but with slightly different meanings are ‘coping style’ (a coherent set of behavioural and physiological stress responses, Koolhaas et al. 1999), ‘coping strategy’ (a set of strategies – escape, remove, search, wait – to cope with aversive situations, Wechsler 1995) and ‘animal temperament’ (characteristics of individuals that describe consistent patterns in feeling, thinking and behaving; in human research: the inherited, early appearing tendencies that continue throughout life and serve as foundation to personality (Gosling 2001, Réale et al. 2007). Here I will use the terms ‘animal personality’ and ‘behavioural syndrome’, which are the terms currently used most commonly in the field.

590

Ralph Bergmüller

BOX 19.1 Definitions of animal personality and behavioural syndromes Animal personality describes the behavioural phenotype that is consistent over time and differs between individuals of the same species, i.e. integrated behavioural phenotypes and stable traits that are consistent over time and across situations (Budaev 1998, Drent et al. 2003). A behavioural syndrome is defined more broadly and involves behavioural consistency within, and also between individuals. Within individual consistency is present when individuals behave in a consistent way in similar situations, i.e., individuals have a behavioural type (i.e. animal personality). Between individual consistency occurs when the rank order differences among individuals are maintained when the environment changes (Fig. 19.1 and 19.2a; but the behavioural type does not need to remain constant in different situations; Bell 2007, Sih and Bell 2008) which would statistically be reflected as a behavioural correlation between behaviours in different contexts or situations (Fig. 19.2b). The three key questions that arise from the phenomena of animal personality and behavioural syndromes are: (1) (2) (3)

Why are individuals consistent in behaviours? Why do individuals within species differ in behaviour? Why do behaviours in different functional contexts correlate among each other?

Behaviour Y

Behavioural type: individuals with a particular combination of behaviour X and Y

Behavioural syndrome: correlation between behaviour X and Y Behaviour X

Fig. 19.1 Definitions of behavioural type and behavioural syndrome. Each data point represents a different individual in the population. (With permission from Bell 2007)

Animal personality and behavioural syndromes a

591

7 6

Behaviour

5 4 3 2 1 0

Environment 1 b

Environment 2

Behaviour in Environment 2

7 6 5 4 3 2 1 0

0

1

2

3

4

5

6

7

Behaviour in Environment 1

Fig. 19.2 (a) Different individuals show different behavioural responses in environments 1 and 2. Horizontal lines depict that individuals remain consistent in their behaviour across environments, while all other lines depict a phenotypic plastic response. Many parallel lines depict that rank order differences are maintained across individuals which reflect a behavioural correlation in graph (b).

To illustrate the difference of within and between individual consistency, it is useful to consider reaction norms, which describe the trait changes of individuals (strictly speaking genotypes) across different environments or situations (Fig. 19.2 redrawn after Sih et al. 2004a). If individuals adjust their behaviour between two different environments, they show a phenotypic plastic response (Fig. 19.2a). Within individual consistency in behaviour across contexts or situations is depicted by a straight

592

Ralph Bergmüller

line in Fig. 19.2a (the individual behaves the same way regardless of changes in the environment or situation). In Fig. 19.2a between individual consistency becomes apparent in the rank order differences between individuals that is maintained across different environments (parallel lines) which result in a cross environment correlation between behaviours in Fig. 19.2b. As reaction norms describe the compromise of individuals to optimise fitness across different environmental conditions, many reaction norms with the same compromise or fitness value can exist at any point of time. 19.1.1.2 What is not personality? There are many reasons for why individuals consistently differ in behaviours for some time. For instance, if behavioural measurements are done in the wild, consistent individual differences can result from persisting differences in the micro-environment of individuals, including their social position (Dingemanse et al. 2009a, Bergmüller and Taborsky in revision). But not all these differences do necessarily correspond to personality differences. Individuals may vary because they differ in state, such as body condition, social rank, life history stage or parasite load. Even though these differences can result in consistently different behaviours among individuals, they are not considered to be personality traits if a difference or change in behavioural tendency can be entirely attributed to a change or difference in state. However, this does not mean that differences in state are irrelevant. Individuals who make different experiences in life may develop different personalities. This can be due to learning, physiological changes or alterations in brain anatomy due to interactions with the environment or critical experiences in life triggering the development of behavioural patterns in the long term. 19.1.2 What is new about animal personality research? Many of the questions that are currently studied using the labels ‘animal personality’ and ‘behavioural syndromes’ have a long research tradition and therefore do not per se describe a fundamentally new field of research. For instance, in the field of behavioural ecology one of the early landmark studies investigating correlations between behaviours, Huntingford (1976) found a positive relationship between aggressiveness towards conspecifics and boldness towards predators in sticklebacks (Gasterosteus aculeatus). Behavioural differences between individual mice have been studied in re-

Animal personality and behavioural syndromes

593

gard to the effects of changing environments (Benus et al. 1987). Evidence for heritability of behaviour and genetic correlations between behaviours has been reported in fish (Bakker and Sevenster 1989, Bakker 1994, Bell 2005) and rodents (Sluyter et al. 1995, Koolhaas et al. 1999). Psychologists have also been studying animal personality already for some decades (Gosling and John 1999). For instance, Stevenson-Hinde et al. (1980) studied personality variation between individual rhesus monkeys (Macaca mulatta) with the help of observer ratings. Finally, much research in regard to animal personality has been performed in studies on domestic animals (Hessing et al. 1993, Wechsler 1995). Although these studies did not directly address evolutionary questions, they provided important information about the underlying proximate bases, the development, and the heritability of behaviours. This is important because evolutionary theory is to a large extent concerned with genetic processes (but see West-Eberhard 2003), but empirical testing in behavioural ecology often involves data collected on behavioural phenotypes, thereby implicitly assuming that phenotypic patterns are good predictors of genetic patterns, i.e. the ‘phenotypic gambit’ (Grafen 1984). Although the relationship between phenotype and genotype seems to be tight in traits with high heritability, such as many morphological traits, there is some debate about the generality of this assumption for traits with moderate heritability (Boake et al. 2002). The last few years have seen an upsurge in research on animal personality in the field of behavioural ecology (for reviews see Wilson et al. 1994, Wilson 1998, Sih et al. 2004a,b, van Oers et al. 2005, Bell 2007, Penke et al. 2007, Réale et al. 2007, Biro and Stamps 2008, Sih and Bell 2008, Bell et al. 2009, Biro and Dingemanse 2009, Dingemanse et al. 2009a, Réale and Dingemanse in press, Schuett et al. 2010, Bergmüller and Taborsky in revision). But why is the study of animal personality now generating so much interest? First, the study of animal personality and behavioural syndromes has led to some interesting switches in perspective, focusing on the variation of behavioural traits between individuals and not only on the presumably optimal mean within a population. Additionally, there is increasing interest in studying the evolution of individual variation in plasticity, i.e. individual by environment interactions (Wilson 1998, Nussey et al. 2007). If personality variation has fitness consequences or acts as constraint, this has potential consequences for all traditional research areas in behavioural ecology, including predator-prey, host-parasite and mutualistic interactions among individuals of different species (e.g. symbioses), competition for resources, group living, sexual conflict, mating, breeding, parental care and cooperation.

594

Ralph Bergmüller

Second, research on animal personality links evolutionary and ecological questions of why animals behave the way they do tightly to developmental questions and to proximate aspects underlying behaviours. Therefore, asking whether individuals differ in behaviours fundamentally integrates proximate and ultimate aspects that are commonly studied more independently of each other (Sih et al. 2004a, Bell 2007). Third, important evolutionary implications can result from behavioural syndromes. Thus, if behavioural traits are correlated among each other due to a common regulating mechanism, selection on one trait could be important in shaping behaviours that are important in other contexts (Sih et al. 2004a). Fourth, individual differences can also be important in regard to experimental issues, e.g. if they generate a bias due to the selection of study individuals (Biro and Dingemanse 2009). Finally, research on animal personality may ultimately also help to better understand the evolutionary origins of variation in human personality.

19.2 Current evolutionary explanations for animal personality and behavioural syndromes As evolutionary research on animal personality and behavioural syndromes is a relatively young field of research, the conceptual basis to explain the existence of personality is still developing. Current explanations focus around ‘constraint hypotheses’ which posit that individuals are less flexible than would be optimal due to costs and limits to flexibility, and ‘adaptive hypotheses’ that seek to explain how consistency in behaviours and specific combinations of different behaviours may be explained adaptively (Sih et al. 2004a, Bell 2005). This distinction can be made for all three questions that are addressed in the context of animal personality (BOX 19.1). Although it seems important to attempt to answer each of the three questions independently, there will obviously be many overlaps between these issues, so that dealing with one question will often involve an indirect answer to one or both of the two other questions. 19.2.1 Why are behaviours consistent? 19.2.1.1 Consistency as a constraint – costs and limits of flexibility Individual behavioural consistency describes the phenomenon that an individual maintains its rank for a behavioural phenotype relative to other phe-

Animal personality and behavioural syndromes

595

notypes in the population, not necessarily that this individual always shows the same phenotypic value (Réale and Dingemanse in press). Hence, consistency is present when an individual does not express the whole range of phenotypic variation of a trait observed in its population for the environmental conditions measured. How can consistency be explained in evolutionary terms? Individuals may be consistent in their behaviours simply because there are costs involved in maintaining flexibility or because there are certain limits to flexibility (DeWitt et al. 1998). Maintenance costs involved in retaining the flexibility to react appropriately to each challenge could explain consistent behaviour. For instance, maintaining the physiological and sensory machinery to react aggressively in one situation, e.g. towards a conspecific competitor but non-aggressive towards a predator. A limit to flexibility prevails if it is not possible to decouple traits that are regulated by the same proximate mechanism, thereby generating a spill-over (see below). Another evolutionary cause for a limit to flexibility is that producing a plastic response would require reliable information about the environment. However, because the environment is often unpredictable, there are limits to maintaining an appropriate flexibility because plastic individuals would express a poor phenotype-environment matching and inflexible individuals with fixed rules of behaviour can be more successful (McElreath and Strimling 2006). 19.2.1.2 Consistency as an adaptive response – benefits of being consistent Behavioural consistency may confer benefits to individuals (a) because specialists can outcompete generalists (b) or because behavioural specialisation serves to reduce conflict among individuals. (c) Another potential benefit of consistency is the idea of commitment, i.e. signalling reliably that not all options available will be used. (d) Finally, consistency may be a means to communicate the quality of the signalling individual. Below these four mechanisms will be treated in some detail. However, some of the examples that will be described do not conform to animal personality but merely serve to illustrate evolutionary mechanisms that may result in consistency in behaviour. (a) Benefits of specialisation can arise when specialists are more efficient than generalists, i.e. because ‘the jack of all trades is a master of none’. For instance, the enormous success of social insects has, in part, been attributed to their specialisation in caste systems and division of labour (Wilson 1987, Korb this volume). In ants with morphological castes,

596

Ralph Bergmüller

certain types of individuals have been found to be more efficient in performing certain tasks than others (Wilson 1985, Beshers and Fewell 2001). Specialisation is also thought to increase efficiency in species without morphological specialisations and many studies simply assume that if there is specialisation, this corresponds to improved performance in the respective task (as has been shown, e.g. by Wolf et al. 2008). With the help of specialisation, individuals may avoid costs arising from switching between tasks, or they can become better in a task due to learning. For instance, in bumblebees novice foragers first extract nectar from a diversity of flower morphs and later specialise on a particular species of flower with a dramatic increase in efficiency (Heinrich 1979). The positive feedback resulting from getting skilled in a certain task can promote individual consistency and divergence. Surprisingly, the idea that specialists are more efficient than generalists has rarely been tested. A recent study that attempted to do so in the ant species Temnothorax albipennis did not find that specialist ants were more efficient than generalists. In three of four tasks, generalists were equally efficient as specialists, in the fourth task generalists even outcompeted specialists (Dornhaus 2008). (b) Specialisation may also decrease costs involved in interactions among individuals by reducing the level of conflict between conspecifics (Bergmüller and Taborsky in revision). For example, feral pigeons (Columba livia) show a higher specialisation in regard to food choice when competition for food is high (Giraldeau and Lefebvre 1985, Inman et al. 1987). In cooperatively breeding species, where sexually mature individuals help others to raise offspring that are not their own, some examples of behavioural specialisation resulting in division of labour have been reported (Arnold et al. 2005, Bergmüller and Taborsky 2007). In a Lake Tanganyika cichlid (Neolamprologus pulcher) helpers show some behavioural specialisation as sexually immature helpers. Individuals that were consistently more explorative over several months also spent more time in territory defence and less time in territory maintenance, compared to helpers showing little explorative behaviour (Bergmüller and Taborsky 2007; Fig. 19.3). These differences in individual trait combinations could be a means to reduce conflict among individuals and reflect specialisation of helpers on different life history pathways (Bergmüller and Taborsky 2007, in revision, Bergmüller et al. 2007b). ‘Stayers’ should be helpful in order to be tolerated in the territory (the ‘pay-to-stay hypothesis’: Gaston 1978, Bergmüller and Taborsky 2005, Bergmüller et al. 2005b), they should exhibit low levels of aggression, and they should show low levels of exploration because they do not need to explore the surroundings

a

female male

40

30

20

10

0

2

4

6

8

Exploration / 15 min

10

Digging / 10 min (dominant present)

Aggressive displays / 10 min

Animal personality and behavioural syndromes

b

597

female male

40

20

0 0

5

10

15

20

25

30

Aggressive displays / 10 min (dominants present)

35

Fig. 19.3 (a) Exploration behaviour (the number of pot halves visited in a novel environment) was positively correlated with the frequency of aggressive displays towards an intruder in a cichlid. This relationship was also significant for female helpers only. (b) The frequency of aggressive displays towards an intruder correlated negatively with the frequency of digging (but only when the dominant breeders were present). This relationship was also marginally significant for females only.

in order to find a vacant territory for independent breeding. In contrast, ‘dispersers’ should display the minimum amount possible of costly help in order to be able to invest all energy into fast growth, they should tend to be aggressive in order to acquire and defend an own territory, and they should be explorative in order to monitor the surroundings and find a vacant breeding spot (Bergmüller et al. 2005a, Bergmüller and Taborsky 2007). Several studies suggest that helpers in N. pulcher show conditional cooperative behaviour (Bergmüller et al. 2007a, Bshary and Bergmüller 2008) in that they pay in order to be allowed to stay in the dominants’ territory. Indeed, helpers may engage in ‘pre-emptive appeasement’ in that they use helping and submissive behaviour in order to pre-empt punishment by dominant individuals (Bergmüller and Taborsky 2005). Interestingly, helping and submissive behaviours have considerable energetic costs (Grantner and Taborsky 1998), and individuals who show more submissive behaviour tend to produce lower levels of circulating stress hormones (Bender et al. 2006). Taken together, these results suggest that individual variation in the propensity to engage in helping and other social behaviours is related to a range of potentially distinct specialised life histories (Bergmüller and Taborsky 2007, Schürch 2008). However, the situation is somewhat

598

Ralph Bergmüller

more complicated, as individual differences in behaviour tend to vary depending on the behaviour measured, but also depending on sex, life history stage, the social situation and interactions between these factors (Schürch 2008). Interestingly, behavioural specialisation may be the result of a behavioural feedback during ontogeny in that initially flexible helpers choose distinct social roles according to niche availability and competitive ability. As specialisation in coping with a particular social challenge often affects how individuals need to deal with other social or ecological challenges, this could lead to cascading effects and specialisations in several behaviours, i.e. adaptive behavioural syndromes (Bergmüller and Taborsky 2007). In case of canalisation of behaviours during ontogeny (age-dependent decline in flexibility in behaviour), distinct behavioural types differing in multiple behaviours may result, even if individuals are genetically identical (Bergmüller and Taborsky submitted). Another possibility how specialisation can reduce conflict has been proposed by Dall et al. (2004), who suggested that if some individuals are eavesdroppers who assess the contest behaviours of others, consistently playing hawk or dove (as opposed to being unpredictable) can reduce costly, escalated fights. The above examples also show that behavioural specialisation provides a potential explanation for consistency and individual differences at the same time. (c) Consistency can also come about due to commitment, which occurs when individuals produce reliable signals about limiting own options, which can be beneficial when the signal alters the behaviour of the receiver to the benefit of the signaller (Schelling 1960). This is also referred to as making credible threats or promises (McNamara and Houston 2002). Examples for commitments include emotions, reproductive restraint, the threat to attack or the promise to cooperate (Nesse 2001). Such commitment should be expected when deception (faking a commitment) is difficult to achieve or not in the long-term interest of an individual (Dall et al. 2005). (d) Finally, consistency in behaviour could be a result of sexual selection, for instance, if females prefer consistent males over inconsistent males or males do better in male-male competition when they are consistent (Schuett et al. 2010). Consistency could either be an indicator of mate quality or a cue of mate compatibility. In case a personality trait is an indicator of genetic quality, the trait should exhibit genetic variation and a positive correlation with fitness. In case the same personality trait is an indicator of genetic quality in both sexes, this should lead to positive assortative mating (see also Sect. 19.3).

Animal personality and behavioural syndromes

599

19.2.2. Why are individuals different in their behaviour? Individual differences in behaviour become apparent when individuals with diverging behaviours are measured repeatedly. Additionally, consistent differences between individuals can also result in different degrees of consistency (Réale and Dingemanse in press). For example, aggressive rodents are more likely to form routines while non-aggressive individuals are more responsive to environmental stimuli (Koolhaas et al. 1999). Two main evolutionary processes are currently discussed to explain the evolution of individual differences in animal personality. The most likely evolutionary mechanism to maintain the existence of different types of personality in a population is balancing selection and the most important types are considered to be frequency-dependent selection (Maynard Smith 1982, Dall et al. 2004) and spatio-temporal environmental heterogeneity (Roff 1997). Other potentially important mechanisms include densitydependent selection (Wilson et al. 1994) and disruptive selection (Wilson 1998). An explanation for individual differences based on phenotypic plastic responses is a positive feedback mechanism (e.g. learning) that can result in different phenotypes in the long run (see above and Sect. 19.4). Since the advent of evolutionary game theory (Maynard Smith 1982), it is well established that the fitness of a particular behavioural strategy depends on the frequencies of competing strategies. In case the rare strategy has an advantage, the resulting negative frequency-dependent selection can lead to stable coexistence of different behavioural types due to balancing selection. Consider for instance a producer-scrounger system in which individuals in foraging groups can make use of one of two different foraging options, i.e. either search actively for food (producers) or exploit the food discovered by others (scroungers). It turns out that scroungers do better in a group with many producers, but producers have an advantage over scroungers when producers become rare (Giraldeau and Dubois 2008). Hence, frequency-dependent selection (Dall et al. 2004) or frequencydependent learning (Giraldeau 1984, West-Eberhard 2003) may result in differences in personality. Another example is provided by a recent model, in which individuals that are responsive to environmental stimuli and unresponsive types are maintained in a population due to negative frequencydependence in combination with a positive feedback of being responsive (Wolf et al. 2008). However, although negative frequency-dependence can maintain the co-existence of different behaviours, this does not necessarily mean that different individual types with a largely consistent personality perform these behaviours. A particular problem is that the same distribution of behaviours can also result when flexible individuals adjust their be-

600

Ralph Bergmüller

haviour depending on the behaviours performed by others (Sih and Bell 2008). When the fitness of a trait varies across time or space, significant genetic variation can be maintained in a population because of balancing selection resulting from spatial or environmental heterogeneity (Roff 1997). While under certain environmental conditions one personality type has a fitness advantage over other types, under different conditions in time and space another type can be favoured (Réale and Festa-Bianchet 2003, Dingemanse et al. 2004). Cote et al. (2008) studied the level of sociability in common lizards (Lacerta vivipara) to test this hypothesis. To determine sociability, juvenile lizards were tested one day after birth for the level of sociability using the proportion of time individuals spent in a shelter together with a paper with conspecific odour. The level of sociability affected fitness differently, depending on population density. Lizards who tended to spend less time in a shelter with conspecific odour survived better in low-density populations, whereas more social lizards tended to survive better in high-density populations. These results support the hypothesis that spatio-temporal variation in environmental conditions is involved in maintaining the coexistence of different personality types. Studies on animal personality usually find that behavioural types are more or less continuously distributed. However, as yet there is little theory to explain the existence of continuous distributions of behavioural types (but see Bergmüller and Taborsky in revision). In contrast, when the distribution of behavioural types is bi- or multi-modal, this indicates disruptive selection, which means that two or more divergent phenotypes are favoured by selection. In such a case, animal personality research converges with research on alternative behavioural strategies (Gross 1996, Taborsky et al. 2008), which have extensively been studied in the context of reproduction, Taborsky and Brockman this volume). Disruptive selection in combination with assortative mating of different types has also been suggested as a potential mechanism of speciation (Wilson 1998). 19.2.3 Why are different behaviours correlated? A proximate mechanism for the existence of individual differences in suites of behaviour is pleiotropy. Evolutionary processes to generate such behavioural trait packages include antagonistic selection, life history tradeoffs and correlational selection. A correlation between two behavioural traits can result from common regulating processes, i.e. common genes or a common physiological basis

Animal personality and behavioural syndromes

601

regulate different behaviours. Such common regulatory mechanisms can act as constraint on the independent evolution of behaviour (Sih et al. 2004a) and may result in a ‘spill over’ of behaviour in one situation to another. At the genetic level, pleiotropy and genetic linkage are two important sources for genetic correlations (Roff 1997, Lynch and Walsh 1998). Such genetic correlations can lead to an evolutionary trade-off, i.e. one trait results in fitness benefits while the other trait results in fitness costs, which is also called antagonistic pleiotropy (Sih and Bell 2008). For instance, if boldness and aggression are positively correlated, then aggressive individuals might do well in some situations (e.g., when competing for resources), but the same individuals would perform poorly when low levels of boldness are advantageous (e.g. under high predation risk; Sih et al. 2004a). An intriguing example for a potential case of pleiotropy is a phenotypic correlation between behaviours in different contexts that has been described in fishing spiders (Dolomedes fimbriatus; Arnqvist and Henriksson 1997). As in some other species of spiders, female fishing spiders show sexual cannibalism (Schneider and Fromhage this volume). In this particular species, it has been shown that sexual cannibalism has costs for females because some individuals appear to cannibalise on males before they even mate. How can the existence of such an apparently maladaptive behaviour be explained? The authors propose that a behaviour that is beneficial in the feeding context may ‘spill over’ to the mating context, thereby leading in the extreme to pre-copulatory sexual cannibalism. Female body size, measured as cephalothorax size (which reflects juvenile growth), was positively related to fecundity. One possible scenario is that strong selection favouring aggressive hunting in juvenile females results in an inappropriately high tendency for those females to attack males when they mature to adulthood. In the funnel-web spider Agelenopsis aperta, several studies have shown that aggression towards prey and conspecifics is positively correlated among individuals (Riechert and Hedrick 1993). There is some evidence that such correlations may have a genetic basis, which would mean that extreme females may be caught in an ‘evolutionary trap’ resulting from strong selection on juveniles. However, it has been proposed that under such conditions modifier genes could evolve that disrupt the pleitropic effect (Sih et al. 2004b). Another evolutionary process that could maintain behavioural correlations between different behaviours in a population is a combination of antagonistic selection and phenotypic plasticity. In sticklebacks, several studies found a positive correlation between boldness and aggressiveness. Interestingly, this correlation was only observed in populations in which predators were present (Bell 2005, Dingemanse et al. 2007). The study of

602

Ralph Bergmüller

a

b 35%

50%

300

Boldness

Boldness

300

200

100

0

45% 0

10

100

0

62%

Aggressiveness

200

20

0

10

20

Aggressiveness

Fig. 19.4 Effect of predation on stickleback behaviour. (a) Prior to exposure to a real predator, aggressiveness was not correlated with boldness (all dots), also not for survivors of the predation treatment alone (survivors are depicted by open circles and predator victims are depicted by close dots). (b) After the predation event surviving individuals showed a positive correlation between both behaviours. (With permission from Bell and Sih 2007)

Bell (2005) found that correlations among the behaviours differed between different populations of sticklebacks, and Dingemanse et al. (2007) demonstrated that the presence or absence of a correlation between boldness and aggressiveness within a population depends indeed predictably on the presence or absence of predators. To test whether predation induces a positive correlation between the behaviours, Bell and Sih (2007) subjected fish from a population without strong experience of predation to a predator for 24 h and measured the behaviours and the persistence of the individuallymarked fish before and after this treatment. While there was no correlation between aggressiveness and boldness before the treatment, the fish that survived the predation treatment showed a positive correlation between both behaviours (Fig. 19.4a, b). The results cannot be explained by correlational selection (see below) because the individuals that survived did not show a correlation between boldness and aggression before they had been exposed to the predator (Fig. 19.3a). Interestingly, both selection and phenotypic plasticity appeared to contribute to the results. As would be expected, bold individuals were more likely to be consumed by the predator. Additionally, more aggressive individuals were more likely to survive exposure to predation. This suggests that antagonistic selection favouring shy and aggressive individuals was involved in generating two positively correlated traits. Additionally, predation induced a phenotypic plastic response in that individuals became generally less aggressive after the predator treatment and also changed

Animal personality and behavioural syndromes

603

their rank order in aggressiveness. This was not the case in regard to boldness of individuals, which remained stable within individuals before and after the treatment. Why certain behaviours are consistent and correlated among each other may also be explained with a life-history trade-off, i.e., the problem of how and when to allocate resources to competing functions. A typical trade-off occurs between growth and survival when individuals forage under predation risk. Individuals that spend much time foraging grow fast but at the risk of dying early. If individuals find different solutions to deal with this trade-off, this may select for different personality types. Life-history trade-offs may be involved in the frequently found correlation between aggressiveness and boldness. In species with indeterminate growth such as fish, the trade-off between growth and survival has been suggested to maintain individual variation in growth rate (Mangel and Stamps 2001). Interestingly, individual differences in growth have been reported to occur, even if individuals are kept under the same conditions, including a similar food regime. The explanation for why these individual differences in growth rate co-exist is that, due to the trade-off between growth and survival, slow and fast growing individuals have overall comparable fitness. It has been proposed that such variation in growth rate may in turn result in variation in personality if high grow rates favour bold and aggressive behaviour and low growth rates favour shy and non-aggressive behaviour (Stamps 2007, Biro and Stamps 2008). A similar relationship between personality and life history was proposed in a model that explored the effects of a trade-off between current and future reproduction (Wolf et al. 2007b). Assuming that individuals with low expected future reproduction are more likely to take risks, such individuals will be bold and aggressive and therefore grow fast and reproduce early. Shy and non-aggressive individuals should result from the opposite life history characteristics. As long as the differences between individuals in future expectations (also called assets) remain stable, the model can explain consistent individual differences in behaviours and correlations among them. However, as has been noted by McElreath et al. (2007), both strategies should tend to converge due to a negative feedback. If individuals with low assets can increase their future expectations by being bold and aggressive (and survive), and shy non-aggressive individuals tend to lose assets, individuals should become more similar over time. Hence, future studies will need to clarify whether differences in future reproductive expectations are sufficient to explain stable individual differences, or whether other potentially important factors such as current payoffs associated with bold and aggressive behaviours need to be invoked (Wolf et al. 2007a).

604

Ralph Bergmüller

Behavioural syndromes can be adaptive if correlational selection favours certain combinations of traits (Lande and Arnold 1983). For example, correlational selection has been shown to favour certain combinations of colour patterns and escape behaviours in garter snakes (Brodie 1992). Until now, only few studies have shown that correlational selection can maintain correlations between personality traits. A study on 1101 Australian postmenopausal women combined the results of a personality survey with the number of children the woman gave birth to throughout their life (Eaves et al. 1990). Interestingly, lifetime reproductive success was highest in women that were highly extravert and low neurotic or low extravert and highly neurotic. In contrast, women with intermediate score combinations had intermediate fitness and high-high and low-low combinations had lowest fitness values. Neither of both personality dimensions had fitness effects on its own. Hence, this study suggests that the combinations of traits are indeed subject to selection (Sinervo and Svensson 2002).

19.3 Heritability and fitness consequences of personality Is personality subject to evolutionary change? For this to be the case, two conditions must be met. First, at least part of the variation in personality must be heritable, and secondly, the traits must affect fitness (Endler 1986). The first condition may at first seem surprising, because according to Fisher’s (1930) fundamental theorem, natural selection will reduce the heritable variance of a trait until one genetic variant becomes fixed (i.e. the heritable variation is zero), provided the direction of selection is constant. However, most morphological traits that have been investigated so far also show non-zero heritable variation (see BOX 19.2), and this seems also to be true for most behaviours (Sinervo and Svensson 2002, van Oers et al. 2005, Réale et al. 2007). One reason for the persistence of heritable genetic variance is that this is influenced by two factors with opposing effects: mutation rate (which increases genetic variance) and the strength and type of selection (i.e. directional or stabilizing selection), which in general decreases genetic variance. If mutation and selection occur at the same rate, there is a mutationselection balance, but this does not seem to be a likely source of personality variation (Penke et al. 2007). Instead, there are several evolutionary mechanisms (frequency-dependent selection, spatio-environmental heterogeneity) that can stabilise the co-existence of different types in a population, and, hence, also provide a source of heritable genetic variation.

Animal personality and behavioural syndromes

605

Single genes can have major effects on behaviour (McGuffin et al. 2001, Fidler et al. 2007). For instance, it has recently been shown that a polymorphism at the Drd4 gene is associated with explorative behaviour in great tits (Fidler et al. 2007). Interestingly, variation at the Drd4 gene is also associated with personality variation in horses, humans and other primates. However, behaviours are usually quantitative traits, i.e. traits that are distributed continuously because they are affected by the expression of many genes. Such genes are called quantitative trait loci (QTLs) because they result in continuous distributions of phenotypes. It is important to note that the effects of genes on behaviour are probabilistic and not deterministic, due to gene x environment interactions (Dingemanse et al. 2009b). Thus, even when the underlying genes are known, it is still a long way to understanding their effects, because existing genes are not always expressed. Since variation in behaviour is at least partly heritable, it can be subject to selection. But how can the evolution of personality, which is usually shaped by many genes, be studied? Here, a statistical technique called quantitative genetics can be employed. Quantitative genetic models use the resemblance among relatives due to shared genotypes to study the structure of inheritance by estimating the proportion of phenotypic variance that is contributed by genetic effects, the environment and gene by environment interactions (Falconer and Mackay 1996, Lynch and Walsh 1998; see BOX 19.2). 19.3.1 Heritability of personality and behavioural syndromes A number of studies have provided evidence for heritability and genetic correlations between behaviours (van Oers et al. 2005, Réale et al. 2007). Overall, they suggest that individual differences in behaviour are moderately heritable (heritability estimates of around 20-40%, range 0-66%), and they can be relatively stable over an individual’s lifetime (Boissy 1995, Koolhaas et al. 1999, Bouchard and Loehlin 2001, van Oers et al. 2005), but this is not always the case (Bell and Stamps 2004, Schürch 2008, Sinn et al. 2008). Great tits (Parus major) have been studied in much detail in regard to heritability and fitness consequences of personality traits in the lab and field. A parent-offspring regression of data on wild-caught birds and their hand-reared offspring showed that 30% of the variation in explorative behaviour in offspring could be attributed to the behaviour of their parents (Drent et al. 2003). Similar findings resulted from a field study in which individuals with known family relationships were investigated (Dingemanse et al. 2002). A two-way selection experiment over four generations

606

Ralph Bergmüller

BOX 19.2 Quantitative genetics The underlying assumption of quantitative genetic analyses is that the variation in the phenotype P can be partitioned into different components, i.e. the sum of genetic (G) and environmental (E) components, and hence: P=G+E Considering the variances and the interaction between genes and the environment this becomes: VP = VG + VE + VGxE + e where VP = phenotypic variance VG = genotypic variance VE = environmental variance VGxE = variance due to gene x environment interactions e = error variance In order to understand whether evolution can act on a behavioural trait, it is useful to estimate heritability. The broad sense heritability is the proportion of genotypic variance in a population that contributes to the observed phenotypic variation (H2 = VG/VP). However, because only additive genetic variance (VA, the variance in breeding values of a trait, i.e. the trait variance transmitted from parents to their offspring) will lead to a response to selection, the genetic variance (VG) needs to be further partitioned into additive variance (VA) and non-additive components (VN, interactions among genes due to dominance and epistasis) (VG = VA + VN). Only VA (the allelic variation of characters transmitted from parents to their offspring (calculated as a parent offspring regression)) is available for a response to selection. This can be used to calculate the narrow sense heritability (h2 = VA/VP), which is a useful measure to estimate the (short term) response of a trait to selection (long term responses will also be affected by non-additive genetic variance). Correlated characters (or a phenotypic correlation) can result from environmental correlations or genetic correlations. An environmental correlation prevails when correlations between two traits are due to shared aspects of the environment, e.g. individuals live in the same group or habitat. A genetic correlation is the proportion of phenotypic variance that two traits share due to genetic causes. Quantitative genetics can distinguish between additive genetic correlations and the residual correlation (a mix of environmental correlation, non-additive correlation and correlated errors that is difficult to disentangle). Any genetic correlation suggests that the same genes contribute to both traits measured (pleiotropy). A positive genetic correlation indicates a concordant effect of the variants of the same genes on two or more traits. A negative genetic correlation indicates a discordant effect of gene variants, and hence a trade-off between two traits.

Animal personality and behavioural syndromes

607

Another potential cause of a genetic correlation is genetic linkage, i.e. several different alleles or loci are inherited jointly because they are located near each other on a chromosome and are therefore relatively unlikely to be disconnected through recombination during meiosis. This is sometimes regarded as less important in evolutionary terms because selection can break up such associations easily. However, as long as genetic linkage produces beneficial or neutral trait combinations, it is very likely to persist. A genetic correlation is estimated in the same way as heritability, but using the covariance between two characters. A correlated response to selection can be determined by artificial selection on a particular behaviour and measuring the response to selection on other behaviours. Genetic correlations may either act as constraints or they may be adaptive because they result from selection on optimal trait combinations. Increasing evidence suggests that individual animals differ in both (a) their average level of behaviour displayed across a range of contexts, but (b) also in their responsiveness to environmental variation (plasticity). As a consequence, the degree and type of behavioural plasticity (in other words, the intercepts and slopes of individual variation in reaction norms) are also subject to evolutionary change (Schlichting and Pigliucci 1998, West-Eberhard 2003, Dingemanse et al. 2009a). This has recently been considered within a single framework, which is based on the concept of ‘behavioural reaction norms’ using quantitative genetics approach to behavioural plasticity (Dingemanse et al. 2009a).

for ‘fast’ and ‘slow’ early explorative behaviour showed a realised heritability of 54% for early exploratory behaviour (Drent et al. 2003). Another selection experiment found a realised heritability of risk-taking behaviour of 19% (van Oers et al. 2004c). Interestingly, exploratory behaviour and risk taking behaviour were positively correlated: birds selected for fast or slow exploratory behaviour also tended to differ in regard to risk taking (van Oers et al. 2004a). To completely understand the inheritance of personality, it is necessary to understand the underlying genetic structure. This was investigated in great tits using original lines, F1 crosses and back-crosses of selection lines for fast and slow early explorative behaviour, which enabled the break-up of some of the components of variation (van Oers et al. 2004b). The results showed that low exploration is dominant over high exploration and that low boldness is dominant over high boldness. This result highlights the fact that apart from additive genetic variance, genetic dominance can also be important in the structure of inheritance of personality, and therefore may affect the evolution of personality in the long-term.

608

Ralph Bergmüller

19.3.2. Fitness consequences of personality Only few studies have investigated the functional aspect of personality, i.e. whether it is related to proxies of fitness (Dingemanse and Réale 2005). In humans, some studies have attempted to establish a potential link with fitness proxies, such as susceptibility to stress or the level of stress hormones (Grossarth-Maticek et al. 1988, Schmitz 1992) and health (Martin 1998). There are only few studies that investigated the fitness consequences of animal personality in the wild. To illuminate the possible fitness consequences of personality, wild great tits were captured, tested individually for exploratory behaviour in the lab and released back at the place of capture (Dingemanse et al. 2002, 2003). Early exploratory behaviour in a novel environment turned out to be repeatable and heritable but unrelated to age, body condition and sex. Additionally, explorative behaviour predicted survival of adult birds under natural conditions in complex ways (Fig. 19.5). In years with relaxed competition for food in winter, fast-exploring adult males and slow-exploring adult females had highest survival. Sex-specific survival seemed to be related to different types of competition. Years with beech masting resulted in relaxed competition for food between females in winter. However, as more birds survived in these years, recruitment rates were high, resulting in intensified intra-sexual competition for territories among males. Hence, the pattern of survival of adult birds was reversed in years with increased competition for food in winter (no beech masting) but subsequent low competition for territorial space. In these years, slow exploring males and fast exploring females had higher survival. Although, sex-specific competition seems to explain why fast-exploring adult birds survive better in years of high competition, it is unclear why they do worse than slowexploring birds in years with low competition. Other patterns of selection resulted from the study of fitness measures, such as offspring production and recruitment. Slow-exploring females produced larger offspring than fast-exploring females. Pairs with assortative personality (i.e. both partners were either fast or slow explorers) produced offspring with a higher body mass (and therefore higher competitive ability) compared to other pair combinations (Both et al. 2005). Interestingly, pairs with assortative personality also were more likely to have chicks resulting from extra-pair paternity (van Oers et al. 2008), indicating that behavioural differences between individuals are also important with regard to reproductive partner preferences and hence sexual selection (see Kempenaers and Schlicht this volume, Schuett et al. 2010). When considering recruitment (the number of offspring that survived and bred in the study area) as a measure of selection, assortative pairs were more success-

Animal personality and behavioural syndromes Winter

Adult survival Males

609

Offspring recruitment

Females

Males

Females

Poor

Rich Slow

Fast Slow

Fast Slow

Fast Slow

Fast

Exploration score

Fig. 19.5 Overview over the fitness consequences of exploratory behaviour in novel environments in great tits. The arrows indicate the type of selection ( directional selection favouring fast explorers;  directional selection favouring slow explorers;  stabilising selection;  disruptive selection (all fitness consequences concern adult annual survival and offspring recruitment). Hatched bars indicate non-significant trends. (With permission from Dingemanse and Reale 2005)

ful in years with beech masting (i.e. resulting in disruptive selection for fast- and slow-exploring types), but pairs of birds exploring with medium speed were more successful in years without beech mast, resulting in stabilising selection for intermediate types (Fig. 19.5). In conclusion, early exploratory behaviour in great tits is heritable (Drent et al. 2003) and affects several components of fitness which could result in evolutionary change. However, the direction of selection is not constant. As masting of beeches occurs only about every 3 years the resulting temporal fluctuations appear to generate a pattern of selection that is overall balancing (Dingemanse et al. 2004), and may generate the persistence of different personality types. Additional mechanisms such as migration in a spatially heterogeneous environment, frequency-dependent selection and sexual selection may additionally be involved in maintaining and shaping the different personality types (Dingemanse and Réale 2005, van Oers et al. 2008).

19.4 Personality development During development, genotypes are transformed into phenotypes and some of the variance in personality among individuals results from the interactions of genetic factors and the environment individuals encounter during different stages of life, i.e. by gene-environment interactions in combina-

610

Ralph Bergmüller

tion with changing levels of phenotypic plasticity throughout their lifetime (West-Eberhard 2003). Environmental influences can shape the phenotype due to phenotypic plasticity, e.g. due to learning or other changes of the phenotype based on experience, and subsequent canalisation (the reduction of phenotypic variation) of behaviours during ontogeny can result in consistent types (West-Eberhard 2003). It is well established that early experience determines personality development in humans, which has been termed ‘early experiential calibration’ (Belsky et al. 1991). It seems likely that this is also the case in many other species. Indeed, several studies suggest that early environmental conditions can have profound effects on behavioural traits expressed later in life (Carere et al. 2005). For example, guppies that had been reared at low densities had a significantly higher shoaling tendency than guppies reared at higher densities (Chapman et al. 2008). This suggests that differences in early experience affected the development of sociability. Other early developmental factors that can be important in shaping personality are food availability, which affects sibling competition in birds (Carere et al. 2005, Groothuis and Carere 2005), number and sex of litter mates (Mendl and Paul 1991, Benus and Henkelmann 1998), and the quality of interactions between mothers and infants (Stevenson-Hinde et al. 1980, Meany 2001). In order to understand how variation in personality develops during ontogeny, behaviours need to be studied longitudinally. Do behaviours and their correlations remain stable throughout life? If not, which social or environmental conditions or developmental switch points (e.g. sexual maturation) contribute to shifts in behaviour? Such questions can be studied well in animals with a short lifespan. In dumpling squid (Euprymna tasmanica), which only live about 5 months (Sinn et al. 2008), feeding behaviour under risk and the response in a threat test were measured at different stages throughout lifetime. Both behaviours were not correlated with each other at any age. However, individuals with different phenotypes showed different amounts of developmental variation during sexual maturation. Individuals that were initially shy in the threat test later remained shyer in the same situation, whereas initially more bold individuals were more plastic during later phases of development. Interestingly, in the feeding context, the pattern was reversed, bolder individuals tended to remain bold, while more shy individuals tended to become even shyer. This shows that two apparently closely related behavioural reactions to stress can be independent in their expression throughout life. Additionally, within the same behavioural reaction, initial phenotypic differences between individuals can result in differences in plasticity of the behavioural phenotype during development.

Animal personality and behavioural syndromes

611

A study on the developmental stability of different behaviours in sticklebacks compared aggressiveness and boldness in individuals from populations with either high or low predation risk (Bell and Stamps 2004). Neither of the behaviours was completely stable throughout ontogeny. Individuals, who were very aggressive as juveniles were not necessarily the most aggressive ones when subadult or adult. In fish from the population with low predation risk, boldness and aggression were positively correlated only in juveniles, but the syndrome disappeared during later stages of development. Hence, initial correlations between behaviours can be uncoupled during ontogeny. However, in fish that originated from the population with more predators, there was a positive correlation between aggression and boldness throughout all three stages of life, i.e. as juveniles, subadults and adults, despite of the fact that individuals were not stable in both behaviours. This study shows that a behavioural syndrome does not necessarily require that individuals remain themselves stable in each behaviour during ontogeny. Instead, a behavioural syndrome can prevail when some individuals that are initially bold and aggressive later become shy and non-aggressive and vice versa.

19.5 Proximate bases of animal personality and behavioural syndromes Proximately, individual differences in behaviour are due to changes in gene-expression that in turn affect the production and release of several types of messengers, such as hormones and neurotransmitters. Different types of messenger molecules affect different processes within organisms that can directly or indirectly modify behaviour. Additionally, the different processes also influence each other and are themselves modulated by behaviour or the social situation (Oliveira 2004). As hormones regularly act on multiple target tissues, it seems likely that they will be critically involved in generating pleitropic effects (Ketterson and Nolan 2000) that may result in behavioural syndromes (Sih et al. 2004a). Below, I will briefly outline how studying proximate aspects of behaviour can help to understand the underlying mechanisms behind animal personality, using some examples of relationships between hormones and behaviour as an example. Despite of the complexity of hormone-behaviour interactions, it can be useful to broadly distinguish between organisational and activating effects (Phoenix et al. 1959). Organisational effects shape morphological and physiological aspects such as the brain neuro-anatomy and therefore act

612

Ralph Bergmüller

primarily early in development, thereby producing relatively fixed underlying structures of behaviour (Groothuis and Carere 2005). For example, differences in the concentration of steroid hormones in the egg yolk of birds correlated with later competitive behaviour between chicks in the nest (Schwabl 1993), which is likely due to organisational effects in the chick embryo. In contrast, activating effects are involved in short-term changes of behaviour, such as seasonal changes in reproductive status or adjustments to current stressors or the social situation. Activating effects of steroids have been suggested to fine-tune aggressive behaviour to the social challenges in the early environment in early territorial behaviour in black-headed gull chicks (Ros 1999). At this early phase during development, activating and organisational effects of hormones on behaviour may closely interact (Ros et al. 2001). When applying the distinction to animal personality and behavioural syndromes, organisational effects should be involved in the development of stable behavioural patterns and also produce correlations between behaviours that prevail through significant periods of time. In contrast, activating effects may be responsible for correlated responses in behaviour on shorter time scales, i.e. behavioural ‘spillovers’ (Sih et al. 2004a). Traditionally, the relationship between hormone concentrations and the expression of traits have been studied. For instance, house mice (Mus musculus), that were selected to respond differently towards stress (coping styles), also showed different neuro-endocrine profiles (Koolhaas et al. 1999). In rats, freezing behaviour showed a positive relationship to the level of corticosterones. Generally, a number of differences in neuroendocrine responses distinguish pro-active and reactive individuals in rodents. Thus, the activity of the hypothalamic-pituitary-adrenal (HPA) axis was found to be lower in pro-active individuals, whereas their testosterone activity was found to be higher (Koolhaas et al. 1999). A number of studies have measured the neuro-endocrine basis of responses to stress in other mammalian species (Boissy 1995) and birds (Cockrem 2007), indicating that the relationship between hormones and behaviour is much more complex than would be expected by a simple relationship between circulating hormone levels and behaviours (Wingfield et al. 1990, Sapolsky et al. 2000, Hirschenhauser and Oliveira 2006, Goymann and Hofer this volume). A good example of this complexity is provided by androgens, which are crucially involved in male aggressive behaviour. The usual notion is that high levels of androgens cause high levels of aggression. But this is not the case. For example, cichlid fish (Oreochromis mossambicus) that engaged in escalated fights with their mirror image did not exhibit increased levels of androgens. These fish could not assess the outcome of a fight, but an-

Animal personality and behavioural syndromes

613

drogens are only released when this serves to maintain social status after the situation has been resolved (Oliveira et al. 2005). Thus, circulating levels of androgens are the result of behaviour, not its cause. Hence, behavioural action is concerted in the brain and steroid hormones often only modulate or facilitate certain behaviours within the context of individual history based on environmental stimuli. Testosterone only mediates behaviour within a cascade of processes in which it is bound to a receptor complex, activates the expression of genes, modulates second messengers such as neuropeptides, which in turn have various effects on behaviour. This complexity highlights the fact that variation in behaviour can have multiple sources, each of which is involved in interactions and feedbacks, including gene activation or interactions with other messenger substances. Additionally, the same hormones can have different effects depending on the type of tissue or receptor they bind to (Sapolsky et al. 2000). In conclusion, while we are only beginning to understand the complex underlying principles, proximate mechanisms are an important and critical key to an understanding why behaviours often show limited plasticity, why individuals differ in behaviours and why they are often correlated among each other.

19.6 Future directions and perspectives Research on animal personality is likely to have important implications for future research on animal behaviour. Important aspects include (a) the question how heritable variation and phenotypic plasticity are jointly involved in shaping personality, (b) the consequences of potential bias resulting from individual differences in behaviour, and (c) a more integrative view of animal behaviour that takes into account the potential effects of intrinsic individual differences. (a) For simplicity, the description of many evolutionary mechanisms above assumed a simple genotype-phenotype relationship. However, when measuring behaviours of an individual, this usually involves taking only a snapshot in the landscape of possible life histories of individuals with the same genetic background (West-Eberhard 2003). Mary Jane West-Eberhard captured the resulting complexity arising in an amazingly simple statement: ‘Evolution of the phenotype is synonymous to evolution of development’ (p. 89). Research on animal personality offers the possibility to explain individual differences in behaviour in terms of gene x environment and phenotype x environment

614

Ralph Bergmüller

interactions, thereby framing the phenomenon of animal personality in terms of life history strategies that depend on what others do. (b) Intrinsic individual differences in behaviour imply that it is important to take them into account when conducting behavioural or ecological research and to avoid sampling bias (Biro and Dingemanse 2009). For instance, if more bold individuals are more easily caught (Réale and Festa-Bianchet 2003, Wilson et al. 1993) using only this subset of a population could critically bias the results. Personality differences can also affect the results obtained from physiological studies, such as measurements of metabolic rate (Careau et al. 2008). If individuals vary in their stress response and reactive individuals ‘freeze’, this could behaviourally be classified as ‘resting’, although these individuals are highly stressed and therefore show an elevated metabolic rate. (c) Research on animal personality and behavioural syndromes suggests that animal behaviour needs to account for the potential ‘package nature’ of behaviours. This should lead to a more integrated approach to animal behaviour which acknowledges the possibility that behaviours such as foraging, mating, predator defence, parental care, dispersal or cooperation may not always be free to evolve independently from each other (Sih 2004a). Additionally, research on individual differences provides the possibility to observe more cryptic behavioural strategies that, for instance, may result in behavioural task sharing in social groups (Bergmüller and Taborsky 2007). Similarly, in the context of sexual selection, different personalities may help to explain why not all females choose the apparently best male (van Oers et al. 2008, Schuett et al. 2010). Currently, we are only beginning to understand how personality and behavioural syndromes may affect the evolution of traits and behaviours previously studied in isolation from each other (Bell 2007). Acknowledging the existence of individual differences, resulting either from constraints or due to adaptive reasons, may pave the way to a more fine-scaled picture for why animals behave the way they do.

Acknowledgements I thank Niels Dingemanse, Katharina Förster, Peter Kappeler, Kees van Oers, Katharina Peer, Albert Ros and one anonymous referee for valuable comments that greatly improved the book chapter and Redouan Bshary, Ian Hamilton, Dik Heg, Roger Schürch and Michael Taborsky for discussion. I was supported by a grant of the Suisse National Science Foundation (to R. Bshary).

Animal personality and behavioural syndromes

615

References Arnold KE, Owens IPF, Goldizen AW (2005) Division of labour within cooperatively breeding groups. Behaviour 142:1577-1590 Arnqvist G, Henriksson S (1997) Sexual cannibalism in the fishing spider and a model for the evolution of sexual cannibalism based on genetic constraints. Evol Ecol 11:255-273 Bakker TCM (1994) Genetic correlations and the control of behaviour, exemplified by agressivness in sticklebacks. Adv Stud Behav 23:135-171 Bakker TCM, Sevenster P (1989) Changes in the sexual tendency accompanying selection for aggressiveness in the three-spined stickleback, Gasterosteus aculeatus L. J Fish Biol 34:233-243 Bell AM (2005) Behavioural differences between individuals and two populations of stickleback (Gasterosteus aculeatus). J Evol Biol 18:464-473 Bell AM (2007) Future directions in behavioural syndromes research. Proc R Soc Lond B 274:755-761 Bell AM, Sih A (2007) Exposure to predation generates personality in three spined sticklebacks (Gasterosteus aculeatus). Ecol Lett 10:828-834 Bell AM, Stamps JA (2004) Development of behavioural differences between individuals and populations of sticklebacks, Gasterosteus aculeatus. Anim Behav 68:1339-1348 Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of behaviour: a meta-analysis. Anim Behav 77:771-783 Belsky J, Steinberg L, Draper P (1991) Childhood experience, interpersonal development, and reproductive strategy – an evolutionary-theory of socialization. Child Dev 62:647-670 Bender N, Heg D, Hamilton IM, Bachar Z, Taborsky M, Oliveira RF (2006) The relationship between social status, behaviour, growth and steroids in male helpers and breeders of a cooperatively breeding cichlid. Horm Behav 50:173182 Benus RF, Henkelmann C (1998) Litter composition influences the development of aggression and behavioural strategy in male Mus domesticus. Behaviour 135:1229-1249 Benus RF, Koolhaas JM, van Oortmerssen GA (1987) Individual differences in behavioral reaction to a changing environment in mice and rats. Behaviour 100:105-122 Bergmüller R, Taborsky M (2005) Experimental manipulation of helping in a cooperative breeder: helpers ‘pay-to-stay’ by pre-emptive appeasement. Anim Behav 69:19-28 Bergmüller R, Taborsky M (2007) Adaptive behavioural syndromes due to strategic niche specialization. BMC Ecol 7:12, doi:10.1186/1472-6785-7-12 Bergmüller R, Taborsky M (in revision) Social niche specialisation – an evolutionary framework to explain consistent individual variation in behaviour. Trends Ecol Evol

616

Ralph Bergmüller

Bergmüller R, Taborsky M (submitted) Animal personality resulting from social niche specialisation. Bergmüller 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 Bergmüller R, Heg D, Taborsky M (2005b) Helpers in a cooperatively breeding cichlid stay and pay or disperse and breed, depending on ecological constraints. Proc R Soc Lond B 272:325-331 Bergmüller R, Johnstone RA, Russell AF, Bshary R (2007a) Integrating cooperative breeding into theoretical concepts of cooperation. Behav Proc 76:61-72 Bergmüller R, Russell AF, Johnstone RA, Bshary R (2007b) On the further integration of cooperative breeding and cooperation theory. Behav Proc 76:170-181 Beshers SN, Fewell JH (2001) Models of division of labor in social insects. Annu Rev Entomol 46:413-440 Biro PA, Dingemanse NJ (2009) Sampling bias resulting from animal personality. Trends Ecol Evol 24:66-67 Biro PA, Stamps JA (2008) Are animal personality traits linked to life-history productivity? Trends Ecol Evol 23:361-368 Boake CRB, Arnold SJ, Breden F, Meffert LM, Ritchie MG, Taylor BJ, Wolf JB, Moore AJ (2002) Genetic tools for studying adaptation and the evolution of behavior. Am Nat 160:S143-S159 Boissy A (1995) Fear and fearfulness in animals. Q Rev Biol 70:165-191 Both C, Dingemanse NJ, Drent PJ, Tinbergen JM (2005) Pairs of extreme avian personalities have highest reproductive success. J Anim Ecol 74:667-674 Bouchard TJ Jr, Loehlin JC (2001) Genes, evolution, and personality. Behav Genet 31:243-273 Brodie ED (1992) Correlational selection for color pattern and antipredator behavior in the garter snake Thamnophis ordinoides. Evolution 46:1284-1298 Bshary R, Bergmüller R (2008) Distinguishing four fundamental approaches to the evolution of helping. J Evol Biol 21:405-420 Budaev SV (1998) How many dimensions are needed to describe temperament in animals: a factor re-analysis of two data sets. Int J Comp Psychol 11:17-29 Careau V, Thomas D, Humphries MM, Réale D (2008) Energy metabolism and animal personality. Oikos 117:641-653 Carere C, Drent PJ, Koolhaas JM, Groothuis TGG (2005) Epigenetic effects on personality traits: early food provisioning and sibling competition. Behaviour 142:1329-1355 Chapman BB, Ward AJW, Krause J (2008) Schooling and learning: early social environment predicts social learning ability in the guppy, Poecilia reticulata. Anim Behav 76:923-929 Cockrem JF (2007) Stress, corticosterone responses and avian personalities. J Ornithol 148:S169-S178 Cote J, Dreiss A, Clobert J (2008) Social personality trait and fitness. Proc R Soc Lond B 275:2851-2858

Animal personality and behavioural syndromes

617

Dall SRX, Houston AI, McNamara JM (2004) The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecol Lett 7:734-739 Dall SRX, Giraldeau L-A, Olsson O, McNamara JM, Stephens DW (2005) Information and its use by animals in evolutionary ecology. Trends Ecol Evol 20:187-193 DeWitt TJ, Sih A, Wilson DS (1998) Costs and limits of phenotypic plasticity. Trends Ecol Evol 13:77-81 Dingemanse NJ, Réale D (2005) Natural selection and animal personality. Behaviour 142:1159-1184 Dingemanse NJ, Both C, Drent PJ, van Oers K, van Noordwijk AJ (2002) Repeatability and heritability of exploratory behaviour in wild great tits from the wild. Anim Behav 64:929-938 Dingemanse NJ, Both C, van Noordwijk AJ, Rutten AL, Drent PJ (2003) Natal dispersal and personalities in great tits (Parus major). Proc R Soc Lond B 270:741-747 Dingemanse NJ, Both C, Drent PJ, Tinbergen JM (2004) Fitness consequences of avian personalities in a fluctuating environment. Proc R Soc Lond B 271:847852 Dingemanse NJ, Wright J, Kazem AJN, Thomas DK, Hickling R, Dawnay N (2007) Behavioural syndromes differ predictably between 12 populations of three-spined stickleback. J Anim Ecol 76:1128-1138 Dingemanse NJ, Kazem AJN, Réale D, Wright J (2009a) Behavioural reaction norms: animal personality meets individual plasticity. Trends Ecol Evol 25:81-89 Dingemanse NJ, van der Plas F, Wright J, Réale D, Schrama M, Roff DA, van der Zee E, Barber I (2009b) Individual experience and evolutionary history of predation affect expression of heritable variation in fish personality and morphology. Proc R Soc Lond B 276:1285-1293 Dornhaus A (2008) Specialization does not predict individual efficiency in an ant. PLoS Biol 6:e285, doi:10.1371/journal.pbio.0060285 Drent PJ, van Oers K, van Noordwijk AJ (2003) Realized heritability of personalities in the great tit. Proc R Soc Lond B 270:45-51 Eaves LJ, Martin NG, Heath AC, Hewitt JK, Neale MC (1990) Personality and reproductive fitness. Behav Genet 20:563-568 Endler JA (1986) Natural Selection in the Wild. Princeton University Press, Princeton Falconer DS, Mackay TFC (1996) Introduction to Quantitative Genetics. Longman, New York Fidler AE, van Oers K, Drent PJ, Kuhn S, Mueller JC, Kempenaers B (2007) Drd4 gene polymorphisms are associated with personality variation in a passerine bird. Proc R Soc Lond B 274:1685-1691 Fischer RA (1930) The Genetical Theory of Natural Selection. Dover, New York Gaston AJ (1978) The evolution of group territorial behavior and cooperative breeding. Am Nat 112:1091-1100

618

Ralph Bergmüller

Giraldeau L-A (1984) Group foraging: the skill pool effect and frequencydependent learning. Am Nat 124:72-79 Giraldeau L-A, Dubois F (2008) Social foraging and the study of exploitative behavior. Adv Stud Behav 38:59-104 Giraldeau L-A, Lefebvre L (1985) Individual feeding preferences in feral groups of rock doves. Can J Zool 63:189-191 Gosling SD (2001) From mice to men: what can we learn about personality from animal research? Psychol Bull 127:45-86 Gosling SD, John OP (1999) Personality dimensions in nonhuman animals: a cross-species review. Curr Dir Psychol Sci 8:69-75 Grafen A (1984) Natural selection, kin selection and group selection. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Sinauer, Sunderland/MA, pp 62-84 Grantner A, Taborsky M (1998) The metabolic rates associated with resting, and with the performance of agonistic, submissive and digging behaviours in the cichlid fish Neolamprologus pulcher (Pisces: Cichlidae). J Comp Physiol B 168:427-433 Groothuis TGG, Carere C (2005) Avian personalities: characterization and epigenesis. Neurosci Biobehav Rev 29:137-150 Gross MR (1996) Alternative reproductive strategies and tactics: diversity within sexes. Trends Ecol Evol 11:92-98 Grossarth-Maticek R, Eysenck HJ, Vetter H (1988) Personality type, smoking habit and their interaction as predictors of cancer and coronary heart-disease. Pers Individ Dif 9:479-495 Heinrich B (1979) Bumblebee Economics. Harvard University Press, Cambridge/MA Hessing MJC, Hageslo AM, van Beck JAM, Wiepkema PR, Schouten WGP, Krukow R (1993) Individual behavioural characteristics in pigs. Appl Anim Behav Sci 37:285-295 Hirschenhauser K, Oliveira RF (2006) Social modulation of androgens in male vertebrates: meta-analyses of the challenge hypothesis. Anim Behav 71:265277 Huntingford FA (1976) The relationship between anti-predator behavior and aggression among conspecifics in three-spined stickleback, Gasterosteus aculeatus. Anim Behav 24:245-260 Inman AJ, Lefebvre L, Giraldeau L-A (1987) Individual diet differences in feral pigeons – evidence for resouce partitioning. Anim Behav 35:1902-1903 Ketterson ED, Nolan V (2000) Adaptation, exaptation, and constraint: a hormonal perspective (vol 154:S4-S25, 1999). Am Nat 155:U7, doi:10.1086/303308 Koolhaas JM, Korte SM, de Boer SF, van der Vegt BJ, van Reenen CG, Hopster H, de Jong IC, Ruis MAW, Blokhuis HJ (1999) Coping styles in animals: current status in behavior and stress-physiology. Neurosci Biobehav Rev 23:925935 Krebs JR, Davies NB (1997) Behavioural Ecology. An Evolutionary Approach. Blackwell Science, Oxford

Animal personality and behavioural syndromes

619

Lande R, Arnold SJ (1983) The measurement of selection on correlated characters. Evolution 37:1210-1226 Lynch M, Walsh B (1998) Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland/MA Mangel M, Stamps J (2001) Trade-offs between growth and mortality and the maintenance of individual variation in growth. Evol Ecol Res 3:583-593 Martin P (1998) The Sickening Mind. Harper Collins, London Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge McElreath R, Strimling P (2006) How noisy information and individual asymmetries can make ‘personality’ an adaptation: a simple model. Anim Behav 72:1135-1139 McElreath R, Luttbeg B, Fogarty SP, Brodin T, Sih A (2007) Evolution of animal personalities. Nature 450:E5-E5 McGuffin P, Riley B, Plomin R (2001) Genomics and behavior: toward behavioral genomics. Science 291:1232-1249 McNamara JM, Houston AI (2002) Credible threats and promises. Philos Trans R Soc Lond B 357:1607-1616 Meany MJ (2001) Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu Rev Neurosci 24:1161-1192 Mendl M, Paul ES (1991) Litter composition affects parental care, offspring growth and the development of aggressive behavior in wild house mice. Behaviour 116:90-108 Nesse RM (2001) Evolution and the Capacity for Commitment. Russell Sage Press, New York Nussey DH, Wilson AJ, Brommer JE (2007) The evolutionary ecology of individual phenotypic plasticity in wild populations. J Evol Biol 20:831-844 Oliveira RF (2004) Hormones, social context and animal communication. In: McGregor PK (ed) Animal Communication Networks. Cambridge University Press, Cambridge Oliveira RF, Carneiro LA, Canario AVM (2005) No hormonal response in tied fights. Nature 437:207-208 Oliveira RF, Taborsky M, Brockmann HJ (2008) Alternative Reproductive Tactics. An Integrative Approach. Cambridge University Press, Cambridge Penke L, Denissen JJA, Miller GF (2007) The evolutionary genetics of personality. Eur J Pers 21:549-587 Phoenix CH, Goy RW, Gerall AA, Young WC (1959) Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology 65:369-382 Réale D, Dingemanse NJ (in press) Personality and individual social specialisation. In: Szelely T, Moore AJ, Komdeur J (eds) Social Behaviour: Genes, Ecology and Evolution. Cambridge University Press, New York Réale D, Festa-Bianchet M (2003) Predator-induced natural selection on temperament in bighorn ewes. Anim Behav 65:463-470

620

Ralph Bergmüller

Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) Integrating animal temperament within ecology and evolution. Biol Rev 82:291-318 Riechert SE, Hedrick AV (1993) A test for correlations among fitness-linked behavioural traits in the spider Agelenopsis aperta (Araneae, Agelenidae). Anim Behav 46:669-675 Roff DA (1997) Evolutionary Quantitative Genetics. Chapman and Hall, New York Ros AFH (1999) Effects of testosterone on growth, plumage pigmentation, and mortality in black-headed gull chicks. Ibis 141:451-459 Ros AFH, Hirschenhauser K, Oliveira RF (2001) The interaction between organizational and activational effects of testosterone in the control of early aggression in birds: a comment to Sasvári, Hegyi and Péczeli. Ethology 107:851-853 Sapolsky RM, Romero LM, Munck AU (2000) How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr Rev 21:55-89 Schelling TC (1960) The Strategy of Conflict. Harvard University Press, Cambridge/MA Schlichting CD, Pigliucci M (1998) Phenotypic Evolution: A Reaction Norm Perspective. Sinauer Associates, Sunderland/MA Schmitz PG (1992) Personality, stress-reactions and disease. Pers Individ Dif 13:683-691 Schürch R (2008) Individual variation in life-history and behavioural type in the highly social cichlid neolamprologus pulcher. PhD, University of Bern Schuett W, Tregenza T, Dall SRX (2010) Sexual selection and animal personality. Biol Rev, doi:10.1111/j.1469-185X.2009.00101.x Schwabl H (1993) Yolk is a source of maternal testosterone for developing birds. Proc Natl Acad Sci USA 90:11446-11450 Sih A, Bell AM (2008) Insights for behavioral ecology from behavioral syndromes. Adv Stud Behav 38:227-281 Sih A, Bell AM, Johnson JC (2004a) Behavioral syndromes: an ecological and evolutionary overview. Trends Ecol Evol 19:372-378 Sih A, Bell AM, Johnson JC, Ziemba RE (2004b) Behavioral syndromes: an integrative overview. Q Rev Biol 79:241-277 Sinervo B, Svensson E (2002) Correlational selection and the evolution of genomic architecture. Heredity 89:329-338 Sinn DL, Gosling SD, Moltschaniwskyj NA (2008) Development of shy/bold behaviour in squid: context-specific phenotypes associated with developmental plasticity. Anim Behav 75:433-442 Sluyter F, Bult A, Lynch CB, Vanoortmerssen GA, Koolhaas JM (1995) A comparison between house mouse lines selected for attack latency or nestbuilding: evidence for a genetic basis of alternative behavioral strategies. Behav Genet 25:247-252 Stamps JA (2007) Growth-mortality tradeoffs and ‘personality traits’ in animals. Ecol Lett 10:355-363 Stevenson-Hinde J, Stillwell-Barnes R, Zunz M (1980) Subjective assessment of rhesus monkeys over four successive years. Primates 21:66-82

Animal personality and behavioural syndromes

621

Taborsky M, Oliveira RF, Brockmann HJ (2008) The evolution of alternative reproductive tactics: concepts and questions. In: Oliveira RF, Taborsky M, Brockmann HJ (eds) Alternative Reproductive Tactics: An Integrative Approach. Cambridge University Press, Cambridge, pp 1-21 van Oers K, de Jong G, Drent PJ, van Noordwijk AJ (2004a) A genetic analysis of avian personality traits: correlated, response to artificial selection. Behav Genet 34:611-619 van Oers K, Drent PJ, de Jong G, van Noordwijk AJ (2004b) Additive and nonadditive genetic variation in avian personality traits. Heredity 93:496-503 van Oers K, Drent PJ, de Goede P, van Noordwijk AJ (2004c) Realized heritability and repeatability of risk-taking behaviour in relation to avian personalities. Proc R Soc Lond B 271:65-73 van Oers K, de Jong G, van Noordwijk AJ, Kempenaers B, Drent PJ (2005) Contribution of genetics to the study of animal personalities: a review of case studies. Behaviour 142:1185-1206 van Oers K, Drent PJ, Dingemanse NJ, Kempenaers B (2008) Personality is associated with extrapair paternity in great tits, Parus major. Anim Behav 76:555563 Verbeek MEM, Drent PJ, Wiepkema PR (1994) Consistent individual differences in early exploratory behaviour of male great tits. Anim Behav 48:1113-1121 Wechsler B (1995) Coping and coping strategies: a behavioural view. Appl Anim Behav Sci 43:123-134 West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford University Press, Oxford Wilson EO (1985) The sociogenesis of insect colonies. Science 228:1489-1495 Wilson EO (1987) Causes of ecological success – the case of the ants. J Anim Ecol 56:1-9 Wilson DS (1998) Adaptive individual differences within single populations. Philos Trans R Soc Lond B 353:199-205 Wilson DS, Coleman K, Clark AB, Biederman L (1993) Shy bold continuum in pumpkinseed sunfish (Lepomis gibbosus): an ecological study of a psychological trait. J Comp Psychol 107:250-260 Wilson DS, Clark AB, Coleman K, Dearstyne T (1994) Shyness and boldness in humans and other animals. Trends Ecol Evol 9:442-446 Wingfield JC, Hegner RE, Dufty AM Jr, Ball GF (1990) The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. Am Nat 136:829-846 Wolf M, van Doorn GS, Leimar O, Weissing FJ (2007a) Evolution of animal personalities – Reply. Nature 450:E5-E6, doi:10.1038/nature06326 Wolf M, van Doorn GS, Leimer O, Weissing FJ (2007b) Life-history trade-offs favour the evolution of animal personalities. Nature 447:581-584 Wolf M, van Doorn GS, Weissing FJ (2008) Evolutionary emergence of responsive and unresponsive personalities. Proc Natl Acad Sci USA 105:1582515830

Chapter 20

Social learning and culture in animals CAREL P. VAN SCHAIK

ABSTRACT Most animals must learn some of the behaviours in their repertoire, and some must learn most. Although learning is often thought of as an individual exercise, in nature much learning is social, i.e. under the influence of conspecifics. Social learners acquire novel information or skills faster and at lower cost, but risk learning false information or useless skills. Social learning can be divided into learning from social information and learning through social interaction. Different species have different mechanisms of learning from social information, ranging from selective attention to the environment due to the presence of others to copying of complete motor sequences. In vertical (or oblique) social learning, naïve individuals often learn skills or knowledge from parents (or other adults), whereas horizontal social learning is from peers, either immatures or adults, and more often concerns eavesdropping and public information use. Because vertical social learning is often adaptive, maturing individuals often have a preference for it over individual exploration. The more cognitively demanding social learning abilities probably evolved in this context, in lineages where offspring show long association with parents and niches are complex. Because horizontal learning can be maladaptive, especially when perishable information has become outdated, animals must decide when to deploy social learning. Social learning of novel skills can lead to distinct traditions or cultures when the innovations are sufficiently rare and effectively transmitted socially. Animal cultures may be common but to date taxonomic coverage is insufficient to know how common. Cultural evolution is potentially powerful, but largely confined to humans, for reasons currently unknown. A general theory of culture is therefore badly needed.

624

Carel P. van Schaik

20.1 Introduction The traditional null model of behavioural ontogeny is that animals acquire their species-specific behavioural repertoires through maturation of genetically anchored developmental programs (instincts), supplemented by individual exploration of environmental affordances (individual learning). The main reason for this position, however, may turn out to be an historical accident: comparative psychology developed as an experimental, laboratorybased, rather than a field-based naturalistic science. In the wild, opportunities for social learning arise whenever animals interact or even are in mere proximity. Learning is a change in the brain produced by experience. We usually call this change knowledge or skill, and infer that it occurred by noting a change in behaviour. Social learning is learning under the influence of conspecifics. More formally, social learning, or socially mediated learning, can be defined as changes in an individual’s behaviour resulting from attending to another individual’s behaviour or its products (Box 1984, Fragaszy and Perry 2003). Social learning is different from communication. Both involve responding to conspecifics (see Schaefer this volume), but communication usually does not involve learning. However, there is some overlap, as in learning through social interaction (see below). Through social learning, animals can acquire information or skills they would otherwise not have obtained, or acquire routine skills faster. It is therefore not surprising that social learning is the default mode of learning by naïve individuals in those animal species where overlapping generations and long parent-offspring association enables this. But, as we will see, that does not necessarily mean that traditions and cultures are widespread, nor that animals will always rely on cues from conspecifics, even if they are available. I first examine social learning by asking a number of proximate questions about how (mechanisms), what (content), when (at what age and from whom) and where (contexts). I then turn to ultimate questions, such as its adaptive significance and the factors favouring the evolution of the more advanced sorts of social learning. But we can also ask an additional set of questions, concerning the consequences of social learning at the level of populations and species. An immediate consequence can be culture, whereas in the long run social-learning skills can co-evolve with tendencies to explore and innovate, and thus intelligence. Throughout, the reader must keep in mind that social learning is studied in very different research traditions that have developed largely independently from each other. Behavioural ecologists have generally focused on

Social learning and culture in animals

625

adults gathering ecologically relevant information (horizontal transmission), whereas psychologists (and primatologists and anthropologists) have generally focussed on the mechanisms of immatures acquiring skills (vertical transmission). These traditions complement each other, and I will therefore try to integrate their approaches.

20.2 Proximate aspects of social learning 20.2.1 How: mechanisms of social learning Social learning is quite heterogeneous in terms of mechanisms. Learning by attending to social information has received most attention. It may be used for the acquisition of social skills through eavesdropping (i.e. observing social interactions between others: e.g. Valone 2007), but it is most studied in the context of acquiring subsistence skills. It encompasses a variety of poorly understood processes, and the technical literature is complicated and confusing. The interested reader is referred to Byrne (2002), Whiten et al. (2004), Subiaul (2007) and Hoppitt and Laland (2008). For the present purpose, I divide learning by attending to social information into a few major categories (Table 20.1). In the simplest category are social or response facilitation (non-specific increases in activity as a result of the proximity or activity of conspecifics) and selective association (following role models around), the latter perhaps simply as a consequence of gregariousness. These forms of social learning share the property that the presence of others biases the exploratory tendency of the naïve individuals. Yet, this mechanism may be sufficient for young animals to acquire the same diet as their mothers, simply by following her around and thus selectively encountering, and learning about, whatever she eats (van der Post and Hogeweg 2006). The next major category includes cases where naïve animals use the presence or behaviour of other animals as pointers to a specific site or object upon which to focus their own independent exploration and learning (enhancement). This may help naïve animals to develop preferences or skills by exposing them to the affordances of objects and foods. Enhancement is closely related to social information use (Valone 2007), where animals base decisions to forage or move or their choice of mates on observation of the presence and activity of conspecifics. Such information use becomes enhancement when the animals doing this learn something new in the process. Indirect forms of enhancement may also be found. Thus, naïve animals may increase exploration upon encountering the feeding sign of con-

626

Carel P. van Schaik

Table 20.1 The main kinds of social learning distinguished here. Learning by attending to social information

Non-observational social learning

1. facilitation / gregariousness 2. enhancement

Observational social learning

3. contextual imitation / observational conditioning 4. production imitation / emulation

Learning through social interaction

1. interaction conditioning 2. teaching

specifics. For instance, aye-ayes (Daubentonia madagascariensis) engage in tap foraging, in which they tap on branches with a specialized longnailed finger and listen for cavities. Upon discovery of a cavity, they gnaw a hole and fish out larvae. Young aye-ayes are attracted to the holes made by other aye-ayes, and explore them selectively (Krakauer 2005). In all these cases, the presence or actions of conspecifics lead naïve animals to learn about the environment, but the latter do not necessarily pay attention to the models’ actions. In the following two major categories, they do (Subiaul 2007). In simple observational learning, species-typical actions are copied. For instance, in mate copying, seen in fishes and birds, females copy the mate choice of other females observed to mate with particular (types of) males (Witte 2006). Another form is so-called observational conditioning, in which learners copy the striking responses of models that accompany alarm calls, and learn to associate them with the releasing stimuli (e.g. predator sightings). Many primates learn the proper use of alarm calls this way (Seyfarth and Cheney 1980). Note that in observational conditioning, the naïve animals are socially fine-tuning what are basically instinctive actions, using the reactions of experienced conspecifics to classify or label other animals or objects. The copying of known motor patterns as seen in many birds and some primates (Zentall 2004, Subiaul 2007) is often referred to as contextual or familiar imitation. This is not a trivial ability because it somehow requires a translation from perception into action, and probably relies on dedicated neural mechanisms, such as the presence of mirror neurons, i.e. neurons that fire both when an action is performed and when the same action is observed in others (Rizzolatti 2005).

Social learning and culture in animals

627

The final major category of social learning is observation learning through production imitation (also: motor or novel imitation), which refers to the copying of novel actions or action sequences (emulation, the copying of the goals of the model’s actions, is probably very similar). This ability is remarkably common among birds (Zentall 2004), but among primates motor imitation has so far only been found in apes while monkeys are capable of contextual imitation at best (Whiten and van Schaik 2007). Subiaul (2007) speculates that this is because production imitation requires neural adaptations that mediate the planning and coordination of gross and fine motor patterns. As to the cognitive abilities required, there is broad consensus that observational learning is cognitively more demanding than non-observational social learning, and that production imitation is more demanding than context imitation. For instance, human infants develop the ability to copy known motor actions before they can copy novel actions or action sequences (Masur 1988). Moreover, among primates, there are taxonomic neurobiological differences that support the presence of production imitation in humans but not in monkeys (reviewed in Subiaul 2007). Finally, at least among mammals, the taxonomic distribution of this ability may turn out to be correlated with the presence of mirror-self-recognition, cognitive empathy, and elements of a theory of mind (de Waal 2001, Heyes 2001). However, because documenting the actual abilities involved in social learning requires careful experiments, we know very little about the taxonomic distribution and the link with other cognitive abilities or even brain size. In addition to learning from social information, Table 20.1 recognises a second major kind of social learning: learning through social interaction. This form of social learning overlaps extensively with communication, but differs in that it also involves conditioning (learning) as a result of the individual’s social interactions: agonistic and affiliative behaviour, as well as social play (e.g. Pellegrini et al. 2007). Learning through social interactions (or socialisation: Box 1984) is probably the major means by which immature animals acquire social skills, as suggested by social deprivation, which produces socially incompetent adults (Harlow and Harlow 1962) and by interspecific cross-fostering, which shows adjustment to reigning social norms (de Waal and Johanowicz 1993). Learning through social interaction normally plays no role in the learning of subsistence skills or ecological knowledge, but teaching, which does, also belongs in this category. In teaching, the model takes an active role. A behaviour pattern qualifies as teaching if certain actions only occur in the presence of a naïve observer, carry some cost but are not immediately beneficial to the teacher, and if the observer profits from these ac-

628

Carel P. van Schaik

tions in that it learns these actions faster than it would otherwise do (Caro and Hauser 1992). This does not require that the teacher is aware of the knowledge state of the naïve animal, i.e. that the teaching is intentional, although that would make it more effective. It should therefore be expected in many species, but although it has over the past few years been found in various species, far more striking is its absence in most. The species in which it is present are largely characterised by one or both of the following two conditions: (i) the presence of complex diets, where naïve individuals need help to familiarise themselves with the affordances of their food, usually fast-moving prey or embedded food items that must be extracted; and (ii) the presence of altruistic tendencies on the part of the teacher, be it through high relatedness or prosocial attitudes or both, as in cooperative breeders (Hoppitt et al. 2008, Burkart and van Schaik 2010, Burkart et al. in press). 20.2.2 What: the content of social learning Dawkins (1976) coined the term ‘meme’ for the fundamental unit of socially transmitted information, and although it has led to some theorising, it has not caught on among those studying social learning (Aunger 2007). One reason is the association with selfish genes, which has led to a theoretical focus on parasitic behavioural innovations to the exclusion of the numerous useful ones (see below). Another reason is the strongly anthropomorphic focus, which meant that meme transmission was equated with production imitation, thus ignoring the variability of socially transmitted behaviours found among animals. I will therefore refer to ‘informational variant’ for the general case and use more detailed words where possible. Here, I propose the following classification of informational variants that can be acquired, with or without modification, through social learning: perishable information, non-perishable information (or labels), skills and knowledge, signal variants, and symbols (after van Schaik et al. 2003). Perishable information refers to the kind of information that is often easily gleaned by associating with another individual, or even observing this individual from afar, summarised as ‘knowing where’. Non-perishable information or labels, on the other hand, is about ‘knowing that’ (also: declarative knowledge). Thus, knowing that a particular kind of red berry is edible is a label, whereas knowing that this week there is a good patch of them at a particular location is perishable information. Similarly, knowing that a particular animal is a predator is a label, but knowing where it roamed yesterday and where it might be today, is perishable information. A skill refers to ‘knowing how,’ e.g. about how to access food, or how to

Social learning and culture in animals

629

swing to another tree (also: procedural knowledge). Some informational variants may be harder to classify. Thus, remembering that these red berries tend to be available at a particular location at this time of year is either knowledge or skill, though certainly not perishable information. The final two categories refer to social communication. A signal variant is an alternative way to signal the same particular message. For instance, vocal dialects contain signal variants, because one assumes that the content of the message conveyed is not affected by the change in acoustical features. Other cases concern the non-vocal domain. Thus, orang-utans (Pongo abelii and P. pygmaeus) produce kiss-squeaks when in distress, and have created geographic variants. In some regions, orang-utans place these kisses on (bundles of) leaves. The meaning is the same, albeit perhaps more pronounced (Hardus et al. 2009). A symbol is a communicative variant that is arbitrary, its meaning having become a local or regional convention, rather than species-wide, as in signals. Thus, orang-utan mothers, in some localities but not others, make acoustically distinct sounds to call in their infants, but the actual sounds vary among populations (A. Lameira et al. unpubl. data). Different social-learning mechanisms may be minimally needed to acquire these different classes of variants, with skills almost certainly requiring observational learning. It is not clear how the communicative variants are learned socially, although vocal learning in birds is known to involve dedicated neural circuits (Prather et al. 2008). All information variants beyond perishable information are thought to arise through innovation, either of the inadvertent, accidental kind or by cognitively more demanding processes such as insight. Little is known, however, of the processes that produce innovation. 20.2.3 From whom: deployment strategies and transmission biases If an animal is learning socially, whom should it pick as a model (Laland 2004)? A crucial distinction here is between (i) vertical or oblique social learning, by immatures from parents (vertical) or other adults (oblique), and (ii) horizontal social learning, which is among peers, and usually studied among adults. Consider vertical social learning first. A simple prediction is that young animals have more to learn from others than adults, and that the motivation for social learning should therefore decline with age. This is well known, but rarely documented in a way one can compare across species, although interesting taxonomic variation is expected. For instance, in species with

630

Carel P. van Schaik

Fig. 20.1 Social learning in nature is often vertical: immatures learn from their mother, as in this Bornean orang-utan mother-infant pair. Photo © Lynda Dunkel.

extensive cultural repertoires, such as great apes, adults may continue to be interested in, and capable of, social learning (e.g. Whiten et al. 2005), although this is not always the case (see Sect. 20.3.3). A second straightforward prediction is that naïve animals should always be selectively attending in situations where novel skills or information can be acquired. In many species, young behave like apprentices: following their mother or other caretakers around with intense curiosity, and paying special attention when difficult skills are being demonstrated (Fig. 20.1). Foraging skills are often learned because infants scrounge partly processed foods from the mother (Terkel 1996, Jaeggi et al. 2008, Rapaport and Brown 2008), or selectively focus on her activities when she targets foods that are rare or difficult to process (Tarnaud and Yamagiwa 2008, Jaeggi et al. 2009). As a result, long-term studies of the development of maturing individuals show that they tend to acquire the variants used by those with whom they associated the most (Perry and Ordoñez 2006; but see Matthews 2009; see also BOX 20.1). Other decisions are also relevant. If there is a choice among models, which of those should a naïve animal attend to? One generally expects

Social learning and culture in animals

631

there to be some transmission bias. The most obvious solution is to focus attention on those individuals that show the best mastery and allow being observed, which is what naïve chimpanzees and capuchin monkeys do in the case of nut cracking (Matsuzawa et al. 2001, Biro et al. 2003, Ottoni et al. 2005) – a tendency known as model-based or prestige bias (Richerson and Boyd 2005). This tendency is also captured in de Waal’s (2001) BIOL rule: bonding- and identification-based observational learning. However, this rule additionally postulates a ‘desire to be like others’ on the part of the naïve immature, i.e. social conformity. This addition may be most pertinent for signal variants and symbols, because they do not differ in content. It is also possible that the naïve individual is exposed to multiple distinct variants. The null hypothesis then is that an animal simply adopts whatever behaviour is demonstrated most commonly, or is demonstrated by the largest number of models. If the learner can evaluate the value of each variant, she may simply adopt the one or ones she finds satisfactory (known as content bias; cf. Galef 1995). This is evident for labels, involving food choice or predator recognition, but may also be relevant where learners carefully evaluate a new feeding skill. However, suboptimal choices may also ensue. First, animals may be conservative, i.e. have a tendency to stick to the technique they know well, even if other, more efficient techniques are shown to them by others in the group (Biro et al. 2003, Hrubesch et al. 2009). Second, they may display a disproportionate tendency to follow the majority in their choice (positive frequency dependence), i.e. conformity (Efferson et al. 2008). A common cause of conformity is obligate gregariousness, which forces all animals to make the same choices when exploring alternatives require leaving the group (Day et al. 2001). Active conformity is common among humans, partly because it is the optimal decision when information about the variants is poor (informational conformity), and partly because it is a sign of group membership or may be a socially imposed norm (normative conformity: Henrich and McElreath 2007). Social learning can also be horizontal. In this context, it is possible that new skills or important knowledge are gained (Page and Ryan 2006), but much horizontal transfer also concerns perishable information about the current state of the environment or of conspecifics. Behavioural ecologists often distinguish between public information use and eavesdropping. In the use of public information (also: inadvertent social information), individuals learn about the current state of the local environment by attending to the behavioural decisions or the success rates of others. The information is gained from inadvertent cues, not signals (see Schaefer this volume).

632

Carel P. van Schaik

Table 20.2 Comparing anthropological and behavioural-ecology approaches. Field

Behavioural Ecology

Emphasis

AnthropologyPsychologyPrimatology

Direction of transfer

vertical and oblique

horizontal

Content

labels and skills

perishable information

Method of social learning studied?

yes

no

Consequences

traditions or culture (often)

not so

Examples include foraging animals using the behaviour of others to determine whether a particular food patch is worth visiting or where to settle for breeding (Valone 2007). Eavesdropping is a form of social learning about the quality of other individuals by attending to their social interactions, e.g. fights, usually focusing on the signals being exchanged. It is social learning because the eavesdropper often subsequently modifies its behaviour, but what is learned is usually perishable information, e.g. the fighting ability of a particular individual or the attractivity of that individual as a mate. Eavesdropping has been demonstrated through elegant experiments (Dabelsteen 2005). It is surprisingly common among birds, and may also be widespread among mammals (Valone 2007). One well-studied example is ‘mate choice copying,’ where females show a preference for mating with a male that they had just seen successfully mate with another female over other, otherwise similar males (Witte 2006). This makes sense when other information to choose between potential mates is absent; indeed, copying is more common when the males are evenly matched or when females are naïve (Valone 2007). There is an interesting contrast in the emphases of different research traditions (Bonnie and Earley 2007). Psychologists (and with them primatologists) are generally interested in vertical and oblique social learning, often with a strong focus on mechanisms, and thus also in cultural phenomena in the wild. Many anthropologists have similar interests, although they take the presence of complex social-learning mechanisms for granted, and instead concentrate on context and content biases in the deployment of these social-learning abilities. Behavioural ecologists, in contrast, are rarely interested in mechanisms, and often focus on horizontal social learn-

Social learning and culture in animals

633

ing of perishable information in field conditions. As a result, they tend to focus on eavesdropping and public information use, rather than culture (Table 20.2). These different traditions have led to some important gaps in our knowledge. Most importantly, psychologists and primatologists rarely study horizontal transmission, especially not in the wild, whereas behavioural ecologists tend not to know which mechanisms their subjects are using, even when they must often go beyond enhancement (cf. Galef and Giraldeau 2001).

20.3 Social learning as an adaptation 20.3.1 When is social learning adaptive? The previous section leaves little doubt that social learning is generally adaptive and that the various social learning mechanisms evolved through natural selection. However, that does not mean that social learning is always the optimum way of acquiring information. Here, I explore in more detail under what conditions the ability to learn socially is adaptive. In general, social learning confers adaptive significance if BSL – CSL > BIL – CIL, where B refers to a fitness benefit, C to a fitness cost and the subscripts SL and IL stand for social and individual learning, respectively. The main reason why social learning is generally adaptive is obvious: CSL << CIL. Social learning frees up individuals from having to decide which environmental stimuli to attend to, thus fundamentally improving the signal-to-noise ratio of environmental inputs for an individual. Moreover, individual learning involves costs (Mery and Kawecki 2004), as well as risks, especially if it involves exploration (Mettke-Hofmann et al. 2006). Social learning thus reduces such costs and risks, while speeding up the acquisition of skills or knowledge, making it an adaptive strategy under a wide range of conditions. This analysis assumes that the models are indifferent. However, from the demonstrator’s perspective, there may also be costs or benefits to being copied. Consider horizontal transmission first. Assume that the possession of the information or the skill provides a clear fitness benefit. Then, if there is negative density dependence in the population, models would lose fitness if they are being copied (i.e. provide public information). Thus, one might expect animals to actively reduce the risk of being copied by making activities inconspicuous to others or even driving others away, unless they

634

Carel P. van Schaik

benefit from attracting others into their proximity or are surrounded by relatives, or there is group augmentation (Kokko et al. 2001). Similarly, some animals may incur a cost from having their signals intercepted by eavesdroppers. For instance, courtship activities in birds often attract male neighbours, who then interfere with mating (Dabelsteen 2005, Valone 2007). As a result, we expect some signals, e.g. courtship vocalisations, to be highly directed, which they indeed are in red-winged blackbirds (Agelaius pheoniceus: Patricelli et al. 2007). Vertical transmission of skills or knowledge, on the other hand, generally involves a stake in the learner’s welfare through kinship or group augmentation, making it in the demonstrator’s interest to ensure that the information or skill is passed on. Unfortunately, very few studies have specifically tested these straightforward predictions. 20.3.2 Vertical social learning Returning to the learner’s perspective, we can ask when skills or information acquired through social learning are adaptive separately for vertical and horizontal social learning. Vertical social learning is most likely to be adaptive because immatures are naïve and must acquire many of their skills though learning. Indeed, maturing individuals in many species may prefer to learn socially. First, such a preference is suggested by examples of the avoidance of novel foods by immatures until a parent has eaten it (meerkats: Thornton 2008; aye-ayes: Krakauer 2005; callitrichids: Rapaport and Brown 2008; chimpanzees: Ueno and Matsuzawa 2005), as well as by the very low rates of independent food exploration among immature wild orang-utans (Jaeggi et al. 2009). Second, it is strongly suggested by cross-fostering experiments. Such experiments are most feasible in birds, where eggs can be exchanged between nests of different species. Growing up with another species had a dramatic impact on the overall behaviour of the transplanted individuals in cockatoos or titmice (Rowley and Chapman 1986, Slagsvold and Wiebe 2006). These findings support the assumption that especially the young and naïve actually show a preference for social learning over individual exploration. Overall, these findings suggest that our null model of individual learning, with social learning added occasionally, is probably utterly wrong for all species in which vertical learning of skills is important. It is certainly wrong for humans (Meltzoff et al. 2009). Nonetheless, even vertical social learning may sometimes be maladaptive to the learner (i.e. BSL << BIL) in some conditions. In humans, young children rely extensively on social learning, cobbling pretty much their whole repertoire together through copying. The adoption of dangerous pas-

Social learning and culture in animals

635

times, such as rock climbing, or costly customs, such as low fertility, shows that the unquestioning copying of behaviour patterns shown by prestigious role models or the whole community need not be adaptive (Richerson and Boyd 2005). Why are human youngsters so uncritical? The most plausible explanation is that our socially learned skills are so numerous and so complicated that the best strategy for a youngster is to simply copy without much individual evaluation. On average, this tendency is clearly adaptive, although maladaptive habits may also arise (Richerson and Boyd 2005). This argument is supported by comparisons between chimpanzees and children: human children tend to imitate irrelevant details of a demonstrated task whereas chimpanzees do not (Horner and Whiten 2005), a phenomenon known as over-imitation (Lyons et al. 2007). 20.3.3 Horizontal social learning Let us now turn to horizontal learning. Here, the asymmetry in knowledge and skills is less obvious than between caretaker and young, and it is more likely that socially acquired information may be misleading, especially when it is perishable. This point is intuitively obvious: because the location of food or the identity of the optimal mate vary over time, simply copying the choices of others will not necessarily produce optimum decisions. Indeed, because public information is often less reliable than individual information, individuals use it as an additional rather than the sole source of information on the environment (Galef 1995, Giraldeau et al. 2002), and tend to ignore public information when they have reliable private information (van Bergen et al. 2004, Valone 2007). Another reason may be the cost of learning when the animal already possesses a certain skill, e.g. how to remove food in an extractive-foraging task. If another individual shows another technique for the same task, adopting the new technique would be adaptive if the potential increase in future yield rate exceeds the learning costs (which also includes a temporary depression in yield). Animals obviously cannot easily estimate the utility of switching to another technique. Research on captive chimpanzees shows that individuals that have mastered one technique to solve a task are reluctant to invest in learning another technique to solve the same task (Marshall-Pescini and Whiten 2008, Hrubesch et al. 2009). Now that this conservatism has been revealed, it is interesting to find out which conditions may compel individuals to invest in learning new skills. Rogers (1988) argued that the value of socially learned information inevitably declines, and eventually the benefit of social learning relative to individual learning disappears (see also Boyd and Richerson 1985). This

636

Carel P. van Schaik

conclusion is too strong. As we saw, it is often adaptive to acquire knowledge (e.g. is this fruit edible?) or skills (how do I extract these seeds?), which are useful, perhaps after some modification or generalisation. Perhaps the safest generalisation at present is that preferential reliance on vertical social learning of skills is often adaptive, whereas preferential reliance on horizontal social learning to acquire perishable information is often maladaptive. 20.3.4 Evolution of social-learning abilities The simplest forms of social learning require no more than selective association or facilitation, and selection could easily favour such a tendency if the information acquired in this way is on average adaptive. However, under what conditions could selection have favoured the evolution of the cognitively more demanding and therefore more costly observational forms of social learning? A naïve answer would be that mere gregariousness or living in larger rather than smaller groups provides sufficient conditions for the evolution of observational-learning abilities. However, there is no evidence for an effect of group size on the frequency of social learning, both within (Huffman and Hirata 2003) or among species (Reader and Laland 2002) of primates and among bird species (Lefebvre et al. 1996). Moreover, many group-living lineages, such as equids, ground squirrels or lemurs show no evidence for advanced social learning skills. Finally, a mathematical analysis suggested that the tolerance of models and the duration of close proximity to them are more important for the acquisition of complex skills than their number (van Schaik and Pradhan 2003). Obviously, when teaching is involved, group size becomes even less important. More likely, the answer has to do with two factors: how easy is it to learn the skills independently, and how tolerant are models? First, acquiring a skill by copying the model’s behaviour or goals is needed when the odds are poor that the naïve individual will independently come up with the skill by simply being pointed to the right situation through association or enhancement (cf. Boyd and Richerson 1996). It is therefore likely that selection for enhanced social-learning skills largely took place in an ecological context where invented skills were especially useful, such as for extractive foragers or hunters (the co-evolution between these two cognitive abilities will be explored later). A major role for ecology is plausible because social skills can often be learned through social interactions and eavesdropping (unlikely to require production imitation).

Social learning and culture in animals

637

Table 20.3 Kinds of learning deployed by orang-utan infants to acquire their diet, and the nature of the foods and the knowledge acquired this way (after Jaeggi et al. 2008, 2009). Social/ individual learning?

Independent exploration

Mechanism

individual trial and error

proximity and enhancement

% items in Pongo diet

< 1%

≈ 54%

Food types Kind of knowledge

Parent-offspring association

common and easy-to-process Labels (what?)

+ informational food transfer ≈ 46%

+ observational learning

+ teaching

≈ 17%

0%

rare and difficult-toprocess

extracted

Skills (how?)

Second, the opportunity for social acquisition of skills will depend on the benefits for models of being copied and how often and how long observational learning is possible. These conditions suggest observational learning abilities in species with slow-paced life history and long associations between parents or caretakers and immatures, which form stable social units with overlapping generations and high social tolerance. There is as yet not enough information on the taxonomic distribution of observational learning to test these suggestions, but the high incidence of simple imitation in birds noted above suggests that tolerant parent-offspring association and difficult diets play a major role. Animals are expected to use their whole repertoire of social-learning abilities, but will rely on the most demanding ones to learn the most difficult skills. Indeed, in wild Bornean orang-utans, immatures use a variety of mechanisms to learn their skill set (Table 20.3), from proximity to scrounging to production imitation and emulation, often in combination, and also often insert bouts of practice after observing their mother perform a difficult skill (Jaeggi et al. 2009).

638

Carel P. van Schaik

20.4 Immediate consequences of social learning: traditions and culture 20.4.1 Animal cultures Geographic variation in animal behaviour is normally explained by invoking genetic or ecological processes. An ecological explanation would claim that all individuals exposed to a geographically localised set of habitat features independently converge on the same behavioural response. A genetic explanation would argue that all individuals in a particular region have a strong genetic predisposition to develop the behaviour. This would generally lead to clear-cut geographic clusters, whose boundaries coincide with subspecies boundaries or long-term dispersal barriers. However, social learning can also produce geographically varying presence or absence of behavioural variants, otherwise known as traditions or cultures. Culture has traditionally been considered a major dividing line between humans and animals. Yet, cultural anthropologists cannot really agree on a definition of culture (Durham 1991), except that it pervades all our actions and even our thinking, and have tended to concentrate on the belief systems, social norms and rituals of a society at the expense of its technology and subsistence pattern (McGrew 2004). A simpler biological definition of culture – socially transmitted skills and knowledge – is both consistent with the phenomena captured in human culture and applicable to animals. To satisfy it, animals in nature must acquire an innovative behaviour (i.e., behaviour that did not arise routinely in a given environment but was instead invented by someone) by learning it from others who already possess this innovation. Because the word culture has long been earmarked for privileged use for the human species, some biologists have proposed to use the term tradition for cultural phenomena among animals. Traditions are defined as enduring distinctive behaviour patterns characteristic of social units and passed on through social learning (Fragaszy and Perry 2003). While this definition of animal cultures averts interdisciplinary spats and can encompass human culture, it also leads us to disregard possible conceptual and phylogenetic continuities. Moreover, the concept of tradition generally requires stability across generations, but animal data usually do not have the time-depth to demonstrate this. Here, I will use the word culture to stress the continuities with those among animals, while acknowledging the radically different uses that evolved among humans. Social learning does not automatically produce culture. Horizontally transmitted perishable information about environment (public information) or conspecifics (eavesdropping) does not lead to culture because any spa-

Social learning and culture in animals

639

tial patterns are transient. Likewise, most vertically transmitted information generally does not produce culture, because animals everywhere will readily stumble upon the same innovation. As a result, culture should largely be limited to taxa that can learn through observation and are intelligent enough to come up with strong innovations (sensu Ramsey et al. 2007). Even among the latter taxa, culture in the spatial sense may not arise if the species’ social organisation is such that all transmission is strictly vertical. In dolphins (Tursiops truncatus), for example, infants learn their foraging specialisations (i.e., innovations) exclusively from their mothers (Mann and Sargent 2003, Krützen et al. 2005), and there may be several different specialisations within a single locality, and indeed in most localities. However, despite the absence of geographic variation, it would still make sense to call the behaviours involved cultural given that immatures acquire them developmentally through social learning. Primate fieldworkers developed an approach to bolster the cultural interpretation of geographic variation, called the ethnographic, geographic or group-contrast method, or method of elimination (Boesch 1996, Whiten et al. 1999, van Schaik 2003). A behavioural variant is considered cultural if it is highly prevalent wherever it occurs, consistent with its spread and maintenance by social learning, but not clearly linked to ecological differences among the areas or genetic differences among the populations. By concentrating on behavioural variants that do not show clear genetic or ecological correlations in their spatial distribution, researchers could eliminate non-cultural explanations. Together, these field-based, nonexperimental methods have served to establish the plausibility of culture in chimpanzees (Whiten et al. 1999), cetaceans (Rendell and Whitehead 2001), orang-utans (van Schaik et al. 2003), and capuchin monkeys (Perry et al. 2003). This interpretation was consistent with the demonstration of sophisticated and highly reliable observational forms of social learning in apes, especially chimpanzees (Whiten et al. 2005). The major weakness to date of these field methods is that they are useful as proof of principle but not to estimate the size of the cultural repertoire and thus to assess the importance of culture in nature (van Schaik 2009). By design, the geographic exclusion method ignores any behaviour that is correlated with ecological variables or genetic discontinuities, e.g. inclusion in the diet of a particular food item, even if the animals are critically dependent on social learning for their maintenance (cf. Humle and Matsuzawa 2002). On the other hand, if it fails to recognise ecological or demographic factors that underlie the behaviour pattern, which therefore need not be socially transmitted, this technique will overestimate cultural repertoires (Laland et al. 2009). That said, most ape researchers would estimate

640

Carel P. van Schaik

that the 40 or so variants described for chimpanzees and the 30 or so described for orang-utans are likely to be the tip of the cultural iceberg. It is important that additional techniques be developed to recognise cultural variants in the wild. BOX 20.1 gives an overview of the methods available to date. BOX 20.1 Assessing culture Some non-experimental approaches to identify socially transmitted variants, based on observations of spatial patterns (Pattern) or evidence for, or indications of, social learning (Process). Based on Perry (2009), Laland et al. (2009), Whitehead (2009) and van Schaik (2009). Pattern (1) Across sites: the geographic or group-contrast method, assuming no ecological or genetic differences between sites, and assuming withinsite homogeneity of the incidence of the behaviour. A quantitative version turns it into a multiple matrix regression, estimating behavioural similarity as a function of genetic similarity and ecological similarity, either for a particular behaviour pattern or a whole repertoire. (2) Across individuals: multiple matrix regression of behavioural similarity among individuals (of one or more populations) as a function of genetic and ecological similarity and past association (to estimate cumulative opportunities for social learning). (3) Features of behaviours: Identify the socially learned behaviour patterns in an individual’s repertoire, by showing that its emergence is associated with use of opportunities for social learning and practice when non-routine behaviours are performed by models. Criteria include: (1) selective attention, e.g. by peering, suggesting observational social learning; (2) practice subsequent to selective attention; (3) begging and scrounging, suggesting socially induced affordance learning; and (4) in the case of techniques, hand specialisation. Process (1) Documenting ontogeny: Link patterns of association and selective attention over time to the acquisition of a particular variant, whenever there is within-population variation. (2) Documenting spread: Examine the spread of an innovation whose origin was witnessed, against null models of association and rate of spread.

Social learning and culture in animals

641

20.4.2 Mapping variation in animal cultures To properly assess the distribution of culture in nature, i.e. which species have which components of culture as well as how humans differ from other animals we need to go beyond semantic debates and take a broad perspective. In principle we could estimate the strength of three potentially independent dimensions of socially transmitted innovations: (i) the complexity of the innovation; (ii) the complexity of the social-learning techniques required to acquire the behaviour socially; and (iii) the extent to which a geographic imprint ensues. At this stage, a simple dichotomy on each dimension will suffice to frame the approach. Thus, innovations can be cognitively simple or cognitively complex, depending on whether they could have arisen by chance and trial and error or instead required some form of insight to have been performed the first time (cf. Whiten and van Schaik 2007). Likewise, they could have been acquired by simple non-observational forms of social learning or by observational forms. Finally, we can divide cultural variants into those that do produce geographic variation with those that do not, either because they are universal and ubiquitous everywhere or are interspersed with other variants at many different sites (due to exclusive vertical learning). Although this procedure still leaves eight possible states, in practice the number of observed states will be much smaller. As was already noted implicitly, the complexity of innovation and social learning will tend to be correlated, because cognitive abilities limit the complexity of both innovation and social learning, and thus the content and extent of cultural repertoires. The variants that are easily innovated and transmitted through simple mechanisms such as social facilitation or stimulus or local enhancement may be most widespread. They may also be linked to the third dimension because where innovations arise so easily we may not find any geographic variation. As the cognitive complexity of the innovation increases, more dedicated mechanisms of social learning are required for social transmission. These make the most interesting cultural variants, because social transmission is likely to be essential for their spread and maintenance, and geographic variation almost inevitably arises. This proposal can be linked to the content-based classification proposed earlier (Fig. 20.2). Perishable information cannot produce cultures. Labels are usually simple innovations requiring simple social learning mechanisms and also need not produce geographic variation. Most cases of vertical social transmission in nature may concern the faster than usual acquisition during development of species-wide foraging patterns or predator recognition, i.e. labels (cf. Galef and Giraldeau 2001). In most species, the

642

Carel P. van Schaik

Geographic variation

Present

Absent

Signals

Labels

Cumulative skills

Symbols

Skills CULTURES

(Label variants)

Most labels

CUMULATIVE CULTURES

3

Exclusively vertically transmitted skills 2

1

Cognitive complexity (of innovation and social learning)

Fig. 20.2 The dimensions of tradition and culture.

learning processes involved need not be any different from the ones an animal would use when alone. Still, label variants may arise, as when different easily recognised foods are included in the diet in different places, although such cases will probably be rare. Skills, signal variants and symbols can more readily produce cultures. Whenever the cultural variant requires innovation, geographic variation is likely, at least for a certain amount of time. Well-known examples are the cultures found among great apes (including humans) and cetaceans, but also those of capuchin monkeys or rats (Terkel 1996), even though in these cases the social-learning mechanism may involve no more than enhancement. One extension immediately comes to mind. Cumulative culture refers to innovations that are beyond the reach of individual inventors and arose through the step-by-step accumulation of modifications to the variant that improved its function but moved it further and further away from the original innovation. These skills or social rules could be called cumulative innovations, which are one step up from complex innovations, and may rely on both production imitation and teaching to be transmitted. This gives us a three-step scale of complexity of innovation and social learning (Fig. 20.2). The value of scheme like that of Fig. 20.2 is heuristic, in that it should serve to focus the debate on the relevant aspects of the whole phenomenon of culture rather than on terminology. For instance, why are there no examples of cultures from such well-studied animals such as baboons,

Social learning and culture in animals

643

horses, great tits or sticklebacks? The scheme suggests that many of the culturally transmitted innovations in those taxa are labels that leave no geographic imprint. If new work necessitates its revision it will have served its purpose.

20.5 Long-term consequences: culture and intelligence 20.5.1 Culture and intelligence Most developing animals need environmental inputs to fine-tune their brains so as to acquire the behavioural repertoire most appropriate in the current conditions, a phenomenon called constructive learning (Quartz 2003). Play is an example of active attempts to generate such inputs that serve to train the motor system (Fairbanks 2000). When it comes to acquiring knowledge or specific skills, social learning should improve the signalto-noise ratio of these inputs because animals learn skills faster when learning socially than when learning (i.e. generating sensory inputs) alone. Indeed, there are clear developmental effects of social learning on an individual’s repertoire of cognitive skills, as shown by comparisons among wild great ape populations (van Schaik et al. 2003) as well as the results of social deprivation and enculturation experiments (van Schaik and Burkart unpubl. data). Thus, if there is a niche-construction element to an individual’s set of learned skills, and if this set approximates intelligence, intelligence is culturally constructed (Tomasello 1999). To the extent that this ontogenetic account is correct, culture should in the long run affect intelligence. Species with frequent opportunities for social learning should more readily respond to selection pressures to add cognitive abilities, so that they should end up with larger relative brain size on average than species that have fewer opportunities for social learning. The cultural intelligence hypothesis (van Schaik 2004) therefore predicts a tight positive interspecific correlation between social learning performance and individual learning ability. There is evidence for this at the level of frequencies of innovation and social learning (Reader and Laland 2002) and at the level of level of maximum complexity achieved (van Schaik and Burkart unpubl. data). It also predicts that cultural species are more intelligent, but there is not enough comparative information to test this prediction yet. Thus, culture could provide a compelling explanation for the evolution of intelligence, complementing accounts that focus exclusively on the fitness benefits of enhanced cognitive abilities, such as the social brain or technical intelligence hypotheses (Byrne and Bates 2007).

644

Carel P. van Schaik

20.5.2 Cultural evolution Cultural transmission systems have some properties that are fundamentally different from regular organic evolution, as summarised in Table 20.4 (after Danchin et al. 2004). Whereas mutations are random with respect to fitness, most innovations are not, but instead improve fitness. Moreover, the transmission of favourable variants, slow and noisy in organic evolution since it has to rely on natural selection of individuals across generations, can be fast and precise in the case of social transmission. Cultural selection is therefore truly Lamarckian, and this new system of inheritance can in principle be far more efficient than classic natural selection. It would appear to be the silver bullet toward almost instantaneous adaptation to local conditions, and thus highly adaptive for species that colonise new habitats or encounter temporal changes in their habitats – in other words, pretty much all species. Why, then, is cultural evolution so rare in nature? We have seen that many species rely on social learning, yet few have traditions and even fewer have cultures, as defined here, and only humans have truly cumulative culture. Thus, the features of this table only apply to very few species, and in its more extreme forms only to humans. Culture is almost synonymous with being human, and may be our most successful adaptation. One reason for this is that the proper adoption of innovations, i.e. potentially novel cultural variants, requires the cognitively most complex forms of observational learning along with teaching. This combination may be limited to humans. The absence of cumulative culture in great apes or other species capable of observational learning (Whiten et al. 2004) suggests that some factor was added during human evolution. One recent hypothesis is that the proactive prosociality induced by the adoption of cooperative breeding during human evolution served to add both the

Table 20.4 The main distinguishing features of regular (organic) and cultural evolution. Organic evolution (Darwinian)

Cultural evolution (Lamarckian)

Raw material

Mutation (non-directed, usually deleterious)

Innovation (usually directed, rarely deleterious)

Spread of favorable variants

through differential lifetime reproductive success across many generations

through adoption, potentially within a single generation

Social learning and culture in animals

645

teaching component and, in the long run, the increased cognitive demands of observational social learning (Burkart et al. 2009a). 20.5.3 Toward a unified science of culture Cultures can be dynamic, changing continuously due to innovation, forgetting, dropping out of fashion, demographic changes, etc. Cultures can remain stagnant due to lack of innovation, but even when innovations are produced they may not spread due to social conformity or conservatism. Thus, a mature theory of culture should contain the following elements:  The rate of innovation: One obvious factor affecting this rate is population size (Henrich 2004), but this will be affected by the degree of conservatism, which in turn may depend on the presence of environmental crises, such as famine or war, and perhaps on demographic factors.  The nature of social transmission mechanisms: The mechanism may critically affect the spread of innovations and the reliability of their transmission (Boyd and Richerson 1996).  Biases in the transmission of innovations: Strong conformity will tend to slow down cultural change, whereas prestige-based copying will tend to speed it up. The more individual evaluation of novel variants encountered by a naïve individual takes place, the slower cultural change (Richerson and Boyd 2005). Demographic factors, such as the presence of a large bulge of adolescent, more willing than others to try out new innovations, may favour their spread.  Conditions affecting transmission: The effective size of the social network plays a major role, which is strongly affected by association patterns and social tolerance among individuals (e.g. van Schaik and Pradhan 2003, Henrich 2004), as should social relations between social units. An additional factor is the intrinsic salience of the innovation, which is partly a function of its domain. For instance, orang-utans perceive comfort innovations as much less salient than those linked to feeding or communication (van Schaik et al. 2006).  Conditions affecting diffusion: The degree to which separate populations are open to diffusion of variants from others may be especially important where populations are small and poorly networked, and variants are therefore prone to extinction. The presence of dispersal barriers (e.g. due to habitat fragmentation), the sex of the dispersers, and the degree to which dispersers are integrated socially all play a role.  Tradeoffs between innovative and social-learning ability: Exclusive reliance on social learning may interfere with directing attention toward

646

Carel P. van Schaik

individual exploration, and thus solving problems. For instance, Burkart et al. (2009b) showed that those individual marmosets (Callithrix jacchus) that were better at socially learning a new foraging task than others were less quick to notice the presence of a hitherto unavailable shortcut to food. Thus, in a population of social-learning specialists, culture may get stuck. However, the presence of different personality types within a population may remove this obstacle to cumulative cultural change. Such a mature theory of culture would not only explain the taxonomic variation in the presence of culture, repertoire sizes within and across species, or the longevity of individual traditions, but also the relative role of culture in such well-known human phenomena as the upper Paleolithic revolution (Powell et al. 2009) or the adoption of agriculture.

20.6 Conclusions and outlook Every developing animal must acquire a number of skills to be successful in its habitat and social organisation through learning, although species may differ dramatically in the actual number of learned skills. Most come equipped with a set of near-instinctive response predispositions, which may be honed by social learning, and thus produce local diets and niches. Where studied, immatures have a preference for social over individual learning, and adults wherever possible make widespread use of eavesdropping or public information. Moreover, several individuals learn complex skills they would otherwise almost certainly not have acquired. Still, cultural evolution in nature seems largely confined to humans. Nonetheless, the reader must have noted from the way findings or patterns were formulated that much of our knowledge in this young field is still tentative. At the same time, it is clear that an improved understanding of cultural phenomena in animals is vital for evolutionary biology and especially evolutionary anthropology. Theorising has run well ahead of empirical research. One important avenue for future research is to unravel the relative importance of genetic endowment, ecological conditions and opportunities for social learning in bringing about an individual’s behavioural repertoire, and to compare the picture across a wide array of species. Tools for this task are as yet inadequate, but they are being developed (Laland et al. 2009; see also BOX 20.1). This analysis needs to be done in a range of species, varying from those with suspected cultural variation, such as great

Social learning and culture in animals

647

apes and cetaceans, to species for which such variation is thought to be absent, such as many birds and fishes. Another very important issue is innovation. Little is known about what produces innovation, probably because it is a by-product of individual exploration (perhaps related to personality: Bergmüller this volume). Yet, the presence of culture relies on it (Fig. 20.2), and cumulative culture is critically dependent on continuing innovation. In order to get some purchase on the role of innovation in the production of animal culture, solid operational definitions are needed. These are being developed (Reader and Laland 2003, Ramsey et al. 2007) and tested (van Schaik et al. 2006, Lehner et al. in press). I expect that a good theory for the conditions that favour or hinder the incidence of innovation and of their spread and maintenance through social learning will contribute to solving the puzzle of the limited taxonomic distribution of culture and the near-absence of cumulative cultural evolution in animals. GLOSSARY Affordances: Learned action possibilities, i.e. properties of objects or environments that can be learned about by exploration, and then be used in subsequent behavioural actions. Culture: In the broadest sense, culture is used as socially transmitted information, in the narrower one as socially transmitted technology and norms that involve conformity (‘the way we do things here’ McGrew 2004). Whiten and van Schaik (2007) separate it from tradition by noting that culture tends to contain multiple traditions in multiple domains. Model: The knowledgeable individual that may (passively) allow a naïve individual to acquire a particular skill or information, or (actively) pass it on through teaching. Social learning: Changes in an individual’s behaviour resulting from attending to another individual’s behaviour or its products. Tradition: An enduring distinctive behaviour patterns characteristic of social units and passed on through social learning (Fragaszy and Perry 2003). Teaching: Teaching has the following characteristics: (i) certain actions only occur in the presence of a naïve observer, (ii) these actions carry some cost but are not immediately beneficial to the teacher, and (iii) the observer profits from these actions in that it learns these actions faster than it would otherwise do.

648

Carel P. van Schaik

Acknowledgements The work on which this chapter is based was supported by the A.H. Schultz Foundation and the Schweizerischer Nationalfonds (31003A111915). I thank Adrian Jaeggi, Rachel Kendal, Peter Kappeler and Maria van Noordwijk for useful comments, and Judith Burkart and Christine Hrubesch for discussion.

References Aunger R (2007) Memes. In: Dunbar RIM, Barrett L (eds) Oxford Handbook of Evolutionary Psychology. Oxford University Press, Oxford, pp 599-604 Biro D, Inoue-Nakamura N, Tonooka R, Yamakoshi G, Sousa C, Matsuzawa T (2003) Cultural innovation and transmission of tool use in wild chimpanzees: evidence from field experiments. Anim Cogn 6:213-223 Boesch C (1996) The emergence of cultures among wild chimpanzees. Proc Brit Acad 88:251-268 Bonnie KE, Earley RL (2007) Expanding the scope for social information use. Anim Behav 74:171-181 Box HO (1984) Primate Behavior and Social Ecology. Chapman and Hall, London Boyd R, Richerson PJ (1985) Culture and the Evolutionary Process. University of Chicago Press, Chicago Boyd R, Richerson PJ (1996) Why culture is common, but cultural evolution is rare. Proc Brit Acad 88:77-93 Burkart JM, van Schaik CP (2010) Cognitive consequences of cooperative breeding in primates. Anim Cogn 13:1-19 Burkart JM, Hrdy SB, van Schaik CP (2009a) Cooperative breeding and human cognitive evolution. Evol Anthropol 18:175-186 Burkart JM, Foglia M, Strasser A (2009b) Trade-offs between social learning and individual innovativeness in common marmosets, Callithrix jacchus. Anim Behav 77:1291-1301 Byrne RW (2002) Imitation of novel complex actions: what does the evidence from animals mean? Adv Stud Behav 31:77-105 Byrne RW, Bates LA (2007) Sociality, evolution and cognition. Curr Biol 17:R714-R723 Caro TM, Hauser MD (1992) Is there teaching in nonhuman animals? Q Rev Biol 67:151-174 Dabelsteen T (2005) Public, private or anonymous? Facilitating and countering eavesdropping. In: McGregor PK (ed) Animal Communication Networks. Cambridge University Press, Cambridge, pp 38-62 Danchin E, Giraldeau L-A, Valone TJ, Wagner RH (2004) Public information: from nosy neighbors to cultural evolution. Science 305:487-491

Social learning and culture in animals

649

Dawkins R (1976) The Selfish Gene. Oxford University Press, Oxford Day RL, MacDonald T, Brown C, Laland KN, Reader SM (2001) Interactions between shoal size and conformity in guppy social foraging. Anim Behav 62:917-925 de Waal FBM (2001) The Ape and the Sushi Master: Cultural Reflections of a Primatologist. Basic Books, New York de Waal FBM, Johanowicz DL (1993) Modification of reconciliation behavior through social experience: an experiment with two macaque species. Child Dev 64:897-908 Durham WH (1991) Coevolution: Genes, Culture, and Human Diversity. Stanford University Press, Stanford/CA Efferson C, Lalive R, Richerson PJ, McElreath R, Lubell M (2008) Conformists and mavericks: the empirics of frequency-dependent cultural transmission. Evol Hum Behav 29:56-64 Fairbanks LA (2000) The developmental timing of primate play. A neural selection model. In: Parker ST, Langer J, McKinney ML (eds) Biology, Brains, and Behavior: The Evolution of Human Development. School of American Research Press, Santa Fe, pp 131-158 Fragaszy DM, Perry S (2003) Towards a biology of traditions. In: Fragaszy DM, Perry S (eds) The Biology of Traditions: Models and Evidence. Cambridge University Press, Cambridge, pp 1-32 Galef BG Jr (1995) Why behaviour patterns that animals learn socially are locally adaptive. Anim Behav 49:1325-1334 Galef BG Jr, Giraldeau L-A (2001) Social influences on foraging in vertebrates: causal mechanisms and adaptive functions. Anim Behav 61:3-15 Giraldeau L-A, Valone TJ, Templeton JJ (2002) Potential disadvantages of using socially acquired information. Philos Trans R Soc Lond B 357:1559-1566 Hardus ME, Lameira AR, van Schaik CP, Wich SA (2009) Tool use in wild orang-utans modifies sound production: a functionally deceptive innovation? Proc R Soc Lond B 276:3689-3694 Harlow HF, Harlow MK (1962) Social deprivation in monkeys. Sci Am 207:136146 Henrich J (2004) Demography and cultural evolution: how adaptive cultural processes can produce maladaptive losses – the Tasmanian case. Am Antiq 69:197-214 Henrich J, McElreath R (2007) Dual-inheritance theory: the evolution of human cultural capacities and cultural evolution. In: Dunbar RIM, Barrett L (eds) Oxford Handbook of Evolutionary Psychology. Oxford University Press, Oxford, pp 555-570 Heyes CM (2001) Causes and consequences of imitation. Trends Cogn Sci 5:253261 Hoppitt WJE, Laland KN (2008) Social processes influencing learning in animals: a review of the evidence. Adv Stud Behav 38:105-165 Hoppitt WJE, Brown GR, Kendal R, Rendell L, Thornton A, Webster MM, Laland KN (2008) Lessons from animal teaching. Trends Ecol Evol 23:486-493

650

Carel P. van Schaik

Horner V, Whiten A (2005) Causal knowledge and imitation/emulation switching in chimpanzees (Pan troglodytes) and children (Homo sapiens). Anim Cogn 8:164-181 Hrubesch C, Preuschoft S, van Schaik CP (2009) Skill mastery inhibits adoption of observed alternative solutions among chimpanzees (Pan troglodytes). Anim Cogn 12:209-216 Humle T, Matsuzawa T (2002) Ant-dipping among chimpanzees of Bossou, Guinea, and some comparisons with other sites. Am J Primatol 58:133-148 Jaeggi AV, van Noordwijk MA, van Schaik CP (2008) Begging for information: mother-offspring food sharing among wild Bornean orangutans. Am J Primatol 70:533-541 Jaeggi AV, Dunkel LP, van Noordwijk MA, Wich SA, Sura AAL, van Schaik CP (2009) Social learning of diet and foraging skills by wild immature Bornean orangutans: implications for culture. Am J Primatol 72:62-71 Kokko H, Johnstone RA, Clutton-Brock TH (2001) The evolution of cooperative breeding through group augmentation. Proc R Soc Lond B 268:187-196 Krakauer EB (2005) Development of Aye-Aye (Daubentonia madagascariensis) foraging skills: independent exploration and social learning. PhD thesis. Duke University, Durham/NC Krützen M, Mann J, Heithaus MR, Connor RC, Bejder L, Sherwin WB (2005) Cultural transmission of tool use in bottlenose dolphins. Proc Natl Acad Sci USA 102:8939-8943 Laland KN (2004) Social learning strategies. Learn Behav 32:4-14 Laland KN, Kendal RL, Kendal JR (2009) Animal culture: problems and solutions. In: Laland KN, Galef BG Jr (eds) The Question of Animal Culture. Harvard University Press, Cambridge/MA, pp 174-197 Lefebvre L, Palameta B, Hatch KK (1996) Is group-living associated with social learning? A comparative test of a gregarious and territorial columbid. Behaviour 133:241-261 Lehner SR, Burkart JM, van Schaik CP (in press) An evaluation of the geographic method for recognizing innovations in nature, using zoo orangutans. Primates Lyons DE, Young AG, Keil FC (2007) The hidden structure of overimitation. Proc Natl Acad Sci USA 104:19751-19756 Mann J, Sargent B (2003) Like mother, like calf: the ontogeny of foraging traditions in wild Indian Ocean bottlenose dolphins (Tursiops sp.). In: Fragaszy D, Perry S (eds) The Biology of Traditions: Models and Evidence. Cambridge University Press, Cambridge, pp 236-266 Marshall-Pescini S, Whiten A (2008) Chimpanzees (Pan troglodytes) and the question of cumulative culture: an experimental approach. Anim Cogn 11:449-456 Masur EF (1988) Infants’ imitation of novel and familiar behaviors. In: Zentall TR, Galef BG Jr (eds) Social Learning: Psychological and Biological Perspectives. Erlbaum, Hillsdale/NJ, pp 301-318

Social learning and culture in animals

651

Matsuzawa T, Biro D, Humle T, Inoue-Nakamura N, Tonooka R, Yamakoshi G (2001) Emergence of culture in wild chimpanzees: education by masterapprenticeship. In: Matsuzawa T (ed) Primate Origins of Human Cognition and Behavior. Springer, Tokyo, pp 557-574 Matthews LJ (2009) Intragroup behavioral variation in white-fronted capuchin monkeys (Cebus albifrons): mixed evidence for social learning inferred from new and established analytical methods. Behaviour 146:295-324 McGrew WC (2004) The Cultured Chimpanzee: Reflections on Cultural Primatology. Cambridge University Press, Cambridge Meltzoff AN, Kuhl PK, Movellan J, Sejnowski TJ (2009) Foundations for a new science of learning. Science 325:284-288 Mery F, Kawecki TJ (2004) An operating cost of learning in Drosophila melanogaster. Anim Behav 68:589-598 Mettke-Hoffman C, Rowe KC, Hayden TJ, Canoine V (2006) Effects of experience and object complexity on exploration in garden warblers (Sylvia borin). J Zool 268:405-413 Ottoni E, de Resende B, Izar P (2005) Watching the best nutcrackers: what capuchin monkeys (Cebus apella) know about others’ tool-using skills. Anim Cogn 8:215-219 Page RA, Ryan MJ (2006) Social transmission of novel foraging behavior in bats: frog calls and their referents. Curr Biol 16:1201-1205 Patricelli GL, Dantzker MS, Bradbury JW (2007) Differences in acoustic directionality among vocalizations of the male red-winged blackbird (Agelaius pheoniceus) are related to function in communication. Behav Ecol Sociobiol 61:1099-1110 Pellegrini AD, Dupuis D, Smith PK (2007) Play in evolution and development. Dev Rev 27:261-276 Perry S (2009) Are nonhuman primates likely to exhibit cultural capacities like those of humans? In: Laland KN, Galef BG Jr (eds) The Question of Animal Culture. Harvard University Press, Cambridge/MA, pp 247-268 Perry S, Ordoñez Jiménez JC (2006) The effects of food size, rarity, and processing complexity on white-faced capuchins’ visual attention to foraging conspecifics. In: Hohmann G, Robbins MM, Boesch C (eds) Feeding Ecology in Apes and Other Primates. Cambridge University Press, Cambridge, pp 203234 Perry S, Panger M, Rose LM, Baker M, Gros-Louis J, Jack K, MacKinnon KC, Manson J, Fedigan L, Pyle K (2003) Traditions in wild white-faced capuchin monkeys. In: Fragaszy DM, Perry S (eds) The Biology of Traditions: Models and Evidence. Cambridge University Press, Cambridge, pp 391-425 Powell A, Shennan S, Thomas MG (2009) Late Pleistocene demography and the appearance of modern human behavior. Science 324:1298-1301 Prather JF, Peters S, Nowicki S, Mooney R (2008) Precise auditory-vocal mirroring in neurons for learned vocal communication. Nature 451:305-310

652

Carel P. van Schaik

Quartz S (2003) Toward a developmental evolutionary psychology: genes, development and the evolution of cognitive architecture. In: Scher SJ, Rauscher M (eds) Evolutionary Psychology: Alternative Approaches. Kluwer, Boston, pp 185-210 Ramsey G, Bastian ML, van Schaik CP (2007) Animal innovation defined and operationalized. Behav Brain Sci 30:393-437 Rapaport LG, Brown GR (2008) Social influences on foraging behavior in young nonhuman primates: learning what, where, and how to eat. Evol Anthropol 17:189-201 Reader SM, Laland KN (2002) Social intelligence, innovation, and enhanced brain size in primates. Proc Natl Acad Sci USA 99:4436-4441 Reader SM, Laland KN (2003) Animal Innovation. Oxford University Press, New York Rendell L, Whitehead H (2001) Culture in whales and dolphins. Behav Brain Sci 24:309-382 Richerson PJ, Boyd R (2005) Not by Genes Alone: How Culture Transformed Human Evolution. University of Chicago Press, Chicago Rizzolatti G (2005) The Mirror Neuron System and Imitation. MIT Press, Cambridge/MA Rogers AR (1988) Does biology constrain culture? Am Anthropol 90:819-831 Rowley I, Chapman G (1986) Cross-fostering, imprinting and learning in two sympatric species of cockatoo. Behaviour 96:1-16 Seyfarth RM, Cheney DL (1980) The ontogeny of vervet monkey alarm calling behavior: a preliminary report. Z Tierpsychol 54:37-56 Slagsvold T, Wiebe KL (2007) Learning the ecological niche. Proc R Soc Lond B 274:19-23 Subiaul F (2007) The imitation faculty in monkeys: evaluating its features, distribution and evolution. J Anthropol Sci 85:35-62 Tarnaud L, Yamagiwa J (2008) Age-dependent patterns of intensive observation on elders by free-ranging juvenile Japanese macaques (Macaca fuscata yakui) within foraging context on Yakushima. Am J Primatol 70:1103-1113 Terkel J (1996) Cultural transmission of feeding behavior in the black rat (Rattus rattus). In: Heyes CM, Galef BG Jr (eds) Social Learning in Animals: The Roots of Culture. Academic Press, San Diego, pp 17-47 Thornton A (2008) Social learning about novel foods in young meerkats. Anim Behav 76:1411-1421 Tomasello M (1999) The Cultural Origins of Human Cognition. Harvard University Press, Cambridge/MA Ueno A, Matsuzawa T (2005) Response to novel food in infant chimpanzees: do infants refer to mothers before ingesting food on their own? Behav Proc 68:85-90 Valone TJ (2007) From eavesdropping on performance to copying the behavior of others: a review of public information use. Behav Ecol Sociobiol 62:1-14 van Bergen Y, Coolen I, Laland KN (2004) Nine-spined sticklebacks exploit the most reliable source when public and private information conflict. Proc R Soc Lond B 271:957-962

Social learning and culture in animals

653

van der Post DJ, Hogeweg P (2006) Resource distributions and diet development by trial-and-error learning. Behav Ecol Sociobiol 61:65-80 van Schaik CP (2003) Local traditions in orangutans and chimpanzees: social learning and social tolerance. In: Fragaszy DM, Perry S (eds) The Biology of Traditions: Models and Evidence. Cambridge University Press, Cambridge, pp 297-328 van Schaik CP (2004) Among Orangutans: Red Apes and the Rise of Human Culture. Harvard University Press, Belknap, Cambridge/MA van Schaik CP (2009) Geographic variation in the behavior of wild great apes: is it really cultural? In: Laland KN, Galef BG Jr (eds) The Question of Animal Culture. Harvard University Press, Cambridge/MA, pp 70-98 van Schaik CP, Pradhan GR (2003) A model for tool-use traditions in primates: implications for the evolution of culture and cognition. J Hum Evol 44:645664 van Schaik CP, Ancrenaz M, Borgen G, Galdikas BFM, Knott CD, Singleton I, Suzuki A, Utami SS, Merrill MY (2003) Orangutan cultures and the evolution of material culture. Science 299:102-105 van Schaik CP, van Noordwijk MA, Wich SA (2006) Innovation in wild Bornean orangutans (Pongo pygmaeus wurmbii). Behaviour 143:839-876 Whitehead H (2009) How might we study culture? A perspective from the ocean. In: Laland KN, Galef BG Jr (eds) The Question of Animal Culture. Harvard University Press, Cambridge/MA, pp 125-151 Whiten A, van Schaik CP (2007) The evolution of animal ‘cultures’ and social intelligence. Philos Trans R Soc Lond B 362:603-620 Whiten A, Goodall J, McGrew WC, Nishida T, Reynolds V, Sugiyama Y, Tutin CEG, Wrangham RW, Boesch C (1999) Cultures in chimpanzees. Nature 399:682-685 Whiten A, Horner V, Litchfield CA, Marshall-Pescini S (2004) How do apes ape? Learn Behav 32:36-52 Whiten A, Horner V, de Waal FBM (2005) Conformity to cultural norms of tool use in chimpanzees. Nature 437:737-740 Witte K (2006) Learning and mate choice. In: Brown C, Laland KN, Krause J (eds) Fish Cognition and Behaviour. Blackwell, Oxford, pp 70-95 Zentall TR (2004) Action imitation in birds. Learn Behav 32:15-23

Chapter 21

Levels and mechanisms of behavioural variability PETER M. KAPPELER AND CORNELIA KRAUS

ABSTRACT The behaviour of animals is the result of adaptations and constraints. To what extent a particular behaviour pattern can be attributed to the relative strengths of these two forces is crucial with respect to understanding the evolution of behaviour. Moreover, in the still young history of the study of animal behaviour, different conceptual approaches have placed very different emphases on the residual behavioural variability beyond adaptations and constraints. In this chapter, we retrace some of these paradigm shifts, offer an overview of different hierarchical levels at which behavioural variability occurs and summarise some of the mechanisms that generate or constrain it. Above the species level, phylogenetic constraints often limit behavioural variability because of a functional relationship between taxonwide life history traits and behaviour, but the exact nature of their underlying mechanisms remains obscure. Phylogenetic constraints exist at different taxonomic levels, and, as several examples from studies of primate behaviour indicate, they are also common in animals with relatively advanced cognitive abilities in which social learning is common. We therefore emphasise the importance of acknowledging the existence of such constraints in behavioural analyses. We also decompose behavioural variability further into variation within species, among individuals and within individuals over time and highlight some of the mechanisms responsible for generating and maintaining this variability. Our review suggests that the specific evolutionary history of a taxon will set the stage at which levels variability can arise, and that cognitive abilities appear to create surprisingly little additional freedom for behavioural variability.

656

Peter M. Kappeler and Cornelia Kraus

21.1 Introduction Evolutionary analyses of behaviour have either examined its adaptive nature as a result of natural, sexual and kin selection, or they have investigated the effects of behaviour on evolutionary processes. With respect to the latter approach, behaviour can sometimes be seen as a pacemaker of evolutionary change because behavioural variability can expose individuals to novel selection pressures, resulting in rapid subsequent adaptations of morphological, physiological and life history traits. Alternatively, behavioural variability can be interpreted as shielding individuals from strong directional selection pressures by allowing them to evade unfavourable conditions (Huey et al. 2003, Duckworth 2009, Kearney et al. 2009). In both types of processes, behavioural variability plays a key role – either as the outcome of adaptations or as a mechanism mediating adaptations of other traits. Behavioural variability has been subject to one of the most remarkable paradigm shifts in the study of animal behaviour. Over the past five decades, our field has witnessed a shift in emphasis from considering behavioural patterns (or ‘innate instincts’) as invariant species-specific traits (Tinbergen 1951) to a recent appreciation of inter- and intra-population (Laland and Janik 2006, van Schaik this volume) and even inter-individual (Sih et al. 2004, Bell 2007, Bergmüller this volume) variation. The intervening years have seen a similar shift; first from the classical ethologists’ perspective towards behavioural ecology’s focus on behavioural adaptations to local conditions and the acknowledgement of individual behavioural strategies (Krebs and Davies 1981, Maynard Smith 1974) to the subsequent recognition that behavioural variability can be constrained by phylogeny (Brooks and McLennan 1991, Harvey and Pagel 1991). In this chapter, we retrace some of these developments, focusing on different levels and sources of behavioural variability as well as the nature and impact of various constraints in behavioural evolution. Our general aim is to complement some of the previous chapters (by Bergmüller, van Schaik, Sachser and Kaiser, Taborsky and Brockman) on behavioural variability to emphasise how more or less independent research programmes on relevant topics revolving around each of Tinbergen’s four questions are conceptually linked. Our specific aims are to emphasise the different levels at which behavioural variation is observed, to summarise the main factors that can influence it and to discuss the underlying mechanisms, because a broad appreciation of these aspects of variability is central to an understanding of the control and function of behaviour, independent of the current conceptual paradigm.

Levels and mechanisms of behavioural variability

657

21.1.1 A historical perspective on behavioural variability ‘Classical’ ethology, as a new biological discipline, was initially faced with the formidable challenge of generating detailed descriptions of the behavioural repertoires of many different species. Species-specific ethograms (see also Glossary for definition of other key terms) provided inventories of innate behaviour patterns, naturally focusing on fine details and their careful description (Tinbergen 1951). The pioneers of behavioural analyses were impressed by the occurrence of recurring, largely invariant behavioural elements in all members of a species, irrespective of their variable ecological backgrounds (Franck 1997). It was therefore not surprising that one application of ethology advocated the use of such instincts in phylogenetic analyses (e.g., Lorenz 1941; see also Wenzel 1992, Blackledge et al. 2009). Being keen observers and ardent naturalists, however, early ethologists were not unaware of intra-specific variation in behaviour, but they were content to describe and classify it as variation among subspecies. Behavioural ecology, whose beginnings were influenced importantly by comparative studies (Winn 1958, Crook 1964, Crook and Gartlan 1966, Jarman 1974), shifted the focus of explanations of behaviour from proximate to ultimate aspects, initially by attempting to correlate species differences in behaviour with differences in ecological factors, such as habitat characteristics, food distribution or predation risk. The central paradigm of behavioural ecology therefore posited that much of the behaviour of animals is expected to reflect adaptations to local ecological and social conditions. Accordingly, through the action of natural and sexual selection, favourable behavioural phenotypes evolve towards a spatial and temporal optimum, thereby also generating intra-specific variation if members of a species inhabit variable environments. In addition, behavioural ecology also introduced individual cost-benefit analyses, optimality models and game theory with its evaluation of individual strategies and tactics, and, hence, an additional level of variation, to the study of behaviour (MacArthur and Pianka 1966, Maynard Smith 1974, 1979, Krebs et al. 1977, Gross 1996). Thus, behavioural ecology fostered an appreciation of intraspecific variation of behaviour that charted new territory in the analysis of animal behaviour. In their famous critique of the ‘adaptationist programme’ Gould and Lewontin (1979) called for the equal consideration of alternative, nonadaptive explanations of biological phenomena such as phylogenetic, developmental or allometric constraints of the ‘Bauplan’ of an organism. Indeed, in the early days of behavioural ecology, the focus on ultimate functions of behaviour came at the expense of studying Tinbergen’s three other

658

Peter M. Kappeler and Cornelia Kraus

questions (Krebs and Davies 1997), even though students of animal behaviour had been long aware of the fact that explanations of a particular behaviour should also include reference to the species’ phylogenetic history (Tinbergen 1963). Accordingly, because selection can only lead to evolutionary change if traits are genetically based, closely related species are expected to share more traits than more distantly-related ones. This similarity is due to the fact that fundamental adaptations tend to get canalised over evolutionary times into reaction norms of variable breadth. For example, some behavioural traits are functionally closely related to particular life history traits and conserved across the members of a particular lineage as part of their Bauplan. Methodological advances in comparative analyses in the 1990s began providing the necessary tools for quantifying these phylogenetic constraints (Harvey and Pagel 1991) and for reconstructing the (co-)evolution of character states (Maddison 1990). These methods helped initiate a wealth of studies exploring phylogenetic resilience and coevolution among behavioural traits above the species level (e.g., Brooks and Mc Lennan 1991, Blomberg et al. 2003). More recently, the focus of many behavioural analyses has shifted to the nature and sources of inter-individual variation. Behavioural syndromes, in particular, have received much recent attention from students of behaviour, evolution, ecology and development alike (Sih et al. 2004, Réale et al. 2007). Behavioural syndromes refer to the correlation between rank-order differences in behaviour patterns between individuals of a given population across time and situations (Bell 2007). Accordingly, individuals within populations can be described as behavioural types exhibiting particular combinations of two or more behavioural traits. This focus on consistent inter-individual differences across contexts and situations has generated considerable interest in fundamental questions about animal behaviour, including the proximate genetic and developmental mechanisms underlying behavioural types and the nature of behavioural syndromes and factors affecting their long-term dynamics. Studies of a few wild vertebrate populations revealed that behavioural types exhibit differential survival under changing ecological conditions (Dingemanse and Réale 2005, Smith and Blumstein 2008), highlighting the ecological and evolutionary relevance of inter-individual variation in behaviour. Epigenetic effects have recently been recognised as an additional source of inter-individual variation in behaviour (Laviola et al. 2003). Epigenetic effects provide a mechanism for generating behavioural diversity by linking environmental stimuli during early ontogeny with stable modifications of gene expression patterns and phenotypes without affecting the underlying genotype (see also Sachser and Kaiser this volume), but studies of these mechanisms have so far been largely limited to a few

Levels and mechanisms of behavioural variability

659

laboratory model species. These recent developments indicate that questions addressing levels and mechanisms of behavioural variability are key to a comprehensive understanding of behaviour in a Tinbergian sense. They also show that the attention of students of animal behaviour has returned to phenomena below the species-level. In the remaining sections of this chapter, we will use behavioural variability at different hierarchical levels as our organising principle (Table 21.1). We will address variation among and within lineages, among species, populations and individuals as well as within individuals. At each level, we will characterise the type of variability, discuss relevant mechanisms, and present a few illustrating examples. Table 21.1 Overview of hierarchical levels, types, underlying mechanisms and examples of behavioural variability. Flexibility in the generic sense increases from top to bottom. Level of variability

Type of variability

Main mechanisms

Example phenomena

Among lineages

Adaptation

Genetic

 Adaptive radiations  Lineage-specific traits  Species-specific traits  Species-specific traits

Among species Canalisation Among populations

Geneenvironment interactions

 Local adaptations

Among individuals

Plasticity

(reaction norms), (epi-)genetic

 Polymorphism  Personality  Predator-induced plasticity  Parental effects

Within individuals

Flexibility

Environmental

 Social flexibility  Situationdependency  (Social) Learning

Note that gradients from top to bottom within columns are not scaled linearly

660

Peter M. Kappeler and Cornelia Kraus

21.2 Constraints on behavioural variability above the species level Adaptation and constraint represent the two ends of a continuum of biological explanations for individual traits (Stearns 1992). Most traits are probably the result of the mixture of the two, so that the challenge for students of animal behaviour lies in the identification of their relative contribution in explaining a particular behaviour pattern. In order to illuminate effects of constraints on behavioural variability, their exact nature needs to be known. Constraints on behavioural phenotypes can arise at different levels of organismal organisation and integration because the laws of physics and chemistry need to be obeyed and the developmental and functional integrity of a phenotype needs to be maintained. Hormonal control systems, which are highly integrated and interact in complex ways with the nervous system (Goymann and Hofer this volume), are obvious candidates for mediating such constraints, but, upon closer inspection, widespread endocrine systems, such as the hypothalamic-pituitary-gonadal axis or the activational effects of sex steroids, exhibit not only phenotypic, but also enough evolutionary flexibility so as to not impede rapid adaptation to changing environments (Adkins-Regan 2008). Whatever their precise nature, such proximately acting developmental constraints can ultimately be equated with phylogenetic constraints (WestEberhard 2003). Wilson (1975:32) highlighted the effects of these constraints in behavioural evolution and introduced them as ‘phylogenetic inertia’ to sociobiology. According to his definition: ‘Phylogenetic inertia … consists of the deeper properties of the population that determine the extent to which its evolution can be deflected in one direction or another, as well as the amount by which its rate can be speeded or slowed’. Several other meanings and definitions of this concept have since been discussed (summarised in Blomberg and Garland 2002; see also Antonovics and van Tienderen 1991), but today phylogenetic inertia is widely used in behavioural and evolutionary biology as the opposite of adaptation and synonymously with constraint. In other words, all individuals of a particular lineage exhibit or lack a certain (behavioural) trait due to their evolutionary history, and this trait exhibits resilience to adaptive evolutionary change. How can these ‘deeper properties of a population’ be identified? Comparative analyses can help to detect patterns of invariance within a lineage. However, the existence of such a ‘phylogenetic signal’ (sensu Blomberg and Garland 2002), is necessary, but not sufficient to show that the focal trait is indeed – as implied by the term ‘constraint’ – limited somehow in its potential to move toward a more adaptive state. For example, is there

Levels and mechanisms of behavioural variability

661

really a lack of genetic variance or is the trait in question preserved, e.g. by stabilising selection? To infer how the observed patterns are maintained, additional knowledge about the underlying ecology and genetics are required (Lynch 1991). Selection experiments are a powerful tool to tease apart alternative hypotheses, but they are not always feasible. Still, observational data can add valuable evidence; for example, if a trait state is maintained, even though selection pressures seem to have changed considerably (e.g. Sih et al. 2000) or if it is constant across a variety of environmental conditions that should plausibly affect its correlation with fitness (e.g. BOX 21.1). Even if such additional evidence is often circumstantial rather than conclusive, it helps to collect support for the competing hypotheses of adaptation versus constraint. 21.2.1 Life history constraints At a functional level, many phylogenetic constraints on behavioural traits arise as the result of relatively invariant life history traits, which characterise the broad features of an organism’s life cycle and provide one contribution to the ‘deeper properties’ of a population. Among others, these traits describe the organism’s size at birth and subsequent growth pattern to sexual maturity at a particular age and size, the number and size of offspring it will produce over a variable number of reproductive events as well as its age- and size-specific mortality patterns. These traits are inter-connected through various trade-offs, vary primarily above the species level (Gittleman 1986, Owens and Bennet 1995, Jeschke and Kokko 2009; but see Reznick et al. 1990), and also place broad constraints on behaviour patterns related to survival, reproduction and parental care. In 1992, Stephen Stearns’ assessment of life history constraints on behaviour was rather reserved: ‘While life history evolution limits the behaviour that can evolve, these limits are not surprising enough to catalyse interest’ (p. 210). While some of these constraints may indeed be trivial, others may be more interesting, and, in our opinion, still underappreciated in the analysis of behaviour. These effects are most striking at higher taxonomic levels and therefore contribute to an understanding of general principles as well as major patterns of behavioural variability. Below, we illustrate this point by highlighting a few examples of how variation in key life history traits predisposes animals to certain behavioural strategies and, hence, constrains their behavioural variability. First, the mode of reproduction, which is deeply engrained in the organismal design of a particular lineage, can have far-reaching consequences for behaviour. Whether a given species has asexual or sexual reproduction,

662

Peter M. Kappeler and Cornelia Kraus

two separate sexes or hermaphrodites, and internal or external fertilisation affects available options for potential mechanisms of sexual selection, the scope for sexual conflict or sex-specific parental care considerably. Consequently, it can explain the corresponding reproductive behaviour of females and males in a basic way when adopting a broad comparative perspective (Anthes this volume). The type of fertilisation, for example, is a good predictor of which sex exhibits parental care in large groups of vertebrates (males in species with external fertilisation), and can therefore offer a basis for explaining important behavioural differences between higher taxa (Gross and Shine 1981, Beck 1998). Similarly, behavioural, physiological and morphological adaptations to sperm competition differ between fishes with internal or external fertilisation (Taborsky 1998, Taborsky and Brockmann this volume). Second, the number and size of offspring is intimately related to patterns of parental care because taxa that produce small litters are much more likely to exhibit parental care than taxa with large numbers of offspring per reproductive event (Clutton-Brock 1991). The number and size of offspring is partly dependent on whether they are produced as eggs or live young. Because eggs and live-born young both facilitate and require different forms of parental care, which are modulated in addition by the presence or absence of endothermy (Case 1978), this contrast explains additional variation in related behavioural patterns at various taxonomic levels. For example, vivipary evolved from ovipary more than 100 times independently in squamate reptiles, once in amphibians, 12 times in bony fishes and 10 times in cartilaginous fishes, and provides a predictor for variation in sex-specific parental care among more closely-related taxa (Goodwin et al. 1998, Reynolds et al. 2002). Moreover, within lineages with a particular reproductive pattern, there can be variation in the developmental state of young at birth or hatching (precocial vs. altricial birds and mammals; Trillmich this volume) or in hatchling and clutch size (Nussbaum 1985), which in turn predict more detailed aspects of parental care at lower taxonomic levels. Third, species differ tremendously in age- and sex-specific mortality patterns, giving rise to variation in longevity, which in turn facilitates or constrains various patterns of social behaviour. Differentiated social relationships and advanced cognitive abilities (Silk 2007) and social learning (Laland 2004) are all more likely to develop in longer-lived organisms. Similarly, stable cooperation through reciprocation is more likely when individuals have the opportunity to interact repeatedly (Trivers 1971, Bshary this volume). Increased longevity also favours certain cognitive abilities, such as remembering the location of crucial resources during rare ecological crises (Kerth this volume). The existence of age-dependent reproduc-

Levels and mechanisms of behavioural variability

663

tive tactics is also much more common in species that live long enough to experience several successive reproductive events (Taborsky and Brockman this volume). A strong correlate of life history variation is body size, which is also conserved at higher taxonomic levels (Jeschke and Kokko 2009). Thus, members of a particular lineage, such as a genus or family, are more similar in body size to each other than to members of other lineages (Clauset and Erwin 2008). Among terrestrial mammals, for example, body size is highly conserved among orders and most congeneric species are extremely similar in body size (Smith et al. 2008). Because of various functional and scaling constraints, body size per se has pervasive influence on behaviour. Hence, variation in body size above the species level has consequences for homeostatic behaviour and behavioural aspects of survival, in particular. Thus, how animals forage and what they eat, how they evade or repel predators, or how they fight with each other is fundamentally predicted by their body size (Dial et al. 2008). Body size also predicts behavioural variation at lower taxonomic levels, however. For example, feeding style of African antelopes (Brashares et al. 2000) or mating systems of grouse (Wiley 1974) vary as a function of body size. 21.2.2 Constraints and socio-cognitive complexity One might be tempted to conclude that constraints on behavioural variability are more pronounced in small, short-lived and small-brained animals, for example because they exhibit fewer behavioural innovations (Lefebvre et al. 2004, Pérez-Barberia et al. 2007) or because they are less successful at colonising new habitats (Sol et al. 2008). Because some mammals and birds, such as corvids, cetaceans and primates, exhibit the opposite combination of traits, live in stable complex societies and exhibit a rich repertoire of behavioural innovations as well as striking cognitive capacities (Cheney et al. 1986, Connor et al. 1998, Marino 2002, Emery and Clayton 2004, Kotrschal et al. this volume), they might be expected to be much less constrained in their behavioural variability by phylogenetic constraints than most other animals. Primates, in particular, have been subject to many detailed long-term behavioural studies of known individuals, and primatologists have synthesised this wealth of data into a formidable theoretical framework that explains variation in female social behaviour in groupliving, pair-living and solitary species as a response to variation in food quality and distribution (Wrangham 1980, van Schaik 1989, Sterck et al. 1997, Schülke 2003, Dammhahn and Kappeler 2009, Koenig and Borries 2009). This socio-ecological model, which has also been applied to other

664

Peter M. Kappeler and Cornelia Kraus

BOX 21.1 The lemur syndrome: a phylogenetic constraint on multiple traits? The lemurs of Madagascar (Lemuriformes) represent the endpoints of an adaptive radiation following a single successful colonisation event more than 50 million years ago. The 100 or so members of this radiation include some of the smallest and largest living primates, nocturnal and diurnal forms, they exhibit a broad range of feeding adaptations, a range of slow to fast life histories, and they include solitary, pair-living and group-living species that inhabit a range of habitats, including arid spiny forests and humid rain forests (Fig. 21.1). Despite this variability in ecology, life history and behaviour, all lemurs, including 16 much larger recently extinct species, exhibit a combination of behavioural, demographic, morphological and physiological traits that distinguish them from other primates. Moreover, these traits also deviate from some predictions of evolutionary theory, raising the question as to whether they represent adaptations to current or past selection pressures.

Fig. 21.1 Extant lemurs are grouped into 5 families, which exhibit broad variation in ecological and behavioural traits. Representatives of the 5 families include a) Daubentonia madagascariensis, b) Microcebus berthae, c) Lepilemur ruficaudatus, d) Eulemur fulvus and e) Propithecus verreauxi

The salient behavioural feature of lemurs is the widespread occurrence of female social dominance. Adult females are able to elicit submissive behaviour from all adult males in all social contexts, and males virtually never direct aggressive behaviour towards females (Kappeler 1993). At a proximate level, female dominance is facilitated by a lack of sexual dimorphism in body and canine size, most notably in polygynous species (Kappeler 1990). Androstenedione levels of females are nearly as high as those of males (Drea

Levels and mechanisms of behavioural variability

665

2007), and may proximately contribute to their prowess, as well as to their masculinised genitals, another component of the syndrome. The lack of sexual dimorphism in polygynous species is striking because small female numbers in group-living species, in combination with seasonal, but asynchronous reproduction ought to enhance male monopolisation potential (Kappeler and Schäffler 2008). However, adult sex ratios are commonly even or male-biased. Even more strikingly, male reproductive skew is high, indicating that male reproductive competition is not eliminated (Kappeler and Port 2008). Adaptive explanations of this syndrome have focused on feeding competition, seasonality and climatic unpredictability (Jolly 1984, Wright 1999, Dewar and Richard 2007), but none of them provides a comprehensive explanation and they fail to explain satisfactorily why only lemurs among primates should respond to these ecological factors in this peculiar manner. An alternative explanation includes historic elements and explains the idiosyncrasies of at least the group-living species as a result of evolutionary disequilibrium (van Schaik and Kappeler 1996). Accordingly, the recent extinction of several large predators and competitors has facilitated a shift of previously nocturnal pair-living species to a gregarious diurnal life style. However, this hypothesis does not account for solitary species, which leaves the possibility that this unusual combination of traits characterised the first colonists, and that it has remained resilient to evolutionary modification for more than 50 million years.

taxa such as equids and hyenas (Linklater 2000, Watts and Holekamp 2007), therefore provides one of the clearest examples of the approach of behavioural ecology to explain behavioural variation as an adaptation to ecological factors (see also Emlen and Oring 1977, Davies 1985, Ims 1988). However, there is ample evidence that different aspects of primate social behaviour are also influenced by phylogenetic constraints (Thierry 2008). For example, variation in the number and developmental state of infants at birth across species predicts whether mothers cache or carry their infants (Kappeler 1998), which in turn is a good predictor of permanent male-female association, because males can protect only infants from infanticide that are associated with their mother (van Schaik and Kappeler 1997). Other aspects of reproductive physiology can also have far-reaching consequences for behaviour. For example, only tamarins and marmosets (Callitrichidae) among primates experience a post-partum oestrus, which annihilates the risk of infanticide by strange males and all associated adaptations and counter-adaptations by both sexes (van Schaik 2000). In other cases, phylogenetic signals are obvious, but their precise bases remain obscure. Gibbons and siamangs (Hylobatidae) are some of our

666

Peter M. Kappeler and Cornelia Kraus

closest relatives, yet their social organisation is highly invariant. All 12 species from 4 genera, with a wide geographical distribution throughout south-east Asia, are pair-living, territorial and duetting (Leighton 1987). Similarly, in a large group of Old World monkeys (Cercopithecoidea) that includes macaques and baboons, aspects of the social system, including group composition, dispersal pattern, female and inter-sexual relationships and mating patterns are strongly conserved, even though these species occupy a range of very different habitats across all of Africa and Southeast Asia, that can only be ‘explained’ by their phylogenetic affiliation (Di Fiore and Rendall 1994). The ‘lemur syndrome’ provides yet another example of phylogenetic clustering of multiple traits in a speciose primate radiation (Box 21.1). Phylogenetic effects on behaviour were also found among the members of a single genus. Phylogenetic analyses of social behaviour traits among macaques (Macaca) revealed that some important features of macaque social organisation correlate with phylogeny, including patterns of male dispersion, female dominance rank acquisition, relative dominance ranks between adult females and males and patterns of reconciliation (Thierry et al. 2000, 2008). A similar study of 12 species and subspecies of Eulemur, which inhabit a variety of habitat types across Madagascar, also found that social organisation showed no relation with habitat, but was correlated with phylogenetic distance among populations (Ossi and Kamilar 2006). Limited experimental studies also revealed a lack in flexible behavioural responses following manipulations of food distribution and abundance within species (Gore 1993), whereas some behavioural flexibility was reported under naturally variable conditions or following cross-fostering of juveniles with another species (de Waal and Johanowicz 1993, Sinha et al. 2005, see below). Thus, even primates with a high potential for sociocognitive flexibility appear to be heavily constrained in core aspects of their social behaviour by factors that arise simply from their affiliation to a particular lineage. Future studies in other taxa will therefore likely find increasing evidence for such effects (e.g., Prum 1994, Sih et al. 2000).

21.3 Behavioural variability below the species level Species are the fundamental category in most aspects of biology. How even closely related species differ in aspects of their behaviour has been amply described since the earliest days of ethology. In fact, some species are morphologically so similar that they can only be reliably distinguished on the basis of their behaviour (Bickford et al. 2007). Species differences

Levels and mechanisms of behavioural variability

667

in behaviour are therefore the default and outside the focus of this review. Because patterns, functions and mechanisms of behavioural variability among populations, among individuals and within individuals over time have been discussed at length in previous chapters (by van Schaik, Taborsky and Brockmann, Bergmüller, Goymann and Hofer, Sachser and Kaiser), we only provide a synoptic summary of these components of variation below the species level. 21.3.1 Variability among populations Populations of a given species often vary substantially with regard to the behavioural solutions their individuals adopt to solve a common problem such as predator avoidance. The proximate mechanisms behind these population-level differences might vary widely, however, from genetic differentiation to social learning. Barrett (2009) recently provided an instructive example of the effects of historical, ecological, social and cognitive contingencies on the social behaviour of local populations of one geographically widespread taxon: the African baboons (Papio ssp.) A conclusion of her and other similar studies (Chapman and Rothman 2009) is that many aspects of the social system reflect a balance between adaptations to the present-day environment and phylogenetic inertia, with interesting, albeit still little-studied variation among taxa in which factors they exhibit how much variability. A comparative study of several European populations of Bechstein’s bats (Myotis bechsteinii) also revealed that the cornerstones of their social system are invariant across populations, independent of their recent history or location, despite local variation in genetic diversity (Kerth et al. 2008). Additional comparative studies of members of the same interbreeding population inhabiting different environments (e.g. striped mice, Rhabdomys pumilo: Schradin and Pillay 2005) or of distantly related taxa exhibiting similar behaviour (e.g. chimpanzees, Pan troglodytes and spider monkeys, Ateles spp.: Chapman et al. 1995) should improve our understanding of the role and importance of ecological pressures in shaping behaviour. Below, we complement the review of van Schaik (this volume) on nongenetic mechanisms by focusing on genetically-based behavioural variability, i.e. local adaptations (e.g. Smith et al. 1997, Foster and Endler 1999, Weitere et al. 2004). If local conditions vary predictably and are reasonably stable over time, we would expect evolution to adjust behavioural traits to the local costbenefit functions. Such local adaptations can then seed speciation over evolutionary time (e.g. Plath and Strecker 2008). In most cases of inter-

668

Peter M. Kappeler and Cornelia Kraus

population variability in behavioural phenotypes it is not known whether genetic divergence has already occurred. In some examples it seems highly unlikely that differences between populations are genetically based, e.g. the various tool use traditions in different primate populations (Whiten and van Schaik 2007). In other cases it is not quite so clear. For example, the different migratory strategies of facultatively anadromous ninespine sticklebacks (Gasterosteus aculeatus) across river habitats (Arai and Goto 2005) and the differences in parental care behaviour of smallmouth bass (Micropterus dolomieu) from different lakes (Gravel and Cooke 2009) might be due to genetic divergence. Common garden experiments – preferably over several generations in several environments – can shed light on the proximate basis of behavioural variation. By transplanting organisms from one environment to another, it is possible to detect the genetic component of population differences. This approach revealed that inter-population differences in behavioural traits, such as male song in field crickets (Teleogryllus oceanicus, Simmons 2004), male mating tactics in guppies (Poecilia reticulata, Kolluru et al. 2007) and hoarding behaviour in degus (Octodon degus, Quispe et al. 2009) are likely of a genetic origin. Extra-genetic inheritance, where epigenetic effects are transmitted over generations (see Sect. 21.3.3) might theoretically mimic genetically based phenotype differences, however it is still far from clear how widespread and stable such effects are over generations and environments. Genetic divergence is necessary, but not sufficient, to infer that behavioural variability among populations is indeed due to local adaptation. Alternatively, genetic drift might have assigned trait values randomly to populations. A first line of circumstantial evidence comes from predictions based on prior knowledge of the functional consequences of the focal trait – if these predictions are met, selection is supported as the driving force behind population divergence (e.g. van Buskirk and Arioli 2005). For example, predation risk and a latitudinal gradient co-vary with activity levels of anuran larvae as predicted based on the local adaptation scenario: those originating from ponds with high predation pressure show decreased activity (van Buskirk and Arioli 2005) and those coming from high latitude habitats maintain comparatively high activity despite the presence of a predator, suggesting that the higher growth rates resulting from high activity levels is necessary due to time constraints in these northern populations (Laurila et al. 2008). Similarly, as predicted, the latency to return to foraging was longer in funnel-web building spiders (Agelenopsis aperta) from populations exposed to high predation pressure (Riechert and Hedrick 1990). In contrast, differences in mating behaviour in guppies from differ-

Levels and mechanisms of behavioural variability

669

ent populations did not show the presumed association with levels of food availability (Kolluru et al. 2007). The next step in demonstrating local adaptation is to link behavioural variation to its fitness consequences. Indeed, the more active tadpoles from high latitudes experienced increased mortality in the presence of a predator, compared to the low-activity ones from southern locations (Laurila et al. 2008). In some cases the lack of inter-population variability clearly does affect fitness components such as survival: mayfly (Callibaetis ferrugineus hageni) larvae developing in ponds harbouring trout did not exhibit reduced activity or increased crypsis. However, whether the observed inflexibility is really maladaptive and due to phylogenetic inertia, or whether gene flow or conflicting selection pressures prevent local adaptation is difficult to establish (Caudill and Peckarsky 2003). If local adaptations consist of a loss of costly behaviour which no longer provides benefits, as, for example, anti-predator behaviour in a predator-free island population of tammar wallabies (Macropus eugenii, Blumstein et al. 2004, Blumstein and Daniel 2005), this might result in an increased risk of extinction (Berger et al. 2001, Gittleman and Gompper 2001). 21.3.2 Behavioural variability among individuals within populations It is intuitive and well established that different local environmental conditions select for local adaptations, thereby generating population level variability. In contrast, individuals from the same population share a common environment and, hence, one might expect selection to stabilise ‘the optimal phenotype’. Yet, behavioural variability among individuals within populations is rampant (Stamps 2003). The distinction between amongindividual variability and the next lower level – within-individual variability – is arguably somewhat fuzzy since the age at which a behavioural phenotype is determined can vary from conception to old age. Here we consider behavioural phenotypes that are consistent (repeatable) within individuals over an appreciable period of their lifetime (cf. Stamps 2003) and are not completely determined by the state of an individual, such as age or sex. Inter-individual differences can be attributed to phenomena ranging from genetically determined discrete behavioural polymorphisms (e.g. Sinervo 2000, see Taborsky and Brockman this volume) over moderately heritable continuously distributed personality traits (Réale et al. 2007, see Bergmüller this volume) to cases of extra-genetic inheritance (Stamps 2003, see below) and socially transmitted traditions (Sanz and Morgan

670

Peter M. Kappeler and Cornelia Kraus

2009, see van Schaik this volume). The evolutionary mechanisms and processes that maintain inter-individual variability include negative frequency-dependent selection, where rare behavioural phenotypes possess a fitness advantage (Dall et al. 2004, Fitzpatrick et al. 2007). Small-scale spatiotemporal heterogeneity in the environment also seems to offer sufficient opportunities for evolution to create and maintain diverse behavioural phenotypes via fluctuating selection (Dingemanse et al. 2004). Additionally, different fitness-equivalent solutions to the same environmental challenges along trade-offs functions can allow behavioural diversity (Arnold 1988). Apart from learned behaviours, consistent variation among individuals is in most cases based on gene x environment interaction (GEI), which can be profitably investigated using a reaction norm approach (Stamps 2003). The relative contribution of GEI to the total phenotypic variability is just beginning to be quantified for simple behaviours (e.g. Sambandan et al. 2008). One of the best-studied examples of GEI-based behavioural variability among individuals is provided by predator-induced phenotypic plasticity in aquatic animals. Many of the common garden experiments mentioned above revealed that behavioural variability implies genetic differentiation among populations, but also environmentally induced phenotypic plasticity, suggesting that population-specific reaction norms have been selected for. Specifically, anuran larvae from different populations differ in baseline activity levels and refuge use, but they additionally respond plastically to the actually perceived predation risk (Relyea 2004, Laurila et al. 2008). Water-borne chemical cues released by the predator are sufficient to alter the developmental trajectories of behaviour and morphology in many aquatic animals (Turner et al. 2000, van Buskirk and Schmidt 2000, Agrawal 2001). The adaptive benefit is obviously an increased chance of survival. As a consequence, feeding rates, and hence growth rates, are often decreased, which in turn can affect reproduction negatively (Sih 1987, Lima 1998). In this case, inducing a particular behavioural phenotype only when the benefits outweigh the costs seems a more flexible solution than adaptation to the average of some environmental factor. Some studies have documented an apparent lack of phenotypic plasticity. For example, mayfly larvae, Callibaetis ferrugineus hageni, developing under high predation pressure did not did not show any evidence of induced antipredator behaviour, suggesting phylogentic inertia or conflicting selection pressures (Caudill and Pecharsky 2003). Beyond GEIs, on the next proximate level, (steroid) hormones are often functionally related to behavioural diversity (Hau 2007). Maternal effects mediated by steroid hormones have been extensively studied in recent years, especially in birds (reviewed in Schawbl et al. 2007, Groothuis and

Levels and mechanisms of behavioural variability

671

Schwabl 2008; see also Kotrschal et al. this volume). Still, sex steroids are certainly not the only possible mediators of inter- and intra-individual trade-offs. Behavioural variability due to parental effects has been documented in a set of fascinating studies on the effects of maternal pup-licking in rats (Curley et al. 2008, Champagne and Curley 2009, see also Sachser and Kaiser this volume). Here, early experience (different intensities of licking modulated by oxytocin) triggers epigenetic effects that shape individual behaviour, including later maternal behaviour by female offspring, so that this variation is transmitted across generations. As yet, however, few studies have focused on the molecular underpinnings of socially flexible phenotypes in vertebrates (Hofmann 2003). A study of the functional genomics of an African cichlid (Astatotilapia burtoni) revealed sufficient variation in transcript levels of candidate genes of the neuroendrocrine pathways that are associated with dominance and reproductive status to prevent a proper classification of dominant and subordinate males (Renn et al. 2008), but more comparative studies of different systems are required for a deeper understanding of the relationship between gene activity and inter-individual behavioural variation. Finally, various mechanisms of individual and social learning to which individuals are exposed can have lasting effects on stable behavioural variation among individuals (Immelmann 1975, Thornton and Hodge 2009, van Schaik this volume). This can lead to stable, semi-closed social units, where groups or other sub-populations within a local population may exhibit behavioural idiosyncrasies. Local innovation and subsequent spread of a novel behaviour pattern among all members of a social unit through social learning may lead to stable differences among neighbouring groups that are transmitted culturally across generations. Some examples include foraging style of Bonnet macaques (Macaca radiata: Sinha 2005), tool use in chimpanzees (Biro et al. 2003), cooperative behaviour in wild chimpanzees (Mitani 2006), and arbitrary experimentally introduced behavioural conventions in captive chimpanzees (Bonnie et al. 2007) and wild common marmosets (Callithrix jacchus: Pesendorfer et al. 2009). Finally, variation in group size within a population may itself provide a mechanism for generating behavioural variation among groups (Liker and Bokony 2009). 21.3.3 Behavioural flexibility within individuals The type of consistent, rather irreversible behavioural variability discussed above can arguably be seen as an individual constraint – the behavioural trajectory determined by the (epi)genetic and/or environmental factors

672

Peter M. Kappeler and Cornelia Kraus

might bind an animal to a sub-optimal behavioural solution in the future (DeWitt et al 1998). One the one hand, such individual stability is a defining feature of behavioural types (Bergmüller this volume) and flexibility itself can carry costs (‘the jack of all trades is master of none’-idea: Dall et al. 2004). On the other hand, it is well established that individuals modify their behavioural phenotypes according to environmental cues, but also as a result of individual development, maturation and learning (Sachser and Kaiser this volume, Taborsky and Brockman this volume, Bshary this volume). The prediction and analysis of the extent of individual flexibility possible despite all the factors generating various constraints described above is what makes the study of animal behaviour so challenging and fascinating. Intra-individual changes in behaviour can be more or less flexible, likely depending on the predictability of the internal and external cues triggering them. Seasonal changes in the environment are usually predictable and behavioural phenotypes are accordingly well adapted to such changes. For example, food-storing behaviour, but also learning ability and the associated brain structures in adult birds are phenotypically plastic between different seasons (Jacobs 1996). Similarly, the annual cycles of migratory birds are defined by fixed stages and transitions between non-breeding, spring migration, breeding, moult and autumn migration, where each stage is distinguished by unique behavioural (and physiological and morphological) characteristics (Wingfield 2005). Seasonal variation in day length also acts as an important ecological constraint on daily time budgets of most temperate species (e.g. Hill et al. 2003). Such adaptations to predictable environmental variables are presumably proximately facilitated by endogenous circadian or circannual clocks (Dawson et al. 2001). Temporal polyethism, where individuals perform different tasks at different times of their life, is a prominent characteristic of many social insects (Wilson 1976) that (also) depends importantly on physiological control; in this case on a developmental process (see Muscedere et al. 2009). If environmental cues are less predictable, context- or state-dependent decisions are necessary. At the inter-individual level, the reproductive tactic employed by male side-blotched lizards (Uta stansburiana) is determined to a large part by male genotype, but it can still be adjusted to some degree to the current situation. Late in the breeding season, as territorial vacancies arise, some formerly non-territorial sneaker males of the yellow morph transform to a blue morph and become territorial (Sinervo 2000). This change seems to be mediated by testosterone and gonadotropin hormones (Mills et al. 2008). The extent of such flexibility can be studied by exposing animals experimentally to novel stimuli or situations. For example, Artic charr (Savelinus alpinus) that were reared under different dietary

Levels and mechanisms of behavioural variability

673

regimes in captivity cannibalised smaller conspecifics after being released into a natural lake, and their behaviour could be explained by relative availability of conspecific vs. alternative prey (Svenning and Borgstrom 2005). Similarly, water fleas (Daphnia pulicaria) fine-tuned their swimming behaviour in response to the simulated presence of predatory fish with different hunting techniques under differing light conditions (Szulkin et al. 2006). Invasive species provide another opportunity to study behavioural changes of the same individuals in different habitats, and there is evidence that Argentine ants (Linepithema humile), for example, modify their aggressive behaviour according to local conditions (Sagata and Lester 2009). An endogenous variable changing inevitably within individuals is age and the associated experience, which tends to affect fitness-relevant behaviour, such as foraging efficiency and adaptive foraging effort, as e.g. demonstrated in European shags (Phalacrocorax aristotelis, Daunt et al. 2007). Furthermore, individuals also adapt their behaviour to changes in the social environment. Banded mongoose pups (Mungos mungo), for example, respond with strategic adjustments in their begging behaviour to variation in their helpers’ condition (Bell 2007). Finally, there is growing evidence that animals respond in a flexible manner to the law of supply and demand in social commodities. Grooming behaviour of female vervet monkeys (Chlorocebus aethiops), for example, was fine-tuned to experimentally induced changes in value of conspecifics that could provide them with access to food (Fruteau et al. 2009, see also Barrett 2009). These arbitrary examples and numerous other studies demonstrate that animals at all levels of social organisation and cognitive sophistication are capable of subtle, local behavioural flexibility. However, the limits of flexibility are clearly evident: individuals rarely switch between more than two alternative systems. 21.3.4 Integration Most behavioural phenotypes we observe emerge from a complex interplay of mechanisms creating varying levels of variability at different levels. Life history theory and socio-ecological theory predict and explain how functional constraints and the spatiotemporal distribution of resources shape the social system of a species. Lott (1984, 1991) offered a comprehensive review of intra-specific variability in vertebrate social systems, the underlying ecological factors and potential mechanisms responsible for variation at that level. In general, the environmental factors shaping the shifts in social systems within species are the same ones implicated in inter-specific variability, e.g. a high predation pressure leading to group-

674

Peter M. Kappeler and Cornelia Kraus

living or increased population density to a more polygynous mating system. These parallels suggest that social flexibility is associated with similar benefits and costs between and within species. As exemplified by the population-specific reaction norms in antipredator response in aquatic animals, population-level variability and interindividual phenotypic plasticity often jointly determine the behavioural phenotype. Similarly, stable inter-individual differences can often be somewhat adjusted according to current conditions, as for example in the male lizard morphs mentioned above. Even individual variability in behavioural flexibility itself is known to vary, i.e. to depend on the early rearing environment in cod (Gadus morhua, Salvanes et al. 2007) and steelhead (Oncorhynchus mykiss, Lee and Berejikan 2008), for example. Thus, behavioural variation among and within individuals appears to be responsive to more fine-grained variation in internal state, condition and experience as evaluated against the current ecological and social environment. The benefits and costs of behavioural variability at the different levels involved in producing the observed phenotype have however rarely been measured (but see Duckworth 2006, Duckworth and Kruuk 2009). The quantification of the relative contribution of the different levels of variability in producing the diversity of behavioural phenotypes continues to present an intriguing puzzle for future research. The realised patterns of variation are shaped by the selective environment against the phylogenetic background of a species. In a rather stable and predictable environment, canalisation into well-adapted genetically determined behavioural phenotypes might represent the most cost-efficient solution, whereas fluctuating, unpredictable environments might require more flexible short-term adjustments, such as ongoing learning processes. The specific evolutionary history, however, will set the stage at which levels variability can arise, and cognitive abilities appear to facilitate surprisingly little behavioural variability, except, perhaps, in humans (Kappeler and Silk 2010).

Acknowledgements We are grateful for very helpful and constructive comments on this chapter from Melanie Dammhahn, Peter Klopfer, Fritz Trillmich and Carel van Schaik. We thank Manfred Eberle, Claudia Fichtel, David Haring and Roland Hilgartner for providing photos.

Levels and mechanisms of behavioural variability

675

GLOSSARY Behavioural syndrome: correlation between rank-order differences in behaviour patterns between individuals of a given population across time and situations (Bell 2007) Behavioural types: sets of individuals within populations exhibiting particular combinations of two or more behavioural traits consistently across contexts and situations (Bell 2007) Canalisation: species-typical regularities in behavioural development as a result of normal experiences (Gottlieb 1991) or strong genetic control (Waddington 1942) Developmental plasticity: individual variation that results from processes during development as a consequence of variation in the environment (Pigliucci 2001, Piersma and Drent 2003) Epigenetic effect: linkage between environmental stimuli during early ontogeny with stable modifications of gene expression patterns and phenotypes that does not affect the underlying genotype Ethogram: inventory and definition of all species-typical behaviour patterns Instinct: a hierarchically organised nervous mechanism, who responds to internal and external alerting, eliciting and orienting impulses with well coordinated motor patterns that have a survival value (Tinbergen 1951); unmodifiable, genetically determined behaviour pattern that is found in every member of the species Parental effects: the phenotype of an organism is not only influenced by his genotype and the environment in which it grows up, but also by the phenotype of its mother/father Phenotypic flexibility: reversible intra-individual phenotypic transformations (Piersma and Drent 2003) Phenotypic plasticity: a change in the expressed phenotype of a genotype as a function of the environment (Pigliucci 2001, Scheiner 1993) Polymorphism: multiple alleles of a gene produce variable genotypes Reaction norm: the array of phenotypes developed by a given genotype over an array of environments (Schmalhausen 1949, Stearns 1992) Strategy: genetically-based (behavioural) programme (Gross 1996) Tactic: one phenotype that results from a strategy (Gross 1996)

References Adkins-Regan E (2008) Do hormonal control systems produce evolutionary inertia? Philos Trans R Soc Lond B 363:1599-1609 Agrawal AA (2001) Phenotypic plasticity in the interactions and evolution of species. Science 294:321-326 Antonovics J, van Tienderen PH (1991) Ontoecogenophyloconstraints? The chaos of constraint terminology. Trends Ecol Evol 6:166-167

676

Peter M. Kappeler and Cornelia Kraus

Arai T, Goto A (2005) Flexible life history strategies of ninespine sticklebacks, genus Pungitius. Environ Biol Fishes 74:43-50 Arnold SJ (1988). Behavior, energy and fitness. Am Zool 28:815-827 Barrett L (2009) A guide to practical babooning: historical, social, and cognitive contingency. Evol Anthropol 18:91-102 Beck CW (1998) Mode of fertilization and parental care in anurans. Anim Behav 55:439-449 Bell AM (2007) Future directions in behavioural syndromes research. Proc R Soc Lond B 274:755-761 Bell MBV (2008) Strategic adjustment of begging effort by banded mongoose pups. Proc R Soc Lond B 275:1313-1319 Berger J, Swenson JE, Persson I-L (2001) Recolonizing carnivores and naive prey: conservation lessons from Pleistocene extinctions. Science 291:10361039 Bickford D, Lohman DJ, Sodhi NS, Ng PKL, Meier R, Winker K, Ingram KK, Das I (2007) Cryptic species as a window on diversity and conservation. Trends Ecol Evol 22:148-155 Biro D, Inoue-Nakamura N, Tonooka R, Yamakoshi G, Sousa C, Matsuzawa T (2003) Cultural innovation and transmission of tool use in wild chimpanzees: evidence from field experiments. Anim Cogn 6:213-223 Blackledge TA, Scharff N, Coddington JA, Szüts T, Wenzel JW, Hayashi CY, Agnarsson I (2009) Reconstructing web evolution and spider diversification in the molecular era. Proc Natl Acad Sci USA 106:5229-5234 Blomberg SP, Garland T Jr, Ives AR (2003) Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57:717-745 Blumstein DT, Daniel JC (2005) The loss of anti-predator behaviour following isolation on islands. Proc R Soc Lond B 272:1663-1668 Blumstein DT, Daniel JC, Springett BP (2004) A test of the multi-predator hypothesis: rapid loss of antipredator behavior after 130 years of isolation. Ethology 110:919-934 Bonnie KE, Horner V, Whiten A, de Waal FBM (2007) Spread of arbitrary conventions among chimpanzees: a controlled experiment. Proc R Soc Lond B 274:367-372 Brashares JS, Garland T Jr, Arcese P (2000) Phylogenetic analysis of coadaptation in behavior, diet, and body size in the African antelope. Behav Ecol 11:452463 Brooks DR, McLennan DA (1991) Phylogeny, Ecology, and Behavior: A Research Program in Comparative Biology. University of Chicago Press, Chicago Case TJ (1978) Endothermy and parental care in the terrestrial vertebrates. Am Nat 112:861-874 Caudill CC, Peckarsky BL (2003) Lack of appropriate behavioral or developmental responses by mayfly larvae to trout predators. Ecology 84:2133-2144 Champagne FA, Curley JP (2009) Epigenetic mechanisms mediating the longterm effects of maternal care on development. Neurosci Biobehav Rev 33:593-600

Levels and mechanisms of behavioural variability

677

Chapman CA, Rothman J (2009) Within-species differences in primate social structure: evolution of plasticity and phylogenetic constraints. Primates 50:1222 Chapman CA, Wrangham RW, Chapman LJ (1995) Ecological constraints on group size: an analysis of spider monkey and chimpanzee subgroups. Behav Ecol Sociobiol 36:59-70 Cheney DL, Seyfarth RM, Smuts BB (1986) Social relationships and social cognition in non-human primates. Science 234:1361-1366 Clauset A, Erwin DH (2008) The evolution and distribution of species body size. Science 321:399-401 Connor RC, Mann J, Tyack PL, Whitehead H (1998) Social evolution in toothed whales. Trends Ecol Evol 13:228-232 Crook JH (1964) The evolution of social organisation and visual communication in the weaver birds (Ploceinae). Behaviour (suppl) 10:1-178 Crook JH, Gartlan JS (1966) Evolution of primate societies. Nature 210:12001203 Curley JP, Champagne FA, Bateson P, Keverne EB (2008) Transgenerational effects of impaired maternal care on behaviour of offspring and grandoffspring. Anim Behav 75:1551-1561 Dall SRX, Houston AI, McNamara JM (2004) The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecol Lett 7:734-739 Dammhahn M, Kappeler PM (2009) Females go where the food is: does the socioecological model explain variation in social organisation of solitary foragers? Behav Ecol Sociobiol 63:939-952 Daunt F, Wanless S, Harris MP, Money L, Monaghan P (2007) Older and wiser: improvements in breeding success are linked to better foraging performance in European shags. Funct Ecol 21:561-567 Davies NB (1985) Cooperation and conflict among dunnocks, Prunella modularis, in a variable mating system. Anim Behav 33:628-648 Dawson A, King VM, Bentley GE, Ball GF (2001) Photoperiodic control of seasonality in birds. J Biol Rhythms 16:365-380 de Waal FBM, Johanowicz DL (1993) Modification of reconciliation behavior through social experience: an experiment with two macaque species. Child Dev 64:897-908 Dewar RE, Richard AF (2007) Evolution in the hypervariable environment of Madagascar. Proc Natl Acad Sci USA 104:13723-13727 DeWitt TJ, Sih A, Wilson DS (1998) Costs and limits of phenotypic plasticity. Trends Ecol Evol 13:77-81 Di Fiore A, Rendall D (1994) Evolution of social organization: a reappraisal for primates by using phylogenetic methods. Proc Natl Acad Sci USA 91:99419945 Dingemanse NJ, Réale D (2005) Natural selection and animal personality. Behaviour 142:1165-1190

678

Peter M. Kappeler and Cornelia Kraus

Dingemanse NJ, Both C, Drent PJ, van Oers K, van Noordwijk AJ (2002) Repeatability and heritability of exploratory behaviour in great tits from the wild. Anim Behav 64:929-938 Dingemanse NJ, Both C, Drent PJ, Tinbergen JM (2004) Fitness consequences of avian personalities in a fluctuating environment. Proc R Soc Lond B 271:847852 Duckworth RA (2006) Behavioral correlations across breeding contexts provide a mechanism for a cost of aggression Behav Ecol 17:1011-1019. Duckworth RA (2009) The role of behavior in evolution: a search for mechanism. Evol Ecol 23:513-531 Duckworth RA, Kruuk LEB (2009) Evolution of genetic intergration between dispersal and colonization ability in a bird. Evolution 63:968-977 Emery NJ, Clayton NS (2004) The mentality of crows: convergent evolution of intelligence in corvids and apes. Science 306:1903-1907 Emlen ST, Oring LW (1977) Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223 Fitzpatrick MJ, Feder E, Rowe L, Sokolowski MB (2007) Maintaining a behaviour polymorphism by frequency-dependent selection on a single gene. Nature 447:210-212 Foster SA, Endler JA (1999) Geographic Variation in Behavior: Perspectives on Evolutionary Mechanisms. Oxford University Press, New York Franck D (1997) Verhaltensbiologie. Georg Thieme, Stuttgart Fruteau C, Voelkl B, van Damme E, Noë R (2009) Supply and demand determine the market value of food providers in wild vervet monkeys. Proc Natl Acad Sci USA 106:12007-12012 Gittleman JL (1986) Carnivore life history patterns: allometric, phylogenetic and ecological associations. Am Nat 127:744-771 Gittleman JL, Gompper ME (2001) Ecology and evolution – the risk of extinction – what you don’t know will hurt you. Science 291:997-999 Goodwin NB, Balshine-Earn S, Reynolds JD (1998) Evolutionary transitions in parental care in cichlid fish. Proc R Soc Lond B 265:2265-2272 Gore MA (1993) Effects of food distribution on foraging competition in rhesus monkeys, Macaca mulatta, and hamadryas baboons, Papio hamadryas. Anim Behav 45:773-786 Gottlieb G (1991) Experiential canalization of behavioral development: theory. Dev Psychol 27:4-13 Gould SJ, Lewontin RC (1979) The spandrels of San Marco and the Panglossian Paradigm: a critique of the adaptationist programme. Proc R Soc Lond B 205:581-598 Gravel M-A, Cooke SJ (2009) Influence of inter-lake variation in natural nest predation pressure on the parental care behaviour of smallmouth bass (Micropterus dolomieu). Ethology 115:608-616 Groothuis TGG, Schwabl H (2008) Hormone-mediated maternal effects in birds: mechanisms matter but what do we know of them? Philos Trans R Soc Lond B 363:1647-1661

Levels and mechanisms of behavioural variability

679

Gross MR (1996) Alternative reproductive strategies and tactics: diversity within sexes. Trends Ecol Evol 11:92-98 Gross MR, Shine R (1981) Parental care and mode of fertilization in ectothermic vertebrates. Evolution 35:775-793 Harvey PH, Pagel MD (1991) The Comparative Method in Evolutionary Biology. Oxford University Press, Oxford Hau M (2007) Regulation of male traits by testosterone: implications for the evolution of vertebrate life histories. BioEssays 29:133-144 Hill RA, Barrett L, Gaynor D, Weingrill T, Dixon P, Payne H, Henzi SP (2003) Day length, latitude and behavioural (in)flexibility in baboons (Papio cynocephalus ursinus). Behav Ecol Sociobiol 53:278-286 Hofmann HA (2003) Functional genomics of neural and behavioral plasticity. J Neurobiol 54:272-282 Huey RB, Hertz PE, Sinervo B (2003) Behavioral drive versus behavioral inertia in evolution: a null model approach. Am Nat 161:357-366 Immelmann K (1975) Ecological significance of imprinting and early learning. Annu Rev Ecol Syst 6:15-37 Ims RA (1988) Spatial clumping of sexually receptive females induces space sharing among male voles. Nature 335:541-543 Jacobs LF (1996) The economy of winter: phenotypic plasticity in behavior and brain structure. Biol Bull 191:92-100 Jarman PJ (1974) The social organisation of antelope in relation to their ecology. Behaviour 48:215-267 Jeschke JM, Kokko H (2009) The roles of body size and phylogeny in fast and slow life histories. Evol Ecol 23:867-878 Jolly A (1984) The puzzle of female feeding priority. In: Small MF (ed) Female Primates: Studies by Woman Primatologists. A.R. Liss, New York, pp 197215 Kappeler PM (1990) The evolution of sexual size dimorphism in prosimian primates. Am J Primatol 21:201-214 Kappeler PM (1993) Female dominance in primates and other mammals. In: Bateson PPG, Klopfer PH, Thompson NS (eds) Perspectives in Ethology. Vol 10: Behaviour and Evolution. Plenum Press, New York, pp 143-158 Kappeler PM (1998) Nests, tree holes, and the evolution of primate life histories. Am J Primatol 46:7-33 Kappeler PM, Port M (2008) Mutual tolerance or reproductive competition? Patterns of reproductive skew among male redfronted lemurs (Eulemur fulvus rufus). Behav Ecol Sociobiol 62:1477-1488 Kappeler PM, Schäffler L (2008) The lemur syndrome unresolved: extreme male reproductive skew in sifakas (Propithecus verreauxi), a sexually monomorphic primate with female dominance. Behav Ecol Sociobiol 62:1007-1015 Kappeler PM, Silk JB (2010) Mind the Gap: Tracing the Origins of Human Universals. Springer, Heidelberg Kearney M, Shine R, Porter WP (2009) The potential for behavioral thermoregulation to buffer ‘cold-blooded’ animals against climate warming. Proc Natl Acad Sci USA 106:3835-3840

680

Peter M. Kappeler and Cornelia Kraus

Kerth G, Petrov B, Conti A, Anastasov D, Weishaar M, Gazaryan S, Jaquiéry J, König B, Perrin N, Bruyndonckx N (2008) Communally breeding Bechstein’s bats have a stable social system that is independent from the postglacial history and location of the populations. Mol Ecol 17:2368-2381 Koenig A, Borries C (2009) The lost dream of ecological determinism: time to say goodbye? … Or a White Queen’s proposal? Evol Anthropol 18:166-174 Kolluru GR, Grether GF, Contreras H (2007) Environmental and genetic influences on mating strategies along a replicated food availability gradient in guppies (Poecilia reticulata). Behav Ecol Sociobiol 61:689-701 Krebs JR, Davies NB (1987) An Introduction to Behavioral Ecology. Blackwell, Oxford Krebs JR, Davies NB (1997) The evolution of behavioural ecology. In: Krebs JR, Davies NB (eds) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, pp 3-12 Krebs JR, Erichsen JT, Webber MI, Charnov EL (1977) Optimal prey selection in the great tit (Parus major). Anim Behav 25:30-38 Laland KN (2004) Social learning strategies. Learn Behav 32:4-14 Laland KN, Janik VM (2006) The animal cultures debate. Trends Ecol Evol 21:542-547 Laurila A, Lindgren B, Laugen AT (2008) Antipredator defenses along a latitudinal gradient in Rana temporaria. Ecology 89:1399-1413 Laviola G, Macrì S, Morley-Fletcher S, Adriani W (2003) Risk-taking behavior in adolescent mice: psychobiological determinants and early epigenetic influence. Neurosci Biobehav Rev 27:19-31 Lee JSF, Berejikian BA (2008) Effects of the rearing environment on average behaviour and behavioural variation in steelhead. J Fish Biol 72:1736-1749 Lefebvre L, Reader SM, Sol D (2004) Brains, innovations and evolution in birds and primates. Brain Behav Evol 63:233-246 Leighton DR (1987) Gibbons: territoriality and monogamy. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker TT (eds) Primate Societies. University of Chicago Press, Chicago, pp 135-145 Liker A, Bókony V (2009) Larger groups are more successful in innovative problem solving in house sparrows. Proc Natl Acad Sci USA 106:7893-7898 Lima SL (1998) Stress and decision making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Adv Stud Behav 27:215-290 Linklater WL (2000) Adaptive explanation in socio-ecology: lessons from the Equidae. Biol Rev 75:1-20 Lorenz K (1941) Vergleichende Bewegungsstudien an Anatiden. J Ornithol 89:194-293 Lynch M (1991) Methods for the analysis of comparative data in evolutionary biology. Evolution 45:1065-1080 MacArthur RH, Pianka ER (1966) On optimal use of a patchy environment. Am Nat 100:603-609

Levels and mechanisms of behavioural variability

681

Maddison WP (1990) A method for testing the correlated evolution of two binary characters: are gains or losses concentrated on certain branches of a phylogenetic tree? Evolution 44:539-557 Magellan K, Magurran AE (2007) Behavioural profiles: individual consistency in male mating behaviour under varying sex ratios. Anim Behav 74:1545-1550 Marino L (2002) Convergence of complex cognitive abilities in cetaceans and primates. Brain Behav Evol 59:21-32 Maynard Smith J (1974) The theory of games and the evolution of animal conflicts. J Theor Biol 47:209-221 Maynard Smith J (1979) Game theory and the evolution of behaviour. Proc R Soc Lond B 205:475-488 Mitani JC (2006) Demographic influences on the behavior of chimpanzees. Primates 47:6-13 Muscedere ML, Willey TA, Traniello JFA (2009) Age and task efficiency in the ant Pheidole dentata: young minor workers are not specialist nurses. Anim Behav 77:911-918 Nussbaum RA (1985) The Evolution of Parental Care in Salamanders (Miscellaneous Publications 169). Museum of Zoology, University of Michigan, Ann Arbor Ossi K, Kamilar JM (2006) Environmental and phylogenetic correlates of Eulemur behavior and ecology (Primates: Lemuridae). Behav Ecol Sociobiol 61:53-64 Owens IPF, Bennett PM (1995) Ancient ecological diversification explains lifehistory variation among living birds. Proc R Soc Lond B 261:227-232 Pérez-Barbería FJ, Shultz S, Dunbar RIM (2007) Evidence for coevolution of sociality and relative brain size in three orders of mammals. Evolution 61:28112821 Pesendorfer MB, Gunhold T, Schiel N, Souto A, Huber L, Range F (2009) The maintenance of traditions in marmosets: individual habit, not social conformity? A field experiment. PLoS ONE 4:e4472, doi:10.1371/journal.pone.0004472 Pigliucci M (2001) Phenotypic Plasticity: Beyond Nature and Nurture. Johns Hopkins University Press, Baltimore/MD Plath M, Strecker U (2008) Behavioral diversification in a young species flock of pupfish (Cyprionodon spp.): shoaling and aggressive behavior. Behav Ecol Sociobiol 62:1727-1737 Prum RO (1994) Phylogenetic analysis of the evolution of alternative social behavior in the Manakins (Aves: Pipridae). Evolution 48:1657-1675 Quispe R, Villavicencio CP, Cortés A, Vásquez RA (2009) Inter-population variation in hoarding behaviour in degus, Octodon degus. Ethology 115:465-474 Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) Integrating animal temperament within ecology and evolution. Biol Rev 82:291-318 Relyea RA (2002) Costs of phenotypic plasticity. Am Nat 159:272-282 Relyea RA (2004) Fine-tuned phenotypes: tadpole plasticity under 16 combinations of predators and competitors. Ecology 85:172-179

682

Peter M. Kappeler and Cornelia Kraus

Renn SCP, Aubin-Horth N, Hofmann HA (2008) Fish and chips: functional genomics of social plasticity in an African cichlid fish. J Exp Biol 211:30413056 Reznick DA, Bryga H, Endler JA (1990) Experimentally induced life-history evolution in a natural population. Nature 346:357-359 Riechert SE, Hedrick AV (1990) Levels of predation and genetically based antipredator behavior in the spider, Agelenopsis aperta. Anim Behav 40:679-687 Roe JH, Georges A (2008) Maintenance of variable responses for coping with wetland drying in freshwater turtles. Ecology 89:485-494 Sagata K, Lester PJ (2009) Behavioural plasticity associated with propagule size, resources, and the invasion success of the Argentine ant (Linepithema humile). J Appl Ecol 46:19-27 Salvanes AGV, Moberg O, Braithwaite VA (2007) Effects of early experience on group behaviour in fish. Anim Behav 74:805-811 Sambandan D, Carbone MA, Anholt RRH, Mackay TFC (2008) Phenotypic plasticity and genotype by environment interaction for olfactory behavior in Drosophila melanogaster. Genetics 179:1079-1088 Sanz CM, Morgan DB (2009) Flexible and persistent tool-using strategies in honey-gathering by wild chimpanzees. Int J Primatol 30:411-427 Scheiner SM (1993) Genetics and evolution of phenotypic plasticity. Annu Rev Ecol Syst 24:35-68 Schlichting CD, Pigliucci M (1995) Gene regulation, quantitative genetics and the evolution of reaction norms. Evol Ecol 9:154-168 Schmalhausen II (1949) Factors of Evolution: The Theory of Stabilizing Selection. University of Chicago Press, Chicago Schradin C, Pillay N (2005) Intraspecific variation in the spatial and social organization of the African striped mouse. J Mammal 86:99-107 Schülke O (2003) To breed or not to breed – food competition and other factors involved in female breeding decisions in the pair-living nocturnal fork-marked lemur (Phaner furcifer). Behav Ecol Sociobiol 55:11-21 Schwabl H, Palacios MG, Martin TE (2007) Selection for rapid embryo development correlates with embryo exposure to maternal androgens among passerine birds. Am Nat 170:196-206 Sih A (1987) Prey refuges and predator-prey stability. Theor Popul Biol 31:1-12 Sih A, Kats LB, Maurer EF (2000) Does phylogenetic inertia explain the evolution of ineffective antipredator behavior in a sunfish-salamander system? Behav Ecol Sociobiol 49:48-56 Sih A, Bell AM, Johnson JC (2004) Behavioral syndromes: an ecological and evolutionary overview. Trends Ecol Evol 19:372-378 Silk JB (2007) Social components of fitness in primate groups. Science 317:13471351 Simmons LW (2004) Genotypic variation in calling song and female preferences of the field cricket Teleogryllus oceanicus. Anim Behav 68:313-322

Levels and mechanisms of behavioural variability

683

Sinervo B, Miles DB, Frankino WA, Klukowski M, DeNardo DF (2000) Testosterone, endurance, and Darwinian fitness: natural and sexual selection on the physiological bases of alternative male behaviors in side-blotched lizards. Horm Behav 38:222-233 Sinha A (2005) Not in their genes: phenotypic flexibility, behavioural traditions and cultural evolution in wild bonnet macaques. J Biosci 30:51-64 Sinha A, Mukhopadhyay K, Datta-Roy A, Ram S (2005) Ecology proposes, behaviour disposes: ecolocical variability in social organization and male behavioural strategies among wild bonnet macaques. Curr Sci 89:1166-1179 Smith BR, Blumstein DT (2008) Fitness consequences of personality: a metaanalysis. Behav Ecol 19:448-455 Smith TB, Wayne RK, Girman DJ, Bruford MW (1997) A role for ecotones in generating rainforest biodiversity. Science 276:1855-1857 Smith FA, Brown JH, Haskell JP, Lyons SK, Alroy J, Charnov EL, Dayan T, Enquist BJ, Morgan Ernest SK, Hadly EA, Jablonski D, Jones KE, Kaufman DM, Marquet PA, Maurer BA, Niklas KJ, Porter WP, Roy K, Tiffney B, Willig MR (2004) Similarity of mammalian body size across the taxonomic hierarchy and across space and time. Am Nat 163:672-691 Sol D, Bacher S, Reader SM, Lefebvre L (2008) Brain size predicts the success of mammal species introduced into novel environments. Am Nat 172:S63-S71 Stamps J (2003) Behavioural processes affecting development: Tinbergen’s fourth question comes of age. Anim Behav 66:1-13 Stearns SC (1992) The Evolution of Life Histories. Oxford University Press, Oxford Sterck EHM, Watts DP, van Schaik CP (1997) The evolution of female social relationships in nonhuman primates. Behav Ecol Sociobiol 41:291-309 Svenning MA, Borgstrøm R (2005) Cannibalism in Arctic charr: do all individuals have the same propensity to be cannibals? J Fish Biol 66:957-965 Szulkin M, Dawidowicz P, Dodson SI (2006) Behavioural uniformity as a response to cues of predation risk. Anim Behav 71:1013-1019 Taborsky M (1998) Sperm competition in fish: ‘bourgeois’ males and parasitic spawning. Trends Ecol Evol 13:222-227 Thierry B (2008) Primate socioecology, the lost dream of ecological determinism. Evol Anthropol 17:93-96 Thierry B, Iwaniuk AN, Pellis SM (2000) The influence of phylogeny on the social behaviour of macaques (Primates: Cercopithecidae, genus Macaca). Ethology 106:713-728 Thierry B, Aureli F, Nunn CL, Petit O, Abegg C, de Waal FBM (2008) A comparative study of conflict resolution in macaques: insights into the nature of trait covariation. Anim Behav 75:847-860 Thornton A, Hodge SJ (2009) The development of foraging microhabitat preferences in meerkats. Behav Ecol 20:103-110 Tinbergen N (1951) The Study of Instinct. Oxford University Press, Oxford Tinbergen N (1963) On aims and methods of ethology. Z Tierpsychol 20:410-433 Trivers RL (1971) The evolution of reciprocal altruism. Q Rev Biol 46:35-47

684

Peter M. Kappeler and Cornelia Kraus

Turner AM, Bernot RJ, Boes CM (2000) Chemical cues modify species interactions: the ecological consequences of predator avoidance by freshwater snails. Oikos 88:148-158 van Buskirk J, Arioli M (2005) Habitat specialization and adaptive phenotypic divergence of anuran populations. J Evol Biol 18:596-608 van Buskirk J, Schmidt BR (2000). Predator-induced phenotypic plasticity in larval newts: trade-offs, selection, and variation in nature. Ecology 81:30093028 van Schaik CP (1989) The ecology of social relationships amongst female primates. In: Standen V, Foley RA (eds) Comparative Socioecology: The Behavioral Ecology of Humans and Other Mammals. Blackwell, Oxford, pp 195218 van Schaik CP (2000) Vulnerability to infanticide by males: patterns among mammals. In: van Schaik CP, Janson CH (eds) Infanticide by Males and Its Implications. Cambridge University Press, Cambridge, pp 61-71 van Schaik CP, Kappeler PM (1996) The social systems of gregarious lemurs: lack of convergence due to evolutionary disequilibrium? Ethology 102:915941 van Schaik CP, Kappeler PM (1997) Infanticide risk and the evolution of malefemale association in primates. Proc R Soc Lond B 264:1687-1694 Waddington CH (1942) Canalization of development and the inheritance of acquired characters. Nature 150:563-565 Watts HE, Holekamp KE (2007) Hyena societies. Curr Biol 17:R657-R660 Weitere M, Tautz D, Neumann D, Steinfartz S (2004) Adaptive divergence vs. environmental plasticity: tracing local genetic adaptation of metamorphosis traits in salamanders. Mol Ecol 13:1665-1677 Wenzel JW (1992) Behavioral homology and phylogeny. Annu Rev Ecol System 23:361-381 West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford University Press, Oxford Whiten A, van Schaik CP (2007) The evolution of animal ‘cultures’ and social intelligence. Philos Trans R Soc Lond B 362:603-620 Wiley RH (1974) Evolution of social organization and life-history patterns among grouse. Q Rev Biol 49:201-227 Wilson EO (1975) Sociobiology: The New Synthesis. Harvard University Press, Cambridge/MA Wilson EO (1976) Behavioral discretization and the number of castes in an ant species. Behav Ecol Sociobiol 1:141-154 Wingfield JC (2005) Flexibility in annual cycles of birds: implications for endocrine control mechanisms. J Ornithol 146:291-304 Winn HE (1958) Comparative ecology and reproductive behavior and ecology of fourteen species of darters (Pisces-Percidae). Ecol Monogr 28:155-191 Wrangham RW (1980) An ecological model of female-bonded primate groups. Behaviour 75:262-300 Wright PC (1999) Lemur traits and Madagascar ecology: coping with an island environment. Ybk Phys Anthropol 42:31-72

Taxonomic Index

A aardwolf (Proteles cristatus), 271 acarid mites, ‘fighter’/‘scrambler’ males, 549 Achatina fulica, 337 Acromyrmex spp. (leafcutter ants), 420 Adélie penguin (Pygoscelis adeliae), 386 Aeolidiella glauca, 339 African black coucals (Centropus grillii), 476 African buffalos (Syncerus caffer), 248 African cichlid (Astatotilapia burtoni), 671 African wild dog (Lycaon pictus), 480 Agelaius phoeniceus (red-winged blackbird), 372, 634 Agelenopsis aperta (funnel-web spider), 601, 668 Alderia willowi, 347 Amblyrhynchos cristatus (Galápagos marine iguana), 39 Amegilla dawsoni (Dawson’s burrowing bee), 556 American robin (Turdus migratorius), 360 amoeba, social, 65, 70 Anas platyrhynchos (mallard), 364 angler fish, 457 Anolis lizards, dewlap colouration, 20 Anser albifrons (white-fronted geese), 136 Anser anser (greylag geese), 121, 126 Anser domesticus (domestic geese), 42 ants, 70, 151, 164–168, 245, 249, 423, 428, 457, 596, 673 aphids, 154, 182–185 Apis cerana cerana, 111

Apis dorsata (giant Asian honeybee), 422, 428 Apis mellifera (honeybees), 87, 162, 165, 245, 249, 414, 422, 428 Apis mellifera capensis (Cape honeybee), 432 Apis mellifera ligustica, 111 Aplysia sea slugs, 336 Araneidae, 444, 449 Araneoidae, monogyny/multiple males, 454 Arctocephalus galapagoensis (Galápagos fur seals), 276 Argentine ants (Linepithema humile), 673 Argiope aurantia, 456 Argiope bruennichi, 450 Argiope keyserlingi, 450 Argiope lobata, 450 Argiope nephila, 456 Arianta arbustorum, 338, 340, 341, 348 army ants (Eciton burchelli), 245, 415, 422, 423, 427 Astatotilapia burtoni (African cichlid), 671 Ateles spp. (spider monkeys), 667 aye-ayes (Daubentonia madagascariensis), 626, 664 B baboons, 38–43, 138, 249, 642, 666, 667 banded mongoose (Mungos mungo), 34, 268, 280, 673 bank swallows (Riparia riparia), 64 Barbary macaques (Macaca sylvanus), 43, 64 685

686

Taxonomic Index

barnacle geese (Branta leucopsis), 64, 133 barnacles, 457 Bechstein’s bats (Myotis bechsteinii), 254, 256, 667 bees, see also, honeybees Belding’s ground squirrels (Spermophilus belding), 34 Biomphalaria glabrata, 334, 342 birds, 126, 359ff black elder (Sambucus nigra), 21 black scavenger fly (Sepsis cynipsea), 307, 310, 315, 320 black widow (Latrodectus hasselti), 443, 448, 449 black-backed jackal (Canis mesomelas), 289 blackcap (Sylvia atricapilla), 6, 21 Blatta germanica (German cockroach), 245 blenniid fishes, brood care, 270 blue monkeys (Cercopithecus mitis), 34 blue tit (Cyanistes caeruleus), 20, 377, 378, 385 bluegill sunfish (Lepomis macrochirus), 380, 538, 539 bluethroat (Luscinia svecica), 363, 364, 369, 378 bola spiders, 4 Bombus terrestris (bumblebee), 169, 418, 421 bonnet macaques (Macaca radiata), 671 Bos indicus, 248 Branta leucopsis (barnacle geese), 64, 133 brown bears (Ursus arctos), 279 brown hyena (Parahyaena brunnea), 484 brown lemurs (Eulemur fulvus fulvus), 251, 664 bumblebees, 169, 418, 421 burying beetles (Necrophorus vespilloides), 64 butterflies, eyespots, 14 C Caenorhabditis elegans, 276 Cakile edentula (sea rockets), 65 California mouse (Peromyscus californicus), 271, 284, 286 Callibaetis ferrugineus hageni (mayfies), 669, 670 Callithrix jacchus, 671 Callitrichidae, 270, 665

Callorhinus ursinus (northern fur seals), 283 Canis lupus (wolves), 138 Canis mesomelas (black-backed jackal), 289 Cape honeybee (Apis mellifera capensis), 432 capuchin monkeys (Cebus capucinus), 34, 248, 251, 631 Cardiocondyla kagutsuchi, 159, 160 Cardiocondyla obscurior, 160 cats (Felis catus), 488 cattle (Bos indicus), 248 Cavia aperea, 278 cavies, 271, 278, 282, 518, 525 Cebus capucinus (capuchin monkeys), 34, 248, 251 Centris pallida, 550 Centropus grillii (African black coucal), 476 ceratopogonid biting midges, 453, 457 Cercopithecoidea, 666 Cercopithecus mitis (blue monkey), 34 chacma baboons (Papio hamadryas ursinus), 38–42, 249, 251 chameleons, 12 Chelidonura hirundinina, 335, 345, 346 Chelidonura sandrana, 337, 346, 348 chickens (Gallus gallus), 33 chimpanzees (Pan troglodytes), 34, 138, 631, 634, 667, 671 clawed frogs (Xenopus leavis), 65 cleaner wrasse, 224, 225 cliff swallows (Petrochelidon pyrrhonota), 45 Clitaetra spp., 447 coal tit (Parus ater), 369 cockroaches (Blatta germanica), 245 cod (Gadus morhua), 674 Columba livia f. domestica (pigeons) 256 common tern (Sterna hirundo), 277 Cornu aspersum (Helix aspersa), 335, 347 Corvus ossifragus (fish crows), 6, 17 coucals, 476 Crocuta crocuta (spotted hyenas), 484, 486 crows, 11, 17, 126 Cryptomys anselli, 276 Cryptotermes secundus, 183 Cryptotermes spp. 189 cuckoo species, 476 Cyanistes caeruleus (blue tit), 20, 377, 378, 385

Taxonomic Index Cycliophora, 457 Cynictis penicillata, (yellow mongoose), 34 D Dama dama (fallow deer), 555 damselflies, 541, 553, 559 Daphnia pulicaria (water fleas), 673 Daubentonia madagascariensis (aye-ayes), 626, 664 Dawson’s burrowing bees (Amegilla dawsoni), 556 Dayak fruit bat (Dyacopterus spadiceus), 268 degus (Octodon degus), 668 Diacamma spp., 166 Dicrurus adsimilis (fork-tailed drongos), 34 Dictyostelium purpureum, 65 digger wasps (Sphex ichneumoneus), 541 Dinoponera quadriceps, 457 Dinoponera spp., 166 Djungarian hamster (Phodopus campbelli), 271, 284 dogs, wild (Lycaon pictus), 289 Dolichovespula saxonica (Saxon wasp), 161 Dolomedes fimbriatus (fishing spider), 601 dolphins, 639 domestic cat (Felis catus), 488 domestic pigeons (Columba livia f. domestica), 256 Dorylus spp., 415, 416 dragonflies, 555 Drosophila spp., 347 drywood termites, 189 Dugesia gonocephala, 338, 345 dumpling squid (Euprymna tasmanica), 610 dung flies, 307, 310, 312 dunnocks (Prunella modularis), 64 dusky-footed woodrats (Neotoma fuscipes), 470 dwarf mongoose (Helogale parvula), 480 Dyacopterus spadiceus (Dayak fruit bat), 268 E eastern kingbird (Tyrannus tyrannus), 392 Echiura, 457 Eciton burchelli, 245, 416, 427

687

Eisenia andrei, 337 Eisenia fetida, 340 Elacatinus spp., 225 elephants (Loxodonta africana), 34, 39, 43, 249, 251, 258, 289 Emberiza schoeniclus (reed bunting), 369 Entelegynae, 449 eukaryotes, 180, 191–194 Eulemur fulvus fulvus (brown lemurs), 251, 664 Euplectes ardens (red-collared widowbird), 5 Euplectes axiliaris (short-tailed widowbird), 19 Euplectes progne (long-tailed widowbird), 301 Euprymna tasmanica (dumpling squid), 610 European shags (Phalacrocorax aristotelis), 673 F fairy-wren, 396 Falco tinnunculus (kestrels), 5, 7 fallow deer (Dama dama), 555 Felis catus (domestic cat), 488 field crickets (Gryllus integer), 555 fire ants (Solenopsis invicta), 70 fish crows (Corvus ossifragus), 6, 17 fishing spiders (Dolomedes fimbriatus), 601 flatworms, 335 fork-tailed drongos (Dicrurus adsimilis), 34 Fulmarus glacialis (northern fulmar), 363 funnel-web spiders, 601, 668 G Gadus morhua (cod), 674 Galápagos fur seals (Arctocephalus galapagoensis), 276 Galápagos marine iguana (Amblyrhynchos cristatus), 39 Galea monasteriensis (yellow-toothed cavy), 271, 518 Gallinula chloropus (moorhens), 64 Gallus gallus, 33, 364 garden warblers (Sylvia borin), 15 Gasterosteus aculeatus (three-spined stickleback), 65, 250, 592, 668 geese (Anser domesticus), 42 Geospiza forti (ground finch), 314

688

Taxonomic Index

gibbons, 665 golden hamsters (Mesocricetus auratus), 69 golden-collared manakin (Manacus vitellinus), 10, 20 Gorilla gorilla beringei (mountain gorillas), 248 great grey shrike (Lanius excubitor), 372, 373 great tit (Parus major), 280, 386, 387, 605 green woodhoopoes (Phoeniculus purpureus), 42 greylag geese (Anser anser), 121, 126–128 ground finch (Geospiza forti), 314 Gryllus integer (field crickets) 555 Guinea baboons (Papio papio), 43 guinea pigs, 282, 512–517, 523 Gunther’s dik-dik (Madoqua guentheri), 39 guppies (Poecilia reticulata), 11, 668

K Kalotermitidae, 189 kestrels (Falco tinnunculus), 5, 7 killer whales (Orcinus orca), 35 L

iguanas, 563 Ischnura ramburi, 541

Labroides bicolor (cleaner wrasse), 225 Labroides dimidiatus (cleaner wrasse), 225 Lacerta vivipara (common lizards), 600 Lake Tanganyika cichlids, 542, 596 Lake Victoria cichlids, 19 Lamprologus ocellatus (snail cichlid), 270 Lanius excubitor (great grey shrike), 372, 373 Lanyu scops owl (Otus elegans botelensis), 361 Latrodectus geometricus (black widow), 448 Latrodectus hasselti (black widow), 448, 449 Latrodectus spp., 443, 448 leafcutter ants, 420 Lemur catta (ring-tailed lemur), 466 lemurs, 251, 466, 664 Lepilemur ruficaudatus, 664 Lepomis macrochirus (bluegill sunfish), 380, 538, 539 Leptothorax spp., 166–168 lesser bulldog bats (Noctilia albiventris), 43 Limia perugiae, 560 Linanthus parryae, 22 Linepithema humile (Argentine ants), 673 lions (Panthera leo), 34, 138 Liostenogaster flavolineata, 183 lizards, common (Lacerta vivipara), 600 long-tailed widowbird (Euplectes progne), 301 Loxodonta africana (elephants), 34, 39, 43, 249, 251, 258, 289 Lumbricus terrestris, 335, 342, 348 Luscinia svecica, (bluethroat), 363 Lycaon pictus (African wild dog), 289, 480 Lymnaea spp., 334, 340, 344

J

M

Japanese macaques (Macaca fuscata), 42 junglefowl (Gallus gallus), 364

Macaca fuscata (Japanese macaques), 42 Macaca mulatta (rhesus macaques), 35, 36, 511, 518–520

H hanuman langurs (Semnopithecus entellus), 285 harvester ants, 419 Helogale parvula (dwarf mongoose), 480 Herennia multipuncta, 447 Heterocephalus glaber (naked mole-rats), 290 Homo sapiens, kin recognition, 67 honeybees (Apis mellifera), 87, 103, 162, 165, 249, 415, 422, 549 house mice (Mus musculus), 64, 65, 612 house sparrow (Passer domesticus), 308, 311, 367 humans, 233, 235, 290 hyenas, 274, 281, 283, 484–488, 517 Hylobatidae, 665 Hymenoptera, 125, 155, 156, 183–189, 413 I

Taxonomic Index Macaca radiata (bonnet macaques), 671 Macaca sylvanus (Barbary macaques), 43, 64 macaques, 35, 36, 42, 64, 511, 518–520, 666, 671 Macropus eugenii (tammar wallabies), 668 Macrostomum lignano, 334, 335 Macrotermes bellicosus, 199 Madoqua guentheri (Gunther’s dik-dik), 39 mallard (Anas platyrhynchos), 364 Malurus coronatus (purple-crowned fairywren), 396 Malurus cyaneus (superb fairy-wren), 363, 385 Malurus splendens (splendid fairy-wren), 395 Manacus candei (white-collared manakin), 20 Manacus vitellinus (golden-collared manakin), 10, 20 manakins, 10, 20 marine isopods, 558 marmosets (Callithrix jacchus), 665, 671 mayflies, 669, 670 meadow voles (Microtus pennsylvanicus), 482, 485 meerkats (Suricata suricatta), 230, 271, 279, 284, 289 Melipona quadrifasciata, 425 Melipona spp., 188–190 Melospiza melodia (song sparrow), 370 melyrid beetles, 22 Meriones unguiculatus (Mongolian gerbil), 271 Mesocricetus auratus (golden hamsters), 69 mice, 506–511, 515, 523 Microcebus berthae, 664 Microcebus murinus (mouse lemurs), 278 Micropterus dolomieu (smallmouth bass), 668 Microtus montanus (montane vole), 482 Microtus ochrogaster (prairie vole), 268, 271, 482 Microtus pennsylvanicus (meadow vole), 482, 485 Microtus pinetorum (pine vole), 271 Mimulus spp. (monkeyflowers), 12 Mnais costalis, 553, 559 mole-rat (Cryptomys anselli) 276 Mongolian gerbil (Meriones unguiculatus), 271

689

mongoose, 34, 268 monkeyflowers (Mimulus spp.), bees/hummingbirds, 12 montane vole (Microtus montanus), 482 moorhens (Gallinula chloropus), 64 mountain gorillas (Gorilla gorilla beringei), 248 mouse lemurs (Microcebus murinus), 278 Muenster yellow-toothed cavy (Galea monasteriensis), 271, 518 Mungos mungo (banded mongoose), 34, 268, 280, 673 Mus musculus (house mice), 64, 65, 612 Myotis bechsteinii (Bechstein’s bats), 254, 256, 667 N naked mole-rats (Heterocephalus glaber), 185, 187, 290 Nannophya pygmaea, 555 Navanax inermis, 344 Necrophorus vespilloides (burying beetles), 64 Neolamprologus pulcher, 542, 596 Neotoma fuscipes (dusky-footed woodrats), 470 Nephilengys spp., 447 Nephilia clavipes, 446 Nephilia edulis, 446, 456 Nephilia fenestrata, 446, 450, 456 Nephilia plumipes, 446, 456 Nephilidae, 446, 449 New Zealand stitchbird (Hihi, Notiomystis cincta), 385 Noctilia albiventris (lesser bulldog bat), 43 northern fulmar (Fulmarus glacialis), 363 northern fur seals (Callorhinus ursinus), 283 Notiomystis cincta (New Zealand stitchbird, hihi), 385 O Octodon degus, 668 Ocyphaps lophotes (crested pigeons), 43 Oncorhynchus mykiss (steelhead), 674 Onthophagus acuminatus, 538, 547, 548 Ophryotrocha diadema, 334, 335, 343 Opilionidae, 443 orang-utans (Pongo abelii/P. pygmaeus), 629, 634, 637 orb-web spiders (Araneoidae), 454

690

Taxonomic Index

Orchis mascula, 18 Orcinus orca (killer whales), 35 Oreochromis mossambicus, 612 Otus elegans botelensis (Lanyu scops owl), 361 P Pan troglodytes (chimpanzee), 34, 138, 667, 671 Panthera leo (lions), 34 paper wasp, 165, 168 Papio hamadryas ursinus (chacma baboons), 38–42, 249, 251 Papio papio (Guinea baboons), 43, 667 Paracerceis sculpta, 558 Parahyaena brunnea (brown hyena), 484 parrots, 126 Parus ater (coal tit), 369 Parus major (great tit), 280, 386, 387, 605 Passer domesticus (house sparrow), 308, 367 Pavo cristatus (peafowl), 65 peafowl (Pavo cristatus), 65 Pelvicachromis taeniatus, 65 Perissodus microlepis, 554 Peromyscus californicus (California mouse), 271, 284, 286 Peromyscus leucopus (white-footed mice), 555 Petrochelidon pyrrhonota (cliff swallows), 45 Phalacrocorax aristotelis (European shags), 673 Philomachus pugnax, 553 Phodopus campbelli (Djungarian hamster), 271, 284 Phoeniculus purpureus (green woodhoopoes), 42 Physa acuta, 334, 340–342, 344 Physa gyrina, 341 Phytoseiulus persimilis, 64 pied babblers (Turdoides bicolour), 34 pigeons (Columba livia f. domestica), 256 pigs, 280, 281 pine vole (Microtus pinetorum), 271 plains spadefoot toad (Spea bombifrons), 73 Platythyrea punctata, 169 poeciliid fishes, 560 Pogonomyrmex rugosus (harvester ants), 419 poison frogs, 16

Pongo abelii/Pongo pygmaeus (orangutans), 629, 634 prairie vole (Microtus ochrogaster), 268, 271, 482 Presbytis thomasis (Thomas langurs), 34 primates, 42, 258, 270, 665 Pristomyrmex punctatus, 165 Propithecus verreauxi (Verreaux’s sifakas), 43, 251, 664 Proteles cristatus (aardwolf), 271 Prunella modularis (dunnocks), 64 Pseudobiceros bedfordi, 347 Pseudoceros bifurcus, 335, 347 purple-crowned fairy-wren (Malurus coronatus), 396 Pygoscelis adeliae (Adélie penguin), 386 R red fox (Vulpes vulpes), 128 red junglefowl (Gallus gallus) 364 red-collared widowbird (Euplectes ardens), 5 red-winged blackbird (Agelaius phoeniceus), 372, 634 reed bunting (Emberiza schoeniclus), 369, 382 Rhabdomys pumilio (striped mice), 268, 667 rhesus macaques (Macaca mulatta), 35, 36, 511, 518–520 ring-tailed lemur (Lemur catta), 466 Riparia riparia (bank swallows), 64 rough-toothed dolphin (Steno bredanensis), 43 S Sambucus nigra (black elder), 21 Sancassania berlesei, 550, 556 Saxon wasp (Dolichovespula saxonica), 161 scarab beetle (Onthophagus acuminatus), 538, 547, 548, 563 Scathophaga stercoraria (yellow dung flies), 312, 315, 321 Schistocephalus solidus, 338, 341, 342 Schistosoma mansoni, 342 Schistosomatidae, 457 Schmidtea polychroa, 341, 345 sea rockets (Cakile edentula), 65 sea slugs, 335–338

Taxonomic Index Semnopithecus entellus (hanuman langurs), 285 Sepsis cynipsea (black scavenger fly), 307, 310, 315, 320 Serranus tabacarius, 334, 335 short-tailed widowbirds (Euplectes axiliaris), 19 siamangs, 665 side-blotched lizard (Uta stansburiana), 472, 473, 554 slime mold (Dictyostelium), 191 smallmouth bass (Micropterus dolomieu), 668 snail cichlid (Lamprologus ocellatus), 270 Solenopsis invicta (fire ants), 70 solicitation behaviour, 280 song sparrow (Melospiza melodia), 370 Spea bombifrons (plains spadefoot toad), 73 Spermophilus belding (Belding’s ground squirrels), 34 sphecid digger wasps, 538 Sphex ichneumoneus, 541 spider monkeys (Ateles spp.), 667 spiders, 441, 455 splendid fairy-wren (Malurus splendens), 395 spotted hyena (Crocuta crocuta), 484, 486 squirrel monkey, 515 steelhead (Oncorhynchus mykiss), 674 Steno bredanensis (rough-toothed dolphin), 43 Sterna hirundo (common tern), 277 sticklebacks, 20, 65, 250, 592, 601, 611, 668 stingless bee, 166, 169 stitchbird, 385 Streblognathus spp., 166, 169 striped mice (Rhabdomys pumilio), 268, 667 Succinea putris, 338, 340 superb fairy-wren (Malurus cyaneus), 363, 385 Suricata suricatta (meerkats), 230, 271, 284, 289 swallows, 45, 64, 362 swordtail fishes, 19, 553, 559 Sylvia atricapilla (blackcap), 6 Sylvia borin (garden warblers), 15 Symphodus ocellatus (Mediterranean wrasse), 550, 551 Syncerus caffer (African buffalo), 248

691

T Tachycineta bicolor (tree swallow), 362, 363 Taeniopygia guttata (zebra finch), 467 tamarins, 665 tammar wallabies (Macropus eugenii), 668 Teleogryllus oceanicus (field crickets), 668 Temnothorax albipennis, 245, 249, 251, 596 Temnothorax nylanderi, 168 Temnothorax unifasciatus, 164, 167 termites, 154, 180–189, 193 Theridiidae, 448 Thomas langurs (Presbytis thomasis), 34 three-spined stickleback (Gasterosteus aculeatus), 65, 250, 601 thrips, 182, 183, 185 Tidarren cuneolatum, 449 Tiddaren argo, 448 tilapia (Oreochromis mossambicus), 612 titi monkey (Callicebus molloch), 518 tree lizards, 472, 562 tree swallow (Tachycineta bicolor), 362, 363 Turdoides bicolour (pied babblers), 34 Turdus migratorius (American robin), 360 Tursiops truncatus (dolphins) 639 Tyrannus tyrannus (eastern kingbird), 392 U Urosaurus ornatus (tree lizard), 472, 562 Ursus arctos (brown bear), 279 Uta stansburiana (side-blotched lizard), 472, 554 V Verreaux’s sifakas (Propithecus verreauxi), 43, 251, 664 voles, 5, 7, 268, 271, 373, 482, 485, 487 Vulpes vulpes (red fox), 128 W wallabies, 668 wasps, 152, 161, 165, 168, 183–186, 413, 426, 538 sphecid, 538 stenogastrine, 183 water fleas (Daphnia pulicaria), 673

692

Taxonomic Index

white-collared manakin (Manacus candei), 20 white-footed mice (Peromyscus leucopus), 555 white-fronted geese (Anser albifrons), 136 widowbirds, 5, 19, 301 wild dogs (Lycaon pictus), 289 wolves (Canis lupus), 138 woodrats, aggression/testosterone, 470 wrasses, 225, 550, 551, 566 X Xenopus leavis (clawed frogs), 65 Xiphophorus multilineatus, 559

Xiphophorus nigrensis (Panuco swordtail), 553, 559 Xiphophorus spp. (swordtail fishes), 19 Y yeast, 70 yellow dung flies (Scathophaga stercoraria), 312, 315, 321 yellow mongoose (Cynictis penicillata), 34 yellow-toothed cavy (Galea monasteriensis), 271, 518 Z zebra finch (Taeniopygia guttata), 467

Subject Index

A acoustic signals/communication, 30, 31 activational effect, 473 adaptation, 505, 522–528 definition, 61 adaptationist programme, critique (Gould/Lewontin), 657 adaptiveness, horizontal, 629, 635 oblique/vertical, 629, 634 adolescence, 505, 506, 522ff fighting, 524 social interactions, testosterone, 526 stress responsiveness, 522 adoption, 489 adrenal cortex, 123 adrenocorticotropic hormone (ACTH), 489, 490 adult sex ratio (ASR), 269 affordances, 624, 625, 647 aggregation, 242 aggression, deprivation of early maternal care/ physical interactions with peers, 518 early life stress, 518 hormones, 469, 476 male-male, 469, 479 maternal, 288 progesterone, 477 ritualised, 167 routines/adaptiveness, 599 testosterone, woodrats, 470 aggressiveness, 506, 523 behavioural stability, 611 conspecifics, 592

future expectations, 603 monoamine oxidase-A, 506 agonistic interaction, 131, 134, 137 alarm calls, 32–34, 39, 626 allelic compatibility, 379 allelic quality, 379 allocation decisions, 537 alloparenting (alloparental care), 185, 186, 189, 290 alternative reproductive tactics (ARTs), 537ff density dependence, 555 frequency dependence, 552 genetic/environmental, 557 hormones, 562 sequential, 539, 558 altricial young, 272 altruism, 56, 74, 79, 182–190, 202, 203, 214, 215, 219, 231 coercion, 190 enforced, 151, 187 kin selection, 183 weak (by-product mutualism/pseudoreciprocity, n-player games) 231 altruistic punishment, 233 Alzheimer’s disease, 507 Alzheimer-like pathology, 508 amniotic fluid, 287 androgens, 129, 134, 512–517, 523 behaviour, birds, 129 male aggression, 612 receptors, 477 responsiveness, 475, 478 seasonality, 474 androstenedione (AE), 484

693

694

Subject Index

angler fish, monogynous males, 457 animal personality, 587ff animal societies, 246 antagonistic pleiotropy, 547 antennae, 90 antennal boxing, 167 anthocyanins, 6, 17 fungal growth inhibition, 7 anti-predator strategies, 39, 669 anti-recognition hypothesis, 76 anxiety, 507, 509, 520, 522 anxiety-like behaviour, modification, 509 arginine vasopressin, 482–485 arginine vasotocin, 561 armaments/weaponry, 389 armpit effect, 69 associative learning, 93, 110 associative recall, 110, 114 assortative mating, 133, 317, 337, 600 audience, 31–34 effects, 33ff B Bateman gradient, 393–396 Bateman’s principle, 389 Beach, F.A., 468 begging, 35, 44, 167, 279–281, 389, 640, 673 behaviour, altruistic, 56, 74, 79, 182–190, 202, 203, 214, 215, 219, 231 behavioural consistency, 590, 594, 595 behavioural constraints, Bauplan, 657 phylogenetic, 655 behavioural development, 505–512, 521, 527, 528 behavioural ecology, 657 behavioural flexibility, 587 behavioural infantilisation, 514 behavioural phenotype, 129, 130, 669 behavioural polymorphism, 669 behavioural profile, 505ff behavioural ‘spillovers’, 595, 601, 612 behavioural syndromes, 587–590, 611, 658, 675 behavioural tactic, 125 behavioural types, 675 behavioural variability, 655 constraints, 660 benefits, compatibility, 379 direct, 371–375, 383 diversity, 375–376 female, 371–379, 397–398 fertility, 373–374

genetic, 379 good genes, 379 indirect, 367, 369–382, 396 insurance, 373–375 male, 383, 386–387, 396–397 material, 372–373 transfer, 417 Bernstein, I.S., 469 Berthold, A.A., 468 best of a bad job tactic, 550 bet-hedging strategy, 348 between-species cooperation, 191 bias, prestige/model-based/content, 631 transmission, 629 bigyny, spiders, 453 biochemical switches, 561 BIOL rule, 631 biparental care, 360, 381 birds, 126, 359ff cognitive abilities, 126 monogamy, 138 parental care, 361 birthdate effect, 552, 560 body size, 307, 663 female fecundity (clutch size), 315 frequency distribution, 312 bonding mechanisms, 122 bourgeois males/tactics, 537ff, 545–567 vs. parasitic males, damselflies, 553, 559 iguanas, 563 brain, 123 nidopallium, 123, 125 prefrontal cortex, 125 size, 125 breeding, cooperative, 363, 542 sequential ARTs, 543 synchrony, 368 breeding density, 368 brood care, 154, 182, 186, 199, 268, 270, 540 blenniid fishes, 270 cooperative, 289 male fish, 67, 270 male mammals, 284 brood parasitism, intraspecific, 541 by-product mutualism, 219–225, 230–234 bystanders, 29–31, 34, 137, 228 C caller, sacrifice, 46 calling behaviour, begging, 44

Subject Index calls, echo-locating, 43 travel, 32, 42 camouflage, 13 disruptive, 13 canalisation, 598, 610, 675 candidate gene, 671 care, biparental, 360, 361, 381 costs, 361 male facultative response, 381 paternal, 271, 365, 381–384, 388, 442, 518 reduction, 365 carotenoids, immune system, birds, 129 castes, 157, 537, 595 determination, 420, 426 polymorphism, 162 categorisation, bees, 94, 95 Challenge Hypothesis (J. Wingfield), 471, 525 cheaters, 179, 182, 190–193, 197, 202 cleaner wrasses, 225 sperm exchange, 346 choice experiments, 321 chromatic/achromatic information, 8 circadian time, 110 cleaner pairs, joint inspections, cheating, 228 cleaners, cooperating, 225, 227 clients, ectoparasites, 224 cleaning mutualisms, 229 coalitions, 258, 485, 488 decision-making processes, 258 coefficient of relatedness (r), 183 coercion, 187–190 altruism, 190 cognition, 61, 87–92, 103, 125, 126 cognitive abilities, behavioural variability, 252, 655, 663 crows, 126 parrots, 126 cognitive empathy, 627 colony efficiency, 188 colour, visual, 8, 12 colour changes, chameleons, 12 colour contrasts, 18 colour perception, 9 colour polymorphism, 22 colour selection, associative learning, 15 colour vision, 89 colouration, disruptive, 13 commitment, 595, 598 benefit of the signaller, 598 common ancestry, 179, 191, 193, 202

695

common garden experiments, behavioural variation, 668 communication, 4, 30, 31, 44, 624 acoustic, 31 cost-benefit, 44 definition, 4, 31 environmental influence, 20 models, 29, 44 network, 29–32, 46 parent-offspring, 279 social groups, 29 synergy effect, 45 vocal, 29 communication problem, 77 compatibility, 376, 379, 419 genetic, 419 compensatory allocation, 377 competition, avoidance, 57 between-group, 232 male-male, 302, 346, 442, 471 parent-offspring, 279 competitors, ARTs, 538 cooperation, 541–543 eavesdropping, 37 complex learning, 88, 113 complex maze, 103, 114 compound eye, 89 compromise, 256 concorde fallacy, 268, 269 conditional reciprocity, 342 conditionality, 216 condition-dependent switch, 546 conditioned stimulus, 93 conditioning, observational, 626 configural learning, 93 conflict, 151, 155, 246 about parental care, 274 bees, 152 between parents, 274 inter-individual, 246 mating roles, 342 parent-offspring, 272ff, 275 prevention/resolution, Hymenoptera, 155 reduction by specialisation, 598 resolution, 151, 182, 193, 195 sexual, 273 siblings, 273 testosterone, 478 wasps, 152 weaning, 282 within families, 273 conflict behaviour, 122 conformity, obligate gregariousness, 631

696

Subject Index

consensus, 243, 247 cost, 243 conservatism, 635 consistency, benefits, 595 consolation, 135 conspicuousness, 10 constrained female hypothesis, 365 constraints, behavioural, 655, 660 developmental, 200, 555, 657, 660 ecological, 72 life-history, 661 male reproductive interests, 138 monogamy, 390 phylogenetic, 655, 658, 660, 665 physiological, 268, 275, 515, 527 sensory, 19 time, 246, 250, 259, 565, 668 contact calls, 42 content bias, 631 context cues, 106 contextual/familiar imitation, 626 cooperation, 39, 56, 70ff, 153, 179ff, 258 between-species, 191 competitors, 541–543 complexity, 190 evolutionary dilemma, 181 fission-fusion societies, 247 genetic relatedness, 190 relatedness, 184 unrelated individuals, 213 cooperative behaviour, 215 cooperative breeding, 363, 373, 395, 542 cooperative care, humans, 289, 290 cooperators, 34 coordination, 29, 31, 41, 46, 133, 242–248 coping style/strategy, 589, 612 copulation, 359ff, 380 forced, 384–388 duration, spiders, 444 frequent, 380, 384, 388 promiscuity/extra-pair, 359ff successful, 386 copying, 250, 623, 626, 632 corticosteroid binding globulins (CBG), 490 corticosteroids, 130, 489, 520, 523 corticotropin releasing hormone (CRH), 489 cortisol stress responses, 516–519, 523 cost-benefit, signalling networks, 44 costs, female, 365, 382, 380–383 male, 384, 387, 386–387, 396–397 opportunity, 366, 371, 384 reproduction, 276

courtship gifts, within-pair/extra-pair copulations, 373 cross-fostering, 634 interspecific, 627 cross-modal associative recall, 110 cross-modal transfer, 99 cubic spline software, 314 cuckoldry, 76, 365, 377, 380, 399, 546 cues, 7, 30, 31, 43 definition, 31 vs. signals, 6 cultural evolution, 623, 644 cultural group selection, 233 culture, 623, 638, 640, 647 adaptation, 644 assessing, 640 capuchin monkeys, 639, 642 cetaceans, 639, 642 chimpanzees, 639, 642 cumulative, 642 definitions, 638 geographic variations, 639, 640 intelligence, 643 social learning, 623 variation, 641 cuticular hydrocarbons, 164 cytoplasmic male sterility (CMS), 194 D dart shooting, garden snail, 347 Darwinian fitness, 467 decision-making process, 43 decision rules, 241 decisions, group size, 259 shared (democratic) vs. unshared (despotic), 241, 244 state-dependent, 672 delayed match-to-sample (DMTS), 88–92, 95, 98–101, 114 symbolic (SDMTS), 99 delayed non-matching task (DNMTS), 98 demasculinisation, 514, 516 density dependence, 332, 368, 515, 537, 555, 600 despotic societies, 258 developmental plasticity, 675 differential conditioning, 95 discriminating admittance, 57 discrimination, biconditional, 93 definition, 60

Subject Index discrimination learning (peak-shift phenomena), 15 disease, veneral/resistance, 421, 422 display behaviour, 389 divergent selection (Lake Victoria cichlids), water clarity, 19 division of labour, 88, 199–203 reproductive, 154 divorce, 365, 375, 380 DNA, microsatellite, 426, 428 DNA fingerprinting, 361 dominance, 155, 163–167, 170, 485 hierarchies, 166, 487 dopamine receptors, partner preference, 485 Drd4 gene, explorative behaviour in great tits, 605 drone congregation, 422, 429 E ear flagging, elephants, 43 eavesdropping, 5, 29–38, 228, 625, 632–638 competitors, 37 definition, 31, 37 escalated conflicts, 35 interceptive, 31, 37 parasites, 37 predators, 37 social, 31, 37, 38 ecdysone, 564 echo-locating calls, 43 ectoparasite removal, 224 edge detection mechanisms, contrasts, 13 effective sex ratio (ESR), 453 egg size, 129 egg trading, 343 emergent properties, 199–203 emotionality/emotional systems, 122 empathy, 627 emulation, 627 energetic investment, 124 enforcement, 187 enhancement (local or stimulus), 625 environmental enrichment, 507 epigenetic inheritance/effect, 505, 521, 527, 668, 675 ESS, 44, 218, 269, 552 Essential Paternal Care Hypothesis, 479 estradiol, 477 ethograms, 657, 675

697

ethology, 657 Euler-Lotka equation, 304 eusociality, 154–157, 182, 183, 189, 243 haplodiploidy, 156 parental manipulation, 189 termites, 154, 157 evolution, 23, 303, 322 major transitions, 153 evolutionary causes, 371–388 evolutionary consequences, 388–397 female perspective, 371–383 frequency, 361–370 male perspective, 383, 384 male/female driven, 384–388, 399 evolutionary history, 467 evolutionary stable strategies (ESS), 44, 218, 269, 552 evolutionary trap, 601 exaptation, 61 experimental testing, 568 exploratory behaviour, fitness consequences, 609 expression of preference, 244 extrachromosomal factors, 559 extra-group paternity, 363 extra-pair behaviour, 361 diversity benefits, 375 evolutionary causes, 371 female benefits, 371 female costs, 380 female quality, 398 fertility benefits, 374 genetic benefits, 378 insurance benefits, 373 male benefits, 383 male costs, 384 male/female driven, 384 material benefits, 372, 383 extra-pair matings, 359ff, 480, 541 male retaliation, 365 reduced male care, 380 extra-pair paternity (EPP), 359ff adult mortality, 367 assortative personality, 608 fitness, 391 hierarchical explanation, 370 island populations, 367 weather conditions, 369 extra-pair young, fitness-relevant traits, 396 quality genomes, 376 eye mimicry, 14 eyespots, butterflies, 14

698

Subject Index

F facilitation (social), 625 familiarity, direct (prior association), 62, 63, 66 indirect (phenotype matching), 67, 69 fangs, abdomen somersault, male sacrifice, 448 fathers, caring, 271 prolactin, 284 feather whistle, 43 fecundity, 304 selection, 320 female choice, 302 allelic quality/compatibility, 376 cryptic, 364, 386–389, 399 female control/dominance, 384–388, 485 female fitness, extra-pair behaviour, 359 female forays, 384–388 female manipulation hypothesis, 385 female mate choice, 541, 555, 565 female perspective, 482 female pursuit, 384–388 female quality, 368, 369, 398 female resistance, 384–388 female-centred clans, 133 fertilisations, alternative tactics, 537 stealing, 68, 539, 542, 557 ‘fighter’/‘scrambler’ males, acarid mites, 549 fighting, adolescence, 524 over females, 540, 550, 556 fission-fusion behaviour/societies, 241ff, 247, 491 fitness, 303, 304, 322, 390, 517, 527, 587–593, 598–609 components, 318, 322, 391–393 direct, 183, 190, 201, 203 estimators, 304 frequencies of competing strategies, 599 inclusive, 79 indirect, 183–186 relative, 307, 390 selection of combinations, 604 variance, 390–393 flexibility, costs, 594 flower constancy, 99 flowers, rewardless, exploitation of pollinators, 5 foraging lanes, 245 foraging skills, learning, 630 foraging success, colony size, 45

fruit colouration, food selection, 6 fruits, ripe, anthocyanins, 17 G game structure, 217, 218, 225 game theory, 213ff, 545, 552, 599, 657 gamergates, 166, 171 games/partner control mechanisms, 229 gene-by-environment interactions, 505, 508, 511, 609, 670 gene expression, 605, 613 genes, compatible, 379 good genes, 379 social behaviour, 429 genetic bet hedging, 375, 376 genetic compatibility, 419 genetic correlations, 547, 565 genetic divergence, 668 genetic diversity, 375, 376, 423 genetic drift, 303 genetic kin recognition, 69 genetic linkage, 601, 607 genetic polyethism, 420 genetic polymorphisms, 505 genetic predispositions, 507 genetic relatedness, 56, 79 genetic variance, 423, 424, 604 genital damage, 441–446 genitalia, enlarged clitoris/pseudo-penis/ virilisation (hyenas), 484, 486 ‘one-shot’, 443, 458 genotype, 508, 509, 526 gloves-off hypothesis, 447 glucocorticoids, 124, 131–136, 490, 516 gluconeogenesis, 490 gonadotropin-releasing hormone (GnRH), 469, 474, 562 good gene model, 20, 367, 377 gossip, helping, 234 green-beard genes, 56, 69, 70, 79, 191, 203 greylag geese, 121, 126 female-bonded clan, 121 Grünau flock, 128 social system, 127 group augmentation, 289 group cohesion, vocal signals, 41 group coordination, vocal signals, 29, 41 group decisions, 46, 241ff, 251–254 conflict, 252 group-level adaptations, 154 group-living, 487 group movements, tarvel calls, 42

Subject Index group selection, 195, 197, 232 intrademic, 198 grouping benefits, 244 guppy pigmentation, sexual vs. natural selection, 11 H half-sib comparisons, 376, 397 Hamilton’s rule, inclusive fitness theory, 56, 183, 184, 199, 202 haplodiploidy, 156, 161, 183 evolution of eusociality, 156, 183 origin of males, 161 harmonic radar, 92, 113 heart rate, 136 helpers, 185, 596 at the nest, 482 reproductive tactic, 543 helping, conditional, 57 fitness, 213–223, 230–234 heritability, 21, 302, 322, 593, 605–607 hermaphroditism, sequential, 332, 568 simultaneous, 329, 331 heterozygosity, reproductive success/ sexual attractiveness, 377, 378 heuristics, 79, 642 hierarchy, 165 hoarding behaviour, degus, 668 honest signal, 163 honeybees 87, 162, 165, 249, 415, 422 age polyethism, 549 choice of nest-site, 251 complex maze, 103 compound eyes, 89 conditioning, 93 dance language, 200 giant Asian, 422, 428 map-based strategies, 113 memory, 88, 99 monogyny, 456 navigation, 92 numerical abilities, 101 optic flow, 112 polyandry, 414 selection in colonies, 430 threshold response, 427 waggle dance, 245 working memory, 99 horizontal inheritance, 193 hormonal control systems, 660 hormonal synchrony, 135 hormone-behaviour relationships, 465ff hormones, 466, 562, 597, 608–613

699

humans, cooperation, 235 cooperative breeding, 290 tragedy of the commons, 233 Huntington’s disease, 507 hydrocarbons, heptacosane/nonacosane, 163, 164 hyenas, 274, 281, 283, 484–488, 517 enlarged clitoris/pseudo-penis/ virilisation, 484 siblicide, 283, 484 hymenopterans, 125, 155, 156, 183–189 conflict prevention/resolution, 155 polyandry, 413 hypothalamic pituitary adrenocortical (HPA) system, 123, 489, 514, 516, 523, 612 I image scoring, 221, 227, 228, 234 imitation, 627 contextual/familiar, 626, 627 over-imitation, 635 production/motor/novel, 627 immune system, 76, 129, 274, 340, 468, 481, 489, 511 immune tolerance, 76 immunocompetence, 134, 303 inbreeding, 185, 340 avoidance, 56, 57, 74, 78, 185 inclusive fitness, 56, 79, 155, 162, 169, 183, 187, 197, 198, 232 increase, intrinsic rate, 304 indeterminate growth, 544, 558, 566 individual differences, 599 individual flexibility, 672 individual optimisation, 277 individual recognition, 66 individual social performance, 121 individual style, 122 individuality, 154, 170 infant killing/infanticide, 138, 509 infanticide, lions, 138 infantilisation, prenatally stressed sons, 512–516 infidelity, 399 information, 30, 31, 628 cascades, 250 manipulation, 4 perishable/non-perishable, 628, 641 pooling, 244 public, 631, 638 transfer, 200, 254

700

Subject Index

inheritance, horizontal, 193 uniparental, 193 vertical, 193 innovation, 629, 641, 644, 647 insect society, superorganism or police state, 169 insects, altruistic worker castes, 157 conflicts/conflict resolution, 151 cooperation, 153 social, 154, 179, 182, 186, 190–195, 199–202, 243, 245 instinctive social behaviour, 123 instincts, 624, 656, 675 null hypothesis, 624 intelligence, culture, 643 inter-population differences, 668 intrademic group selection, 198 investment, 214ff iterated prisoner’s dilemma, 215, 218, 220, 228 K kin, 79 kin aggregations, 57 kin bias, 56ff, 79 ancillary, 57 non-discriminatory, 57 kin detection, 77 kin discrimination, 58, 77, 79 kin recognition, 55ff, 366 cues, 63 definitions, 59, 63 failure, 76 genetic, 69 Homo sapiens, 67 mechanisms, 62 tag-based, 70 kin selection, 57, 79, 155, 156, 183, 414 conflict/cooperation, 155 patrilines, 414 kinship, 79, 156, 253 decision-making processes, 258 inclusive fitness theory, 156 kinship communication, 77 kinship cues, 71 kiss-squeaks, 629 knowledge, 628 Konrad Lorenz Research Station (KLF), 128 L label, 628, 643 lactation, 252, 268, 270, 275, 286, 288, 509

Lamarck, 644 landmark, 94 lateralisation, sensoric, 90 leadership, 244, 248 personalised, 248 learning, 90, 126, 592, 596, 599, 610, 623 birds, 126 constructive, 643 context-dependent, 106, 111 costs, 633 individual, 624 maze experiments in bees, 90 observational, 626, 627 rules, 98 social information, 625 social interaction, 627 songbirds, 467 lek mating, ruffs, 553 lemur syndrome, 664, 666 life-for-life relatedness, 158 life-history, 365 constraints, 661 tactics, 537 trade-off, 600 light gradients, depth-mediated, divergent evolution, 19 longevity, 168 Lorenz, K., 128 luminance, 8 luteinising hormone (LH), 469, 562 M maintenance costs, 595 major evolutionary transitions, 179, 190, 191, 202 major histocompatibility complex (MHC), 340 kin recognition, 68 male brood care, fish, 270 male care, reduced, 380 male coercion, 384–388 male fighting, lethal (ants), 160 male harm, hermaphrodites, 346 male-male competition, 302, 346, 442, 471, 479, 598 male polymorphism, 554, 559 male sacrifice, abdomen somersault, spiders, 448 male song, field crickets, 668 males, haplodiploidy, 161 morph abundance regulated by females, 556 origin, 161

Subject Index ornamented, 21 parasitic, 539 quality, 368, 369, 377, 388, 482 size dimorphism, 553 ‘top-quality’, 376 mammals, parental care, 267 manipulation, 4, 187, 189 manipulative ejaculation, Drosophila spp., 347 masculinisation, 484, 512–516, 665 social environment during pregnancy, 514 masquerade, 13 mate choice, 336 copying, 632 size-dependent, 336 mate competition, local, 159 mate copying, 626 mate guarding, 360, 361, 368–371, 380, 384, 388, 458 mate quality, 371 mate switching, rapid, 363 maternal aggression, 288 maternal behaviour, 506, 520, 521 maternal care, 505, 511, 518–522, 527 induction, 285 maintenance, 286 olfactory cues, 287 maternal effects, 129, 509, 517, 527, 670 maternal effort, 274 maternal inheritance, 194 mates, larger-size, 19 mating, multiple, costs-benefits, 348, 442 suicidal, 453 mating behaviour, 417 mating competition, 321, 442 mating decisions, neuroendocrine regulation, 482 mating effort, 451, 452 mating frequency, 158, 161, 443, 449, 451 mating plugs, 446 mating roles, conflict, 342 mating strategies, alternative, 472 mating success, 304, 305, 391 mating systems, genetic, 360, 471 hormones, 470, 476 social, 360, 471 spiders, 443 mating tactics, proximate control, 481 matriarchs, 252, 485 matrilines, 132 matrilineal groups, support amongst females, 488 maze, 91

701

meme, socially transmitted information, 628 memory, 15 circadian timed episodic-like, 110 episodic-like, 110 long-term, 99 microsatellite DNA analyses, paternity assignment, 426 migratory tradition, 127 milk-letdown, pigs, 280, 281 mimicking/avoiding males, damselflies, 541 mimicry, 537 mirror neurons, 626 mirror-self-recognition, 627 mobbing calls, 33 model-based bias, 631 molecular tools, 361 monandry, 158, 171, 422 monogamy, 133, 138, 157, 470 biparental care, 133 birds, 360 constraints, 390 genetic, 389 social, 359–361 strict, 389, 397 monogyny, 158, 171, 421, 441, 446–459 functional, 171 male-biased sex ratio, 441 multiple males, Araneoidae, 454 sex ratio/evolution, 454 mortality, 662 sex-specific, 360 mother-infant interactions, aggression by bystanders, 35, 36 multicellular organisms, 179, 180, 191, 195 multilevel selection, 156, 179, 196, 198, 201, 202, 232 multiple choice maze, 95, 96 multiple mating, 161, 348, 359, 427, 442, 453, 541 costs-benefits, 348, 359, 442 mutation-selection balance, 604 mutualism, 214, 541 N natural selection, 301, 303, 306, 322 neonatal identify deception, 76 nepotism, 55–58, 73–79 nest inheritance, 185 network analyses, 202 neuromodulators, 561

702

Subject Index

neuropeptides, 561 neuroticism, genetic factors, 506 neurotransmitter, 611 nidopallium (NCl), birds, 123, 125 numerical abilities, bees, 101 nuptial flight, 429 nut cracking, chimpanzees/capuchin monkeys, 631 nutrition, detectability, 16 O observational conditioning, 626 odour signals, 71 offspring, number/size, 662 parental care, 267 quality, extra-pair behaviour, 359 stress responsiveness, maternal behaviour, 520 offspring begging, 279 olfaction, 90 bees, 89 recognition, 71 omatidium, 89 ontogeny, 467, 505, 526 operational sex ratio (OSR), 269, 555 opsin genes, divergent evolution, 19 optic flow, 89 optimal investment, 277 organisational effect, 473 orientation, 94, 95, 108, 109 ornamentation, 389 over-imitation, 635 oxytocin, 124, 139, 286, 287, 465, 482, 671 receptors, 483 P pair-bond formation, hormones, 465 pair-living, 360 pair partners, 133 pairing, forward, 93 parasite resistance, genetic diversity, 420 parasites, eavesdropping, 37 parasitic male, 538, 539, 547–553, 563 parasitic tactic, 537–541, 545, 549, 554, 558–560, 566, 569 parasitism, reproductive, 387 parent-offspring conflict (POC), 272, 275, 279 parentage analysis, 361, 399 DNA fingerprinting, 361

parental care, 267, 271, 475 asymmetry, 268 conflict, 268, 273, 274 definition, 270 mammals, 268 mechanisms, 283 multiple mating, 269 phylogenetic aspects, 270 sex roles, 268 parental effects, 377, 397, 675 parental effort/input/investment, 274 parenting/parents, 129, 360 parthenogenesis, thelytokous, 165, 432 partner choice/preference, 218–222, 227, 483 partner control mechanisms, 218–221, 225–229 allohormones, 347 paternal care, 271, 366, 382 paternal investment, 274, 382, 451, 452 paternity, 453 analysis, 361 assignment, microsatellite DNA analyses, 426 extra-group, 363 mixed, 361 multiple, 366 protection, 388 uncertainty, 442 pathogen resistance, 421 patrilines, 413, 430 hymenopterans, 413 pay-to-stay hypothesis, 221, 278, 596 pedipalp ectomisation, 448 penis, pseudo-penis (hyenas), 486 penis-fencing, flatworms, 335 perception, visual, 10 personality (traits), 129, 138, 257, 398, 588, 669 assortative, 608 constraint/adaptive hypotheses, 594 definitions, 589, 590 development, 609 evolution, 588 fitness, 608 heritability, 604 sexual, 398 phenotype, adaptive, 517 development, 66, 613, 660, 670, 675 epigenetic factors, 201 individual behavioural, 125, 129, 130 matching (indirect familiarity), 62, 67, 69, 71

Subject Index plasticity, 303, 601, 602, 609, 613, 670, 675 variance, 425 phylogenetic aspects, 270 phylogenetic constraints, 655, 658 phylogenetic inertia, 660 pituitary, 123 plant–animal relationships, 6, 17, 21 plants, seed dispersers, 6 plasticity, phenotypic, 601, 602, 609, 613, 670, 675 play, 627, 643 pleiotropy, 600, 601, 606 antagonistic, 601 plumage colours, conspicuousness, conspecifics vs. avian predators, 11 polarised light, 89 policing, 221 ants, 163–170, 187, 188, 235 pollination, isolation, 23 pollinators, exploitation, flowers rewardless flowers, 5 sensory ecology, 12 polyandry, 158, 171, 413ff, 421, 424, 444, 470 ants, 158, 420 classical, 476 colony structure, 424 extreme, 422, 433 sequential, 363 polyethism, 537 temporal, 672 polygyny, 158, 171, 421, 453, 470 ants, 158, 420 polymorphisms, 675 behavioural, 669 heritable threshold responses, 547 preference traits, 540, 541, 556 polyphenism, 537, 552 population selection, laboratory studies, 320 population standard deviation, 307 populations, behavioural variability, 667 inter-population differences, 668 positive patterning, 93 post-copulatory processes, 399 post-mating investment, 481 post-partum oestrus, tamarins/marmosets, 665 posterior forebrain caudolateral nidopallium (NCl), birds, 123 postnatal, 505, 506, 511, 518, 527 pre-mating investment, 481 precocial young, 272

703

predation, 320 predators, camouflage traits, 13 eavesdropping, 37 prefrontal cortex (PC), 123 pregnancy, 509–517, 527 endocrine change, 286 social environment, 513 prenatal phase, influences on behaviour, 506, 511 social influences, 512 prenatal stress, adaptive value, 517 prestige bias, 631 prey eavesdropping, 37 Price equation, theory of selection, 196 Price’s covariance, 156 primates, parental care, 270 phylogenetic constraints, 665 subordinates forming coalitions, 258 tarvel calls, 42 prisoner’s dilemma, iterated, 214–220, 228 proactives vs. reactives, 129 proboscis extension response (PER), 93 producer–scrounger system, 599 production imitation, 627 progesterone, 284, 286, 473, 477 aggression, 477 prolactin, 284, 286, 563 promiscuity, 359, 471, 483 costs, 359 fitness, 379 genetic predisposition, 368 protandry, 566 spiders, 455 proximate mechanism/causation, 121, 466 pseudo-penis, hyenas, 486 pseudoreciprocity, 219–223, 227–234 public good, 215, 229–234 public information, 30 punishment, 163, 170, 187, 195, 220, 221, 227, 233 altruistic, 233 Q quality female, 368, 369, 398 quality genomes, extra-pair young, 376 quality male, 368, 369, 377, 388 quantitative genetics, 302, 322, 605–607 quantitative trait loci (QTL), 605 queen control, 161 queen number, 158 queen pheromones, 163 queen–worker conflict, 155, 158

704

Subject Index

queens, 154 executed by workers (bees), 189 risk of mating flights, 422 queuing, 488, 491, 524 quorum, 244, 250

restaurant hypothesis, 282 retaliation, 131, 272, 365, 368 rumble vocalisations, elephants, 34, 43

R

salmon, bourgeois/parasitic males, 547, 563 sample stimulus, 92 sanctions, 220–222 satellites, parasitic breeding, 551, 555, 558, 568 scale-eating, Lake Tanganyika fishes, 554 scrounging, 630 seasonal pattern, 128 seasonal variation, 672 seed dispersers, antioxidant rewards, 6 segregation distorters, 194 selection, antagonistic, 600–602 artificial, 306 balancing, 322, 599, 600, 609 coefficients, 306, 316, 322 correlated, 314 correlational, 322, 600–604 density-dependent, 599 differential, 322 direct, 314 directional, 311, 322 disruptive, 306, 314, 322, 537–541, 555, 556, 564–566, 599, 600, 609 diversifying, 306, 322 epigamic, 305, 323 episode, 322 fecundity, 323 frequency-dependent, 269, 537, 538, 552–556, 599, 604, 609 gradients, 307, 311, 322, 394 multivariate, 313 univariate, 308 group, 195–198, 232 intensity, 311, 322 inter-sexual, 323 intra-sexual, 323 levels, 154, 196, 198 linear, 323 multilevel, 156, 179, 196–202, 232 multiple receivers, 11 neuronal processes, 12, 15 non-linear, 323 opportunity, 390–394 positive, 323 response, 302 sexual vs. natural, 11 stabilising, 306, 314, 323, 609

radio frequency identification, 92 rank inheritance, 489 rank order, 165 reaction norm, 539, 546, 547, 550, 557, 591, 592, 607, 674, 675 recall, associative, cross-modal, 110 receiver, 30, 31 recipients, desirable/undesirable, 72 reciprocity, 218–228, 234, 258, 541 indirect, 220, 234 tit-for-tat like, 221, 224 recognition, 65, 71, 278 alleles, 62, 69, 79 failure, 76 individual, 66 state-mediated, 64 vs. discrimination, 60 recruiting support, 488 regression relatedness, 171 regurgitation, 167 relatedness, 155–171, 203, 257, 423 asymmetry, 159, 171 genetic diversity, 423 reproduction, mode, 661 reproductive allocation, 169 Reproductive Ground Plan Hypothesis, 430 reproductive load, 185 reproductive parasitism, 387 reproductive polymorphisms, 537ff reproductive skew, 166, 186, 203 reproductive success, 391, 481 actual/realized/true, 391 apparent, 391 genetic, 391 lifetime, 317 reproductive suppression, 491 reproductive tactics, alternative, 537ff exploitive, 540 reproductive value (RV), 269 resource competition, local, 159 resource holding potential, 555, 558 response, evolutionary, 365, 366 facultative, 365, 366, 380, 381 responsiveness, environmental stimuli, 507

S

Subject Index strength, 389, 390, 394 studies, phenomenological, 322 survival, 323 viability, 323 self-fertilisation, 331 self-inspection (self-referent phenotype matching), 65, 69, 70 self-maintenance, maternal care, 275 self-organisation, 199, 201, 244, 245, 259 self-reference, 67 self-restraint, 163 self-sacrifice, 443 self-serving behaviour, 215 selfing, 334, 341 selfish genes, parasitic behaviour, 628 selfish policing, 169 sender, 30, 31 sensitisation, 285 females to newborns, 285 sensoric lateralisation, 90 sensory exploitation, 16 sequence fallacy, sexual cannibalism, 451 sequential tactics, 539–545, 558–563 serotonin, 506–511 serotonin transporter (5-HTT) gene, 507–511 gene, polymorphism, 506 sex allocation, 158, 169, 331, 334, 419, 537 manipulation, 419 ratio, 158, 171 sex investment ratio, 414 sex locus load, 418 sex pheromones, bola spiders, 4 mimicing, 4 sex ratio, 158–160, 390, 454 adult (ASR), 269 effective (ESR), 453 operational (OSR), 269, 555 optimisation, 159 primary, 158 sex roles, 268, 361, 441 egalitarian, 360 reversal, 476 sex steroids, organising effects, 511 sex switching, 228 sex-determining factors/locus, 418, 559 sexes, hermaphroditism, 329 sexual cannibalism, 441, 443, 449–453 pre-copulatory, 601 sexual conflict, 330, 371, 379, 380, 384–388, 399, 458, 541 intralocus, 379

705

sexual dimorphism, 389–391, 396, 441, 444 sex ratio, 454 sexual mimicry, 569 sexual monopolisation, 458 sexual personality, 398 sexual selection, 11, 302, 305, 322, 389–397, 441 Bateman gradient, 393–396 consistency, 598 extra-pair behaviour, 359 variance in mating success, 389 vs. natural selection, 11 shared (democratic) decisions, 241, 244 vs. unshared (despotic) decisions, 241 siblicide, 283, 484 sibling competition/rivalry, 375, 380, 383, 389, 484 signals, 5, 11, 16, 30–32 acoustic, 31 communication, 5 conspicuousness, 10 costly/cheap, 44 definition, 31 mate attraction, 32 territorial defence, 32 variants, 628, 629, 642 simultaneous hermaphroditism, 329 simultaneous tactics, 543, 544, 566 size, effect, 566 size-fecundity, 336 skills, 628, 642 snapshot, 99 sneaker males (sneaky matings), 39, 541, 550, 551 sociability, 139, 600 social amoeba, 65, 70 social behaviour, instinctive, 123 network, 122 social competence, 124 social context, 121, 123, 488 social defeat, 511, 512, 521 social eavesdropping, 31 social efficiency, 124 social environment, 505, 511–514, 522, 527 social evolution theory, 216 social experience, 505, 509, 519–523, 527 social insects, 154, 179, 182, 186, 190– 195, 199–202 castes, 157, 537, 595 monogamy, 157 social instability, 512, 515, 522 life span reduction, 515

706

Subject Index

social isolation, aggression/aberrant behaviour, 518 social knowledge, 38, 39 social learning, 623ff, 647, 671 adaptive, 633, 634 evolution, 636 group size, 636 horizontal, 631, 635 vertical, 630, 634 social life, 203 social organisation, 203, 247 social performance, individual, 121, 124 social prestige, 220, 221, 228 social queues, ranks, 488 social rank, 127, 491 social relationship, 518 social stability, prenatal effect, 512 social status, 488 social stress, 122, 487 heart rate, 136 reduction, 139 status, 491 social support, 130–133, 137–139, 512, 518 monogamous pair-bonds, 138 passive, 138 social systems, birds, 121, 125, 138 cognition-based, 125 socialisation, 129, 506, 518 sociality, 122 syndromes, 184–186, 190 society, 242 socio-ecological model, 663 socio-sexual diencephalic and tegmental network, 123 soldiers, 182, 185, 189, 199, 200 solicitation behaviour, 280 specialisation, 595–598 speciation, 564 disruptive selection, 600 species, close, behavioural differences, 666 spectral sensitivities, 8, 9 speed-accuracy trade-offs, 15 sperm competition, 302, 339, 364, 388, 399, 415 hermaphrodites, 339 sperm injection, traumatic/hypodermic, 347 sperm limitation, extreme polyandry, 416, 424 sperm selection, 419 sperm storage, 363, 388 sperm trading, 344

spermatheca, 161, 166 spiders, copulation duration, 444 genital damage, 441 monogynous mating, 441ff protandry, 455 sexual cannibalism, 441 spiteful behaviour, 70 split sex ratio, 158 SSD, protandry, 455 status-dependent switch, 546 steroid hormone receptor gene, epigenetic modification, 521 sticklebacks, 65, 250, 592, 601, 611, 668 agressiveness, 602, 611 breeding colouration, 20 migratory strategies, 668 stimuli, conditional, 429 conditioned/unconditioned, 93 matching, 92 stolen fertilisations, 68 strategies, insect societies, 125 strategy vs. tactic, 214–218, 224–227, 545, 675 stress, 122, 131, 136, 139, 487, 505, 507, 512–527 axes, 123 coping, 122 prenatal, 512 responsiveness, sex differences, 523 social, 122, 487 stress-related pathologies, subordinate animals, 491 stressors, 137, 487, 511–523 non-social, 512 pregnancy, 511 social interactions, 137 subfamily, 413 submissive behaviours, energetic costs, 597 suicidal mating, 453, 458 superorganism, 153, 169, 170 support, recruitment, 488 surrogate mothers, spotted hyenas, 489 survival, 304, 603 switch, biochemical, 561 condition-/status-dependent, 560 switch points, selection, 546 symbiotic associations, 193 symbolic delayed match-to-sample (SDMTS), 99 symbolic matching, 98 symbols, 628, 629 symmetry, 94

Subject Index sympathico-adrenergic axis (SAA), 123 synergy, 44, 46, 399, 433

707

travel calls, 32, 42 triumph ceremony, greylag geese, 127 twinning, callitrichids, 270

T tactics, best of a bad job, 550 bourgeois, 545 conditional, 545, 549–554 fixed, 544ff, 558 flexible, 544ff, 561–565 parasitic, 538, 545 plastic, 545, 557 switching, 537 vertebrate societies, 125 vs. strategy, 545 tadpoles, colouration, 12 task specialisation, 425, 429 teaching, 245, 626, 627, 636, 642–647 ants, 245 temperament, 589 testis size, 465 testosterone, 134, 467, 471, 476, 481, 484, 512, 516, 523–526, 563, 612 5α-dihydrotestosterone (DHT), 484 extra-pair behaviour (birds), 481 11-keto-testosterone, 468 male mating tactics, 481 risk of injury/mortality, 481 testosterone-co-variation (TC), 134 thelytoky, 165, 432 theory of mind, 627 thermoregulation, pigments, 7 third-party relationships, 489 threshold response/switching, 426, 427, 537, 546–549, 556–565 time constraints, 250 Tinbergen, N., 267, 270, 466, 656 tit-for-tat like reciprocity, 221, 224 tool use, chimpanzees, 671 toxic animals, contrasting colours, 14 trade-offs, 259, 537, 541, 546, 564, 565 trading (conditional reciprocity), 343 traditions, 638, 647, 669 tragedy of the commons, 154, 162, 213, 215, 229–235 trait-group selection, 198 transgenic mice, 507 transmission bias, 629 conformity, 631, 645 content bias, 631 prestige bias, 631, 645 transposons, 194 traumatic sperm injection, 347

U uncertainty/unpredictability, stress responses, 515 UV reflectance, visible cue, 5 V vagino-cervical stimulation (VCS), oxytocin, 287 variance partitioning, see, fitness components vasopressin, 465, 561 viability selection, 320 visual generalisation, 101 mumber-based, 101 visual system, divergent evolution, 19 vitellogenin, 430 vocal communication, 29 vocalisations, travel, 42 W warning colouration, 16 poison frogs, 16 weaning, 274, 282 conflict, 282 weaponry/armaments, 389 weather conditions, extra-pair paternity, 369 Westermarck effect, 66, 67 Wingfield, J., Challenge Hypothesis, 469 worker policing, 90, 162, 164 workers, 154, 182, 185–191, 199, 202 egg-laying, 432 reproduction, 162, 187, 188, 432 sex ratio manipulating, 158 task specialisation, 429 working memory, 91, 99–101 Y Y-maze, 91, 98, 99, 103, 106 Yalow, R. 469 yeast, cooperation, 70 Z z-score, 307